FATE OF MERCURY IN THE ARCTIC Michael Evan ... - COGCI

Fate of Mercury in the Arctic 

FATE OF MERCURY IN THE ARCTIC 

Michael Evan Goodsite 

Ph.D. Thesis 

Submitted to the Department of Chemistry at 

the University of Copenhagen, June 2003

Fate of Mercury in the Arctic 2 

ISBN XX-XXXX-XXX-X


Dansk sammenfatning 

I denne Ph.D. er kviksølvs skæbne i Arktis undersøgt gennem feltmålinger og teoretiske 

beregninger. Ph.D.`en er udført på COGCI Ph.D. skole, Kemisk Institut, Københavns Universitet i 

samarbejde med Danmarks Miljøundersøgelser, DMU, Afdeling for Atmosfærisk Miljø, ATMI, 

med finansieringsmidler fra Forskningsstyrelsen og DMU. 

I arbejdet er der målt kviksølv og kviksølvfluxe over snedækket samt i troposfæren over Arktis. 

Der er desuden målt kviksølv i tørveprofiler fra diverse arktiske områder. 

I atmosfæren over Arktis sker der efter den polare solopgang en oxidation af gasformig 

elementært kviksølv, Hg 0 , til divalent gasformig kviksølv, såkaldt RGM, hvilket let deponeres. Ved 

hjælp af en ”relaxed eddy accumulation” flux-maskine udviklet i dette Ph.D. projekt, er målinger af 

denne deponering for første gang rapporteret. Flux-maskinen er specielt designet til måling af RGM 

under arktiske betingelser. Baseret på disse fluxmålinger samt andre rapporterede felt- og 

laboratoriemålinger er en teoretisk kemisk mekanisme for oxidation af Hg 0 til RGM i troposfæren 

efter polar solopgang rapporteret. 

Med det formål at etablere en tidsserie er der målt kviksølv i tørveprofiler fra arktiske områder. 

Udviklingen af en ny dateringsmetode under denne Ph.D. muliggjorde dette med en meget høj tids- 

opløsning inden for de sidste 50 år. Der er desuden udviklet en metode til analyse af kviksølv i tørv 

samt et boresystem til anvendelse ved frossen tørv. 

Undersøgelserne bekræfter, at kviksølvforekomsten er steget i Arktis siden starten af 

industrialiseringen, og at der har været en faldende tendens de seneste 10 år. Modelberegninger 

foretaget på baggrund af data fra dette Ph.D. projekt tyder på en fordobling af 

kviksølvdeponeringen til de arktiske områder, når forårets oxidation af Hg 0 til RGM med 

efterfølgende deponering tages i betragtning. Yderligere undersøgelser af dette fænomen samt dets 

effekt på det arktiske miljø er derfor nødvendige i fremtiden.


Abstract 

Using a micrometeorological system, relaxed eddy accumulation, with a heated sampling 

system specifically designed for Arctic use, the dry deposition of reactive gaseous mercury (RGM) 

is quantified after polar sunrise, in Barrow, Alaska. KCl coated manually analysed annular denuders 

were used as the accumulators. At 3 m above the snow pack significant RGM fluxes measured 

during March 29 th – April 12 th 2000 were directed toward the snow surface. Overall mean flux was 

found to be - 0.4 ± 0.2 pg m -2 s -1 ; N=9, ± SE, where the negative sign convention denotes 

deposition. Using measured total RGM concentrations; depositional velocities were then calculated 

and found on average to be 1 cm s -1 . 

Examining the experimental data from this field campaign. as well as current published field 

data and reaction kinetics data, a plausible mechanism for the oxidation of gaseous Hg (0) to the 

divalent gaseous form is proposed. The mechanism is consistent with the kinetics, thermodynamics 

and field observations, but the final end products are still not known, due to the measurement 

method in the field, i.e., reactive gaseous mercury is operationally determined and defined. The 

hypothesized mechanism is not the same as the mechanisms otherwise derived, where the depletion 

of atmospheric boundary-layer mercury is said to be due to a reaction between gaseous elemental 

mercury, GEM, and BrO free radicals or mechanisms resulting in the formulation of HgCl2. 

The proposed reaction mechanism is that gaseous elemental mercury, Hg (0) combines with Br 

atoms, called X, coming from the polar sunrise destruction of ozone, in a reversible reaction, 

forming the energised HgBr*. Through a third body reaction, M, where M is N2 or O2, the HgBr 

radical is formed. The HgBr radical can live long enough at the low temperatures of the Arctic to 

combine with O2 forming the HgBrOO peroxy radical or can combine with Br forming HgBr2. It is 

not likely to react with Cl, since this reaction would be endothermic. Similarly, the product cannot 

be Hg2Br2 since this would imply a tri molecular reaction, which is highly unlikely to occur in the


atmosphere given the very low concentrations of elemental mercury, bromine atoms, and HgBr. Nor 

would the product be HgO, since this formation is similarly thermodynamically not favourable. The 

final product is the divalent gaseous mercury family of unknowns, HgXY, proposed here to be 

HgBr2. By modelling the reaction of Hg and Br in the atmosphere, with current reaction constant 

data, and BrO and Br measured concentrations, we find, assuming 20 ppt BrO and 2 ppt Br in the 

atmosphere during a depletion event, that the lifetime of Hg is 4.6 hrs against forming HgBr, The 

lifetime of HgBr is 0.35 hrs. against forming HgBr2; comparing with the lifetime of HgBr of 0.75 

hrs means that 68% of the time HgBr will form HgBr2. Thus the overall lifetime of removal of Hg 

to HgBr2 is 4.6 hrs. / .68 = 6.7 hrs. This is in good agreement with the observed 10 hr lifetime of Hg 

under depletion. 

Mercury concentrations were investigated in peat to determine background levels of mercury in 

the Arctic area. Since the types of wetlands found in the Arctic region are not raised, ombrotrophic, 

rainwater nourished, bogs but rather minerotrophic, groundwater nourished, fens, comparison was 

made with a raised bog from Denmark. To ensure that post depositional processes are comparable, 

especially in the oxidation/reduction transition zone near the surface, a high-time resolution direct 

dating technique, based on accelerator mass spectrometry, AMS, measurements of 14 C levels in the 

profile were made, and correlated to the elevated atmospheric 14 C levels from past nuclear bomb 

testing. 

In southern Greenland, and the Faroe Islands, the chronology of Hg accumulation is similar to 

that of the bog in Denmark, suggesting that in remote areas, Hg is supplied to the wetlands 

primarily via atmospheric deposition, demonstrating that given the proper conditions, minerotrophic 

sediments provide a history of atmospheric deposition as consistent as one provided by a raised bog. 

Hg fluxes in the Greenland core (0.3 to 0.5 µg m -2 yr -1 ) were found in peats dating from AD 

550 to AD 975, compared to the maximum of 164 µg m -2 yr -1 in 1953. Atmospheric Hg


accumulation rates have since declined with the value for 1995, 14 µg m -2 yr -1 comparable to those 

values published in 1995 from the Danish Eulerian Hemispheric Model (Christensen et al., 2002, 

Skov et al., 2003, Appendix C), of 12 µg m -2 yr -1 for southern Greenland. 

In Denmark, the greatest rate of atmospheric Hg accumulation is found in 1953, 184 µg m -2 yr - 

1 , comparable to that of Greenland with the flux going into sharp decline, with an accumulation rate 

for 1994 as 14 µg m -2 yr -1 . This compares well with the modelled rate of 18 µg m -2 yr -1 for all of 

Denmark. On the Faroe Islands, the maximum mercury concentration was 498 ng g -1 . Dated to be 

in 1954 +/-2, with a 210 Pb constant rate of supply dating model, in good agreement with historical 

maximums in 1953 in S. Greenland and Denmark. Depositional maximum was in 1985, 34.3 µg m -2 

yr -1 , with the 1995 value of 18 µg m -2 yr -1 , comparing with Denmark and the modelled value of 

approximately 10 µg m -2 yr -1 . The present day value of approximately 10 µg m -2 yr -1 compares well 

with recently reported measurements of 7 µg m -2 yr -1 (Daugaard, 2003). Long-term accumulation 

rate of mercury in this core was 0.95 ± 0.36 µg m -2 yr -1 for the period of 4200 B.C. to AD 833; n = 

61, in agreement with the Danish and South Greenland peat cores, peat cores from Switzerland and 

what is known about the global mercury cycle, and what would be expected in a remote area. The 

Hg concentrations in the Faroe Islands are higher than those found in cores from other sites, but the 

net Hg accumulation rates are comparable. 

The depositional records in the peat are in good agreement with what would be predicted: 

lower depositional rates at sites further North from European industries; and in agreement with 

previous atmospheric measurements that showed that total gaseous mercury in Europe reached an 

average annual maximum in the late 1980’s and have been falling since 1990 decreasing globally 

by 22% from 1990 - 1994 (Slemr et al., 1995 and references therein). The results are discussed 

with respect to the available literature and their potential implication for our understanding of the 

global mercury cycle.


Preface and Acknowledgements 

This thesis was submitted to the Department of Chemistry at the University of Copenhagen in 

partial fulfillment of the requirements for obtaining the Ph.D. degree through the Copenhagen 

Global Change Initiative, COGCI, Ph.D. school. 

The subject of this thesis is the experimental investigation of the fate of mercury in the Arctic 

by determination of mercury flux and background concentrations through studies in peat, and 

determination of Hg conversion and subsequent deposition of reactive gaseous mercury, RGM, to 

the snow pack. 

The work for this thesis was performed during the period May 2000 through May 2003 with 

Professor Ole John Nielsen from the University of Copenhagen and Dr. Henrik Skov from NERI 

Denmark as supervisors. I performed my work at NERI, and spent 6 months at Oak Ridge National 

Laboratory, under the supervision of Dr. Steven Lindberg; with concurrent stay at the NOAA, 

ATDD, Oak Ridge, under the supervision of Dr.’s Steve Brooks and Tilden Meyers. Work on peat 

cores was performed under the supervision of Professor William Shotyk, University of Heidelberg. 

My Ph.D. work has been supported by the Danish Research Academy, NERI Denmark, and the 

University of Southern Denmark, Department of Chemistry. Research funding for fieldwork was 

primarily through DANCEA. 

Most of the results presented herein have been published elsewhere, or are in the process of 

being published. Papers and Manuscripts are inserted as supplementary material in Appendix C. 

There are a number of people whom I wish to acknowledge. In particular my family and 

extended family, who have allowed me to be absent for over one year, during field and research 

work, my academic supervisors and research collaborators and the Department of Chemistry at the


University of Southern Denmark, who have supported my well being and development by hiring me 

and allowing me to finish my Ph.D. dissertation over the last year. 

Dedicated to the people of the North, who have welcomed me and supported my research. I 

hope that my work will in a small way contribute to efforts to keep your homelands pristine and 

your traditions preserved. Quanaaq!


Table of Contents 

1. Introduction 10 

1.1 The aim of this work 12 

1.2 Summary of results 13 

1.3 Sampling and Arctic fieldwork 17 

2. Overview 18 

2.1 General Mercury 18 

2.2 Arctic Mercury 19 

2.3 Mercury recorded in environmental archives 20 

2.4 Overview of Arctic atmospheric chemistry “The polar sunrise mercury 

depletion event” 22 

3. Experimental Methods, Equipment and Procedures 27 

3.2 Measurement of atmospheric elemental mercury and ozone 30 

3.3 Ozone measurements 30 

3.4 Gaseous elemental mercury measurements 30 

3.5 RGM determination 33 

3.6 Flux measurements of reactive gaseous mercury by relaxed eddy 

accumulation 39 

3.7 Peat analysis 57 

3.8 Age dating of peat profiles 61 

4. Results and Discussion 63 

4.1 Ozone and GEM Measurements 69 

4.2 RGM concentrations and Flux 71 

4.3 High resolution dating of peat archives and mercury in peat 77 

5. Conclusion and future work 117 

Glossary, acronyms and abbreviations 119 

Bibliography 121 

Appendix A List of Papers 128 

Appendix B Field Work: planning and post expedition report format132 

Appendix C Supplementary material 142


1. Introduction 

Elemental mercury is unique when compared to other trace metals found in the atmosphere, in 

that it is approximately 95% in the gaseous elemental form (Slemr et al., 1985, Schroeder and 

Munthe 1998), where other metals such as lead, are primarily associated in the atmosphere as 

aerosols. Its characteristics, such as low aqueous solubility, mean that it is relatively non reactive 

and stable and therefore has a long atmospheric residence time, enabling global transport. Its vapour 

pressure allows it to be deposited and re-emitted, as does bacteriologic conversion and subsequent 

emission. All of these factors contribute to its spread throughout the globe to areas where there are 

no natural or man-made inputs, such as the Arctic. 

In 1998 Schroeder et al., reported on their 1995 discovery of the springtime depletion of 

tropospheric gaseous mercury in the high Canadian Arctic. This perennial phenomenon is since 

dubbed atmospheric mercury depletion episodes, AMDEs (Schroeder et al., 2003), and has since 

been shown to be a polar and sub-polar phenomenon. Apart from Alert, AMDEs have been reported 

at other Arctic locations (Lindberg et al., 2001, Berg et al., 2001, 2003, Skov et al., 2001, Steffen et 

al., 2002) as well as in the Antarctic (Ebinghaus et al., 2002) and sub-arctic (Poissant and Pilote, 

2001). The phenomenon attracted attention and concern in Denmark, since Greenland and the Faroe 

Islands are part of the Danish Kingdom and toxic exposure to mercury from eating a traditional 

marine diet may result. All forms of mercury such as elemental, inorganic and organic, are toxic. 

However once in the ecosystem, elemental mercury is subject to methylation, increased uptake and 

toxicity. 

Methylmercury compounds are bioaccumulated and can present a health hazard to people 

eating from affected food sources. 

Background levels of mercury in the Arctic regions are not well known, since environmental 

archives typically studied are ice cores, and ice is incredibly difficult and slow to obtain reliable Hg


analyses from. In the Arctic, these cores are taken from areas where AMDEs do not occur, typically 

at altitudes greater than 2 km above sea level on the inland ice. AMDEs are coastal phenomenon, so 

it was necessary to find a suitable environmental archive to investigate background and recent 

levels of mercury in the Arctic. For this work, peat was chosen, since there are many unanswered 

issues regarding Hg diagenesis in aquatic environments and sedimentation rates in Arctic limnic 

systems are very slow. 

In general, the state of Hg research is relatively immature, compared with other trace metals, 

especially lead. The global and Arctic Hg cycle is poorly understood, and source receptor 

relationships need to be well formulated in order to apply the science to real world problems. The 

most actual present day large-scale mercury problem is faced in the Arctic, i.e., where ecosystems 

and local communities are fragile due to high exposure. Mercury’s toxicity has already had 

measurable effects in the Faroe Islands (Grandjean et al., 1992, 1995, 1997; Sørensen et al., 1999; 

Steuerwald et al., 2000). In 1986 in north Greenland more than 80% of the population exceeded the 

benchmark level of concern for the United States, 50 µg Hg l -1 blood and 16% exceeded the World 

Health Organization, WHO, minimum toxic blood concentration in non-pregnant adults, 200 µg Hg 

l -1 blood (Hansen and Pedersen, 1986). Weihe et al., 2002 suggest, but do not conclude, that 

observed neurobehavioral deficits in Inuit children from Qaanaqq, NW Greenland might be related 

to dietary mercury exposure. 

Therefore, the National Environmental Research Institute in co-ordination with the University 

of Copenhagen Global Change Initiative, COGCI, initiated Arctic mercury monitoring and research 

programmes, commencing shortly after the Canadian discovery. This Ph.D. is a part of these 

national efforts, and is focused on the atmospheric chemistry of mercury, its gaseous deposition and 

subsequent and historical accumulation in the Arctic.


1.1 The aim of this work 

The purpose of this work is to better understand the temporal, and spatial patterns of mercury 

deposition and accumulation in the Arctic, gaining an understanding of the chemical and 

depositional processes so that they might be applied to models of atmospheric transport and 

deposition and eventually policy decisions. To meet this aim, specific scientific questions were 

addressed. These specified questions contained implied questions or tasks that must be answered or 

addressed in order to investigate the specification. 

The specified and implied questions addressed in this Ph.D. work are as follows: 

Specified 1. What is the deposition of reactive gaseous mercury to the Arctic following a 

mercury depletion incident, and what is the depositional velocity of RGM? 

Implied 1.1. Reactive gaseous mercury, RGM, flux must be measured concurrently with total 

concentration of RGM. 

Implied 1.2. A method to measure RGM flux for the Arctic must be developed. 

Specified 2. What is a plausible reaction mechanism for the oxidation of gaseous elemental 

mercury to reactive gaseous mercury? 

Implied 2.1. Since the measurement technique applied for RGM concentration measurement, 

KCl coated annular denuders, does not allow for the subsequent chemical determination of RGM, 

HgXY, post field study analysis must be applied using relative rate theory studies and known 

reaction kinetics and thermodynamics. 

Specified 3. What are the pre and post-industrial values of Hg in the Arctic? 

Implied 3.1. A suitable environmental archive must be chosen; in this case peat. 

Implied 3.1.2 Arctic peat must be found and shown to provide a faithful record of atmospheric 

mercury deposition.


Implied 3.1.3 A method must be developed to reliably and reproducibly analyse very non- 

homogeneous sediment such as peat for Hg content. 

Implied 3.3. Arctic peat must be compared with traditionally studied raised bog, peat profiles. 

Implied 3.3.1 Arctic peat must be dated with high enough time resolution for comparison with 

raised bog profiles. The traditional method of dating highly mineral peat cores such as Arctic peat, 

i.e. 210 Pb chronologic modelling is sufficiently accurate for Arctic peat studies alone, but not 

accurate enough for dating raised bog profiles. A direct dating method must therefore be developed 

to allow comparison in the geochemically important upper layers of the peat profiles at each site. 

Implied 3.3.2 Techniques must be developed for safe, effective and efficient sampling of peat in 

the Arctic. 

1.2 Summary of results 

Bearing in mind the specified questions and implied tasks as noted, the main findings of my 

Ph.D. work are as follows. Papers referred to are included in Appendix C. 

Gaseous mercury is oxidized in the Arctic troposphere at polar sunrise. The resulting divalent 

mercury product species, HgXY, suggested as HgBr2 and/or HgBrOH, is deposited to the 

snow pack. The evasion of HgXY is also observed. 

The micrometeorological technique conditional sampling or “relaxed eddy accumulation”, 

REA, was developed for measurement of the flux of the operationally defined inorganic divalent 

gaseous mercury compounds HgXY, reactive gaseous mercury, in the Arctic (paper 1, Goodsite et 

al., submitted). 

The first measurements and quantification of HgXY surface flux and depositional velocity to 

the snow pack during polar sunrise was made, and an initial reaction mechanism suggested (paper 

2, Lindberg et al., 2002).


After polar sunrise, in Barrow, Alaska, KCl coated manual RGM denuders were used as the 

accumulator with the REA system. At 3 m above the snow pack significant RGM fluxes measured 

during March 29 th – April 12 th 2000 were directed toward the snow surface. Overall mean 

deposition was found to be - 0.4 +/- 0.2 pg m -2 s -1 ; N=9, +/1 SE and re-evasion was also observed. 

Using measured total RGM concentrations; depositional velocities were then computed and found 

to be on the order of 1 cm s -1 . Upon closer examination of field data, a relative rate study (paper 3, 

Skov et al., submitted) and utilizing knowledge from recently published reaction kinetics (Ariya et 

al., 2002), an updated mechanism is suggested (paper 4, Goodsite et al., submitted). 



forming the energised HgBr*. Through a third body reaction, M, where M is N2 or O2, the HgBr 

radical is formed. The HgBr radical can live long enough at the low temperatures of the Arctic to 

combine with O2 forming the HgBrOO peroxy radical or can combine with Br forming HgBr2. It is 

not likely to react with Cl, since this reaction would be endothermic. Similarly, the product cannot 

be Hg2Br2 since this would imply a tri molecular reaction, which is highly unlikely to occur in the 

atmosphere given the very low concentrations of the reactants. Nor would the end product be HgO, 

since this formation is similarly thermodynamically not favourable. The final product is the divalent 

gaseous mercury unknown, HgXY. By modelling the reaction of Hg and Br in the atmosphere, with 

current reaction constant data, and BrO and Br measurements, we find that assuming 20 ppt BrO 

and 2 ppt in the atmosphere during a depletion event that the lifetime of Hg is 4.6 hrs against 

forming HgBr, The lifetime of HgBr is 0.35 hrs. against forming HgBr2; comparing with the 

lifetime of HgBr of 0.75 hrs means that 68% of the time HgBr will form HgBr2. Thus the overall 

lifetime of removal of Hg to HgBr2 is 4.6 hrs. / .68 = 6.7 hrs. This is in good agreement with the 

observed 10 hr lifetime of Hg under depletion.


We learned from our experiences measuring gaseous elemental mercury on the Faroe Islands 

that measuring gaseous elemental mercury is not trivial. Care must be taken in choosing appropriate 

measurement procedures (paper 5, Skov et al., accepted). 

Mercury in environmental archives, peat in the Arctic 

Mercury was measured in peat cores from S. Greenland, the Faroe Islands, Denmark, and from 

the high Arctic (Bathurst Island and Carey Islands) samples from Bathurst Islands were given to N. 

Givelet, University of Berne, and will not be discussed. Samples from Nordvestø, Carey Islands, 

Greenland, are still being investigated, so preliminary results only will be treated. Cores from S. 

Greenland, Denmark and the Faroe Islands had sufficient new growth to examine the geochemistry 

during the last 50 years with high resolution. 

To do this, a new, direct, approach to high time resolution dating was developed and applied to 

peat from sub-arctic southern Greenland and Denmark. The advantages of direct, high time 

resolution dating are thoroughly discussed (paper 6, Goodsite et al., 2002). 

An analytical method was developed to best determine the amount of mercury in the peat 

(paper 7, Roos-Barraclough et al., 2002). 

The first long term terrestrial record of mercury on Greenland (paper 8, Shotyk et al., accepted) 

and the Faroe Islands are produced (Shotyk, Goodsite et al., unpublished data, manuscript in 

preparation). Hg fluxes in the Greenland core (0.3 to 0.5 µg m -2 yr -1 ) were found in peats dating 

from AD 550 to AD 975, compared to the maximum of 164 µg m -2 yr -1 in 1953. Atmospheric Hg 

accumulation rates have since declined with the value for 1995, 14 µg m -2 yr -1 comparable to 

published depositional rates. Lead and stable lead isotopes were measured in these cores, showing 

coal burning was the predominant source of lead. Mercury depositional trends in the cores follow 

the lead trends suggesting that the profile is dominated with mercury from coal burning and that the 

peat cores are faithfully reproducing depositional trends, in agreement with what is known about


predominant anthropogenic mercury emissions at the regional and global levels and their 

anthropogenic dependence on emissions such as coal burning (Slemr et al., 1992 and references 

therein). 

In Denmark, the greatest net rate of atmospheric Hg accumulation is found in 1953, 184 µg m -2 

yr -1 , comparable to that of Greenland with the flux going into sharp decline, with an accumulation 

rate for 1994 as 14 µg m -2 yr -1 . 

On the Faroe Islands, the maximum mercury concentration of 498 ng g -1 dw., is seen in the 

depth dated to be 1954 +/-2, with a 210 Pb dating model, in good agreement with the historical 

maximums in 1953 in S. Greenland and Denmark. Long-term accumulation rate of mercury in this 

core was 0.95 +/- 0.36 µg m -2 yr -1 for the period of 4200 B.C. to AD 833; n = 61, in agreement with 

what is known about the global mercury cycle, and what would be expected in a remote area. The 

Hg concentrations in the Faroe Islands are higher than those found in cores from other sites, but the 

net Hg accumulation rates are comparable. 

Mercury was also investigated in high arctic peat, requiring development of an improved corer 

for sampling (Paper 9, Noernberg et al., submitted). 

In addition to the above, in-progress, high Arctic studies, a study of Station Nord, NE 

Greenland was carried out for the Royal Danish Air Force. This is an engineering study conceived 

and carried out by this author studying a high Arctic station as a point source. It includes trace 

metals and mercury data from lake and marine sediments, as well as poly-aromatic hydrocarbon 

(PAH) data. It is an internally reviewed NERI report to the Royal Danish Air Force, (Goodsite et 

al., 2003, not included in this PhD but available from NERI). The results for mercury in the marine 

and lake sediments are in good agreement with other published results, as discussed in the report.


1.3 Sampling and Arctic fieldwork 

Much Ph.D. work was spent in the field. Safe and effective planning of fieldwork is discussed. 

A logical template meeting the reporting requirements for reporting Environmental Arctic fieldwork 

in Greenland, Canada and the US was developed. This template has since been accepted and 

utilized by research groups in Germany, Switzerland and Finland. Planning and executing fieldwork 

is discussed in Appendix B.


2. Overview 

The purpose of this chapter is to give a general background for the rest of this thesis, in order to 

understand the approach used to address the scientific questions. The overview should also help 

place the results of this work in the context of atmospheric chemistry and importance for the global 

mercury cycle and the Kingdom of Denmark, and therefore the Danish Arctic perspective. 

2.1 General Mercury 

Metallic mercury, Nr. 80 in the periodic system, density 13.5 g/cm 3 has been known since 400 

B.C. when a disciple of Aristotle (Theophrastus) described a method of producing what he 

described as “liquid silver” by rubbing cinnabar in vinegar. It has since been widely used in 

metallurgy and other industrial processes. 

As Hg is an element, man can neither create nor destroy it. 

Mercury is toxic and the toxicity of mercury vapour and its affects on the central nervous 

system has been observed and well known through time, for example, the Romans equivocated 

sending someone to work in the mercury mines of Spain as a death sentence. Most have heard of 

the “mad hatters”, felt workers who became insane from working with mercury, popularized by 

Lewis Carrol (“Alice in Wonderland”). 

Mercury use has been found in various cultures as a part of dyes, cosmetics and medicines (e.g. 

Ernst and Coon, 2001 or Garvey et al., 2001). Its ancient cultural use can be seen in examples from 

the old, new and Asian cultures. For example during the Kofun period (6 th century) Japan, it was 

thought that mercury was painted onto the body of a dead person (Yamada et al., 1995). 

The world’s largest producer of mercury has been and still is the Almaden area of Spain. 

Mercury has always been important to metallurgy. It was mined and sent to the new world for


extracting primarily silver, as well as gold and may account for 1/3 of anthropogenic emissions 

(Hylander and Meili, 2003). 

2.2 Arctic Mercury 

The Arctic has relatively few sources of mercury. Natural sources include the ocean, volcanoes, 

and reemission from snow and soil. Anthropogenic sources are typically point sources such as 

towns, settlements, mines and military bases. Most of the mercury arriving to the Arctic is via long- 

range atmospheric transport following anthropogenic or natural release in other parts of the world. 

The form of the mercury released depends on source type and other factors, such as if the 

source has any form for emissions control. Aside from gaseous elemental mercury, which 

constitutes the majority of emissions, typically emitted from coal burning power plants or mercury 

mines, the remaining emission of mercury are in the form of gaseous inorganic ionic mercury 

species typically mercuric chloride, with a much less fraction bound to emitted particles. Due to the 

reactivity and size of these forms, they will have a shorter atmospheric lifetime but still contribute 

to the global problem through transformation, reemission and subsequent long-range transport. 

Mercury is released into the environment from 4 primary sources and emission estimates vary 

widely. Most recent estimates (e.g., Lamborg et al., 2002, or Pacyna and Pacyna, 2000) estimate 

anthropogenic emissions as approximately 3000 metric tons per year, including 400 tons per year as 

reemission from anthropogenic mercury previously deposited to the Oceans, and approximately 

1400 metric tons per year from natural emissions. These estimates may be seen as probable lower 

limits with upper limits being up to twice as much. Anthropogenic Hg from fossil fuels and waste 

amount to more than 1500 t Hg year -1 , but these emissions may now be effectively reduced, i.e. 

more than 97% of Hg, by the use of a three stage system for flue gas treatment (Hylander et al., 

2003). This is a new and expensive technology however.


Given the amount of RGM measured in this study, when extrapolated to the Arctic by the 

Danish Eulerian Hemispheric Model, the amount of mercury deposited to the Arctic doubles to 

about 200 tons per year, or approximately 4.5% of the global total emissions as a result of the 

springtime arctic mercury depletion episode (Skov et al., 2003, Appendix C). 

It is hypothesized that in the Arctic, this mercury is thereafter released at snowmelt into an 

ecosystem starving for nourishment. It is therefore subject to immediate uptake into the marine food 

web with subsequent methylation and bioaccumulation to follow. 

Little is known of the magnitude of the transformation and subsequent re-emission of mercury 

as Hg (0). Reactive species, operationally known as reactive gaseous mercury (RGM) identified as 

HgXY are typically HgCl2 from power plants, in industrial areas, as well as total particulate 

mercury (TPM) will eventually deposit via dry or wet deposition. The true product or products, 

HgXY, have not yet been identified in the Arctic region. A summary of what is known with respect 

to Arctic atmospheric mercury, as well as identified needs and data gaps are found in Schroeder et 

al., 2003. 

2.3 Mercury recorded in environmental archives 

Post depositional processes for Hg or HgXY include biological or physical fixation, i.e. 

adsorption, and subsequent diagenesis, to include chemical oxidation or reduction and subsequent 

reemission of at least a portion of the mercury. Gaseous reemission, especially in tropical areas, or 

physical reemission through aeolian transport, adsorbed to particles. Reemission of mercury is one 

of the ongoing research areas and is very relevant to the global cycle. 

Due to a portion of the deposited mercury being reemitted, it must be understood that 

environmental archives of mercury represent a net record of accumulation, one that has been subject 

to the deposited mercury being subjected to physical and chemical processes that affect the overall 

amount of mercury that remains in the archive from year to year. Depending on the archive, and the


hydrological and geochemical conditions it is exposed to, there may also be post depositional 

vertical or horizontal transport. 

Working with environmental archives is therefore extremely dynamic and complex, with 

several functions that should be taken into consideration: the transfer function from air to plant, soil, 

ice, water or snow, the sedimentation function, which describes what happens to the mercury as it 

undergoes sedimentation, the fixation function or how the mercury is preserved in the archive, the 

diagenesis function which is how the mercury is chemically or physically affected over time and the 

reemission function which accounts for losses from the depositional area under consideration. We 

have therefore no way of knowing what the actual concentration of atmospheric gaseous elemental 

mercury was prior to the industrial age, or through the Holocene and previous geologic history. 

However, we can, by studying environmental archives such as lake, marine sediments, peat 

sediments and ice cores or historical records (e.g. Hylander and Meili, 2003) relatively determine 

that anthropogenic activities have increased levels of mercury in the atmosphere by a factor of 3 

(Munthe et al., 2001) or more, as noted in this work. 

As also seen from this work, actual spatial loading of mercury over geologic time is very 

consistent between different regions, thus indicating the global character of mercury transport and 

deposition. 

Based on emission inventories from 1995, it is seen that the majority of mercury being released 

to the Northern hemisphere and thus most burdening the Arctic is coming from anthropogenic 

emissions from coal fired combustion plants in Asia with primary burden in Asia coming from 

China (Pacyna and Pacyna, 2000). 

The total annual load of mercury transported to the Arctic, is a significant burden to the fragile 

and pristine environment. The Arctic weather system and cold as well as the highly productive and 

efficient post snow melt marine ecosystem, may represent the ultimate sink for mercury.


Alternatively, processes such as the start of one of the great Oceanic “conveyor belts”, may 

mean that much of the mercury brought to the Arctic is transported out again. Many questions 

remain about the Arctic mercury cycle. Therefore understanding Arctic mercury depletion events 

are important for understanding global mercury cycling. The Arctic may be a much more important 

sink or cycling node for global mercury than previously known. 

2.4 Overview of Arctic atmospheric chemistry “The polar sunrise mercury 

depletion event” 

The polar sunrise mercury depletion event, AMDE, or “the mercury sunrise”, is a highly 

elevated deposition of mercury which occurs during the first few months following the polar sunrise 

and was first reported at Alert, Canada by (Schroeder et al. 1998) having been discovered there by 

his team in 1995. 

In the Arctic during spring, the lifetime of gaseous elemental mercury is significantly shorter 

than 1 year, as mercury in the boundary layer is observed to be depleted in less than one day, during 

atmospheric mercury depletion episodes, AMDEs (Schroeder et al. 1998, Lindberg et al. 2002, 

Appendix C, Berg et al. 2001, 2003, Skov et al., 2003, Appendix C). 

Gaseous elemental mercury is converted to oxidized mercury in the gas phase to the 

operationally defined reactive gaseous mercury, RGM, product HgXY. RGM is relatively quickly, 

lifetime on the order of approximately 10 hours or less, deposited to the ground, as quantified in this 

work. 

It should be emphasized that it is not yet actually known what HgXY is in the Arctic. 

The mercury depletion happens simultaneously with the start of surface-level ozone depletion. 

The springtime destruction of surface ozone in the Arctic is a separate phenomenon from ozone 

depletion in the stratosphere, though halogen chemistry activation appears to be just as important 

for the tropospheric depletion.


The tropospheric depletion of ozone in the Arctic has been observed to be depleted during 

spring since, at least, 1966 (Tarasick and Bottenheim, 2002). It is generally accepted that the 

destruction of ozone is caused by catalytic reactions involving halogens, especially bromine, which 

originate from the polar oceans (Barrie et al., 1988 and Barrie and Platt, 1997). Our measurements 

suggest that the solar activity and present ice crystals influence the atmospheric transformation of 

elemental gaseous mercury to divalent mercury, which is more rapidly deposited (Lindberg et al. 

2002, paper 2, Appendix C). Satellite observations of BrO (as discussed in paper 2) suggest a 

correlation between mercury depletion and the total amount of BrO in the atmosphere, though later 

work shows the previously suggested mechanism to be more plausible, with the mechanisms 

presented in Lindberg et al., 2002 being thermodynamically unfavourable i.e. it is not likely that 

BrO/ClO + Hg 0 → HgO + Br/Cl radicals. Or that Hg 0 + 2Br/Cl radicals → HgBr2 or HgCl2. 

The most likely source of the halogens is the sea ice. Mercury depletion events therefore appear 

correlated with the presence of sea ice and are therefore probably found wherever there is sea ice 

and the potential for open leads. Today, typical levels of gaseous elemental mercury in the remote 

Northern hemisphere atmosphere and Arctic are about 1.5 ng m -3 (see paper 2). This level is a little 

bit less in the Southern Hemisphere, since most of the anthropogenic sources of gaseous mercury, 

are found in the Northern Hemisphere, and there is a lag time in transequatorial mixing. 

How long have mercury depletion events been occurring? Tarasick and Bottenheim (2002) 

have looked at surface ozone depletion events in the Arctic and Antarctic from historical ozone 

sonde records, and have noted ozone depletions since 1966, with an increasing frequency through 

2000 “explaining the apparent increase of Hg in Arctic biota in recent times”. They further note that 

“If the key to GEM depletion is the presence of reactive bromine, which can only be sustained via 

BrO recycling and hence ozone depletion, the increase of mercury in Arctic biota is a direct result 

of the increase in occurrence of ozone depletion episodes. “ Work accomplished by the present


research team indeed does suggest reactive bromine as one step in the mercury depletion 

mechanism and shows that Hg deposition has been decreasing. This researcher however, is not 

ready to directly correlate the measured increase of Hg in Arctic biota directly to ozone depletion 

events. This is a hypothesis that must be tested with future research. 

The work done in this Ph.D. on peat in fact suggests that global Hg deposition has peaked, and 

that net deposition is in decline. With an appropriate archive, the high resolution dating technique 

developed as a part of this work will provide a useful tool for investigating in detail, deposition to 

the Arctic coast since 1960. The chance of finding an appropriate high Arctic archive for this time 

period are however, slim, due to freeze thaw destruction of the upper layers of soil. 

The post snow melt flux into the marine environment is not yet well described, nor is the total 

mass balance for mercury cycling in the Arctic, though it is known that once deposited, the RGM is 

methylized, primarily by microbial metabolism, to methylmercury, which has the capacity to collect 

in organisms i.e., bioaccumulate, and to concentrate up food chains i.e., biomagnify or 

bioconcentrate, especially in the marine food chain. 

Atmospheric mercury has been of interest since the 60’s with an overview of mercury in the 

atmosphere published by Williston in 1968. Slemr et al., 1985, and Lindqvist and Rodhe (1985) 

give one of the first comprehensive distribution, speciation and budget of atmospheric mercury. 

Since then, two major reviews of the atmospheric chemistry of mercury have been published: 

Schroeder and Munthe, 1998; and Lin and Pekhonen, 1999. These studies are cardinal references 

with respect to atmospheric mercury. Lin and Pekhonen establish the lifetime of gaseous elemental 

mercury to be about 1 year, given what is known of the atmospheric chemistry of mercury at the 

time. The work in this Ph.D. suggests that globally, this may need to be re-evaluated, taking into 

consideration the extremely short lifetime of Hg (0) during Arctic polar sunrise.


The two reviews suggest that much like other trace constituents in the atmosphere, the 

atmospheric chemistry of mercury involves 5 probable families of reaction mechanisms/pathways: 

1. Gas phase reactions; 2. Aqueous phase reactions, in cloud and fog droplets and deliquesced 

aerosol particles; 3. Partitioning of elemental and oxidised mercury species between the gas and 

solid phases; 4. Partitioning between the gas and aqueous phases; 5. Partitioning between the solid 

and aqueous phases for insoluble particulate matter scavenged by fog or cloud droplets. 

Since the high Arctic is so dry, the predominating reactions must occur in the gas phase and the 

predominating deposition in the spring will be dry deposition. Dry depositional removal of gases 

from the atmosphere to a surface is one of the main processes by which pollutants are removed 

from the atmosphere. Dry deposition refers to the combined effects of three transfers: 1. Transfer 

through the turbulent layer of the atmosphere, 2. Molecular transfer through the viscous layer and 3. 

Transfer to the surface as a result of contact adsorption, dissolution and other contact phenomena 

(Karlsson and Nyholm, 1998). With respect to dry deposition on snow, Valdez et al. (1987) found 

that depositional velocity, Vd, most importantly is influenced by the liquid water content of the 

snow, i.e. the liquid water-to-air ratio in the snow, and lesser influenced by the density i.e. age of, 

and the degree of metamorphism of the snow pack, as well as sun light and temperature. Vd 

increases for increased values of snow water content, decreased values of Henry’s law Constant, H, 

and increased diffusion coefficient. The corollary is that surface resistance, rs, decreases. Vd 

decreases, i.e. rs increases, with decreased snow temperature, or increased time. The increased time 

exposure reflect that one expects desorption to occur as atmospheric concentrations are depleted 

and the concentrations in the snow come into equilibrium with the atmosphere.


Given the water solubility and reactivity of RGM, the maximum Vd expected would be on the 

order of 10 cm s -1 , with minimum Vd of about 0.5 cm s -1 .


3. Experimental Methods, Equipment and Procedures 

The purpose of this section is to describe the methods used to answer the scientific questions 

raised. The experiments presented in this work were carried out at different laboratories and field 

locations, utilizing and developing methods for air sampling and analysis as well as peat sampling 

and analysis. It is appropriate to credit the list of collaborators and visits that have made the 

experimental work possible and this list is included as a preface to Appendix C. The limit of 

detection, where mentioned refers to that concentration which gives an instrument signal 

significantly different from the blank or background signal. There is no full agreement between 

statutory bodies or researchers as to what is best to use as the limit of detection. Throughout this 

study it is defined in one of the most commonly encountered ways: as the blank signal plus three 

standard deviations of the blank. It is important to determine the limit of detection to avoid 

reporting the presence of the analyte when it is actually present, or avoid claiming the presence of 

an analyte when it is actually absent.


3.1 Study Sites 

Barrow 71 0 18’N, 56 0 

40’W: RGM flux 

Bathurst Island: Peat 

Given to N. Givelet, Berne, for 

analysis along with peat from 

Continental Canada and USA. 

Nordvestø, Carey 

Islands, Greenland 76 0 

44’ N, 73 0 12’ W: 

Guanogenic peat, 

minerogenic peat. 

Preliminary Hg, dates 

only Nuuk 64 0 02’ N, 51 0 

07’ W: Hg(0), O3. 

Ongoing 

S. Greenland, 

Peat 

Figure 1. Sites of field campaigns. Map:CIA public regional reference map. 

Station Nord 81 0 36’ N, 16 0 39’ 

W: Hg(0), O3, RGM; Lake and 

Marine Sediments 

Faroe Islands: 

Peat 

Hg(0), O3 

Denmark 

(Funen): 

Peat


Atmospheric mercury and ozone measurements were made by the NERI team at two sites in 

Greenland: Station Nord and Nuuk, and on the Faroe Islands, in Thorshavn; and by the 

ORNL/NOAA/US EPA team at Barrow, Alaska. 

Manual measurements of reactive gaseous mercury and experiments were performed at Oak 

Ridge National Laboratory, Environmental Sciences Division, and at Walker Branch Research 

Station, US National Oceanic and Atmospheric Administration, NOAA, Atmospheric Turbulence 

and Diffusion Division, ATDD, Oak Ridge, USA. RGM measurements in the spring, 2001 

campaign in Barrow Alaska, were performed by the US team. 

The relaxed eddy accumulation system was developed at NERI, with trial and developmental 

experiments carried out in Oak Ridge. The system was deployed for the first ever measurements of 

RGM flux in the Arctic at the Barrow Arctic Mercury Study site, during the spring of 2001, with 

following deployments for developmental purposes only, at Walker Branch in Oak Ridge, as well as 

Lille Valby, a NERI experimental site, at an agricultural field north of Roskilde, Denmark. The 

improved NOAA relaxed eddy accumulation system was deployed at Lille Valby to compare with 

the initial REA system, and for the Spring 2002, Station Nord campaign. The Station Nord 

campaign deployed the first NERI system as well, as a back up. 

Peat cores were collected from a fen located in southern Greenland, a raised bog in Denmark, a 

blanket bog on the Faroe Islands, and two Arctic island locations: on Bathurst Island, Canada and 

on Nordvestø, Carey Islands, Greenland, located between the northwest corer of Greenland and 

Ellesmere Island, Canada. Samples from Bathurst Island and continental Canada and the USA were 

given to another student, Nicolas Givelet, Berne, for primary investigation, and will not be 

discussed in this thesis. Samples from Nordvestø are unique in that they are from peat that is 

guarnogenic, i.e., nourished by sea bird droppings, as they sit on the island cliffs looking for prey. 

They therefore have the potential of providing a primarily biogenic Holocene record of mercury


accumulation. These samples will be preliminarily discussed since at the time of completion of this 

PhD, research is still in progress. 

3.2 Measurement of atmospheric elemental mercury and ozone 

As seen from measurements in this work, and first reported by Shroeder et. al, 1998, the 

destruction of ozone coincides with an atmospheric mercury depletion event. Therefore ozone 

monitoring was an integral component of the campaigns investigating atmospheric mercury in the 

Arctic, as is the measurement of total gaseous mercury, or gaseous elemental mercury. 

The monitor used for atmospheric mercury measurements is the same analyser employed for on 

site analysis of manual reactive gaseous mercury measurements, though there is also now, a fully 

automated RGM monitor available, which combined with a particulate mercury unit and TGM 

monitor, provide automated data of mercury fractionation. This type of unit will greatly improve the 

effectiveness of future studies. 

Ozone and GEM were measured at Station Nord, NE Greenland, until June 2002. On the Faroe 

Islands, from May 2000 forward one year, and in Nuuk, Greenland from March 2002 and are still 

ongoing. GEM was only measured at Station Nord from February until July each year. 

3.3 Ozone measurements 

Ground level ozone was measured using an UV absorption monitor, Monitorlab 8810, with a 

detection limit of 1 ppbv and an uncertainty of 3 % for concentrations above 10 ppbv and 6 % for 

concentrations below 10 ppbv, (all uncertainties are at a 95% confidence interval). 

3.4 Gaseous elemental mercury measurements 

Gaseous elemental mercury, GEM, was measured with high time resolution using an automated 

dual channel, single amalgation, cold vapour atomic fluorescence spectroscopy (CVAFS) mercury 

analyser, Tekran Analyser Model 2537 A, Tekran Inc., Toronto, Canada, following a standard 

operating procedure (SOP) written by Alexandra Steffen and W. Schroeder, for use at the high


Arctic station, Alert, and distributed to other Arctic mercury research teams for standardization, as 

broadly as possible (Steffen and Schroeder, 1999). 

NERI adopted this SOP to the situation and resources that were available to us for our 

measurement campaigns. The major changes were the type of heated sampling line and quality 

control procedures. As we were not on site for the entire campaign period, we did not conduct a 

daily manual calibration of the machine, relying instead upon on the permeation tube as a secondary 

standard. 

By directly sampling the air, the machine measures total gaseous mercury, TGM, utilizing two 

channels. One channel is being sampled while the other is being thermally desorbed and then the 

function switches, allowing continuous sampling. 

Through introduction of a soda lime trap into the sampling chain, as well as a 45 mm diameter 

Teflon filter, to protect from the introduction of particulate matter, only the gaseous elemental 

mercury fraction is measured. 

Measuring GEM can also be achieved by introducing a KCl coated denuder into the sampling 

train, Figure 2., page 33, which captures only reactive gaseous mercury, as described below. 

Although it is the method utilized by Tekran for their automated mercury speciation unit, Model 

1130, it was not utilized as part of the present study’s mercury sampling chain, except by the 

American teams at Barrow, when reporting GEM. 

The principle of the Tekran 2537A instrument is cold vapour atomic fluorescence spectroscopy 

(CVAFS) and is described as follows: A measured volume of sample air is drawn through a gold 

trap that retains elemental mercury. The collected mercury is desorbed by heat and is transferred by 

argon into the detection chamber where the amount of mercury is detected by CVAFS.


The detection limit is 0.1 ng/m 3 and the reproducibility for concentrations above 0.5 ng/m 3 is 

within 20% on a 95% confidence interval. This is sufficient for quantification of RGM, and 

therefore also for using the annular denuders for flux measurements. 

Volatile organic mercury species such as dimethylmercury and monomethylmercurychloride 

may also react with the gold traps, however their contribution to the TGM fraction can be neglected 

under background conditions (Ebinghaus et al., 2001). 

In order to protect the instrument against passivation of the gold traps due to for example, 

humidity and sea salt in the air, a soda-lime trap, soda lime pellets in a Teflon cartridge, was placed 

in the sample line just before the 2137A analyser, as proposed by M. Landis, US EPA (personal 

communication) just before the 2001 season as humidity and sea salt can cause serious artefacts 

(Skov et al. 2003, paper 5, Appendix C). However, there was not observed any change in the 

general level of GEM after the installation of the trap. Experiments in Denmark, at 56 0 N with 

parallel measurements of GEM with and without soda lime trap showed a perfect agreement within 

the uncertainty established previously. A heated coil around the Teflon sampling line, imparted an 

effect of 50 0 C above ambient temperature in order to assure that mercury did not condense on the 

sampling lines.


3.5 RGM determination 

RGM was measured and analysed by the method developed by Landis et al. (2002). Dr. Landis 

along with Dr. Robert Stevens, trained this author on the coating and manual measurements and 

desorption technique at the US EPA at Research Triangle Park, prior to going to Barrow Alaska. 

The Landis et al., method for RGM determination uses a KCl coated quartz annular denuder 

Figure 2. The quartz denuder system from University Research Corporation (URG), North Carolina, 

used for the collection of RGM in ambient air. Diagram from URG catalogue. The KCl coating is in 

the annulus space, in the grey, sandblasted region. The inlet and outlet was carefully rinsed of KCl, 

and the coating was uniform, with no large crystals. Heating mantle not shown.


sampling chain heated to 50 0 C, to sample air, Figure 2., page 33, and is described in detail as cited 

so the sampling system and method will therefore only be briefly reviewed. 

At the inlet, there is an elutriator and an impactor, with impactor plate. The elutriator is coated 

in cross linked teflon. The elutriator straightens the flow and accelerates it, by forcing it through an 

orifice onto a roughened impactor plate with no coating applied. The cut-off diameter is 2.5 µm, so 

only the fine fraction of particles flows past the active area of the denuder. The sample flow is 10 

litres per minute. The flow is controlled just prior to the denuder chain with a “dry cal” flow meter 

prior to and after sampling. 

Immediately following the impactor, there is a dead space prior to the annulus. This allows for 

expansion of air, from ambient to 50 o C, since KCl optimally collects RGM at this temperature 

(Landis et. al, 2002); as well as proper development of laminar flow, which is a necessary condition 

for proper functioning of denuders. 

9 annular denuders were available: 6 new, belonging to NERI and 3 approximately 1 year old, 

belonging to ORNL; however, there were only 4 complete sampling systems as depicted in Figure 

2., page 33. 

Measurements made in Barrow were therefore with US EPA denuders and sampling systems, 

to include the annular denuder system employed in the reactive gaseous mercury flux system. 

The denuders were characterized in the laboratory of Dr. Lindberg at Oak Ridge National lab 

prior to field deployment. 

Upon receipt, denuders were physically inspected. The physical inspection is an important 

aspect of quality control prior to using the denuders and includes checking for cracks, and ensuring 

that the threads produce a tight seal. Each new denuder was then numbered 1 – 6, by carefully 

etching with a diamond pen. The 3 older denuders were numbered X1 – X3.


The denuders were then cleaned, coated, dried, and prepared according to the protocol for 

manual denuder measurements provided by Dr. Landis. As published in Landis et al., 2002. 

Each denuder was then tested for eventual variation of the blanks of the denuders. Denuders 

were also concurrently sampled in pairs to ensure that collocated precision was < 15%, in 

agreement with the findings reported for the method in Landis et. al., 2002. 

Denuders were exposed to laboratory air with relatively high RGM concentrations, 

approximately twice that expected in ambient air, due to the past and present use of HgCl2 in the 

building at Oak Ridge National Lab; and ambient air at Walker Branch Watershed Research Area, 

Oak Ridge, Tennessee with and without end caps to test if there was any appreciable passive 

diffusion. This is important since only one of three denuders is actively sampled in the flux system 

at any given time. Over the sampling period it is expected that the denuder will be idle, subject only 

to passive uptake two thirds of the total sampling time. 

The heating mantels are different than those used by Landis, et al., 2000, since theirs were 

judged to be too bulky for use in the flux system where at least two would be required. The heating 

mantels used in this work are originally developed while at NOAA, under guidance of chief 

engineer, Mark Hall. The original mantles are seen hanging from the tower in Figure 3., page 39. 

The heating mantles are described as follows: high temperature polyvinylchlorid, PVC, pipes, 

which allow for a 2 cm airspace around the outside of the annular denuder, and encloses the 

denuder from the tip of the inlet to the top of a filter pack at the outlet. The outer portion of the pipe 

is wrapped with a silver tape to ensure heat transfer from self-regulating heat tape. The heat tape is 

uniformly wrapped around the pipe and grounded on the metal tape. This is then covered with 

insulating tape. The length of the heating tape is manufacturer and electrical voltage dependent, but 

is cut long enough to heat the air inside the tube to 50 o C and keep it at that temperature.


Based on experience gained at Barrow, the heating mantles lost too much heat, so further 

improvement was made at NERI to the heating mantles prior to the Station Nord, 2002 campaign. 

The improvements are specifically: an extra layer of insulation, an extra hard shell for 

enhanced wind protection, silicone sealant to protect against water or melted snow and insulated 

end caps to warm from the inlet onwards. The original heating mantle assembly was placed inside a 

larger PVC pipe, allowing 5 cm spacing between sidewalls, giving an overall diameter of 

approximately 9.6 cm. The space between the two shells was then filled with self-expanding 

polyurethane foam thermal insulation. Upon drying, the insulation was cut, so that top and bottom 

end caps will fit snugly. The inside of the top and bottom caps are filled with foam cut to fit the 

inlet and outlet of the denuder. The heating mantle is sealed at each end with silicone, so that it is 

watertight. Other shells than PVC may be used, for example, metal tubing. This work chose 

thermally rated PVC since it was cost efficient. 

The result is a self-regulating heating mantle capable of producing the heating required for an 

efficient active coating temperature of 50 o C in the Arctic spring. 

page 53. 

The improved heating mantles are seen on the tower in the photo from Station Nord: Figure 5., 

Once the air in the heating cap is heated and the denuder is heated to equilibrium, there is no 

heating gradient directly across the mantle wall to the denuder, as there is in heating mantles that 

employs direct application of heating tape to the denuder walls, since the airspace functions as an 

insulator, thus avoiding uneven heating and activation of the denuder coating, or over heating the 

coating. 

However, there is certainly a gradient and therefore expansion of air, as an air parcel travels 

through the denuder and becomes warmer. This is a source of experimental error that we


investigated as part of this work, at Station Nord, 2002, by comparing RGM measurements in 

heated versus non-heated denuders. 

Never the less, for all flux measurements heating was kept constant, since the heat might affect 

the laminar flow within the denuder, the gas diffusion coefficient and the relative adherence of the 

gas to the KCl surface. 

After sampling, the quartz denuders were closed immediately with plastic caps equipped with 

teflon inner seals, and taken into the laboratory for thermal desorption: 

During sampling RGM is converted by KCl to HgCl2. HgCl2 is pyrolized to Hg(0) at 

temperatures above 370 o C, so heating the quartz tubes to 540 o C in a Lindberg tube oven, Lindberg 

Blue Tube Furnace Model Mini Mite, no. TF55030A-1, Serial Number Z11K-508436-ZK, ensures 

a complete pyrolization of RGM to Hg (0) over 15 minutes with a flow of purified air. Quartz filters 

are used to ensure a good seal between the annular denuder and the clamshell furnace, to help 

prevent direct heat escape, as are coaxial cooling fans, set to cool the denuder ends as the denuder is 

being heated. The effect of using and not using cooling fans was tested. 

The sample is introduced with charcoal filter purified air directly into a TEKRAN 2537A 

mercury analyser and quantified as follows: 

Three sampling runs were made prior to heating in order to test for leaks, and establish a blank 

value followed by three heated runs and then two runs at ambient temperature again. Each run is for 

5 minutes duration. If zero readings on the TEKRAN were not obtained after 3 heated runs then the 

denuders were washed and recoated in accordance with the procedure in Landis et al., 2002. 

Otherwise, the denuders were considered blanked, and ready for use again. Care was taken to 

keep the ends of the denuders cooled while the active area was being heated, by plugging the ends 

of the oven with quartz pads and using co-axial fans prior to and after the oven, so that only RGM 

captured on the active area was desorbed.


For all measurements a field blank was obtained by handling a denuder in the field. Hg Mass 

from this field blank was subtracted from the measured Hg masses on the exposed denuders. 

If there was any indication of Hg (0) adsorption, for example with sudden sharp increases in Hg 

amounts then the denuder was cleaned and re-coated, since as pointed out by Sheu and Mason, 

2001, just 1% of Hg (0) adsorption on a denuder is enough to compromise RGM measurements. It 

is therefore very important to carefully follow the prescribed methods, especially since annular 

denuders are used as part of the reactive gaseous mercury flux measurement system. 

All accuracies with the denuders were found to be in good agreement with those reported by 

Landis et al., 2002. In the Barrow 2001 campaign, the EPA manual denuders exhibited a precision 

of 10%, based on co-located parallel measurements. 

During sampling the denuders were kept constant at approximately 50 o C above ambient 

temperatures with the initial heating caps, or at 50 o C, with the improved heating caps. Temperature 

was verified by sliding a thermometer probe into the air surrounding the denuder between the outer 

wall of the denuder and the heating mantel, and measuring the air temperature. It is assumed that 

the denuder coating temperature is in equilibrium with the air temperature.


3.6 Flux measurements of reactive gaseous mercury by relaxed eddy accumulation 

Figure 3 Relaxed eddy accumulation system deployed on a tower support pole at NOAA CMDL, Barrow Alaska. 

The photo is toward Point Barrow, with the system oriented northeast, in the direction of prevailing 

winds coming from the Beaufort Sea, approximately 2 km from the station. NOAA CMDL is 

approximately 9 m above sea level, with a flat terrain profile. Metek Sonic anemometer mounted 

independently on 1 side of the support. 3 original heating mantels mounted behind the sonic 

anemometer. The two heating mantels to the right are for up and down channels, the heating cap to 

the left is for mid channel denuders. Original heating caps could warm the denuders to 50 0 C above 

ambient such that typical temperatures were 20 0 C. Black neoprene air hoses go to the black box that 

will be buried in the snow to hold the REA switches and other components warm. From left to right 

in the photo, S. Brooks, NOAA, M. Goodsite, NERI and M. Landis, US EPA. REA control system 

from Metsupport, Denmark. Courtesy of S. Lindberg, ORNL. 

Relaxed eddy accumulation, REA, is a micrometeorological method for trace gas flux 

determination. All micrometeorological systems require the measurement of the turbulent 

component of air and the measurement of the trace gas of interest. REA “relaxes” the requirement 

for instantaneous gas analysis by preferentially collecting air over time into some type of 

accumulator and analysing the trace gas in the collected sample after the sampling period.


Therefore any REA system requires collection of the trace gas of interest over time. When 

RGM is the trace gas of interest heated, KCl coated, annular denuders may be used as accumulators. 

Collection and measurement of RGM was discussed in the previous section. This section will 

present the measurement of the reactive gaseous mercury flux using the relaxed eddy accumulation 

system and its components. 

RGM flux measurements were performed from 29 March, 2001 through April 12, 2001, with 

the system set up at approximately 3 m above the snow pack surface on a guy-wire support of the 

NOAA, Barrow, CMDL tower, oriented into the prevailing winds, arriving from the Beaufort Sea. 

The measurements were made using a micrometeorological flux measurement system built by 

METSUPPORT aps, Denmark in January 2001 for this campaign. The system was designed in 

consultation with this author and Dr. Skov, together with the Director of METSUPPORT, Dr. P. 

Hummelshøj, with this new, Arctic application of REA in mind. 

METSUPPORT dubbed the system the Mobile REA Data Acquisition System, Model 1005-02, 

serial no. 10012, and shipped it directly to the author in Oak Ridge, where its components were 

tested at NOAA, ATDD and Walker Branch, prior to shipping to Barrow. 

This system was coupled with the Landis et al., 2002, annular denuder method for measuring 

RGM, and the previously described heating caps, as the sampling front end. 

The total system consists of 4 primary components: 

1. The sampling end, consisting of three annular denuder sampling trains and associated 

heating caps, with neoprene tubing running to the switches. 

2. A METEK, Germany, model USA-1 heated, sonic anemometer, serial no. 2000 12 

003/01 with a separate weather tight electronic box. The effect of the heating is 55 W, 

sufficient to prevent rime ice build up under Arctic conditions. The sonic provides 

airflow information as a 10 Hz serial data stream.


3. A data acquisition suitcase. The central component of the suitcase is an Advantech 4823 

single board computer, with a PC 104 bus. There is a relay card mounted on the bus to 

control switching valves and a 12 bit analogue input card for monitoring up to 6 

analogue input signals plus the supply voltage and an analogue zero channel. A RS422 

serial port on the Advantech computer allows serial data input from the sonic 

anemometer. A 3 ½ inch floppy disc drive is provided for data download and back-up. 

Automatic data back up is a function of the software. A 6 GB hard disk with a DOS 

operating system controls the REA data acquisition software, DAQ, 

METSUPPORT/Risø DAQ version 4.10, with integrated REA control. The initial 

software set up provides for automatic booting once power is supplied to the system. A 

60-minute run, made up of two minute archived statistical data blocks will start at either 

clock time 00 or clock time 30, unless manually bypassed. The REA control software 

has a deadband, based on a moving 10-minute standard deviation of the sonic vertical 

signal and an adaptive correction for tilt angle. The average tilt angle is calculated for 

every four-degree sector and the vertical component of the wind is corrected using this 

angle. The software used only the most recent 10 minutes to determine the average tilt 

angle, to best account for changes in flow regimes due to varying speed and stability. 

Therefore, the REA system will first start switching 10 minutes after it is started. The 

above software is designed to maximize the REA measurements. The deadband is 

discussed later in this thesis. The software output was written to write just basic data to 

the floppy drive at the end of each hour, i.e. run. This data includes those factors needed 

to calculate the flux and control the measurement including: run name, defined 

conveniently as the date and start time, average temperature for the run, average wind 

speed for the run, average wind direction for the run, the standard deviation of the


vertical wind velocity, needed to calculate the REA flux, see later, and the value beta, 

also needed to calculate the vertical flux and the number of times each valve was 

switched in during the run. The data could only be retrieved by stopping the system, so 

data was seen only first after the sampling period. The text output file was not 

comparable with the file saved for each individual run, the .ACU extension files. In the 

text output file, the total number of counts did not add up to the total run time, and the 

value for the standard deviation of the vertical wind velocity and sometimes, beta, was 

different for the run than the value found in the .ACU file. Therefore, the .ACU values 

were used for consistency throughout analysis, since the counts in the .ACU file 

equalled the total run time. The data from the .ACU file for each run was imported by 

hand-typing into a spreadsheet where it was averaged for the entire sampling period and 

total sampling time for each denuder calculated by multiplying the switching time by the 

sampling frequency. The software allows adjustment of the sampling frequency and for 

the third channel, known as the MID channel to either remain open for constant, co- 

located sampling, or to open recording the air in the dead band, such that the sum of the 

three channels should equal concentrations contained from a separate constant sampling 

system, allowing some aspect of quality control, and employed for this study. The 

computer time was not internally changed for this study, but left at its original setting, 4 

hours ahead of Barrow, i.e., a run name with a start time noted as 20:00 was started at 

16:00 local time. 

4. A REA valve box, METSUPPORT model 1006-01, serial no. 10013, built to interface 

with the sampling front end with fast response valves, that are meant to run with 5 m 

long cables between the valve box and the switches, so that they might be mounted as 

close to the outgoing end of the denuder as possible. It was thought that heat from


switching would keep the valves warm enough from freezing in the Arctic. This proved 

to be a faulty assumption, valves froze on the second day, so from the third campaign 

day forward, the valves along with the valve control box, sonic control box, mass flow 

controller, and pump were buried in the snow pack inside of a box, beside the tower to 

keep them warm enough to function. 

5. The system did not come with a power supply, just a 5 m long power cable, which was 

connected to a 24 volt DC power supply at Barrow by CMDL engineer Glenn 

McConville. The machine pulled 24 volts at 1.2 amperes. A Sonic cable, connecting the 

sonic control box, which was buried beside the tower, to the REA control suitcase, 

located in the garage of CMDL as well as a 50 m long cable connecting the control box 

to the valve switching box were included. Their plastic covers made them moveable 

only with pre-warming once laid out in the Arctic terrain, and great care was taken not to 

break them. 

A similar METSUPPORT system and components have been previously deployed by NERI for 

measuring volatile organic compounds (VOCs) and is described in detail in Christensen et al. 2000. 

The REA flux measurement system was set up on a steel rod affixed to the CMDL tower guy- 

line support pole with the head of the sonic anemometer facing into the predominant wind direction, 

and oriented towards North. See Figure 3., page 39 and Figure 4., page 44. The support beam was 

hung such that the heating caps and therefore inlets of the sampling system were perpendicular to, 

and 1 meter behind the centrum of the sonic head. The inlets were 1-2 mm longer than flush 

protruding from the bottom of the heating caps. The denuders for the up and down draft were co- 

located nearest the centre of the mast, while the parallel measurements or denuders sampling the air 

that is not coming either as down or up, were located near the edges of the mast. The denuders are 

hung vertically so that particles do not fall onto the denuder walls. The denuders rest with gravity in


the heating caps, due to the filter packs. A quartz filter was kept in each of the filter packs, to ensure 

a constant pressure drop. Flow was measured with a mass flow meter and a dry cal mass flow 

meter, prior to and after sampling, as was temperature in the heating caps. The dry cal meter and 

electronic mass flow meter were attached to the end of the sampling line with a quick connect, in 

the same manner the sampling line is attached from the denuder. A piece of Teflon tubing was used 

to connect the mass flow meter to a denuder inlet to ensure the proper sampling rate and adjustment 

of the mass flow controller. This adjustment was made at Walker Branch prior to system 

deployment and confirmed at Barrow. 

Figure 4. REA RGM system schematic 

3m above 

the snow 

pack 

Prevailing 

wind 

direction → 

REA control 

computer (up to 50 

m from the tower) or 

in a box beneath the 

tower, in the case of 

the NOAA system at 

Station Nord. 

1 m 

Sonic 

control 

box 

Switch 

control 

box 

Cross support perpendicular to 

predominant wind direction. Sonic is 

oriented North, and in the prevailing 

wind direction. Neoprene sampling 

lines between end of denuders for the 

up, down and mid channels and 

solenoid switches. Solenoids kept in 

a box under the snow pack to keep 

from freezing 

120 v power cables for: 

heating caps, pump, 

mass flow controller not 

depicted 

Mass flow 

controller


From the quick connect at the top of the filter pack were connected 3.2 m long neoprene hoses into 

3 fast response switching valves supplied by MetSupport. From behind the switches, the three 

valves were connected into one sampling line using a simple 3 inlet manifold constructed with 2 T- 

type locking copper hose connectors in series and 1 L type locking hose connector as the end piece. 

Coming out of the manifold, a locking ball valve, was used to adjust and fix the flow, as a back 

up to the mass flow controller. Between the pump and the valve, a Tylan mass flow controller was 

used to ensure a flow prior to the manifold of just over 10 litre min -1 , so that the flow was measured 

as 10 litre min -1 at the denuder inlet. Pressure loss was minimal in the manifold and through the 

sampling lines. Once the system was running, the lag time from when the switch opened to when 

the flow started out the denuder was very small compared to the air sampling switching frequency 

of 1 Hz. 

Normally flux systems operate at air sampling switching rates of 10 Hz, switching as fast as the 

air flow is sampled with the sonic anemometer. The only lag time between the air and the switching 

comes from the software and physical switching process, including flow development in systems 

that do not have zero air induced into the sampling inlet system to stabilize the flow. This means 

that within one second the flow can effectively be changed between one accumulator to another 10 

times, sufficient for total air accumulators such as canisters or bags. Denuders are selective 

accumulators, collecting only the trace gas of interest, and require laminar flow. Given the 

geometry of the URG annular denuders, the 10 Hz sampling switching rate nominally allows for 

full laminar flow development and escape of an air packet through the end of the annular denuder. 

From initial measurements, it was immediately clear, that the annular denuders were not 

collecting efficiently while switching at 10 Hz, with RGM mass just above detection limits being 

collected. Therefore tests were carried out in Oak Ridge, on top of a 50 m tower at Walker Branch, 

to determine if the switching rate required for good reproduction of total RGM ambient


concentration in a flux system sampling in three channels: up drafts, down drafts and all else, as 

compared to levels measured in parallel annular denuders under constant sampling conditions, 

could be set at 1 Hz. 

In these experiments flow rates were controlled by a Tylan mass flow controller and checked 

prior to and after the sampling period. The results of these experiments led to the sampling 

switching rate being set in the REA system to 1 Hz with the 10 litre per minute flow rate maintained 

by a mass flow controller. The air sampling-switching rate is so long compared to errors that could 

be induced from physical switching and software, that these are considered negligible. Zero air was 

not added to the sampling inlet system since this would mean that ambient air could not be directly 

sampled into the annular denuder, but that ambient air would have to flow through an additional 

inlet first. RGM is so reactive that it was thought best to directly sample it into the annular 

denuders. This is possible given the 1 Hz switching rate. 

Therefore, by sampling at 1 Hz, approximately 95% of the turbulence is captured and the best 

compromise between the meteorological measurements and chemical sampling is obtained in order 

to ensure a laminar flow in the annular denuders and thereby measure RGM flux most accurately 

Measurements were made as follows: The heating mantles are turned on approximately 1 hour 

prior to measurement. 4 annular denuders were prepared and blanked immediately prior to use: 3 

for the system and one as the field blank. Inlets are baked prior to field deployment unless 

contaminated by handling. Powder-free rubber gloves are used when handling the inlets. The 

denuders are then taken outdoors and carefully placed into the heating caps. Care is taken not to 

disturb the sonic alignment, or crack a denuder due to handling, nor handle the sampling inlet since 

the annular denuders are relatively fragile under Arctic weather conditions, and the inlets may 

easily be contaminated. Flow is checked in the sampling lines and temperature measured in the 

denuders. Once the temperature of the denuders has stabilized, the REA system is started by


applying power, since there is an automatic start routine. A floppy disc is inserted into the floppy 

disk drive for data output. Air is then sampled at least for 4 hours, i.e. four REA system consecutive 

runs, sampling switching at 1 Hz in all three channels, if a separate RGM ambient air concentration 

is being measured. If the system is used to obtain the ambient concentration of RGM from constant 

sampling, then just the up and down channels are switched, and the mid channel remains open. The 

mass flow is adjusted in the manifold to just over 20 litres per minute. 

After the 4 hour sampling period, the system is turned off. The floppy disk is removed and data 

is imported and analysed as described previously. The annular denuders are taken off of the tower 

and tightly capped. The denuders are then analysed for RGM concentration as previously described. 

Flux and depositional velocity may then be calculated as described below. 

The best way to measure flux with a meteorological system, is instantaneously measuring a 

trace gas in an air parcel going up or down and correlating it with air mass exchange data, called 

“eddy correlation”. With present analytical techniques, this is not possible for reactive gaseous 

mercury, so relaxed eddy accumulation was used. Desjardins (1972) proposed a means of 

overcoming the problem of not being able to instantaneously measure the trace gas exchange 

simultaneously with the mass exchange of the air by proposing accumulating the air and airflow 

data over time, and completing the analysis after a finite sampling period, called “eddy 

accumulation”. Businger and Oncley (1990) further simplified this method by replacing the fast 

response trace gas analyser with fast response valves accumulators. This allows trace gases to be 

collected over time, each into their own accumulator: gases on the way up, or gases on the way 

down; and analysed after an appropriately long sampling time to ensure enough trace gas was 

accumulated to know if the observed difference in the up and down reservoirs was significant or 

not, and thus determine the flux.


The decision making data for the switches, i.e., sample into the up or down accumulator, from 

an instrument that provides airflow data to the REA control computer. 

All modern systems including the present system use a sonic anemometer. A sonic anemometer 

measures the three components of the wind velocity by determining the windspeed from the flight 

times of ultrasonic pulses across a fixed path of length l. Opposing transducers alternatively act as 

transmitters and receivers, sending ultrasonic pulses between themselves in direction 1 and 2. Thus 

the flight times in each direction are directly measured (2): 

t1 = l/(c+v) and t2 = l/(c-v) (2) 

where c is the speed of sound and v is the windspeed along the sound path. 

Solving for the windspeed v: 

v =(l/2)(1/t1-1/t2) (3) 

(2) may also be solved for the speed of sound c 

cmeasured = (l/2) (1/t1 + 1/t2) (4) 

This is then corrected to find the actual speed of sound, based on the fact that cmeasured is 

distorted by the component of wind perpendicular (normal) to the sound path, expressed by Kaimal 

and Gaynor (1991) as: 

cmeasured = ( ctrue 2 – vnormal 2 ) 1/2 (5) 

Via the ideal gas law, the pressure is related to the temperature. C is related to the adiabatic 

compressibility and therefore to the dry air properties. Without solving (5) further for ctrue = c , it 

can be seen that software algorithms readily can calculate c and thus calculate the temperature. 

However, to calculate c effectively, the humidity must be known. 

We did not measure humidity for this experimental set-up, since the pilot project was carried 

out in the arid high Arctic.


The temperature reading was however, used to control that the sonic anemometer system was 

running properly, since the temperature calculation is related to the flight time determination, as 

seen in (4), and if the path length determination in (4) manifested in unusual temperature readings, 

then necessarily the vector determination was affected, as seen from (3). 

Throughout the experiments, the sonic anemometer functioned well, though as will be 

discussed later, in Barrow, the sonic was not set on the tower properly, causing the down channel to 

be preferentially sampled. This was corrected analytically after the measurements as discussed later. 

It can be seen from (3) that the wind speed is not affected by the speed of sound in air and is 

therefore independent of temperature and thus pressure, as well as humidity, though temperature 

determination is humidity dependent. 

The METEK sonic anemometer, like all 3 dimensional sonic anemometers, arranges three pairs 

of transducers such that the three dimensional wind vector can be unambiguously derived for the 

local vector going through the centrum of the sonic anemometers plane. It does this by a non- 

orthogonal arrangement around the vertical axis of the instrument which has a rotational symmetry 

= 120 o , with each path having an angle of 45 o with reference to the horizontal. 

The instrument is set up completely level with a reference point on the axis aligned to magnetic 

North, so that information about the arrival direction of the wind parcel is properly recorded and the 

software provided transforms the information into the Cartesian coordinate system, which is sent to 

the REA system and translated to a switching decision. 

The system utilized had software that allowed it to run for 10 minutes, correcting for the tilt of 

the terrain. The tilt of the terrain creates a flow distortion that will bias results if not corrected for. If 

the instrument is not set up completely level, as turned out to be the case in Barrow, an artificial 

flow bias will be introduced, since the terrain correction program assumes the system is level.


Also, stopping the system every hour, to download and check data, negates the self-learning 

function of the system, thus causing it to have to “re-learn the terrain”, possibly introducing bias. 

The instrument is oriented to be exposed to the dominant wind direction, so that most flow arrives 

undisturbed over the terrain, with the sampling inlets as close to the centrum of the instrument as 

possible. 

With our large heating mantles, we effectively blocked one complete 180 0 wind sector however 

Prior to deployment to the field, we checked in Oak Ridge to ensure that there was no blocking 

effect or attenuation due to formation of a turbulent wake behind the instrument. 

This was done empirically with a blow dryer, observing the visual response to the system, 

pointing the blower at different directions around the system and adjusting the distances on the 

boom so that the system responded with switching as expected. 

We found that we could have the sampling system 1 meter behind the sonic anemometer and 

not effect flow from the 180 0 sector in front of the sampling system plane and assumed there were 

no observable effects from flow distortion due to convection from the heating mantles. However, 

due to the possibility of flow disturbance from the heating mantles, all data coming from behind the 

sampling train was considered suspect. 

The 10 Hz measurements of the vertical wind velocity control fast switching valves on three 

channels. When the wind is an updraft, upward air is sampled at 1 Hz in channel 1, when it is a 

downdraft, downward air is sampled at 1 Hz in channel 3 and when it is neither up nor down (or not 

blowing) it is considered in the deadband “0” and it is sampled at 1 Hz in channel 2. 

The area around 0 is called the dead band and is dynamic, and chosen so each channel is open 

about 1/3 of the time. Defining the deadband is an essential part of sampling control. The purpose 

of the deadband is to increase the conditional mean difference in the trace gases collected in the up


draft and the down draft reservoirs, allowing for the most reliable determination of the difference in 

concentration in the two reservoirs and thus the best flux determination. 

An inherent disadvantage of using a constant deadband is a reduction of the effective sampling 

time for the updraft and the downdraft and thus this can also negatively impact concentration 

determination, and prolong total sampling time required. 

To maximize the advantages of the deadband, the Metsupport REA system uses a dynamic 

deadband, i.e. one which dynamically changes with the wind conditions, adjusting to the standard 

deviation in the vertical wind direction: σw. 

Many of the REA studies use a dynamic deadband around zero vertical velocity (e.g., Oncley et 

al., 1993, Guenther et al., 1996). Some use a constant deadband in order to create a constant and 

stable REA proportionality constant β. Bowling et al., (1999) use CO2 flux to maximize the 

difference in the scalar, not just the turbulence. In theory, there is no significant disadvantage for 

deadband values (wd) smaller than σw i.e. wd/ σw < 1, so it is not a necessity to use a system with a 

dynamic deadband in a system that switches as frequently as the sonic samples. The present system 

samples air flow at 10 Hz and switches at 1 Hz, so, to truly maximize the difference of RGM in the 

up and down channels, and thus lend confidence in the flux measurement, the dead band approach 

is best. 

Once RGM concentrations are obtained for the sampling period, the surface flux F of RGM is 

calculated from equation 6. 

F = βσv(C1-C3) (6), 

where β is an empirical coefficient, the “proportionality constant”, dependent on wind speed 

and turbulence, generally 0.6 for a fixed deadband and approximately 0.3 for a dynamic deadband,


and calculated via the heatflux with the Metsupport system. σv is the standard deviation of the 

vertical wind velocity, approximately 0.18 over snow (Tilden Meyers, personal communication and 

empirical results from this work): both values are obtained directly as output from the REA system; 

C1 and C3 are the concentrations of RGM in upward and downward air masses, respectively. 

From the flux measurements the depositional velocity of RGM can be calculated if the ambient 

concentration for RGM is known: 

vd = F/C (7), 

where C is the concentration of RGM, and F is from 6. The corollary to depositional velocity is 

surface resistance. As depositional velocity increases, surface resistance decreases. The data 

required to calculate surface resistance was not readily available from the Metsupport REA system 

output file, but can be modelled given the depositional velocity and snow conditions. Modeling of 

surface resistance is discussed later. 

Due to the uncertainty of the concentration measurements in Barrow: 10 % and of the 

meteorological measurements, 10 % for β, σw for the Metsupport system (Christensen et al., 2000) 

the uncertainty on the flux measurements using the Landis et al., 2002, sampling end, heating caps 

and Metsupport REA system are estimated to be within 40% on a 95% confidence interval, though 

20% would be a more realistic estimate of denuder precision, given the fact that flow at 10 lpm is 

not instantaneously developed, so there is necessarily some flow and turbulence information lost 

when switching at 1 Hz. This gives a conservative sampling error for flux as 60% for the above 

system in Barrow. 

As a quality assurance check for the REA flux measurements, the total concentration of the 

three annular denuders for each run was compared with ambient concentration a collocated manual


denuder system. Results varied with differences from 3% to 78%. On average, outliers excluded, 

they were within 25% of each other. 

The system as deployed at Station Nord, Greenland with the improved heating caps, software 

and control system is seen in Figure 5., page 53. 

Figure 5. The improved NOAA REA system installed at Station Nord, April 2002 at 3m above 

the snow pack, with the Metsupport REA system deployed as a backup, Note METEK sonic 

anemometer to lower left, heating mantles and actinic flux monitor to the right. Switches, 

pumps and controls in the black box buried in the snow pack to keep from freezing. Author on 

the tower, blocking view of the RM Young sonic anemometer used in the NOAA system. The 

tower was set up during the summer prior to the expedition in May 2002. Terrain with little 

snowdrift, and 1-week-old footprints, attest to the stability of the weather during the 

campaign. Ridges in background are pressure ridges from sea ice on the Wandel Sea. Photo: 

Henrik Skov, NERI 

To summarize: RGM is collected in three heated KCl coated annular denuders as part of the 

RGM flux measurement system, since in the flux system air is collected into two reservoirs: one for 

up drafts and one for down drafts. Each of these therefore requires its own annular denuder


sampling train and heating cap. Sampling was over a 4 hr. sampling period, to minimize the chance 

of break through, a possibility because of the relatively high RGM levels expected in the Arctic 

during mercury depletion events. Break through is what occurs when the active surface is used up, 

or passivated, and can no longer adsorb the gas; thus the gas is said to “break through” the upstream 

end of the denuder and lower than actual levels are measured. 

RGM collected in the denuders is analysed by thermal desorption following the sampling 

period. Manual measurements of reactive gaseous mercury were therefore necessary to obtain the 

mercury mass in each annular denuder in the flux measurement system with the RGM concentration 

in the sampling period for each denuder subsequently calculated based on sampling time 

information and flow rate from the REA system. The difference in concentration found for the up 

channel denuder and down channel denuder are multiplied by two factors obtained for the sampling 

period as output from the REA control system: the standard deviation of the vertical wind velocity 

and the coefficient beta. Flux is then calculated. By dividing with the total ambient RGM 

concentration for the sampling period, concurrently measured with the third channel in the flux 

machine, or obtained as a sum of the three channels, the depositional velocity was calculated. 

For the flux measurements with the REA system, the two collocated denuders acting as 

accumulators for collected RGM over the sampling period must provide faithful representations of 

the amount of RGM in air masses that are either moving up, or coming down, since differences in 

these two denuders are used to calculate the flux. 

REA systems that have used denuders as the trace gas accumulators, e.g., Zhu et al., 2000, 

generally have three denuders in each of the up and mid channel, so that the difference in 

concentration, required for the REA flux calculation, is the difference between the average of the 

concentration determined in three denuders in each channel, providing more confidence to the 

measurement. The present system was a pilot system, with resources for just one annular denuder


for each channel, at best two in the up and down, without using the mid channel. Two denuders 

sampling trains in the up and down channels and none in the mid was not satisfactory. First, an 

average of two measurements does not provide more confidence than just a single measurement, 

second, we would be giving up the ability to indirectly control our measurements as follows: The 

third collocated denuder in the mid channel, sampled in the deadband, so that RGM total ambient 

concentration, could be determined afterwards as the sum of the concentration in the three bands 

and compared with co-located RGM concentration measurements. It is assumed that if these two 

total concentration values were close to each other, then the system was functioning properly. 

However, it is realized that controlling the flux measurement indirectly in this manner does not 

prove that flux measurements were actually representative of the true flux. 

Upon returning from Barrow, the system was taken to Walker Branch watershed for further 

experiments relating to establishing the sampling rate. At a meeting in Oak Ridge in April, 2001, it 

was decided between the author, Dr. Skov, and Dr.’s Meyers and Brooks that NOAA ATDD would 

develop an improved REA system for the collaborative efforts in the Arctic, taking advantage of the 

lessons learned from the Barrow campaign, and improve the software and control system. 

Dr. Brooks brought the NOAA system to NERI in November 2001 where it was intercompared 

with the Metsupport system, and the NERI team was trained on its use. Major improvements 

included software: button type operation, with online graphical data, so that any sampling problems 

are seen during the run, and not post sampling period. The data in run files matched the data in the 

output file. The entire unit except for the sampling is self-contained in just one box. The new 

control box has the capability of hooking up to any laptop or PC with standard serial cables. The 

software can be downloaded on any LINUX operating system PC, alleviating the hardware 

dependency with the previous system. The system is set up to use a different sonic anemometer,


from RM Young. It utilizes the same front-end sampling system, a dynamic deadband, and also 

employs switching at 1 Hz.


3.7 Peat analysis 

The author was the manager and one of the principle investigators for the international and 

interdisciplinary peat project “long term records of atmospheric deposition of Hg, Pb, Cd and POPs 

in the Arctic” with Prof. William Shotyk as the chief scientist. 

Peat localities were chosen, primarily based on pre-existing geological or botanical studies 

for the area, so that any trace metal knowledge gained would be complementary to previous studies, 

and as the data from previous studies would be supplemental to the trace metal studies. 

Once reconnoitred, the sites judged to be the best, i.e., apparently topographically and 

hydrologically isolated, deep, non-disturbed, ombrotrophic type mosses, were sampled, typically on 

the lawn of the peat hummock, i.e., halfway between the hummock and hollow. Peat monoliths 

were taken with a titanium Wardenaar corer (Wardenaar, 1987). 

In the high Arctic, due to the need for automated coring through the permafrost, samples 

were cored from the top of the hummock. Hollow peat was sampled as well, since it represents 

modern growth. Frozen arctic peat does not allow readily sampling, so an automated peat sampler, 

was developed for this work. The version taken to Bathurst improved, by coating with teflon, and 

redesigning the cutting blades and top cap of the bore, prior to being taken to Nordvestø. Details of 

this sampler and its use are given in Noernberg et al, 2003, Appendix C. 

After sampling in the field peat was packaged and frozen for and during transport, where it 

was then processed in the laboratory at either University of Berne, Geological institute, or in the 

case of high Arctic peat, at the University of Heidelberg Institute for Environmental Geochemistry. 

The cores to be analysed for trace elements were sliced while still frozen, every 1 cm and sub 

sampled. 

The "zero" depth of the monolith was taken to represent the interface between the living plant 

material at the top, and the underlying dead plant matter, peat. For high Arctic peat, the coring


started after clearing a block of very freeze-thaw cracked peat representing the active layer, 

typically 8 to 10 cm above the permafrost. 

Individual slices were trimmed of ca. 2.5 cm of edge material for monoliths and 1 cm edge 

material for high Arctic cores, assumed to be contaminated. Those slices where analyses would take 

place, typically every centimeter in the top portion of the core, and every second or third centimeter 

in the lower portion were subsampled using a stainless steel microcorer, ca. 16 mm ID, to recover 

ca. 2 cm 3 plugs from every centimeter slice set to be analyzed. The microcores taken from the 

center of the samples were reserved for dating purposes, to allow for a continuous stratigraphy. 

Microcores taken for mercury analyses were otherwise taken from different quadrants of the slice. 

These cores were air-dried overnight at room temperature in a Class 100 laminar flow clean air 

cabinet, to constant weight. Mercury concentrations were measured in these plugs using the 

methods described below. The Hg concentration profiles presented in the papers in Appendix C, 

except for Goodsite et al., 2002, are made by Dr. Roos, University of Berne, as they expanded upon 

the preliminary measurements made by this author. However, as reported in the papers, this 

author’s measurements are in excellent agreement with the latest measurement series, produced by 

her. Mercury in cores from Nordvestø, Carey Islands has been analyzed in one core, by the author 

and Ms. Nicole Rausch, University of Berne, Institute of Environmental Geology. Additional cores 

from Nordvestø are presently being analyzed at the NERI laboratory and the results are not 

available at the time of this writing. 

Peat samples were analysed for mercury at two laboratories by this author, at the Geological 

Institute, University of Bern, Switzerland, and Institute of Environmental Geochemistry, University 

of Heidelberg. Using solid sample atomic absorption spectroscopy (Salvato and Pirola, 1996) with a 

LECO AMA 254 total mercury analysis instrument. Peat is sliced and sub sampled frozen as 

described later, using care not to contaminate the samples. Sub samples are dried, homogenized and


weighed into Ni sampling boats that have been blanked in the AMA 254 prior to use. Samples are 

then placed into the machine, one sample at a time. They enter a combustion chamber where they 

are dried, then are thermally decomposed in a stream of pure oxygen. Gases from the thermal 

decomposition are then carried in the oxygen stream into a catalyst chamber, which fully 

decomposes the gases at a temperature of 750°C. Mercury is thereafter carried with the oxygen 

stream onto and trapped on a gold amalgamator situated immediately after the catalyst chamber. 

The amalgamator is then heated to 500°C to release the Hg, which is carried with the oxygen stream 

into a cuvette, and mercury determined using atomic absorption spectrometry, with the mercury 

absorption line at 254 nm. 

The detection limit of the instrument is 0.01 ng Hg and the working range is 0.05-600 ng Hg, 

with reproducibility being


a permanent pink colour was obtained, in order to maintain an oxidizing environment and prevent 

loss of Hg. The samples were then diluted with 18 MΩ water to approximately 25 g in polyethylene 

bottles. Mercury concentrations were determined following reduction with sodium borohydride in a 

flow injection AAS system, Perkin Elmer FIMS. The data set obtained using this procedure is in 

good agreement with the data obtained using the LECO AMA 254, on average, for Southern 

Greenland and Denmark, within 15%, r2 = 0.727, n=65. The total mercury data for Nordvestø is 

also being controlled in the above manner. Total mercury in peat from the Faroe Islands was 

controlled at the University of Maine Department of Geology, Dr. Steve Norton, using hydride 

generation atomic fluorescence spectrophotometry after acid dissolution. Each of these 

collaborators also independently measured additional material from the sampling sites as part of the 

total investigation. 

Total lead and other trace elements were determined using X-ray fluorescence on powders. 

Subsamples were dried overnight at 105 0 C in a drying oven, or in the case of high Arctic peat, 

freeze dried, and milled in a centrifugal mill equipped with a Ti rotor and 0.25 mm Ti sieve, 

Ultracentrifugal Mill ZM 1-T, F. K. Retsch GmbH and Co., Haan, Germany. The Energy-dispersive 

Miniprobe Multielement Analyzer (EMMA) using Mo Kβ as the exciting radiation was thereafter 

used to non-destructively measure selected major and trace elements (Pb, As, Fe, Mn, Ti, Zr, K, Ca, 

Cr, Ni, Cu, Zn, Br, Rb, Sr, Y). Calibration, lower limits of detection, accuracy, and precision of the 

method are given elsewhere (Shotyk et al., 2000). 

Additional physical and chemical analyses of the samples were performed to verify the 

trophic state of the peat. These are detailed in the papers in Appendix C and will not be discussed 

here.


Stable lead isotope measurements were also made as a tool to help explain and convince that 

the profiles provide reliable reconstructions of atmospheric mercury and lead. The procedures used 

by the project colleagues who performed these analyses are detailed in papers 8 and 9. 

One of the most difficult determinations to make in sampled peat is dry bulk density, DBD, 

determination. Dry bulk density is essential however to later calculating trace metal accumulation 

rates, or for modeling the dates based on peat accumulation. As will be discussed later, the accurate 

determination of density in peat samples is a subject that requires further research. 

3.8 Age dating of peat profiles 

The 14 C bomb pulse method 

The author completed the development of a direct high resolution dating method, shown to be 

feasible during the authors’ MSc thesis. The method is 14 C based, and uses macrofossils from the 

peat profile. It is published and for details the reader is referred to Goodsite et al., 2002, Appendix 

C. 

The expertise of a paleo-botanist, W.O. van der Knaap, Univeristy of Bern, Institute of Plant 

Science, University of Berne, or paleo ecologist, Ole Bennike, Geological Survey of Denmark and 

Greenland (GEUS), in the case of the Carey Islands, was necessary for appropriate selection of 

plant macrofossils identified in sub samples from the cores where trace metal analysis would be 

carried out. 

To prevent modern fungal growth, the macrofossils were processed within one week of sub- 

sampling, at the AMS 14 C Dating Laboratory, University of Aarhus, using a standard procedure for 

plant material: washed, acid-base-acid treatment, this author performed the analysis for the samples 

from Denmark and Greenland. 

Percent modern carbon in the macrofossils were determined using 14 C Accelerator Mass 

Spectrometry (AMS). By comparing the measured percent modern carbon in the samples to a 

calibration curve of atmospheric and terrestrial measurements for the northern most northern


hemisphere (Goodsite et al., 2002) individual samples more recent than AD 1950 could be dated 

directly with high time resolution, ± 2 years. Samples that stopped respiration between this period 

and 1995 have elevated levels of carbon 14, since thermonuclear testing led to elevated 14 C levels in 

the atmosphere. Using AMS, these levels can be directly measured and correlated with the curve, 

not utilizing the conventional method of half-life decay of 14 C. 

Selected samples from deeper layers, dating from before AD 1950, could not use this method 

and were AMS 14 C dated by the usual tree-ring calibration method, with much larger uncertainties 

to follow or by 210 Pb dating. 

Pb-210 dating 

1,5 grams of solid sample powders along with water content and dry bulk density information 

were sent to Professor Peter G. Appleby, University of Liverpool, for age dating using the 210 Pb 

constant rate of supply model (Appleby and Oldfield, 1978). This method results in a model 

providing chronological information spanning the last approximately 200 years. Further details are 

discussed in the papers where employed. Comparison of 210 Pb age dating with the Goodsite et al., 

2002 atmospheric bomb pulse method of the same set of peat samples, is found in the paper, 

Appendix C.


4. Results and Discussion 

Laboratory experiments 

The laboratory experiments were carried out in the lab of Dr. Steven Lindberg, in the 

Environmental Sciences Division, Oak Ridge National Laboratory, USA. The building has a 

common ventilation system, turned off after work hours to save energy, and many scientists in the 

building have used mercuric chloride for their experiments. Therefore RGM levels are 

approximately twice as high as what is expected for ambient, though comparable with 

measurements at Station Nord, just prior to an AMDE. None of the denuders were heated in the 

laboratory; the denuders were at room temperature, approximately 23 0 C. 

In Figure 6., page 64, the annular denuders were hung in parallel, in the laboratory; and allowed to 

sample for 1 to 2 hours during the day and for 16 hours during the night and into the weekend. 

Concentration levels may be slightly underestimated since the denuders were not heated for the 

experiments. The figure shows that denuders that are 1 year old, X1 and X2 appear to be 

functioning just as well as denuders that are new: numbers 1-5. Reproducibility is within 10%, 

except for the first experiment. Denuders were cleaned and recoated after each use. The results 

show that under laboratory conditions, where HgCl2 is expected to be the dominating RGM species, 

that non-heated KCl coated denuders are able to collect a significant amount of RGM at ambient 

temperature in a reproducible manner.


pg Hg/m3 

140 

120 

100 

80 

60 

40 

20 

0 

1:40 

Threw out X1, could not 

obtain a low zero prior to 

development 

2:00 

Active sampling 

16:00 (Night) 

50 min 

40 min 

1 2 3 4 5 X1 X3 X2 1 2 4 5 

2 denuders per run, Sampling time as noted 

16:00 

(Night) 

Figure 6. Parallel determination of reactive gaseous mercury concentration over one week inside of Dr. 

Steven Lindberg’s mercury laboratory, Environmental Science Division Building, Oak Ridge 

National Laboratory; during night time and daytime, January 2001. RGM is collected over 

non-heated annular denuders; with a 10 lpm flow rate. The times shown are hours or minutes 

exposed. Method used is by Landis et al., 2002, except that heating mantles were not 

employed.


Pg Hg/hr 

0,14 

0,12 

0,1 

0,08 

0,06 

0,04 

0,02 

0 

One cap off Horiz. 

72 hrs. 

Passive absorption rate 

Filt. Packs, Impactors, Vertical 

24.5 hrs. 

Both caps off Horizontal, 

Night 

17 hrs 

Caps, impactors, 

Vert. 

217 hrs. 

1 2 3 4 1 3 X3 

2 denuders per run. Run 1 was over the weekend, 2:3 was 

one complete day, 4:1 was overnight, 3:X3 was passively 

sampled ca. 9 days 

Figure 7. Annular denuders were passively exposed, both in vertical and horizontal positions, with and 

without end caps to laboratory air in Dr. Steven Lindbergs’ Laboratory, Environmental 

Sciences Division, Oak Ridge National Laboratory in January, 2001; average concentration, 

60 -80 pg m -3 is a factor ten lower than in the Arctic under mercury depletion. The run with 

denuders 2 and 3 are set up, as denuders would be in the field for RGM determination. 

Comma used as a decimal holder. 

The relaxed eddy accumulation system is set up by virtue of its dynamic deadband to sample 

each channel approximately 1/3 of the total run time. This means that for 1 channel during a total 

run of 6 hours, the annular denuder will be actively sampled 2 hours, and the other 4 hours, exposed 

to possible passive uptake. Therefore, it was important to quantify any passive uptake in the 

denuders. After obtaining an analytical zero, annular denuders were tested horizontally and 

vertically with and without end caps or filters for passive uptake. They were deployed without 

pumping for a number of hours, taken down and analysed. Passive uptake was defined as the total 

mass minus the blank value, divided by the number of hours deployed. Results were variable, but


generally reproducible as seen with the parallel measurements. The highest passive uptake resulted 

from denuders deployed in the total sampling train, from inlet to filter pack with filter, exposed for 

24.5 hrs in the laboratory, during the workweek, non heated. They were started in the morning and 

analysed the next morning, resulting in a passive uptake of 0.12 pg Hg per hour. 

80 

70 

60 

50 

40 

30 

20 

10 

0 

RGM mass (pg Hg) 

6 1 2 3 4 Average Std. Dev. 

Figure 8. Mass of RGM collected concurrently in 4 annular denuders, 2 co-located parallel measurements, 

in the laboratory of Dr. Steve Lindberg, Oak Ridge National Laboratory. 

Figure 8., page 66, is a reproducibility experiment run with 4 co-located, non-heated denuder 

sampling trains. Denuders 1 – 4 are actively sampled while denuder 6 is blanked handled and hung 

without any sampling. Denuders are connected to a single pump via 2 T-type connectors as flow 

splitters. Flow was controlled in each inlet with a mass flow meter. Denuder 6 is the field blank and 

reported concentrations have the filed blank subtracted. Of the 4 denuders, only denuder three had a 

value less than the three other denuders, 10% error taken into consideration.


pg/m 3 

pg/m 3 

90.00 

80.00 

70.00 

60.00 

50.00 

40.00 

30.00 

20.00 

10.00 

Cold and warm heating Mantles, RGM measurements 2002 

Cold heating mantle 

Warm heating mantle 

0.00 

16-apr 17-apr 18-apr 19-apr 20-apr 21-apr 22-apr 23-apr 24-apr 25-apr 

Time 

Figure 9. Parallel measurements of RGM at Station Nord, Greenland. Warm denuders were +50 0 C cold 

denuders were at ambient temperature, approximately -30 0 C each point represents the 

averaged concentration of two denuders. Note: Results near 0.00 pg m -3 cover each other, for 

cold and warm denuders. 

In Figure 9., page 67, it is seen that the annular denuders heated to 50 0 C using the improved 

heating caps, give higher RGM concentration values, when RGM concentrations are higher than 

approximately 10 pg m -3 and when there is apparently little, 5 pg m -3 or no RGM, the heated or non- 

heated denuders report the same amounts. Each point represents the average of parallel 

measurements. 

Controlling the heating of the denuders is not just important in the field, but as seen from Figure 

10., page 68, also important during laboratory analysis. Landis et al., 2002 note the need for coaxial 

fans and good insulation to limit the heat escaping from the clamshell denuder oven and warming 

the teflon sampling lines connecting the denuder to the charcoal filter on the inlet side, and the 

mercury monitor on the outlet side. Figure 10., page 68, is an experiment where a single denuder,


denuder 1, is analysed for its blank value, with and without use of coaxial fans. Without use of fans, 

mass values are seen to increase, probably as a result of ambient uptake through the non-cooled, 

therefore warm to the touch, teflon sampling lines. 

It is seen that heating or non-heating of the denuders may bias the distribution. In the field the 

denuders should be heated to a constant temperature, especially for parallel measurements, or the 

values may not be comparable. Such is the case as well for intercomparison of data between two 

Arctic sites using different heating systems. In the laboratory, the method of Landis et al., 2002 

must be followed to avoid introducing sampling artefacts. 

Pg RGM (mass) 

3 

2,5 

2 

1,5 

1 

0,5 

0 

Analytical Zero test 

1 prior 1 1 1 1 1 

Denuder 1 sequentially run (once temp. fell to 50 deg. C) from left to right, not removed from machine 

Figure 10. Denuder nr. 1 analysed for analytical blank value in series with and without coaxial cooling. The 

first result shown is from the previous analysis of denuder 1 prior to the experiment.; the next 

two are without using a coaxial heating fan, the next two with, and the last two without. The 

denuder was not removed from the oven. The oven was allowed to cool to 50 0 C prior to start 

of next analysis.


4.1 Ozone and GEM Measurements 

The results of ozone and GEM measurements for Station Nord are shown in Figure 11, page 

69. It is seen that ozone is relatively stable from September/October until the end of 

February/beginning of March, then a highly perturbed period occurs, where ozone and GEM are 

simultaneously depleted to 0, from respectively about 40 ppbv and 1.5 ng/m 3 within hours. They 

remain at 0 for up to several days when they suddenly rise again, to levels above normal ambient 

levels seen in July. In July the ozone concentration stabilises slightly above 20 ppbv followed by a 

slow increase to about 40 ppbv in September/October. 

GEM was only measured from February to the end of July or beginning of August. The 

measurements were focusing on the description of the AMDE. Previously Schroeder et al. (1998) 

Ozone, ppbv 

60 

50 

40 

30 

20 

10 

0 

1999 

2000 

2001 

Time 

Ozone 

GEM 

1/10*fBr 

Figure 11. Hourly ozone mixing ratios and weekly concentration of filterable bromine, fBr, 

measured from 1999 to 2002 at Station Nord, Northeast Greenland. GEM is measured in the 

period from 25 September 1999 to 23 August 2000; 14 February 2001 to 23 August 2001 and 

26 April to 29 June 2002. In general the half-life of GEM at Station Nord is approximately 10 

hours during AMDE. From Skov et al., 2003, submitted, Appendix C. 

2002 

2003 

3 

2 

1 

0 

GEM,Br, ng/m 3


have described that ozone and GEM are simultaneously depleted and that their depletion are highly 

correlated. The Station Nord data confirms this, see Figure 12., page 70, After the depletion period 

some very high concentrations of GEM appeared with values up to above 2 ng/m 3 and in 2002 up to 

4.5 ng/m 3 . In 2001, similar observations were made at Barrow (Lindberg et al. 2002, Appendix C). 

The high values after a depletion event are attributed to re-emission of mercury to the atmosphere 

(Lindberg et al., 2002). The interruption of the GEM time series at Station Nord in the middle of the 

summer makes it difficult to further interpret the importance of re-emission at Station Nord. 

GEM, ng/m 3 

2 

1.5 

1 

0.5 

0 

-0.5 

y = 0.039x - 0.095 

R 2 = 0.800 

0 10 20 30 40 50 60 

Ozone, ppbv 

Figure 12. GEM against ozone concentrations at Station Nord, Northeast Greenland including 

a regression line obtained by orthogonal regression analysis. Data was selected from Fig. 11 

where at least 3 consecutive concentrations were decreasing on both ozone and GEM and the 

initial GEM concentration was above 0.4 ng m -3 . Only data from 2000 and 2001 were used as 

high concentrations in 2002 indicates the presence of other processes than in 2000 and 2001. 

The R 2 of approximately 0.8 was typical for the 2000 and 2001 campaigns. From Skov et al., 

2003, submitted, Appendix C.


4.2 RGM concentrations and Flux 

Figure 13., page 71, shows the result of the RGM measurements at Station Nord in 2002. The 

measurements were made by manually sampling and using the improved heating caps, with 

temperatures in the heating caps approximately 40 0 C, or some 80 0 C above ambient. Bar thickness 

is proportional to exposure time. The thicker bars in the histogram were exposed more hours than 

the thinner bars. Data is field blank corrected. The concentration varies between values below 

detection limit and up to 75 ng m -3 . These values are comparable to those found at Barrow before 

the start of strong AMDE’s (Lindberg et al. 2002) and very typical for calm weather conditions. 

pg/m 3 

Pg m-3 

Figure 13. Manual RGM measurements made during April,2002 at Station Nord 

80.00 80 

70.00 70 

60.00 60 

50.00 50 

40.00 40 

30.00 30 

20.00 20 

10.00 10 

0.00 0 

7-apr 9-apr 11-apr 13-apr 15-apr 17-apr 19-apr 21-apr 23-apr 25-a 

Time Date 

The relaxed eddy accumulation system operating at 10 m above the tree top canopy on the 

tower at Walker Branch Research Area, Oak Ridge showed no mass uptake in the up or down tubes 

operating at 10 Hz; n = 3. The frequency was adjusted to 1 Hz and the measurable flux; n=3; 1 Hz


and the additive product of the denuders, up, mid, down were within 25% of the constantly sampled 

denuder. There was a good mass balance between the up and down channel. 

The results of the 2001 campaign in Alaska (Barrow Arctic Mercury Study (BAMS)) are 

shown in Table 1. page 73 and Figure 14 page 77. The largest deposition velocity in 2001 was 

approximately 2.8 cm s -1 and the average was close to 1 cm s -1 . Concentrations of the three denuder 

tubes, up, mid and down, were added together to determine the total concentration, and 

concentration was compared with the on site Tekran automatic RGM monitor, MODEL 1130, 

described in Lindberg et al., 2002, Appendix C. Since the machines were not running absolutely 

simultaneously, a linear interpretation was made to compare concentrations for the same time 

frame. The heating system for the annular denuder heating system for the 1130 kept the denuders 

warmer than the heating mantles on the REA system. Not taking run two into account, since a 

valve was frozen open, it is seen that the percent difference varies, with an average for the 

campaign of 24% with a standard deviation of 42. Percent difference between the two 

measurements was as small as 3% and as large as 78%. It is seen that in all but 2 of the runs, run 2 

excluded, the RGM REA total is less than the 1130 value, perhaps because of differences in the 

heating mantle system or in differences of flow calibration between the two systems. 

The REA system when properly on the tower, after a period of time should report the same 

number of counts per up and down channel, due to conservation of mass, and did so prior to 

deployment in Barrow at Walker Branch. In Barrow, there was always a greater number, up to 30% 

more, of counts on the down channel, suggesting that the sonic anemometer was not positioned 

properly. Therefore the channel count data was corrected to ensure mass conservation as follows: 

1) Corrected counts in Down = registered counts in Down - (registered counts in Down – counts in up). 

Control: total counts for all channels registered = total counts with down corrected. 

2) Corrected Counts for Mid = registered Counts for Mid - (Registered Counts for Down - corrected 

counts for Down); The lost mass must be placed in mid channel for later concentration calculation. 

Control: Corrected counts Mid > registered counts mid, but total counts all channels still the same. 

3) Corrected Counts Up = Old Counts up Up, since down was preferred during sampling.


4) New Total Volume for each denuder = New count * 10 l per second (fixed) 

Control: registered total Volume = Corrected total volume 

Control: registered total mass, RGM in ng Hg (0) = new total mass RGM, ng Hg (0). 

6) New concentration: New mass / New volume 

Run dates and time, reported as Greenwich Mean Time, not local time, in accordance with 

the other monitors at CMDL Barrow. The sonic anemometer consistently showed an ambient 

temperature of 2 degrees higher than other ambient temperature instruments at CMDL. This 

indicates that one ore more of the heads was slightly out of line after shipping. This should not 

affect the measurements since the proportionality constant, β, is based on the heat flux, so it is the 

difference in temperature that matters. This is assuming, that there was no significant heat flux. 

Table 1. RGM mass raw data and accumulated concentrations corrected for blank values, 1130 

are automatic RGM concentration measurements from the NOAA TEKRAN speciation unit, for 

comparison to REA total as % difference, run 2 not included in compared avg. and std. deviation. 

Date time group RGM mass (pg) Concentration Blank. Corrected (pg/m3) 

Run Dt/time(Z) Down mid up down mid up Total 1130 % difference 

1 29 2200-30 0100 MAR 29,3 15,3 12,3 67,7 19,0 33,7 121 126 4 

2 30 1800-30 2200 MAR 28,2 353,9 24,7 44,6 310,6 46,7 402 58 (-593) 

3 02 2000-03 0000 APR 31,0 23,7 21,0 39,9 23,2 42,2 105 93 -13 

4 04 2015-05 0015 APR 11,8 22,9 9,4 17,4 22,5 15,6 56 65 14 

5 05 1839-06 0039 APR 31,0 56,4 27,2 33,8 32,7 31,7 98 448 78 

6 07 1850-08 0050 APR 61,4 92,6 38,6 57,5 58,0 46,2 162 99 -64 

7 08 2030-09 0030 APR 25,5 24,6 13,7 33,5 23,8 27,0 84 172 51 

8 09 1804-10 0104 APR 44,8 68,1 35,9 35,9 36,2 36,9 109 335 67 

9 10 1839-10 2339 APR 29,9 41,7 22,7 33,5 30,3 36,1 100 166 40 

10 11 2000-12 0000 APR 27,1 34,8 16,4 39,8 32,8 29,4 102 121 16 

12 12 2100-13 0000 APR 26,7 41,8 21,4 31,2 30,0 32,8 94 97 3 

AVERAGE 31,5 70,5 22,1 39,5 56,3 34,4 24 

Std. Dev. 12,47 96,72 9,26 13,49 84,99 8,97 42 

11 (night) 12 0800-12 1100 APR 4,80 6,70 2,2 6,4 6,4 4,5 17 48 65 

Table 2, page 75, shows other average data for the run, as recorded by the sonic 

anemometer. The clean air sector for NOAA, CMDL, Barrow, Alaska is defined as wind arriving


from the coast, as opposed from the town of Barrow, and is 45 0 to 135 0 . The system was oriented to 

360 0 . The way the data was reported by the system was not sufficient for in depth analysis, but 

sufficient enough for the pilot measurements. 

In the new REA system, there are graphs correlating temperature, wind speed and direction 

with the turbulence. This is an important improvement since little can be done with average data for 

a run, since gusts for example, may carry much of the RGM mass from the coastline, while the 

prevailing winds are coming from another location. This will not be seen when simply taking 

averages into consideration. Due to the nature of micrometeorological measurements, the REA 

system works best at wind speeds near 5 m s -1 since there needs to be good turbulence to sample. 

Wind speeds less than 2 m s -1 are nominal. During the campaign, on average, the wind is coming 

outside of the CMDL defined clean air sector. This means that during the measurements, the wind 

was uncharacteristically coming from the town of Barrow. There are no known sources of RGM or 

elemental mercury in the town, so this should not have affected the measurements. The standard 

deviation in the vertical wind component is actually much higher than expected. It was expected 

that we would find 0.18, with little variability, given the stable terrain, and weather conditions, and 

instead found a campaign average value of 0.29 with a standard deviation of 0.11. The 

proportionality constant was also higher than the 0.3 expected, though lower than 0.6, as it should 

be, for a dynamic deadband. The campaign average was 0.41 with a standard deviation of 0.01. 

Indicating no significant variation in the heat flux. The machine reported all numbers to four 

decimals; results have been rounded to two.


Table 2. Average temperature, wind speed, wind direction, standard deviation of the vertical wind 

velocity and the proportionality constant based on heat flux as reported by the sonic anemometer 

result file runname.txt. Temperature was 2 0 higher than what other on site instruments reported. 

DAYHOURMON avg.temp avg.WS avg.Wdir In σv β 

Run Dt/time(Z) deg. C m/s deg Sector? average Average 

1 29 2200-30 0100 MAR -22,2 2,6 160 N 0,27 0,41 

2 30 1800-30 2200 MAR -22,6 5,5 132 Y 0,19 0,41 

3 02 2000-03 0000 APR -21,4 2,5 236 N 0,25 0,41 

4 04 2015-05 0015 APR -14,8 7,9 15 N 0,27 0,42 

5 05 1839-06 0039 APR -17,2 7,8 65 Y 0,26 0,41 

6 07 1850-08 0050 APR -9,3 5,9 93 Y 0,24 0,42 

7 08 2030-09 0030 APR -6,3 4,9 208 N 0,59 0,39 

8 09 1804-10 0104 APR -12,3 4,9 210 N 0,40 0,42 

9 10 1839-10 2339 APR -9,6 6,8 51 Y 0,27 0,41 

10 11 2000-12 0000 APR -8,8 5,2 41 N 0,18 0,42 

12 12 2100-13 0000 APR -8,5 3,3 247 N 0,25 0,41 

11 

0,11 0,42 

(night) 12 0800-12 1100 APR -8,7 2,7 36 N 

After correcting the mass, and taking the difference between the down and the up channel, 

then multiplying with the proportionality constant and the standard deviation in the vertical wind 

turbulence, the flux is calculated, as shown in Figure 14, page 77. The values for uncorrected mass 

are also shown. If the forcing was natural, then this assumes that the system was set up correctly, 

but that since, on average the wind came from the city, and hence flowed over the CMDL building 

prior to meeting the tower, that the wind flow was forced primarily downward. Except for the first 

measurement, the trends appear to be the same. Depositional velocity was found by dividing the 

flux with the sum of the concentrations in the denuders in the three channels. The depositional 

velocity for reactive gaseous mercury is approximately 1 cm s -1 for the mass corrected data and 

approximately 0.5 cm s -1 for the non-mass corrected data. Average depositional flux for Barrow 

was 1.3 ± 0.7 ng m -2 h -1 for the mass corrected and approximately half that for non-mass corrected 

values. For comparison purposes, Schroeder et al., 1998 used a dry depositional velocity of 0.5 cm s - 

1 when estimating for Alert, an average springtime dry-deposition flux for mercury of 2.5 ± 0.5 ng


m -2 h -1 based on their measurements of TGM. Flux and depositional velocities are in Table 3, page 

76, and Figure 14., page 77. 

Table 3. Results of REA RGM vertical flux measurements, Barrow, Alaska, 2001; with computed 

dry depositional velocities for RGM. Runs 1 and 7 are excluded due to outlying depositional 

velocities, indicating a problem with micrometeorological measurements. Run 2 is excluded since 

the mid-channel froze open, and the mass balance could therefore not be corrected. 11 excluded 

from average, since it is a nighttime run. 

Vertical Flux (Fc) Vd 

Run Nr. Date time GMT (pg m^-2 s^-1) ng/m2/hr m s^-1 cm s^-1 

1 29 2200-30 0100 MAR -5,6 -20 -0,2 -15,7 

2 30 1800-30 2200 MAR 4,57 16,44 0,03 2,58 

3 02 2000-03 0000 APR -0,8 -2,8 0,0 -2,4 

4 04 2015-05 0015 APR -0,2 -0,6 0,0 -0,8 

5 05 1839-06 0039 APR -0,3 -0,9 0,0 -0,8 

6 07 1850-08 0050 APR -1,2 -4,3 0,0 -2,2 

7 08 2030-09 0030 APR -2,79 -10,05 -0,1 -10,06 

8 09 1804-10 0104 APR 0,2 0,7 0,0 0,6 

9 10 1839-10 2339 APR 0,1 0,4 0,0 0,4 

10 11 2000-12 0000 APR -1,0 -3,5 0,0 -2,8 

12 12 2100-13 0000 APR 0,2 0,6 0,0 0,6 

11 (night) 12 0800-12 1100 APR -0,08 -0,3 -0,01 -1,41 

Average Excl. 1,2,7,11 -0,4 -1,3 0,0 -0,9 

Std. Dev. Excl. 1,2,7,11 0,6 2,0 0,0 1,4


Figure 14. Surface flux of RGM to the snow pack on a 3 m tower, at Barrow, Alaska, spring, 2001. 

RGM vertical flux (pg m-2 s-1) and 

Depositional Velocity of RGM cm/s 

1,5 

1,0 

0,5 

0,0 

-0,5 

-1,0 

-1,5 

-2,0 

-2,5 

-3,0 

-3,5 

02-apr 

04-apr 

RGM flux: BAMS 2001 

05-apr 

07-apr 

08-apr 

09-apr 

Date, average; 

error bars reflect 1 standard error 

10-apr 

12-apr 

Vertical Flux (Fc) (pg m^-2 s^-1) 

Depositional Velocity (Vd) cm s^-1 

Vertical flux without mass correction 

Depositional velocity without mass correction 

Average 

For the air surface exchange, the deposition is dominating but periods were observed with 

emission of RGM as well. 

4.3 High resolution dating of peat archives and mercury in peat 

Nuclear testing raised the levels of 14 C in the atmosphere, test ban treaties stopped this 

anthropogenic production of 14 C and levels began to fall. To take advantage of this knowledge as a


dating tool, it was necessary to create a calibration curve valid for the northernmost northern 

hemisphere. The method relies on the fact that living moss material, later becoming peat, has taken 

up 14 C levels equal to the amount found in the air or similar archives during its growth season. This 

is the same principle as radiocarbon dating, except where classical radiocarbon dating employs the 

half-life of 14 C, the present method employs knowledge of the amount of 14 C present in the 

sample. Figure 15., page 79, shows the atmospheric bomb-pulse calibration curve, in percent 

modern carbon units, pMC, created by accumulating previously published and new terrestrial and 

atmospheric 14 C data for the period of interest from the northern most northern hemisphere, 30 o -90 o 

N latitude band. The inset figure shows the global input by the atmospheric nuclear weapons’ tests 

measured as TNT energy equivalent, and the corresponding uptake by the biosphere and oceans 

using an exponential decay time of 18.70 ± 0.15 yr (Levin and Hesshaimer 2000). The curve is 

valid from 1954 to the present. Near the top of the curve, 1963, the date has a larger uncertainty 

than other portions of the curve. The curve otherwise provides a date directly determined by the 

measured amount of 14 C in the sample to ± 2 years.


Figure 15. The atmospheric bomb pulse northernmost northern hemisphere calibration curve from Goodsite et al, 

2002, Appendix C. 

Peat cores are sliced in 1-centimetre slices. In each slice, a macrofossil of peat moss is removed 

and processed for determination of the level of 14 C, expressed as percent modern carbon, pMC, by 

accelerator mass spectrometry. As one goes deeper in the profile, 14 C levels rise to a peak of 

approximately 180 pMC. This level is where the 14 C bomb pulse peaked in 1963. Levels then begin 

to fall until the early 1950’s. There are a couple of features in the curve related to the temporary 

stop in testing. Figure 16., page 80, shows that peat profiles from Denmark and Greenland could 

reproduce the calibration curve without dampening. Dampening is when the top of the curve is 

found to be much lower than 180 pMC.


Figure 16. Samples from peat cores from Denmark and S. Greenland reproduce the bomb pulse curve well. From 

Goodsite et al., 2002, Appendix C. 

Atmospheric Hg accumulation rates in southern Greenland 

The dated profile can be analysed for trace elements or for paleo-ecological purposes and can 

be used to calculate an age-depth model to estimate peat accumulation rates. For example by using 

the dates determined by the bomb pulse dating method, from AD 1950 to ca. 1976, the 

accumulation rate was 0.68 cm yr -1 , and since ca. 1976 the rate has been 0.20 cm yr -1 . 

The age depth model allows the atmospheric Hg accumulation rate to be estimated as the 

product of the volumetric Hg concentrations, ng cm -3 , Figures 17 a and b, page 83, and the peat 

accumulation rate, cm yr -1 , when re-emission is not taken into account. Figure 18 a and b, page 84. 

Results for deeper portions of the profile, see Shotyk et al., 2003, Appendix C, show that pre- 

industrial Hg accumulation rate ranged from 0.3 to 3 µg m -2 yr -1 when clearly minerogenic 

enrichments are disregarded, such as seen at the bottom of the south Greenland profile at a depth of


approximately 80 cm. This then would be an indication of an upper limit for natural flux values. 

Since this is a minerogenic peat, with geogenic input, this rate is expected to be higher than 

deposition due to atmospheric supply alone. 

From Figure 18a. page 84, it is seen that the accumulation rate in south Greenland reached its 

maximum in 1953 at 164 µg m -2 yr -1 , where it thereafter falls and begins to rise again in the 70’s 

with a fall in the late 80’s. 

The value in 1995: 14.1 µg m -2 yr -1 is in excellent agreement with the deposition predicted by 

the Danish Eulerian Hemispheric Model, DEHM: 12.0 µg m -2 yr -1 for South Greenland (Christensen 

et al., 2002, Skov et al., 2003 submitted). 

Present day values appear to be declining and are presently an order of magnitude higher than 

pre industrial values at the conservative limit, and a factor 3 to 4 at the positive limit. 

Atmospheric Hg accumulation rates in southern Denmark 

By using dates determined by the high resolution dating method (Goodsite et al., 2002, 

Appendix C) an average peat accumulation rate from AD 1950 to AD 1980 was 0.47 cm/y, and 

since AD 1980 was 0.21 cm/yr (Shotyk et al., 2003, accepted, Appendix C). 

Using these peat accumulation rates, the net atmospheric Hg accumulation rate was calculated 

in the same way as for south Greenland, by multiplying the volumetric concentration by the peat 

accumulation. 

In Figure 18b. page 84, it is seen that in Denmark the maximum Hg accumulation rate was also 

found in 1953, at 184 µg m -2 yr -1 . The overall trend in Denmark is the same as that seen in 

Greenland. For 1994 the deposition of 14 µg m -2 yr -1 is very comparable to the Danish Eulerian 

Hemisphere Model prediction of 18 µg m -2 yr -1 in 1995 (Christensen et al., 2002, Skov et al., 2003 

submitted, Appendix C).


After coring, analysis showed that the Danish core had been disturbed by peat cutting from 

during WWII and prior, so it is difficult to conclusively interpret the profile. It is however evident 

that pre-industrial levels were much lower than post WWII levels. It is difficult to find a non- 

disturbed location in Denmark since surface signs are not always evident, and historical records of 

peat digging not always complete. 

In both cores, the error associated with the Hg accumulation is calculated to be 21%, based on 

conservative estimates of the errors associated with the 14 C bomb pulse curve age dates, ca. 5%; Hg 

concentrations, ca. 5%; and bulk density measurements, ca. 20%. 

One would expect the bulk density to increase rather smoothly with depth in the core as the 

peat becomes more humified and compacted under its own weight. However, in practice, bulk 

density determination now introduces the greatest source of error into accumulation determination 

as seen from the variation in the bulk density as shown in Figure’s 17 a and b, page 83, and the 

error calculation shown above.


Figure 17a and b. Mercury results from Southern Greenland and Denmark, from Shotyk et al., 2003, Appendix C.


Figure 18a Mercury accumulation rate for minerogenic peat from South Greenland. Accumulation rate as 

determined by the high-resolution bomb pulse dating (Goodsite et al., 2002, Appendix C) upper and 

lower limits depicted. Curve as submitted in Shotyk et al., 2003 (accepted), Appendix C 

Figure 18b Mercury accumulation rate from an ombrogenic peat profile from a raised bog in Denmark. 

Accumulation rate as determined by the high-resolution bomb pulse dating (Goodsite et al., 2002, 

Appendix C). Upper and lower limits depicted. Curve as submitted in Shotyk et al., 2003 (accepted), 

Appendix C.


Figure 19. Mercury in a peat profile from Myrarnar bog, Faroe Islands, Shotyk, Goodsite et al., unpublished 

preliminary results (manuscript in preparation). Coloured markers are used for discussion and 

notational purposes while manuscript is being prepared. 

Atmospheric Hg accumulation rates on the Faroe Islands 

Peat on the Faroe Islands was very decomposed and compact, already at the surface. Therefore, 

high-resolution dating was not possible, making the profile much more difficult to interpret and 

compare with the Danish and S. Greenland core. The large amount of minerogenic material made it 

possible however to use the lead-210 constant rate of supply model. Figure 19. page 85, shows the 

peak of mercury concentration to be in the 1954 layer, within reason and error, at the same time as 

the maximum in S. Greenland and Denmark. There are also discrete mercury enrichments in the 

profile that may be associated with volcanic ash fall. Figure 20., page 86, shows the mercury 

accumulation rate.


Figure 20. Preliminary mercury accumulation rate on the Faroe Islands. Unpublished data (manuscript in 

preparation by Shotyk, Goodsite et al). a. is the total profile, b looks in detail at the deeper 

depths. The arrows denote discrete ash layers, deposited from eruptions from nearby Icelandic 

volcanoes. c. is the uppermost 15 cm of the profile with the background value plotted for 

comparison purposes Dates determined by 210-Pb constant rate of supply model. 

The mercury accumulation rate cannot achieve the same detail, given the fact that a dating model is 

used and the peat is so compact and dense that a one-centimetre slice may represent 5 to 10 years of 

growth already towards the top of the profile. Never the less, the same trend is observed, with 

mercury accumulation peaking approximately 20 years ago, and since falling. The back ground 

accumulation rate of approximately 1 µg m -2 yr -1 is approximately the same as that found in 

southern Greenland and Denmark. The total loading, that is, the total amount of mercury considered 

with the number of years of sedimentation is not higher than the other two locations studied, though 

on initial inspection, one notes very high concentrations, and an accumulation rate over 30 times 

higher than background.


Total Hg ng/g (freeze dried weight) 

45 

40 

35 

30 

25 

20 

15 

10 

5 

Nordvestø Carey Islands, Greenland 

M.Goodsite/N. Rausch 

4520 BP 5065 BP 5414 BP 5940 BP 

0 

0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 200 210 

Depth (cm) 

6205 BP 

Figure 21. Preliminary results for mercury for Nordvestø, Carey Islands, Greenland. The permafrost starts 

at 8 cm. Peat was sampled using a new rotary corer developed for frozen peat (Noernberg et 

al., 2003, Appendix C) Prelimary dating results show that the peat had a very linear 

accumulation, and that core represents approximately 1700 years of peat accumulation from 

the mid Holocene, with a 1 cm slice representing approximately 10 years of peat 

accumulation. Unpublished data. 

The mercury concentrations found in the peat profile from Nordvestø, Carey Islands, NW 

Greenland are shown in Figure 21., page 87. The study of peat from the Carey Islands is still 

ongoing at the time of completion of this Ph.D. The active layer extends to the depth of 8 cm. The 

core is just over 2 meters long and preliminary dates show that the core covers the mid-Holocene. 

The peat formed from moss that grew on bird droppings from seabirds that sat on the cliff, as 

evidenced by feathers found in each centimetre layer. The feathers have been sent to the Danish 

National Zoological Museum for species determination. There are very distinct discrete peaks in the 

core, for example between 120 and 130 cm deep. The concentration trend seems to be rising. The


highest peak is found in the active layer, though chronologically dating this layer, which is exposed 

to freeze thaw cycles, cracking and physical transport is not realistic, since this sediment is very 

cracked and disturbed. The concentration value is however a factor three to four over what is seen 

in deeper layers. This must be assumed to be due to anthropogenic input. At the very first slice, 

where there was little green growth, the concentration was as low as values deeper in the core. This 

core is challenging to analyse, since the common trace metal used for comparison, lead, is found in 

values under the detection limit in the core. Analyses are being carried out for cadmium to help 

interpret the signal.


Discussion 

It is important to have reliable RGM concentrations because the RGM levels represent the 

amount of GEM oxidised and thereafter available for fast deposition from the Arctic troposphere 

during mercury depletion events, with resultant wash out to the marine environment once the snow 

melts. We are in need of reliable RGM observations in order to accurately measure flux, verify 

atmospheric models e.g. DEHM on both a temporal and spatial scale. Experimental data from this 

work confirms that whether deployed in the laboratory, or the Arctic, the annular denuder method is 

one that when deployed in accordance with Landis et al., 2002, reproducibly measures the 

operationally defined RGM with a 15% precision within 3 standard deviations. 

RGM annular denuder tests 

Annular denuders were tested under a variety of circumstances to determine how well the 

denuders could reproduce the same measurement and if they were subject to any passive uptake. 

The denuders are hand crafted, and are fragile, therefore the denuders are inspected and categorized 

prior to use. 

The RGM annular denuder tests performed in Oak Ridge at Station Nord taught many valuable 

lessons about the operational measurement of reactive gaseous mercury and have raised questions 

that should be addressed as this method improves. 

The primary lesson is that RGM measurements are not trivial. In fact they are difficult to 

perform. They require routine, and strict adherence to protocols to be able to compare them with 

other data. The blanking, analysis and zeroing of a single denuder takes approximately 1 hour. 

The results of this work show, that in the future, manual methods for measuring RGM should 

be performed using triplicate co-located denuder measurements, and the automatic RGM analyser 

should be deployed when a high-resolution time series is needed.


The annular denuders are subjected to numerous analytical steps to inspect, clean and coat 

them, expose them and afterwards analyse them. The physical principle of these denuders is so 

simple, that it should make them ideal for use. However, in practice they are so difficult to employ 

that of the measurements for GEM, particulate mercury and RGM; RGM is the one that scientists 

working with atmospheric mercury in the Arctic have least confidence in (Schroeder et al., 2003). 

Why is this? As shown in Figure 6., page 64, in general, parallel measurements in these 

experiments are reproducible within the 15% documented by Landis et al., 2002, except for the first 

measured replicate. In Lindberg et al., 2002, Appendix C., manual denuders were compared with 

the automated system and 5 of 6 replicates gave excellent results. However, a single outlier is 

enough to document the need for continued improvement. Especially since the only proper 

validation so far, is to make multiple replicate measurements, as there is at present no standard 

method of the calibrating the annular denuders prior to use. This is especially important since only 

two denuders are used as accumulators in the REA RGM system. In principle, if these two denuders 

cannot measure the same amount of RGM, then they will necessarily produce a measurement 

artefact that will result in an erroneous flux measurement. For example if the two denuders in the 

first parallel measurement shown in Figure 6., page 64, represented denuders that were hung on the 

up and down channel, then this would show as a significant difference in concentration, and thus a 

difference in flux. This single observation reinforces the need for a calibration system, especially if 

the system is to be deployed for further REA flux measurements. 

Figure 7., page 65 confirms that the physical geometry of the denuder sampling train as well as 

the fact that the parts are made out of quartz make the sampling train only negligibly exposed to 

passive uptake. Thus fulfilling a pre-requisite for using the annular denuders as RGM REA 

accumulators. With the passive uptake rate found for the observed total concentration, this means 

that over a typical 4 hour REA sampling period that the RGM due to passive uptake does not


amount to more than the 10 to 15 % error from analysing the annular denuders, or the error in their 

reproducibility, and may therefore be used for RGM REA measurements. 

While at Station Nord, it was pointed out that the denuder sampling technique is very 

sensitive to the vapour pressure of the compound, especially compounds distributed between gas 

and particulate phase as demonstrated in studies of polycyclic aromatic hydrocarbons (PAH) (e.g. 

Feilberg et al. 2000) experiments were therefore made to test the effect of measurements with and 

without heating in the Arctic, Figure 9., page 67. 

It seen that for higher RGM concentrations, there is an apparent difference, with heated 

denuders reporting more RGM. However the reason for this observation is not clear. The central 

question is if TPM is converted to RGM in the denuder or if the denuder is not functioning well at 

low temperatures, as described by Landis et al. (2002). At lower concentrations, 5 pg m -3 , heated 

and non heated denuders report approximately the same. This level may therefore be an expression 

of the detection limit of the KCl coated annular denuder, rather than true ambient values. 

This work confirms that the Landis et al., 2002 method not only is sensitive to differences in 

field application, but also is sensitive to differences in analytical application. For example, Landis et 

al., 2002 prescribes the use of co-axial fans in the analytical desorption train. Figure 10., page 68, 

shows that this is a consideration that will affect results, with higher values coming from systems 

that do not employ coaxial fans. How much cooling effect is needed from the fan? This is probably 

dependent on the type of sampling line used. Subjectively, the sampling lines in this experiment 

were warm to the touch, when co-axial fans were not used. 

The results from the tests confirm the recommendation made in Schroeder et al., 2003, for the 

need of a robust calibration system. The denuders were designed based on data from only HgCl2 

(Landis et al., 2002) and to date, there are no published methods for calibrating the annular 

denuders, though 2 research groups, Eric Prestbo and Julia Lu, report on calibrating the annular


denuders by using HgCl2 generated in respectively either a permeation tube or a diffusion tube, as 

reported in Schroeder et al., 2003. For the present study, they were not “calibrated” they were 

coated and used if a low analytical blank value could be provided prior to exposure, and if there was 

no significant Hg (0) signature from the denuder, prior to starting the heating cycle. 

The question remains: when we sample RGM using annular denuders in the Arctic, what is it 

we are measuring? 

We can systematically look at what we know may create problematic measurements. As 

pointed out in Schroeder et al., 2003, the use of the inertial impactor in the front of the sampling 

chain means that only the gas phase and fine, aerodynamic diameter of 2.5 um or less, particulate 

matter are sampled past the active surface of the denuder, with the coarse particulate matter being 

impacted on the impact plate. It is not known how much mercury is lost in the inlet or on the 

impactor plate. In a heated denuder, it is possible that any mercury on particles on the impactor 

plate is degassed. For example, snow and ice crystals will melt in the temperatures in the heated 

denuder. The air as it travels through the denuder, becomes warmer and continues to expand. This 

might disturb the laminar flow, forcing the gas towards the KCl coated walls. This may actually 

improve the RGM trapping efficiency of the annual denuder, but at the same time, possibly induce 

measurement artefacts due to fine particle capture on the coated wall surface. 

When should denuders be recoated? Recoating of denuders is defined operationally as when 

blank values begin to rise, indicating that the surface cannot be “cleaned” well enough through 

pyrolization. What is it that is inactivating the surface in the Arctic? When should denuders be 

replaced? As long as denuders are providing a good seal, then they are most likely fine for 

continued use. In this work, denuders approximately one year old were still functioning well. 

With respect to breakthrough under Arctic conditions, Landis et al., 2002 showed that for 

automatic sampling up to 2 hours, there was no breakthrough. When sampling over two hours,


unless two denuders are used in series, then one will simply not know if there was breakthrough or 

not. 

Are the denuders appropriate to be used as the sampling end in the REA flux machine? The 

tests from the denuders suggest that they are suitable to use as accumulators for the REA flux 

machine, in the Arctic, as long as the denuders are used under the same flow and heated conditions, 

since with the flux measurements, it is the difference in concentration that is needed. 

The RGM levels, should be higher than ambient background levels, for successful flux 

determination, otherwise a trustworthy difference between the denuder used in the up channel and 

the denuder used in the down channel may not be found. 

The denuder test results suggest that concentration measurements should be made in at least 

triplicate. In fact this is generally the case for REA denuder based sampling systems (e.g., Zhu et 

al., 2000) which employ at least three denuders for each of the up and down channel, then use the 

difference in concentration as the difference between the average concentration between the two 

channels. This approach was not possible for this pilot work, given the number of sampling systems 

available. However, the parallel and quadruplicate measurements suggested that a single annular 

denuder could reliably accumulate RGM in a REA system deployed in the Arctic, without being 

subject to significant passive uptake of RGM, when idle. 

Could other RGM accumulators have been used? Three methods aside from annular denuders 

have been developed for measurement of RGM: refluxing mist chambers, ion-exchange membranes 

behind particulate filters and potassium chloride, KCl, coated tubular denuders (Landis et al., 2002 

and citations therein). Of the four RGM measurement methods, only the annular denuder method by 

Landis et al., had the necessary characteristics of being able to operate under arctic conditions, with 

the flow dynamics essential for the REA flux measurements in this work. The annular denuder has 

the ability to quickly develop laminar flow, i.e. flow with a Reynolds number, Re, under 2000. This


is maintained even though the flow is started and stopped at a certain frequency in response to up or 

down drafts. Under constant, steady state flow conditions, developed within 0.1 second for the 

denuder, the Reynolds number falls to less than 400 (Landis et al. 2002). The Reynolds number for 

an annular denuders’ geometry is defined as: Re = 4Q/((γπ)d1+d2) where Q is the flow rate, γ is the 

kinematic viscosity of air and d1 and d2 are the internal and external diameters of the annulus. 

This feature is especially important compared with tubular denuders since as can be derived 

from the modified Gormley Kennedy equation (Gormley and Kennedy, 1949), discussed below, the 

annular denuders allow a high flow rate and increase the collection efficiency per unit length by a 

factor of 30 (Possanzini et al., 1983). 

Denuders, also known as diffusion denuders, collect gases based on the diffusion properties of 

gases, as compared to particles. The separation of gas and particulate sampling is important so that 

appropriate masses, and thereafter concentrations may be determined. It is especially important to 

this work, when measured concentration differences in two collocated denuders are used as the 

basis for flux determination. As laminar air is pulled through a tube, in the case of an annular 

denuder, there is a concentric tube inside the outer tube, providing extra surface area, and the air is 

pulled through the 1 mm “annular” space between the two tubes, the “annulus”. Gases diffuse 

towards the walls of the tubes, due to their relatively high diffusivity, while the high inertia of 

particles will cause them to continue travelling through the tube (e.g., Hering et al., 1988). The 

inside of the tubes are coated with a substance that causes the gas to “stick” to the walls, and the 

gases are removed from the flow stream. In this case, the coating substance is potassium chloride, 

KCl, and the adsorbed species is reactive gaseous mercury, HgXY. 

The relative removal of the gas is a function of distance, x, travelled through the tube, and is 

given by modifying the Gormley-Kennedy equation for tubular denuders (Gormley and Kennedy, 

1949) to the geometry of the annular denuder (Allegrini et al., 1987):


Cx/Co = 0.82 exp ( - 22.53 (πDL/4Q)(d1+d2/d2-d1) (1) 

As seen from the equation, it is valid for collection efficiencies of 100% where the value of the 

mass concentration of the gas at a distance, Cx, over the mass concentration of the gas at the 

entrance, Co is less than 0.82. D is the diffusion coefficient of the gas in the air, and Q is the 

volumetric flow rate, d1 and d2 are the internal and external diameters of the annulus. Important 

assumptions for the above therefore are: 1., laminar flow must be obtained in the tube; 2., all the 

molecules reaching the wall must be trapped and should not diffuse chromatographically further 

into the tube; 3., The capacity of the trapping medium limits collection time, since eventually the 

active surface will become saturated and the gases will no longer be trapped on the wall, being 

carried through the denuder instead. This is what is known as breakthrough (Bemgård et al., 1996). 

These conditions must be satisfied for annular denuders to work as the RGM reservoir for the REA 

sampling system. 

Considering the results from the denuder tests, with the previously mentioned considerations 

and uncertainties, when sampling at 1 Hz, the above requirements are met. The denuders were 

simply not able to sample RGM at a sampling flow rate of 10 Hz. To do so, their geometry would 

need to be redesigned in accordance with the Gormley Kennedy equation. This would mean that 

they would need to be shorter and thicker. This would limit the development of laminar flow and is 

not rudimentary redesign work. 

One of the ideas for the future could then be to indirectly measure RGM using KCl coated 

beads in a quartz denuder tube, of the same dimension of the annular denuders in Landis et al., 2002 

and thereby facilitating, using the same analytical sampling chain and heating caps. This system 

would collect TPM + RGM. TPM could be measured alone with a co-located TPM collector and 

subtracted from the RGM+TPM, giving an indirect measurement of RGM. A TPM sampler could 

also be modified to collect RGM. This would take advantage of there being greater confidence in


PM measurements than RGM measurements in Arctic atmospheric mercury studies (Schroeder et 

al., 2003) though might introduce greater experimental error. The effect on flow from heating the 

denuders should be studied in greater detail. 

Ozone and gaseous elemental mercury measurements 

The net exchange rate of mercury to and from the snow surface in the Arctic is a result of many 

physical and chemical processes. This work confirms previous observations that the results vary 

seasonally and diurnally, with depositional events occurring perennially across the Arctic and that 

there is an apparent accumulation of mercury in the Arctic, as evidenced by the measurements in 

peat. 

Schroeder et al. (1998) described the first AMDEs, observed at the Canadian military station, 

Alert since 1995. They showed that ozone and GEM are simultaneously depleted and that they are 

highly correlated during AMDEs. Data in Figure 11., page 69, confirms this as well for Station 

Nord for the years of 2000, 2001 and 2002. If it is assumed that Hg was depleted when ozone was 

destroyed, then it can be inferred from the ozone data from Station Nord, that mercury depletions 

occurred there since 1998, and probably prior to that as well. 

After the depletion period, very high concentrations of GEM appeared with values up to 4.5 ng 

m -3 appearing in 2002 (off the scale in Figure 11., page 69). This must be due to reemission of 

mercury from the snow surface to the atmosphere. The measurements at Station Nord are in 

agreement with what was seen in Barrow (Lindberg et al. 2002, Appendix C), though the value of 

4.5 ng m -3 (+) is higher than seen in 2000 and 2001, and more than triple that of normal ambient 

levels, once elemental mercury finds its equilibrium in the late summer atmosphere, at Station 

Nord. What is happening to this mercury? This is a question that still cannot be satisfactorily 

answered and is the subject of future research.


That ozone and GEM are directly dependent on one another, or on a mutual factor, can be seen 

in Figure 12., page 70, where R 2 = 0.8 this is comparable with what is found at the other Arctic 

stations. given the strong correlation between ozone and GEM, observed at Station Nord during 

AMDE’s. A direct reaction between ozone and GEM can be excluded due to the approximately 1 

year lifetime of GEM with respect to the present ozone concentrations (Lin and Pekonen 1999). 

However, a speculatively plausible reaction mechanism can be reasoned from what is presently 

known. 

A plausible mechanism for the oxidation of elemental mercury to divalent gaseous mercury 

after polar sunrise in the Arctic 

Bottenheim (personal communication) showed that Bromine builds up due to the “bromine 

explosion” mechanism ((3) – (6)): 

O Br ⎯⎯→O 

+ BrO 

3 + 2 

(3) 

BrO + 

+ BrO ⎯⎯→ 

2Br O2 

(4) 

BrO + HO2 

⎯⎯→ 

HOBr 

(5) 

− 

Br + HOBr ⎯⎯→ 

2Br 

therefore either Br or BrO is a candidate for GEM removal is. An analogue mechanism may occur 

with Cl, however Cl and ClO cannot initially be ruled out, as significant Cl removal of organic 

compounds have been observed during AMDE (e.g. Boudries and Bottenheim, 2000). Taking 

thermodynamics into consideration however, shows that the reaction: Hg (O) with Cl and radical 

ClO are less likely to occur than with Br and radical BrO. 

The lifetime of GEM is observed to be typically about 10 hours during AMDE as confirmed 

also by this data from Station Nord. Hausmann and Platt observed up to 20 pptv of ClO and BrO 

(6)


and thus the resulting rate constants for the reactions can be estimated to be in the order of 6x10 -14 

cm 3 molec -1 sec -1 . 

Skov et al., 2003, Appendix C, analysed the data using a relative rate study, and found direct 

evidence in a strong linear correlation (>99.9% significance) that ozone and GEM have a mutual 

dependence that cannot be explained solely by meteorology. 

Goodsite et al., 2003, Appendix C, propose a plausible mechanism for the oxidation of gaseous 

Hg (0) to the divalent gaseous form. The mechanism is modeled from the limited data known of Br 

and Hg reactions in the Arctic, given the data from Station Nord, with a predicted lifetime of Hg to 

be approximately 10 hours during AMDEs. 

The mechanism is consistent with the kinetics, thermodynamics and field observations, but the 

final end product is still not definitively known, due to the measurement method in the field, i.e., 

reactive gaseous mercury is operationally determined and defined. 

The hypothesized mechanism is not the same as the mechanisms otherwise derived (e.g. 

Lindberg et al., 2002, Appendix C), where the depletion of atmospheric boundary-layer mercury is 

said to be due to a reaction between gaseous elemental mercury, GEM, and BrO free radicals or 

mechanisms resulting in the product HgCl2. 



forming the energised HgBr*. 

Through a third body reaction, M, where M is N2 or O2, the HgBr radical is formed. 

The HgBr radical can live long enough at the low temperatures of the Arctic to combine with 

O2 forming the HgBrOO peroxy radical or can combine with Br forming HgBr2. 

It is not likely to react with Cl, since this reaction would be endothermic.


Similarly, the product cannot be Hg2Br2 since this would imply a tri-molecular reaction, which 

statistically is highly unlikely to occur in the atmosphere due to the very low concentrations of 

elemental mercury and bromine atoms. The combination of Hg and Nr atoms and the dimerization 

of HgBr is discussed in Grieg et al., (1970). 

Nor would the product be HgO, since this formation is similarly thermodynamically not 

favourable. 

The final product is the divalent gaseous mercury unknown, HgXY, proposed to be HgBr2. 

By modelling the reaction of Hg and Br in the atmosphere, with current rate constant data, and 

BrO measurements, i.e., assuming 20 ppt BrO and 2 ppt Br in the atmosphere during a depletion 

event, it is calculated that the lifetime of Hg is 4.6 hrs against forming HgBr, 

The lifetime of HgBr is 0.35 hrs. against forming HgBr2; comparing with the lifetime of HgBr 

of 0.75 hrs, means that 68% of the time HgBr will form HgBr2. 

Thus the overall lifetime of removal of Hg to HgBr2 is 4.6 hrs. / .68 = 6.7 hrs. This is in good 

agreement with the observed 10 hr lifetime of Hg under depletion, and therefore the above reaction 

mechanism is plausible as one of the limiting reactions for the oxidation of elemental mercury to 

reactive gaseous mercury in the post polar sunrise atmosphere, though it is realized that given the 

dynamics of atmospheric chemistry, that the above scheme is based on a simple model with limited 

laboratory kinetics data to carry out the calculations and that HgBr may also be reacting with O2 or 

many other elements or compounds in the Arctic atmosphere. 

RGM concentration measurements 

RGM measurements performed in parallel at Station Nord in 2002 show that the concentration 

varies from between values below detection limit and up to 75 ng/m 3 , Figure 13., page 71. These 

values are comparable to those found at Barrow before the start of AMDE’s and are comparable to 

levels measured in the Dr. Lindbergs’ laboratory in Oak Ridge. These values were measured prior


to the start of a depletion event, as we had to leave the station prior to when a pronounced AMDE 

occurred. The concentrations are an order of magnitude lower than RGM maximums measured in 

Barrow in 2001 (Lindberg et al., 2002, Appendix C). 

Barrow and Station Nord are however, difficult to compare since at Station Nord, the prevailing 

winds are katabatic, bringing cold, dense air from the inland ice towards the coast. This means that 

many of the conditions observed at the other Arctic stations, with prevailing winds from the coast, 

are not observed at Nord. This also implies that the typical air mass at Station Nord is thinned with 

relatively clean air. 

RGM flux measurements in Barrow 

Table 1 has the mass, and corresponding concentration for the measurement periods, as well as 

the corrected mass and concentration and a comparison of concentration between the total found in 

the three annular denuders with the measured concentration with the TEKRAN automatic speciation 

unit, MODEL 1130. It was necessary to correct the mass and concentration to have conservation of 

mass. A properly set up micrometeorological system will automatically produce a time series, 

sampling turbulence such that mass is conserved. As seen in the non-corrected data from the 

Barrow campaign, this was not the case. Due to limited space on the 10m tower at CMDL, Barrow, 

it was necessary to deploy the REA RGM system on a tower guy-wire support pole. 

The sonic anemometer was set up on a pipe that was bought in Barrow, and whose outer 

diameter did not snugly fit the METEK sonic. While every effort was made to create a stable 

platform for the sonic, it was evident after the first three measurements that the sonic was 

preferentially sampling the down channel. This meant that either the sonic was not totally level, or 

that the air was being forced downward, by the CMDL station. Given the position of the sonic, in 

relation to the station, and the unorthodox method of setting up the sonic, it was determined that it 

was most likely the sonic that was not properly aligned. During the first two sampling runs, the


sampling valves froze, and their data is not further considered. It was therefore necessary to 

reposition the sampling valves, into a shipping box, buried in the snow for insulation. The third run 

also showed sampling which favoured the down channel. One expects a micrometeorological 

system to equally sample turbulence, as seen by similar number of counts in the up and down 

channel. One channel being preferred over another is a sign that either the wind flow is being forced 

in one direction of another, or that the instrumentation is not properly set up. Even though the 

results indicated preferential sampling, the decision was made not to adjust the sonic. It was not 

certain that any adjustments would make the system any more level, and the first three 

measurements were consistent, so mass adjustments could be reasonably applied. Stopping the 

system each hour to read the data caused the tilt function to reset. This could also be a source of 

bias. In the new generation system, it is no longer necessary to stop the system to monitor and 

download the data. 

The total concentration is taken to be the sum of the up, down and mid masses multiplied by 

total sampling time and the flow rate. The concentrations compare sometimes very well with the 

RGM monitor, and other times not as well. There does not seem to be any general trend, except that 

the REA system total is generally less RGM than the 1130 concentration, Figure 22., page 102.


RGM pg m-3 

500 

450 

run nr. 2, valve stuck open in REA 

system 

400 

350 

300 

250 

200 

150 

100 

50 

0 

Comparison of total RGM concentrations, Barrow 2001 

1130 total 

RGM REA total 

1 2 3 4 5 6 7 8 9 10 12 11(dark) 

REA flux run number 

Figure 22 A comparison of RGM air concentrations between the REA system and a co-located TEKRAN model 

1130 mercury speciation unit. Due to the different heating mantles, it is expected that the REA 

system will have lower RGM values, since the annular denuders were only heated to 50 0 C above 

ambient. and did so in all but 2 cases, excluding run nr. 2. The average percent difference between 

the two measurements is 24% with a standard deviation of 42, excluding run nr. 2. RGM 

measurements from S. Brooks TEKRAN 1130, this author used a linear relationship between time 

and concentration for the 2537A since the run start times for the two instruments were different. 

The flux was divided by the total concentration to determine the depositional velocity. Table 2. 

page 75, has the average values reported in the REA system data output file runname.txt. It is seen 

that for the measurement periods, the average temperature was rising, during the campaign, while 

wind speed was on average, generally consistent. Wind direction was more variable coming from 

the direction of Barrow most of the time, instead of the coast. The standard deviation of the vertical 

wind velocity and the value for the proportionality constant, are on average, stable, as should be 

expected for the weather conditions.


The REA pilot system was designed and built prior to deployment with the best intentions in 

mind. The system did what it was built to do, reporting what it was designed to report. There were 

several areas that were found to need improvement and these improvements were incorporated into 

the NOAA REA system, built for use in Station Nord. In the system used in Barrow, in the 

runname.txt file, channel count data did not equal sampling time, when multiplied by the sampling 

frequency. The count data did, in the raw data file, runname.acu. The fact that the data was reported 

as a sampling average provided limited data available for analysis. Average wind direction, while 

nice to know, is not the same as saying the prevailing wind direction when mass was sampled over 

the up and down channels, or even better, breaking up the mass in terms of a wind rose. This is 

now possible with the new REA system. Since the system needed to be shut down for data analysis, 

back-up and recovery, the self-learning function of the sonic was effectively reset each time the 

system was restarted, having to re-learn the terrain at the start of each sampling period. For this type 

of system, the sonic should be left running. This has been remedied with the new system. 

Figure 14. page 77, shows the flux and depositional velocity for the Barrow campaign. Mass 

non-corrected values are shown for informational purposes only. Comparing the flux with the 

average meteorological conditions in Table 2., page 75, shows no direct correlation. One would 

expect that emissions would increase as a function of the rising temperatures, but this can not said 

to be readily apparent. What seems to be apparent is that following a large depositional event; there 

is a small re-emission. This is probably the snow pack regaining equilibrium with the atmosphere. 

Over the course of the campaign, there is an average deposition; this however is for just one Arctic 

location, during a very short period of time. A much longer time series, spanning the 4 seasons is 

required to know more about annual trends and to parameterize models.


Depositional velocity and surface resistance for RGM in Barrow 

Numerical depositional models require the surface resistance and depositional velocity for 

proper model parameterization. Zannetti, 1990, defines an operational definition of dry deposition 

velocity as: 

1 

Vd = 

ra + rm + rs 

(7) 

Where ra is the aerodynamic resistance, depending on the atmospheric turbulent transfer, rm is 

the molecular resistance in the atmospheric viscous sub-layer and rs is the surface resistance of the 

snow, which depends on the flux into the snow layer. 

Zannetti defines the flux operationally as: 

F ( z = 0, 

t) 

= Vd( 

z1) 

Cg( 

z1, 

t) 

(8) 

Where the flux to the surface, when F is negative, or from the surface, when F is positive, at a 

reference height z1 as the depositional velocity found at the reference height multiplied by the 

concentration of the trace gas found at the reference height. In this study, we experimentally 

determine the flux, and the concentration and use this data to solve for depositional velocity; where 

our reference height in Barrow is 3 m above the snow pack. 

Substituting (7) into (8) and solving for a surface resistance, which can then be modeled, rs 

(modeled): 

Cg( 

z1, 

t) 

rs(mod eled) 

= 

− ra − rm 

(9) 

F( 

z = 0, 

t) 

If ra and rm are taken to be small in relation to the first term, then it can be seen that surface 

resistance is the inverse of the depositional velocity. The modeled surface resistance is a necessary 

parameterization term for depositional modeling, see Skov et al., 2003, Appendix C, and for this 

study, as seen by taking the inverse of the depositional velocities shown in Figure 14., page 77, it is 

found to be on average 1 cm -1 s.


A more exact determination of the surface resistance was not possible with the available output 

data from the pilot REA system. However, the new NOAA system produces the necessary output 

factors to allow modeled calculation of ra and rm in accordance with the Karlsson and Nyholm dry 

deposition and desorption model (Karlsson and Nyholm, 1998); i.e., friction velocity, Obukov’s 

length and the stratification function as long as a diffusion coefficient for RGM is known. Therefore 

campaigns with the new system will provide a better approximation of depositional velocity and 

surface resistance. 

As seen in Figure 14., page 77, the depositional velocities noted for depositional events are 

fast, around 2 cm s -1 this is what would be expected for a very reactive gaseous species such as 

HNO3. On average, the depositional velocity is approximately 1 cm s -1 . Comparing with measured 

dry deposition velocities over snow for HNO3, reviewed in Karlson and Nyholm, 1998, show that 

the measured dry depositional velocities for RGM may be an order of magnitude higher than that 

for HNO3, given the snow surface temperature of < 2 0 C. However, there are also reported values 

that fall within the range found for RGM in this study. For example, Cress et al.,1995, as cited in 

Karlsson and Nyholm, found the depositional velocity about 0 0 C to be in the range of 0.88-3.79, 

though the air concentration is not given. It is seen that as ambient concentrations of HNO3 

decreased, depositional velocities apparently increased: for a concentration in the air of 6-15 µg m -3 

Johansson and Granath, 1986, find a depositional velocity for air temperatures < -2 0 C of HNO3 to 

be 0.02 – 0.1 cm s -1 ; near 0 0 C the deposition velocity increases, to 0.6 cm s -1 . Cadle et al., 1985, 

as cited in Karlsson and Nyholm, report a depositional velocity averaged for all temperatures, with 

an air concentration of 6 µg m -3 to be 1.4 cm s -1 . 

Comparing the depositional velocities found in this study with the experimental data reported 

for HNO3 shows that while the expected values could have been an order of magnitude lower, the


values found are within reported ranges for a comparatively highly reactive gaseous species. 

Though RGM Vd may be better compared with NO2 Vd over snow. 

Karlsson and Nyholm note that at temperatures under –2 0 C the depositional velocity appears to 

be controlled by the surface resistance. The surface resistance decreases as the temperature warms 

due to increasing amounts of available water. Measurements with this system need to be made 

around 0 0 C to see if this is also the case for RGM. 

Slinn et. al., 1978, show that if dry deposition is the only removal mechanism for a substance 

not affected by chemical transformation, that the atmospheric residence time for the substance can 

be estimated as the deposition velocity multiplied by the height. If it assumed that Hg, once 

converted within the MBL is deposited almost as quickly as converted, as shown by the averaged 

measured depositional velocity of 1 cm s -1 then RGM would have a lifetime in Nord of around 10 

hrs. This would correspond to an MBL height of 360 m which is reasonable. One sees from the 

above relationship, that as the height of the MBL decreases, the atmospheric residence time will 

exponentially decrease. The measured values may reflect upper limits of surface resistance over the 

snow. 

Comparison of RGM flux with RGM ambient concentrations 

Figure 23., page 107, are the monitored results provided by NOAA, for RGM concentrations 

during the Barrow 2001 flux measurement campaign. Comparing with Figure 14., page 77, or the 

values in Table 3., page 76, show that when there is a deposition recorded by the REA machine, 

there are correspondingly low RGM ambient values. The air has been apparently depleted of RGM 

due to deposition. Trends of RGM rising in the air on the 8 th , 9 th and 12 th of April are recorded by 

the REA system as reemission, perhaps indicating that RGM can be re-volatized from the snow 

surface.


RGM pg m-3 

350 

300 

250 

200 

150 

100 

50 

0 

03-26 

00:00 

03-28 

00:00 

03-30 

00:00 

Automatic RGM measurements, Barrow 2001 

04-01 

00:00 

04-03 

00:00 

04-05 

00:00 

Date and time 

04-07 

00:00 

04-09 

00:00 

04-11 

00:00 

04-13 

00:00 

04-15 

00:00 

Figure 23. Automatic RGM measurements, Barrow Arctic mercury study, BAMS 2001, courtesy of 

Steve Brooks. 

The RGM REA system is providing results in accordance with what is known about RGM. 

RGM appears to have been deposited, when post mercury depletion event levels are low, and 

reemitted, when RGM ambient concentration levels are high. The exact processes need to studied in 

greater detail and a data set from a longer period is needed to make judgments about trends in the 

Arctic. 

Deciding how to make the flux measurements 

There are two types of deposition: dry deposition, a continuous process for gaseous species and 

particles, generally dominating in the arid high Arctic and wet deposition, gas in dissolved form in 

fog or precipitation, occurring periodically, and relatively infrequently in the high Arctic due to its 

arid climate.


Gaseous elemental mercury is approximately 95% of the total gaseous mercury in the 

atmosphere, TGM, (Schroeder and Munthe, 1998). Much of the GEM is oxidised RGM (e.g. 

Schroeder et al., 1998, Lindberg et al., 2002) during an AMDE. Thus RGM is likely subject to 

subsequent fast dry deposition to the Arctic snow surface during an AMDE (Schroeder et al., 1998). 

Therefore this work developed a method to quantify the deposition during an AMDE, so that the 

data could be used for further flux measurement development and model calibration and 

verification. 

Given the challenges encountered with the micrometeorological system, it is worth reviewing 

options considered for measuring RGM flux. Flux measurements could be as “simple” as setting a 

small environmental chamber over the snow surface, however this would give limited information 

as to the pattern and exclude natural processes and deposition. A chamber is not able to fully 

measure representative exchange rates, and there has not been a chamber method developed yet, for 

the flux of reactive gaseous mercury in the Arctic. Development of a chamber method for RGM is 

hindered by the need to have a warm KCl coating, if KCl is used as the active surface for RGM. 

This heating will liquefy the snow surface. 

The work instead chose a micrometeorological method, REA. 

REA has been used for the determination of elemental mercury flux (Cobos et al., 2002). REA 

allows for a covering a relatively large area, known as the fetch, i.e., the fetch is approximately = 

100 x height of system; in this case 300 m; since the system was set up 3 m above the snow pack 

surface. 

However, as discussed below, there are also drawbacks with using REA to investigate mercury 

depletion events in the Arctic, particularly that weather conditions favourable for the system, i.e., 

high turbulence, are not favourable for mercury depletion events to occur, and are the exception to 

stable weather normally encountered on a flat Arctic coastal plane.


REA measurements rely on the transport of mass near the surface of the Earth where 

turbulence in the air is the main transport and mixing mechanism that eventually causes deposition. 

Fluid dynamics show that at any rough boundary, the friction causes velocity of a fluid (e.g. air) to 

go to zero. As this occurs, the air velocity vectors straighten out and are laminar in a plane a few 

millimeters thick over the Earth. Once gas molecules get into this plane they will either diffuse 

towards the surface, or evade towards the atmosphere, depending on the concentration gradient 

between this layer, known as the quasi-laminar boundary layer. Under higher turbulent conditions, 

the amount of mass transported greatly exceeds the amount of mass diffused, and generally 

turbulent transport is the most important physical method for deposition. A REA system samples 

this mass, as it is transported up from emission, or deposited. A valid flux determination relies on 

the principle that the mass of the trace gas measured, in this case RGM, is unique to that air mass. 

This means that it wasn’t produced or destroyed in the air mass. 

From the GEM concentration results from the Station Nord campaign, it can however be seen 

in the high Arctic, that under stable weather conditions that gaseous diffusion may be important. 

This is an area requiring further investigation in the future. 

Shear stress from surface friction is one of the main driving forces of turbulence; the other 

driving force is the change in buoyancy with the change in air temperature and thus density with 

height. This force may be important over the Arctic snow surface, however the cold surface cause 

stable cold air, that is difficult to move, and results in calm conditions, with wind speed under 2 m 

per second, as witnessed at Station Nord in 2002. 

These stable weather conditions are the greatest draw back to deploying the RGM system in the 

Arctic for investigating mercury depletion events. Lu et al., 2001, summarize that the environmental 

conditions favouring mercury depletion events at high latitudes are: 1, marine/maritime location; 2, 

calm weather, low wind speeds, non-turbulent air flow; 3, the existence of a temperature inversion;


4, sunlight and 5, sub-zero temperatures. Conditions 2 and 3 unfortunately imply poor operational 

characteristics for micrometeorological systems. 

The most important requirement however, for a micrometeorological flux system is a surface 

that isolates the vertical fluxes, by being able to neglect horizontal exchanges. This is a “flat” 

horizontally homogenous surface. The coast of the Arctic at Barrow and Station Nord are very good 

in fulfilling this criterion. 

Micrometeorological methods also make the assumption that the surface exchange flux is equal 

to the vertical turbulent flux at any height above the surface, as long as one remains inside the 

inertial sublayer. The inertial sublayer is the layer above the roughness sublayer, closest to the 

surface, and is the bottom tenth of the planetary boundary layer, PBL, generally known as the 

surface layer, SL. This is an application of conservation of mass. 

The PBL grows in height during the day, expanding as the air warms, and contracts at night as 

the air cools. The PBL is generally very stable in the Arctic and is the region of atmosphere closest 

to earth where direct effects from the planet no longer affect the physical properties of air. The 

marine boundary layer, MBL, is the layer between which air masses from the land and sea 

converge. It moves towards and from the land, as well as in height both on diurnal and seasonal 

patterns just as the PBL. In this study, we were within kilometres of the shoreline in either Nord, or 

Alaska, and therefore, for all intensive purposes, always within the Arctic MBL. This is important 

as a source of halogens driving the oxidation previously discussed. 

The system chosen and used, the REA RGM flux system with KCl coated annular denuders as 

accumulators gave the possibility for good flux measurements, which would provide necessary data 

for transport and depositional model parameterization. The flux system could take advantage of 

already developed methods, to accumulate the RGM. Since the deployment of this system, REA 

was successfully applied to measure gaseous elemental mercury flux (Cobos et al., 2002).


Global, Arctic and atmospheric implications 

Reactive gaseous mercury is the operationally defined product of the post polar sunrise 

oxidation of elemental mercury in Arctic areas as shown at two trans-Arctic locations, Station Nord 

and Barrow. This work shows that in accordance with what would be expected for a reactive 

gaseous species, that it is quickly deposited to the snow surface, though there are also measured 

emissions, implying that there are processes in the snow surface capable of releasing RGM, or re- 

emitting the deposited RGM. On average, the mercury was deposited to the snow pack for the short 

spring measurement period. There was little RGM at night; supporting that RGM production in the 

Arctic atmosphere is driven through photochemistry. 

The values found in this work may be preliminarily used for transport and deposition modeling, 

as may the proposed reaction mechanism. The experience gained with the pilot system lead to 

development of a new system. Experience with the annular denuder method for RGM collection 

(Landis et al., 2002), confirms that it is a sensitive method for measuring RGM, but must be applied 

carefully to ensure comparable results. The work supports the need for an RGM calibration unit, 

and an RGM inter-calibration and standardization study, to be carried out in the spring, 2003, in 

Svalbard (see Schroeder et al., 2003). 

The atmospheric bomb pulse curve 

A high time resolution series of trace metals is urgently needed for pollutants such as Hg to 

evaluate the effects of emission controls, and to help calibrate atmospheric transport models. A 

necessary pre-requisite to establishing this time series is development of a dating method that 

provides this type of time resolution. 

For a successful dating method, first there must be an accepted way to calibrate results, and 

then show that the calibration, when applied experimentally, produces reliable results. Figure 15., 

page 79, shows several records of 14 C covering the bomb-pulse period, for the northernmost


northern hemisphere, 20-90 0 N latitude. None of the so far published records from tree rings, as 

seen in the figure, covered the whole second half of the 20 th century, Therefore, tree ring and 

cottonseed data from Arizona as well as data from Denmark was added to complete the curve. 

Previously unpublished Atmospheric measurements provided by I. Levin, University of Heidelberg, 

consisting of the annually averaged atmospheric 14 CO2 curve for the northernmost northern 

hemisphere is plotted and recommended for use as the general bomb-pulse calibration curve for the 

northern hemisphere for 4 reasons: 1) the data from Arizona closely follow the curve but do not 

provide annual resolution, this is the limitation of any of the data from particular points or terrestrial 

sources; 2) the curve provides consistent, carefully checked atmospheric data covering the whole 

pulse until the present, 3) The portion of this data set up through 1997, is published and widely used 

(Levin et al., 1997); differences for the other curves results in calibrated age differences of only 1-2 

years from the data provided by Levin. 

As then seen in Figure 15., page 79, for the northernmost northern hemisphere it is most 

appropriate to use the Levin data as a calibration tool in General, though depending on location, 

there are local features, e.g. testing around China that this curve now provides a ready reference to. 

For more detail discussion, see Goodsite et al., 2002, Appendix C. this work is the first time that 

bomb-pulse dating was successfully applied to peat. This was primarily a function of having the 

appropriate calibration curve, and sampling and handling the peat appropriately, as discussed in the 

paper. 

The successful development of this method means that now, when it is able to be applied and 

used, the greatest source of error in determining a chronological series of a trace metal in peat no 

longer comes from the date determination, but rather comes from sampling and handling. With 

especially the mundane determination of bulk density in peat now providing the greatest source of 

error, when calculating accumulation rates.


This method can be applied where standard-dating methods cannot be, for example in peat 

disturbed by digging or cutting. The limitation of this method is that it only covers the last half of 

the 20 th century. The primary gain is that the chronology is an order of magnitude better than 

standard chronological dating methods. 

Figure 16., page 80, shows high resolution dating is applied successfully for the first time in 

Denmark and Greenland. In Denmark, due to peat digging during WWII, the Pb-210 dating method 

could only be applied with error an order of magnitude higher than that provided by the bomb-pulse 

method. In Greenland application of the method allowed for the first coastal high time resolution 

investigation of trace metal accumulation in an environmental archive in the Arctic. 

In Figure 17a. and b., page 83; 18a. and 18b., page 84; 19., page 85; and 20., page 86, it is seen 

that in the minerotrophic southern Greenland, and the blanket bog on the Faroe Islands, that the 

chronology of Hg accumulation is similar to that of the ombrotrophic bog in Denmark, during the 

last 50 years. This suggests that in remote areas, Hg is supplied to the wetlands primarily via 

atmospheric deposition, demonstrating that given the proper conditions, minerotrophic sediments 

provide a history of atmospheric deposition as consistent as one provided by a raised bog. It is also 

seen that dry bulk density is very variable, most likely the result of sampling and determination 

methods. 

Hg fluxes in the Greenland core (0.3 to 0.5 µg m -2 yr -1 ) were found in peats dating from AD 

550 to AD 975, compared to the maximum of 164 µg m -2 yr -1 in 1953. Atmospheric Hg 

accumulation rates have since declined with the value for 1995, 14 µg m -2 yr -1 comparable to 

published 1995 values from the Danish Eulerian Hemispheric Model (Christensen et al., 2002, also 

found in Skov et al., 2003, Appendix C), of 12 µg m -2 yr -1 for southern Greenland. 

In Denmark, the greatest rate of atmospheric Hg accumulation is found in 1953, 184 µg m -2 yr - 

1 , comparable to that of Greenland with the flux going into sharp decline, with an accumulation rate


for 1994 as 14 µg m -2 yr -1 . This compares well with the modelled rate of 18 µg m -2 yr -1 for all of 

Denmark as well as previous studies (e.g., Madsen, 1981). On the Faroe Islands, the maximum 

mercury concentration was 498 ng g -1 , dated to be in 1954 +/-2, with a 210 Pb constant rate of supply 

dating model, in good agreement with historical maximums in 1953 in S. Greenland and Denmark. 

Depositional maximum was in 1985, 34.3 µg m -2 yr -1 , with the 1995 value of 18 µg m -2 yr -1 , 

comparing with Denmark and the modelled value of approximately 10 µg m -2 yr -1 . The present day 

value of approximately 10 µg m -2 yr -1 compares well with recently reported wet deposition 

measurements of 7 µg m -2 yr -1 (Daugaard, 2003). 

Long-term accumulation rate of mercury in this core was 0.95 ± 0.36 µg m -2 yr -1 for the period 

of 4200 B.C. to AD 833; n = 61, in agreement with other reported peat studies, and what would be 

expected in a remote area. The Hg concentrations in the Faroe Islands are higher than those found in 

cores from other sites, but the net Hg accumulation rates are comparable. 

The depositional records in the peat are in good agreement with what would be predicted: 

lower depositional rates at sites further North from European industries; and in agreement with 

previous atmospheric measurements that showed that total gaseous mercury in Europe reached an 

average annual maximum in the late 1980’s and have been falling since 1990 decreasing globally 

by 22% from 1990 - 1994 (Slemr et al., 1995 and references therein, Slemr et al., 2003). Schuster et 

al. (2002) used a glacial ice-core record to study atmospheric mercury deposition during the last 270 

years, concluding that the anthropogenic contribution during the last 100 years rose to 70 percent of 

the total, a 20 fold increase over background, yet falling the last decade. The mercury depositional 

rates and probable sources are determined in Denmark, Southern Greenland, and the Faroe Islands, 

using lead and stable lead isotope signatures. Surprisingly, on Nordvestø, Carey Islands, lead was 

below detection limits. This is the first time observed in such low levels in peat. Therefore the peat


record from Nordvestø is providing a unique, perhaps biogenic record of the mid Holocene. The 

study is on going and more cores are being analyzed from Nordvestø. 

Global and Arctic significance 

The peat studies in this work document that not only has anthropogenic activity caused mercury 

to accumulate in the Arctic, but also provides apparent support to the fact that anthropogenic 

actions, for example closing of chlor alkali plants, in the latest decades has caused a significant 

decline in mercury accumulation in the Arctic environment. The peat studies support that the 

observed rise and fall in mercury accumulation in recent times is strongly related to coal use 

(Shotyk et al., 2003, Appendix C). 

Unfortunately, the peat profiles that could be investigated with high time resolution come from 

areas of the Arctic where mercury depletion events have not been observed, therefore not answering 

the question of if these profiles would look the same, given that they take place in a AMDE active 

area. 

The core from the Carey Islands, while from an area expected to have AMDEs, is too disturbed 

in the active layer to investigate this. Lake and Marine sediments from the high Arctic, coastal site 

near Station Nord, may provide some insight into this, once dating results are completed, so that the 

cores can be analyzed further, but preliminary comparison with other Arctic sediments, does not 

indicate an increased rate of accumulation. This means that mercury being deposited to the Arctic 

may be coming out again. 

There are three routes that deposited mercury could be redistributed, so that there is no 

terrestrial accumulation signature: 1 uptake into the biosystem, 2 transport away from the Arctic 

after emission into the marine system or 3. re-emission into the atmospheric system, with 

subsequent transport away from the Arctic outside the spring months. Determining a mercury mass 

balance for the Arctic should be a future priority.


In the high Arctic, due to the slow sedimentation of environmental archives, one of the ways to 

see any tendency is to continue long-term environmental monitoring studies. 

The background levels of mercury in the peat core are very consistent with what is understood 

of the global cycling of mercury and its long atmospheric residence time. The cores show some 

variation due to natural process, geogenic or climatic, and the reasons for these pre industrial 

variations should be more closely examined in the future, since it may provide insight into the 

question of how the expected warming climate will affect mercury deposition to the Arctic in the 

future, having observed how it was affected in the past, though without human effects. 

Even though mercury deposition may be declining, there are recently still negative effects 

observed such as attenuated growth of breast fed children exposed to increased concentrations of 

methyl mercury and other contaminants such as polychlorinated biphenyls (PCBs) (Grandjean et al., 

2003). 

Munthe et al., (in press) summarized that the global nature of mercury transport affects not only 

local areas, but regions such as Europe. Therefore control measures in Europe will not reduce 

atmospheric deposition to acceptable levels in Northern Europe. They state that “there is a need to 

assess the hemispherical and global background levels and to what extent these are impacted by 

anthropogenic emission.” This work shows that by coming peat archives with stable lead isotope 

measurements and high time resolution dating, that future assessments may be made using this type 

of environmental archive. 

Gårdfelt et al., (in press) measured and modeled that 66 tonnes of mercury are released to the 

atmosphere from the Mediterranean Sea during the summer. They corroborate measurements in the 

Atlantic showing that this type of evasion is not confined to the temperate regions. Oceanic evasion 

should therefore be looked at in the Arctic region as a possible source of Hg to the Arctic.


5. Conclusion and future work 

This work provides a better understanding of the temporal, and spatial patterns of mercury 

deposition and accumulation in the Arctic, gaining information of the chemical and depositional 

processes so that they might be applied to parameterise models of atmospheric transport and 

deposition and eventually aid in making policy decisions. 

The post solar sunrise deposition of reactive gaseous mercury to the Arctic, and the 

depositional velocity of RGM was quantified during a campaign in Barrow Alaska. To do this, a 

RGM REA micrometeorological flux system was developed for Arctic use, and RGM flux 

measurements made for the first time in the Arctic. Lessons learned from the first campaign were 

applied to redevelop the system. Based on experimental observations a plausible reaction 

mechanism for the oxidation of gaseous elemental mercury to reactive gaseous mercury was 

presented. Continued improvement of the annular denuder measurement method for Arctic use is 

needed, as is standardization, an RGM calibration system and an intercomparison campaign, to 

ensure robust, inter-comparable and accurate measurements in the future. An annual time series of 

wet and dry depositional flux is needed. Since the development of automatic RGM sampling 

systems, it may be possible to automate a REA RGM flux measurement system, readily enabling 

this monitoring. Laboratory kinetic studies of Hg and the halogens, especially Br are needed. It 

should be a goal to find out exactly what RGM is in the Arctic. 

Peat archives of environmental pollution were located and investigated in the Arctic, and in 

Denmark so that the pre and post-industrial values of Hg in the Arctic were determined. A high time 

resolution dating method was developed, allowing the peat profiles to be intercompared and 

providing information as to accumulation over the last 50 years. Methods were developed and 

advanced for determination of mercury in peat and for the sampling of peat in the Arctic. Detailed


studies need to be carried out in suitable environmental archives in areas where mercury depletion 

events occur to determine if there is an enhanced sedimentary Hg accumulation.


Glossary, acronyms and abbreviations 

The terms, acronyms and abbreviations below may appear in this thesis. Whenever possible, they 

have been harmonized with those terms used in the UNEP global mercury assessment report for 

standardization purposes. 

< - less than; 

> - greater than; 

°C - degree Celsius (centigrade); 

µg – microgram (10 -6 gram); 

µg kg -1 body weight per day – micrograms per kilogram body weight per day; units used for 

describing intakes (or doses) of mercury such as intakes that are considered safe for humans; ADI - 

acceptable daily intake; 

AMAP - The Arctic Monitoring and Assessment Programme; 

AMDE – Atmospheric mercury depletion episode, latest name for mercury depletion episodes 

(MDEs, mercury depletion incidents (MDIs) pertaining to the perennial oxidation of gaseous 

elemental mercury to the operationally defined reactive gaseous mercury, first observed in the 

Arctcic in 1995; 

ATSDR – USA Agency for Toxic Substances and Disease Registry; 

Balance (=budget) - totality of quantitative estimates of input and output substance fluxes for a 

given geophysical reservoir or societal entity; 

Dry deposition - process of species transport from the atmosphere to the underlying surface at their 

direct (without precipitation) physical-chemical interaction with elements of the underlying surface; 

dry deposition is of a continuous character independent of the occurrence or absence of atmospheric 

precipitation; 

FIS – one of the Faroe Island mercury exposure studies; 

Flux - amount of mercury deposited over a defined are per a defined time interval; 

Hg – mercury; 

Hg 0 or Hg(0) - elemental mercury; 

Hg 2+ or Hg(II) - divalent mercury - the dominating mercury form in organic and inorganic mercury 

compounds. In the atmosphere, mercury species with divalent mercury are more easily washed out 

of the air with precipitation and deposited than elemental mercury; 

Hgp - particulate mercury - mercury bound in, or adsorbed on, particulate material. In the 

atmosphere, particulate mercury is deposited much faster than elemental mercury; 

kg – kilogram; 

l or L – litre; 

Life-time - In atmospheric physical-chemistry: Time during which the first order processes (or 

totality of the first order processes) of scavenging results in mercury species mass reduction in e 

times in a geophysical reservoir; for a reservoir with homogeneous mercury species distribution the 

life-time is equal to the ratio of the mass contained in the reservoir to scavenging rate. Since the 

mass of mercury in the reservoir left to be reacted or removed decreases over time, the amount 

reacted or removed per unit of time decreases in a natural logarithmic fashion. For example, a


lifetime of mercury of one year, does not mean that it would all be gone in one year if emissions 

were zero. It means that the rate of removal at the start of the time period in terms of mass per unit 

time would remove it all in one year, but since the rate of removal decreases as the mass of mercury 

left decreased, the amount of mercury left after one year would be (1/e) times the initial mass, 

where "e" is 2.71828183 defined to 8 decimals; 

Load - the intensity of input of pollutants to a given ecosystem from the environment; atmospheric 

load - the intensity of input from the atmosphere; 

m – meter; 

MBL – marine boundary layer; the air right over the ocean surface, where exchange of mercury 

between the two compartments takes place; 

MethylHg or MeHg or MHg – methylmercury; CH3Hg; 

metric ton – 1000 kg; 

mg – milligram (10 -3 gram); 

Natural emission - mercury input to the atmosphere, which is not connected with current or 

previous human activity; 

ng – nanogram (10 -9 gram); 

pg – picogram (10 -12 gram); 

POPs - Persistent Organic Pollutants; 

ppb – parts per billion; 

ppm - parts per million; 

Pre-industrial state - a conventional term implying the state of the natural mercury cycle before 

the beginning of human industrial activity; in Europe the beginning of a noticeable production and 

consumption of mercury is related to medieval centuries; 

Re-emission - secondary input of mercury to the atmosphere from geochemical reservoirs (soil, sea 

water, fresh water bodies) where mercury has been accumulating as a result of previous and current 

human activity; 

REA – Relaxed Eddy Accumulation, micrometeorological method of measuring flux that allows 

calculating flux based on differences in concentration in up drafts and down drafts, accumulated in 

a reservoir over time. REA is generally employed for trace gases that cannot be analyzed 

instantaneously; 

RGM – Reactive Gaseous Mercury; an operationally defined term for gaseous divalent mercury 

compounds of the type HgXY (typically HgCl2) though in the Arctic perhaps HgBr2; 

Ton = 1 metric tonne see metric tonne; 

UNEP - United Nations Environment Programme; 

US EPA – Environmental Protection Agency of the United States of America; 

USA – United States of America; 

Wet deposition - flux of substance from the atmosphere onto the underlying surface with 

atmospheric precipitation; 

WHO - World Health Organization;


Bibliography 

1. Allegrini, I., De Santis, F., Di Palo, V., Febo, A., Perrino, C., Possanzini, M., and Liberti, 

A., (1987): Annular denuder method for sampling of atmospheric pollutants. Science of the 

Total Environment 67, 1–16. 

2. Appleby,P.G. and Oldfield, F., (1978): The calculation of 210 Pb dates assuming a constant 

rate of supply of unsupported 210 Pb to the sediment. Catena, 5:1-8 

3. Ariya, P.A., Ghalizov, A. and Gidas, A., (2002): Reactions of gaseous mercury with atomic 

and molecular halogens: Kinetics, products studies and atmospheric implications. Journal of 

Physical Chemistry-A, 106, 7310-7320. 

4. Barrie, L. A.; Bottenheim, J. W.; Schnell, R. C.; Crutzen, P. J.; Rasmussen, R. A., (1988): 

Ozone destruction and photochemical reactions at polar sunrise in the lower Arctic 

atmosphere. Nature (London, United Kingdom), 334(6178), 138-41. 

5. Barrie, L.; Platt, U., (1997): Arctic tropospheric chemistry. An overview. Tellus, Series B: 

Chemical and Physical Meteorology, 49B(5), 450-454. 

6. Bemgård A, Colmsjö A, Melin J., (1996): Assessing breakthrough times for denuder 

samplers with emphasis on volatile organic compounds. Journal of Chromatography, 723, 

301-311. 

7. Berg, T.; Bartnicki, J.; Munthe, J.; Lattila, H.; Hrehoruk, J.; Mazur, A., (2001): 

Atmospheric mercury species in the European Arctic: Measurements and modelling. 

Atmospheric Environment, 35(14), 2569-2582. 

8. Berg, T., Sekkesæter, S., Steinnes, E., Valdal, A-K, and Wibetoe, G., (2003):Springtime 

depletion of mercury in the European Arctic as observed at Svalbard, The Science of The 

Total Environment, Volume 304, Issues 1-3, 20 March 2003, Pages 43-51. 

9. Bottenheim, J.W., Personal communication (2002): presented 10 Nov 02, Workshop on 

Ocean-Sea Ice-Snowpack-Atmosphere Interactions Research, Purdue University, USA. 

10. Boudries, H. and Bottenheim, J. W., (2000): Cl and Br atom concentrations during a surface 

boundary layer ozone depletion event in the Canadian High Arctic. Geophysical Research 

Letters 27, 517-520. 

11. Bowling, D. R.; Delany, A. C.; Turnipseed, A. A.; Baldocchi, D. D.; Monson, R. K., (1999): 

Modification of the relaxed eddy accumulation technique to maximize measured scalar 

mixing ratio differences in updrafts and downdrafts. Journal of Geophysical Research, 

[Atmospheres], 104(D8), 9121-9133. 

12. Businger J.A. and Oncley, S.P., (1990): Flux measurement with conditional sampling. 

Journal of Atmospheric and Oceanic Technology 7, pp. 349–352. 

13. Cadle, S.H., Dasch, J.M., and Mulawa, P.A., (1985): Atmospheric concentrations and the 

deposition velocity to snow of nitric acid, sulfur dioxide and various particulate species. 

Atmospheric Environment (19), 1819 - 1827. 

14. Christensen, C.S.; Hummelshoj, P.; Jensen, N. O.; Larsen, B.; Lohse, C.; Pilegaard, K.; 

Skov, H., (2000): Determination of the terpene flux from orange species and Norway 

spruce by relaxed eddy accumulation. Atmospheric Environment, 34(19), 3057-3067.


15. Christensen, J.H. Brandt, J. Frohn, L.M. and Skov, H. “Modelling of mercury with the 

Danish Eulerian Hemispheric Model” Submitted to Atmospheric Chemistry and Physics 

April 2002. 

16. Cobos, Douglas R.; Baker, John M.; Nater, Edward A., (2002): Conditional sampling for 

measuring mercury vapor fluxes. Atmospheric Environment, 36(27), 4309-4321. 

17. Cress, R.G., Williams, M.W., Sievering, H., (1995): Dry deposition loading of nitrogen to 

an alpine snowpack, Niwot Ridge, CO, in: Biogeochemistry of Seasonal Snow-Covered 

Catchments, IAHS press, publication no. 228, Wallington, UK, pp. 33–40. 

18. Desjardins, R.L., (1972): A study of carbon dioxide and sensible heat fluxes using the eddy 

correlation technique. Ph.D. dissertation, Cornell University. 

19. Daugaard, B., (2003) Mercury in the North Atlantic, M.Sc. Thesis, University of Southern 

Denmark, Department of Chemistry. 

20. Ebinghaus, R., Kock, H. H., Schmolke, S. R., (2001): Measurements of atmospheric 

mercury with high time resolution: Recent applications in environmental research and 

monitoring, Fresenius Journal of Analytical Chemistry, Vol 371, 6, 806-815. 

21. Ebinghaus, R., Kock, H.H., Temme, C., Einax, J.W., Lowe, A.G., Richter, A., Burrows, J.P. 

and Schroeder, W.H. (2002): Antarctic Springtime Depletion of Atmospheric Mercury. 

Environmental Science and Technology 36, 1238-1244. 

22. Ernst, E. and Coon, J.T. (2001): Heavy metals in traditional Chinese medicines: A 

systematic review. Clinical Pharmacology and Therapeutics 2001, Vol. 70; Number 6: 497- 

504. 

23. Feilberg A., (2000): Atmospheric Chemistry of Polycyclic Aromatic Compounds with 

Special Emphasis on Nitro Derivates, Ph. D Dissertation, Information Service Department 

Risø. 

24. Garvey, J.G., Hahn, G., Lee, R.V. and Harbison, R.D., (2001): Heavy metal hazards of 

Asian traditional remedies. International Journal of Environmental Health Research, 2001, 

11: 63-71. 

25. Goodsite, Michael E.; Rom, Werner; Heinemeier, Jan; Lange, Todd; Ooi, Suat; Appleby, 

Peter G.; Shotyk, William; van der Knaap, W. O.; Lohse, Christian; Hansen, Torben S. 

High-resolution AMS 14C dating of post-bomb peat archives of atmospheric pollutants. 

Radiocarbon (2001), 43(2B), 495-515. 

26. Goodsite, M.E., Brooks, S.B., Lindberg, S.E. Meyers, T.P Skov, H. and Larsen, M.R.B., 

(2003): The fluxes of reactive gaseous mercury measured with a newly developed method 

using relaxed eddy accumulation. Submitted to Atmospheric Environment, June 2003. 

27. Goodsite, M.E. Plane, J. Skov, H., (2003): A theoretical study of the oxidation of Hg 0 to 

HgBr2 in the troposphere. Submitted to Environmental Science and Technology, June 2003. 

28. Gormley P., and Kennedy, M., (1949): Diffusion for a stream flowing through a cylindrical 

tube. Proceedings Royal Irish Academy, 52A:163-167. 

29. Grandjean P, Weihe P, Jørgensen PJ, Clarkson T, Cernichiari E, Viderø T., (1992): Impact 

of maternal seafood diet on fetal exposure to mercury, selenium, and lead, Archives of 

Environmental Health, 47: 185-95. 

30. Grandjean P, Weihe P, Needham LL, Burse VW, Patterson DG Jr, Sampson EJ, Jørgensen 

PJ, Vahter M., (1995): Effect of a seafood diet on mercury, selenium, arsenic, and PCBs and 

other organochlorines in human milk. Environmental Research, 71: 29-38.


31. Grandjean, P., Weihe, P., White, R.F., Deves, F., Araki, S., Yokoyama, K., Murata, K., 

Sorensen, N., Dahl, R. and Jorgensen, P.J. (1997): Neurotoxicology and Teratology 1997, 

20, 1. 

32. Grandjean, Philippe; Budtz-Jorgensen, Esben; Steuerwald, Ulrike; Heinzow, Birger; 

Needham, Larry L.; Jorgensen, Poul J.; Weihe, Pal., (2003): Attenuated growth of breast-fed 

children exposed to increased concentrations of methylmercury and polychlorinated 

biphenyls. FASEB Journal (2003), 17(6), 699-701. 

33. Grieg, G. Gunning, H.E. and Strausz. (1970) Reactions of metal atoms. II. Combination of 

Mercury and Bromine Atoms and the Dimerization of HgBr.The J. Chem. Phys Lett. 52, 7. 

3684-3690. 

34. Guenther, Alex; Baugh, William; Davis, Ken; Hampton, Gary; Harley, Peter; Klinger, Lee; 

Vierling, Lee; Zimmerman, Patrick; Allwine, Eugene; et al. Isoprene fluxes measured by 

enclosure, relaxed eddy accumulation, surface layer gradient, mixed layer gradient, and 

mixed layer mass balance techniques. Journal of Geophysical Research, [Atmospheres] 

(1996), 101(D13), 18555-18567. 

35. Gårdfeldt, K., Sommar, J., Ferrara, R., Ceccarini, C., Lanzillotta, E., Munthe, J., Wängberg, 

I., Lindqvist, O., Pirrone, N., Sprovieri, F., et al., Evasion of mercury from coastal and open 

waters of the Atlantic Ocean and the Mediterranean Sea, Atmospheric Environment, In 

Press, Corrected Proof, Available online 23 May 2003, . 

36. Hansen, J.C. and Pedersen, H.S. (1986): Environmental exposure to heavy metals in North 

Greenland. Arctic Medical Research 41: 21-34. As cited in AMAP, 1998. 

37. Hering, S. V.; Lawson, D. R.; Allegrini, I.; Febo, A.; Perrino, C.; Possanzini, M.; Sickles, J. 

E., II; Anlauf, K. G.; Wiebe, A.; et al. (1988): The nitric acid shootout: field comparison of 

measurement methods. Atmospheric Environment (1967-1989), 22(8), 1519-39. 

38. Hylander, L.D., and Meili, M., (2003): 500 years of mercury production: global annual 

inventory by region until 2000 and associated emissions, The Science of The Total 

Environment, Volume 304, Issues 1-3, Pages 13-27. 

39. Hylander, L.D., Sollenberg, H., and Westas, H., (2003): A three-stage system to remove 

mercury and dioxins in flue gases, The Science of The Total Environment, Volume 304, 

Issues 1-3, Pages 137-144. 

40. Johansson C., and Granath, L., (1986): An experimental study of the dry deposition of 

gaseous nitric acid to snow. Atmospheric Environment 20, 1165 – 1170. 

41. Kaimal, J.C., and Gaynor, J.E., (1991): Another look at sonic thermometry. Boundary Layer 

Meteorology, 56, 401-410. 

42. Karlsson, E., and Nyholm, S., (1998): Dry deposition and desorption of toxic gases to and 

from snow surfaces, Journal of Hazardous Materials, Volume 60, Issue 3, Pages 227-245. 

43. Lamborg, C. H., Fitzgerald, W. F., O’Donnell, J. and Torgersen, T. (2002): A non-steadystate 

compartmental model of global-scale mercury biogeochemistry with interhemispheric 

atmospheric gradients. Geochimica et Cosmochimica Acta 66 (7), 1105-1118. 

44. Landis, M.S., Personal communication (2001): US EPA, Research Triangle Park, North 

Carolina, USA. 

45. Landis, M.S., Stevens, R.K., Schaedlich, F. and Prestbo, E.M., (2002): Development and 

characterization of an annular denuder methodology for the measurement of divalent 

inorganic reactive gaseous mercury in ambient air. Environmental Science and Technology 

36, pp. 3000–3009.


46. Levin I, Kromer B, Schoch-Fischer H, Bruns M, Münnich M, Berdau D, Vogel JC, Münnich 

KO. (1997): (Personal Communication) 14 CO2 records from two sites in Central— 

Schauinsland & Vermunt. URL: . 

47. Levin I, Hesshaimer V., (2000): Radiocarbon—a unique tracer of global carbon cycle 

dynamics. Radiocarbon 42(1):69-80. 

48. Lin, C-J. and Pehkonen, S. O. (1999): The chemistry of Atmospheric Mercury: a review. 

Atmospheric Environment, 33, 2067-2079. 

49. Lindberg, S.E., Brooks, S., Lin, C.-J., Scott, K., Meyers, T., Chambers, L., Landis, M. and 

Stevens, R., 2001. Formation of reactive gaseous mercury in the Arctic: evidence of 

oxidation of Hg 0 to gas-phase Hg-II compounds after Arctic sunrise. Water, Air, and Soil 

Pollution: Focus 1, pp. 295–302. 

50. Lindberg, S. E., Brooks, S., Lin, C. J., Scott, K. J., Landis, M. S., Stevens, R. K., Goodsite, 

M. and Richter, A. (2002): Dynamic oxidation of gaseous mercury in the Arctic troposphere 

at polar sunrise. Environmental Science & Technology 36, 1245-1256. 

51. Lindqvist, O., and Rodhe, H. (1985): Atmospheric Mercury – a Review. Tellus Series B – 

Chem. Phys. Meteorol., 37(3), 136-159. 

52. Lu, J. Y., Schroeder, W. H., Barrie, L. A., Steffen, A., Welch, H. E., Martin, K., Lockhart, 

L., Hunt, R. V., Boila, G. and Richter, A., (2001): Magnification of atmospheric mercury 

deposition to polar regions in springtime: the link to tropospheric ozone depletion chemistry. 

Geophysical Research Letters 28, 3219-3222. 

53. Madsen, Poul Pheiffer. (1981): Peat bog records of atmospheric mercury deposition. Nature 

293 (5828), 127-30. 

54. Meyers, T., (2001, 2003): Personal communication, NOAA, ATDD, Oak Ridge TN. 

55. Munthe, J., Wängberg, I., Pirrone, N.,. Iverfeldt, Å, Ferrara, R., Ebinghaus, R., Feng, X., 

Gårdfeldt, K., Keeler, G., Lanzillotta E., et al., (2001): Intercomparison of methods for 

sampling and analysis of atmospheric mercury species, Atmospheric Environment, Volume 

35, Issue 17, Pages 3007-3017. 

56. Munthe, J., Wängberg, I., Iverfeldt, Å., Lindqvist, O., Strömberg, D., Sommar, J., Gårdfeldt, 

K., Petersen, G., Ebinghaus, R., Prestbo, E., et al., Distribution of atmospheric mercury 

species in Northern Europe: final results from the MOE project, Atmospheric Environment, 

In Press, Corrected Proof, Available online 23 May 2003. 

57. Noernberg, T., Goodsite, M.E., Shotyk, W., (in review): An Improved Motorized Corer and 

Sample Processing System for Frozen Peat. In review at Arctic. 

58. Oncley, Steven P.; Delany, Anthony C.; Horst, Thomas W.; Tans, Pieter P. Verification of 

flux measurement using relaxed eddy accumulation. Atmospheric Environment, Part A: 

General Topics (1993), 27A(15), 2417-26. 

59. Pacyna, E., Pacyna, J.M. and Pirrone, N. (2000): Atmospheric Mercury Emissions in Europe 

from Anthropogenic Sources. Atmospheric Environment 35, 2987-2996. 

60. Poissant, L., Pilote, M., (2001): Atmospheric mercury and ozone depletion events observed 

at low latitude along the Hudson Bay in northern Quebec (Kuujjuarapik: 55°N). Book of 

Abstracts (AT-23), Sixth International Conference on Mercury as a Global Pollutant, 

Minamata, Japan, October 15–19. 

61. Possanzini, M., Febo, A., and Liberti, A. New design of a high performance denuder for the 

sampling of atmospheric pollutants. Atmospheric Environment 17 (1983), pp. 2605–2610.


62. Roos-Barraclough F; Givelet N; Martinez-Cortizas A; Goodsite M E; Biester H; Shotyk W., 

(2002): An analytical protocol for the determination of total mercury concentrations in solid 

peat samples. Science of the Total Environment, 292(1-2), 129-39. 

63. Salvato N, Pirola C., (1996): Analysis of mercury traces by means of solid samples atomic 

absorption spectrometry, Microchimica Acta, 123(1–4):63 –71. 

64. Schroeder, W. H. and Munthe, J. (1998): Atmospheric Mercury - An Overview. 

Atmospheric Environment 32, 809-822. 

65. Schroeder, W. H., Anlauf, K. G., Barrie, L.A., Lu, J.Y. and Steffen, A. (1998): Arctic 

springtime depletion of mercury. Nature 394, 331-332. 

66. Schroeder, W. H., Steffen, A., Scott, K., Bender, T., Prestbo, E., Ebinghaus, R., Lu J. Y., 

and Lindberg, S. E., (2003): Summary report: first international Arctic atmospheric mercury 

research workshop, Atmospheric Environment, Volume 37, Issue 18, Pages 2551-2555. 

67. Schuster, P.F., Krabbenhoft, D.P., Naftz, D.L., Cecil, L.D., Olson, M.L., Dewild, J.F., 

Susong, D.D., Green, J.R. and Abbott, M.L., (2002): Atmospheric Mercury Deposition 

during the Last 270 Years: A Glacial Ice Core Record of Natural and Anthropogenic 

Sources. Environmental Science and Technology 36 (11), 2303 -2310. 

68. Seigneur, C., Karamchandani, P., Lohman, K., Vijayaraghavan, K. and Shia, R-L. (2001): 

Multiscale modeling of the atmospheric fate and transport of mercury. Journal of 

Geophysical Research 106 (D21), 27795-27809. 

69. Sheu, G.P., and Mason, R.P., (2001): An Examination of Methods for the Measurements of 

Reactive Gaseous Mercury in the Atmosphere. Environmental Science and Technology, 35 

(6), 1209 -1216. 

70. Shotyk, W., Goodsite, M.E., Roos-Barraclough, F., Frei, R., Heinemeier, J., Asmund, G., 

Lohse, C., Hansen, T.S., (in-press): Anthropogenic contributions to atmospheric Hg, Pb and 

As accumulation recorded by peat cores from southern Greenland and Denmark dated using 

the 14C “bomb pulse curve”. Geochimica et Cosmochimica Acta accepted 02 June, 2003. 

71. Shotyk, W., Goodsite, M.E., Roos-Barraclough, F., Givelet, N., LeRoux, G., Weiss, D., 

Norton, S., Knudsen, K., and Lohse, C., (2003): Unpublished data. Manuscript in 

preparation “Atmospheric Mercury and Lead Accumulation Since 5420 14 C yr BP at 

Myrarnar, Faroe Islands” 

72. Skov, H. (2001) Transport of atmospheric mercury from mid-latitudes to the Arctic A model 

and Measurements study. In: EUROTRAC-2 MEPOP Atmospheric cycling of mercury and 

persistent organic pollutants subproject description. International Scientific Secretariat GSF- 

Forschungszentrum für Umwelt and Gesundheit GmbH; Munich, Germany. 

73. Skov, H. Nielsdóttir, M.C. Goodsite, M.E. Christensen, J. Skjøth, C.A. Geernaert, G. Hertel, 

O. Olsen, J. (2003 1 ) “Measurements and modelling of gaseous elemental mercury on the 

Faroe Islands”. Accepted. Asian Chemistry Letters. 

74. Skov, H., Christensen, J., Goodsite, M.E., Heidam, N.Z., Jensen, B., Wåhlin, P., Geernaert, 

G. The Fate of Elemental Mercury in Arctic during Atmospheric Mercury Depletion 

Episodes and the Load of Atmospheric Mercury to Arctic. Submitted to Environmental 

Science and Technology, June 2003. 

75. Slemr, F., Schuster, G. and Seiler, W., (1985): Distribution, speciation and budget of 

atmospheric mercury. Journal of Atmospheric Chemistry 3, pp. 407–434.


76. Slemr, F.; Langer, E., (1992): Increase in global atmospheric concentrations of mercury 

inferred from measurements over the Atlantic Ocean. Nature (London, United Kingdom), 

355(6359), 434-7. 

77. Slemr, F.; Junkermann, W.; Schmidt, R.W.H.; Sladkovic, R., (1995): Indication of change in 

global and regional trends of atmospheric mercury concentrations. Geophysical Research 

Letters, 22(16), 2143-6. 

78. Slemr, F., E.-G. Brunke, R. Ebinghaus, C. Temme, J. Munthe, I. Wängberg, W. Schroeder, 

A. Steffen, and T. Berg, (2003): Worldwide trend of atmospheric mercury since 1977, 

Geophysícal Research Letters, 30(10), 1516. 

79. Slinn, W.G., Hasse, L., Hicks, B.B., Hogan, A.W., Lal, D., Liss, P.S., Munnich, K.O., G.A. 

Sehmel, G.A., and Vittori, O. (1978): Some aspects of the transfer of atmospheric trace 

constituents past the air–sea interface. Atmospheric Environment 15, p. 863. 

80. Steffen A., and Schroeder, W., (1999): Standard Operation Procedures Manual For Gaseous 

Elemental Mercury Measurements. Canadian Atmospheric Environment Service, 

Meteorological Service of Canada, 4905 Dufferin Street, Toronto, Ontario, Canada M3H 

5T4. 

81. Steffen, A., Schroeder, W., Bottenheim, J., Narayan, J., and Fuentes, J.D., (2002): 

Atmospheric mercury concentrations: measurements and profiles near snow and ice surfaces 

in the Canadian Arctic during Alert 2000, Atmospheric Environment, Volume 36, Issues 15- 

16, Pages 2653-2661. 

82. Steuerwald, Ulrike; Weihe, Pal; Jorgensen, Poul J.; Bjerve, Kristian; Brock, John; Heinzow, 

Birger; Budtz-Jorgensen, Esben; Grandjean, Philippe., (2000): Maternal seafood diet, 

methylmercury exposure, and neonatal neurologic function. Journal of Pediatrics (St. Louis), 

136(5), 599-605. 

83. Sørensen, N., Murata, K., Budtz-Jorgensen, E., Weihe, P. and Grandjean, P. (1999): Prenatal 

methylmercury exposure as a cardiovascular risk factor at seven years of age. Epidemiology 

1999; 10: 370-375. 

84. Tarasick, D. W.; Bottenheim, J. W., (2002): Surface ozone depletion episodes in the Arctic 

and Antarctic from historical ozonesonde records. Atmospheric Chemistry and Physics 

[online computer file], 2: 197-205. 

85. Valdez, M.P., Bales, R.C., Stanley D.A., and Dawson, G.A., (1987): Gaseous deposition to 

snow: 1. Experimental study of SO2 and NO2 deposition. Journal of. Geophysical Research 

92, p. 9779 – 9787. 

86. Wardenaar ECP., (1987): A new hand tool for cutting peat. Canadian Journal of Botany, 

65:1772 –1773. 

87. Weihe, P., Hansen, JC., Murata, K., Debes, F., Jørgensen, PJ., Steuerwald, U., White, RF. 

And Grandjean, P. (2002): Neurobehavioral Performance of Inuit Children with Increased 

Prenatal Exposure to Methylmercury. International Journal of Circumpolar Health 2002; 

61: 41-9. 

88. Williston, S.H., (1968): Mercury in the atmosphere. Journal of Geophysical Research, 73: 

7051-7055. 

89. Yamada, M., Tohno, S., Tohno, Y., Minami, T., Ichii, M., Okazaki, Y. (1995): 

Accumulation of mercury in excavated bones of two natives in Japan. Science of the Total 

Environment 1995; 162:253-256.


90. Zannetti, P., Air Pollution Modelling. Theories, Computational Methods and Available 

Software, Van Nostrand-Reinhold, New York, 1990. 

91. Zhu, T.; Pattey, E.; Desjardins, R. L. Relaxed Eddy-Accumulation Technique for 

Measuring Ammonia Volatilization. Environmental Science and Technology (2000), 

34(1), 199-203.


Appendix A List of Papers


Paper 1: Goodsite, M.E., Brooks, S.B., Lindberg, S.E. Meyers, T.P Skov, H. and Larsen, M.R.B. 

(2003) The Fluxes of Reactive Gaseous Mercury measured with a newly developed method 

using Relaxed Eddy Accumulation. Submitted to Atmospheric Environment, June 2003. 

Paper 2: Lindberg, Steve E.; Brooks, Steve; Lin, C.-J.; Scott, Karen J.; Landis, Matthew S.; 

Stevens, Robert K.; Goodsite, Mike; Richter, Andreas. Dynamic Oxidation of Gaseous 

Mercury in the Arctic Troposphere at Polar Sunrise. Environmental Science and Technology 

(2002), 36(6), 1245-1256. 

Paper 3: Skov, H., Christensen, J., Goodsite, M.E., Heidam, N.Z., Jensen, B., Wåhlin, P., 

Geernaert, G. The Fate of Elemental Mercury in Arctic during Atmospheric Mercury 

Depletion Episodes and the Load of Atmospheric Mercury to Arctic. Submitted to 

Environmental Science and Technology, June 2003. 

Paper 4: Goodsite, M.E., Plane, J. Skov, H. 2003. A theoretical study of the oxidation of Hg 0 to 

HgBr2 in the troposphere. Submitted to Environmental Science and Technology, June 2003. 

Paper 5: Skov, H., Nielsdóttir, M.C., Goodsite, M.E., Christensen, J., Skjøth, C.A., Geernaert, G., 

Hertel, O. Olsen, J., Measurements and Modelling of gaseous elemental mercury (GEM) on 

the Faroe Islands; a case study of the difficulties of measuring GEM. Accepted Asian 

Chemistry Letters (2003). 

Paper 6: Goodsite, Michael E.; Rom, Werner; Heinemeier, Jan; Lange, Todd; Ooi, Suat; Appleby, 

Peter G.; Shotyk, William; van der Knaap, W. O.; Lohse, Christian; Hansen, Torben S. Highresolution 

AMS 14C dating of post-bomb peat archives of atmospheric pollutants. 

Radiocarbon (2001), 43(2B), 495-515. 

Paper 7: Roos-Barraclough F; Givelet N; Martinez-Cortizas A; Goodsite M E; Biester H; Shotyk 

W. An analytical protocol for the determination of total mercury concentrations in solid 

peat samples. Science of the Total Environment (2002 Jun 20), 292(1-2), 129-39. 

Paper 8: Shotyk, W., Goodsite, M.E., Roos-Barraclough, F., Frei, R., Heinemeier, J., Asmund, G., 

Lohse, C., Hansen, T.S. Anthropogenic contributions to atmospheric Hg, Pb and As 

accumulation recorded by peat cores from southern Greenland and Denmark dated using 

the 14C “bomb pulse curve”. Geochimica et Cosmochimica Acta accepted 02 June, 2003. 

Paper 9: Noernberg, T., Goodsite, M.E., Shotyk, W., An Improved Motorized Corer and 

Sample Processing System for Frozen Peat. In review at Arctic.


Other relevant publications not included in this thesis: 

Peer reviewed papers, not including those in preparation. 

Ferrari, C.P., Dommergue, A., Boutron, C., Skov, H., Goodsite, M.E., Jensen, B. Night 

Production of Elemental Gaseous Mercury in Interstitial Air of Snow at Station Nord, 

Greenland Shortly After Polar Sunrise. In revision at Atmospheric Environment. 

Technical reports 

Goodsite, M.E., Hermanson, M., Scholten, C., Asmund, G., Skov, H., Bennike, O., Heidam, N.Z., 

Feilberg, A., Geernaert, G.L., Report for the Royal Danish Air Force, Tactical Air Command 

Contaminant impact on the high Arctic environment from a small military airstrip and 

station: Station Nord, NE Greenland. National Environmental Research Institute, Department 

of Atmospheric Environment, Roskilde, Denmark, 2003 

Skov, H., Christensen, J.H. Goodsite, M.E. Petersen, M.C. Zeuthen-Heidam, N. Geernaert, G. and 

Olsen, J. (2001) Dynamics and chemistry of atmospheric mercury. Danish contribution to 

EUROTRAC MEPOP report, 2000. 

Brooks, S.B. Lindberg, S.E., Goodsite, M., Meyers, T.P., McConville, G. Springtime Deposition 

Rates of Atmospheric Mercury at Barrow, Alaska Followed by Partial Re-emission at 

Snowmelt. The National Oceanic and Atmospheric Research Administration 2000 scientific 

review and report. 

Brooks, S.B. Lindberg, S.E., Goodsite, M., Stevens, R.K., Landis, M., Scott, K., Meyers, T.P., Lin, 

J., and McConville, G. Barrow Arctic Mercury Study The National Oceanic and Atmospheric 

Research Administration 2000 scientific review and report. 

Shotyk, W., Goodsite, M., Givelet, N., Roos-Barraclough, F., Knudsen, K., Asmund, G., 

Cheburkin, A., and Heinemeier, J. Peat Core Records of Natural and Anthropogenic 

Atmospheric Mercury in the Arctic. a short report prepared for the Danish Coorperation for 

Environment in the Arctic (DANCEA). 

N. Givelet, W, Shotyk and M. Goodsite. Long term records of atmospheric deposition of Hg, 

Cd, Pb and Persistent Organic Pollutants (POPs) in Peat Cores from Arctic peatlands 

(Bathurst Island). A progress report to the International Arctic Research Center (IARC). 

Roos, F., Goodsite, M.E., Knudsen, K., and W. Shotyk. The Investigation and Dating of 

transboundary air pollution found in the Faroe Islands. A progress report to the 

International Arctic Research Center (IARC). 

Goodsite, M.E., Shotyk, W. (2000) Chronology of atmospheric Pb fluxes recorded by peat 

profiles in Southern Greenland and Denmark.) Final report, submitted to H. von Storch, 

GKSS (Germany). 

Field Reports 

Goodsite, M.E., Bennike, O., Warncke, E., Nørnberg, T. Post Expedition Field and Status 

Report: Long term records of atmospheric deposition of Hg, Cd, Pb and Persistent 

Organic Pollutants (POPs) in Peat Cores from Arctic peatlands (Nordvestø, Carey Islands, 

Greenland). A field report to the Danish Polar Center, and DANCEA (2001). 

Goodsite, M.E., Shotyk, W., Post Expedition Field and Status Report: Long term records of 

atmospheric deposition of Hg, Cd, Pb and Persistent Organic Pollutants (POPs) in Peat 

Cores from Arctic peatlands (Bathurst Island). A field report to the International Arctic 

Research Center (IARC) and the Polar Continental Shelf Project. (2000).


F. Roos-Baraclough, K. Knudsen, M. Goodsite, and W. Shotyk. Post Expedition Field and Status 

Report: The Investigation and Dating of Transboundary Air Pollution found on the Faroe 

Islands. A field report to the Danish Polar Center and DANCEA. (2000). 

M. Goodsite. Post Expedition Field and Status Report: The Investigation and Dating of 

Transboundary Air Pollution found in Southern Greenland. A field report to the Danish 

Polar Center and DANCEA. (1999). 

Popular Science 

Goodsite, M., Contributing expert in: Arctic Pollution 2002, AMAP, OSLO. ISBN 82-7971-015-9 

Goodsite, M., Shotyk, W. & Nielsdóttir, M.C. (2002): Notat: “Torv kanningar av kyksilvuri í 

Føroyum”. In: Mikkelsen, B., Hoydal, K., Dam, M. og Danielsen, J. 2002. ”Føroya Umhvørvi í 

Tølum 2001” Heilsufrøðiliga Starvstovan, rapport nr. 2002:1, pp 22 (in Faroese). 

Goodsite, M.E., “Atombombsprængninger kan bruges til datering af forurening i de arktiske 

områder” Fønix, Nr. 4. 2001, (in Danish), NERI, Denmark. 

Goodsite, M.E., ”Få bedre forskningsøkonomi ved at tænke som en sportsklub” Fønix, Nr 3. 2001, 

(in Danish), NERI, Denmark.


Appendix B Field Work: planning and post expedition report format


The purpose of conducting environmental sampling field work in the arctic is to expeditiously, 

effectively and efficiently collect the environmental samples required to carry out the proposed and 

permitted research with a minimum impact to the fragile and pristine environment. During the 

conduct of this Ph.D. field work was conducted at three separate high Arctic locations under 3 

different country jurisdictions as well as sub arctic or continental locations in 4 different countries. 

While the expedition guidelines discussed below can in the strictest sense be applied to any 

operation, it is up to the leader/planner to adjust the guidelines to ensure success in their own 

assignments. It is furthermore cautioned that these suggestions do not replace any agency’s 

guidelines or training; they are here to give an impression of the process used while carrying out the 

field work in this Ph.D. thesis. 

The high Arctic can is an unforgiving environment, apparently devoid of most resources 

normally associated as essential to survival (such as firewood). It is a demanding work environment 

and requires careful planning for success. All Arctic research support agencies have advisors 

available to assist with planning and logistics, and these agencies also have their own set of 

operating guidelines, though it should be noted that nearly all guidelines build upon the same 

purpose and principles mentioned above. Never the less it should be noted that even the most 

carefully developed plans must be readily open for change and back up sites and plans are essential. 

Arctic weather conditions WILL change schedules and tax men and material resources. Flexibility 

and patience must be key words when working in the Arctic. Attitude is an often forgotten attribute. 

The will to survive, as well as believing that there are “no problems, only challenges” will carry one 

through many situations, even if training and routine are not completely up to speed, though 

certainly good training instills routine and confidence!


It is incumbent of the expedition leader (EL), who may or may not also be the primary 

scientific investigator to coordinate with the appropriate permitting and control agencies in the area 

of operations well in advance. Some typical permitting requirements include taking of geologic or 

biologic samples, weapons and radio permits, wildlife permits, over-flight permits, military base 

transit permits, insurance...each area in the Arctic has specific requirements and many have separate 

permit issuing coordinators or jurisdiction depending if the purpose is scientific exploration or 

mineralogical (economic geological or private) exploration. 

It is appropriate to note that submitting an application to one of these agencies as the responsible EL 

carries responsibility that must careful be considered namely: the expedition leader is responsible to 

the agency for the safety of the team and the protection of the environment, as well as adherence to 

the conditions of the permit. The expedition leader may (and in cases of larger or extended 

expeditions should) delegate authority to carry out routine tasks (such as radio or wildlife watches) 

but should be aware that responsibility can not be delegated and therefore the leader will be 

ultimately responsible for all the team accomplishes or fails to accomplish. The 11 general planning 

steps utilized are described as below. These steps expand upon and modify a military (8 step) troop- 

leading procedure: 

1) Receive the assignment: 

Ensure that the assignment is clear and well defined. Conduct an initial analysis of the assignment 

defining the: Purpose of the field work, the minimum equipment requirements to accomplish the 

field work, the type of terrain and weather in the area of operations (check this against maps, photos 

and meteorological data available for the area) and cross check this with equipment requirements; 

the personnel and expertise required to carry out the proposed field work. Identify absolutely 

assignment essential and secondary personnel to the team. Team members must know their roles


and must be chosen carefully. In the event of unplanned circumstances the team must be able to 

function effectively they must be able to cross function in the event someone is not able to function. 

In times when the work is hindered due to weather, the team should be able to spend many hours 

together in typically confined environments without too much mental discomfort. Identify the 

amount of time available for planning and for the operation its-self. Use no more than 1/3 of the 

planning time available for your own preliminary purposes. Leave 2/3 of the time to finish 

preparing for the field work with your team. It is generally helpful to think through the above 

process using reverse planning: i.e., use the desired end-state as a planning start point, and bring 

back the planning to the present, establishing a list over the tasks that must be accomplished prior to 

leaving for the field to ensure the team is prepared for success. At this stage of the planning process, 

the desired end state should be getting to the field location successfully. Lastly look at the planned 

budget give the tentative resource needed list. Make an honest judgement at this point in time if the 

desired work can be accomplished with the resources available. Contacting agencies and verify 

with experts that there is a reasonable chance for the operation to obtain permits at the area, with the 

time and resources available: Have a time range to work with and alternate sites to propose to the 

granting agencies. Solidify a tentative timeline based on no adverse information relating to wildlife, 

weather or logistical factors which the agencies know to be problematic for the area of operation. 

2) Alert your team: 

Address items that must be prepared for by the team to accomplish the assignment. Involve 

alternates already at this stage. A team that is used to working together, or has well defined job 

assignments and routines will know to initiate these routines at this point in time. Ensure that your 

team is aware of the nature and purpose of assignment, any special instructions (and to whom) and 

when the planned date is. Remind them to check already now that they have current certifications 

and passports (at the planned time of the field work).


3) Make a tentative plan to accomplish your assignment: 

Accomplish a more careful and complete analysis with your assistant expedition leader, find and 

define the explicit tasks to accomplish the field work (e.g., sample peat) and the implied tasks (i.e., 

need a peat sampler if don’t already have one!). Remember to take into account the weight of 

samples leaving when planning for what can be brought out coming. Consult with a pilot if flying to 

determine if fuel weight loss compensates from the in-bound flight compensates for this (if they 

will fly with less than a full tank). Determine the need for fuel and store depots and make a plan for 

their pre-placement (to be initiated upon receipt of permits). Estimate how weather and terrain will 

affect your mobility and survivability. Task-organize your planned field work and teams. Allow for 

flexibility! Develop primary and alternate courses of action. Talk these through with your assistant. 

Make a decision as to the primary plan. Update the appropriate agencies of your finalized plan and 

team list. 

4) Initiate transport of equipment and personnel 

Cargo must often be requisitioned long in advance, if borrowed from agency stores. Other cargo 

should be shipped well in advance since cargo will have a lower priority on flights in and out of the 

Arctic than personal equipment (due to safety considerations). Flights (commercial or military) in 

and out of the Arctic are always weather dependent and personnel movement should allow and plan 

for delays prior to the start of the assignment. Ensure that control agencies, charters or guides in the 

fields are informed of any changes that may occur. Upon arrival to the Arctic airport, check 

equipment and personnel. Ensure that all assignment and survival essential gear is functioning. 

Communications and weapons are key tools. Ensure that everyone on the team knows how they 

function. Don’t mix equipment – your equipment should function as a back-up to the other, so that 

you always have one set functioning. Communications should be checked with the local controlling 

authority. Leave a map with your routes and plan and leave your frequency and contact plan with


them – write it on the map and ensure they have at least two copies. Ensure that you and your team 

understand the procedures that the controlling authorities will initiate in case of: emergency or if the 

team fails to report in with a periodic radio report (QSO). Ensure that all know the call signs and 

proper radio procedures. If you will need to talk with a satellite camp, pre-plan and request and 

permit internal frequencies off the QSO net. Check you primary navigational and survival aids. Are 

GPS’s set for the same datum and coordinate system as your maps (and just as importantly, the 

control agencies maps!) are your compasses properly corrected for declination? 

5) When possible, reconnoitre to verify the tentative plan: 

If nothing else, talk with the people who live there or have just returned form the field, get their 

opinion. Listen more than you speak. 

6) Complete the plan: 

At this point you have all of the input you need for successful fieldwork. 

7) Make a written day to plan and issue it to your team: 

Ensure that they keep it with their important papers. The plan should carry frequencies and 

reminders of emergency procedures. It should include basic medical data (at least blood type) and 

next of kin information for each of the members. All should keep this in the same place – so that it 

is easily found in case of need. Ensure the controlling authorities receive copies of the plan. 

8) Prior to leaving inspect one final time: 

Personal survival kits, medical/insurance cards and working permits/passports, weapons and 

ammunition, clothing (all should have what is on the packing list, and not more or less, if this is a 

weight consideration), equipment (sleeping bags: zippers work? Tents enough pegs etc., assignment 

essential equipment and back-up to this equipment. Check Radios, extra batteries, communication 

frequencies and protocols. Also ensure Non-radio (visual) communication protocols and sets (signal


mirrors, flags, smoke and flares), as well as first aid kits, food and water. While inspecting ask your 

team about the assignment and their responsibilities ensure that they stay focussed. 

9) Conduct the field work in accordance with permits and the work plan: 

Note deviations to the plan as they (and they will in the Arctic) arrive. Supervise, remain proactive! 

10) Return the base and notify the control authorities of your safe return: 

Initiate return cargo and proper keeping or pre-processing of samples. Recuperate. 

11) Document the expedition with a post expedition field and status report: 

Return to authorities per their instruction (generally 30-90 days post expedition). Use their specified 

format. I have developed the following format based on my experience and it is generally 

acceptable for turn in (given the proper cover sheets) to Danish, Canadian or American scientific 

permitting authorities. 

The field report is submitted by the Expedition Leader to the Project Leader (who is ultimately 

responsible to the funding agencies) and permitting agencies in fulfilment of the requirements as set 

forth in the permit from the issuing authorities. It states why the expedition was carried out and 

what is to follow. It is submitted to the respective agencies, which have supported the project or 

have substantial interest in its results. It primarily provides a public (or status-file) record of the 

expedition, detailing how and where environmental samples were taken for later analysis. The 

description should be complete enough to allow another team to resample or “reproduce” the 

expedition. It should clearly show that all permitting requirements were upheld and that the 

environment was not subjected to damage. In the case of accident or damage, it should clearly 

document the circumstances seen from the team’s point of view, and provide evidence that the 

proper post-incident procedures and notifications were followed.


Additionally, the present status of the samples is given. The complete project descriptions and 

research plans are generally not included (or included only as an appendix) as they are available 

through the project leaders and respective funding agencies, and where probably submitted with 

permit requests. The report should acknowledge the help of those who made it successful and of the 

local population whose land one is investigating as well as support and funding. 

An example table of contents follows with the main points and appendixes included in the report. 

Example Table of Contents 

Preface 

Table of Contents 

Abstract 

Introduction and Background 

Site localization 

Expeditionary Team and Plan 

Project Timeline and Plan (Summary) 

General Notes 

Day for Day Chronology (Log) 

Recon and field notes 

Conclusion 

Literature 

Appendix 

1. Complete scientific project timeline and plan 

2. Complete field work operation and sampling plan


3. Photo list 

4. Map sheets/air photos 

5. Observations (Weather and Biota) 

6. Equipment and packing list 

7. Copies of permits 

GENERAL NOTES: Nearly all reports discuss certain generalities as follow: 

1. TRASH: There was at no time food or trash left trash in the field. None of the party 

members were smokers, so this includes cigarette butts as well. 

2. ELECTRICITY/FUEL: On site fuel supplies were used but replaced with fresh fuel flown 

into the site in accordance with the permit. 

3. LODGING/DEPOTS: Field camp/depots established in accordance with permits with 

minimum impact on the environment. Camp fires established how and why, cleaned up 

how. 

4. ROUTE PLANNING: Routes were reviewed and approved by local authorities and were not 

deviated from with out proper notification. 

5. SAFETY: All party members carried maps, compass, medical cards and personal first aid 

kits. The party always had 2 GPS with, as well as an ANNA (Arctic signaling and survival) 

kit (2 if there were four in the Field) if we were four in the field or over water movements, 

our Guide had VHF radio and mobile telephone with as well), Fresh water, and emergency 

rations. The authorities were informed of the routes, and planned time of return. Return 

times were held. Party members moved in groups of 2. Weather was monitored constantly, 

with weather reports being viewed nightly, and the weather station readings monitored


continuously, and reviewed for trends nightly. Weapons were handled in accordance with 

standard operating procedures submitted with the weapons permit request. 

6. SANITATION AND HYGIENE: Urination or defecation in the field was not done into 

waterways or water sources or within 50 m upland from them. Defecation was into bags and 

carried out (or burned) in accordance with permit. 

7. NAVIGATION: All points read with a XXXX Portable GPS Ser. Nr. XXXX, set to Datum 

XX and the XXXX coordinate system: Points read were plotted in with back azimuths onto 

the (topographic maps) or photos (set to the local declination) by the expedition leader, and 

controlled thereafter by the assistant expedition leader. 

8. CONTACT WITH WILDLIFE: (Unless permitted to do so) wildlife was avoided as were 

there trails. No contact or feeding of wildlife. Trash and waste handled appropriately to 

minimize risk for attracting wildlife or predator encounters. 

Successful field work is conducted following good planning. A good plan will include a definite 

course of action and define a method of execution. The plan will include a risk-analysis and 

measures to minimize risks, as well as a cost-benefit analysis. The planning process includes 

training in skills that might be perishable. Part of a successful expedition is good documentation 

upon return. This aids the team members who were not a part of the field team and can serve as a 

training aid for future teams. As with a laboratory log documenting experiments, the field report 

should be complete enough to reproduce the field work.


Appendix C Supplementary material


This appendix contains papers produced relative to this thesis as where this author was first 

author or a co-author. The work done in this thesis was international, and involved many different 

scientific expertise. It is therefore appropriate to review the scientists, without whose efforts, much 

of this work could not be accomplished. 

The Relaxed Eddy Accumulation development was carried out under the total supervision of Dr. 

Henrik Skov at The National Environmental Research Institute of Denmark, Department of 

Atmospheric Environment, with subsequent development in Oak Ridge, USA at the National 

Oceanic and Atmospheric Administration (NOAA), Atmospheric Turbulence and Diffusion 

Division (ATDD), under the supervision of Dr.’s Tilden Meyers and Steve Brooks. Dr. Alex 

Guenther, US National Center for Atmospheric Research, NCAR, Boulder provided schematics 

photos and field advice for the NCAR REA system, during at a visit to NCAR. 

The RGM KCL coated annular denuder sampling system was tested prior to deployment at Oak 

Ridge National Laboratory, under the supervision and guidance of Dr. Steve Lindberg. Dr. 

Lindberg functioned as this Ph.D.’s overall external supervisor during the stay in Oak Ridge. The 

system was deployed at Walker Branch Research Area under the supervision of Dr. Lindberg, for 

testing prior to deployment to the NOAA Climate Monitoring and Data Laboratory, Barrow, 

Alaska. 

While in Alaska, the system was set up under the supervision of Dr. Brooks, and Denuder 

Analysis under supervision of Dr. Matt Landis, US Environmental Protection Agency. Dr. Robert 

Stevens, Florida Liaison to the US EPA, also advised in sampling system development and 

deployment in the United States and Alaska. 

Initial exposure to GEM and RGM sampling and flux measurement systems was during a visit to 

the Laboratory of Dr. Ralf Ebbinghaus, GKSS, Geesthacht, Germany. Annular denuder preparation 

and analysis was initially instructed by Dr. Weijin Dong, Oak Ridge National Laboratory and


comprehensively taught later by Dr. Landis and Stevens at US EPA National Exposure Laboratory, 

Research Triangle Park, North Carolina. Automatic RGM and total particulate mercury systems 

were learned from TEKRAN, Canada. 

Other aspects of elemental mercury sampling were learned initially from a course at the Swedish 

Environmental Research Institute, IVL, Göteberg, conducted by Dr’s John Munthe, and Ingevar 

Wangberg and comprehensively by Dr. Skov at NERI and through the experience gained in setting 

up, monitoring, maintaining and recovering, gaseous elemental mercury monitoring stations in the 

Faroe Islands, Station Nord, northeast Greenland and Nuuk, west Greenland. The sampling protocol 

employed was developed for MET Canada and provided by its authors, Ms. Sandy Steffen and Dr. 

W. Schroeder. 

The thermodynamics of the proposed mechanism for the oxidation of mercury in the Arctic is 

collaboratively developed with Professor John Plane, University of East Anglia and Henrik Skov, 

while at a research visit at the University of East Anglia. 

Peat research was carried out under the overall supervision of Professor W. Shotyk, formerly of 

the University of Berne, and presently at Heidelberg University. Mercury preparation and analyses 

were carried out at the University of Berne and the University of Heidelberg after comprehensive 

instruction from Dr. Fiona Roos-Barraclough. The idea for trying the bomb pulse in Greenland peat 

is credited to Professor Christian Lohse, University of Southern Denmark. Preparation of materials 

and radiocarbon dating and analyses were performed at the Danish National Radiocarbon AMS 

Laboratory, under the supervision of Professor Jan Heinemeier with additional guidance from 

Professor Emeritus Henrik Loft. Additional samples were analysed collaboratively at the University 

of Arizona National Science Foundation Radiocarbon AMS Laboratory with guidance from 

Professor emeritus Donahue and staff for their time and guidance. Stable lead isotope analyses were


carried out collaboratively by and after consultation with the Danish Isotope Centre, Professor 

Robert Frei.


Paper 1: Goodsite, M.E., Brooks, S.B., Lindberg, S.E. Meyers, T.P Skov, H. and Larsen, M.R.B. 

(2003) The Fluxes of Reactive Gaseous Mercury measured with a newly developed method 

using Relaxed Eddy Accumulation. Submitted to Atmospheric Environment, June 2003.

Submitted to Atmospheric Environment 

The Fluxes of Reactive Gaseous Mercury Measured with a Newly Developed Method using 

Relaxed Eddy Accumulation 

Michael E. Goodsite a,b,d,e* , Henrik Skov a , Steve B. Brooks c , Steve E. Lindberg d , Tilden P. 

Meyers e , Matt Landis f , Michael R.B. Larsen a,e , Glen McConville g 

a National Environmental Research Institute, Frederiksborgvej 399, 4000 Roskilde, Denmark 

b Present address: MEG: Department of Chemistry University of Southern Denmark, Campusvej 55, 

5230 Odense M., Denmark 

c Oak Ridge Associated Universities, P.O. Box 117, 

Oak Ridge, Tennessee 37831-0117 

d Environmental Sciences Division, Oak Ridge National 

Laboratory, Oak Ridge, Tennessee 37831-6038 

e Atmospheric Turbulence and Diffusion Division, National Oceanic and Atmospheric 

Administration, Oak Ridge, Tennessee 37831-0117 

f U.S. EPA, 79 TW Alexander Drive, 

Human Exposure Analysis Branch, MD-56, 

Research Triangle Park, North Carolina 27711 

g CLIMATE MONITORING AND DATA LABORATORY, National Oceanic and Atmospheric 

Administration,Barrow, Alaska 

* 

Corresponding author. Tel.: +45 65 50 25 57; fax: +45- 

65 50 12 14. 

E-mail address: meg@chem.sdu.dk (Mike Goodsite). 

1

Abstract 

2 

The conditional sampling or relaxed eddy accumulation, REA, technique represents the first 

opportunity to directly measure fluxes of the divalent, gaseous mercury compounds HgXY to the 

snowpack in the Arctic. Using a micrometeorological relaxed eddy accumulation system, with a 

heated sampling system specifically designed for Arctic use, the dry deposition of reactive gaseous 

mercury, RGM, is quantified for the first time after polar sunrise, in Barrow, Alaska. Heated KCl 

coated manual RGM annular denuders were used as the accumulators with an impactor, elutriator 

inlet allowing only fine (cut-off = 2.5 µm) particles to pass. At 3 m above the snow pack significant 

RGM fluxes measured during March 29 th – April 12 th 2000 were directed toward the snow surface. 

Overall mean flux was found to be - 0.4 ± 0.2 pg m -2 s -1 ; N=9, ± 1 SE, where the negative sign 

convention denotes deposition. Using measured total RGM concentrations; depositional velocities 

were then computed and averaged 1 cm s -1 . 

Keywords: Arctic; Conditional sampling; Divalent Gaseous Mercury; Emission; Snow–air 

exchange.


Mercury in the atmosphere, is approximately 95% in the gaseous elemental form (Slemr et al., 

1985, Schroeder and Munthe, 1998). Its characteristics, such as low aqueous solubility, mean that it 

is relatively non reactive and stable and therefore has a long atmospheric residence time, enabling 

global transport. 

In 1998 Schroeder et al., reported on their 1995 discovery of the springtime depletion of 

tropospheric gaseous mercury in the high Canadian Arctic. This perennial phenomenon is since 

dubbed atmospheric mercury depletion episodes, AMDEs and has since been shown to be a polar 

and sub-polar phenomenon (Schroeder et al., 2003 and references therein). The mercury depleted 

from the atmosphere is oxidized to divalent gaseous mercury and exists during AMDEs in 

concentrations up to 900 pg m -3 (Lindberg et al., 2002, Skov et al. 2003). The divalent gaseous 

mercury species are operationally defined as reactive gaseous mercury, RGM, since the analytical 

methods, in this case, thermally desorbed KCL coated annular denuders (Landis et al., 2002) only 

allow quantification after reduction to gaseous elemental mercury. Therefore, any speciation 

information is lost. 

It is necessary to quantify the flux of RGM in Arctic areas and determine the depositional 

velocity in order to better understand the temporal, and spatial patterns of mercury deposition and 

accumulation in the Arctic, and gain an understanding of the chemical and dry depositional 

processes so that they might be applied to parameterization of atmospheric transport and deposition 

models and eventually policy decisions. 

The conditional sampling, or relaxed eddy accumulation technique (Businger and Oncley, 

1989) has been used to determine the flux of elemental gaseous mercury (Cobos et al., 2002). Their 

work thoroughly discusses the advantages of using a micrometeorological technique over other flux 

measurement methods. In this paper, RGM flux is determined with a relaxed eddy accumulation 

3

control system (Metsupport, Denmark) coupled with a RGM manual sampling system utilizing 

heated KCl coated annular denuders to capture RGM (Landis et al., 2002). This system uses heating 

mantles specifically built for Arctic use. 

2. Methods 

2.1 Study Site 

4 

Flux measurements were made in March – April, 2001 at the NOAA Climate Monitoring and 

Diagnostic Laboratory (CMDL) in Barrow, AK. Barrow is geographically the northern-most point 

in Alaska, located at 71°19’ N, 156°37’ W, and the CMDL is approximately 9 m above mean sea 

level, providing a very flat fetch for micrometeorological measurements. The air sampled at the 

station is characteristically part of the marine boundary layer during the sunlit hours. The Beaufort 

Sea was frozen, with periodic leads opening during the campaign. The weather was very stable, as 

expected for a high Arctic coastal area. A more detailed description of the study site is provided in 

Lindberg et al., 2002. 

2.2 RGM determination 

RGM was measured and analysed by the method developed by Landis et al. (2002), employing 

a KCl coated quartz annular denuder sampling chain heated to 50 0 C, to sample air. At the inlet, 

there is an elutriator and an impactor, with impactor plate. The elutriator accelerates the flow, by 

forcing it through an orifice onto a roughened, non-coated impactor plate. The cut-off diameter is 

2.5 µm, so only the fine fraction of particles flows through the denuder. The flow rate is 10 litres 

per minute and was controlled prior to and after sampling sampling with a “dry cal” flow meter 

prior attached just before the denuder sampling train. 

Denuders had a collocated precision of < 15.0 ± 9.3 % with 3x std. dev., in agreement with the 

findings reported in Landis et. al., 2002.

For all flux measurements, heating was kept constant, since the heat will affect the laminar flow 

within the denuder, the gas diffusion coefficient and the relative adherence of the gas to the KCl 

surface. The temperature was checked prior to and after measurement. The heating mantles 

employed are high temperature polyvinylchlorid, PVC, pipes, which allow a 2 cm airspace around 

the outside of the annular denuder, and enclose the denuder from the tip of the inlet to the top of a 

filter pack at the outlet. The outer portion of the pipe is wrapped with a silver tape to ensure heat 

transfer from self-regulating heat tape. The length of the heating tape is manufacturer and electrical 

voltage dependent, but is cut long enough to heat the air inside the tube to 50 o C and keep it at that 

temperature. From campaign experience, we found it necessary to improve the heating mantles to 

achieve 50 0 C. The original heating mantle assembly was placed inside a larger PVC pipe, allowing 

5 cm spacing between sidewalls, giving an overall diameter of approximately 9.6 cm. The space 

between the two shells was then filled with self-expanding polyurethane foam thermal insulation. 

After sampling, the denuders were taken into the laboratory for thermal desorption in 

accordance with Landis et al. 2002, using a TEKRAN 2537A mercury analyser, to quantify the 

RGM as Hg 0 . 

For all measurements a field blank was obtained by handling a denuder in the field. Hg mass 

from this field blank was subtracted from the measured Hg masses on the exposed denuders. If 

there was any indication of Hg(0) adsorption, then the denuder was cleaned and re-coated, since as 

pointed out by Sheu and Mason, 2001, just 1% of Hg(0) adsorption on a denuder is enough to 

compromise RGM measurements. All accuracies with the denuders were found to be in good 

agreement with those reported by Landis et al., 2002. In the Barrow 2001 campaign, the manual 

denuders exhibited a precision of 10%, based on co-located parallel measurements, and were on 

average within 25% of the automated RGM sampling system running separately and independently, 

on the roof of Barrow CMDL. 

5

2.3 The RGM REA system and flux measurement 

6 

Relaxed eddy accumulation, REA, is a micrometeorological method for trace gas flux 

determination. REA “relaxes” the requirement for instantaneous gas analysis by preferentially 

collecting air over time into some type of accumulator for up and down drafts, and all other air, the 

mid-channel. The trace gas in the collected sample is analysed after the sampling period. 

RGM flux measurements were performed from 29 March, 2001 through April 12, 2001, with 

the system set up at approximately 3 m above the snow pack surface on a guy-wire support of the 

NOAA, Barrow, CMDL tower, oriented into the prevailing winds, arriving from the Beaufort Sea. 

The measurements were made using a micrometeorological flux measurement system built by 

METSUPPORT aps, Denmark in January 2001 for this campaign. This system was coupled with 

the Landis et al., 2002 annular denuder method for measuring RGM, and the previously described 

heating caps, as the sampling front end. A similar METSUPPORT system and components have 

been previously deployed by NERI for measuring volatile organic compounds (VOCs) and is 

described in detail in Christensen et al. 2000. 

The REA flux measurement system was set up on a steel rod affixed to the CMDL tower guy- 

line support pole with the head of the sonic anemometer facing into the predominant wind direction, 

and oriented towards North. The support beam was hung such that the heating caps and therefore 

inlets of the sampling system were perpendicular to, and 1 meter behind the centrum of the sonic 

head. The inlets were 1-2 mm longer than flush protruding from the bottom of the heating caps. The 

denuders for the up and down draft were co-located nearest the centre of the mast, while the parallel 

measurements or denuders sampling the air that is not coming either as down or up, were located 

near the edges of the mast. A quartz filter was kept in each of the filter packs, to ensure a constant 

pressure drop. Temperature in the heating caps was measured prior to and after sampling. From the 

quick connect at the top of the filter pack were connected 3.2 m long neoprene hoses into 3 fast

esponse switching valves supplied by MetSupport. From behind the switches, the three valves were 

connected into one sampling line using a simple 3 inlet manifold constructed with 2 T-type locking 

copper hose connectors in series and 1 L type locking hose connector as the end piece. Coming out 

of the manifold, a locking ball valve was used to adjust and fix the flow, as a back up to the mass 

flow controller. Between the pump and the valve, a Tylan mass flow controller was used to ensure a 

flow prior to the manifold of just over 10-litre min -1 , so that the flow was measured as 10-litre min -1 

at the denuder inlet. Pressure loss was minimal in the manifold and through the sampling lines. 

Once the system was running, the lag time from when the switch opened to when the flow started 

out the denuder was very small compared to the air sampling switching frequency of 1 Hz. 

Normally flux systems operate at air sampling switching rates of 10 Hz, switching as fast as the 

air flow is sampled with the sonic anemometer. The only lag time between the air and the switching 

comes from the software and physical switching process, including flow development in systems. 

Denuders are selective accumulators, that need a laminar flow in order to work properly. Given the 

geometry of the URG annular denuders, the 10 Hz sampling switching rate nominally allowed for 

full laminar flow development and escape of an air packet through the end of the annular denuder. 

Therefore the sampling switching rate was set in the REA system to 1 Hz with the 10 litre per 

minute flow rate maintained by a mass flow controller. The air sampling-switching rate is so long 

compared to errors that could be induced from physical switching and software, that these are 

considered negligible. By sampling at 1 Hz, we expect that approximately 95% of the turbulence is 

captured and the best compromise between the meteorological measurements and chemical 

sampling is obtained in order to ensure a laminar flow in the annular denuders and thereby measure 

RGM flux most accurately. 

Once RGM concentrations were obtained for the sampling period for each denuder channel, the 

surface flux F of RGM was calculated from equation 1. 

7

8 

F = βσv(C1-C3) (1), 

where β is an empirical coefficient, the “proportionality constant”, dependent on wind speed and 

turbulence, generally 0.6 for a fixed deadband and approximately 0.3 for a dynamic deadband, and 

calculated via the heat flux with the Metsupport system. σv is the standard deviation of the vertical 

wind velocity: both values are obtained directly as output from the REA system; C1 and C3 are the 

concentrations of RGM in upward and downward air masses, respectively. 

From the flux measurements the depositional velocity of RGM can be calculated if the ambient 

concentration for RGM is known: 

vd = F/C (2), 

where C is the concentration of RGM, and F is from 1. 

Due to the uncertainty of the concentration measurements in Barrow: 10 % and of the 

meteorological measurements, 10 % for β, σw for the Metsupport system (Christensen et al., 2000) 

the uncertainty on the flux measurements using the Landis et al. 2002, KCl coated annular denuder 

sampling end, heating caps and Metsupport REA system are estimated to be within 40% on a 95% 

confidence interval, though 20% would be a more conservative estimate of denuder precision, given 

the fact that flow at 10 lpm is not instantaneously developed, so there is necessarily some flow and 

turbulence information lost when switching at 1 Hz. This gives a conservative sampling error for 

flux as 50 % for the above system in Barrow. 

As a quality assurance check for the REA flux measurements, the total concentration of the 

three annular denuders for each run was compared with ambient concentration from a CMDL 

collocated automated RGM monitoring system described in Lindberg et al., 2002. Results were 

favourable, within 25% of each other on average. 

3. Results and Discussion

3.1 RGM flux measurements 

The results of the 2001 campaign in Alaska: Barrow Arctic Mercury Study (BAMS -2001) are 

shown in Table 1 and Fig. 1. The largest deposition velocity in 2001 was 2.7 cm s -1 and the average 

was close to 1 cm s -1 . Mean dry depositional flux was found to be - 0.4 ± 0.2 pg m -2 s -1 . 

Concentrations of the three denuder tubes, up, mid and down, were added together to determine the 

total concentration, and concentration was compared with the on site Tekran automatic RGM 

monitor, MODEL 1130, described in Lindberg et al., 2002 (Table 1). 

The machines were not running absolutely simultaneously so a linear interpretation was made 

to compare concentrations. Not taking run two into account, since a valve was frozen open, it is 

seen that the percent difference varies, with an average for the campaign of 24% with a standard 

deviation of 42%. The percent difference between the two measurements was between 3% and 

78%. This variance can be explained as a combination of the comparison method, the systems were 

not started or stopped simultaneously, and the oxidation of gaseous elemental mercury to RGM is 

dynamic, thus a linear extrapolation of concentration during a sampling period, is conservative. It is 

seen that in all but 2 of the runs, run 2 excluded, the RGM REA total is less than the TEKRAN 

1130 value, perhaps because of differences in the heating mantle system, but this observed bias 

could also be because of different flow calibration systems. 

The REA system when properly on the tower, after a period of time should report the same 

number of counts per up and down channel, due to conservation of mass, and did so when initially 

inventoried and tested during pre-deployment trials in Oak Ridge. In Barrow, there was always a 

greater number, up to 30% more, of counts on the down channel, suggesting that the sonic 

anemometer was not positioned properly, or a forcing of the turbulence. Therefore the channel 

count data was corrected to ensure mass conservation as follows: 

1) Corrected counts in Down = registered counts in Down - (registered counts in Down – 

corrected Down); where corrected counts in down = counts in up channel. 

9

Control: total counts for all channels registered = total counts with down corrected. 

2) Corrected Counts for Mid = registered Counts for Mid - (Registered Counts for Down - 

corrected counts for Down); The lost mass must be placed in mid for later concentration 

calculation. 

Control: Corrected counts Mid > registered counts mid, but total counts all channels still the same. 

3) Corrected Counts Up = Old Counts up Up 

4) New Total Volume for each denuder = New count * 10 l per second (fixed) 

Control: registered total Volume = Corrected total volume 

Control: registered total mass, RGM in ng Hg (0) = new total mass RGM, ng Hg (0). 

6) New concentration: New mass / New volume 

10 

Run dates and time, were reported as Greenwich Mean Time in accordance with the other 

monitors at CMDL Barrow. The sonic anemometer consistently showed an ambient temperature of 

2 degrees higher than other ambient temperature instruments at CMDL. This should not affect the 

measurements since the proportionality constant, β, is based on the heat flux. 

Table 2 shows other average data for the run, as recorded by the sonic anemometer. The clean 

air sector for NOAA, CMDL, Barrow, Alaska is defined as wind arriving from the coast, as 

opposed from the town of Barrow, and is 45 0 to 135 0 . The system was oriented to 360 0 . 

Due to the nature of micrometeorological measurements, the REA system works best at wind 

speeds near 5 m s -1 since there needs to be good turbulence to sample. Wind speeds less than 2 m s -1 

are nominal. During the campaign, the wind generally came from outside of the CMDL defined 

clean air sector. This means that during the measurements, the wind was uncharacteristically 

coming from the town of Barrow. The average standard deviation in the vertical wind component is 

0.29 with a standard deviation of 0.11. The proportionality constant averaged 0.41 with a standard 

deviation of 0.01. Indicating a stable heat flux. The REA system reported all numbers to four 

decimals; results have been rounded to two. Fig. 1, show that the average depositional velocity for 

reactive gaseous mercury is approximately 1 cm s -1 for the mass corrected data and approximately 

0.5 cm s -1 for the non-mass corrected data. Average depositional flux for Barrow was 1.3 ± 0.7 ng 

m -2 h -1 for the mass corrected and approximately half that for non-mass corrected values. For 

comparison purposes: Schroeder et al., 1998 used a dry depositional velocity of 0.5 cm s -1 when

estimating for Alert an average springtime dry-deposition flux for mercury of 2.5 ± 0.5 ng m -2 h -1 

based on their measurements of TGM. Comparing the flux with the average meteorological 

conditions in Table 2., gives no direct correlations. One would expect that emissions would increase 

as a function of the rising temperatures, but this can not said to be readily apparent. What seems to 

be apparent is that following a large depositional event; there is a small re-emission. This is 

probably the snow pack regaining equilibrium with the atmosphere. 

As seen in Fig. 1., the depositional velocities noted for depositional events are fast, around 2 

cm s -1 this is what would be expected for a very reactive gaseous species such as HNO3. On 

average, the depositional velocity is approximately 1 cm s -1 . Comparing with measured dry 

deposition velocities over snow for HNO3, reviewed in Karlson and Nyholm, 1998, show that the 

measured dry depositional velocities for RGM may be an order of magnitude higher than that for 

HNO3, given the snow surface temperature of < 2 0 C. 

3.2 Comparison of RGM flux with RGM ambient concentrations 

Fig. 2, are the monitored results RGM concentrations during the Barrow 2001flux measurement 

campaign. Comparing with Fig. 1., show that when there is a deposition recorded by the REA 

machine, there are correspondingly low RGM ambient values. The air has been apparently depleted 

of RGM due to deposition. Trends of RGM rising in the air on the 8 th , 9 th and 12 th of April are 

recorded by the REA system as reemission, perhaps indicating that RGM can be re-volatized from 

the snow surface. 

3.3 Drawbacks and other possible accumulators 

Stable weather conditions are the greatest drawback to deploying the RGM system in the Arctic 

for investigating mercury depletion events. Lu et al., 2001, summarize that the environmental 

conditions favouring mercury depletion events at high latitudes are: 1, marine/maritime location; 2, 

calm weather, low wind speeds, non-turbulent air flow; 3, the existence of a temperature inversion; 

11

4, sunlight and 5, sub-zero temperatures. Condition 2 implies poor operational characteristics for 

micrometeorological systems. We needed to compromise the sampling frequency to achieve 

laminar flow. We examined other RGM accumulators: three methods aside from annular denuders 

have been developed for measurement of RGM: refluxing mist chambers, ion-exchange membranes 

behind particulate filters and potassium chloride, KCl, coated tubular denuders (Landis et al., 2002 

and citations therein). However, only the annular denuder method by Landis et al., had the 

necessary characteristics of being able to operate under arctic conditions, with the flow dynamics 

essential for the REA flux measurements in this work. 

4. Conclusions 

12 

This work reports the first measurement series of RGM flux in the Arctic providing data for 

transport and depositional model parameterization. It shows that RGM is quickly deposited to the 

snow surface, however, there were also emissions, implying that there are processes in the snow 

surface capable of releasing RGM, or re-emitting the deposited RGM. The flux system takes 

advantage of already developed methods, to accumulate the RGM (Landis et al., 2002) and a 

micrometeorological method, REA, that has successfully been employed to measure gaseous 

elemental mercury flux (Cobos et al., 2002). 

Acknowledgements 

This work was carried out under basic NERI funding, as well as funding from the Danish 

Environmental Protection Agency, DANCEA program. MEG gratefully acknowledges the 

Copenhagen Global Change Initiative graduate research fellowship from the Danish Research 

Agency and NERI. He would like to thank NOAA, ATDD and ORNL, ESD for allowing him to be 

a visiting student and supporting his research, and to NOAA CMDL (Dan Endres), for support 

during the campaign.

References 

1. Businger J.A. and Oncley, S.P., (1990): Flux measurement with conditional sampling. Journal of 

Atmospheric and Oceanic Technology 7, pp. 349–352. 

2. Christensen, C.S.; Hummelshoj, P.; Jensen, N. O.; Larsen, B.; Lohse, C.; Pilegaard, K.; Skov, 

H., (2000): Determination of the terpene flux from orange species and Norway spruce by 

relaxed eddy accumulation. Atmospheric Environment, 34(19), 3057-3067. 

3. Cobos, Douglas R.; Baker, John M.; Nater, Edward A., (2002): Conditional sampling for 

measuring mercury vapor fluxes. Atmospheric Environment, 36(27), 4309-4321. 

4. Karlsson, E., and Nyholm, S., (1998): Dry deposition and desorption of toxic gases to and from 

snow surfaces, Journal of Hazardous Materials, Volume 60, Issue 3, Pages 227-245. 

5. Landis, M.S., Stevens, R.K., Schaedlich, F. and Prestbo, E.M., (2002): Development and 

characterization of an annular denuder methodology for the measurement of divalent 

inorganic reactive gaseous mercury in ambient air. Environmental Science and Technology 

36, pp. 3000–3009. 

6. Lindberg, S. E., Brooks, S., Lin, C. J., Scott, K. J., Landis, M. S., Stevens, R. K., Goodsite, M. 

and Richter, A. (2002): Dynamic oxidation of gaseous mercury in the Arctic troposphere at 

polar sunrise. Environmental Science & Technology 36, 1245-1256. 

7. Lu, J. Y., Schroeder, W. H., Barrie, L. A., Steffen, A., Welch, H. E., Martin, K., Lockhart, L., 

Hunt, R. V., Boila, G. and Richter, A., (2001): Magnification of atmospheric mercury 

deposition to polar regions in springtime: the link to tropospheric ozone depletion chemistry. 

Geophysical Research Letters 28, 3219-3222. 

8. Oncley, Steven P.; Delany, Anthony C.; Horst, Thomas W.; Tans, Pieter P. Verification of flux 

measurement using relaxed eddy accumulation. Atmospheric Environment, Part A: General 

Topics (1993), 27A(15), 2417-26. 

9. Schroeder, W. H. and Munthe, J. (1998): Atmospheric Mercury - An Overview. Atmospheric 

Environment 32, 809-822. 

10. Schroeder, W. H., Anlauf, K. G., Barrie, L.A., Lu, J.Y. and Steffen, A. (1998): Arctic 

springtime depletion of mercury. Nature 394, 331-332. 

11. Schroeder, W. H., Steffen, A., Scott, K., Bender, T., Prestbo, E., Ebinghaus, R., Lu J. Y., and 

Lindberg, S. E., (2003): Summary report: first international Arctic atmospheric mercury 

research workshop, Atmospheric Environment, Volume 37, Issue 18, Pages 2551-2555. 

12. Sheu, G.P., and Mason, R.P., (2001): An Examination of Methods for the Measurements of 

Reactive Gaseous Mercury in the Atmosphere. Environmental Science and Technology, 35 

(6), 1209 -1216. 

13. Skov, H., Christensen, J., Goodsite, M.E., Heidam, N.Z., Jensen, B., Wåhlin, P., Geernaert, G. 

The Fate of Elemental Mercury in Arctic during Atmospheric Mercury Depletion Episodes 

and the Load of Atmospheric Mercury to Arctic. Submitted to Environmental Science and 

Technology, June 2003. 

14. Slemr, F., Schuster, G. and Seiler, W., (1985): Distribution, speciation and budget of 

atmospheric mercury. Journal of Atmospheric Chemistry 3, pp. 407–434. 

13

Figure captions 

Figure 1. Surface flux of RGM to the snow pack on a 3 m tower, at Barrow, Alaska, spring, 2001 

using REA. Sampling time approximately 4 hrs at midday. KCl coated annular denuder tubes, flow 

rate = 10 lpm (Landis et al., 2002) used as accumulators. 

Figure 2. RGM ambient concentrations monitored at CMDL during flux campaign period. 

14

Table captions 

Table 1. RGM mass raw data and accumulated concentrations corrected for blank values, 1130 are 

automatic RGM concentration measurements from the NOAA TEKRAN speciation unit, for 

comparison to REA total as % difference, run 2 not included in compared avg. and std. deviation. 

Table 2. Average temperature, wind speed, wind direction, standard deviation of the vertical wind 

velocity and the proportionality constant based on heat flux as reported by the sonic anemometer 

result file runname.txt. Temperature was 2 0 higher than what other on site instruments reported. 

Table 3. Results of REA RGM vertical flux measurements, Barrow, Alaska, 2001; with computed 

dry depositional velocities for RGM. Runs 1 and 7 are excluded due to outlying depositional 

velocities, indicating a problem with micrometeorological measurements. Run 2 is excluded since 

the mid-channel froze open, and the mass balance could therefore not be corrected. 11 excluded 

from average, since it is a nighttime run. 

15

Figures and Tables 

Figure 1. 

16 

RGM vertical flux (pg m-2 s-1) and 

Depositional Velocity of RGM cm/s 

1,5 

1,0 

0,5 

0,0 

-0,5 

-1,0 

-1,5 

-2,0 

-2,5 

-3,0 

-3,5 

02-apr 

04-apr 

RGM flux: BAMS 2001 

05-apr 

07-apr 

08-apr 

09-apr 

Date, average; 

error bars reflect 1 standard error 

10-apr 

12-apr 

Vertical Flux (Fc) (pg m^-2 s^-1) 

Depositional Velocity (Vd) cm s^-1 

Vertical flux without mass correction 

Depositional velocity without mass correction 

Average

Figure 2. 

RGM pg m-3 

350 

300 

250 

200 

150 

100 

50 

0 

03-26 

00:00 

03-28 

00:00 

03-30 

00:00 

Automatic RGM measurements, Barrow 2001 

04-01 

00:00 

04-03 

00:00 

04-05 

00:00 

Date and time 

04-07 

00:00 

04-09 

00:00 

04-11 

00:00 

04-13 

00:00 

04-15 

00:00 

17

Table 1. 

18 

Date time group RGM mass (pg) Concentration Blank. Corrected (pg/m3) 

Run Dt/time(Z) Down mid up down mid up Total 1130 % difference 

1 29 2200-30 0100 MAR 29,3 15,3 12,3 67,7 19,0 33,7 121 126 4 

2 30 1800-30 2200 MAR 28,2 353,9 24,7 44,6 310,6 46,7 402 58 (-593) 

3 02 2000-03 0000 APR 31,0 23,7 21,0 39,9 23,2 42,2 105 93 -13 

4 04 2015-05 0015 APR 11,8 22,9 9,4 17,4 22,5 15,6 56 65 14 

5 05 1839-06 0039 APR 31,0 56,4 27,2 33,8 32,7 31,7 98 448 78 

6 07 1850-08 0050 APR 61,4 92,6 38,6 57,5 58,0 46,2 162 99 -64 

7 08 2030-09 0030 APR 25,5 24,6 13,7 33,5 23,8 27,0 84 172 51 

8 09 1804-10 0104 APR 44,8 68,1 35,9 35,9 36,2 36,9 109 335 67 

9 10 1839-10 2339 APR 29,9 41,7 22,7 33,5 30,3 36,1 100 166 40 

10 11 2000-12 0000 APR 27,1 34,8 16,4 39,8 32,8 29,4 102 121 16 

12 12 2100-13 0000 APR 26,7 41,8 21,4 31,2 30,0 32,8 94 97 3 

AVERAGE 31,5 70,5 22,1 39,5 56,3 34,4 24 

Std. Dev. 12,47 96,72 9,26 13,49 84,99 8,97 42 

11 (night) 12 0800-12 1100 APR 4,80 6,70 2,2 6,4 6,4 4,5 17 48 65

Table 2 

DAYHOURMON avg.temp avg.WS avg.Wdir In σv β 

Run Dt/time(Z) deg. C m/s deg Sector? average Average 

1 29 2200-30 0100 MAR -22,2 2,6 160 N 0,27 0,41 

2 30 1800-30 2200 MAR -22,6 5,5 132 Y 0,19 0,41 

3 02 2000-03 0000 APR -21,4 2,5 236 N 0,25 0,41 

4 04 2015-05 0015 APR -14,8 7,9 15 N 0,27 0,42 

5 05 1839-06 0039 APR -17,2 7,8 65 Y 0,26 0,41 

6 07 1850-08 0050 APR -9,3 5,9 93 Y 0,24 0,42 

7 08 2030-09 0030 APR -6,3 4,9 208 N 0,59 0,39 

8 09 1804-10 0104 APR -12,3 4,9 210 N 0,40 0,42 

9 10 1839-10 2339 APR -9,6 6,8 51 Y 0,27 0,41 

10 11 2000-12 0000 APR -8,8 5,2 41 N 0,18 0,42 

12 12 2100-13 0000 APR -8,5 3,3 247 N 0,25 0,41 

11 

0,11 0,42 

(night) 12 0800-12 1100 APR -8,7 2,7 36 N 

19

Table 3. 

Vertical Flux (Fc) Vd 

Run Nr. Date time GMT (pg m^-2 s^-1) ng/m2/hr m s^-1 cm s^-1 

20 

1 29 2200-30 0100 MAR -5,6 -20 -0,2 -15,7 

2 30 1800-30 2200 MAR 4,57 16,44 0,03 2,58 

3 02 2000-03 0000 APR -0,8 -2,8 0,0 -2,4 

4 04 2015-05 0015 APR -0,2 -0,6 0,0 -0,8 

5 05 1839-06 0039 APR -0,3 -0,9 0,0 -0,8 

6 07 1850-08 0050 APR -1,2 -4,3 0,0 -2,2 

7 08 2030-09 0030 APR -2,79 -10,05 -0,1 -10,06 

8 09 1804-10 0104 APR 0,2 0,7 0,0 0,6 

9 10 1839-10 2339 APR 0,1 0,4 0,0 0,4 

10 11 2000-12 0000 APR -1,0 -3,5 0,0 -2,8 

12 12 2100-13 0000 APR 0,2 0,6 0,0 0,6 

11 (night) 12 0800-12 1100 APR -0,08 -0,3 -0,01 -1,41 

Average Excl. 1,2,7,11 -0,4 -1,3 0,0 -0,9 

Std. Dev. Excl. 1,2,7,11 0,6 2,0 0,0 1,4


Paper 2: Lindberg, Steve E.; Brooks, Steve; Lin, C.-J.; Scott, Karen J.; Landis, Matthew S.; 

Stevens, Robert K.; Goodsite, Mike; Richter, Andreas. Dynamic Oxidation of Gaseous 

Mercury in the Arctic Troposphere at Polar Sunrise. Environmental Science and Technology 

(2002), 36(6), 1245-1256.

Dynamic Oxidation of Gaseous 

Mercury in the Arctic Troposphere 

at Polar Sunrise 

STEVE E. LINDBERG* 

Environmental Sciences Division, Oak Ridge National 

Laboratory, Oak Ridge, Tennessee 37831-6038 

STEVE BROOKS 

Oak Ridge Associated Universities, P.O. Box 117, 

Oak Ridge, Tennessee 37831-0117 

C.-J. LIN 

Department of Civil Engineering, P. O. Box 10024, 

Lamar University, Beaumont, Texas 77710 

KAREN J. SCOTT 

Department of Microbiology, University of Manitoba, 

Winnipeg, Manitoba R3T 2N2, Canada 

MATTHEW S. LANDIS 

U.S. EPA, 79 TW Alexander Drive, 

Human Exposure Analysis Branch, MD-56, 

Research Triangle Park, North Carolina 27711 

ROBERT K. STEVENS 

Florida Department of Environmental Protection, 

2600 Blair Stone Road, Tallahassee, Florida 32399 

MIKE GOODSITE 

National Environmental Research Institute of Denmark, 

Copenhagen, Denmark 

ANDREAS RICHTER 

Institute of Environmental Physics, University of Bremen, 

D-28359 Bremen, Germany 

Gaseous elemental mercury (Hg 0 ) is a globally distributed 

air toxin with a long atmospheric residence time. Any 

process that reduces its atmospheric lifetime increases 

its potential accumulation in the biosphere. Our data from 

Barrow, AK, at 71° N show that rapid, photochemically 

driven oxidation of boundary-layer Hg 0 after polar sunrise, 

probably by reactive halogens, creates a rapidly depositing 

species of oxidized gaseous mercury in the remote Arctic 

troposphere at concentrations in excess of 900 pg m -3 . 

This mercury accumulates in the snowpack during polar 

spring at an accelerated rate in a form that is bioavailable 

to bacteria and is released with snowmelt during the 

summer emergence of the Arctic ecosystem. Evidence 

suggests that this is a recent phenomenon that may be 

occurring throughout the earth’s polar regions. 

Introduction 

Mercury has been targeted for global concern as a highly 

toxic contaminant. Exposures are thought to be increasing, 

* Corresponding author phone: (865)574-7857; fax: (865)576-8646; 

e-mail: Lindbergse@ornl.gov. 

Environ. Sci. Technol. 2002, 36, 1245-1256 

especially among indigenous populations who consume fish 

and piscivorus species contaminated with methylmercury 

(1, 2), a neurotoxin that biomagnifies in aquatic food chains. 

Environmental mercury levels are known to be elevated in 

the Arctic, to generally increase with latitude, and to have 

increased over time (3, 4). Extensive wetlands exist in the 

Arctic (5), and such areas are now recognized as important 

sources of methylmercury (6). 

In addition to Hg 0 , somewhat lesser amounts of oxidized 

reactive gaseous mercury (RGM) species are now known to 

be emitted from industrial sources (7). RGM species are watersoluble, 

exhibit a much shorter atmospheric lifetime than 

Hg 0 , and their potential contribution to atmospheric deposition 

is widely recognized (8, 9). Although RGM compounds 

represent only a few percent of the overall gaseous mercury 

in typical ambient air, their dry deposition velocities and 

scavenging ratios exceed those of Hg 0 by more than an order 

of magnitude (8). Since industrial emissions of water-soluble 

RGM compounds are controllable to some extent (7), the 

demonstration of direct production of RGM in the atmosphere 

has profound implications on ecosystem and human 

exposure. 

The concept of global fugacity has been used to explain 

the condensation and accumulation of organic toxins in Arctic 

regions (10), but Hg 0 does not effectively “condense out” 

even at -50 °C. However, other forms of airborne Hg, 

especially oxidized Hg, might strongly partition from gas to 

solid phases at low temperatures (11). Recent reports of 

ground-level ozone (O3) and Hg 0 depletions at Alert in the 

Canadian High Arctic (∼82° N) provided the first evidence 

that conditions may exist in the upper Arctic following polar 

sunrise that promote depletion of airborne Hg 0 (12). Unlike 

O3, which is chemically destroyed, Hg only changes its 

oxidation state, and depletion from the airmass implies 

accumulation elsewhere. While the original Alert study did 

not include measurements of the fate of the depleted Hg, a 

recent paper reports elevated levels of Hg in snow throughout 

the Canadian Arctic (13). The characterization of the associated 

Hg species and their ultimate fate in the Arctic 

ecosystem is critical to our understanding of mercury 

depletion events (MDEs). 

In 1998, we initiated Hg 0 measurements as part of the 

Barrow Arctic Mercury Study (BAMS; 14) to determine the 

geographic extent and reaction mechanism of MDEs. In 1999, 

we added the first semi-continuous measurements of RGM 

in the Arctic and measured both species during and after 

polar sunrise through June 2001. In an intensive campaign 

during April-June 2001, we also collected particulate Hg (Hgp) 

samples at Barrow, aircraft measurements of RGM, and 

aerodynamic measurements of RGM fluxes over snow. We 

report here the first detailed analysis of the BAMS data and 

demonstrate that Hg 0 is being “depleted” as a result of in-air 

oxidation reactions that produce some form of oxidized 

gaseous mercury that is rapidly deposited to the snow surface 

at Barrow. 

Methods 

Study Site. All of the data reported here were collected at the 

NOAA Climate Monitoring and Diagnostic Laboratory (CMDL) 

in Barrow, AK. There are no Hg sources within the CMDL, 

no emission points on the roof where our intakes were 

located, nor any known major Hg point sources in the town 

of Barrow, which is located ∼10 km southwest of the CMDL. 

Prevailing winds are from the northeast across the Beaufort 

Sea. The CMDL is located near the peninsula at Point Barrow, 

∼2 km from the shoreline, and is surrounded primarily by 

10.1021/es0111941 CCC: $22.00 © 2002 American Chemical Society VOL. 36, NO. 6, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 9 1245 

Published on Web 02/13/2002

FIGURE 1. Schematic diagram of the sample flow path in the automated Tekran model 2537A Hg 0 analyzer with the 1130 gaseous speciation 

denuder module and 1135 particulate Hg pyrolysis unit. 

water to the north, east, and west. Barrow is geographically 

the northern-most point in Alaska, located at 71°19′ N, 156°37′ 

W, and the CMDL is ∼9 m above mean sea level. In latitude, 

Barrow is ∼1600 km south of Alert. We initiated atmospheric 

Hg sampling at Barrow in September 1998 and began 

additional parallel sampling of speciated Hg in air 1 yr later 

using the methods described below. These measurements 

will continue through at least 2003. We also collected fresh 

surface snow and snow cores for Hg analysis from the tundra 

in the CMDL clean air sector during 2000 and 2001. Particulate 

Hg, eddy flux, and aircraft Hg measurements were added 

during the intensive BAMS-2001 campaign in March-June 

2001 at the Barrow lab. Ancillary data available at the CMDL 

include routine meteorological data and trace gases such as 

ozone (15). 

Sampling and Analytical Approaches. Atmospheric Hg 0 

was determined using a Tekran 2537A vapor-phase mercury 

analyzer. Descriptions of the Tekran and its operating 

parameters have been published (16), and it has been the 

subject of a number of intercomparison studies (17-19). 

The 2537A instrument utilizes two parallel solid gold traps 

to preconcentrate Hg 0 that is subsequently thermally desorbed 

into a cold vapor atomic fluorescence spectrometer 

(20). The instrument was configured to sample at a 5-min 

time resolution using a heated Teflon inlet line mounted ∼5 

m above the ground on a mast ∼1 m above the roof of the 

NOAA CMDL building. The insulated sampling line was 

maintained at 50 °C by a PID temperature controller. 

Atmospheric Hg speciation was determined by integrating 

a Tekran 1130 speciation unit with the Tekran 2537A. The 

model 1130 speciation unit consists of a heated denuder 

module, a pump module, and a controller module. The model 

1130 controller module integrates the analytical capabilities 

of the Tekran model 2537A unit with the 1130 speciation 

module allowing for continuous measurement of both Hg 0 

1246 9 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 36, NO. 6, 2002 

and RGM at pg m-3 concentrations. The 1130 speciation unit 

was configured to collect 2-h RGM samples onto a KCl-coated 

quartz annular denuder at a 10 L/min flow rate. During the 

2-h sampling period, 5-min Hg0 samples were continuously 

quantified by the 2537A analyzer. After the 2-h sampling 

period, the 1130 system was flushed with Hg-free air, and 

the annular denuder was heated to 500 °C. The RGM collected 

on the denuder was thermally decomposed into a Hg-free 

airstream and subsequently analyzed as Hg0 . The denuders 

collect oxidized reactive gaseous mercury compounds with 

a diffusion coefficient >0.1 cm2 /s that readily adhere to a 

KCl coating at 50 °C (21). The most probable candidate 

compounds are HgCl2 and HgBr2; HgO is less likely. In this 

configuration, the fine particulate phase mercury (Hg-p; e2.5 

µm) that passes the impactor inlet was collected on a 

downstream quartz fiber filter but was not analyzed. 

For 3 months during the BAMS-2001 spring intensive 

campaign, our EPA colleagues deployed additional mercury 

monitoring equipment to allow separate quantification of 

Hg0 , RGM, and also Hg-p. The equipment included a Tekran 

2537A/1130 configuration described above with the addition 

of a prototype Tekran 1135 particulate pyrolysis unit. One of 

the greatest problems with conventional Hg-p measurement 

methods is that RGM has been shown to adsorb to filter 

media and previously collected aerosols. This can result in 

large and variable measurement artifacts (21). The Tekran 

1130/1135 speciation system avoids this problem by collecting 

RGM prior to collecting the Hg-p onto a quartz fiber 

filter. The quartz filter and quartz denuder components were 

sequentially desorbed during the analysis phase at 500 and 

650 °C, respectively. During the filter heating step, the 

pyrolyzer was maintained at 650 °C to ensure complete 

decomposition of all Hg-p compounds evolved during the 

filter temperature ramp-up to Hg0 . Figure 1 shows a schematic 

of the flow path for the overall 1130/1135 system as integrated

with the 2537A. During desorption, the sampling system was 

flooded with zero air to eliminate background air and achieve 

good analysis blanks. This zero air also acts as the carrier gas 

during subsequent analysis steps. The pyrolyzer for the quartz 

regenerable filter was preheated to convert to elemental form 

any mercury compounds that are eluted during subsequent 

steps. As the regenerable particulate trap was heated, the 

particle-bound mercury captured on the trap was desorbed 

and quantified by the Tekran 2537A. The heating process 

also reconditioned the trap for subsequent cycles. After 

desorption was complete, the entire sampling train was 

cooled to 50 °C. After being cooled, the denuder and 

particulate trap were ready for another measurement cycle. 

Atmospheric Hg fluxes were quantified during the BAMS- 

2001 intensive by deployment of a relaxed eddy accumulation 

(REA) system developed for RGM. The Danish REA system 

uses a METEK sonic anemometer coupled with three heated 

manual RGM denuders and filter packs (22). The basis of the 

REA logging and control system used is as described in the 

literature (23). The system was operated over the snow during 

April 2001 to collect 12 4-h samples at a 10 L/min flow rate 

and a switching frequency of 1 Hz to quantify the RGM 

concentrations in the upflow and downflow eddies and a 

deadband (we realize that there is decreased resolution of 

the turbulence at 1 Hz; however, we feel that the data 

represent a general trend and are in good agreement with 

fluxes calculated based on the constant sampling during the 

same time periods). The vertical flux was calculated as 

described in ref 24: F ) σW(C_up - C_down)�, where σW is 

SD (vertical wind speed), � is a proportionality constant (both 

measured in real time), and C_up - C_down is the direct 

difference in concentration between the upflow and downflow 

denuders for each period. The denuders were analyzed 

manually as described in refs 22 and 25, a method based on 

the automated approach described above. Development 

continues on the REA system to increase the sampling 

frequency and maintain a constant laminar flow in the 

denuders. 

Surface snow samples (upper 10 cm) were collected in 

acid-washed Teflon bottles and maintained frozen until 

analysis. Subsurface samples were similarly collected from 

the freshly exposed snowpack face. Snowmelt runoff was 

collected in prefired Pyrex bottles with Teflon caps. Total 

Hg and methylmercury analyses were performed at Flett 

Research Ltd. (Winnipeg, MB) and at ORNL using cold vapor 

atomic fluorescence spectrophotometry (26). Assays for 

bioavailable mercury (bioHg) were performed on a number 

of snow samples using mer-lux bioreporters, genetically 

engineered bacteria that produce light when divalent inorganic 

mercury enters their cells (27). Samples were melted 

in the dark and analyzed immediately. Sample preparation 

was done in a Class 100 laminar flow hood in a HEPA-filtered 

Hg clean lab at the Freshwater Institute (Winnipeg, MB). 

The general assay method employed was as described in ref 

27 with the following modifications: Vibrio anguillarum was 

the host species, not Escherichia coli; cells were grown in 

Glucose Minimal Medium; the assay medium was modified 

to include 5 mM glucose and 18 mM (NH4)2SO4; in addition, 

3 mM sodium/potassium phosphate buffer (pH 6), used to 

wash and resuspend the mer-lux cells, was added to the 

sample in the final cell suspension. The specific growth 

requirements and cell preparation of V. anguillarum are 

described in ref 28. The primary Hg standard used was a 

1 µgmL -1 Hg(NO3)2 solution prepared by Flett Research Ltd. 

(Winnipeg, MB). Light production was measured with a Beckman 

LS 6500 scintillation counter in noncoincidental mode. 

Maps of BrO distribution in the total column were 

generated from GOME satellite data (Global Ozone Monitoring 

Experiment). GOME is a 4-channel UV/visible grating 

spectrometer on board the ESA satellite ERS-2. GOME 

radiances and irradiances were used with a published 

algorithm (29) to derive total column BrO. GOME measurements 

do not yield profile information but reflect the BrO 

distribution throughout the column. Since stratospheric BrO 

(which mainly depends on photolysis of the reservoirs and 

total stratospheric BrY) is relatively constant with time and 

shows little spatial variation, as supported by model calculations 

and ground-based observations of stratospheric BrO 

(29), the BrO plots show mainly the boundary-layer variations 

plus a more or less constant offset. This idea is supported 

by enhanced boundary-layer BrO observed independently 

at other polar sites using ground-based instruments that 

match with enhanced BrO values in the GOME data (30). 

Quality Assurance Activities. At Barrow, the Tekran 2537A 

routinely undergoes automated periodic recalibrations daily 

using an internal permeation source. Two-point calibrations 

(zero and span) are performed separately for each cartridge. 

A Commercial permeation tube (VICI Metronics) provides 

approximately 1 pg/s at +50.0 °C. Since it is not practical to 

certify these low rates gravimetrically, manual injections were 

used in our home lab before shipment to Barrow to initially 

calibrate the tube against a saturated mercury vapor standard 

(16, 31). Since then, this procedure has been repeated once 

or twice annually. The adjustment for perm tube drift has 

been on the order of 1-2% per year. During routine 

operations at Barrow, the Tekran was also subject to periodic 

zero checks and spikes of ambient air with a known amount 

of Hg 0 . During 1999, we performed spikes and zero checks 

every 25 h using an automated system that externally 

controlled the perm source to deliver 16 pg of Hg 0 into the 

Teflon sampling line. Over the year, spike recoveries ranged 

from 85 to 114% and averaged 102 ( 3% recovery. There was 

no significant difference in recovery during the MDE period 

as compared to the rest of the year. Zero checks were 

consistently at the detection limit (0.09 ( 0.01 ng/m 3 ). 

Although the 1130 RGM speciation system does not yet 

include a direct calibration method with known amounts of 

RGM, all analyses are performed with the 2537A after thermal 

desorption and conversion of the RGM to Hg 0 , a process 

demonstrated to be quantitative with an efficiency of 100% 

(21). Field intercomparisons of paired 1130 instruments 

performed elsewhere over a range of RGM from ∼20 to 400 

pg m -3 showed that the denuder method exhibits good 

precision (( 15%) and that no significant RGM breakthrough 

occurs for the 2-h denuder samples collected over the RGM 

concentration range seen at Barrow (21). We also performed 

an intercomparison for RGM in Barrow using the Tekran 

1130 automated denuder operated adjacent to a manual 

quartz annular denuder of similar design (21). During May 

1-6, 2000, these systems were operated simultaneously on 

the roof of the Barrow CMDL to collect six sets of replicate 

samples. The Tekran 1130 denuder was analyzed as described 

above, while the manual denuders were thermally desorbed 

in a tube furnace, and the exhaust gas was directed into a 

second 2537A for an independent analysis of desorbed Hg 

(21, 25). Five of the six replicates gave excellent results; the 

data were well correlated (r 2 ) 0.91), and the means were 

within ∼5% (Tekran ) 108 ( 23 pg/m 3 , manual ) 115 ( 24 

pg/m 3 ). One set differed by a factor of 2 (manual lower), but 

RGM was increasing rapidly during this period, and the two 

samples were slightly offset in time. 

During the REA flux measurements, the total concentration 

of the three denuder tubes for each run compared 

favorably with a collocated manual denuder sampling system 

as a QA check. On the basis of parallel measurements in the 

lab and field study, the manual denuders exhibit a precision 

of 10%, which infers a sampling error of 40% (95% confidence 

level) for the RGM flux determined by the REA system, given 

an allowed 10% error in the micrometeorological determinations 

of � and σw (23). Flow in the denuders was set by a 

VOL. 36, NO. 6, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 9 1247

mass flow controller to 10 L/min and controlled prior to and 

after each measurement. However, because of flow development 

during switching, a more cautious number for the REA 

system would be a denuder precision of 20%, thus giving a 

sampling error for flux as determined by the REA system of 

∼60% (95% confidence level; 22). 

Approximately half of the Barrow snow samples were split 

and analyzed in duplicate by each analytical lab with good 

agreement ((15%). Assays for bioavailable mercury (bioHg) 

in snow samples were routinely run in replicate with a 

precision on the order of 5-10%. 

Results and Discussion 

Dynamics and Chemistry of Depletion Events. The 1999 

data (Figure 2) provide the first confirmation of MDEs at this 

more southerly Arctic site (Barrow is ∼1600 km south of Alert), 

and the 2000 data show that RGM is produced during MDEs 

(Figure 3). The early 2001 data exhibit comparable trends. 

This is the first evidence that the MDEs could be a widespread 

phenomenon of Arctic dawn and that RGM only appears at 

significant levels when Hg 0 is being depleted. Although others 

have suggested that the depleted Hg 0 at Alert accumulates 

in the aerosol-phase Hg (12), our data clearly indicate an 

important change in gaseous speciation during MDEs, 

producing levels of RGM unprecedented at remote and rural 

sites (8, 18, 25). 

Depletion events begin within a few days of polar sunrise 

(late January) and persist until snowmelt (early June, Figure 

2), suggesting a role of both sunlight and frozen surfaces. 

During this period, Hg 0 exhibits a strong correlation with O3 

(e.g., r 2 ) 0.76 for the period graphed) as also seen at Alert 

(12). Surface O3 destruction is a common feature at Barrow 

during Arctic spring (15). There is no correlation between O3 

and Hg 0 in the months before polar sunrise (r 2 < 0.1). 

Gaseous and aerosol Br also exhibit strong seasonal cycles 

at Barrow and, like RGM, peak annually between January 

and June (32). During this period, aerosol Br increases nearly 

20-fold over typical concentrations and can exceed 100 ng 

m -3 . Hypotheses for the sources of this Br include aerosol 

enrichment by bubble bursting from the sea-surface microlayer, 

gaseous reactions resulting from organic Br emissions 

from marine algae (e.g., bromoform is thought to be 

emitted by ice algae), and/or other aerosol-related reactions. 

The most probable mechanism involves heterogeneous 

reactions at the interface of hygroscopic sea-salt aerosols 

(15), many of which are initiated in the surface microlayer 

of snowflakes or the snowpack (33). Several gaseous reactive 

halogen species (e.g., BrO, see Figure 2) may result with the 

potential to oxidize Hg 0 to gaseous Hg(II) (RGM) compounds. 

Many of these halogen compounds exhibit a strong diel 

pattern (e.g., BrO, ClO, Br, and Cl), indicating the importance 

of sunlight and photochemical reactions (34), as reflected in 

the diel cycle of the airborne Hg species. Our data indicate 

that peak RGM production and Hg 0 depletion generally occur 

at midday under maximum UV (Figure 4). 

We hypothesize that RGM is formed through a rapid, 

in-situ oxidation of Hg 0 in the gaseous phase during MDEs. 

Production of RGM may be attributed to the same photochemically 

active halogen species involved in surface O3 

destruction (15), suggesting that the overall process is 

heterogeneous. On the basis of reported halogen activation 

mechanisms in the remote marine boundary layer 

(34-39), we propose the physicochemical pathways conceptualized 

in Figure 5 for our observation of RGM during 

MDEs. In the reaction mechanism, bromine and chlorine 

radicals are produced autocatalytically from a heterogeneous 

photochemical mechanism involving sea-salt aerosol. 

The halogen radicals (Br/Cl) and halogen oxide radicals (BrO, 

ClO) produced from the ozone destruction reaction (reaction 

1) exhibit strong diurnal patterns with solar radiation (34, 


36) and serve as the primary oxidants that produce RGM 

(Figure 5): 

and/or 

Br/Cl + O 3 f ClO/BrO + O 2 

BrO/ClO + Hg 0 f HgO + Br • /Cl • 

Hg 0 + 2Br • /Cl • f HgBr 2 /HgCl 2 

These proposed mechanisms are plausible for the excellent 

correlation between Hg 0 and O3 concentrations during MDEs. 

There are other reactive halogen species present in the 

Arctic that are thermodynamically favorable in oxidizing Hg 0 

to form RGM in the gaseous phase, including Cl2,Br2, HOCl, 

HOBr, and BrCl. However, we feel that molecular Cl2, Br2, 

and BrCl are not likely to cause the in-situ RGM formation 

seen here as they can be rapidly photolyzed under sunlit 

conditions (34-36). HOCl and HOBr are more resistant to 

solar irradiation. However, since they do not exhibit a strong 

diel cycle in the remote marine boundary layer (35-38), they 

may not account for the observed RGM at Barrow. Figure 5 

identifies two RGM species as proposed products: mercury 

oxide (HgO) and mercury halides (HgBr2/HgCl2). Considering 

the extremely low concentration of all reacting species (both 

Hg 0 and reactive halogen), the bimolecular oxidation of Hg 0 

by halogen oxide would be the favorable RGM formation 

pathway, yielding HgO as the product. However, since the 

denuders may preferentially collect HgBr2/HgCl2 over HgO 

(21), there may also be a significant contribution from halogen 

radicals in the trimolecular oxidation of Hg 0 . On the basis of 

published studies of reactive halogens in the Arctic (15, 32- 

40), HgBr2 should be favored. Research to determine the 

predominant species of RGM at Barrow continues. 

The reaction scheme in Figure 5 explains the strong Hg/ 

O3 correlation and also the extreme seasonality of MDEs: 

the depletions of O3 and Hg 0 require both sunlight and a 

frozen aerosol or snow surface. The reactive halogens are 

initiated by the production of HOBr (or HOCl) from hydroxyl 

radical reactions with Br - (or Cl - ), which are highly concentrated 

in the surface layer of a frozen water droplet or 

snow crystal (40). The abrupt end of the MDEs precisely at 

snowmelt suggests that these reactions do not proceed when 

the droplets deliquesce, decreasing the surface Br - and Cl - 

concentrations as the droplets become homogeneously 

mixed (Figure 5). 

Meteorological factors strongly influence the extreme 

levels of RGM measured at Barrow. We developed a simple 

predictive model for airborne Hg depletion and dry deposition 

to snowpack using local meteorological data and an 

inverse boundary-layer approach (41). Taken together, 

boundary-layer entrainment rates and deposition velocities 

(both modeled as a function of wind speed, UV-B, and air 

temperature) explained ∼70% and ∼80% of the measured 

variance in airborne Hg 0 and RGM, respectively. The highest 

RGM concentrations consistently occurred during periods 

of reduced wind speed and maximum UV-B. Elevated RGM 

also coincided with periods characterized by increased levels 

of BrO in the vicinity of Barrow (e.g., Figure 2). These 

conditions often follow periods of elevated wave activity and 

sea-salt aerosol generation. For example, an extensive area 

of elevated BrO occurred on March 29, 2000, when the 

Beaufort gyre had opened a series of leads north of Barrow. 

Under these conditions, RGM reached 900-950 pg/m 3 (e.g., 

Figure 3, days 91 and 138-148), levels that exceed those 

measured near industrial point sources (8, 42). Even the 

average daytime RGM of 180 pg m -3 at Barrow during March- 

May is several times higher than typical event levels at rural 

sites (8, 25). 

(1) 

(2) 

(3)

FIGURE 2. Trends in Hg 0 at Barrow during 1999, showing a strong correlation between Hg 0 and O3 depletions, and development of a zone 

of elevated BrO near Barrow during this period (mean vertical column BrO is expressed as molecule cm -2 , derived from GOME satellite 

data). 


FIGURE 3. Trends in Hg 0 (upper plot) and reactive gaseous mercury (lower) at Barrow during 2000, showing parallel behavior in total 

Hg in the surface snowpack (lower) and mean daily UV-B (lower). 

Fate of RGM and Dynamics of Hg in Snow. RGM is formed 

continuously as long as O3 and reactive Br are present during 

polar spring, and the observed air concentrations reflect the 

dynamic balance between formation and deposition. Integrating 

the data in Figure 4 suggests that ∼30-40% of the 

depleted Hg 0 appears as airborne RGM and that the 

remainder is deposited to the snow surface directly as RGM 

and/or is scavenged by fine aerosols. Our inverse boundarylayer 

model of the local sink strength (41) predicted Hg 

deposition fluxes of ∼40 µgm -2 at Barrow during February- 

May 1999 and ∼55 µgm -2 during January-May 2000. These 

estimated 5-month fluxes are much higher than annual wet 

deposition rates measured in the northeastern United States 

(∼5-15 µg m -2 yr -1 ; 9) and can be supported only by direct 

deposition of RGM since submicron aerosols are subject to 

far less local deposition than is RGM (8). 

Our year 2000 snow chemistry data confirm these modeled 

rates of Hg accumulation (Figure 3). The measured Hg 

concentrations in the snowpack and calculated snowpack 

depth (86 cm) prior to melt in 2000 (assuming a 70% water 


equivalency) yield an estimated average flux of ∼50-60 µg 

m -2 . This flux is reflected in total Hg in snow, which increased 

steadily from 90 ng/L over this period (Figure 3). 

Such dramatic increases have not been seen in snow from 

lower latitudes (43), but 1998 data from a ship frozen in the 

Beaufort Sea 550 km north of Barrow showed similar temporal 

trends in surface snow (although at lower concentrations; 

44), suggesting that this phenomenon should be investigated 

at higher latitudes. These concentrations of Hg in the May 

snowpack are not only unprecedented for non-Arctic remote 

sites but also exceed typical levels near industrialized regions 

(9, 43). 

The fraction of the total Hg pool in snow that was 

biologically available to bacteria (bioavailable mercury) was 

determined using the mer-lux bioreporter V. anguillarum 

pRB28 and its constitutive control V. anguillarum pRB27. 

The mer-lux bioreporters are genetically engineered bacteria 

that produce light when divalent inorganic mercury (Hg(II)) 

enters their cells (27), thereby distinguishing biologically labile 

from biologically inert Hg(II) species that do not enter the

FIGURE 4. Typical diel cycle of tropospheric gaseous Hg species 

and UV-B at Barrow (UV is measured in near-realtime, while Hg 0 

and RGM represent integrated samples of 5 min and 2 h, respectively, 

as described in the text). 

FIGURE 5. Conceptual diagram of proposed Hg 0 oxidation reaction 

sequences in the Arctic at Barrow. As explained in the text, there 

are several possible pathways and products. This schematic presents 

those that appear most favorable given the observations (dashed 

lines represent inter-phase transport; solid lines are reaction 

pathways). 

bacterial cell. This measurement is of interest because 

microorganisms play an important role in the transformation 

of Hg(II) in the environment (6). In January (2000), prior to 

polar sunrise, bioavailable Hg(II) was undetectable in Barrow 

snow. It then increased from 0.22 ng/L (∼1% of total Hg) in 

March to 8.8 ng/L (nearly 13% of the total Hg) in May. Prior 

to this study, bioavailable Hg(II) had never been measured 

in the Arctic. However, concentrations in snow and precipitation 

in a remote Boreal site in northern Canada were 

on average

FIGURE 6. Trends in several Hg species in the atmosphere and in the snowpack at Barrow around the period of annual snowmelt [during 

the June 4-10, 2000, snowmelt (days 156-162), slushy snow was collected from atop the frozen snowpack for analysis]. Inset photographs 

show the conditions of the snowpack at midday, sky conditions, and mean daily air temperature. 

may be responsible for creating Hg-p as compared to RGM 

and that the Hg-p produced at night is photosensitive. One 

candidate reaction would involve aerosol-bound BrCl that 

would readily oxidize any sorbed Hg 0 but that is rapidly 

decomposed under sunlight (36). However, upon the advent 

of 24-h sunlight in mid-May, the Hg-p/RGM ratio dropped 

to ∼0.1, and the two species were now positively correlated 

(48). The authors speculate that the Hg-p detected after 24-h 

sun reflects RGM sorbed onto the existing aerosol. These 

observations may help explain why the air at Alert appears 

to be characterized by a larger Hg-p/RGM ratio than at Barrow 

(13). The surface reactivity of airborne RGM suggests that it 

would readily partition to the aerosol phase upon formation. 

Hence, Hg-p/RGM ratios may be useful indicators of the age 

(time since oxidation of Hg 0 ), and hence transport distance, 

of depleted air masses. As noted here, our data suggest that 

at least some RGM is being formed in situ at ground level 

at the Barrow site, while Hg 0 in the air sampled at Alert may 

have undergone significant depletion/oxidation events over 

the sea ice prior to being sampled at the Alert station. 

To determine the vertical extent of RGM formation, upper 

air RGM was sampled with heated RGM denuders attached 

to the outer strut of a Cessna 207 and to a mass flow meter/ 

vacuum pump system. Two successive 1-h midday flights 

were conducted on three separate days in late March and 

early April at 1000 m (exterior to the boundary layer) and 100 

m altitude (within the boundary layer) immediately northeast 

of Point Barrow (48). The six aircraft surveys consistently 

showed that RGM exists primarily in ground-level air below 


the marine boundary layer (concentrations decreased from 

an average ∼70 to ∼20 to ∼2 pgm -3 from 5 to 100 to 1000 

m), supporting the hypothesis that the Hg oxidation reactions 

are occurring in the near-surface boundary layer driven by 

halogen compounds derived from sea-salt aerosols. 

We also performed the first RGM dry deposition measurements 

in the Arctic at 3 m above the snowpack using a 

tower-based relaxed eddy accumulation (REA) approach (22, 

24) with manual RGM denuders, as outlined above. Significant 

RGM fluxes measured during March 29-April 12 were 

directed toward the snow surface (overall mean net deposition 

)-0.4 ( 0.2 pg m -2 s -1 ; N ) 9, (1 SE). Computed dry 

deposition velocities for RGM were high, on the order of 1 

cm/s, and agree with those predicted by our inverse 

boundary-layer model (41). Quantifying the extent of areas 

of elevated Hg deposition is clearly an important research 

need, as is understanding its fate. 

Implications of the Barrow Study: Is This a Recent 

Phenomenon and What Is Its Extent? We have convincing 

evidence that tropospheric Hg 0 from the global background 

pool and long-range transport is being depleted from the air 

and converted to a form of rapidly depositing reactive gaseous 

mercury (RGM) following polar sunrise at Barrow, AK. 

Tropospheric oxidation by sea-salt-derived reactive halogens 

involved in O3 depletion (Br and Cl) generates peak levels of 

RGM at this remote site not observed at more southerly 

locations, including those near major point sources. This 

reactive Hg species has a very short atmospheric lifetime 

and accumulates in the Barrow snowpack in forms that

are bioavailable to bacteria. Mercury concentrations and 

accumulation rates in snowpack prior to snowmelt greatly 

exceed those in source regions such as eastern North America, 

and some of this Hg reaches the Arctic tundra ecosystem at 

the initiation of its annual growth cycle. 

Recent reports suggest this Hg oxidation phenomenon 

may exist at many Arctic sites as well as in the Antarctic (12, 

49-51) and could represent an important sink in the global 

cycle of Hg 0 (13). The implications of polar MDEs may be 

assessed by addressing two frequently asked questions: Is 

the phenomenon recent? and Are the polar regions an 

important sink for Hg in the global cycle or likely to become 

so? There are lines of evidence that suggest the answer to 

both questions is yes. 

Is This a Recent Phenomenon? Several data sets suggest 

that there has been a recent increase in Hg levels in Arctic 

biota despite a 20-yr decrease in global atmospheric Hg 

emissions of ∼30% (52). Mercury levels in seabird populations 

monitored within Arctic Canada have roughly doubled in 

the last 20-30 years (53), while Hg accumulation in ringed 

seals and beluga whales has also increased over the last two 

decades (54, 55). Mercury emissions within the Arctic are 

not thought to be increasing (52), and with global emissions 

clearly decreasing, another explanation must be sought. 

We suggest that Arctic MDEs are recent phenomena, 

resulting from changes in Arctic climate that have increased 

atmospheric transport of photooxidants and production of 

reactive halogens (Br/Cl) in the Arctic. Observations show 

that the Arctic region has undergone dramatic physical 

changes in climate over the last 30-40 years, including a 

decreasing trend in multi-year ice coverage, related increases 

in annual ice coverage, later timing of snowfall and earlier 

timing of snowmelt, increasing ocean temperature, and 

increasing atmospheric circulation and temperature (56). The 

changes related to ice formation can impact the dynamics 

of MDEs. The GOME satellite data suggest that BrO enhancements 

are generally absent over multi-year ice (notably 

within the Canadian basin) where ice thickness and windblown 

dust accumulation make sunlight conditions under 

the ice insufficient for algal primary productivity (one source 

of photolyzable Br). As multi-year ice is decreasing, annual 

ice is increasing. The reactive Br surface source is this polar 

annual sea ice region where ice thinness and optical 

transparency support rich under-ice biotic communities. 

Photolyzable bromine (a waste product of ice algae) builds 

up under the ice and escapes through constantly changing 

patterns of open leads and polynyas (open water in an actively 

upwelling region). These dynamic open water areas are also 

sources of sea-salt aerosols, water vapor, and heat from the 

comparatively warm ocean waters. All these products remain 

concentrated in the near surface air due to the lack of vertical 

convection (caused by limited solar input, the high-albedo 

snow/ice surfaces involved, and a positive temperature 

inversion strength (57)), where they react with O3 and other 

photooxidants, leading to oxidation of Hg 0 as described 

earlier. 

Changes in the chemical climate of the Arctic may also 

enhance Hg oxidation reactions. Satellite total ozone mapping 

(TOMS) data indicate an ∼20% decrease in total column 

ozone amounts over the Arctic since 1971, and decreased 

ozone leads to increased surface UV-B exposure (58). The 

link between Hg behavior and UV is clear from our data: 

near-surface RGM during the March-April period at Barrow 

is strongly correlated with a function of incident solar UV-B 

(which controls production of BrO from photolyzable Br) 

and wind speed (which controls the turbulent deposition 

rate) (r 2 ) 0.82; 41). Increased UV radiation reaching the 

troposphere may also result in increased levels of the OH 

radical through photolysis of tropospheric ozone (59). In the 

Arctic atmosphere, increasing OH levels could lead to even 

greater oxidation of Hg 0 because of a positive feedback 

between increasing OH and production of reactive halogens 

(Figure 5). If MDE-enhanced mercury deposition in the Arctic 

is a relatively recent phenomenon (as a result of increased 

synoptic activity and increased annual ice area, for example), 

this could explain the data sets showing a recent increase in 

Hg accumulation in Arctic biota, despite the decrease in global 

atmospheric emissions of Hg in recent decades. 

Are the Polar Regions an Important Sink for Hg in the 

Global Cycle? To address this question, one needs to assess 

the evidence for the spatial extent of the MDE phenomena 

and the extent to which deposited Hg is being re-emitted 

back into the atmosphere during and after snowmelt. 

Depletion events have now been recorded at five widely 

dispersed, primarily coastal, polar sites (12, 14, 49-51). One 

potential indicator of the overall spatial extent of these events 

is illustrated in the monthly GOME maps of BrO distribution. 

The average column BrO concentrations over the Arctic for 

April 2000 are shown in Figure 7. These and related maps 

(13) clearly suggest that MDEs and associated RGM production 

should be concentrated in coastal zones and in areas of 

active open water and might not be expected in other 

locations (e.g., continental Greenland). The bromine source 

regions are concentrated in the dynamic areas of annual sea 

ice, and emission products from these areas are advected 

downwind where reactive halogen compounds form under 

sunlight conditions (e.g., ref 36). The maps suggest that 

horizontal advection of Br compounds to inland and iceshelf 

regions is controlled by prevailing winds and is 

effectively dammed by topographic features such as the 

Brooks, Anadyr, and Rocky Mountain ranges as well as by 

the location of the polar front. The front tends to follow the 

permafrost contours around the pole; the BrO map follows 

roughly these same contours (Figure 7). Note that air over 

the ice-covered Greenland and Ellesmere Islands is relatively 

free of BrO enhancements because the predominating 

katabatic (outward flowing) winds over the icecaps block 

significant inland advection. Oxidation of Hg 0 and enhanced 

deposition of RGM would not be expected in these areas, a 

hypothesis that could be tested by future snow surveys. 

However, coastal locations, such as Nord and Alert, are 

affected by the local marine environment and do experience 

episodic BrO enhancements along with the associated 

mercury depletion events and ozone losses (12, 49). We expect 

that production of oxidized gaseous Hg species will also be 

reported for these areas once new measurements are 

underway in 2002. 

Recent surveys of environmental Hg levels near Barrow 

also indicate similarities in the spatial trends of enhanced 

BrO and Hg accumulation, as would be expected if RGM 

production is dependent on BrO. The concentrations of 

marine-related reaction products taper off with distance from 

the coastline, and Figure 7 illustrates a well-defined inland 

gradient in BrO in Alaska. Mercury levels there are also 

anticorrelated with distance from the coast: Landers et al. 

(3) reported such trends for Hg levels in Arctic Alaskan 

vegetation, and Snyder-Conn et al. (60) reported similar 

trends in total mercury levels in Arctic Alaskan snow. More 

recently, Garbarino et al. (61) showed that mercury concentrations 

in snow over sea ice were highest in the 

predominately downwind direction of the open water leads 

and polynyas surrounding Point Barrow (e.g., to the west), 

an area that often shows enhanced BrO (e.g., Figure 2). 

Comparable Data Exist for the Canadian Arctic. A recent 

report shows that locations of high total mercury concentrations 

in snow are well correlated with areas of high 

atmospheric BrO concentrations, especially in the Canadian 

archipelago (13). Mercury levels in biotic surveys also follow 

these trends; total mercury in Glaucous Gull eggs sampled 

at four coastal locations in Canada are highest in the Canadian 


FIGURE 7. Spatial patterns in monthly mean BrO over the Arctic for April 2000 showing locations of recorded mercury depletion events 

at Barrow, AK; Alert, NWT; Ny-A° lesund, Spitzbergen; and Station Nord, Greenland (mean vertical column BrO is expressed as molecule 

cm -2 , derived from GOME satellite data). MDEs have also been recorded at Neumeyer station, Antarctica, in an area of elevated BrO (not 

shown). 

archipelago where BrO is enhanced (53). The Canadian 

archipelago is dominated by annual ice and open water 

polynyas and leads, and the extensive shorelines and ocean 

currents between the islands create shear zones between 

the “fast” ice grounded to shore and the pack ice moving 

with the ocean currents. This interface area is dominated by 

the open leads that are probable sources of bromine and 

marine products to the near-surface air. A recent estimate 

of the gross atmospheric Hg loading to northern waters in 

this region was 50 T/yr (13). This estimate was based on Hg 

levels in snowpack that were generally lower than those 

reported near Barrow. Other estimates from models (49) and 

our preliminary scaling from GOME BrO data such as Figure 

7(63) range from ∼150-300 T/yr, but all such estimates of 

gross fluxes carry a high uncertainty. 

To assess the overall net strength of the so-called missing 

polar sink (14) using the GOME satellite BrO maps, we must 

fully understand the importance of Hg re-emission during 

snowmelt. Although re-emission is apparent (e.g., Figure 6, 

also ref 12), we and the group working at Alert (62) are yet 

unable to quantify its overall effect on the net accumulation 

of Hg in the Arctic. A simple analysis based on the upslope 

of the Hg 0 concentration in air during Barrow snowmelt (as 

an indicator of the re-emission signal, Figures 1, 2, and 6) 

suggests that melt-related re-emission represents ∼10-20% 

of the deposited Hg (63), and our measurements of Hg in 

runoff indicate that Hg is being transported to the tundra 

during snowmelt. Quantifying the net effect of re-emission 


in the Arctic is clearly an important goal in understanding 

the fate of the deposited Hg. 

Since several lines of evidence support the hypothesis 

that elevated Hg levels exist in both abiotic and biotic pools 

in areas that are characterized by enhanced levels of BrO, an 

additional question arises: What will be the severity and extent 

of mercury depletion/oxidation events in the future? It is 

important to understand how the global mercury cycle will 

be affected by changes within the Arctic, as well as changes 

in atmospheric transport, and future and ongoing domestic/ 

worldwide Hg emission reductions. Since a recent modeling 

study concluded that the concentrations of Hg in the Arctic 

atmosphere were indistinguishable from the global background 

(50), changes in physical climate might actually have 

a greater impact on the arctic Hg cycle than changes in global 

emissions. 

Multi-year ice thickness in the central arctic ocean, as 

measured by U.S. Navy submarines over the last two decades, 

has shown a remarkable 43% reduction in thickness (64). At 

this rate, the Arctic Ocean may become seasonally free of sea 

ice within 30-40 years. If this occurs, it will effectively double 

annual ice coverage, thereby doubling the total area affected 

by mercury depletion/oxidation and enhanced deposition. 

One likely scenario is that climate-driven reductions in multiyear 

ice coverage in favor of increased annual ice coverage 

throughout the Arctic will increase marine primary productivity 

(including ice algal communities). This scenario would 

result in production and release of more photolyzable

omine into the near-surface air. When these surface 

emissions of photolyzable bromine encounter an airmass 

containing Hg 0 emissions from southern latitudes under 

sunlight conditions, mercury depletion/deposition events 

will occur. The springtime mercury deposition rates in the 

Arctic could therefore be related (in the simplest sense) to 

a function of the spatial coverage of annual sea ice, the airmass 

transport of mercury emissions to this region, and local 

airmass circulation. These phenomena are, in turn, controlled 

by average spring and summertime temperatures (as a 

surrogate for melting multi-year ice), by multi-year ice 

coverage (which is decreasing; 65), by synoptic activity 

(increasing), and by variations in the position of the polar 

front (56, 66). There is a clear need for increased research on 

all these phenomena, especially over the Arctic Ocean, to 

determine if mercury depletion/oxidation events in the Arctic, 

and possibly also in the free troposphere at mid-latitudes 

(48, 67), could play an ever increasing role as an important 

sink in the global Hg cycle (13, 14). 

Acknowledgments 

We thank the sponsors of this project for their continued 

support [NOAA Office of Arctic Research, and the U.S. EPA 

Offices of International Affairs (methods speciation work at 

ORNL and Barrow) and Research and Development (BAMS- 

2001 aircraft and Hg-p sampling)]; Dan Endres, Malcolm 

Gaylord, and Glen McConville for local support at CMDL; F. 

Schaedlich for extended help with Tekran equipment; S. 

Oltmans for Barrow ozone data; and Lala Chambers, George 

Southworth, and Mary Anna Bogle for snow analyses and 

data management. GOME spectra have been supplied to A.R. 

by the ESA through the German Aerospace Centre (DFD/ 

DLR Oberpfaffenhofen Germany). The REA system was 

developed under the supervision of H. Skov of NERI-DK and 

T. Meyers of NOAA. K.J.S. was supported by a scholarship 

from the Arctic Institute of North America and by a fellowship 

from the University of Manitoba. M.G. is supported by the 

Danish Research Agency and the Department of Atmospheric 

Environment, NERI-DK. This is Publication No. Hg-72 of 

the ORNL Hg Group. 

Literature Cited 

(1) Wheatley, B.; Wheatley, M. Sci. Total Environ. 2000, 259, 23- 

30. 

(2) Wheatley, B.; Wheatley, M. Arctic Med. Res. 1988, 47 (Suppl. 1), 

163-167. 

(3) Landers, D. H.; Ford, J.; Gubala, C.; Monetti, M.; Lasorsa, B. K.; 

Martinson, J. Water Air Soil Pollut. 1995, 80, 591-601. 

(4) Wagemann, R.; Lockhart, W. L.; Welch, H.; Innes, S. Water Air 

Soil Pollut. 1995, 80, 683-693. 

(5) Zimov, S. A.; Voropaev, Y. V.; Semiletov, I. P.; Davidov, S. P.; 

Prosiannikov, S. F.; Chapin, F. S., III; Chapin, M. C.; Trumbore, 

S.; Tyler, S. Science 1977, 277, 800-803. 

(6) St. Louis, V. L.; Rudd, J. M. W.; Kelly, C. A.; Beaty, K. G.; Bloom, 

N. S.; Flett, R. J. Can. J. Fish, Aquat. Sci. 1994, 51, 1065-1076. 

(7) Fahlke, J.; Bursik, A. Water Air Soil Pollut. 1995, 80, 209-215. 

(8) Lindberg, S. E.; Stratton, W. J. Envir. Sci. Technol. 1998, 32, 49. 

(9) Bullock, O. R., Jr. Sci. Total Environ. 2000, 259, 145. 

(10) Wania, F.; Mackay, D. Ambio 1993, 27, 2079. 

(11) Lin, C.-J.; Pehkonen, S. O. Atmos. Environ. 1999, 33, 2067. 

(12) Schroeder, W. H.; Anlauf, K. G.; Barrie, L. A.; Lu, J. Y.; Steffen, 

A.; Schneeberger, D. R.; Berg, T. Nature 1998, 394, 331. 

(13) Lu, J. Y.; Schroeder, W. H.; Barrie, L.; Steffan, A.; Welch, H.; 

Martin, K.; Lockhart, L.; Hunt, R.; Bolia, G.; Richter, A. Geophys. 

Res. Lett. 2001, 28, 3219-3222. 

(14) Lindberg, S. E.; Brooks, S.; Lin, C.-J.; Scott, K.; Meyers, T.; 

Chambers, L.; Landis, M.; Stevens, R. Water Air Soil Pollut.: 

Focus 2001, 1, 295-302. 

(15) Oltmans, S. J.; Schnell, R. C.; Sheridan, P. J.; Peterson, R. E.; Li, 

S.-M.; Winchester, J. W.; Tans, P. P.; Sturges, W. T.; Kahl, J. D.; 

Barrie, L. A. Atmos. Environ. 1988, 23, 2431. 

(16) Lindberg, S. E.; Vette, A.; Miles, C.; Schaedlich, F. Biogeochemistry 

2000, 48 (2), 237. 

(17) Schroeder, W. H.; Keeler, G.; Kock, H.; Roussel, P.; Schneeberger, 

D.; Schaedlich, F. Water, Air, Soil, Pollut. 1994, 80, 611-620. 

(18) Ebinghaus, R.; Jennings, S. G.; Schroeder, W. H.; Berg, T.; 

Donaghy, T.; Guentzel, J.; Kenny, C.; Kock, H. H.; Kvietkus, K.; 

Landing, W.; Mühleck, T.; Munthe, J.; Prestbo, E. M.; Schneeberger, 

D.; Slemr, F.; Sommar, J.; Urba, A.; Wallschläger, D.; 

Xiao, Z. Atmos. Environ. 1999, 33, 3063. 

(19) Gustin, M.-S.; Lindberg, S. E.; Casimir, A.; Ebinghaus, R.; 

Edwards, G.; Fitzgerald, C.; Kemp, J.; Kock, H. H.; London, J.; 

Majewski, M.; Owens, J.; Marsik, F.; Poissant, L.; Pilote, M.; 

Rasmussen, P.; Schaedlich, F.; Schneeberger, D.; Sommar, J.; 

Turner, R.; Vette, A.; Walshlager, D.; Xiao, Z.; Zhang, H. J. Geophys. 

Res. 1999, 104, 21831-21844. 

(20) Ng, A. C. W.; Corbridge, M. D.; Schneeberger, D. R.; Schaedlich, 

F. H. J. Air Waste Manage. Assoc. 1993, No. 93-TA-39.07, 8. 

(21) Landis, M.; Stevens, R. K.; Schaedlich, F.; Prestbo, P. Environ. 

Sci. Technol. (submitted for publication). 

(22) Goodsite, M. E. Fate of Mercury in the Arctic. Ph.D. Thesis, 

Danish Ministry of Environment and Energy, National Environmental 

Research Institute, 2002. 

(23) Christensen, C. S.; Hummelshøj, P.; Jensen, N. O.; Larsen, B.; 

Lohse, C.; Pilegaard, K.; Skov, H. Atmos. Environ. 2000, 34, 3057- 

3067. 

(24) Businger, J. A.; Oncley, S. P. J. Atmos. Oceanic Technol. 1990, 

7, 349. 

(25) Munthe, J.; Wängberg, I.; Pirrone, N.; Iverfeldt, A° .; Ferrara, R.; 

Costa, P.; Ebinghaus, R.; Feng, X.; Gårdfelt, K.; Keeler, G.; 

Lanzillotta, E.; Lindberg, S. E.; Lu, J.; Mamane, Y.; Nucaro, E.; 

Prestbo, E.; Schmolke, S.; Schroeder, W. H.; Sommar, J.; Sprovieri, 

F.; Stevens, R. K.; Stratton, W.; Tuncel, G.; Urba, A. Atmos. 

Environ. 2001, 35, 3007-3017. 

(26) Bloom, N. S. Can. J. Fish. Aquat. Sci. 1989, 46, 1131. 

(27) Selifonova, O.; Burlage, R.; Barkay, T. Appl. Environ. Microbiol. 

1993, 59, 3083-3090. 

(28) Scott, K. J.; Hudson, R. J. M.; Kelly, C. A.; Rudd, J. W. M.; Barkay, 

T. (manuscript in preparation). 

(29) Richter, A.; Wittrock, F.; Eisinger, M.; Burrows, J. P. Geophys. 

Res. Lett. 1998, 25, 2683. 

(30) Wittrock, F.; Richter, A.; Burrows, J. P. European Symposium on 

Atmospheric Measurements from Space, European Space 

Agency: 1999; WPP-161; pp 735-738. 

(31) Dumarey, R. Anal. Chim. Acta 1985, 170, 337-340. 

(32) Berg, W. W.; Sperry, P. D.; Rahn, K. A.; Gladney, E. S. J. Geophys. 

Res. 1983, 88, 6719. 

(33) Barrie, L.; Platt, U. Tellus 1997, 49B, 450. 

(34) Vogt, R.; Crutzen, P. J.; Sander, R. Nature 1996, 383, 327. 

(35) Dickerson, R. R.; Rhoads, K. P.; Carsey, T. P.; Oltmans, S. J.; 

Burrows, J. P.; Crutzen, P. J. J. Geophys. Res. 1999, 104, 21, 385. 

(36) Fan, S.-M.; Jacob, D. J. Nature 1992, 359, 522. 

(37) McConnell, J. C.; Henderson, G. S.; Barrie, L.; Bottenheim, J.; 

Niki, H.; Langford, C. H.; Templeton, E. M. Nature 1992, 355, 

150. 

(38) Mozurkewich, M. J. Geophys. Res. 1995, 100, 14, 199. 

(39) Richter, A.; Wittrock, F.; Eisinger, M.; Burrows, J. P. Geophys. 

Res. Lett. 1998, 25, 2683. 

(40) Foster, K. L.; Plastridge, R. A.; Bottenheim, J. W.; Shepson, P. B.; 

Finlayson-Pitts, B. J.; Spicer, C. W. Science 2001, 291, 471. 

(41) Brooks, S.; Lindberg, S. E. J. Geophys. Res. (in press). 

(42) Sheu, G.-R.; Mason, R. P. Environ. Sci. Technol. 2001, 35, 1209. 

(43) St. Louis, V. L.; Rudd, J. W. M.; Kelly, C. A.; Barrie, L. A. Water 

Air Soil Pollut. 1995, 80, 405. 

(44) Welch, H. K.; Martin, K.; Lockhart, W. L.; Hunt, R. V.; Boila, G. 

In Synopsis of Research Conducted under the 1997/98 Northern 

Contaminants Program; Jensen, J., Ed.; Department of Indian 

Affairs and Northern Development: Ottawa, 1999; p 93. 

(45) Scott, K. J. Ph.D. Thesis, University of Manitoba (in preparation.). 

(46) Amyot, M.; Mierle, G.; Lean, O. R. S.; McQueen, D. J. Environ. 

Sci. Technol. 1994, 28, 2366. 

(47) Zhang, H.; Lindberg, S. E. Environ. Sci. Technol. 2001, 35, 928- 

935. 

(48) Landis, M. S.; Stevens, R. K.; McConville, G.; Brooks, S. R. 

Presented at the 6th International Conference on Mercury as 

a Global Pollutant, Minamata, Japan, October 2001 (manuscript 

in preparation). 

(49) Skov, H.; Christensen, J. H.; Goodsite, M. E.; Petersen, M. C.; 

Zeuthen-Heidam, N.; Geernaert, G.; Olsen, J. Dynamics and 

chemistry of atmospheric mercury; Danish contribution to 

EUROTRAC MEPOP report, Near surface conversion and fluxes 

of gaseous elemental mercury to reactive gaseous mercury in 

the Arctic, 2001. 

(50) Berg, T.; Sekkeseter, S.; Steiness, E.; Valdal, A.; Wibtoe, G. 

Presented at the 6th International Conference on Mercury as 

a Global Pollutant, Minamata, Japan, October 2001. 


(51) Ebinghaus, R.; Kock, H. H.; Temme, C.; Einax, J. W.; Löwe, A. 

G.; Richter, A.; Burrows, J. P.; Schroeder, W. H. Environ. Sci. 

Technol. 2002, 36, 1238-1244. 

(52) Pacyna, J. M.; Pacyna, E. G. Environ. Rev. (in press). 

(53) Braune, B. M.; Donaldson, G. M.; Hobson, K. A. Environ. Pollut. 

2001, 114, 39-54; followup (-II) paper by same authors 

(in press). 

(54) Hyatt, C. K.; Trebacz, E.; Metner, D. A.; Wagemann, R.; Lockhart, 

W. L. Presented at the 5th International Conference on Mercury 

as a Global Pollutant, Rio de Janeiro, May 1999. 

(55) Wagemann, R.; Innes, S.; Richard, P. R. Sci. Total Environ. 1996, 

186, 41-66. 

(56) Dickson, R. R. Nature 1999, 397, 389-391. 

(57) Kahl, J. D. Int. J. Climatol. 1990, 10, 537-548. 

(58) Cahill, C.; Weatherhead, E. Chron. NSF Arctic Sci. Prog. 2001, 

8, 1. 

(59) Madronich, S.; Granier, C. Geophys. Res. Lett. 1992, 19, 465. 

(60) Snyder-Conn, E.; Garbarino, J. R.; Hoffman, G. L.; Oelkers, A. 

Arctic 1997, 50, 201-215. 

(61) Garbarino, J. R.; Snyder-Conn, E.; Leiker, T. J.; Hoffman, G. L. 

Water Air Soil Pollut. (in review). 


(62) Schroeder, W. H. Meteorological Service of Canada., Toronto, 

personal communication. 

(63) Lindberg, S.; Brooks, S.; Lin, C.-J.; Scott, K.; Landis, M.; Stevens, 

R. Presented at the 6th International Conference on Mercury 

as a Global Pollutant, Minamata, Japan, October 2001. 

(64) Rothrock, D. A.; Yu, Y.; Maykut, G. A. Geophys. Res. Lett. 1999, 

26, 3469-3472. 

(65) Maslanik, J. A.; Serreze, M. C.; Agnew, T. Geophys. Res. Lett. 

1999, 26, 1905-1908. 

(66) Serreze, M. C.; Box, J.; Barry, R. G.; Walsh, J. E. Meteorol. Atmos. 

Phys. 1993, 51, 147-64. 

(67) Murphy, D. M.; Thompson, D. S. Geophys. Res. Lett. 2000, 27, 

3217. 

Received for review August 15, 2001. Revised manuscript 

received December 19, 2001. Accepted December 20, 2001. 

ES0111941


Paper 3: Skov, H., Christensen, J., Goodsite, M.E., Heidam, N.Z., Jensen, B., Wåhlin, P., 

Geernaert, G. The Fate of Elemental Mercury in Arctic during Atmospheric Mercury 

Depletion Episodes and the Load of Atmospheric Mercury to Arctic. Submitted to 

Environmental Science and Technology, June 2003.

The fate of elemental mercury in Arctic during atmospheric mercury depletion 

episodes and the load of atmospheric mercury to Arctic 

Authors: Henrik Skov*, Jesper Christensen, Michael E. Goodsite 1 , Niels Z. Heidam, Bjarne Jensen, 

Peter Wåhlin and Gerald Geernaert 2 . 

*Corresponding author. 

National Environmental Research Institute, Frederiksborgvej 399, 4000 Roskilde, Denmark. 

2 

Now at Dept. of Chemsitry, Univ. of Southern Denmark, Campusvej 55, DK-5230 Odense M, 

Denmark 

1 

Now at Inst. of Geophysics and Planetary Physics, Los Alamos National Laboratory, MS C-305 

Los Alamos, NM 87545, USA. 

Abstract 

Atmospheric mercury depletion episodes (AMDE’s) have been studied at Station Nord, Northeast 

Greenland, 81 0 36'N, 16 0 40'W, during Arctic Spring. Gaseous elemental mercury (GEM) and ozone 

were measured starting from 1998 and 1999, respectively until August 2002. GEM was measured 

with a TEKRAN 2735A automatic mercury analyser based on pre-concentration of mercury on a 

gold trap followed by detection using fluorescence spectroscopy. Ozone was measured by UV 

absorption. A scatter plot of GEM and ozone confirmed that also at Station Nord GEM and ozone 

are linearly correlated during AMDEs. The relationship between ozone and GEM is further 

investigated in this paper using basic reaction kinetics, i.e., Cl, ClO, Br and BrO have been 

suggested as reactants for GEM. The analyses in this paper show that GEM most probably is 

reacting with Br. 

Based on the experimental results of this paper and results from the literature a simple 

parameterisation for AMDE was included into the Danish Eulerian Hemispheric Model (DEHM). 

GEM is converted linearly to reactive gaseous mercury (RGM) over sea ice with temperature less 

than – 4 o C with a lifetime of 3 to 10 hours. The new AMDE parameterisation was used together 

with the parameterisation of mercury chemistry (21). The obtained model results were compared 

with measurements of GEM at Station Nord. There was good agreement between the start and 

general features of an AMDE, though the fast concentration changes could not be reproduced by the 

model. Furthermore, the modelled RGM concentrations over the Arctic were found to agree well 

with the temporal and geographical variability of the surface column of BrO observed by the 

GOME satellite. 

Scenario calculations were carried out with and without an AMDE. The resulting mercury 

deposition was calculated at various locations in the Arctic and sub-Arctic and was found to be of 

the same general magnitude. For the area north of the Polar Circle the mercury deposition was 

demonstrated to increase from 89 tons/year for calculations without an AMDE to 208 tons/year 

while with the AMDE, 208 tons/year represents the upper limit for the mercury load. 

Introduction 

Mercury is found at high levels in marine animals at many places in the Arctic and North Atlantic 

Ocean (1). It has been shown that the present levels of mercury in sea animals have a negative 

effect also on the health of the local populations, who in turn depend on these animals as food 

supply (2). Furthermore, the input of atmospheric mercury to the Arctic environment has at least 

tripled compared to pre-industrial times (3). 

The lifetime of gaseous elemental mercury (GEM) in the atmosphere (95% of atmospheric 

mercury) is in general about 1 year (4). However, in the Arctic during spring, the lifetime of GEM 

is significant shorter and GEM is observed to be depleted in less than one day during mercury 

1

depletion episodes (AMDE), (5,6,7,). During an AMDE, GEM is converted to oxidised mercury in 

the gas phase the so-called reactive gaseous mercury (RGM) that deposits fast to the ground (6,8,). 

AMDE occurs typically in the Arctic from March to June, which coincidentally occurs when the 

marine arctic ecosystems are also extremely active due to the increasing solar flux combined with 

the melting of sea ice (9). Therefore, it is hypothesised that there is a higher efficiency of 

bioaccumulation of mercury than would be expected from extrapolating data from mid-latitudes to 

the Arctic. Thus it is very important to get a fully understanding of the processes responsible for the 

AMDE’s. 

Ozone has been observed to be depleted during the Arctic spring for nearly 15 years (10) and it is 

well accepted that it is due to photochemical degradation after the polar sunrise. More recently it 

has been demonstrated that GEM is depleted as well and that GEM is strongly correlated with 

ozone during an AMDE (5). Ref. (6) showed that solar radiation and surface temperature of marine 

ice either control or are proxies for controls driving the depletion of GEM. However, very little is 

known about the reaction(s) transforming GEM to RGM. Several hypotheses have been proposed 

where Cl and/or Br atoms or ClO and/or BrO are common candidates initiated by the heterogeneous 

reaction between ozone and sea salt chloride or bromide on the surface of sea ice with temperatures 

below –4 o C. Thus AMDE is expected to be limited to areas exposed to marine air. Though the ratio 

is 300/1 of Cl - /Br - in seawater, a series of physical and chemical processes favours the liberation of 

bromine compared to chlorine from the sea ice (11). 

Measurements of GEM and ozone at Station Nord, Northeast Greenland are presented in this paper. 

The data are treated using basic reaction kinetic theory and physical theory for mixing of gases in 

the atmosphere. The results of the measurements presented here and results from other studies are 

used to make a simple parameterisation of AMDE, which was then used in the DEHM (12, 13). The 

results of the model calculations are compared with measured values of GEM in this study and with 

satellite observations of BrO (from the GOME satellite). The model is used to calculate the burden 

of atmospheric mercury to some coastal areas in the Arctic and Sub-arctic. Finally the importance 

of AMDE for the burden of atmospheric mercury to the Arctic is determined by scenario 

calculations with and without AMDE. 

Experimental 

Measurements 

Weekly average concentrations of atmospheric bromine were determined from samples collected at 

40 l/min through a 42 mm inlet diameter on a particle filter, the first of a series of 47 mm filters in a 

filter pack. A detailed description of the filter pack system is given elsewhere, (14). The resulting 

filters were transported to the laboratory where they were analysed by proton induced x-ray 

emission (PIXE) that is capable of detecting elements heavier than Aluminium (15)). The results of 

the analysis showed that the concentration of bromine atoms was the same on the front and reverse 

side of the filter. If bromine were present only in particle phase, it would have been observed only 

on the front side. Therefore a significant fraction of the measured bromine must have been present 

in the gas phase (16). The measured bromine is thus called filterable bromine, fBr. The uncertainty 

of the method is estimated to be 25 % at a 95% confidence interval. 

Ozone and GEM were measured starting in 1998 and 1999 respectively and until June 2002. GEM 

was only measured from February through mid-summer (July or August) each year. The monitoring 

site is at the Danish military base at Station Nord, Northeast Greenland located at 81 0 36'N, 

16 0 40'W, see Fig. 1a. The measurements were performed at the Danish AMAP site, ‘Flyger’s 

Hytte’, a laboratory hut located approximately 3 km south of the central complex of buildings, as 

shown on the map in the Fig. 1b. The temperature in the hut was constant at 15±1 0 C. 

2

Ozone was measured with an UV absorption monitor, API , with a detection limit of 1 ppbv and an 

uncertainty of 3 % for concentrations above 10 ppbv and 6 % for concentrations below 10 ppbv, (all 

uncertainties are at a 95% confidence interval) (17). 

GEM was measured by a TEKRAN 2537A mercury analyser. The principle of the instrument is as 

follows: A measured volume of sample air is drawn through a gold trap that quantitatively retains 

elemental mercury. The collected mercury is desorbed from the gold trap by heat and is transferred 

by argon into the detection chamber, where the amount of mercury is detected by cold vapour 

atomic fluorescence spectroscopy. The detection limit is 0.1 ng/m 3 and the reproducibility for 

concentrations above 0.5 ng/m 3 is within 20% on a 95% confidence interval based on parallel 

measurements with two TEKRAN 2537A mercury analysers. It is not at present reasonable to give 

the combined uncertainty of the method following the guidelines of ISO 14956, as the exact identity 

of the measured mercury is unknown, though GEM is determined as the dominant compound (18). 

In order to protect the instrument against humidity and sea salt, a soda-lime trap was placed in the 

sample line just in front of the analyser before the 2001 season in order to avoid passivation of the 

gold traps (19). However, no change in the level of GEM at Station Nord was observed after the 

installation of the trap. Parallel measurements of GEM in Denmark at a site not directly influenced 

by sea spray with and without soda lime trap showed a perfect agreement within the uncertainty 

established previously. 

The model 

The Danish Eulerian Hemispheric Model System (DEHM) is described in detail elsewhere (12,13), 

so only a short description is given here. The model system consists of two parts: a meteorological 

part based on the PSU/NCAR Mesoscale Model version 5 (MM5) (20) and an air pollution model 

part, the DEHM model. The model system is driven by the global meteorological data obtained 

from the European Centre for Medium-range Weather Forecasts (ECMWF) on a 2.5 o x 2.5 o grid 

with a time resolution of 12 hours. 

The DEHM model is based on a set of coupled full three-dimensional advection-diffusion 

equations, one equation for each species. The horizontal mother domain of the model is defined on 

a regular 96x96 grid that covers most of the Northern Hemisphere with a grid resolution of 150 km 

× 150 km at 60 o N. The vertical resolution is defined on an irregular grid with 20 layers up to about 

15 km reflecting the structure of the atmosphere. 

The chemistry scheme of mercury in the atmosphere includes 13 mercury species: 3 in gas-phase 

(Hg 0 , HgO and HgCl2), 9 species in the aqueous-phase and 1 in particulate phase and is adopted 

from the literature (21). Furthermore an additional 1. order reaction of GEM was added where GEM 

is reacting to form RGM inside the boundary layer over sea ice during sunny conditions in order to 

mimic an AMDE. The reaction rate constant for the 1. order removal is based on the observed 

removal rates of mercury. The measurements gave a 1. order lifetime of GEM between 3 and 10 

hours, so scenario calculations with this range of lifetime were carried out. The fast 1. order 

oxidation was stopped, when surface temperature exceeded -4 o C as this temperature appears to be 

crucial for the presence of AMDE as Br2 and BrCl are formed at the surfaces of re-freezing leads 

and leads are representing the most important halogen source during AMDE (6). Except for the 

AMDE, the fate of GEM was controlled by a slow chemical removal in the gas phase and uptake by 

cloud water. 

The dry deposition velocities of the reactive gaseous mercury species are based on the resistance 

method, where the surface resistance is similar to nitric-acid, i.e. 0 s m -1 (8). The wet deposition of 

reactive and particulate mercury is parameterised by using a simple scavenging coefficients 

formulation with different in-cloud and below-cloud scavenging coefficients (12). 

The emissions of anthropogenic mercury are based on the global inventory of mercury emissions 

for 1995 on a 1 o x1 o grid (22) including emissions of GEM, RGM and total particulate mercury 

3

(TPM). The model does not contain any re-emissions from land and oceans. Instead, a background 

concentration of 1.5 ng/m 3 of Hg 0 is used as initial concentrations and boundary conditions. The 

mercury model has been run for the period October 1998 to October 2002. 

Results and discussion 

Measurements 

The results of ozone and GEM measurements are shown in Fig. 2 together with concentrations of 

fBr. It is seen that ozone and GEM are rather stable from September/October until the end of 

February/beginning of March. Then a highly perturbed period is occurring where ozone and GEM 

within hours are both depleted to 0 from about 40 ppbv and 1.5 ng/m 3 , respectively. The 

concentrations remain at 0 for periods that may last at some hours up to several days before 

suddenly rising again. At the same time fBr increases and reaches a maximum of about 10 ng/m 3 . In 

July, the ozone concentration stabilises just above 20 ppbv and then it slowly increases to about 40 

ppbv in September/October and fBr decreases to values close to zero. 

GEM was measured from February to the end of July or to the beginning of August. The 

measurements were focusing on the description of the AMDE. Previously a investigation has 

described that ozone and GEM are simultaneously depleted and that they are highly correlated 

during these depletion episodes (5). This is indeed confirmed by the present data set, Fig. 3. After 

the depletion period some very high concentrations of GEM appeared with values above 2 ng/m 3 . 

Similar observations were also done at Alert (5), at Barrow (6) and at Svalbard (7) and they are 

attributed to reemission of mercury to the atmosphere. 

The strong correlation between ozone and GEM suggests that they are dependent on a mutual 

factor. A direct reaction between ozone and GEM can be ruled out due to the long lifetime of GEM 

with respect to the present ozone concentrations (4). In a field study (23) BrO was observed to build 

up when ozone is decreased due to the reaction: 

O Br ⎯⎯→O 

+ BrO 

3 + 2 

(1). 

Therefore serious candidates for GEM removal are Br or BrO. However Cl and ClO cannot be 

ignored as significant Cl removal of organic compounds have been observed during AMDE (e.g. 

ref. (24) and the importance of these species depends on their concentrations and their reactivity 

towards GEM. 

The lifetime of GEM is observed to be typically about 10 hours during AMDE. Up to 30 ppt of ClO 

and 30 ppt BrO have been observed (25) and thus the resulting rate constants for the reactions 

between GEM and BrO and/or ClO can be estimated to be in the order of 4Χ10 -14 cm 3 molec -1 sec -1 , 

see reaction 2 and 3 respectively: 

Hg + BrO ⎯⎯→ 

HgO + Br 

(2) 

and/or 

Hg + ClO ⎯⎯→ 

HgO + Cl 

(3) 

The most probable product of reaction 2 and/or reaction 3 is the formation of HgO, which has a 

very low vapour pressure (9Χ10 -12 Pa, (26)). Thus the reaction would lead to the formation TPM 

and not RGM as observed (6). Furthermore in a thermodynamic study calculations were carried out 

4

demonstrating that BrO and ClO with GEM are endothermic (27) and thus reaction 2 and 3 are most 

probably not important for the removal of Hg o in the atmosphere. 

Instead of BrO and ClO, GEM may react with Cl and/or Br. Therefore the data were analysed 

assuming relative rate conditions between ozone and GEM. The method is widely used under 

laboratory conditions, where the reaction of interest is proceeding in competition with another 

reaction with a well-known reaction rate constant. So the system here is: 

and 

O + X → product 

3 (4) 

Hg + X → product 

(5), 

where X is either Br or Cl and assuming that all other reactions are of negligible importance for the 

removal of ozone and GEM. 

The kinetic equations for reaction 4 and 5 are: 

and 

[ GEM ] 

∫ 

[ GEM ] 

[ O ] 

3 

∫ 

[ O ] 

t 

t 

d ln [ O ] = −k 

[ X ]dt 

(6) 

3 0 

3 

t 

∫ 

4 

0 

t 

d ln [ GEM ] = −k 

[ X ]dt 

(7), 

0 

respectively. 

∫ 

5 

0 

Integrating equation and dividing equation 6 with 7 the relative rate expression is obtained. 

[ GEM ] 

[ GEM ] 

[ ] 

[ ] ⎟⎟ 

O ⎞ 3 0 

O 

⎛ 

⎛ 

0 ⎞ k5 

ln ⎜ ⎟ = • ln⎜ 

⎜ 

(8). 

⎝ 

t ⎠ k4 

⎝ 3 t ⎠ 

A plot of ln([GEM]0/[GEM]t against ln([O3]0/[O3]t should give a straight line with intercept 0 and a 

slope equal to k5/k4. Fig. 4 shows such a plot using the data from Station Nord. Data were selected 

for periods where the initial concentration of GEM was above 0.4 ng/m 3 in order to ensure good 

signal to noise ratio and where three consecutive measurements of both ozone and GEM are 

decreasing. All measurement included were in periods with 24 hour daylight. There is a strong 

linear correlation (

the literature of the reactions of Hg o with halogen atoms are also listed. All laboratory results are 

obtained using relative rate conditions. 

The reactions of halogens with Hg are independent of temperature (29,30) and therefore results of 

the various studies should be directly comparable. On the other hand ozone reactions with halogen 

atoms are temperature dependent and thus the rate constants obtained here of Hg are calculated at 

233 K and 263 K representative for the conditions in Arctic during mercury depletion. 

In general the half-life of GEM at Station Nord is 3 to 10 hours during an AMDE. Using rate 

constants of the latest study (31) this lifetime corresponds to a concentration of Br or Cl at 1-3 pptv 

and 0.1-0.3 pptv, respectively. Laboratories report Cl concentrations in the Arctic during AMDEs 

from 0.001 to 0.004 ppt (24,32,33), at least a factor of 25 lower in concentration than needed for the 

observed GEM depletion, whereas Br in the ppt level is reported by many authors (24,25,32,33). 

This implies that Br most probably is the key species leading to mercury depletion. 

Based on the results presented above the most important reactant for GEM removal is Br and the 

results in the study indicate a second order reaction rate constant of about 1•10 -12 cm 3 molec -1 sec -1 . 

The above result needs confirmation in the laboratory and though the result is a strong indication of 

the removal channel for GEM other possibilities needs to be examined before a definite answer can 

be given. In particular, it is important to clarify the role of heterogeneous chemistry during 

AMDE’s. Furthermore, there is a factor 16 difference in the reaction rate constant of the reaction 

between Br and Hg o reported (34,31) and in both cases determined by the relative rate technique. 

The difference shows that secondary reactions in the laboratory systems clearly plays a role and 

thus the reaction rate constant needs to be determined by an absolute method. To the knowledge of 

the authors there is not any study of the reaction rates constant of the reactions between Hg and ClO 

or BrO in the literature. Interesting enough there is good agreement between the theoretical 

calculated rate constant (27) and the rate constant extracted from the field measurements presented 

here. However, that might be a coincidence. 

The reaction between Br and Hg o leads to the formation of a radical. 

Hg + Br → HgBr • 

(9). 

The further fate of this radical is at the moment speculative but is the subject for a theoretical study 

(27,35). 

Modelling 

The mercury model has been run for October 1998 to October 2002 and the model results for GEM 

(Hg o ) are compared with measurements from Station Nord. The model does a good job reproducing 

the occurrence and length of AMDE but the simple parameterisation does not and cannot describe 

the fast variations in the period during the spring period, (Fig. 5). However, the main structure is 

reproduced and the results present a large step forward in the understanding of the fate of mercury 

in the Arctic atmosphere. The reproduction of the main structure is a strong indication for that the 

limiting factor is the surface conditions, which has to be sea ice with surface temperature below –4 

o C. A comparison of three model runs with observed GEM is shown in Fig. 5. The three model runs 

are: 1) without depletion, 2) with depletion where lifetime during depletion is 10 hours, and 3) with 

depletion and lifetime on 3 hours. The main difference between the two last runs with depletion is 

only a slightly higher minimum level in the case of 10 hour life time. The variations and duration of 

episodes are not changed. 

6

In the model the removal of GEM leads to build-up of RGM as observed in the field (6). The 

calculated concentrations of RGM have been compared with measured integrated surface column of 

BrO as obtained from the GOME satellite (36). In Fig. 6 the mean BrO column near the surface (in 

the boundary layer) and RGM concentrations for each month are shown for the period of January to 

June 2000. The figures show clearly that BrO and RGM have the same general temporal and 

geographical variability and reach their maximum level and extension in April to May. This finding 

supports that the conversion of GEM in fact is connected to sea-ice with temperatures below –4 o C 

and to the chemistry of Br. Notwithstanding there is general good agreement between RGM and 

BrO, some clear discrepancies can be observed. In May the largest BrO concentrations are found 

along the coast of the Beaufort Sea (North of Canada and Alaska), whereas maximum RGM 

concentrations are predicted North of Greenland. At present there is no explanation for this 

observation but most probably it reflects the rough assumption that RGM is produced ubiquitously 

above surfaces with temperatures below –4 o C. Bromine is most probably formed on the surface of 

re-freezing leads (6). These leads form and disappear again more or less randomly around the Arctic 

Ocean depending on the oceanic currents, wind, temperature and solar flux. Therefore large 

variation in the concentrations of bromine is expected during spring and thus also in the removal of 

GEM and build up of RGM concentration. This feature is in fact clearly seen in the measurements 

of GEM, Fig. 2 and 5; and it explains the discrepancy between the model results giving a smooth 

depletion event extending for the whole depletion period, whereas measurements show a long series 

of shorter depletion episodes during the depletion period. 

The deposition of atmospheric mercury calculated with DEHM is shown in Fig. 7 for some selected 

locations in the Arctic: Station Nord (Greenland), Barrow (Alaska), Alert (Canada), Thule 

(Greenland), Spitzbergen (Norway), in the sub-Arctic, Nuuk area (south Greenland), and on Faeroe 

Islands and Denmark. The total deposition is divided into three components: the contribution from 

deposition of; RGM, photo-chemically formed TPM, and directly emitted TPM. It is clearly seen 

how important the depletion phenomenon is for the deposition of mercury in the high Arctic, 

whereas it has practically no importance in the Faroe Islands. However, the Nuuk area appears to be 

influenced by AMDEs and there are on going field activities to confirm experimentally if AMDEs 

extends to this sub-arctic area. The importance of AMDE for the mercury load is seen for example 

for the Thule area, where the total deposition of mercury is increased by a factor 3, while for the 

Faeroe Islands the depletion phenomena only lead to a 10% increase of the mercury deposition. The 

contribution from directly emitted particulate mercury is very small in the high Arctic. It is mainly 

the large atmospheric reservoir of elemental mercury, which contributes through its chemical 

conversion to RGM followed by fast deposition of RGM. For all places close to the sea the total 

deposition is at the same levels. However, the large contribution of RGM deposition in the Arctic 

occurs only over a 4 month period from March to June where the algae bloom occurs (9). This fact 

might lead to a higher uptake of mercury in the food chain than would be expected if one simply 

extrapolated data from mid latitudes to the Arctic. 

The deposition of mercury for 1999 and 2000 is shown in Fig. 8 for the Northern Hemisphere with 

and without AMDE’s. The largest depositions are found close to the sources in Asia, Europe and 

North America mainly due to the deposition of primarily emitted RGM and TPM that is removed 

fast mainly due to dry deposition and washout by rain. The calculations with and without AMDE’s 

show again the importance of AMDE in the Arctic for the total deposition of mercury. Here the 

photochemically formed RGM is removed mainly by dry deposition as the Arctic is characterized 

by its very dry climate. The total annual deposition increases in the whole Arctic, and for the area 

north of the Polar Circle the total deposition of mercury increases from 89 to 208 tons/year due to 

the depletion. 

7

While we believe that we have made a major advance in understanding the chemistry governing 

MDEs, these results associated with mercury load estimates have to be taken with caution for 

several reasons. 

• There is far from full understanding of the chemical processes controlling atmospheric mercury 

and as a consequence the parameterization of AMDE in the model is not at present adequate for 

a reliable quantitative calculation of the mercury burden. 

• The source of Br is not well described and thus the temporal and geographical variability is not 

well described 

• Evidence is reported for re-emission of mercury to the atmosphere, which is not included in the 

model 

For these reasons the 208 tons/year represents only an estimate of the upper limit for the mercury 

load to the Arctic area. 

Acknowledgement 

A. Richter is acknowledged for providing us with the BrO data from the GOME satellite. 

The Danish Environmental Protection Agency financially supported this work with means from the 

MIKA/DANCEA funds for Environmental Support to the Arctic Region. The findings and 

conclusions presented here do not necessarily reflect the views of the Agency. 

The Royal Danish Air Force is acknowledged for providing free transport to Station Nord and the 

staff at Station Nord are specially acknowledged for an excellent support 

Michael E. Goodsite was financially supported by NERI and the Danish Research Council. 

References 

1. AMAP Greenland 1994-1996, (1997) Publisher: Ministry of the Environment and Energy, 

Danish Environmental Protection Agency, ISBN 87-7810-762-8. 

2. Grandjean, P. Weihe, P. White, R.F and Debes, F. Env. Res. Sec A. (1998) 77, 165-172. 

3. Shotyk, W. Goodsite, M.E. Roos-Barraclough, F. Frei, R. Heinemeier, J. Asmund, G. Lohse, C. 

Hansen, T.S. Anthropogenic contributions to atmospheric Hg, Pb and As accumulation recorded 

by peat cores from southern Greenland and Denmark dated using the 14C “bomb pulse curve” 

Geochimica et Cosmochimica Acta Accepted 02 June, 2003 

4. Lin C-J. and Pehkonen, S.O. Atm. Envir.333 (1999) pp. 2067-2070. 

5. Schroeder, W.H., Anlauf, K.G., Barrie, L.Y., Lu, A., Schneebeerger, D.R. and Berg, T. Nature 

(1998) 394, 331-332. 

6. Lindberg, S.E. Brooks, S. C-J. Lin, C-J. Scott, K.J. Landis, M.S. Stevens R.K. Goodsite, M. and 

Richter, A. E. S.& T. (2002) 36, 1245-1256. 

7. Berg, T. Batnicki, J. Munthe, J. Lattila, H. Hrehoruk, J. and Mazur, A. Atm. Envir. (2001) 35, 

2569-2582. 

8. Goodsite, M.E. Brooks, S.B.Lindberg, S.E. Meyers, T.P Skov, H. and Larsen, M.R.B. (2003) 

Measuring reactive gaseous mercury flux by relaxed eddy accumulation using KCl coated 

annular denuders. Under preparation. 

9. Hansen, A.S. Nielsen, T.G. Levinsen, H. Madsen, S.D. Thingstad, T.F. and Hansen, B.W. 

Deep-Sea Res. Part I (2003) 50, 171-187. 

10. Barrie, L.A. Bottenheim, J.W. Schnell, R,C. Crutzen, P.J. and Rasmussen, R.A. Nature (1988) 

334, 138-141. 

11. Foster, K.L. Plastridge Atm. Env. , R.A. Bottenheim, J.W. Shepson, P.B. Finlayson-Pitts, B.J. 

Spicer C.W. Science (2001) 291, 471-474. 

12. Christensen, J. Atm. Env. (1997) 31, 4169-4191. 

13. Christensen, J.H. Brandt, J. Frohn, L.M. and Skov, H. Submitted to Atm. Chem. Phys. April 

2003. 

14. Heidam, N.Z. Wåhlin, P. and Christensen, J.H. J. Atm. Sciences, (1999) 56, 261-278. 

8

15. Kemp, K. and Wåhlin, P. Application of Accelerators in Research and Industry, Proceedings of 

the 15. International Conference (eds. J.L. Duggan and J.L. Morgan), AIP, Woodbury, New 

York (1999) 472-475. 

16. Impey, G.A. Mikele, C.M. Anlauf, K.G. Barrie,L.A. Hastie, D.R. and Shepson, P.B. J. At.Chem. 

(1999) 34, 21-34. 

17. H. Skov, A.H. Egeløv, K. Granby and T. Nielsen. Atm. Environ. 1997, vol. 31 No. 5, p 685- 

691. 

18. Ebinghaus, R. kock, H.H. Schmolke, S.R. Fres. Jour. Anal. Chem. (2001) 371, 806-815 

19. Skov, H. Nielsdóttir, M.C. Goodsite, M.E. Christensen, J. Skjøth, C.A. Geernaert, G. Hertel, O. 

Olsen, J. Accept. Asian Chemistry Letters (2003). 

20. Grell, G. A., Dudhia J. and Stauffer D. R., A Description of the Fifth-Generation Penn State/NCAR 

Mesoscale Model (MM5). NCAR/TN-398+STR. NCAR Technical Note. June 1995. Mesoscale 

and Microscale Meteorology Division. National Center for Atmospheric Research. Boulder, 

Colorado, pp. 122, 1995. 

21. Petersen, G. Munthe, J. Pleijel, K. Bloxam, R. and Vinod Kumar, A., Atm. Env. (1998) 32, 829- 

843. 

22. Pacyna E.G. and Pacyna J.M. Water, Air and Soil Pollution (2002) 137, 149-165. 

23. Hausmann, M. and Platt, U. J. Geophys. Res. (1994) 99, D12, 25,399-25,413. 

24. Boudries, H. and Bottenheim, J.W. Geophys. Res. Let. (2000) 27,4. 517-520. 

25. Tuckermann, M. Ackermann, R. Gölz, C. Lorenzen-Schmidt, H. Senne, T. Stutz, J, Trost, B. 

Unold, W. and Platt, U. Tellus, (1997) 49B, 533-555. 

26. Schroeder, W. and Munthe, J. Atm. Envir. (1998) 32, 809-822. 

27. Goodsite, M.E. Plane, J. Skov, H. 2003. A thermodynamic modelling study of the oxidation of 

mercury during polar sunrise, in preparation for Atmospheric Environment, June 2003 

28. DeMore, W.B. Sander, S.P. Golden, D.M. Hampson, R.F. Kurylo, M.J. Howard, C.J. 

Ravinshankara, A.R. Kolb, C.E. Milina, M.J. (1997) Chemical Kinetics and Photochemical 

Data for Use in Stratospheric Modeling. Evaluation 12. NASA panel evaluation. 

29. Horne, D.G. Gosavi, R. and Strausz, O.P. The J. Chem. Phys. Lett. (1968) 48, 4758-4764. 

30. Grieg, G. Gunning, H.E. and Strausz. The J. Chem. Phys Lett. (1970) 52, 3684-3690. 

31. Ariya, P.A. Khalizov, A. and Gidas, A. J. Phys. Chem. (2002) 106, 7310-7320. 

32. Röckmann, T. Brenninkmeijer, C.A.M. Crutzen, P.J. and Platt, U. J. Geophys. Res. (1999) 104, 

1691-1697. 

33. Jobson, B.T. Niki, H. Yokouchi, Y. Bottenheim, F. Hopper, F. and Leaitch, R. J. Geophys. Res. 

(1994) 99, 25,355-25,368. 

34. Sommar, J. Gårdfelt, K. Feng, X. and Lindquist, O. Rate coefficient for gas-phase 

oxidation of elemental mercury by bromine and hydroxyl radicals. Paper presented at 

the 5th International Conference on mercury as aglobal pollutant, Rio de Janeiro, 1999. 

35. Calvert, J.G. and Lindberg, S.E. A modelling study of the mechanism of the halogen-oxygenmercury 

homogeneous reactions in the troposphere during the Arctic spring. Submitted to 

Atmospheric Environment 2003. 

36. Richter, A. Wittrock, F. Eisinger, M. and Burrows, J.P. Geophys. Res. Lett., (1998) 25, 2683- 

2686. 

9

Figures 

Summit 

Fig. 1a Greenland with the location 

of Station Nord. 

Ozone, ppbv 

60 

50 

40 

30 

20 

10 

0 

1999 

Station Nord 

2000 

Fig. 2 Hourly ozone mixing ratios and weekly concentration of fBr measured from 1999 to 2002 at Station 

Nord, Northeast Greenland. GEM is measured in the period from 25 September 1999 to 23 August 2000; 14 

February 2001 to 23 August 2001 and 26 April to 29 June 2002. 

2001 

Time 

50 m 

100 m 

glacier 

Fig. 1b The position of the air monitoring site Flyger’s Hytte at Station Nord. 

Ozone 

GEM 

1/10*fBr 

2002 

2003 

3 

2 

1 

0 

GEM,Br, ng/m 3 

km 

10

GEM, ng/m 3 

Fig 3. GEM against ozone concentrations at Station Nord, Northeast Greenland including a regression line 

obtained by orthogonal regression analysis. Data was selected from Fig. 2 where at least 3 consecutive 

concentrations were decreasing on both ozone and GEM and where the initial GEM concentration was above 

0.4 ng/m 3 . Only data from 2000 and 2001 were used as high concentrations in 2002 indicates the presence of 

other processes than in 2000 and 2001. 

ln([GEM] 0/[GEM] t) 

2 

1.5 

1 

0.5 

-0.5 

4.00 

3.50 

3.00 

2.50 

2.00 

1.50 

1.00 

0.50 

y = 0.039x - 0.095 

R 2 = 0.800 

0 

0 10 20 30 40 50 60 

Ozone, ppbv 

y = 1.437x + 0.006 

R 2 = 0.901 

0.00 

0.00 0.50 1.00 1.50 

ln([ozone] 0/[ozone] t) 

2.00 2.50 3.00 

Fig. 4. The natural logarithm to the relative concentrations of GEM and ozone during Depletion episodes in 

2000 and 2001. The regression analysis is carried out by orthogonal regression analysis. Only censored 

data are included where three consecutive measurements of both ozone and GEM are decreasing and where 

the initial concentration of GEM is larger than 0.4 ng/m 3 . 

11

Table 1. The calculated reaction rate constants for the reactions between GEM and Br and Cl based on Fig. 4 and the reaction rate 

of Cl and Br with ozone (28). The rate constants obtained in this study are calculated using a reaction rate for ozone at 233 K and 

263 to be representative for the conditions in Arctic. 

Reactant 10 

*Correct within a factor 3 

-12 cm 3 molec -1 sec -1 Kelvin Reference 

Br 0.8 233 This Study 

Br 1.2 263 This study 

Br 0.2 ± 0.08 295 34 

Br 0.3* 120-170 30 

Br 3.2 ± 0.4 298 31 

Cl 14 233 This Study 

Cl 16 263 This study 

Cl 15* 120-170 29 

Cl 10 ± 4 298 31 

5 

4.5 

4 

3.5 

3 

2.5 

2 

1.5 

1 

0.5 

0 

01-09-1999 

3 hours lifetime 

10 hours lifetime 

no depletion 

observations 

01-11-1999 

01-01-2000 

01-03-2000 

01-05-2000 

01-07-2000 

01-09-2000 

01-11-2000 

01-01-2001 

Fig- 5 Comparisons between observed (black curve) of GEM and calculated daily means of Hg 0 for three model versions, one 

without depletion (orange), two with depletion, where blue curve is with 3 hours lifetime and red curve with 10 hours lifetime 

during the depletion. 

01-03-2001 

01-05-2001 

01-07-2001 

01-09-2001 

01-11-2001 

01-01-2002 

01-03-2002 

01-05-2002 

01-07-2002 

12

Fig. 6 Comparison of measured surface BrO column from the GOME satellite and the calculated concentrations of RGM for the 

Spring 2000. 

14

ug Hg/m 2 /year 

20 

18 

16 

14 

12 

10 

8 

6 

4 

2 

0 

without depletion 

NOR BAR ALE THU SPI NUU FAE DEN 

TPM(Emitted) 

TPM(chemical) 

RGM 

ug Hg/m 2 /year 

20 

18 

16 

14 

12 

10 

8 

6 

4 

2 

0 

With depletion 

NOR BAR ALE THU SPI NUU FAE DEN 

TPM(Emitted) 

TPM(chemical) 

Fig. 7. Total deposition of Mercury in µg Hg/m 2 /year divided into deposition of Reactive Gaseous Mercury, chemically produced 

particulate mercury and directly emitted particulate mercury for two different model runs: without depletion (left) and with depletion 

(right), and for 8 different localities: Station Nord, Greenland (NOR), Barrow, Alaska (BAR), Alert, Canada (ALE), Thule (THU), 

Greenland, Spitzbergen, Norway (SPI), Nuuk area, Greenland (NUU), Faeroe Island (FAE) and Denmark (DEN) 

Fig. 8 The total annual average deposition of mercury without Arctic mercury depletion (left) and with (right) in µg Hg/m 2 /year 

for the years 1999 and 2000. The total annual deposition north of the polar circle is indicated as well. 

RGM 

15


Paper 4: Goodsite, M.E., Plane, J. Skov, H. 2003. A theoretical study of the oxidation of Hg 0 to HgBr2 

in the troposphere. Submitted to Environmental Science and technology, June 2003.

SUBMITTED TO 

ENVIRONMENTAL SCIENCE AND TECHNOLOGY 

A theoretical study of the oxidation of Hg 0 to HgBr2 in the troposphere 

M. E. Goodsite, 1,3* J. M. C. Plane, 2 and H. Skov 1 

1. National Environmental Research Institute, Department of Atmospheric Environment, 

Roskilde Denmark 

2. University of East Anglia, School of Environmental Sciences, Norwich, United Kingdom 

*Corresponding author, contact information: 

University of Southern Denmark, Department of Chemistry, Odense, Denmark 

Campusvej 55 

DK-5230 Odense M 

Denmark 

Tel. (45) 6550 2557 

Fax. (45) 6615 8780 

Email: meg@chem.sdu.dk 

3. M.E.G. present address: University of Southern Denmark, Department of Chemistry, 

Odense, Denmark 

Submitted June 30 th 2003 

Abstract 

The oxidation of elemental mercury (Hg 0 ) to the divalent gaseous mercury dibromide (HgBr2) has 

been proposed to account for the removal of Hg 0 during depletion events in the springtime Arctic. 

The mechanism of this process is explored in this paper by theoretical calculations of the relevant 

rate coefficients. Rice-Ramsberger-Kassel-Marcus (RRKM) theory, together with ab initio 

quantum calculations where required, are used to estimate the recombination rate coefficients of 

Hg with Br, I and OH; the thermal dissociation rate coefficient of HgBr; and the recombination 

rate coefficients of HgBr with Br, I, OH and O2. It is shown that a mechanism based on the initial 

recombination of Hg with Br, followed by addition of a second radical (Br, I or OH) in 

competition with thermal dissociation of HgBr, is able to account for the observed rate of Hg 0 

removal, both in Arctic depletion events and at lower latitudes. 

1

Introduction 

The perennial oxidation of mercury in the Arctic, which occurs simultaneously with the 

post solar sunrise destruction of ozone (1), potentially doubles the loading of mercury to the Arctic 

(2). These atmospheric mercury depletion episodes, AMDEs, were discovered in 1995 at Alert in 

the Canadian Arctic (1), and have since been observed at circum-Arctic locations, the Antarctic, 

and sub-polar locations near sea ice (3 and citations therein). Tarasick and Bottenheim (4) have 

noted that the frequency of occurrence of boundary-layer ozone depletion episodes has increased 

since the 1960’s, particularly at Resolute in the Canadian Arctic (the only site with a sufficiently 

long record for proper trend analysis). Those authors postulate that this increase could have arisen 

from of an increase in open leads in the Arctic ice cover, possibly because of climate change 

induced by increasing levels of greenhouse gases. Furthermore, the increase in frequency of ozone 

depletion events may explain the increase in mercury levels observed in Arctic biota over the last 

few decades (4). 

The current knowledge of AMDEs is summarised in (3). It is clearly important to 

understand the detailed mechanism of mercury oxidation, so that transport and deposition models 

can be properly parameterised. Mercury exists in the atmosphere primarily in gaseous elemental 

form, Hg 0 , which has an atmospheric residence time of approximately 1 year, allowing it to be 

globally transported (5,6,7). Hg 0 can thus be transported to the polar regions, where it is removed 

from the boundary layer during an MADE with an e-folding lifetime of less than 10 hrs, being 

converted to an inorganic oxidised gaseous mercury compound HgXY (2,3). This compound, 

which is commonly referred to as reactive gaseous mercury, RGM, is operationally determined in 

the Arctic. The common method for measuring RGM is by collection onto a KCl-coated annular 

denuder, followed by the pyrolytic reduction of the captured RGM to Hg 0 (8). This technique 

results in information about the composition of the HgXY family being lost. A number of 

environmental conditions favourable for AMDEs at high latitudes have been identified. These 

2

include: a marine/maritime location; calm weather, low wind speeds, and non-turbulent airflow; 

the existence of a temperature inversion; sunlight; and sub-zero temperatures (9). These 

conditions are also favourable to the photochemically initiated heterogeneous production of 

halogen atoms (Br and Cl) and halogen oxide radicals (BrO and ClO), which are assumed to be 

involved in the mercury oxidative mechanism (9,10,11). BrO is produced in large quantities (> 20 

pptv) after polar sunrise in the marine boundary layer, through the so-called “bromine explosion” 

(4, 12, 13). Gas-phase HOBr reacts with a Br - ion in the saline sea ice surface to yield Br2, which 

is then photolysed. The resulting Br atoms react with O3 to form BrO, which in turn react with 

HO2 radicals to regenerate two HOBr, and the cycle repeats. 

Several mechanisms have been proposed to explain the oxidation of gaseous elemental 

mercury to RGM (10,11). These include the reactions between Hg 0 and halogen oxides or 

halogen atoms to produce HgO, HgBr2 and HgCl2: 

Br (Cl) + O3 → ClO (BrO) + O2 (1) 

BrO (ClO) + Hg → HgO + Br (Cl) (2) 

Hg + Br (Cl) → HgBr (HgCl) →→ HgBr2 (HgCl2) (3) 

and the possible role of oxidants such as OH, HO2, O( 1 D and 3 P), and NO3 that are associated with 

high levels of NO resulting from photodenitrification processes in the snowpack (14). A recent 

review of the atmospheric chemistry of mercury can be found in (15). 

In this paper we will consider the following mechanism for producing HgBrY: 

Hg + Br (+ M) → HgBr (M = third body) (4) 

HgBr (+ M) → Hg + Br (-4) 

HgBr + Y (+ M) → HgBrY (Y = Br, I, OH, O2 etc.) (5) 

3

Ariya et al. (11) have shown recently that reaction 4, the recombination reaction between Hg and 

Br, is surprisingly fast, and so this is our prime candidate to initiate the oxidation of Hg 0 . However, 

the analogous reaction of atomic I may also be important, following observations of active iodine 

oxide chemistry in the mid- and low-latitude marine boundary layer (16, 17). Bauer et al. (18) 

have recently reported that the reaction Hg + OH is very slow, although two previous studies 

obtained rate coefficients for this reaction that varied by 2 orders of magnitude. We will therefore 

investigate the reactions of Hg with both I and OH in this paper. In contrast, the concentration of 

atomic Cl is extremely low (

Theoretical calculations 

Table 1. Optimised geometries and molecular parameters for HgBr, HgI, HgOH, HgBr2, HgBrI, 

HgBrOH and HgBrO2, calculated at the B3LYP/CEP-121G level of theory. Experimental values, 

where available, are given in parentheses 

Species (term 

symbol) 

HgBr( 2 Σ) 

Geometry a Dipole 

r(Hg-Br) = 2.69 

[2.62 f ] 

HgI( 2 Σ) r(Hg-I) = 2.89 

[2.81 g ] 

HgOH( 2 A′) r(Hg-O) = 2.25 

r(O-H) = 0.99 

∠Hg-O-H = 106.8 o 

HgBr2( 1 Σg) r(Hg-Br) = 2.08 

∠Br-Hg-Br = 180 o 

HgBrI( 1 Σ) 

r(Hg-Br) = 2.08 

∠Br-Hg-Br = 180 o 

HgBrOH( 1 A′) r(Hg-Br) = 2.49 

r(Hg-O) = 2.04 

r(O-H) = 0.98 

∠Br-Hg-O = 174.9 o 

∠Hg-O-H = 113.0 o 

∠Br-Hg-O-H = 

180.0 o 

HgBrO2 ( 2 A) r(Hg-Br) = 2.50 

r(Hg-O) = 2.145 

r(O-O) = 1.37 

∠Br-Hg-O = 179.3 o 

∠Hg-O-O = 114.6 o 

∠Br-Hg-O-O = 

162.9 o 

Moment b 

Rotational 

constants c 

Vibrational 

frequencies d 

Bond energy e 

3.10 1.23 [1.30 f ] 146 [187 f ] D0(Hg-Br) = 63.8 

[74.9 f ] 

2.67 0.776 

[0.821 g ] 

1.89 600, 6.24, 

6.18 

106 [125 g ] D0(Hg-I) = 46.3 

[34.7 g ] 

297, 705, 

3536 

0.0 0.514 58, 58, 109, 

269 

0.82 0.376 52, 52, 164, 

250 

1.97 549, 1.07, 

1.07 

1.15 30.3, 0.78, 

0.76 

99, 115, 

235, 555, 

864, 3634 

53, 89, 188, 

250, 455, 

1070 

D0(Hg-OH) = 39.4 

D0(HgBr-Br) = 247 

D0(HgBr-I) = 227 

D0(HgBr-OH) = 

226 

D0(HgBr-O2) = 30 

a Bond lengths in Å; b In Debye (= 3.336 x 10 -30 Cm). c In GHz. d In cm -1 . e In kJ mol -1 . f Ref. 21. 

g Ref. 22 

5

Table 1 lists the binding energies and molecular parameters required to apply RRKM 

theory to reactions 4, -4 and 5. Because there does not appear to be accurate experimental data 

available on the HgXY species, we have calculated these ab initio using the hybrid density 

functional / Hartree-Fock B3LYP method from within the Gaussian 98 suite of programs (23). 

The molecular geometries were first optimised using the Stevens-Basch-Krauss triple-split CEP- 

121G basis set (24). This is a standard basis set for calculations on post-third row atoms: the inner 

electrons on the heavy atoms are treated using effective core potentials, and some relativistic 

effects are included. The resulting geometries, dipole moments, rotational constants and 

vibrational frequencies are listed in Table 1, together with the relevant bond energies. The peroxy 

radical HgBrO2 with doublet spin multiplicity is found to be 30 kJ mol -1 more stable than the 

quartet form. Theoretical calculations on HgBr and HgI are also included in the Table, and 

compared with accurate experimental bond energies, bond lengths and vibrational frequencies. 

The agreement is quite satisfactory, especially for the bond energies and bond lengths, bearing in 

mind the heavy nuclei in these molecules. 

We now apply RRKM theory, using a master equation (ME) formalism (25) that we have 

applied extensively to recombination reactions (26, 27). Briefly, a recombination reaction is 

considered to proceed via the following mechanism (exemplified by reaction 4 with atomic Br): 

Hg + Br → HgBr * (4.1) 

HgBr * → Hg + Br (4.2) 

HgBr * + M → HgBr + M (M = N2) (4.3) 

where HgBr* denotes that the nascent HgBr formed in reaction 4.1 has sufficient internal energy to 

dissociate back to the reactants (reaction 4.2). The energy of the adduct HgBr was first divided 

into a contiguous set of grains (width 30 cm -1 ), each containing a bundle of rovibrational states of 

average energy, Ei. Each grain was then assigned a microcanonical rate coefficient for 

dissociation, k-4,i. The ME describes the evolution with time of the grain populations 

d 

i ( t ) 

dt 

ρ 

∑ 

= Pij 

j ( t ) − i ( t ) − k − 4, i i 

j 

ρ 

ωρ ρ ω 

( t ) + 

R 

i 

(I) 

6

where Ri is the rate of population of HgBr(Ei) via reaction 4.1, ω is the frequency of collisions 

between HgBr * and N2, and Pij is the probability of transfer of HgBr from grain j to grain i on 

collision with N2. The individual Pij were estimated using the exponential down model (28). The 

average energy for downward transitions (i < j), down, was set to be 400 cm -1 for N2 (28), and 

assumed to be independent of temperature. The parameters σ and ε /k, which describe the 

intermolecular potential between HgBr and N2 from which ω is calculated, were set to typical 

values of 4 Å and 400 K, respectively (28). For upward transitions where j > i, Pij was calculated 

by detailed balance. In order to simulate irreversible stabilization of HgBr via reaction 4.3, an 

absorbing boundary was set 24 kJ mol -1 below the energy of the reactants, so that collisional 

energization from the boundary to the threshold was highly improbable. The rate of population of 

grain i, Ri, is given by detailed balance between reactions 4.1 and 4.2: 

Ri = krec,∞ [Hg] [Br] ηi (II) 

where krec,∞ is the limiting high-pressure association rate coefficient (reaction 4.1) and 

η i = 

k − 4 , i f i 

∑ k − 4 , i f i 

i 

(III) 

where fi is the equilibrium Boltzmann distribution of HgBr(Ei). 

The microcanonical rate coefficients for dissociation of HgBr were determined using 

inverse Laplace transformation (25), which links k-1(Ei) directly to krec,∞. In the present case, krec,∞ 

was expressed in the Arrhenius form A ∞ exp(-E ∞ /R T). Assuming that collisions between Hg and 

Br are governed by the long-range attractive dispersion force, then A ∞ = 1.67 x 10 -10 cm 3 molecule - 

1 -1 ∞ -1 

s and E = -423 J mol . 

The microcanonical rate coefficient for dissociation is then given by 

∞ 

( 

i 

3/ 

2 E −E 

−∆H 

∞ o 0. 

5 

− ∆H0 

) − x] 

∞ o 

A 2πµ 

) 

i 

0 

k−4, i = 

N p( 

x)[( 

Ei 

E 

3 ∫ 

− 

N( 

E ) Γ( 

1. 

5) 

h 0 

where the density of states of HgBr at energy Ei, N(Ei), was calculated using a combination of the 

Beyer-Swinehart algorithm for the vibrational modes (including a correction for anharmonicity) 

dx 

(V) 

7

and a classical densities of states treatment for the rotational modes; Np(Ei) is the convoluted 

density of states of Hg and Br; ∆H0 o is the Hg-Br bond energy; and µ is the reduced mass of Hg 

and Br. The ME was expressed in matrix form and then solved to yield k4, the bimolecular 

recombination rate constant at a specified pressure and temperature. The dissociation rate 

coefficient, k5, was calculated by detailed balance with k4. Note that for these calculations of k4 

and k5, the experimental parameters in Table 1 were employed. 

Fig. 1 illustrates the calculated rate coefficients k4, k5 and k6 at a pressure of 1 atm N2 over 

the temperature range 180 – 400 K. This shows that k4 and k6 have small negative temperature 

dependences, as expected for recombination reactions. In contrast, k5 has a large positive 

activation energy, approximately equal to the Hg-Br bond energy. The rate coefficients at 1 atm 

pressure are: 

k4(Hg + Br → HgBr, 180 – 400 K) = 1.1 x 10 -12 (T / 298 K) -2.37 cm 3 molecule -1 s -1 

k5(HgBr → Hg + Br, 180 – 400 K) = 1.2 x 10 10 exp(-8357 / T) s -1 

k6(HgBr + Br → HgBr2, 180 – 400 K) = 2.5 x 10 -10 (T / 298 K) -0.57 cm 3 molecule -1 s -1 

Reaction 6 is close to the high-pressure limit at 1 atm. Inspection of Table 1 shows that atomic I 

and OH bond only slightly less strongly to HgBr, so the rate coefficients for these reactions are 

very similar to k6, essentially at their high-pressure limits. Note that the products HgBr2, HgBrI 

and HgBrOH (Table 1) are extremely stable against thermal dissociation at temperatures below 

400 K. 

For the recombination reactions of Hg with I and OH, and the dissociation of HgI and 

HgOH, application of RRKM theory using the data in Table 1 yields (pressure = 1 atm N2): 

k(Hg + I → HgI, 180 – 400 K) = 4.0 x 10 -13 (T / 298 K) -2.38 cm 3 molecule -1 s -1 

k(Hg + OH → HgOH, 180 – 400 K) = 3.2 x 10 -13 (T /298 K) -3.06 cm 3 molecule -1 s -1 

k(HgI → Hg + I, 180 – 400 K) = 3.0 x 10 9 exp(-3742 / T) s -1 

k(HgOH → Hg + OH, 180 – 400 K) = 2.7 x 10 9 exp(-4061 / T) s -1 

The temperature dependences of these four reactions are also illustrated in Figure 1. 

8

Discussion 

Figure 1 demonstrates several important points with respect to the oxidation of Hg. First, 

the recombination of Hg with Br is surprisingly fast for an atom-atom recombination. The reason 

is the high density of rovibrational states arising from the low vibrational frequency and small 

rotational constant of HgBr (Table 1). Interestingly, the theoretical estimate of k4 is about a factor 

of 3 lower than the recent experimental measurement (11). In fact, we can only match the 

experimental value if the bond energy of HgBr is increased to over 100 kJ mol -1 , about 30 kJ mol -1 

higher than the current experimental measurement of 74.9 ± 4 kJ mol -1 (22, 29). However, the 

recent experimental estimate of k4 was a relative rate measurement that required several significant 

correction factors (11). Since the vibrational frequency and rotational constant of HgBr used in the 

present application of RRKM theory are known precisely from laser induced fluorescence 

spectroscopy in a supersonic jet (22), we prefer the theoretical estimate of k4. Another 

experimental measurement of this rate coefficient would clearly be very desirable. In the case of 

reaction 6, the addition of the second bromine to HgBr is predicted to be a very fast reaction, 

proceeding close to the high pressure limit (essentially the collision number) at atmospheric 

pressure. 

The second point that emerges from Figure 1 is that the recombination reactions of Hg with 

I and OH are a factor of 3 to 4 times slower than reaction 4. The principal reason is the smaller 

binding energies of HgI and HgOH. It should be noted that there is a very large discrepancy in the 

literature regarding k(Hg + OH), with estimates ranging from 8.7 x 10 -14 to 1.6 x 10 -11 cm 3 

molecule -1 s -1 at close to 300 K (18). The current theoretical calculations are in good agreement 

with the most recent upper limit of 1.2 x 10 -13 cm 3 molecule -1 s -1 (18). Note, however, that the 

lifetime of HgOH is only 280 µs at 298 K, so that a true kinetic measurement of the recombination 

reaction would be difficult to achieve in practice. 

9

The third point demonstrated in Figure 1 is that the thermal dissociation of HgBr is more 

than 10 6 times slower than the thermal dissociation of HgI or HgOH, at temperatures below 300 K. 

This enormous difference arises from the stronger Hg-Br bond. The dissociation lifetimes of HgI 

and HgOH are less than 1 s at temperatures above 200 K. Hence, these species will not play a 

significant role in Hg 0 removal. 

We therefore conclude that Hg 0 is oxidised to Hg II by recombination with Br. There is then 

a competition between further addition of Br to form HgBr2, or thermal decomposition of HgBr. 

The addition of I to HgBr may also be significant in some marine locations; however, the OH 

concentration in the clean marine boundary layer (typically less than 10 6 cm -3 ), is probably too low 

for OH addition to HgBr to be significant. As shown in Table 1, the addition of O2 to form 

HgBrO2 will not be an important process, because this peroxy radical is so weakly bound that it 

will dissociate rapidly even at Arctic temperatures. 

The lifetime of Hg 0 , against conversion to HgBr2, is then given by: 

k5 + k6[ Br] 

τ = 

2 

kk[ Br] 

4 6 

Figure 2 illustrates τ as a function of [Br] and temperature. During springtime in the Arctic, the 

temperature ranges from about 230 to 260 K. During Hg depletion events [Br] is estimated to vary 

from 0.2 ppt, when τ will range from 35 to 60 hours, to 6 ppt, when τ will be only 0.7 to 1.5 hours 

(30). A typically observed 10 hr lifetime of Hg (2,10) would correspond to [Br] = 0.7 ppt at an 

“average” temperature of 245 K. 

At temperatures above 280 K, k2 becomes very fast and so τ increases significantly. In the 

mid-latitude marine boundary layer, where the concentration of BrO has recently been measured to 

be around 2 ppt during daytime [A. Saiz-Lopez and J. M. C. Plane, University of East Anglia, pers. 

comm.], the atomic Br concentration under photochemical steady-state will be ≤ 0.1 ppt. This is 

(VI) 

10

similar to the atomic I (16) and OH concentrations, so that these radicals may also play a role at 

mid-latitudes, in contrast to the Arctic. Nevertheless, Figure 2 shows that under these conditions τ 

increases to > 4000 hours. This is in sensible accord with the observed global lifetime of more 

than 1 year (5,6,7), bearing in mind that HgBr has a rich UV/visible spectroscopy (31) and may 

therefore have a significant photodissociation rate in the troposphere, which would extend τ even 

more. Note that if the measured value of k4 (11) is used in place of the present theoretical estimate 

(and k5 is estimated by detailed balance to maintain the equilibrium between HgBr and Hg + Br), 

then the lifetimes calculated by equation (VI) are little changed. 

In conclusion, we have shown that a mechanism based on the initial recombination of Hg 

with Br, followed by addition of a second radical in competition with thermal dissociation, is able 

to account for the observed rate of Hg 0 removal, both in Arctic depletion events and on a global 

scale. 


The Authors wish to thank the Danish Cooperation for Environment in the Arctic (DANCEA), the 

Danish Research Agency (SNF), and the U.K. National Environmental Research Council (NERC). 

M.G. is supported by a COGCI graduate research studentship from the Danish Research Agency 

and the Department of Atmospheric Environment, NERI-DK. 

Literature Cited 

1. Schroeder, W. H.; Anlauf, K. G.; Barrie, L. A.; Lu, J. Y.; Steffen, A.; Schneeberger, D. R.; 

Berg, T. Nature 1998, 394, 331-332. 

2. Skov, H., Christensen, J., Goodsite, M.E., Heidam, N.Z., Jensen, B., Wåhlin, P., and 

Geernaert, G. Environ Sci. Technol. (submitted for publication). 

3. Schroeder, W.H., Steffen, A., Scott, K., Bender, T., Prestbo, E., Ebinghaus, R., Lu, J.Y., 

Lindberg, S.E. Atmos. Environ. 2003, 37, 2551-2555. 

4. Tarasick, D.W., and Bottenheim, J.W. Atmos. Chem. Phys. 2002, 2, 197–205. 

5. Slemr, F.; Schuster, G.; Seiler, W. J. Atmos. Chem. 1985, 3, 407-434. 

6. Schroeder, W. H.; Jackson, R. A. Chemosphere 1987, 16, 183-199. 

11

7. Lamborg, C. H.; Fitzgerald, W.; O'Donnell, J.; Torgersen, T. Geochim. Cosmochim. Acta 

2002, 66 (7), 1105-1118. 

8. Landis, M. S.; Stevens, R. K.; Schaedlich, F.; Prestbo, E. M. Environ. Sci. Technol. 2002, 

36, 3000-3009. 

9. Lu, J. Y.; Schroeder, W. H.; Barrie, L. A.; Steffen, A.; Welch, H. E.; Martin, K.; Lockhart, 

W. L.; Hunt, R. V.; Boila, G.; Richter, A. Geophys. Res. Lett. 2001, 28, 3219-3222. 

10. Lindberg, S. E.; Brooks, S.; Lin, C.-J.; Scott, K. J.; Landis, M. S.; Stevens, R. K.; Goodsite, 

M.; Richter A. Environ. Sci. Technol. 2002, 36, 1245-1256. 

11. Ariya, P.A., Khalizov, A., Gidas, A. Jour. Phys. Chem. A. 2002, 106, 7310-7320. 

12. Barrie, L.A., Platt, U. Tellus. 1997, 49 B, 450-454. 

13. Foster, K.L., Plastridge, R.A., Bottenheim, J.W., Shepson, P.B., Finlayson-Pitts, B.J., 

Spicer, C.W. Science. 2001, 291, 471-474. 

14. Temme, C., Einax, J.W., Ebinghaus, R. Environ. Sci. Technol, 2003, 37 (1), 22 -31. 

15. Lin, C.-J., Pehkonen, S.O. Atmos. Environ. 1999, 33, 2067-2079 

16. McFiggans, G., Plane, J. M. C., Allan, B. J., Carpenter, L. J., Coe, H., O'Dowd, C. J. 

Geophys. Res., 2000, 105, 14371-14385. 

17. Allan, B. J., Plane, J. M. C., McFiggans, G. Geophys. Res. Lett., 2001, 28, 1945-1948. 

18. Bauer, D., D’Ottone, L., Campuzaon-Jost, P., Hynes, A. J. J. Photochem. Photobiology A – 

chemistry, 2003, 157, 247-256. 

19. McFiggans G., Cox, R. A., Mossinger, J. C., Allan, B. J., Plane J. M. C., J. Geophys. Res.- 

Atmospheres , 2002, 107, 4271-4280. 

20. Steinfeld, J. I, Francisco, J. S., Hase, W. L. Chemical Kinetics and Dynamics, Prentice 

Hall, Hew Jersey, 1989. 

21. Lipson, R. H., Jordan, K. J., Bascal, H. A., J. Chem. Phys., 1992, 98, 959-967. 

22. Jordan, K .J., Bascal, H. A., Lipson, R. H., Melchior, M., J. Molec. Spectrosc., 1993, 159, 

144-155. 

23. Frisch, M. J., Trucks, G. W., Schlegel, H. B., Scuseria, G. E., Robb, M. A., Cheeseman, J. 

R., Zakrzewski, V. G., Montgomery, J. A. Jr., Stratmann, R. E., Burant, J. C., Dapprich, S., 

Millam, J. M., Daniels, A. D., Kudin, K. N., Strain, M. C., Farkas, O., Tomasi, J., Barone, 

V., Cossi, M., Cammi, R., Mennucci, B., Pomelli, C., Adamo, C., Clifford, S., Ochterski, J., 

Petersson, G. A., Ayala, P. Y., Cui, Q., Morokuma, K., Malick, D. K., Rabuck, A. D., 

Raghavachari, K., Foresman, J. B., Cioslowski, J., Ortiz, J. V., Baboul, A. G., Stefanov, B. 

B., Liu, G., Liashenko, A., Piskorz, P., Komaromi, I., Gomperts, R., Martin, R. L., Fox, D. 

J., Keith, T., Al-Laham, M. A., Peng, C. Y., Nanayakkara, A., Gonzalez, C., Challacombe, 

M., Gill, P. M. W., Johnson, B., Chen, W., Wong, M. W., Andres, J. L., Gonzalez, C., 

Head-Gordon, M., Replogle, E. S., Pople, J. A. Gaussian 98, Revision A.7, Gaussian, Inc., 

Pittsburgh PA, 1998. 

24. Cundari, T. R., Stevens, W. J., J. Chem. Phys. 1993, 98, 5555-5565. 

12

25. De Avillez Pereira, R., Baulch, D. L., Pilling, M. J., Robertson, S. H., Zeng, G. J. Phys. 

Chem. 1997, 101, 9681-9690. 

26. Rollason, R. J., Plane, J. M. C. Phys. Chem. Chem. Phys., 2000, 2, 2335-2343. 

27. Self D. E., Plane, J. M. C. Phys. Chem. Chem. Phys. 2003, 5, 1407-1418. 

28. Gilbert, R. G., Smith, S. C. Theory of Unimolecular and Recombination Reactions, 

Blackwell, Oxford, 1990. 

29. Handbook of Physics and Chemistry, ed. Lide, D.R., CRC Press, Boca Raton, 78th. edn., 

1997. 

30. Tuckermann, M. Ackermann, R. Gölz, C. Lorenzen-Schmidt, H. Senne, T. Stutz, J, Trost, 

B. Unold, W. and Platt, U. Tellus, (1997) 49B, 533-555. 

31. NIST Chemistry Webbo ok, http://webbook.nist.gov/. 

13

Figure 1. 

k rec / cm 3 molecule -1 s -1 

10 -9 

10 -10 

10 -11 

10 -12 

10 -13 

200 250 300 350 400 

HgBr + Br recombination 

HgOH dissociation 

Hg + I recombination 

Hg + OH recombination 

HgI dissociation 

HgBr dissociation 

Hg + Br recombination 

200 250 300 350 400 

T / K 

Figure 1. Rate coefficients calculated using RRKM theory, plotted as a function of temperature for 

10 6 

10 4 

10 2 

10 0 

10 -2 

10 -4 

10 -6 

10 -8 

10 -10 

10 -12 

10 -14 

the recombination of Hg with Br, I and OH, and of HgBr with Br (solid lines, left-hand ordinate); 

and for the thermal dissociation of HgBr, HgI and HgOH (broken lines, right-hand ordinate). 

k diss / s -1 

14

Figure 2. 

Temperature / K 

300 

280 

260 

240 

220 

200 

180 

10000 9000 8000 7000 6000 5000 4000 3000 

2000 

1000 900 800 700 600 500 

400 

300 

200 

100 

90 

80 

706050 

40 

0.1 1 10 

300 

80 

706050 

40 

30 

20 

100 

90 

30 

200 

20 

80 70605040 

30 

109 

8 7 6 

5 

109 

8 7 6 

4 

20 

1 

0.9 

0.8 

1 

0.9 

0.8 

280 

260 

240 

220 

200 

180 

0.1 1 10 

[Br] / ppt 

Figure 2. Contour plot of the lifetime in hours for Hg 0 oxidation to HgBr2, plotted as a function of 

[Br] and temperature. 

3 

5 

2 

4 

3 

0.7 

2 

0.6 

0.5 

0.4 

0.7 

0.3 

15


Paper 5: Skov, H., Nielsdóttir, M.C., Goodsite, M.E., Christensen, J., Skjøth, C.A., Geernaert, G., 

Hertel, O. Olsen, J., Measurements and Modelling of gaseous elemental mercury (GEM) on 

the Faroe Islands; a case study of the difficulties of measuring GEM. Accepted Asian 

Chemistry Letters (2003).

Asian Chemistry Letters vol. XX, No 1 (2003) XX-XX 

Measurements and Modelling of gaseous elemental mercury (GEM) on the 

Faroe Islands; a case study of the difficulties of measuring GEM 

H.Skov 1,# , M.C.Nielsdóttir 2 , M.E.Goodsite 3 , J.Christensen 1 , C.A.Skjøth 1 , G.Geernaert 4 , O. Hertel 1 , 

J.Olsen 5 . 

1 National Environmental Research Institute. Frederiksborgvej 399, DK 4000 Roskilde, Denmark 

# Corresponding author, e-mail Address: Henrik.Skov@DMU.DK 

2 Present address: University of East Anglia, Norwich England 

3 Present address: Department of Chemistry, University of Southern Denmark, Campusvej 55, DK 5230 Odense 

M, Denmark 

4 Present address: Inst. of Geophysics and Planetary Physics, Los Alamos National Laboratory, MS C-305 

Los Alamos, NM 87545, USA 

5 Food and Environmental Agency, Faroe Islands 

___________________________________________________________________________________________ 

Gaseous Elemental Mercury (GEM) in the atmosphere was measured on the Faroe Islands from May 2000 to 

March 2001. The measured data were analysed together with basic meteorology, trajectories, and modelled GEM 

concentrations using the Danish Eulerian Hemispheric Model (DEHM). The measured air concentration time 

series shows periods with elevated (>1.5 ng/m 3 of Hg, the generally accepted global background average) 

mercury concentrations. We determined that there are two potential natural causes for the higher than expected 

levels: local sources and/or long-range transport. After a detailed analysis, it was determined that neither local nor 

long range sources were sufficiently responsible for the observed levels and the pattern could not be adequately 

reproduced by the DEHM. Measurement artefacts are the most plausible explanation of much of the measurement 

pattern. The nature of the artefact is discussed and recommendation for future measurements of GEM is given. 

___________________________________________________________________________________________ 

This paper is dedicated to Professor Thorvald Pedersen who has just retired from his position at the 

Institute of Chemistry, University of Copenhagen. Through the years we have benefited from his 

genuine interest and knowledge in spectroscopy and atmospheric chemistry. 

1.Introduction 

Mercury on the Faroe Islands is of both scientific and public concern due to the high concentrations in 

e.g. pilot whales and other higher predators of fish, where up to 3 ppm (µg/g) Hg has been measured 1 . 

Furthermore, it has been shown that the present levels of mercury in sea animals also have a negative 

effect on the health of the local populations, since these animals are an important food supply 2 . High 

concentrations of mercury (up to 700 ng/g) have also been observed in peat cores taken on the Faroe 

Islands 3 . The profiles suggest that local geological input does not significantly contribute to the 

mercury inventory, so the Hg must be primarily supplied by atmospheric deposition, and that the level 

of Hg on the Faroe Islands has through time, been higher than levels reported for other European sites. 

However the concentrations cannot be directly linked to atmospheric concentrations or deposition, as 

the levels are not only a function of deposition, but also bioaccumulation, the runoff area size and the 

geochemistry of the profile and it will therefore be important to quantify the rate of atmospheric 

accumulation 3 . It has also been discovered that trout from fish farming activities on the Faroe Islands 

also contain unacceptably high levels of mercury 4 . In spite of strong indications of high mercury 

exposure to marine food chains and the possible high exposure in terrestrial ecosystems (as implied by 

the peat core data), there has not been any study reported to date, which explains the sources 

responsible for the high mercury levels and/or how to mitigate the problem. 

1


The aim of this study is to report a recently collected time series of atmospheric concentrations of 

gaseous elemental mercury (GEM) collected during roughly a one year period starting in May 2000, 

and to explain the observations by, e.g., exploring relationships between anthropogenic mercury 

source regions and deposition to the Faroe Islands. The results are compared with model calculations 

using the Danish Eulerian Hemispheric Model (DEHM) including scenario calculations. Furthermore 

trajectory calculations were carried out using the trajectory model developed for the Atmospheric 

Chemistry and Deposition model 5 (ACDEP). The quality of the measurements presented here is 

disturbed due to various artefacts and these artefacts are discussed and recommendations are given for 

modifications that will prevent such artefacts in future measurements. 

2. Experimental 

The measurements were performed with a Tekran Model 2537A Mercury Vapour Analyser. The 

instrument is equipped with an internal permeation source ensuring the stability of the measurements 

through a daily automatic addition of this span gas as well as zero air. GEM in ambient air is adsorbed 

on a gold trap and after sampling the adsorbed mercury is thermally desorbed and detected by Cold 

Vapour Atomic Fluorescence Spectrophotometry. The monitor is equipped with two parallel gold 

traps, so continuous samples were taken with 5 minutes resolution, which for practical reasons (for 

reporting) was averaged to 1 hour mean values. Measurements were carried out based on our 

modified version of the Standard Operating Procedure (SOP) Manual for Total Gaseous Mercury 

Measurements for the Canadian Atmospheric Mercury Measurement Network 6 . We modified this 

SOP based on the equipment and resources that we had available. For example, we did not use heated 

inlet lines, and we did not conduct a daily manual calibration of the machine, relying instead upon on 

the permeation tube as a secondary standard. The estimated uncertainty from manual calibration, 

collection efficiency, repeatability etc. is estimated to 10 % (2 times standard deviation) for values 

above 1 ng/m 3 . However, complications significantly effected the measurements on the Faroe Islands. 

The very high relative humidity (above 95%) may have affected the measurements (Matthew Landis, 

Private communication, 2001, TEKRAN manual p. 2-1) and high sea spray concentrations have in 

previous studies been observed to effect the measurements of GEM 7 . These complications 

significantly increased the uncertainty of the measurements, as will be discussed. 

3. Danish Eulerian Hemispheric Model 

The concentration of GEM on the Faroe Islands, based on northern hemispheric emissions and 

atmospheric transport pathways, was calculated by the Danish Eulerian Hemispheric Model (DEHM), 

a 3 dimensional eulerian model. The model is described in detail elsewhere 8,9,10 . 

In the current version of DEHM, emissions of anthropogenic mercury are based on the new global 

inventory of mercury emissions for 1995 on a 1 o x1 o grid 11 , which includes emissions of Hg 0 , reactive 

gaseous mercury and particulate mercury. There are no representations of re-emissions from land and 

oceans; instead a background concentration of 1.5 ng/m 3 of Hg 0 is used as the initial concentration and 

in the boundary conditions. 

The chemical reaction scheme 12 includes 13 mercury species, 3 in the gas-phase (Hg 0 , HgO and 

HgCl2), 9 species in the aqueous-phase and 1 in particulate phase. 

Two scenario calculations were carried out. In the first scenario, the chemistry described above is 

assumed to be valid over the entire Northern Hemisphere. In the second scenario, an additional fast 

oxidation rate of Hg 0 to HgO is assumed during the polar sunrise in the Arctic, producing a depletion 

2


of atmospheric mercury, see later. Therefore, inside the atmospheric boundary layer over sea ice 

during sunny conditions, it is assumed that there is an additional first order oxidation rate of ¼ hour -1 

converting Hg 0 to HgO. The fast oxidation stops, when surface temperature exceeds -4 o C (the freezing 

point saltwater). Removals of Hg 0 are due to the chemistry and the uptake by cloud water. 

Dry deposition velocities of the reactive gaseous mercury species (in the model assumed to be HgO 

and HgCl2) are based on the resistance method, where the surface resistance similar to HNO3 is used 

based on the flux measurements. Dry deposition velocities for RGM have been measured and reported 

from Barrow and are similar to those for HNO3 13,14 . Wet deposition of reactive and particulate 

mercury is parameterized by using a simple formulation with different in-cloud and below-cloud 

scavenging coefficients (see 9 ). 

4. Site 

The monitor was set up in a residential suburban area on the eastside of Tórshavn, the capital on Faroe 

Islands, located 62 o 01’ N and 6 o 47’ W approximately 400 km north of Scotland, 600 km west of 

Norway and 500 km east of Iceland. Because of this suburban location, we tested the influence of 

possible local sources by examining the time series based on wind direction and concentration. No 

correlation between these quantities was found, during both quiescent periods and during higher than 

background level episodes, thus allowing us to conclude that any local sources were insignificant. 

5. Results 

Measurements from May 2000 to March 2001 of GEM are shown in Figure 1. The data vary between 

a general level of about 2 ng/m 3 in the beginning of the campaign (in May) decreasing to 

approximately 0.5 ng/m 3 in July and August and increasing slightly to approximately in January and 

February of about 1.7 ng/m 3 . During the campaign we measured some values peaking at 3.4 ng/m 3 . 

For example we observed a notable high concentration episode from June 21 to 25, which was 

followed by a period where the concentration levels decreased to values between 0.8 to 1.3 ng/m 3 . The 

daily mean temperature in the periods varied from 5.7 0 C in May and June to 12.1 0 C in August. The 

average relative humidity was always close to 95%. During the rest of the year the concentrations 

slowly increased to a level of 1.7 ng/m 3 in January and February 2001. 

3


GEM, ng/m3 

4.5 

4 

3.5 

3 

2.5 

2 

1.5 

1 

0.5 

0 

14-maj-00 3-jul-00 22-aug-00 11-okt-00 30-nov-00 

Time 

Figure 1 GEM concentrations at Faroe Islands from May 2000 to March 2001. Values are 1 hour averages. 

The model calculations of daily average Hg 0 concentrations together with measured daily average GEM 

concentrations are shown in Figure 2. Two scenario calculations were carried out; one with mercury 

depletion episodes (MDE) in the Arctic and one without. The calculated concentrations without MDE 

are close to a constant level of 1.6 ng/m 3 throughout the period. The results of the calculations with 

MDE show a decrease in April and June from about 1.4 ng/m 3 without MDE and to 1.0 ng/m 3 with 

MDE. 

4


GEM, ng/m 3 

5 

4.5 

4 

3.5 

3 

2.5 

2 

1.5 

1 

0.5 

Modelled values with MDE 

Measured values 

Modelled values without MDE 

0 

01-maj-00 20-jun-00 09-aug-00 28-sep-00 17-nov-00 

Day 

06-jan-01 25-feb-01 16-apr-01 05-jun-01 

Figure 2 Comparison of daily mean GEM concentrations at Faroe Islands measured by a TEKRAN Hg analyser and Hg 0 

obtained by DEHM. Modelled values reflect two scenarios: concentrations with a springtime Arctic Mercury Depletion 

Event (MDE) and another without MDE. 

6. Discussion 

The measurements of GEM on the Faroe Islands are comparable in magnitude with the levels in 1999 

from Mace Head on the west coast of Ireland 7 and with the values from Harwell an inland location in 

southern England in June 1995 to May 1996 15 . 

The average concentrations at the three localities are 1.68 ng/m 3 at Mace Head, 1.7 ng/m 3 at Harwell 

and 1.4 ng/m 3 on the Faroe Islands. The largest average concentration levels at Harwell and at Mace 

Head are practically identical, with the concentrations on Faroe Islands somewhat smaller. It is 

generally believed that the background concentration of mercury in the Northern Hemisphere is about 

1.5 ng/m 3 and fairly constant. The concentrations at Mace Head are fairly constant for westerly winds; 

concentrations are generally elevated when air masses come from Europe. At Harwell the 

measurements are much more affected by local sources. The concentrations on the Faroe Islands vary 

much more abruptly, which normally indicates problems with the instrument, e.g., due to passivation 

of the gold traps. Therefore the central question is if the measurements on Faroe Islands reflect the 

actual ambient concentrations or if they are influenced by sensor interference. 

Ref. 7 mentioned passivation of the gold traps as a measurement problem at Mace Head caused by sea 

spray. The TEKRAN manual notes that low values are generally due to trap passivation since a 

passified gold trap has a less active surface area for capturing Hg. A similar complication could be the 

reason for the relative low values at Faroe Islands. However 15 reported similar low values at the 

inland station where sea spray should be of minor importance. 

5


If the low values observed on the Faroe Islands are real, one could also speculate that they might be 

explained by marine air with significant halogen chemistry (of Br, BrO, Cl, ClO) where GEM is 

converted to reactive gaseous mercury (RGM) in a similar way to what is observed in the high Arctic 

during polar spring 13,16,17,18,19 . However, efforts must be taken to ensure that passivation of sample 

gold cartridges do not occur in the future. Therefore all GEM measurements since Spring 2001 in our 

laboratory have been carried out, by sampling through a soda-lime (variable mixture of sodium 

hydroxide and calcium oxide/hydroxide, Aldrich 26,643-4) trap (recommended by Matthew S. 

Landis, US EPA, private communication, 2001). 

Artefacts might also explain the high peak of 3.4 ng/m 3 observed from 21 to 25 June 2000. 

Condensation of water in the sample line might work as a trap for mercury, which is liberated 

afterwards when the droplets evaporate. The instrument is also sensitive to variation in voltage. At 

slightly lower voltages than 220V used in this study, the gold traps are not fully desorbed and there is 

carry over of Hg, which will lead to build up of Hg within the instrument (“memory effect”). In fact 

there has been a decrease in voltage during periods in Thorshavn over the last years due to the 

expansion in the general economy. The voltage decrease can only be hypothesised here as an artefact 

as precise information about it is not available. 

As an alternate to sampling complications, a possible natural cause of the observed high 

concentrations of up to 3.4 ng/m 3 was examined in more detail. The high levels of GEM observed on 

the Faroe Islands were initially hypothesised to originate from local anthropogenic sources. We 

therefore obtained information from public and private sources by distributing a questionnaire survey. 

All questionnaires were returned answered from the small community of Tórshavn. Three potential 

sources were identified: the dentists, the hospital and waste incinerators. We discovered that the 

dentists send their waste (silver amalgam) to Denmark, where the metal is reused. The hospital 

reported that it has eliminated most mercury containing instruments. The little waste that remains is 

sent to the waste incinerator. From control measurements made by the incinerator managers, the 

concentration of mercury in the flue gas from the waste incinerator is less than 30 µg/m 3 . 

Approximately 20 kton/year waste is combusted yearly. Batteries and chemical waste are sent to 

Denmark. The position of the waste incinerator and its stack height make it highly unlikely that it can 

influence the measured GEM concentrations at our measurement site. Overall, based on our analysis, 

we determined that local point sources emit an insignificant amount of mercury, thus local sources are 

eliminated as a cause of elevated ambient mercury concentrations. As a second potential natural 

reason for the observed high concentrations, we hypothesised that long range transport of GEM from 

the source regions on the British Isles and Europe could potentially be the cause of observed episodes 

of high GEM concentrations. In order to test this hypothesis, the origin of air masses during the 

episode from June 21 to 25 were determined by a trajectory model (from the Atmospheric chemical 

and deposition model, ACDEP 5 ). Four-day back trajectories were calculated with starting points at 

nine grid points 150 km apart of one another with the central point on Faroe islands and the others 

distributed around Faroe Islands, see Figure 3. The nine points were chosen in order to reduce the 

uncertainty of the single trajectory calculation. When all trajectories follow the same path, the central 

trajectory is considered representative for the atmospheric transport. Otherwise the trajectory 

calculation is considered highly uncertain and is omitted from the further analysis. All the trajectories 

arrive from the middle of the Atlantic Ocean bringing in marine background air that we expect 

contains about 1.5 ng/m 3 of GEM. Therefore there is a high possibility that measurement artefacts 

caused the increased levels observed for the period of 21-25 June. However, the transport investigated 

6


with the trajectory model was only conducted 4 days back in time. Therefore there may be transport 

from locations further away than the British Isles. However, deducing long range transport from 

trajectory models is highly uncertain. 

The model calculations using the DEHM, showed a near constant level of GEM throughout the 

period in accordance with the general belief that GEM has a long atmospheric lifetime of about 1 

year 19 and in agreement with the reported observations at Mace Head 7 . The results of calculations 

including MDE gave concentrations 0.2 to 0.4 ng/m 3 lower in April and May than calculations 

without MDE, see Figure 2. This indicates that MDE may be influencing the general Hg level in sub- 

Arctic areas, given the synoptic behaviour of the Arctic weather system. 

Figure 3 Four-day back tractory calculations with starting points at nine grid points 150 km apart of each 

other with the central point on Faroe islands and the other distributed around Faroe Islands. Starting 18:00 

hrs, 23 June, 2000. 

Mercury depletion in the Arctic has been hypothesised to be caused by Cl, ClO and/or Br, BrO 

chemistry within the Arctic area (e.g. 13,18 ). If those halogen species are important sinks also in the 

marine boundary layer, then this will shorten the lifetime of GEM, significantly creating a more 

variable time series of GEM in accordance with our observations. The two very different, i.e., poorly 

correlated, time series’ obtained by the model and by measurements therefore reinforces our belief 

that the Hg chemistry in the atmosphere outside the Arctic is much more complex than previously 

thought. However due to the lack of a protecting soda-lime trap the observations can very well be 

explained by passivation of the gold traps. This suggests that more measurements at the Faroe Islands 

are needed with a soda-lime trap before any definite conclusions can be drawn on the causes of natural 

variability and strength and timing of mercury episodes. 

7


7. Conclusion and future work 

The analysis of the concentrations of GEM obtained on the Faroe Islands shows that the average 

levels are slightly lower than those observed at Harwell and at Mace Head and the concentrations vary 

much more abruptly. In short episodes, high concentrations were observed but neither local sources 

nor long range transport can explain the observations. Model calculations based on DEHM gave a 

constant level of approximately 1.5 ng/m 3 with slightly lower levels in April and May and could not 

support the appearance of high observed levels. Measurement artefacts might again be part of the 

explanation. 

Furthermore, there is also doubt about the low concentrations. If the low atmospheric concentrations 

observed on the Faroe Islands reflect realistic levels, then they indicate that GEM has a shorter 

lifetime within the marine boundary layer than is generally believed. This implies that the air sea 

exchange of mercury outside the Arctic may be much faster than previously believed, substantially 

reducing the estimated lifetime of atmospheric mercury from about 1 year to a much shorter period. 

Based on our results, it is strongly recommended that future measurements must be performed by 

sampling air through a soda lime trap at all sites. The power supply has to be stable e.g. with use of an 

Uninterrupted Power Supply (UPS). From the above discussion it is evident that more atmospheric 

studies on the Faroe Islands are needed in order to definitively answer the question of the atmospheric 

lifetime of GEM and the connection between atmospheric GEM concentrations and the high Hg levels 

measured in peat and in marine mammals on/near the Faroe Islands. Measurements of GEM even 

with sophisticated monitors such as a TEKRAN, are not trivial. Each measurement site has 

characteristics, which will need to be taken into consideration to ensure a campaign that produces 

accurate results and minimise artefacts. 


We wish to thank Bjarne Jensen and Hanne Langberg, NERI, for technical support, and the National 

Historic Museum of the Faroe Islands for lending us their laboratory. Per Løfstrøm is acknowledged 

for his advice concerning meteorology and Niels Zeuthen Heidam for administration of the Danish 

contribution to the Arctic Monitoring and Assessment programme. The Danish Environmental 

Protection Agency financially supported this work with means from the MIKA/DANCEA funds for 

Environmental Support to the Arctic Region. The results and conclusions presented are those of the 

authors alone and do not necessarily reflect the opinions of our employers or funding agencies. 

References 

1. AMAP (1998): AMAP Assessment Report: Arctic Pollution Issues. Arctic Monitoring and Assessment Programme 

(AMAP) P.O. Box 8100 Dep. N-0032 Oslo, Norway. ISBN 82-7655-061-4. 

2. Grandjean, P. Weihe, P. White, R.F and Debes, F (1998) Env. Res. Sec A. 77, 165-172. 

3. Shotyk, W. Goodsite, M.E. Roos-Barraclough, F. Givelet, N. Leroux, G. Weiss, D. Norton, S. Knudsen, K. and Lohse, 

C. Atmospheric Mercury and Lead Accumulation Since 5410 14 C yr BP at Myrarnar, Faroe Islands, Under preparation 

(2003). 

4. Larsen, R.B. and Dam, M. (1999). AMAP phase I report, The Faroe Islands. 

8


5. Hertel, O. Christensen, J. Runge, E.H. Asman, W.A.H. Berkowicz, R. Hovmand, M.F. Hov, Ø. (1995). Atm. Env. 29. 

1267-1290. 

6. Steffen, S. Schroeder, B. (1999). Standard Operating Procedures Manual for Total Gaseous Mercury 

Measurements CANADIAN ATMOSPHERIC MERCURY MEASUREMENT NETWORK 

(CAMNET)VERSION 4.0. Environment Canada Atmospheric Environment Service,4905 Dufferin Street, 

Toronto Ontario M3H 5T4 

7. Ebinghaus, R. Kock, H.-H. and Hempel, M. (2000). Gefahrstoffe – Reinhaltung der luft, Vol 60, No. 5, pp 205-211. 

8. Kämäri, J., P. Joki—Heiskala, J. Christensen, E. Degerman, J. Derome, R. Hoff and A.-M Kähkönen: Acidifying 

Pollutants, Arctic Haze, and Acidifications in the Arctic, Chapter 9 in: AMAP Assesment Report: Arctic Pollution Issues. 

Arctic Monitoring and Assessment Programme (AMAP). S. Wilson, J. Murray and H. Huntington, Ed, 1998. 

9. Christensen, J. (1997). Atm. Env. Vol.31, No.24, pp.4169-4191. 

10. Christensen, J. (1999). An overview of Modelling the Arctic mass budget of metals and sulphur: Emphasis 

on source apportionment of atmospheric burden and deposition. In: Modelling and sources: A workshop on 

Techniques and associated uncertainties in quantifying the origin and long-range transport of contaminants to 

the Arctic. Report and extended abstracts of the workshop, Bergen, 14-16 June 1999. AMAP report 99:4. see 

also http://www.amap.no/ 

11. Pacyna J.M., Pacyna E.G., Steenhuisen F., Wilson S., (2002). Water, Air and Soil Pollution 137, 149-165. 

12. Petersen, G. Munthe, J. Pleijel, K. Bloxam, R. and Vinod Kumar, A. (1998). Atm. Env. 32, 829-843. 

13. Lindberg, S.E. Brooks, S. C-J. Lin, C-J. Scott, K.J. Landis, M.S. Stevens R.K. Goodsite, M and Richter, A., 

(2002). Environmental Science and Technology 36, 1245-1256. 

14.Goodsite, M.E. Brooks, S.B.Lindberg, S.E. Meyers, T.P Skov, H. and Larsen, M.R.B. Measuring reactive 

gaseous mercury flux by relaxed eddy accumulation using KCl coated annular denuders. Under preparation 

(2003). 

15. Lee, D.S., Dollard, G.J, Pepler, S., (1998). Atm. Env. Vol.32 No.5. pp. 855-864.. 

16. Schroeder W. H. Anlauf K.G. Barrie L.A. Lu J.Y. Steffen A. Schneeberger D.R. Berg T. (1998). Nature 394, 

331-332. 

17. Berg, T. Batnicki, J. Munthe, J. Lattila, H. Hrehoruk, J. and Mazur, A. (2001) Atm. Env. 35, 2569-2582. 

18. Skov, H. Goodsite, M.E. Christensen, J. Geernaert, G. Heidam, N.Z. and Jensen, B.Atmospheric Mercury and ozone 

at Station Nord, Northeast Greenland. Under preparation (2003). 

19. Lin C-J. and Pehkonen, S.O., (1999). Atm. Env. 33 pp. 2067-2079. 

9


Paper 6: Goodsite, Michael E.; Rom, Werner; Heinemeier, Jan; Lange, Todd; Ooi, Suat; Appleby, 

Peter G.; Shotyk, William; van der Knaap, W. O.; Lohse, Christian; Hansen, Torben S. Highresolution 

AMS 14C dating of post-bomb peat archives of atmospheric pollutants. 

Radiocarbon (2001), 43(2B), 495-515.

HIGH-RESOLUTION AMS 14 C DATING OF POST-BOMB PEAT ARCHIVES OF 

ATMOSPHERIC POLLUTANTS 

Michael E Goodsite 1,2 � Werner Rom 3 � Jan Heinemeier 3 � Todd Lange 4 � Suat Ooi 4 � Peter G 

Appleby 5 � William Shotyk 6 � W O van der Knaap 7 � Christian Lohse 1 � Torben S Hansen 1 

ABSTRACT. Peat deposits in Greenland and Denmark were investigated to show that high-resolution dating of these 

archives of atmospheric deposition can be provided for the last 50 years by radiocarbon dating using the atmospheric bomb 

pulse. 14 C was determined in macrofossils from sequential one cm slices using accelerator mass spectrometry (AMS). Values 

were calibrated with a general-purpose curve derived from annually averaged atmospheric 14 CO 2 values in the northernmost 

northern hemisphere (NNH, 30°–90°N). We present a thorough review of 14 C bomb-pulse data from the NNH including our 

own measurements made in tree rings and seeds from Arizona as well as other previously published data. We show that our 

general-purpose calibration curve is valid for the whole NNH producing accurate dates within 1–2 years. In consequence, 14 C 

AMS can precisely date individual points in recent peat deposits within the range of the bomb-pulse (from the mid-1950s on). 

Comparing the 14 C AMS results with the customary dating method for recent peat profiles by 210 Pb, we show that the use of 

137 Cs to validate and correct 210 Pb dates proves to be more problematic than previously supposed. 

As a unique example of our technique, we show how this chronometer can be applied to identify temporal changes in Hg concentrations 

from Danish and Greenland peat cores. 

INTRODUCTION 

Recent scientific work has demonstrated the feasibility of using peat sediments as a global atmospheric 

archive for heavy metal and organic contaminants. Thus, peat has been shown to be a reliable 

archive of atmospheric Pb (Shotyk et al. 1998), and there is evidence that Hg is also effectively 

immobile in peat, though the question of how faithful an archive peat is for the volatile element Hg 

is still under investigation (see Benoit et al. 1998). Peat has also yielded a long-term climatic record 

(Cortizas et al. 1999) and has provided a high-resolution record of atmospheric CO 2 content (White 

et al. 1994). A high-resolution time series during the last 50 years is urgently needed for pollutants 

such as Hg to evaluate the effects of emission controls, and to help calibrate atmospheric transport 

models. Such time series are especially needed from the Arctic, as the most significant gap at the 

present time in Arctic contaminant research is the “lack of temporal trend information for most contaminants” 

(Braune et al. 1999). 

Although there have been studies where the 14 C from the atmospheric bomb pulse has been used to 

date the top layers of a peat profile (see Gedyé 1998; Arslanov et al. 1999), typically in peat studies 

the upper layers are dated with radiometric methods, customarily 210 Pb (see Appleby et al. 1997). 

1Environmental Chemistry Research Group, Department of Chemistry, University of Southern Denmark, Odense University, 

Campusvej 55, DK-5230 Odense M, Denmark 

2Present affiliation: National Environmental Research Institute of Denmark, Department of Atmospheric Environment, Frederiksborgvej 

399, P.O. Box 358, DK-4000 Roskilde, Denmark. Email: mgo@dmu.dk. 

3AMS 14C Dating Laboratory, Institute for Physics and Astronomy, Aarhus University, Ny Munkegade, DK-8000 Århus C, 

Denmark 

4NSF-Arizona AMS Facility, Department of Physics, University of Arizona, Physics Building, 1118 East Fourth St, P.O. Box 

210081, Tucson, Arizona, 85721-0081, USA 

5Environmental Radioactivity Research Centre, Department of Mathematical Sciences, University of Liverpool, P.O. Box 

147, Liverpool L69 3BX, England 

6Geological Institute, University of Berne, Baltzerstrasse 1, CH-3012 Berne, Switzerland 

7Institute of Plant Sciences, University of Berne, Altenbergrain 21, CH-3013 Berne, Switzerland 

© 2001 by the Arizona Board of Regents on behalf of the University of Arizona 

RADIOCARBON, Vol 43, Nr 2B, 2001, p 495–515 

Proceedings of the 17th International 14 C Conference, edited by I Carmi and E Boaretto 495

496 M E Goodsite et al. 

In the present study we investigate the feasibility of using the bomb-pulse 14 C content to date peat 

cores from Denmark and Greenland for the period of 1950 to the present. For comparison, the cores 

were also dated using the customary 210 Pb method. 

It is the first time that peat from Greenland is used in a high-resolution contaminant study (Goodsite 

2000). Peat provided the opportunity to obtain a relatively inexpensive terrestrial archive from the 

Arctic. The use of accelerator mass spectrometry (AMS) allows dating of single year growth increments 

in individual plant macrofossils. We have used the dating results to compare the concentration 

profiles of Hg in Denmark and Greenland to the published North American Hg emission records 

(Pirrone et al. 1998). 

The details of the concentration profiles of Hg and other metal contaminants in the peat cores have 

been treated elsewhere (Goodsite 2000; Shotyk et al. 2001). 

METHODS 

Two distinctly geochemically and trophically different peat lands, one in Denmark and one in 

Greenland, were selected for study. In Denmark, we selected the raised bog at Storelung, Staaby, 

Funen, Denmark (55°15.5¢ N, 10°15.5¢ E). This is an ombrotrophic bog, nourished by the atmosphere 

since it is raised above the water table. Such bogs are well established as archives of atmospheric 

deposition. Three cores of predominantly Sphagnum peat were sampled in October 1999 and processed. 

To the best of our knowledge, no one has located an ombrotrophic peat bog in Greenland. Therefore, 

it was decided to find and investigate a suitable minerotrophic (groundwater nourished) fen. Small 

mires (a generic term for unclassified peat lands) were located and sampled in September 1999 on 

the Narsaq peninsula, southern Greenland (Tasiusaq, Narsaq: 61°08.3¢ N, 45°33.7¢ W) (Goodsite 

2000). The mires had Carex peat accumulation ranging from 20 cm to approximately 100 cm deep. 

Since they received at least some of their water supply and nutrients from the mineral groundwater 

table surrounding their landscape, they are classified as fens. 

One-meter long cores (monoliths) of peat spanning approximately three thousand years of deposition 

were taken from each location. As the peat deposits were similarly sampled, and the cores were 

processed in the same way, only the Greenland cores will be described in some detail. Three (15 cm 

× 15 cm by approximately 100 cm) replicate monoliths of peat from each of two sites in Greenland 

(only one site in Denmark) were cored using a Ti Wardenaar peat sampler (Wardenaar 1987). At 

each site the three replicate cores were taken at a distance of approximately 1.5 m from each other. 

Further analysis was carried out on cores from only one site, with cores from the other site being frozen 

and stored. The choice of Greenland site to be analyzed immediately was based on pH profiles 

of the peat pore water measured in the field. The site chosen had a higher acidity in the upper 20 cm 

than the other site, which was near neutral pH throughout. Since pH is a typical indicator of trophic 

status, with ombrotrophic bogs typically having pH of 3 to 4, it was hoped that this upper region 

might prove to be ombrotrophic, though later analyses showed it was not. 

The three cores (labelled A, B, C) from each site were frozen soon thereafter and shipped to the 

Trace Metals Lab, Geological Institute, University of Berne for further processing and analysis. The 

zero point on the depth scale is defined by visual inspection as the point where it appears that the living 

(green material) stops. Hg and metals analyses are as described in Shotyk et al. (2001). 

Core A was sliced into 3 cm slices by hand using a stainless steel knife prior to freezing. Pore water 

was manually squeezed out of the slices, filtered and then stored cool. Portions of the slices were

14 C Dating of Post Bomb Peat Archives 497 

dried overnight at 105 °C in a drying oven and milled in a Ti mill. The milled powder from pieces 

of each centimeter slice was then manually homogenized prior to using the powder for further analysis. 

Lead and 19 other elements were then determined using X-ray fluorescence spectrometry 

(XRFS) at EMMA Analytical, Canada, by Dr Andriy Cheburkin (see Shotyk et al. 2001). 

Core B was cut while frozen into 1 cm slices using a stainless steel band saw, and selected portions 

of the slices were then dried and milled as above. Samples were then analyzed as before using 

XRFS. Powders were selected for stable lead isotope analysis using thermal ionization mass spectrometry, 

based on the Pb concentration profile obtained using XRFS. Plant macrofossils were 

removed from the centers of the slices from Cores B at the Institute of Plant Science, University of 

Berne, where they were cleaned and dried at 60 °C. Within one week from selection they were processed 

at the AMS 14 C Dating Laboratory, University of Aarhus, for 14 C dating with AMS using a 

standard procedure for plant material (washed, acid-base-acid treatment). AMS was run on the samples 

to reproduce the atmospheric bomb pulse curve and to date peaks in the profile. We would like 

to stress that we use well-defined and carefully selected macrofossils for our study, so all the wellknown 

effects of getting too old or highly varying ages from dating peat water, bulk material or 

humic acid, humin and fulvic acid fractions (see Olsson 1986; Shore et al. 1995) do not apply. 

For the construction of a terrestrial bomb-pulse calibration curve, we used two different materials 

from southern Arizona (USA), Douglas fir and cottonseeds, measured at the NSF-Arizona AMS 

Facility. A cross-section of Douglas fir was cut, smoothed with sandpaper and individual rings were 

then sampled. The cottonseeds were harvested in the year they were produced and archived for later 

use. Both sample types received the following acid-base-acid pretreatment: They were soaked in 3N 

HCl overnight to remove inorganic carbon, afterwards they were rinsed to neutral pH with ASTM 

Type I water, soaked in 2% NaOH overnight to remove mobile carbon (i.e. humic or fulvic acids), 

rinsed to neutral pH with Type I water, soaked in 3N HCl to neutralize any remaining NaOH, and 

finally rinsed to neutral pH with Type I water. 

Conversion of all the 14 C ages to calendar ages was performed via a Bayesian calibration program 

(Puchegger et al. 2000) using cubic spline interpolation for the calibration curve. Our calibration 

curve (see solid line in Figure 1) was constructed as follows: For the period before 1956 we used the 

tree-ring data from the INTCAL98 14 C calibration curve (Stuiver et al. 1998). For the period from 

1956 till now we used annually averaged atmospheric 14 CO 2 data for the latitude band 30°–90° N 

provided by I Levin (personal communication): for the period of 1955 to 1959 data compiled by 

Tans (1981), from 1959 to 1984 data from Vermunt from Levin et al. (1985); after 1988 the arithmetic 

mean values from the three northern hemispheric stations Izaña (1985–1997), Jungfraujoch 

(1987–1997) and Alert (1989–1997) were taken. Values for 1998 and 1999 were obtained by extrapolating 

the almost exactly exponential global decrease in 14 CO 2 since 1982 (Levin and Hesshaimer 

2000). 

Core B was also dated with 210 Pb and 137 Cs at the Liverpool University Environmental Radiometric 

Laboratory. Dried and ground samples from the profile were analyzed for 210 Pb, 226 Ra, 137 Cs, and 

241 Am by direct gamma assay using Ortec HPGe GWL series well-type coaxial low background 

intrinsic germanium detectors (Appleby et al. 1986). Corrections were made for the effect of selfabsorption 

of low-energy gamma ray within the sample (Appleby et al. 1992). 210 Pb dates were calculated 

using the CRS (constant rate of supply) dating model (Appleby and Oldfield 1978) and corrected 

to agree with the 137 Cs signal using the methodology described in Appleby (1998). 

Core C remains frozen as an archive at the Trace Metals Lab, Geological Institute, University of 

Berne.


Figure 1 The bomb-pulse curve. a) The annually averaged atmospheric 14 CO 2 curve for the 30°–90°N latitude band provided 

by I Levin (personal communication). b) Atmospheric 14 CO 2 data from Vermunt, Austria (47°N) and Schauinsland, 

Germany (48°N) (Levin et al. 1997) but only averaged for April to August, the growing season of oaks (Rom 2000). c) See 

Table 1. d) Trees from the Mackenzie Delta, Canada, and Mingyin, Yunnan Province, China (Dai and Fan 1986). e) Mainly 

ears (but also grains or green parts) of barley, wheat, rye and oats from the Copenhagen area (Tauber 1967). Note that the 

(annually averaged) values given are only approximate values since the original data had been published as D using a physical 

14 C half-life of 5570 yr and 1958 as the reference year. Regarding the rather high value for 1959 Tauber (1967) mentions 

that in that year the polar front had an extreme northern position at European latitudes during the whole summer. f) 

Tree from Obrigheim, Germany (Levin et al. 1985). g) Tree from Dailing, Heilongjiang Province, China (Dai et al. 1992). 

h) Different sections and different chemical fractions in tree-rings from an oak from Uppsala, Sweden (Olsson and Possnert 

1992). 

Inset: Global input by the atmospheric nuclear weapons’ tests measured as TNT energy equivalent (thin black line) and 

cumulative input using an exponential decay time of 18.70 ± 0.15 yr corresponding to the uptake by the biosphere and the 

oceans (see Levin and Hesshaimer 2000). Values are given in arbitrary units. 

The “History” of the Bomb-Pulse (see Bennett et al. 2000; USNT 2000), a report on the US bomb tests published by the 

US Department of Energy, in a few cases states slightly different dates). 1) May 1952: The first major thermonuclear device 

is successfully tested by the USA (Eniwetok Atoll). Aug 1953: The first major thermonuclear device is successfully tested 

by the USSR (Semipalatinsk). 2) Feb. 1954: The USA test their first thermonuclear device at the Bikini Atoll, the highestyield 

test site of the USA (yield maximum in 1954/6/8). Further relevant test sites with high yields for the USA: 1952 Eniwetok 

Atoll, 1958/62 Johnston Island, 1962 Christmas Islands. 3) Sept. 1957: The USSR test their first thermonuclear 

device at the Novaya Zemlya, the highest-yield test site of the USSR (yield maxima in 1958 and 1961/2). 4) Oct/Nov 1958: 

Due to the “Geneva convention of Experts for the Discontinuance of Nuclear Weapons” the USA and the USSR stop their 

atmospheric tests. There is a general moratorium till France starts its atmospheric tests in Algeria in Feb 1960. 5) 1961– 

62: The USSR and the USA resume testing in September 1961 and April 1962, respectively. 6) Nov/Dec 1962 Due to the 

“Partial Test Ban Treaty” the USA and the USSR stop their atmospheric tests forever. Only underground tests are allowed, 

and only if no radioactive debris is spread beyond territorial limits. 7) Oct 1964: China starts its atmospheric testing series 

at Lob Nor (yield maxima in 1967–70 and 1973/6). 8) Jul 1966: France starts its atmospheric testing series at Muroroa and 

Fangataufa (yield maxima in 1968 and 1970–72, stopped in Sep 1974). 9) Oct 1980: China performs the last atmospheric 

test ever. 10) April 1986: The nuclear power plant at Chernobyl, USSR explodes. 11) Sept 1996: The Comprehensive Test 

Ban Treaty is opened for signature, but so far (Nov 2000) has not been ratified by the USA or India. 

The decline of the bomb 14 C after stopping the atmospheric nuclear weapons tests is mainly driven by uptake into the 

oceans and the biosphere. In addition, emissions of fossil fuel CO 2 , emissions of 14 C by nuclear power plants, and possibly 

nuclear underground tests contribute (Levin and Hesshaimer 2000).

RESULTS AND DISCUSSION 

The Bomb-Pulse Curve for the Northern Hemisphere 

General Considerations 


Figure 1 shows several 14 C records covering the bomb-pulse period, which were obtained from 

either atmospheric 14 CO 2 measurements or from tree-rings and seeds in the northernmost part of the 

northern hemisphere (NNH), i.e. the region north of about 30°N excluding the tropical (Hadley) 

convection cell. Now a crucial question arises in connection with any 14 C calibration curve: is it only 

of local validity or does it apply to the entire NNH. In the following we address this question and 

demonstrate that it is possible to construct a calibration curve, which is valid for the whole NNH 

producing accurate dates within 1–2 years. 

Nydal and Lövseth (1983) investigated atmospheric 14 CO 2 concentration patterns for the period 

1962–1980 showing that these patterns are basically the same for the southern tip of Norway (58°N) 

and Spitsbergen (78°N), and also 14 CO 2 concentrations on the Canary Islands (27–28°N) show similar 

results. Similarly, for the period 1980–1992 Nydal and Gislefoss (1996) find no significant deviation 

between the Nordkapp (71°N) and the Canary Islands. From these data there is no evidence for 

a significant latitude dependence of atmospheric 14 CO 2 concentrations for regions ranging from subtropical 

to subpolar/polar. However, a decrease of several percent in atmospheric 14 CO 2 concentrations 

(around the bomb-pulse maximum) when going to northern hemisphere tropical regions (9– 

15°N) is clearly documented in Nydal and Lövseth (1983). The same effect is also reflected in trees 

from the tropics (see Murphy et al. 1997). 

Dai and Fan (1986) and Dai et al. (1992) compared 14 C concentrations in tree-rings from spruce 

trees grown at different longitudes and latitudes in the northern hemisphere. They see good agreement 

for tree-rings grown at the same latitude, which reflects the rapid zonal mixing of the troposphere 

within about 1 month (e.g. Ehhalt 1999). On the other hand, these authors claim a clearly visible 

latitude dependence of 14 C in trees grown in 1961–1967 at 27°N, 47°N, and 68°N (see Figure 

1). This may be explained by the fact that meridional atmospheric mixing takes several months 

(Ehhalt 1999), and injection of air from the stratosphere, enriched in 14 CO 2 compared to the troposphere, 

is not equally distributed in time and latitude but mainly takes place during spring/early summer 

at mid-latitudes (Levin et al. 1985; Dai and Fan 1986). Subsequently, the injected 14 CO 2 “diffuses” 

north and south. Since trees also mainly grow in the spring/summer season, this together with 

the sufficiently slow meridional mixing (compared to zonal mixing) may lead to a significantly 

higher 14 C concentration taken up into tree-rings relative to the annual atmospheric average, and also 

may generate a latitudinal gradient in atmospheric 14 CO 2 concentrations. However, Olsson and Possnert 

(1992) regard some of the values measured by Dai and Fan (1986) in a white spruce from Canada 

(68°N) (see Figure 1) as unexpectedly high compared to atmospheric measurements from Spitsbergen 

(78°N) and Abisko, Sweden (68°N), and they point out that finer details of the sample 

pretreatment are missing (see also discussion below). They therefore advise to treat these results 

with caution. 

For the rising part of the bomb pulse Tauber (1967) compared numerous 14 C data in both atmospheric 

CO 2 and plants at mid-latitude regions all over the world and found no clear latitude dependence. 

He supposed that the amplitude and extent of a possible latitude effect may depend on annually 

varying meteorological factors (see also caption of Figure 1). Grass samples taken within one 

week in June 1963, the year with the highest input of bomb-produced 14 C, also did not reveal any 

significant latitudinal gradient all over Scandinavia (56–70°N). However, for the same month he


observed a clear difference in grass samples from Greenland relative to Scandinavia of more than 

14% in 14 C level, corresponding to a delay of somewhat more than one month. Furthermore, a clear 

gradient with high values in the south was observed for Greenland. This general difference between 

Scandinavia and Greenland has been explained by the location of the polar front: In summer times 

the polar front is over southern Scandinavia with frequent shift toward north and south, but over the 

Atlantic it is generally far south of Greenland. Therefore stratospheric 14 CO 2 injected at mid-latitudes 

and diffusing toward north and south is distributed differently. 

Note that the production of 14 C by the atmospheric nuclear weapons tests was maximum in 1962 

(see Bennett et al. 2000), whereas the maximum 14 C concentration in tropospheric measurements 

show up in August/September 1963 and in June 1964 for the extratropical and the tropical NH, 

respectively (see Nydal and Lövseth 1983, and Nydal and Lövseth 1996 with slightly corrected and 

also updated data; see also inset in Figure 1). This clearly shows a) the (seasonally dependent) injection 

of 14 CO 2 from the stratosphere, and b) a delay due to the different transport patterns between 

tropics (Hadley cells) and extratropics (Ferrel and polar cells). 

Several decades after the bomb-pulse maximum Olsson (1989) finds a still slightly lower 14 C level 

in plants from subarctic and arctic areas (including Iceland, the Faroe Islands, Greenland or Spitsbergen) 

compared to Sweden (e.g. about 2% in Spitsbergen for 1980). Also atmospheric 14 CO 2 measurements 

in Abisko, Sweden (68°N) and Kapp Linné, Spitsbergen (78°N) during the 1980s show 

consistently lower values for Spitsbergen, however in the range of less than 1.5% (Olsson 1993). 

Another factor influencing the 14 C level is the dilution with fossil fuel derived CO 2: For the abovementioned 

oak tree from Sweden studied in Olsson and Possnert (1992) (see Figure 1) a local industrial 

effect in the range of 2% was claimed. Also for the pine tree from Germany studied in I. Levin 

et al. (1985) (see Figure 1) a 14 C depression of 1.5% by fossil fuel contamination has been ascribed, 

although in this case not as a local but as a general ground level effect. Olsson (1989) states, that 

clean air does not exist any longer, and therefore e.g. “clean air” 14 C values from Germany, which 

were generally lower than in Sweden (about 2–3% for the late 1970s/early 1980s), should reflect a 

higher fossil fuel contamination. However, this difference is gone during the 1980s. 

For the present-day northern hemisphere the maximum of fossil fuel CO 2 emissions at mid-latitudes 

leads to a corresponding minimum in atmospheric 14 C activity. E.g. measurements at the Alpine station 

Jungfraujoch, Switzerland (47°N) from 1993–4 show 14 C values, which are about 1‰ and 

1.5‰ lower than values obtained at the remote stations Alert, Nunavut, Canada (82°N) and Izaña, 

Canary Islands (28°N) (Levin and Hesshaimer 2000). Values in the tropics (Llano del Hato, Venezuela, 

8°N) are even higher (2‰), which may be due to reemission of bomb-pulse 14 C from the 

highly active biosphere, which has a carbon turnover time of about 30 years (Levin and Hesshaimer 

2000). 

To study the influence of the sample selection and pretreatment regarding tree-rings Olsson and Possnert 

(1992) investigated different sections (early wood vs. late wood) and different chemical fractions 

in tree-rings from an oak tree from Uppsala, Sweden (60°N). After removal of the soluble fraction, 

which corresponds roughly to applying the commonly used acid-base-acid method, they find 

no clear evidence for a delay between the atmospheric and tree-ring 14 C values on a timescale of 

weeks. For the years 1962–63, i.e. the steepest rise in the bomb pulse, the difference between earlywood 

and late-wood values is evident (see Figure 1). Olsson and Possnert (1992) do not ascribe this 

to the temporally dependent diffusion of injected stratospheric 14 CO 2 but to nutrients from the 

respective preceding year, which are stored in the roots and are used for the formation of the early 

wood of the following year.


Figure 1 summarizes a large part of the data discussed above. When comparing the curve of 14 C concentrations 

in plants and atmospheric 14 CO 2 averaged over the growing season of oaks (dashed line) 

with the curve of the respective annually averaged atmospheric concentrations (solid line), the two 

curves coincide except for a few years around the bomb-pulse maximum. This difference can be 

ascribed to the great seasonal oscillations caused by stratospheric injection of excess (bomb) 14 C 

leading to regional differences within the extratropical northern hemisphere (NNH). However, these 

differences are insignificant after about 1971 (Nydal and Lövseth 1983) (after about 1978 according 

to Olsson 1993), where basically equilibrium between the stratosphere and the troposphere has been 

reached. (The deviation of the spruce tree from China from atmospheric concentrations between 

1976 and 1982 shown in Figure 1 clearly reflects the Chinese atmospheric atomic bomb tests during 

that time (Dai et al. 1992).) 

Finally we want to emphasize that—even including all the effects discussed above (except for the 

period 1976–82 in China as discussed above)—any of the bomb-pulse records shown in Figure 1 

when used for calibration from the mid 1950s on will generally give the same calibrated age within 

1–2 years. 

A Complete Tree-Ring/Seed Bomb-Pulse Record and an Atmospheric Bomb-Pulse Calibration Curve 

for the Extratropical Northern Hemisphere 

None of the so far published bomb-pulse records from tree-rings shown in Figure 1 covers the whole 

second half of the 20th century. We present a first (although not annual) record from plant material 

covering the whole bomb-peak period till now. The 14 C data were obtained from Douglas fir treerings 

and cottonseeds from southern Arizona (32°N), which are summarized in Table 1 and shown 

in Figure 1. As can be seen from Figure 1, the values closely follow the annually averaged atmospheric 

14 CO 2 concentration curve for the NNH (solid line) provided by I Levin (personal communication). 

It is interesting to note that the data from Arizona do not show significantly higher 14 C 

concentration values for the period around the bomb-pulse maximum (see also the previous section), 

i.e. they do not follow the seasonal variations as expected for plants compared to annually averaged 

atmospheric values. For comparison, we constructed a calibration curve using atmospheric 14 CO 2 

data from Vermunt, Austria and Schauinsland, Germany (Levin et al. 1997) but only averaged for 

April to August, the growing season of oak trees (see Wrobel and Eckstein 1992), to simulate the 

relevant annual portion of 14 CO 2 that is taken up by the plant. Such an “atmospheric” calibration 

curve (Rom 2000, see dashed line in Figure 1) clearly shows the features expected from the stratospheric 

injection of excess 14 C in spring, which is surprisingly well matched by the cereals data from 

Denmark (Tauber 1967; see Figure 1). From this, the Arizona data, when compared to the annually 

averaged atmospheric 14 CO 2 data, may either point towards a latitudinal dependence (see preceding 

section) or rather seem to be unaffected by the seasonal variations due to the injection of stratospheric 

14 CO 2. We checked the gap between 1992 and 1998 in our tree-ring/seed record and measured 

green leaves of an (unspecified) tree collected at the Brorfelde Observatory, Zealand, Denmark 

(56°N) resulting in (110.5 ± 0.5) pMC (AAR-3339). Similar to the extrapolation of the 

atmospheric calibration curve by I. Levin (as described in the Methods section) we interpolated the 

1996 value for our Arizona data using an exponential fit to the data points from 1983–1998. The corresponding 

error was obtained by averaging the variances for the given period. The interpolated 

value of (110.8 ± 0.6) pMC perfectly matches the measured value. 

We finally decided to use the annually averaged atmospheric 14 CO 2 curve for the 30–90°N latitude 

band provided by I Levin (personal communication) (solid line in Figure 1) as our general purpose 

bomb-pulse calibration curve since a) the Arizona data closely follow this curve but do not provide


Table 1 14C determination in plants from Arizona, USA to establish a terrestrial bomb-pulse calibration curvea Lab nr Sample speciesb Specification Sample yr Sample location d 13C (‰) pMCc AA6665 Douglas Fir Tree ring 1955 Santa Catalina Mountains, Arizona, USA - 25.0d 100.2 ± 0.6 

AA6667 Douglas Fir Tree ring 1957 Santa Catalina Mountains, Arizona, USA - 25.0d 108.4 ± 0.6 











AA11846 Cotton Seed 1972 Southern Arizona, USA - 27.0 145.8 ± 0.6 



1973 (weighted average) 144.1 ± 0.4 













AA6690 Douglas Fir Tree ring 1981 Santa Catalina Mountains, Arizona, USA - 25.0d 127.1 ± 0.7

Table 1 14 C determination in plants from Arizona, USA to establish a terrestrial bomb-pulse calibration curve a (Continued) 

Specification Sample yr Sample location d 13 C (‰) pMC c 

Lab nr Sample species b 


AA6692 Douglas Fir Tree ring 1983 Santa Catalina Mountains, - 25.0 

Arizona, USA 

d 124.5 ± 0.7 




Arizona, USA 

d 123.8 ± 0.8 



Arizona, USA 

d 119.6 ± 0.8 


— e 

Tree ring 1998 Arizona, USA 110.0 ± 0.7 

aWhenever there is more than one sample in an individual year, first a weighted average for samples from the same sample material (Douglas fir or 

cottonseeds) was made. These sample-material averages were then combined in a weighted average to give the final result for the respective year. 

Since most calibration programs use conventional 14C ages as input, for the convenience of reader we give the formula to convert pMC values into 

14C age T: 


T – 8033 Ln pMC 

= 

× æ ------------ ö 

è 100 ø 

All the bomb-pulse data will then result in negative ages. We know that there is some reluctance in the 14C community to use negative 14C ages, but 

we think they might be a suited extension of the pre-bomb calibration curve into the bomb and post-bomb period without changing units. Note that 

the conventional positive 14C ages are also not “true” ages but only a quite arbitrary though unambiguous mathematical transformation of the 14C activity of the sample, which is still kept for historical reasons. Furthermore using 1950 as a reference year is not based on physical/natural parameters 

but also just a convention. By using negative 14C ages one would avoid to publish 14C datings before 1950 in yrs BP, whereas they have to be given 

as pMC or 14C afterwards. 

bDouglas Fir (Pseudotsuga menziesii); Cotton (Gossypium species) 

cpMC (percent Modern Carbon ) = ( 14C concentration of the sample)/(0.95 × Oxalic Acid I 14C concentration) × 100. Both 14C concentrations are d 13C corrected and refer to the same year. 

dA d 13C of - 25.0 was assumed for all Douglas Fir samples 

e Private communication by D Donahue


annual resolution, b) the curve provides consistent, carefully checked data covering the whole bomb 

pulse till now, c) most of the data used for constructing this curve are widely used among the 14 C 

community (see Levin et al. 1997 for the 14 CO 2 data from Vermunt, Austria), and d) differences to 

the other extratropical northern hemisphere curves shown in Figure 1 generally result only in calibrated 

age differences of 1–2 years as stated in the previous section. 

For calibration no uncertainties were assumed for our NNH curve. If not otherwise stated, all the calibrated 

14 C ages in the Tables and Figures of this paper are given as 95%-confidence intervals (often 

denoted as “2-s intervals”). 

The Peat Cores from Greenland and Denmark 

Figures 2 and 3 and Table 2 show our results on plant material from the two peat cores Storelungmose 

in Denmark and Tasiusaq in Greenland. Both 14 C and modelled 210 Pb results are given. 

14 C Dating 

Regarding the 14 C measurements in the Greenland core (see Figure 2 and Table 2), one can see two 

important features: first, dating of samples from the peat surface gives results that are consistent 

with the year of the sampling, and second, the sample with the maximum 14 C content (AAR-5626) 

of 179.1 ± 0.8 pMC shows rather good agreement with the maximum of the calibration curve of 

184.0 pMC. The peak value in our data from the Greenland core corresponds also well to the D 14 C 

value of (776 ± 8)‰, i.e. (178.2 ± 0.8) pMC, measured in a grass sample from Narsaq, the very same 

peninsula where our peat samples come from, in July 1963 (Tauber 1967). Since no significant natural 

enrichment process for 14 C is known, this agreement between the peak value in our Greenland 

core and the atmospheric record cannot be just coincidence but ensures that we have no significant 

dampening for the Greenland peat core. 

This is in contrast to results of Jungner et al. (1995) who measured 14 C in peat hummocks, i.e. raised 

surfaces on the peat land as opposed to “hollows”, from central and eastern Finland (61°N and 62°N, 

respectively) using AMS on well defined stems of Sphagnum fuscum moss. For the peat core from 

central Finland the maximum 14 C value in the cellulose fraction of a moss sample representing a single-year 

fraction shows a D 14 C of (660 ± 13)‰, i.e. (166.4 ± 1.3) pMC, a moss sample from the eastern 

Finland peat core representing a 3–5 yr average yielded a value as low as (564 ± 11)‰, i.e. 

(156.8 ± 1.1) pMC. Jungner et al. (1995) inferred that this effect was due to CO 2 emitted from 

decaying layers below the surface (i.e. older layers). An alternative explanation might be that the 

samples showing maximum 14 C content do not represent the bomb-pulse maximum. 

Regarding the 14 C measurements in the Denmark core (see Figure 2 and Table 2), one also can see 

two important features: First, dating of samples from the peat surface gives results that are consistent 

with the year of the sampling, which is similar to the Greenland core, and second, the sample with 

the maximum 14 C content (AAR-5614) of 152.7 ± 0.8 pMC clearly does not agree with the maximum 

of the calibration curve of 184.0 pMC. More samples between 2.5 and 8.5 cm depth have to be 

dated in the near future to find out whether we so far just missed the peak of the bomb-pulse or 

whether there is significant dampening in the Denmark core, which we cannot exclude so far. 

Preliminary results of additional determinations are now available for Denmark and Greenland. A 

macrofossil from 9.5 cm from the Danish core (AAR-6860) has a pMC of 176.8 ± 0.7, which 

approaches the expected maximum and additional samples above and below the maximum support 

a well developed curve. Therefore we originally just missed the peak of the bomb-pulse and no significant 

dampening is seen in the Danish core.

Table 2a Sample specifications, 13C values and 14C AMS datings of carefully selected and processed macrofossils from the peat cores from 

the Storelungmose, Denmark, and Tasiusaq, Greenland—Macrofossils from the Storelung Mose, Ståby, Fyn, Denmark (55°15.384¢ N, 

10°15.336¢ E) 

Calibrated age ranges (yr AD) 

(95% confidence intervals) 

pMC b 

d 13C (‰) a 

Depth 

(cm) 

Lab nr 

(AAR–) Sample species Specification 


5611 Andromeda sp. Fresh leaves 0c - 27.5 111.36 ± 0.54 1957 

1995–1999 

5612 Leucobryum sp. Branches and leaves 000.5 - 26.1 111.31 ± 0.61 1957 

1995–1999 


1989–1992 


1970–1971 

6612 Sphagnum sp. Leaves 010.5 - 23.0 136.99 ± 0.74 1962 

1975–1976 

6613 Sphagnum sp. Leaves 011.5 - 24.0 123.68 ± 0.65 1959–1961 

1982–1984 

6614 Sphagnum sp. Leaves 012.5 - 23.6 127.09 ± 0.67 1962 

1980–1981 

6615 Sphagnum sp. Leaves and stems 013.5 - 24.3 109.74 ± 0.65 1957 

1995–1999 

5615 Leucobryum sp. Branches and leaves 014.5 - 24.7 120.19 ± 0.55 1958; 1960 

1984–1987 

5616 Leucobryum sp. Branches and leaves 015.5 - 25.7 120.58 ± 0.56 1958–1961 

1984–1987 

5617 Cyperaceae Leaves 016.5 - 27.0 100.12 ± 0.53 1693–1726; 1813–1850 

1862–1918; 1951–1956 

5618 Sphagnum sp. Branches and leaves 018.5 - 24.2 99.43 ± 0.49 1687–1729; 1810–1923 

1950–1955 

5619 Sphagnum sp. Branches (?) and leaves 078.5 - 24.5 68.39 ± 0.38 1425–1424 BC; 1412–1211 BC 

1201–1192 BC; 1179–1163 BC 

1141–1132 BC 

ad 13C values have been measured by Árny E Sveimsbjörnsdóttir, Science Institute, the University of Iceland 

bSee Table 1 for definition 

cThese samples were taken from slightly above ground and correspond to the Hg concentrations values at negative depths shown in Figure 4


Table 2b Sample specifications, 13C values and 14C AMS datings of carefully selected and processed macrofossils from the peat cores from 

the Storelungmose, Denmark, and Tasiusaq, Greenland—Macrofossils from Tasiusaq, Narssaq, Greenland (61°08.314¢ N, 45°33.703¢ E). 

Calibrated age ranges (yr AD) 

(95% confidence intervals) 

pMC b 

d 13 C 

(‰) a 

Depth 

(cm) 

Lab nr 

(AAR–) Sample species Specification 

5620 Sphagnum sp. Branches and leaves 00c - 26.2 110.99 ± 0.57 1957 

1996–1999 

5621 Sphagnum sp. Branches and leaves 0 - 28.3 111.88 ± 0.62 1957 

1994–1999 

5622 Sphagnum sp. Branches and leaves 00.5 - 26.8 114.34 ± 0.57 1957–1958 

1991–1994 

5623 Sphagnum sp. Branches and leaves 05.5 - 26.3 140.33 ± 0.61 1962 

1973–1974 


1973 


1967 



1980–1981 


5629 Dwarf bush Twigs 018.5 - 27.1 102.52 ± 0.54 1956 

5630 Vascular plant — 087.5 - 26.7 069.52 ± 0.44 1292–1280 BC 

1263–973 BC 

959-941 BC 

ad 13C values have been measured by Árny E Sveimsbjörnsdóttir, Science Institute, the University of Iceland 

bSee Table 1 for definition 

cThese samples were taken from slightly above ground and correspond to the Hg concentrations values at negative depths shown in Figure 4


Figure 2 14 C AMS dating results for the Storelungmose (Denmark) and the Tasiusaq (Greenland) peat cores. 

The solid line represents the annually averaged atmospheric 14 CO 2 curve for the northernmost northern hemisphere 

provided by I Levin, Heidelberg. This curve was used for calibrating all our data from the peat cores. 

Symbols denote the maximum-probability age for the respective sample. Dotted symbols mark the two possible 

solutions for a sample (AAR-5614, see Table 2a) for which it was impossible to decide on one of these 

solutions from stratigraphic evidence. Horizontal bars denote 95%-confidence intervals, coherent portions in 

agreement with stratigraphic order are shown in black, portions contradicting stratigraphic order that therefore 

were rejected are shown in gray. The enclosed circles at 176.8 ± 0.7 pMC show in fact that no dampening 

occurred in the Danish bog. 

In the following we just speculate about the possible consequences of the dampening seen in the 

Danish core prior to additional dating. Since we measure living moss samples from the surface to be 

in agreement with atmosphere values this case would then be similar to Jungner et al. (1995), and 

similarly we may conclude that the mean 14 C activity of the CO 2 emitted from decaying sub-surface 

peat layers is not significantly below 100 pMC. Diluting the atmospheric values with about 35% of 

emitted CO 2 at 100 pMC (which gives the maximum measured in the Denmark core) may produce 

a significant shift of up to four years towards older ages regarding the maximum-probability age for 

the samples AAR-5615, 5616 and 6613 (see Table 2), which are centered around the wiggle in atmospheric 

14 CO 2 activity at the end of the 1950s/beginning of 1960s. However, these shifted ages lie 

only one year outside the confidence intervals (marked in black in Figures 2 and 3) of the respective 

samples. The calibrated ages of all the other samples on the rising part of the bomb pulse would be 

much less affected by such a dampening effect. 

Emanation of CO 2 from decomposition of layers close to the surface might lead to an apparent time 

lag for the bomb-pulse peak value in the peat, but in any conceivable scenario pre-peak layers will 

be dated too old, when calibrated with our general-purpose calibration curve, not too young. For 

comparison with 210 Pb results see later.


Going from 11.5 cm to 15.5 cm there is a significant dip in the 14 C content of the respective moss 

plants (see sample AAR-6615 at 13.5 cm, Table 2a). The order of this excursion (more than 10 percent 

relative to natural level) corresponds to measurements in cereals from Denmark for the period 

1959–62 (Tauber 1967; see Figure 1), which also show a pronounced wiggle (about 130 pMC in 

1959, 123 and 121 pMC in 1960 and 1961, respectively, and 137 pMC in 1962). However, the absolute 

values in our peat core are about 10 percent lower. Such a depletion cannot be explained by a 

simple model of admixture of 35% peat-derived CO 2 with an activity of 100 pMC as discussed 

above, but a more elaborate model would be required to cover all the features of a possibly “dampened” 

curve for a core. 

The 14 C data from both the Denmark and the Greenland peat core suggest a significantly different 

accumulation rate between the topmost layers and lower layers (see Figure 3) A linear regression for 

all 14 C data in the Greenland and the Denmark core (omitting only the unresolved two-fold solution 

for AAR-5614, see Table 2) gives an average accumulation rate of 4.3 and 3.7 mm yr - 1 , respectively, 

whereas for the lower layers as shown in Figure 3 one gets 6.9 and 8.2 mm yr - 1 , respectively. So the 

peat layers close to the top seem to comprise more years per cm than the lower layers. This is just 

opposite to what one expects if gravitational compression takes place. One might assume that 

increased decomposition in the peat layers close to the surface compared to the lower layers is 

responsible for this effect. However, dry bulk density (DBD) vs. depth profiles, which show a high 

degree of similarity in both cores and which are also highly correlated with Hg concentration profiles 

down to about 18 cm (see Goodsite 2000), do not consistently support this assumption. To clarify 

this question it will be necessary to measure more samples from different depths close to the surface. 

Since we have taken 1 cm slices from the individual peat cores, it is evident from the accumulationrate 

data given above that on average a single slice contains more than two years. So to get annual 

resolution for the Hg profiles it therefore might be preferable to measure both 14 C and Hg in the 

annual growth increment of the very same moss plant. 

14 C dating of macrofossils in a peat core makes it possible to selectively date objects from any depth 

of the profile in the bomb-pulse period (and possibly also before; see below). (This, of course, is the 

case only if significant dampening can be excluded, which clearly is the case for both of our cores.) 

14 C therefore is able to pick up details of the peat evolution, e.g. changes of the accumulation rate, 

which may serve as an important input for the 210 Pb modelling. Moreover, the immediate need for a 

continuous chronology, i.e. the need to date an entire column with other radiometric means such as 

210 Pb, to get a date for a certain peat layer is eliminated. However, for flux calculations of e.g. Hg in 

the associated layers a continuous chronology is still required. Flux calculations derived from 14 C 

measurements should be more precise though account may need to be taken of possible migration of 

the pollutant relative to the peat matrix. (Regarding a possible different basic trend between 14 C and 

210 Pb—especially in the data for the Denmark core—see the discussion of the 210 Pb data below.) 

For a given 14 C concentration in a sample there is always an (at least) twofold solution in the bombpulse 

period regarding calibrated-age ranges. Therefore it is necessary to measure at least two points 

of a profile from the bomb-peak period, which are close to each other in depth, in order to determine 

which side of the bomb-pulse one’s points are on. Then—by assuming an undisturbed stratigraphic 

order of the peat layers—it is generally possible to discard one of the solutions for each sample. 

Although 14 C dating is commonly regarded as impossible after 1650 (and before the bomb-peak 

era), we think this is not completely true. Especially the period from 1900 to 1950 shows an almost 

perfect monotonic decrease in 14 C. Single-year data from tree-rings from Douglas firs (grown on the


Olympic Peninsula, 47°N) show a decrease from about - 2‰ to - 25‰ (Stuiver and Quay 1981) corresponding 

to a change of about 120 14 C years within 50 cal yr. If therefore another method (such as 

210 Pb dating) may provide evidence that a peat sample stems from the first half of the 20th century, 

high-resolution dating with 14 C can be performed for this period. 

210 Pb Dating 

Figure 3 Age-depth profiles for the Storelungmose (Denmark) and the Tasiusaq (Greenland) peat 

cores. Both 14 C AMS dating results and data from 210 Pb modelling (corrected according to 137 Cs) 

are shown. 14 C results correspond to those shown in Figure 2 (for symbols and bars see caption of 

Figure 2). A linear regression has been applied to the ages of highest probability at 10.5–18.5 cm 

depth in the Denmark core and at 5.5–18.5 cm depth in the Greenland core, showing a high linear 

correlation between age and depth (especially for the Greenland core with a correlation of more 

than 99%), although one has to be aware that in general peat accumulation rates may deviate a lot 

from linearity. Data at the top clearly deviate from this regression lines indicating a slower accumulation 

(see text). 210 Pb values and the respective uncertainties (1-s ) are shown as crosses and 

thin lines. Note the significant deviations in the Denmark core between the 14 C data and the 

respective 210 Pb data. Better agreement is found for the Greenland core, though in both cases there 

are significant discrepancies in the pre-1963 sections. 

210 Pb is a naturally occurring fallout radionuclide that is deposited on the bog surface from the atmosphere 

and incorporated in the bog archive, along with records of atmospherically delivered pollutants 

such as stable Pb and Hg. Concentrations of 210 Pb at different depths in the bog depend on the 

atmospheric flux at the time the layer was at the bog surface, the net peat accumulation rate (original 

growth rate minus subsequent losses by organic decay), and the age of the layer. Although globally 

the atmospheric 210 Pb flux may vary spatially by up to an order of magnitude (Appleby and Oldfield 

1992), its value at a given location is generally considered to be fairly constant, at least on an annually 

averaged basis. This assumption is supported by measurements comparing the contemporary 

flux via rainfall with the long-term flux via soil cores (unpublished data). Because of the effects of 

varying net peat accumulation rates, the unsupported (atmospherically deposited) 210 Pb activity does 

not usually follow a simple exponential reduction with depth, even when plotted against cumulative


dry mass. The most widely used method for calculating dates in cores with non-exponential records 

is the CRS (constant rate of 210 Pb supply) model (Appleby and Oldfield 1978). This involves measuring 

210 Pb at regular intervals down to the depth at which 210 Pb reaches equilibrium with the supporting 

226 Ra (ca. 130 years). The results are presented as a continuous set of dates spanning this 

period. Because of small-scale irregularities over the surface of the bog, the efficiency with which 

fallout 210 Pb is trapped at a given site in the bog may vary through time, causing errors in the CRS 

model dates. In such cases, independently dated horizons, usually based on records of the artificial 

fallout radionuclides 137 Cs and 241 Am from the atmospheric testing of nuclear weapons (peaking in 

1963) or the Chernobyl accident (1986), are used to make corrections to the 210 Pb dates (Appleby 

1998). 

The Danish (Storelungmose) and Greenland (Tasiusaq) peat cores had relatively similar 210 Pb 

records. In both cores, 210 Pb/ 226 Ra equilibrium was reached at depths of between 24–26 cm. The 

unsupported 210 Pb activity-versus-depth profiles were approximately exponential though with a 

shallower gradient in the upper part of the core. The profiles from both cores had small non-monotonic 

features in the top 15 cm. 210 Pb dates calculated using the CRS model indicated episodes of 

rapid peat growth during the past 40 years, though the general trend was one of declining net accumulation 

rates in the older sections reflecting losses from the peat matrix. 

Results for the Denmark and Greenland peat cores revealed significant differences between the 

210 Pb dates and the 1963 depths indicated by the 137 Cs record. In both cores the uncorrected 210 Pb 

dates placed 1963 at a depth of 12.5 cm. The 137 Cs stratigraphy suggested that it was significantly 

deeper in the core, at 14–16 cm in Storelungmose and 15.5 cm in Tasiusaq. (The latter also has a second 

more recent peak at 3.5 cm that may record fallout from the 1986 Chernobyl accident.) Figure 

3 shows the corrected 210 Pb dates for each core using the 1963 137 Cs date as a reference point 

(Appleby 1998). Uncertainties are given as 1-s intervals, corresponding to 68%-confidence intervals. 

Note that shifting the reference point influences basically all the modelled 210 Pb dates. E.g. placing 

the 1963 peak to 12.5 cm instead of 15 cm means that all the dates above this level are also shifted 

in age although the shift gradually decreases when approaching the surface, which is used as another 

reference point. Dates below the 1963 level are all shifted by a constant value (6 yr in the case of the 

Denmark core). 

The 14 C measurements now offer a means for testing the validity of the 210 Pb dating procedure, and 

in particular the use of 137 Cs as an independent time marker to correct the 210 Pb results. Since 137 Cs 

fallout from atmospheric weapons tests peaked in 1963, the same year as the NNH atmospheric 14 C 

concentration reached its maximum value, it might be expected that bog records of these two radionuclides 

should have peaks at similar depths. This, however, is not the case. The 14 C peaks occur at 

9.5 cm in Storelungmose, and at 12.5 cm in Tasiusaq. The uncorrected 210 Pb dates are thus in better 

agreement with the 14 C dates than those corrected by 137 Cs. 

Although pore-water diffusion of 137 Cs in peat and sediment cores is well documented, this is normally 

assumed to mainly affect the tail of the 137 Cs profile. Until now there has been very little direct 

evidence for a significant displacement of the peak. Two possible causes of this are an initial advective 

displacement through partially saturated surface vegetation at the time of deposition on the bog 

surface, or downward diffusion and preferential adsorption onto a layer containing higher concentrations 

of clay minerals. The apparently larger displacement in the Storelungmose core may be due 

to its lower mineral content, evidenced by very low 226 Ra concentrations (


226 Ra concentrations in the Tasiusaq core (mean value 21 Bq kg - 1 ) suggest a significant minerogenic 

input, possibly as wind-blown dust. 

A further contributory factor that cannot be discounted is distortion of the 137 Cs record by changing 

peat accumulation rates. Since this also affects 210 Pb, the local maximum value of the 137 Cs/ 210 Pb 

activity ratio may be a better indicator of the depth of the 1963 fallout maximum than the 137 Cs peak 

itself. In both cores the 137 Cs/ 210 Pb ratio has a maximum value at 12.5 cm. Regardless, the results 

presented here show that the use of 137 Cs records to validate and correct 210 Pb dates is more problematic 

than previously supposed. Since 210 Pb records, and also those of other trace metal pollutants, 

might be similarly affected, the interpretation of pollutant records in peat bogs, and in particular 

their relationship to the peat matrix, is an issue that needs to be addressed in greater detail. Bomb 

14 C, which offers an accurate means for dating the matrix itself, will be an invaluable tool for inves- 

tigating this relationship. 

One problem that needs further investigation is the decrepancy between 14 C and 210 Pb in pre-1963 

sections of the core (see Figure 3) where 14 C dates get progressively younger than 210 Pb dates (corrected 

or uncorrected). This cannot be explained either by downward migration of 210 Pb (since this 

would produce younger 210 Pb dates) or by emanation of CO 2 from decay of sub-surface layers (see 

Jungner et al. 1995). As mentioned above, this would lead to a shift towards higher ages. 

Mercury 

Figure 4 shows an application of the 14 C bomb-pulse dating method to Hg concentration profiles 

measured in the peat cores from Denmark and Greenland. The chronology of concentration changes 

at the two sites is similar. The similarity of the two curves from two geochemically different mires 

and different climate regimes is consistent with current views, which suggest global transport of Hg. 

Our profiles also show a fair correlation with Hg emissions in the northern hemisphere (Pirrone et 

al. 1998), and this similarity also suggests that peat may be a suitable archive for recording atmospheric 

Hg, though no conclusions can be drawn about how faithfully the archive preserves Hg concentrations. 

E.g. the peaks observed in the mid 1960s and the late 1950s for the Greenland core 

appear to correspond with maxima observed in the ice record (Boutron et al. 1998), though concentrations 

in the peat are approximately 10,000 times higher. Declines in North American emissions 

after 1989 have been reported (Pirrone et al. 1998) and declines seen in Hg archives representing the 

last decade may be related to the closing of major former East German chlor alkali plants and coalfired 

power plants (Gerhard Petersen, pers. comm.) At this time we cannot exclude that the declines 

seen in our peat profiles may be artifacts due to some combination of physical, chemical, and biological 

processes, but the possibility remains that the declining Hg concentrations since the late 

1940s and more sharply during the last decade, reflect a real decline in atmospheric Hg concentrations. 

The same Hg concentration shape, with the same fall in Hg concentration in the approximate 

top ten cm has been seen in many other studies including a recent study at the far south latitude location 

of Patagonia (Biester et al. 2000). 

We do not present Hg flux calculations in this paper since more 14 C dates are required for samples 

from a) depths between the surface layers and lower layers to obtain the proper change of the accumulation 

rate in this region, and b) depths around the 1963 bomb-pulse peak to find out whether 

dampening takes place in the Denmark core.


Figure 4 Hg concentration profiles for the Storelungmose (Denmark) and the Tasiusaq (Greenland) peat 

cores. Hg concentrations are normalized to dry bulk density. Horizontal dashed lines indicate the depth where 

the macrofossils used for 14 C dating were taken. The corresponding calibrated ages are given as numbers 

close to the lines. Where two Hg concentration measurements were performed unweighted mean values and 

error bars corresponding to the uncertainty of the mean are shown. 

CONCLUSION 

By comparing data sets from all over the extratropical northern hemisphere, the present paper shows 

that 14 C dating of plants during the bomb-pulse period is generally possible at a precision of 1–2 

years. For the first time, we were able to clearly reproduce the whole atmospheric 14 C bomb-pulse 

curve in peat cores by measuring 14 C in macrofossils in peat from Greenland and Denmark, and we 

could exclude any significant dampening effect for these cores. We compared the 14 C bomb-pulse 

dating method, which allows precise dating of single points in the peat matrix, with the more familiar 

techniques based on records of fallout radionuclides. 14 C is actively taken up into the living material 

from the surrounding atmosphere and gets fixed via photosynthetic activity along with the stable 

isotopes 12 C and 13 C, which provide normalization of the 14 C concentrations, allowing direct dating 

of the material. In contrast, records of fallout 210 Pb, together with those of Pb and Hg, may be subject 

to small displacements at the time of deposition on the bog surface. Comparisons between 14 C 

and 210 Pb offer a means for determining a more precise interpretation of pollution records in bog 

archives. 

As an example of the usefulness of the 14 C bomb-pulse dating method combined with peat core analysis, 

we applied the 14 C bomb-pulse dating method to two peat cores from Greenland and Denmark 

to obtain high-resolution dates for Hg concentration profiles for the second half of the 20th century.


The 14 C bomb-pulse method is currently being evaluated in peat from the Faroe Islands and another 

site (Store Vildmose) in Denmark. Aside from the above sites, it will be used to obtain a high-resolution 

profile of contaminants through time in peat from locations such as Bathurst Island (Canada), 

and Carey Æ erne (the Carey Islands, Greenland) in the high Arctic. 

ACKNOWLEDGMENTS 

The financial support from the following sources is acknowledged and greatly appreciated: The 

Danish Cooperation for Environment in the Arctic (DANCEA) and the Danish Ministry of the Environment, 

the Danish Natural Science Research Council (SNF), the Swiss National Science Foundation 

(Grant Number 21-55669.98 to W Shotyk), Hans von Storch of the GKSS Institute for Hydrophysics, 

and Robert Frei of the Danish Isotope Center. We especially want to thank Andriy 

Cheburkin of EMMA analytical, Harald Biester, University of Heidelberg, Fiona Roos, University 

of Berne for their expertise in Hg measurements, and Henrik Loft Nielsen for exellent discussions 

throughout the project as well as improvements to the manuscript. Thanks to Douglas Donahue, 

Mariza Costa-Cabral, Ole John Nielsen, Henrik Skov, Steve Lindberg, and Gary Geernaert for discussion 

and encouragement; Rikke Brandt, Tommy Nørnberg and Otto Frederiksen for excellent 

field assistance. Special thanks to the people of Tasiusaq, Narsaq, Greenland, the Greenland Homerule, 

and the Danish Polar Center for their support. The ideas and opinions expressed in this article 

are the authors’ alone. They do not necessarily reflect the opinions or ideas of the funding agencies, 

sponsors, or employers. 

We are indebted to Ingeborg Levin, University of Heidelberg, for providing both 14 CO 2 measurement 

and model data. Data (a part of which has not been published so far) may be directly requested 

from I Levin (email: Ingeborg.Levin@iup.uni-heidelberg.de). 

REFERENCES 

Appleby, PG. 1998. Dating recent sediments by 210 Pb: 

Problems and solutions. Proc. 2nd NKS/EKO-1 Seminar, 

Helsinki, 2-4 April 1997, STUK, Helsinki. p 7–24 

Appleby PG, Oldfield F. 1978. The calculation of 210 Pb 

dates assuming a constant rate of supply of unsupported 

210 Pb to the sediment. Catena 5:1–8. 

Appleby PG, Oldfield F. 1992. Application of 210 Pb to 

sedimentation studies. In: Ivanovich M, Harmon RS, 

editors. Uranium-series disequilibrium: applications 

to earth, marine & environmental sciences. Oxford: 

Oxford University Press. p 731–78. 

Appleby PG, Nolan PJ, Gifford DW, Godfrey MJ, Oldfield 

F, Anderson NJ, Battarbee RW. 1986. 210 Pb dating 

by low background gamma counting. Hydrobiologia 

141:21–27. 

Appleby PG, Richardson N, Nolan PJ. 1992. Self-absorption 

corrections for well-type germaniun detectors. 

Nuclear Instruments and Methods in Physics Research 

B 71:228–33. 

Appleby PG, Shotyk W, Fankhauser A. 1997. 210 Pb age 

dating of three peat cores in the Jura Mountains, Switzerland. 

Water Air and Soil Pollution 100 (3/4):223– 

31. 

Arslanov KA, Saveljeva LA, Gey NA, Klimanov VA, 

Chernov SB, Chernova GM, Kuzmin GF, Tertychnaya 

TV, Subetto DA, Denisenkov VP. 1999. Chronology of 

vegetation and paleoclimatic stages of northwestern 

Russia during the late glacial and holocene. Radiocarbon 

41(1):24–5. 

Bennett BB, De Geer LE, Doury A. 2000. Nuclear weapons 

test programmes of the different countries. In: 

Warner F, Kirchmann RJC, editors. Nuclear test explosions—environmental 

and human impacts. Chichester, 

New York, Weinheim, Brisbane, Singapore, Toronto: 

John Wiley & Sons, Ltd. p 13–32. 

Benoit JM, Fitzgerald WF, Damman AWH. 1998. The 

biogeochemistry of an ombrotrophic bog: evaluation 

of use as an archive of atmospheric mercury deposition. 

Environmental Research, Section A 78:118–33. 

Biester H, Kilian R, Hertel C, Schöler HF. 2000. Elevated 

mercury concentrations in southern Patagonian peat 

bogs—An anthropogenic signal? Oral presentation at 

the Silver Anniversary International Conference on 

Heavy Metals in the Environment in Ann Arbor, USA, 

6–10 August 2000. 

Boutron CF, Vandal GM, Fitzgerald WF, Ferrari CP. 

1998. A 40-year record of mercury in central Greenland 

snow. Geophysical Research Letters 25(17): 

3315–8. 

Braune B, Muir D, Demarch B, Gamberg M, Poole K, 

Currie R, Dodd M, Duschenko W, Eamer J, Elkin B, 

Evans M, Grundy S, Hebert C, Johnstone R, Kidd K,


Koenig B, Lockhart L, Marshall H, Reimer K, Sanderson 

J, Shutt L. 1999. Spatial and temporal trends of 

contaminants in Canadian Arctic freshwater and terrestrial 

ecosystems: a review. Science of the Total Environment 

230:145–207. 

Cortizas AM, Ponteedra Pomba X, Garcia-Rodeja E, Novoa 

Munoz JC, Shotyk W. 1999. Mercury in a Spanish 

peat bog: archive of climate change and atmospheric 

metal deposition. Science 284:939–42. 

Dai KM, Fan CY. 1986. Bomb produced 14C content in 

tree rings grown at different latitudes. Radiocarbon 

28(2A):346-9. 

Dai KM, Qian Y, Fan CY. 1992. Bomb-produced 14C in 

tree rings. Radiocarbon 34(3):753–6. 

Ehhalt DH. 1999. Gas Phase Chemistry of the Troposphere. 

In: Baumgärtel H, Grünbein W, Hensel F, editors. 

Global aspects of atmospheric chemistry. Darmstadt: 

Steinkopff Verlag and New York: Springer 

Verlag. p 21–109. 

Gedyé SJ. 1998. Mass balance in recent peats. Unpublished 

PhD thesis, Liverpool University. 

Goodsite MG. 2000. Determination of heavy-metal deposition 

by correlation with 14-C. Thesis. The University 

of Southern Denmark, Odense University, Department 

of Chemistry, March 2000. 

Jungner H, Sonninen E, Possnert G, Tolonen K. 1995. 

Use of bomb-produced 14C to evaluate the amount of 

CO2 emanating from two peat bogs in Finland. Radiocarbon 

37(2):567–73. 

Levin I, Kromer B, Schoch-Fischer H, Bruns M, Münnich 

M, Berdau D, Vogel JC, Münnich KO. 1997. 

14CO2 records from two sites in Central—Schauinsland 

& Vermunt. URL: . 

Levin I, Hesshaimer V. 2000. Radiocarbon—a unique 

tracer of global carbon cycle dynamics. Radiocarbon 

42(1):69-80. 

Levin I, Kromer B, Schoch-Fischer H, Bruns M, Münnich 

M, Berdau D, Vogel JC, Münnich KO. 1985. 25 

years of tropospheric 14C observations in Central Europe. 

Radiocarbon 27(1):1–19. 

Murphy JO, Lawson EW, Fink D, Hotchkis MAC, Hua Q, 

Jacobsen GE, Smith AM, Tuniz C. 1997. 14C measurements 

of the bomb pulse in N- and S-hemisphere tropical 

trees. Nuclear Instruments and Methods in Physics 

Research B123:447–50. 

Nydal R, Lövseth K. 1983. Tracing bomb 14C in the atmosphere 

1962-1980. Journal of Geophysical Research 

88 (C6):3621–42. 

Nydal R, Lövseth K. 1996. Carbon-14 measurements in 

atmospheric CO2 from northern and southern hemisphere 

sites, 1962–1993. URL: . 

Nydal R, Gislefoss JS. 1996. Further application of bomb 

14C as a tracer in the atmosphere and ocean. Radiocarbon 

38(3):389–406. 

Olsson IU. 1986. A study of errors in 14C dates of Peat 

and Sediment. Radiocarbon 28(2A):429–35. 

Olsson IU. 1989. Recent 14 C activity in the atmosphere, 

“clean air” and the Chernobyl effect. Radiocarbon 

31(3):740–6. 

Olsson IU, Possnert G. 1992. 14 C activity in different sections 

and chemical fractions of oak tree rings, AD 

1938–1981. Radiocarbon 34(3):757–67. 

Olsson IU. 1993. A ten-year record of the different levels 

of the 14 C activities over Sweden and the Arctic. Tellus 

45B:479–81. 

Pirrone N, Allegrini I, Keeler GJ, Nriagu J, Rossman R, 

Robbins JA. 1998. Historical atmospheric mercury 

emissions and depositions in North America compared 

to mercury accumulations in sedimentary 

records. Atmospheric Environment 32(5):929–40. 

Puchegger S, Rom W, Steier P. 2000. Automated Evaluation 

of 14 C Measurements. Nuclear Instruments and 

Methods in Physics Research B 172:274–80. 

Rom W. 2000. 14 C Accelerator Mass Spectrometry - Applications 

in Archaeology, Biomedicine and in the Atmospheric 

Sciences. Thesis, University of Vienna, 

Austria. 

Shore JS, Bartley DD, Harkness DD. 1995. Problems encountered 

with the 14 C dating of peat. Quaternary Science 

Reviews (Quaternary Geochronology) 14:373– 

83. 

Shotyk W, Weiss D, Appleby PG, Cheburkin AK, Frei R, 

Gloor M, Kramers JD, Reese S, van der Knaap WO. 

1998. History of atmospheric lead deposition since 

12,370 14 C yr BP recorded in a peat bog profile, Jura 

Mountains, Switzerland. Science 281:1635–40. 

Shotyk W, Goodsite ME, Lohse C, Hansen TS, Roos F, 

Biester H, Cheburkin AK, Heinemeier H, Appleby 

PG, Reese S. 2001. Continuous, 3000-year peat core 

records of Atmospheric Hg in Greenland and Denmark. 

Earth and Plantetary Science Letters. Submitted. 

Stuiver M, Quay PD. 1981. Atmospheric 14 C changes resulting 

from fossil fuel CO 2 release and cosmic ray 

flux. Earth and Planetary Science Letters 53:349–62. 

Stuiver M, Reimer PJ, Bard E, Beck JW, Burr GS, 

Hughen KA, Kromer B, McCormac G, van der Plicht 

J, Spurk M. 1998. INTCAL98 radiocarbon age calibration, 

24,000–0 cal BP. Radiocarbon 40(3):1041- 

83. Data are available at URL: . 

Tans PP. 1981. A compilation of bomb 14 C data for use in 

global carbon model calculations. In: Bolin B, editor. 

SCOPE 16, Carbon cycle modelling. Chichester New 

York Brisbane Toronto; Wiley. p 131–57. 

Tauber H. 1967. Copenhagen radiocarbon measurements 

VIII—geographic variations in atmospheric C 14 activity. 

Radiocarbon 9:246–56. 

USNT. 2000. United States Nuclear Tests—July 1945 

through September 1992. Report. DOE/NV—209- 

REV 15. United States Department of Energy, Nevada 

Operations Office. Las Vegas, Nevada. URL: http://

www.nv.doe.gov/news&pubs/publications/historyreports/pdfs/DOENV209_REV15.pdf. 

Wardenaar ECP. 1987. A new hand tool for cutting peat. 

Canadian Journal of Botany 65:1772–3. 

White JWC, Ciaia P, Figge RA, Kenny R, Markgraf V. 

1994. A high resolution record of atmospheric CO 2 


content from carbon isotopes in peat. Nature 367:153– 

6. 

Wrobel S, Eckstein D. 1992. Determining Time and Environment 

from Tree Rings. In: Hackens T, Jungner H, 

Carpelan C, editors. PACT 36. Rixensart: PACT Belgium. 

p 33–49.


Paper 7: Roos-Barraclough F; Givelet N; Martinez-Cortizas A; Goodsite M E; Biester H; Shotyk 

W An analytical protocol for the determination of total mercury concentrations in solid 

peat samples. Science of the Total Environment (2002 Jun 20), 292(1-2), 129-39.

The Science of the Total Environment 292 (2002) 129–139 

An analytical protocol for the determination of total mercury 

concentrations in solid peat samples 

a, a b c,1 d 

F.Roos-Barraclough *, N.Givelet , A.Martinez-Cortizas , M.E.Goodsite , H.Biester , 

W.Shotyka,d Abstract 

aGeological 

Institute, University of Berne, Baltzerstrasse 1, 3012 Berne, Switzerland 

Department of Soil Science, University of Santiago di Compostela, Santiago di Compostela, Spain 

cDepartment 

of Chemistry, Odense University, Odense, Denmark 

dInstitute 

of Environmental Geochemistry, University of Heidelberg, Heidelberg, Germany 

b 

Received 10 November 2000; accepted 20 December 2001 

Traditional peat sample preparation methods such as drying at high temperatures and milling may be unsuitable 

for Hg concentration determination in peats due to the possible presence of volatile Hg species, which could be lost 

during drying.Here, the effects of sample preparation and natural variation on measured Hg concentrations are 

investigated.Slight increases in mercury concentrations were observed in samples dried at room temperature and at 

y1 y1 

30 8C (6.7 and 2.48 ng kg h , respectively), and slight decreases were observed in samples dried at 60, 90 and 

y1 y1 

105 8C (2.36, 3.12 and 8.52 ng kg h , respectively).Fertilising the peat slightly increased Hg loss (3.08 ng 

y1 y1 y1 y1 

kg h in NPK-fertilised peat compared to 0.28 ng kg h in unfertilised peat, when averaged over all 

temperatures used).Homogenising samples by grinding in a machine also caused a loss of Hg.A comparison of two 

Hg profiles from an Arctic peat core, measured in frozen samples and in air-dried samples, revealed that no Hg losses 

occurred upon air-drying.A comparison of Hg concentrations in several plant species that make up peat, showed that 

some species (Pinus mugo, Sphagnum recurvum and Pseudevernia furfuracea) are particularly efficient Hg retainers. 

The disproportionally high Hg concentrations in these species can cause considerable variation in Hg concentrations 

within a peat slice.The variation of water content (1.6% throughout 17-cm core, 0.97% in a 10=10 cm slice), bulk 

density (40% throughout 17-cm core, 15.6% in a 10=10 cm slice) and Hg concentration (20% in a 10=10 cm 

slice) in ombrotrophic peat were quantified in order to determine their relative importance as sources of analytical 

error.Experiments were carried out to determine a suitable peat analysis program using the Leco AMA 254, capable 

of determining mercury concentrations in solid samples.Finally, an analytical protocol for the determination of Hg 

concentrations in solid peat samples is proposed.This method allows correction for variation in factors such as 

vegetation type, bulk density, water content and Hg concentration in individual peat slices.Several subsamples from 

*Corresponding author.Tel.: q41-31-6318761; fax: q41-31-631-4843. 

E-mail address: fiona.roos@geo.unibe.ch (F.Roos-Barraclough). 

1 

Present address: Department of Atmospheric Environment, National Environmental Research, Denmark. 

0048-9697/02/$ - see front matter � 2002 Elsevier Science B.V. All rights reserved. 

PII: S0048-9697Ž 02. 00035-9

130 F. Roos-Barraclough et al. / The Science of the Total Environment 292 (2002) 129–139 

each peat slice are air dried, combined and measured for Hg using the AMA254, using a program of 30 s (drying), 

125 s (decomposition) and 45 s (waiting).Bulk density and water content measurements are performed on every 

slice using separate subsamples. � 2002 Elsevier Science B.V. All rights reserved. 

Keywords: Mercury; Peat bog archives; Atmospheric pollution 


Records of net accumulation of atmospheric Hg 

are well-preserved in peat cores from ombrotrophic 

bogs (Madsen, 1981; Jensen and Jensen, 1991; 

Benoit et al., 1998; Burgess et al., 1998; Martinez- 

Cortizas et al., 1999; Shotyk et al., 2001).By 

measuring the concentrations of Hg in peat extending 

back in time to pre-anthropogenic periods, 

natural ‘background values’ and their climate- 

(Martinez-Cortizas et al., 1999) and volcanismrelated 

variations can be quantified and used to 

identify the effects of recent increases due to 

human activities.During traditional acid digestion 

of peat samples, various sources of contamination 

must be considered and clean techniques such as 

those employed by Weiss et al. (1999a) used. 

Analysis of solid samples by combustion and 

subsequent trapping of Hg on gold before analysis 

by AAS (Salvato and Pirola, 1996) using the Leco 

AMA 254 (Martinez-Cortizas et al., 1999) not 

only avoids several possible sources of contamination 

(acids, digestion vessels) but is also safer 

and less expensive (no acids required) and allows 

a high sample throughput.The ash from the 

combusted samples can be recovered and used for 

further analysis, for example determination of 

refractory trace metals such as Ti and Zr by X-ray 

fluorescence spectrometry (XRF).However, the 

volatility of Hg (0) combined with the natural 

variability of peat structural components (water 

content, bulk density, ash content) mean that care 

must also be taken to achieve accurate results 

when analysing solid samples. 

2. Methods 

All tests were carried out in the trace metals 

laboratory of the Geological Institute of the Uni- 

versity of Berne.Mercury concentrations were 

determined by AAS using the Leco AMA 254. 

The bog at Etang de la Gruere (EGR) in the 

Swiss Jura mountains is ombrotrophic, with 6.5 m 

14 

of peat dating from 12 370 C yr BP (Shotyk et 

al., 1998).Peat from the ombrotrophic, pre-anthropogenic 

part of the EGR2A core was homogenised 

and used to determine losses upon drying at 

different times and temperatures, in both unfertilised 

and artificially fertilised peat.Twenty-three 

different plant species were collected over a period 

of 4 years from the surface of the bog at Etang de 

la Gruere ` and their mercury concentrations were 

determined using the Leco AMA 254.The range 

in water content in air-dried peat subsamples 

(cylindrical ‘plugs’, 16-mm diameter) was also 

determined using samples from this core. 

The Hg concentration profile was measured 

twice in a peat core from the High Arctic of 

Canada, using (1) frozen samples and (2) samples 

which had been air-dried in a clean-air cabinet 

overnight at room temperature (RT) to determine 

whether Hg losses occurred upon drying. 

A peat monolith from Schoepfenwaldmoor 

(SWM), Switzerland (Weiss et al., 1999b), was 

used to investigate the variation of water content, 

bulk density and Hg concentration in a typical 

ombrotrophic peat.Water content and bulk density 

were studied both in a short core (17 cm) and 

within an individual peat slice (10=10=1 cm). 

Variation of Hg concentrations within a single 

slice was also recorded.The investigation of the 

effect of air drying at room temperature was also 

carried out using samples from this core. 

The effect of grinding dry samples and of 

alterations to the Leco AMA 254 analysis program 

were studied using an in-house peat standard (OGS 

1878P), which had previously been dried at 105 

8C and homogenised.

F. Roos-Barraclough et al. / The Science of the Total Environment 292 (2002) 129–139 

2.1. Determination of a suitable peat analysis 

program for Hg analysis using the Leco AMA 254 

The Leco AMA 254 is fully compliant with 

EPA (1998) method 7473.Samples contained in 

nickel sample holders enter a sealed drying and 

combustion furnace, where they are dried in a 

stream of oxygen before being thermally decomposed.Gases 

from the thermally decomposed sample 

are swept in the stream of oxygen through a 

catalyst furnace at 750 8C, which fully decomposes 

the gases and traps NO 2, SO2 and halogens.Mercury 

is trapped on a gold amalgamator situated at 

the end of the furnace.Waste gases are removed 

from the system by the gas stream.The amalgamator 

is then heated to 500 8C to release the 

mercury, which is measured by atomic absorption 

spectrometry.The detection limit of the instrument 

is 0.01 ng Hg and the working range is 0.05–600 

ng Hg, with repeatability being -1.5%. 

The instrument was calibrated using liquid standards 

prepared from Merck mercury standard solu- 

y1 

tion, 1000 mg l .Solutions of concentration 10 

y1 

and 1000 ng ml were used to dose the instru- 

ment.A 10-point calibration was made from 0.00 

to 29 ng Hg.The equation used to calculate the 

calibration curve is: 

1 

kŽ NST. s µ AjiŽ NST. ymji∂ 

8ji n 

Where k is the constant of proportionality, NST is 

the relative non-absorbable radiation flux, A is the 

corrected absorbance and m is the quantity of 

mercury in the cell.For the calibration obtained, 

slope (k)s42.64"0.62 ng. 

To test the effect of increased drying time (and 

therefore increased Hg passing through the apparatus 

from oxygen supply), blanks were run on 

the Leco AMA 254 using drying times of 9 and 

500 s (other parameters being kept constant: 

decomposition time 150 s, waiting time 45 s). 

The validity of the recommended drying time 

was also tested w0.7=vol.water (ml)x s.Ten 

samples of the in-house peat standard OGS 1878 

P were moistened (made up to 95% water, 

RH OG18 

V) and analysed using the recommend- 

2 

ed drying time.Ten dry samples were also analysed 

using a drying time of 20 s. 

131 

A suitable decomposition time was established 

by measuring Hg in nine samples of OGS 1878 P 

at 100, 125, 150 and 175 s decomposition time 

(drying and waiting time being constant at 30 and 

45 s, respectively).One coal and five plant-derived 

certified standard reference materials (SRMs) were 

analysed using this program. 

2.2. Effects of sample preparation: drying times 

and temperatures 

Homogenisation of the sample is difficult if the 

peat is wet but can easily be carried out by hand 

using dry peat.However, drying peat, particularly 

samples from cold regions, could results in a loss 

of Hg by volatilisation.For this reason, tests of 

the effect of air-drying peat at room temperature 

on Hg content were carried out.The Hg profile of 

an Arctic peat core was measured twice; once 

using samples which had been kept frozen since 

collection and once using samples which had been 

air-dried overnight in a class 100 clean-air cabinet. 

All samples were measured in duplicate using the 

Leco AMA 254 program recommended here.Bulk 

density was determined for each sample as 

described below and from this, the volumetric 

y3 

concentrations (ng cm ) were calculated. 

To investigate the effect of prolonged drying at 

high temperatures and the effect of fertilisation 

upon Hg losses upon drying, bulk samples of peat 

from the ombrotrophic, pre-anthropogenic section 

(88–234 cm) of the EGR2A core was homogenised 

using a food mixer.The mixture was divided 

into five sections of 150 g each.Three sections 

were artificially fertilised with NH NO (1 mM), 

4 3 

Ca (PO ) (1mM), KCO (1 mM), respectively, 

3 4 2 2 3 

and a fourth was fertilised with a mixture of the 

three above additives (1 mM each).The fifth 

section was left unfertilised.These sections were 

then subsampled into five parts, which were dried 

at the following temperatures; RT, 30, 60, 90 and 

105 8C for the following times; 0, 19, 24, 48, 120, 

168 and 336 h.The drying was carried out in 

acid-cleaned, Teflon bowls, each containing 3.5 g 

of peat slurry.After drying, the samples were 

stored frozen, sealed in air-tight plastic bags until 

AAS Hg analysis using the Leco AMA 254.


2.3. Effects of sample preparation: grinding 

samples 

Five samples of an in-house peat reference 

material were ground using a coffee mill.These 

samples were analysed for Hg using the Leco 

AMA-254 and the values obtained were compared 

to those obtained using five unground samples of 

the same material. 

2.4. Effects of sample preparation: water content 

in air-dried plugs 

The EGR2A core (co-ordinates CH 570.525, 

232.150) was collected on 26 August 1991.The 

core was taken using a Livingston corer (Aaby 

and Digerfeldt, 1986).The cylindrical core was 

initially taken in 1-m sections (total length 692 

cm, ds8 cm).The core was then sliced, frozen, 

into 2-cm slices.The slices were stored in individual 

airtight plastic bags at y18 8C until analysis. 

Slices from 46 to 592 cm were thawed before 

plugs were taken.Four plugs were taken from 

each slice.One plug from each slice was weighed 

wet.All the plugs were allowed to air dry in a 

class 100 clean air cabinet for 20 h.The air-dried 

plugs, which had been weighed wet, were reweighed 

after 20 h and then placed in an oven at 

105 8C until constant weight was obtained.The 

percentage water remaining in the air-dried plugs 

was then calculated from the air-dried weight and 

the constant dry weight of the plugs.This value 

for water content of air-dried samples was used to 

calculate the theoretical dry weight of the remaining 

air-dried plugs, which were subsequently analysed 

for Hg. 

2.5. Effects of peat properties: Hg concentrations 

in different bog plant species 

Twenty-three plant species were collected from 

the surface of the ombrotrophic bog at Etang de 

la Gruere ` in the Swiss Jura Mountains (1005 m) 

between July 1997 and September 2000.The 

samples were stored in air-tight plastic bags immediately 

after collection and were kept frozen until 

analysis.Samples were then air dried in a class 

100 laminar flow cabinet overnight and analysed 

for Hg content by AAS, using the Leco AMA 

254. 

2.6. Bulk density and water content in a peat 

monolith 

A peat monolith was collected at SWM (coordinates 

CH 631.250, 177.000) on 28 August, 

1991, using a Wardenaar corer (Wardenaar, 1987). 

The monolith obtained (10=10=100 cm) was 

stored at y18 8C after collection.It was then cut 

(frozen) into 1-cm slices, using a stainless steel 

band saw and the individual slices were again 

stored at y18 8C until analysis.The top 17 slices 

were used in the following analyses.A stainless 

steel tube with a sharpened end, of diameter 16mm 

was used to remove three plugs of known 

volume from each 1-cm slice.The wet weight of 

each plug was recorded and the plugs were then 

dried to constant weight in an oven at 105 8C.The 

dry weights were recorded and the plugs were then 

stored in air-tight plastic boxes, which had been 

soaked for 1 h in 10% HNO and rinsed six times 

3 

with water (R G18 V). 

H O 

2 

2.7. Effects of peat properties: variation in Hg 

concentrations within one 10=10=1 cm slice 

The slice 15 cm from the surface of SWM core 

was used to investigate the diversity of Hg concentrations 

within a single 10=10=1 cm slice of 

peat.The slice appeared to be homogenous in 

terms of composition of plant material.However, 

wood was removed from the samples before analysis.Sixteen 

plugs (four rows of four) were 

extracted as described above, at even distance 

from one another.The samples were analysed for 

Hg using the Leco AMA 254.The theoretical dry 

weight of the samples was calculated using their 

wet weights and the average water content of slice 

15, previously determined to be 93.0"0.4%. The 

Hg concentration was calculated and expressed as 

y1 ng g in dry weight of peat. 

3. Results 

3.1. Suitable peat analysis program using the Leco 

AMA 254 

Increasing the drying time from 9 to 500 s only 

increased the blank value from 0.020 Hg (ns2)


Table 1 

Results of the analysis of six certified standard reference materials 

using the suggested program on the Leco AMA 254 

SRM Measured wHgx n Certified wHgx 

(ngyg) (ngyg) 

NIST 45.1"0.8 13 44"4 

1515 

NIST 32.3"1.4 17 31"7 

1547 

NIST 118.9"0.9 2 150"50 

1575 

NIST 99.2"2.7 2 93.8"3.7 

1630a 

BCR 16.6"0.5 2 21"2 

281 

IAEA 170.2"0.1 2 200"40 

336 

to 0.025 ng Hg (ns2).It was therefore concluded 

that the effect of the length of program was 

negligible with respect to increased blank values 

due to increased Hg from the oxygen supply during 

longer programs. 

The result of 10 analyses of in-house peat 

standard OGS 1878 P, which were made up to 

95% water (RH OG18 

V), using the recommended 

2 

y1 

drying times, was 84.0"18.0 ng g Hg: a relative 

standard deviation of 22%.The result of 10 anal- 

y1 

yses of the dry material was 76.6"2 ng g Hg: 

a relative standard deviation of 8%.Thus, analysis 

of dried samples appears to reduce the standard 

deviation of the results. 

Analysis using a decomposition time of 125 s 

resulted in the lowest standard deviation among 

10 analyses of OGS 1878P (1.9% compared to 

5.15, 8.87 and 9.68% at 100, 150 and 175 s, 

respectively).Six SRMs analysed using this program 

gave the results shown in Table 1. 

3.2. Effect of prolonged drying at high temperatures 

in unfertilised and artificially fertilised peat 

The results (Fig.1) show overall trends in 

changes in Hg concentration at a rate of 6.7 ng 

y1 kg y1 h y1 at RT, 2.48 ng kg y1 h at 30 8C, y 

y1 2.36 ng kg y1 h y1 at 60 8C, y3.12 ng kg y1 h 

y1 at 90 8C and y8.52 ng kg y1 h at 105 8C for 

these samples (Table 2).Therefore, it appears that 

there is an increase over time in wHgx in peat, 

133 

which is left to air dry at low temperatures, 

probably due to absorbance of Hg from the lab 

air.In contrast, there is a decrease in Hg concentrations 

in samples dried at higher temperatures, 

probably due to loss of Hg by volatilisation.This 

loss is so gradual that drying at low temperatures 

for a few hours should not measurably affect the 

Hg concentrations.However, prolonged drying at 

higher temperatures could result in a significant 

loss of Hg and extensive drying at room temperature 

could give rise to significant Hg 

contamination. 

Artificial fertilisation of the peat was carried out 

to stimulate microbial activity.Artificial fertilisation 

has been found in the past to increase Hg loss 

by volatilisation (Lodenius et al., 1983).Average 

changes in Hg concentrations over the whole 

y1 y1 

temperature range were y0.26 ng kg h for 

y1 y1 

N-fertilised peat, y2.38 ng kg h for P- 

y1 y1 

fertilised peat, 1.18 ng kg h for K-fertilised 

y1 y1 

peat, y3.08 ng kg h for N, P and K-fertilised 

y1 y1 

peat and y0.28 ng kg h for unfertilised peat. 

These results indicate that the addition of phosphorus 

in particular increases mercury losses, probably 

due to increased microbial activity.The 

addition of potassium, however, appeared to 

decrease Hg loss. 

3.3. Effect of air-drying at room temperature on 

Hg content of peat 

The Hg concentration profile in the Arctic peat 

core was measured once using samples which had 

remained frozen since collection and once using 

samples which had been air-dried at RT in a cleanair 

cabinet overnight.Both sets of samples were 

corrected for bulk density and the results are 

displayed as volumetric concentrations (ng cm , y3 

Fig.2).The average difference in Hg concentration 

between Hg content of the samples was 0.3"0.9 

y3 ng cm , ns63, with the wet samples being on 

average slightly lower in Hg than the air-dried 

samples.This difference can be accounted for by 

a combination of the error involved in the bulk 

density determination and uptake of Hg from the 

lab air during air-drying.However, the previous 

experiment suggested a rate of increase in Hg 

y1 y1 

concentration of 6.7 pg g h during air drying


Fig.1.Hg concentrations measured in both unfertilised and artificially fertilised peat dried at RT, 30, 60, 90 and 105 8C for up to 

2 weeks. 

at RT and would therefore account for an increase 

y1 

of only 0.9 ng g , or 2%, during 14 h of air 

drying.It is therefore likely that most of the 

observed increase stems from error in the bulk 

density determination, which was high in this case 

due to the presence of ice in the peat. 

3.4. Effect of grinding samples 

Grinding promotes homogenisation of the sample 

by reducing particle size and by mixing.It 

also allows a greater mass of sample to be analysed 

at one time, as more material can be placed into 

each sample vessel.Therefore grinding allows a 

more representative value to be obtained. 

The average Hg concentration for unground 

y1 

samples was 47.9"1. 2 ng g (ns5) and the 

average for ground samples was 45.2"1.2 ng 

y1 g (ns5).It is therefore possible that a slight 

loss of Hg occurs on grinding, possibly due to 

elevated temperature, coupled with smaller particle


Table 2 

Trends of Hg lossygain during drying in both unfertilised and artificially fertilised peat samples over time (up to 2 weeks) at various 

temperatures 

Trendline gradients 

Temperature N P K NPK Unfertilised Average Av.lossygain 

gradient in ngykg per h 

22 0.0045 0.0068 0.0165 0.005 0.0007 0.0067 6.7 

30 y0.0034 0.0036 0.012 y0.0025 0.0027 0.00248 2.48 

60 0.0004 y0.0063 y0.0097 0.0016 0.0022 y0.00236 y2.36 

90 y0.0014 y0.0047 y0.0005 y0.0062 y0.0028 y0.00312 y3.12 

105 y0.0014 y0.0113 y0.0124 y0.0133 y0.0042 y0.00852 y8.52 

av.gradient y0.00026 y0.00238 0.00118 y0.00308 y0.00028 

av.lossygain y0.26 y2.38 1.18 y3.08 y0.28 

in ngykg per h 

size within the grinder.To achieve a homogenised 

sample without loss of Hg, dried samples can be 

crushed by hand in a sealed plastic sample bag.In 

this way, a fine-grained powder can be obtained 

from the mossy part of the sample, although fibres 

from grass and wood remain intact.This facilitates 

the selection of the mossy part of the samples for 

Hg analysis, reducing the amount of grassy and 

woody material in the samples and therefore also 

reducing the variation in Hg concentration values 

caused by variation in peat type.The relative 

standard deviation in Hg concentration results, 

using samples homogenised in this way, measured 

in duplicate was 4.0% (ns265 duplicate pairs). 

3.5. Variation of water content in air dried plugs 

After drying overnight at RT in the clean air 

cabinet, the average water content of the plugs 

extracted from the EGR core was 9.1"2.0% (ns 

296).Water content of samples to be analysed can 

thus be calculated by comparison with a similarly 

sized and shaped sample, which has been allowed 

to dry for the same length of time and is subsequently 

dried to constant weight. 

3.6. Comparison of Hg concentrations in bog plant 

species 

The comparison of Hg concentrations in a collection 

of 23 plant species all collected from the 

surface of the one bog over a period of 4 years 

(Fig.3) shows an overall agreement among most 

135 

species in each year of collection (1997: 207"44 

y1 y1 

ng g , ns5; 1999: 47"29 ng g , ns10; 2000: 

y1 

50"45 ng g , ns20).Species which showed 

unusually high Hg concentrations are the mosses 

Sphagnum recurvum and Pseudivernia furfuracea 

and the needles and twigs, but not cones or bark, 

of the tree Pinus mugo.Omission of these from 

the data results in average concentrations of 

y1 

36"19 ng g , ns8 for 1999 and 32"14 ng 

y1 g , ns17 for 2000, i.e. reducing the standard 

deviation from 62 to 53% and 90 to 44%, 

respectively. 

3.7. Inconsistency in Hg concentration in a lateral 

peat slice 

Mercury concentrations of the 16 samples taken 

from one SWM peat slice were found to vary by 

y1 y1 

over 20%: 69.2"14.7 ng g , min. 41.4 ng g , 

y1 

max.98.7 ng g Hg.These results indicate the 

importance of homogenising the peat moss prior 

to analysis andyor analysing several samples from 

each slice and averaging the results in order to 

obtain values representative of the whole slice. 

3.8. Bulk density and water content 

The average water content of the 17 slices of 

the SWM core analysed was 94.0"1.6%. The 

average relative standard deviation of the percentage 

of water in a single slice was 0.97%. These 

results indicate that water content fluctuation is 

negligible in the part of the core studied.


Fig.2.Comparison of Hg concentration profiles measured in (a) frozen and (b) air-dried samples from a peat core from the High 

Arctic of Canada. 

However, the average bulk density for all the 

y3 

samples studied was 0.05"0.02 g cm and the 

average relative standard deviation of bulk density 

within a slice was 15.6% (ns16).This is a 

significant fluctuation, indicating that bulk density 

variation should be taken into account when comparing 

Hg values from different parts of a core 

and should be determined throughout the whole 

core if its values are to be used to calculate Hg 

flux data. 

4. Discussion 

The results of the experiments above indicate 

that great care must be taken during peat sample 

preparation and Hg analysis in order to obtain 

accurate and representative results. 

Although the Leco AMA 254 is a suitable 

instrument for Hg analysis in both wet and dry 

peat samples, drying facilitates homogenisation of 

the peat and also allows a greater amount of 

sample to be analysed and therefore a more representative 

value to be obtained.Analysis of dry 

samples also resulted in a lower standard deviation 

than analysis of wet samples.Air drying at room 

temperature overnight appeared, in this set of 

experiments, not to result in a significant loss or 

gain of Hg in the samples, even in peat from the 

High Arctic.However, prolonged drying was 

shown to lead to significant increases in Hg concentrations, 

due to uptake of Hg from the lab air. 

Prolonged drying at higher temperatures lead to an 

overall decrease in Hg concentrations and should 

therefore be avoided in total Hg studies.This is in


Fig.3.Hg concentrations measured in 23 peat-forming plant species collected from the surface of one bog over a period of 4 years. 

N.B. Only Sphagnum collected in 1997. 

agreement with the findings of Norton et al. (1997) 

and Martinez-Cortizas et al. (1999).Fertilisation 

with phosphorus increased Hg losses, probably due 

to increased microbial activity.Variation in the 

water content of air-dried plugs is small and can 

be corrected by measuring a comparable plug from 

the same slice, which has been dried in the same 

manner. 

Mercury concentrations vary greatly between 

different bog plants and even between different 

137 

parts of the same plant species, with certain mosses 

and trees having Hg concentrations much higher 

than average.Differences in vegetation composition 

can exist between bogs or even within the 

same bog at different times.This could potentially 

affect the Hg record and also the comparability of 

records from different bogs.It is therefore suggested 

that leaves, grass and wood be removed 

from the samples before Hg analysis.However, 

even Hg concentrations between different moss


species or between mosses growing within different 

microclimates on the surface of a bog could 

be inconsistent (Norton et al., 1997).Mercury 

concentrations within one small horizontal section 

of peat may be uneven.Several subsamples should 

be taken from each peat slice in order to obtain a 

representative Hg concentration for the whole 

slice.Because large fluctuations in bulk density 

values can occur throughout a peat core, bulk 

density should be determined in each slice to allow 

Hg flux rates to be calculated accurately. 

Finally, mixing together the moss component of 

the subsamples increases the homogeneity of the 

material to be analysed.Here, grinding in a coffee 

mill resulted in a slight loss of Hg.However, 

simple hand mixing of the dried samples held in 

a plastic bag resulted in low standard deviation of 

duplicate pairs. 

5. Conclusions 

The following protocol is proposed for the 

determination of mercury concentrations in solid 

peat samples: 

Cores should be transported frozen from the 

field to the laboratory, if possible, to avoid unnecessary 

compaction and losses of water or Hg.The 

peat should be kept in clean (wrapped airtight in 

plastic), cool conditions until analysis.The edge 

of the core should be removed prior to analysis in 

case of contamination (by smearing) during core 

collection.Several subsamples should be analysed 

from each peat slice in order to obtain a representative 

value.These subsamples can be air-dried at 

room temperature in a clean environment, such as 

a Class 100 laminar flow cabinet, without significant 

loss of mercury.During this period, the 

cabinet should be darkened to avoid light-enhanced 

evasion of mercury from the samples (Gustin and 

Maxey, 1998).Dried subsamples can be homogenised 

by crushing air-dried plugs of peat together 

into a powder and mixing; this can be done by 

hand, with the samples in a sealed plastic bag. 

One or more subsamples from each slice should 

be used for bulk density determination, which is 

required to allow mercury deposition fluxes to be 

calculated.Bulk density can be determined from 

the dry weight of a peat subsample of known 

volume.Air drying this subsample with the subsamples 

to be used for mercury analysis (of the 

same shape and volume) before drying it in a 

drying oven to constant weight allows the water 

content of the air-dry samples to be accurately 

estimated; this information can be used to correct 

the masses of the samples analysed for Hg to their 

true dry weights. 

Suggested times for a Leco AMA 254 air-dry 

peat analysis program are: 30 s (drying), 125 s 

(decomposition) and 45 s (waiting for emission 

of waste gases before Hg content determination by 

AAS). 

Acknowledgments 

Financial support for this work, including graduate 

student assistantships to F.R. and N.G., was 

provided by the Swiss National Science Foundation 

(grants 21-55669.98 and 21-061688.00) to 

W.S. Peat core collection in the High Arctic of 

Canada was made possible by a research grant to 

W.S., Heinfried Scholer ¨ (University of Heidelberg) 

and Stephen Norton (University of Maine) by the 

International Arctic Research Centre, Fairbanks, 

Alaska.Many thanks to Drs W.O.Van der Knaap, 

E.Feldmeyer, A.Gruenig and A.Holzer for plant 

identification and B.Eilrich for considerable field 

assistance. 

References 

Aaby B, Digerfeldt G.Handbook of Holocene Paleoecology 

and Paleohydrology.Sampling techniques for lakes and bogs 

New York: John Wiley and Sons, 1986.p.181 –194. 

Benoit JM, Fitzgerald WF, Damman AWH.The biogeochemistry 

of an ombrotrophic peat bog: evaluation of use as an 

archive of atmospheric mercury deposition.Environ Res 

Sect A 1998;78:118 –133. 

Burgess, N., Beauchamp,S., Brun, G., Clair, T., Roberts, C., 

Rutherford, L., Tordon, R.,Vaidya,O. Mercury in Atlantic 

Canada A Progress Report, Environment Canada, Atlantic 

Region.Environment Canada Report, 1998. 

Environmental Protection Agency.EPA method 7473:Mercury 

in solids and solutions by thermal decomposition, amalgamation, 

and atomic absorption spectrophotometry.1998. 

Gustin MS, Maxey R.Mechanisms influencing the volatile 

loss of mercury from soil.Proceedings of Air and Waste 

Management Association: Measurement of toxic and related 

air pollutants.Carty, NC Sept.1–3, 1998.


Jensen A, Jensen A.Historical deposition rates of mercury in 

Scandinavia estimated by dating and measurement of mercury 

in cores of peat bogs.Water Air Soil Pollut 

1991;56:769 –777. 

Lodenius M, Ari S, Antti U-R.Sorption and mobilisation of 

mercury in peat soil.Chemosphere 1983;12(11y12):1575 – 

1581. 

Madsen PP.Peat bog records of atmospheric mercury deposition.Nature 

1981;293:127 –130. 

Martinez-Cortizas A, Pontvedra-Pombal X, Garcia-Rodeja E, 

Novoa-Munoz JC, Shotyk W.Mercury in a Spanish peat 

bog: archive of climate change and atmospheric metal 

deposition.Science 1999;284:939 –942. 

Norton SA, Evans GC, Kahl JS.Comparison of Hg and Pb 

fluxes to hummocks and hollows of ombrotrophic Big Heath 

Bog and to nearby Sargent Mt.Pond, Maine, USA.Water 

Air Soil Pollut 1997;100:271 –286. 

Salvato N, Pirola C.Analysis of mercury traces by means of 

solid samples atomic absorption spectrometry.Microchim 

Acta 1996;123(1–4):63 –71. 

Shotyk W, Weiss D, Appleby PG, Cheburkin AK, Frei R, 

Gloor M, Kramers JD, Reese S, van der Knaap WO.History 

139 

of Atmospheric Lead Deposition Since 12,370 14C yr BP 

from a Peat Bog, Jura Mountains, Switzerland.Science 

1998;281:1635 –1640. 

Shotyk W, Goodsite ME, Roos F, Heinemeier J, Rom W, 

Appleby PG, Frei R, Van der Knaap WO, Cheburkin AK, 

Reese S,Biester H, Lohse C, Hansen TS.Atmospheric Hg, 

Pb and As in peat cores from Greenland and Denmark dated 

210 

using the 14C AMS bomb pulse technique and Pb: 

concentrations, natural and anthropogenic enrichments, and 

fluxes.Submitted to Environ Planetary Sci Lett, 2001. 

Wardenaar ECP.A new hand tool for cutting peat.Can J Bot 

1987;65:1772 –1773. 

Weiss D, Shotyk W, Schaefer H, Loyall U, Grollimund E, 

Gloor M.Microwave digestion of ancient peat and determination 

of lead by voltammetry.Fresenius J Anal Chem 

1999a;363:300 –305. 

Weiss D, Shotyk W, Appleby PG, Cheburkin AK, Kramers 

JD.Atmospheric Pb deposition since the Industrial Revolution 

recorded by five Swiss peat profiles: enrichment 

factors, fluxes, isotopic composition, and sources.Environ 

Sci Technol 1999b;33:1340 –1352.


Paper 8: Shotyk, W., Goodsite, M.E., Roos-Barraclough, F., Frei, R., Heinemeier, J., Asmund, G., 

Lohse, C., Hansen, T.S. Anthropogenic contributions to atmospheric Hg, Pb and As 

accumulation recorded by peat cores from southern Greenland and Denmark dated using 

the 14C “bomb pulse curve”. Geochimica et Cosmochimica Acta accepted 02 June, 2003.

Monday, June 2, 2003 

Anthropogenic contributions to atmospheric Hg, Pb and As 

accumulation recorded by peat cores from southern Greenland and 

Denmark dated using the 14 C “bomb pulse curve” 

W. SHOTYK 1* , M.E. GOODSITE 2# , F. ROOS-BARRACLOUGH 3$ , R. FREI 4 , J. HEINEMEIER 5 , 

G. ASMUND 6 , C. LOHSE 7 , T.S. HANSEN 7 

*1, Institute of Environmental Geochemistry, University of Heidelberg, INF 236, D-69120 

Heidelberg, GERMANY 

tel +49 (6221) 54 4803 fax +49 (6221) 54 5228 

email: shotyk@ugc.uni-heidelberg.de 

2, Department of Atmospheric Environment, National Environmental Research Institute of 

Denmark, Frederiksborgvej 399, P.O. Box 358, DK-4000 Roskilde, Denmark 

3, Institute of Geological Sciences, University of Berne, Baltzerstrasse 1-3, CH-3012 Berne, 

Switzerland 

4, Geological Institute, University of Copenhagen, Øster Voldgade 10, DK-1350 Copenhagen 

Denmark 

5, AMS 14C Dating Laboratory, IFA, Aarhus University, Denmark 

6, Department of Arctic Environment, National Environmental Research Institute, 

Frederiksborgvej 399, P.O. Box 358, DK-4000 Roskilde, Denmark 

7, Environmental Chemistry Research Group, Department of Chemistry, University of 

Southern Denmark, Odense University, Campusvej 55, Odense M, Denmark 

# present address (M.E.G.): Environmental Chemistry Research Group, Department of 

Chemistry, University of Southern Denmark, Odense University, Campusvej 55, Odense M, 

Denmark 

$ present address (F. R-B.): Chemical Analytical R&D, Cilag AG, Hochstrasse 201, CH-8205 

Schaffhausen, Switzerland 

first submitted to Geochimica et Cosmochimica Acta 17.9.02 

REVISED VERSION SUBMITTED 15. APRIL 2003; provisionally accepted 24.4.03 

EDITORIAL IMPROVEMENTS TO TEXT COMPLETED 2.6.03 ACCEPTED 2.6.03

ABSTRACT 

Mercury concentrations are clearly elevated in the surface and sub-surface layers of 

peat cores collected from a minerotrophic (“groundwater-fed”) fen in southern Greenland (GL) 

and an ombrotrophic (“rainwater-fed”) bog in Denmark (DK). Using 14 C to precisely date 

samples since ca. AD 1950 using the “atmospheric bomb pulse”, the chronology of Hg 

accumulation in GL is remarkably similar to the bog in DK where Hg was supplied only by 

atmospheric deposition: this suggests not only that Hg has been supplied to the surface layers 

of the minerotrophic core (GL) primarily by atmospheric inputs, but also that the peat cores have 

preserved a consistent record of the changing rates of atmospheric Hg accumulation. The lowest 

Hg fluxes in the GL core (0.3 to 0.5 µg/m 2 /yr) were found in peats dating from AD 550 to AD 

975, compared to the maximum of 164 µg/m 2 /yr in AD 1953. Atmospheric Hg accumulation 

rates have since declined, with the value for 1995 (14 µg/m 2 /yr) comparable to the value for 

1995 obtained by published studies of atmospheric transport modelling (12 µg/m 2 /yr). 

The greatest rates of atmospheric Hg accumulation in the DK core are also found in 

the sample dating from AD 1953 and are comparable in magnitude (184 µg/m 2 /yr) to the GL 

core; again, the fluxes have since gone into strong decline. The accumulation rates recorded by 

the peat core for AD 1994 (14 µg/m 2 /yr) are also comparable to the value for 1995 obtained by 

atmospheric transport modelling (18 µg/m 2 /yr). Comparing the Pb/Ti and As/Ti ratios of the DK 

samples with the corresponding crustal ratios (or “natural background values” for pre- 

anthropogenic peat) shows that the samples dating from 1953 also contain the maximum 

concentration of “excess” Pb and As. The synchroneity of the enrichments of all three elements 

(Hg, Pb, and As) suggests a common source, with coal-burning the most likely candidate. 

Independent support for this interpretation was obtained from the Pb isotope data ( 206 Pb/ 207 Pb 

= 1.1481 ± 0.0002 in the leached fraction and 1.1505 ± 0.0002 in the residual fraction) which 

2

is too radiogenic to be explained in terms of gasoline lead alone, but compares well with values 

for U.K. coals. In contrast, the lowest values for 206 Pb/ 207 Pb in the DK profile (1.1370 ± 0.0003 

in the leached fraction and 1.1408 ± 0.0003 in the residual fraction) is found in the sample dating 

from AD 1979: this shows that the maximum contribution of leaded gasoline occurred 

approximately 25 years after the zenith in total anthropogenic Pb deposition. 

3


Atmospheric pollution in the industrial regions of the northern hemisphere is 

increasingly recognised as a potential threat to many life forms in the Arctic (AMAP, 1998). The 

present state of the Arctic environment has been summarised as follows (Braune et al., 1999): 

1) some native groups are among the most exposed populations in the world to certain 

environmental pollutants; 2) the levels of organic and inorganic contaminants in birds and 

mammals may exceed thresholds associated with reproductive, immunosuppressive, and 

neurobehavioral effects; 3) mercury seems to be increasing in aquatic sediments and in marine 

mammals, and tends to accumulate in marine food webs. The most significant knowledge gap 

at the present time is the “lack of temporal trend information for most contaminants” (Braune 

et al., 1999). 

In contrast to the other "heavy metals" of environmental concern, Hg can be singled 

out as a truly global pollutant because 1) more than 95% of atmospheric Hg is in the vapour 

phase where it has a residence time of at least one year (Lin and Pehkonen, 1999) and can be 

transported thousands of kilometres (Schroeder and Munthe, 1998), and 2) it is the only metal 

which indisputably biomagnifies through the food chain, as inorganic forms of the metal are 

methylated by bacteria (Lindberg et al., 1987). Accumulation of Hg in the Arctic environment 

is of particular concern because of bioaccumulation of its methylated forms (Morel et al., 1998) 

and their potential toxicity (Fitzgerald and Clarkson, 1991). However, the relative importance 

of natural versus anthropogenic sources of Hg to the Arctic is poorly understood and there is an 

urgent need for long term records to quantify fluxes due to anthropogenic emissions. 

To quantify the effects of human activities of the atmospheric geochemical cycle of 

Hg, the natural variability in the Hg cycle must be known, and this can only be obtained using 

long-term records of Hg accumulation. Ombrotrophic peat bogs receive inputs exclusively from 

4

the air (Clymo, 1987). Boyarkina et al. (1980) suggested that peat cores from these types of bogs 

preserve a record of the changing rates of atmospheric Hg deposition. This view has been 

supported by independent experimental evidence (Oechsle, 1982), and by several subsequent 

studies of Hg concentrations in dated peat cores from temperate bogs (Pheiffer-Madsen, 1981; 

Jensen and Jensen, 1991; Norton et al., 1997; Benoit et al., 1997). The first long-term 

reconstruction of atmospheric Hg deposition was obtained using a Spanish bog which has been 

accumulating peat since since 4070 14 C yr BP; this study not only showed that anthropogenic 

sources have exceeded natural contributions for more than a millenium (Martinez-Cortizas et al., 

1999), but also that cold climate phases promote atmospheric Hg accumulation which has 

important implications for polar regions. More recently, a peat bog in Switzerland (Etang de la 

Gruère in the Jura Mountains) has been used to create a high-resolution reconstruction of 

atmospheric Hg deposition extending back 14,500 calendar years (Roos-Barraclough et al., 

2002): here, the natural range of Hg fluxes (0.3 to 8 µg/m 2 /yr) was found to be impacted both 

by Holocene climate change and volcanic emissions. Anthropogenic Hg was quantified using 

the natural range of Hg to Br to calculate “excess” Hg; this revealed the appearance of 

anthropogenic Hg beginning in the late Medieval Period. A subsequent study of a neighbouring 

peat bog in the Swiss Jura provided a comparable chronology of atmospheric Hg accumulation, 

even though the second bog (La Tourbière des Genevez) became predominately minerotrophic 

with increasing depth (Roos-Barraclough and Shotyk, 2003). Thus, the chemical weathering 

reactions which have been taking place in the basal sediment underlying the peat bog had not 

contributed significantly to the Hg inventory of the profile. The maximum rates of atmospheric 

Hg accumulation in the Swiss peat bog profiles was 29-43 µg/m 2 /yr, with higher rates 

consistently found in the forested bog (TGE). 

To date, there are no long term records of atmospheric Hg deposition for the Arctic, 

5

and the relative importance of natural versus anthropogenic sources to such remote areas has 

aroused controversy (Rasmussen, 1994; Fitzgerald et al., 1997). To address this problem, we 

measured Hg concentration profiles in mires from Greenland and Denmark which have been 

accumulating peat for more than 3000 years (Goodsite, 2000). In fact, the south Greenland site 

is sub-Arctic but, the core collected here is of special importance because the past 50 years of 

peat accumulation has been precisely dated with 14 C using the “atmospheric bomb pulse” 

(Goodsite et al., 2002); this offers the promise of a detailed reconstruction of recent changes. As 

the peat profile from Denmark was also age dated using the same, high precision approach 

(Goodsite et al., 2002), it was also included in our study for comparison. 

To compare with Hg, we have also studied Pb, which is transported primarily in the 

fine (sub-micron) aerosol fraction (Puxbaum, 1991). Several characteristics of Pb render this 

element a particularly useful tracer for studying atmospheric deposition of trace metals and 

organic contaminants. First, it is known to be well preserved in ombrotrophic bogs (Vile et al., 

1995, 1999; Shotyk et al., 1996, 1997; MacKenzie et al., 1997, 1998; Weiss et al., 1999a,b), with 

bogs recording Pb chronologies which are comparable with lake sediment archives (Bränvall et 

al., 1997; Cortizas et al., 1997; Farmer et al., 1997; Norton et al., 1997) and historical records 

of ancient Pb mining (Kempter et al., 1997; Kempter and Frenzel, 1999, 2000). Second, long- 

term changes in the environmental pollution record of this element in Europe is comparatively 

well established (e.g. Shotyk et al., 1998, 2001, 2003). Third, Pb has a number of radiogenic 

isotopes which can be used to “fingerprint” the predominant sources of atmospheric pollutants 

(Kober et al., 1999; Bollhöfer and Rosman, 2001; Flament et al., 2002; Reuer and Weiss, 2002). 

Finally, we also measured As as this element is commonly enriched in coal (Bouska, 1981; 

Valkovic, 1983; Swaine, 1990). 

METHODS 

6

Description of the peat deposits 

The two peatlands studied are distinctly different with respect to hydrology, 

geochemistry and trophic status: an acidic (pH 4) ombrotrophic bog in Denmark contrasts 

strongly with a circumneutral (pH 7) minerotrophic fen in Greenland. 

The peatland near the village of Tasiusaq in southern Greenland (GL) is a small, 

confined subarctic fen on the Narsaq peninusla (61 o 08.314` N, 45 o 33.703` W) of southern 

Greenland (Fig. 1). The average annual temperature and rainfall at Narsarsuaq airport (opposite 

the fjord) were 0.9 o C and 615 mm, respectively, from 1961 to 1990 (Danish Meteorological 

Institute, Technical Report 00-18). The site was cored in September 1999, towards the end of the 

growing season. Representative plant species and photos of the fen as well as the general area 

are given elsewhere (Goodsite, 2000). The maximum thickness of peat accumulation is ca. 1 m. 

The basal material is predominately clay, with aquatic plant species dominating the deepest 10 

cm of peat accumulation (W.O. van der Knaap, University of Berne, personal communication); 

ascending from this zone, the peat consists predominately of mosses. The fen developed between 

two small lakes, and may have formed through the terrestrialization of a shallow lake basin. A 

small brook, ca. 1m wide and 1 m deep, runs through the fen, connecting the lakes. 

Topographically the mire is in a valley between steep mountains - but there was no visual 

indication that the peat deposit may have formed from a landslide. The fen surface today is 

characterized by small hummocks (20 to 40 cm high), and the ground is very spongy. 

The bedrock geology of this part of S. Greenland belongs to the Qassiarsuk complex 

which consists of a sequence of alkaline silicate tuffs and extrusive carbonatites interlayered with 

sandstones and their subvolcanic equivalents (Andersen, 1997). This complex is located in a 

roughly E-W trending graben structure between the village of Qassiarsuk and the settlement of 

Tasiusaq in the northern part of the Precambrian Gardar rift. Uranium mineralisations 

7

corresponding to the alkaline igneous activities have been found, with the alkaline, high 

conductivity waters marked by anomalous U concentrations both in the aqueous phases and in 

the sediments (Armour-Brown et al., 1983). 

Storelung Mose in Denmark is a typical raised Sphagnum bog (55 o 15.38' N, 10 o 15.37' 

E) on the island of Funen (Fig. 1). The average annual temperature and rainfall were 8.1 o C and 

639 mm, respectively, from 1961 to 1990 (Danish Meteorological Institute, Technical Report 

99-5). The bog had a history of peat digging lasting through the end of WWII (Bent Aaby, 

Danish National Museum, personal communication): this fact had been established prior to core 

collection, but the condition was considered tolerable as this study emphasises the chronology 

of atmospheric Hg, Pb, and As deposition since AD 1950 (which is datable using the 14 C 

atmospheric bomb pulse procedure). With most of the peat bogs in Denmark having been 

destroyed by man, the number of remaining possible coring sites which are suitable for 

atmospheric deposition studies really is rather limited. The bog is situated in a rural farming 

community, and when viewed from the edge, is seen to rise above the landscape. The peat 

deposit is 5 to 6 m deep with a relatively flat microtopography. Today's bog bears visible scars 

of disturbance, with an overgrown cart road through the middle of the bog, and excavation pits 

close to the trail. The pits are now overgrown with mats of moss. The sampling site chosen was 

as far from visible signs of peat digging as possible, in the NE section of the bog. Samples were 

collected in October, 1999 from zones intermediate in relief between hummocks and hollows. 

Sample collection and preparation 

Coring sites were located with obvious peat accumulation and three (15 cm x 15 cm 

by approximately 100 cm) monoliths of peat were cored at each site using a Ti Wardenaar peat 

sampler (Wardenaar, 1987). Three cores (labelled A, B, and C) were taken at each site, with each 

ca. 1.5 m apart, and forming a triangle. "A" cores were cut into 3 cm slices by hand in the field, 

8

and porewaters were expressed by hand for subsequent chemical analyses. 

Chemical analyses of porewaters and the trophic status of the mires 

Pore water samples were kept cool. Upon returning to the lab, all water samples were 

filtered to exclude material greater than 0.2 :m using polysulfone membrane filters (Acrodisc, 

Gelman), and refrigerated. Anions (Cl - and Br - ) and cations (Na + , K + , Mg 2+ , Ca 2+ ) in the waters 

were measured using chemically suppressed ion chromatography with conductivity detection 

(Steinmann and Shotyk, 1997). Each of the porewater samples was analyzed in duplicate both 

for anions and for cations. 

The pH of the surface waters at GL is neutral to alkaline and the porewaters typically 

contain ca. 20 mg/l Ca 2+ ; this is typical of minerotrophic fens and indicates the importance of 

mineral dissolution reactions, both within the peatland and in the surrounding basin, which 

control the composition of the waters (Shotyk, 1988). In contrast to the GL fen, the pH of the bog 

waters from DK is acidic (pH 4) and the porewaters contain only ca. 2 mg/l Ca 2+ ; this is typical 

of ombrotrophic bogs, and shows that mineral dissolution reactions are quantitatively 

insignificant: this is true because of the limited supply of minerals for reaction with the surface 

waters of the bog. The low pH value and low concentrations of Ca 2+ in the porewaters, therefore, 

indicates that mineral matter is supplied to the DK core only from the atmosphere, and this must 

be true also of trace metals such as Hg and Pb. 

Measurement of Pb and other trace element analyses using XRF 

Peat cores were frozen within one day of sample collection, and shipped frozen to the 

lab in Berne. Individual slices from the “A” cores were partially thawed and a plexiglass 

template (10 x 10 cm) was used to allow the outer 2.5 cm of each slice to be trimmed away using 

an acid rinsed ceramic knife on a plastic cutting board; clean lab procedures were followed, 

cleaning the cutting board and knife with deionized water three times between each slice. The 

9

edges cut off the slices were dried overnight at 105 o C in a drying oven and milled in a 

centrifugal mill equipped with a Ti rotor and 0.25 mm Ti sieve (Ultracentrifugal Mill ZM 1-T, 

F. K. Retsch GmbH and Co., Haan, Germany). The milled powder from pieces of each slice was 

then manually homogenized prior to using the powder for further analysis. Selected major and 

trace elements (K, Ca, Ti, Cr, Mn, Fe, Ni, Cu, Zn, As, Br, Rb, Sr, Y, Zr and Pb) were measured 

using the Energy-dispersive Miniprobe Multielement Analyzer (EMMA) using Mo K$ as the 

exciting radiation. The lower limits of detection for the elements reported here, on a dry weight 

basis, are Mn 12, Fe 5, Cu 1, As 3.0, Se 0.4, Br 0.7, Rb 0.5, Sr 0.6, Y 1.0, Zr 1.5, Pb 0.4 and U 

2 µg/g. Calibration of the instrument using international, certified, standard reference materials 

(SRMs), and the accuracy and precision of the trace metal measurements, are described in detail 

elsewhere (Cheburkin and Shotyk, 1996; Shotyk et al., 2000). Using the EMMA XRF, As 

concentrations are calculated as ([Pb]+[As]) - [As] but this is acceptable considering the 

abundance of As in the surface layers of the DK core (see below). Moreover, As is used here 

only to compare its distribution with that of Hg and Pb. 

The XRF data from the 3 cm slices (core A) not only provides quantitative information 

about the concentrations of various elements in the peat profiles, but also offers a general 

indication of the trophic status of the sites. 

Ash contents were measured by combustion at 550 o C overnight. 

Measurement of Hg 

The second set of cores (GL and DK “B” cores) were used for Hg analyses and age 

dating. One cm slices were cut while frozen using a stainless steel band saw. The "zero" depth 

of the monolith was taken to represent the interface between the living plant material at the top, 

and the underlying dead plant matter (peat); this transition was typically 3 cm below the top of 

the core. Individual slices were subsampled using a stainless steel microcorer (ca. 16 mm ID) 

10

to recover ca. 2 cm 3 plugs from every cm down to 40 cm, and every 2 nd cm below 40 cm; these 

were air-dried overnight at room temperature in a Class 100 laminar flow clean air cabinet. 

Mercury concentrations were measured in these plugs using solid sample atomic absorption 

spectroscopy (Salvato and Pirola, 1996) with a LECO AMA 254 as described in detail elsewhere 

(Roos-Barraclough et al., 2002). The instrument was calibrated using liquid Hg working 

standards prepared from a Merck 1000 mg/l Hg standard solution. Every 12th measurement was 

a certified reference material, either NIST 1515 (Apple Leaves) or NIST 1547 (Peach Leaves). 

The average measured values were 45.0 ± 0.7 ng/g (n=11) for NIST 1515 (certified value 44 ± 

4 ng/g) and 32.1 ± 1.4 ng/g (n=15) for NIST 1547 (certified value 31 ± 7 ng/g). The Hg 

concentration profiles presented here (analyst: F. R-B.) are in excellent agreement with those 

reported previously (analyst: M.E.G.) for selected samples from the same cores (Goodsite, 2000). 

As an independent check on the data which are presented here, Hg concentrations were 

also measured in a complete set of subsamples at the Department of Arctic Environment, Danish 

National Environmental Research Institute (analyst: G.A.). Approximately 1 g fresh sample and 

4 ml concentrated Merck Suprapur nitric acid were added to Teflon bombs with stainless steel 

caps, which were then heated for 12 hours at 150/C. After cooling the dissolved samples were 

left uncovered until the majority of the nitrous oxides had evaporated. Before analysing for Hg, 

potassium permanganate solution was added until a permanent pink colour was obtained in order 

to maintain an oxidising environment and to prevent loss of Hg. The samples were then diluted 

with 18 MS water to approximately 25 g in polyethylene bottles. Mercury concentrations were 

determined following reduction with sodium borohydride in a flow injection AAS system 

(Perkin Elmer FIMS). The data set obtained using this procedure is in good agreement with the 

data which is presented in this paper, obtained using the LECO AMA 254 (on average, within 

15%, r 2 = 0.727, n=65). 

11

Age dating using 14 C 

Plant macrofossils identified in selected samples from each “B” core were 14 C dated 

by using Accelerator Mass Spectrometry (AMS). The macrofossils were taken from the centers 

of selected one-cm slices at the Institute of Plant Science, University of Berne, where they were 

also cleaned and dried at 60 /C. Within one week of selection they were processed at the AMS 

14 C Dating Laboratory, University of Aarhus, using a standard procedure for plant material 

(washed, acid-base-acid treatment). Individual samples more recent than AD 1950 were dated 

directly by comparing the measured 14 C concentrations in the samples with the atmospheric 

concentrations of 14 C recorded since the beginning of thermonuclear bomb testing; these tests 

greatly increased the amount of 14 C in the atmosphere, resulting in an “atmospheric bomb pulse” 

(ABP). The high resolution age dates obtained this way are ± 2 years. All details describing the 

AMS 14 C dating method and the application of the atmospheric bomb pulse, as well as a 

comparison with 210 Pb age dating of the same set of peat samples, is given elsewhere (Goodsite 

et al., 2002). Since the publication of that paper (Goodsite et al., 2002), approximately twice the 

number of samples has been age dated, with all results summarised here (Tables 1, 2). Selected 

samples from deeper layers (before AD 1950) were AMS 14 C dated by the usual tree-ring 

calibration method; here, the uncertainties are much greater (Tables 1, 2). 

In addition to the 14 C age dates obtained using AMS, the basal peat sample from each 

“A” core was dated using 14 C decay counting (Physics Institute, University of Berne) which 

yielded ages of 3540 ± 30 and 2790 ± 40 14 C yr BP for the bottom sample of the GL (78cm) and 

DK (84cm) cores, respectively. 

Stable Pb isotopes 

Powdered samples of peat were analysed at the Danish Center for Isotope Geology for 

stable lead isotopes ( 204 Pb, 206 Pb, 207 Pb, 208 Pb) in two fractions: weak acid leachates which are 

12

designed to recover predominantly atmospheric lead adhering to the plant material, and from the 

corresponding residues which primarily represents geogenic lead hosted in the inorganic, 

minerotrophic matrix, as well as atmospherically-derived soil dust (Table 3). Leaching of the 

peat material was performed with 2N HCl for two hours. The samples were then centrifuged and 

the supernatant was carefully pipetted off. The residues were attacked by 8N HBr for 1 day, then 

dried and subsequently re-attacked using an HF-HNO 3 mixture for 2 days in Savillex Teflon 

beakers. This procedure has been shown to be effective in dissolving both phosphates and 

silicates (Frei et al., 1997; Schaller et al., 1998). Lead was processed and separated over 0.5 ml 

glass columns charged with a 100 mesh AG-1 x8 anion exchange resin (BIORAD) and purified 

in a second clean-up over 200 µl Teflon columns containing the same resin. Liquid aliquots 

of both leachates and residues were doped with a 204 Pb spike for isotope dilution concentration 

measurements. Lead separates were loaded with a 1M H 3 PO 4 - silica-gel mix and measured 

from 20 µm Re-filaments on a VG-54 Sector-IT thermal ionization mass spectrometer at the 

Geological Institute, University of Copenhagen. Lead isotopes were analyzed in static multi- 

collection mode. Procedural blanks remained below 110 pg Pb: this is insignificant relative to 

the amount of Pb contained in the samples. Isotopic fractionation of Pb was monitored by 

repeated analyses of the NBS-SRM 981 Pb standard, and the measured data were corrected for 

mass bias using the values of Todt (1993). 

RESULTS 

Mercury concentrations in relation to ash contents and dry bulk density 

The large differences in ash contents illustrate the fundamental hydrological 

differences between the two peat profiles. In the GL core (Fig. 2a), ash concentrations are in the 

range of ca. 10 to 40% whereas the peat samples below 30 cm in the DK core typically contain 

ca. 2 % ash (Fig. 2b); these values are typical of minerotrophic and ombrotrophic peats, 

13

espectively (Naucke et al., 1993). There is an exceptional zone of elevated ash content in the 

DK core at ca. 18cm (Fig. 2b), which may reflect the disturbance of the bog surface by peat 

cutting during WWII. Bulk density values are generally higher in the GL core (Fig. 2a) compared 

to DK core (Fig. 2b). 

The Hg concentration profiles reveal elevated Hg concentrations in both the upper and 

lower sections of the GL core (Fig. 2b), but only in the upper section of the DK core (Fig. 2b). 

To take into account the large differences in bulk density within and between the two cores, the 

Hg concentrations are also expressed on a volumetric basis (Fig. 2). The variation in volumetric 

Hg concentrations above 30 cm show a remarkable similarity between the two cores, with a very 

intense peak in volumetric Hg concentration in the GL core at 21 cm (Fig. 2a) and in the DK 

core at 17 cm (Fig. 2b), respectively. The selected age dates shown in Fig. 2 indicate that the 

zone of greatest Hg concentration in each core dates from the 1950`s. The very old radiocarbon 

ages in the DK peat samples below 29 cm (which was dated at 2395 ± 45 14 C yr BP) is evidence 

that some part of the original peat surface has been lost due to peat cutting. These old sections 

of the DK profile will not be considered further. 

Mercury concentrations in ombrotrophic (DK) versus minerotrophic (GL) peat profiles 

The chronology of changes in volumetric Hg concentrations is remarkably similar at 

the two sites (Fig. 2). The DK core is ombrotrophic, therefore Hg was supplied to this profile 

exclusively by atmospheric deposition. The greatest Hg concentrations are found in samples 

dating from the 1950`s. As there are no known natural geochemical processes which could have 

led to this pronounced Hg enrichment, we assume that the elevated Hg concentrations in the 

surface layers, compared to deeper, older peats, reflect increased rates of atmospheric Hg 

deposition caused by industrial activities. Similarly, in the GL core the greatest Hg 

concentrations are found in samples dating from the 1950`s. In contrast to the DK core, the GL 

14

profile is minerotrophic, and the possible importance of Hg provided by mineral-water reactions 

has to be carefully considered. However, we know of no natural geochemical processes which 

could have enriched the surface of this core with Hg without also enriching the middle section 

of the core, between ca. 30 and 55 cm, which contain the lowest Hg concentrations. We further 

assume, therefore, that the elevated Hg concentrations in the surface layers of the GL core also 

can be attributed to atmospheric Hg inputs. Other possible causes of the changes in Hg 

concentrations in the GL core will be discussed later in the paper. 

Atmospheric Hg accumulation rates in southern Greenland 

The age dates given in Table 1 can be used to calculate an age-depth model to estimate 

peat accumulation rates (Fig. 3). This process identifies three regions, indicated by the three 

regression lines (Fig. 3). These regression lines serve as an age model from which the age of any 

depth can be deduced. In the ca. 3000 year period before AD 1950, the average peat 

accumulation rate was 0.019 cm/yr (Fig. 3a); from AD 1950 to ca. 1976, the accumulation rate 

was 0.68 cm/yr, and since ca. 1976 the rate has been 0.20 cm/yr (Fig. 3b). The age depth model 

allows the atmospheric Hg accumulation rate to be estimated as the product of the volumetric 

Hg concentrations (ng/cm3) and the peat accumulation rate (cm/yr). Strictly speaking, the 

atmospheric Hg accumulation rate calculated in this way is equal to the depositional flux minus 

re-emission of Hg from the peatland surface. Experimental studies using peat (Lodenius, 1983) 

have shown that Hg re-emissions are small (ie. < 0.01% of total Hg added as 203 Hg ) so that the 

atmospheric Hg accumulation rates shown here (in units of :g/m 2 /yr versus age in Figure 4) are 

comparable to the depositional fluxes. These results show that the net, pre-industrial Hg 

accumulation rate ranged from 0.3 to 3 :g/m 2 /yr (Fig. 4a) which is comparable to the range (0.3 

to 8 :g/m 2 /yr) obtained from ca. 12,500 BC to 1300 AD using a Swiss peat bog profile (Roos- 

Barraclough et al., 2002). The graph shows further that the net Hg accumulation rate prior to AD 

15

200 was in the range 1 to 3 :g/m 2 /yr, whereas samples from AD 550 to AD 975 were in the 

range 0.3 to 0.5 :g/m 2 /yr (Fig. 4a). While some part of this variation in pre-industrial 

accumulation rates may be due to Holocene climate change (Martinez-Cortizas et al., 1999; 

Roos-Barraclough et al., 2002), the elevated Hg concentrations in samples below 55 cm (dated 

at AD 938 ± 48 (Fig. 2b) are found in minerotrophic peats, and we cannot yet separate the Hg 

concentrations into atmospheric, aquatic, and terrestrial components. However, it is clear that 

the pre-industrial, net rates of Hg accumulation in GL were in the range 0.3 to 3 :g/m 2 /yr (Fig. 

4a), placing an upper limit on the natural atmospheric Hg flux. 

In contrast to the “natural background” Hg accumulation rate, the flux in GL reached 

164 :g/m 2 /yr in 1953 (Fig. 4b). The value in 1995 (14.1 :g/m 2 /yr) is an order of magnitude 

lower, but still exceeds by a wide margin the natural range in net Hg accumulation rate shown 

in Fig. 4b. The value obtained here for 1995 is in good agreement with the Danish Eulerian 

Hemisphere model calculations (12.0 :g/m 2 /yr in 1995) published for South Greenland 

(Christensen et al., 2002), supporting the approach used here to estimate the atmospheric Hg 

fluxes. The error associated with the Hg fluxes (Fig. 4) is calculated to be 21%, based on 

conservative estimates of the errors associated with the 14 C bomb pulse curve age dates (ca. 5%), 

Hg concentrations (ca. 5%), and bulk density measurements (ca. 20%). Even after the upper and 

lower limits on the flux data are added to Fig. 4, the temporal trends in net atmospheric Hg 

accumulation rates are clear. 

Atmospheric Hg accumulation rates in southern Denmark 

While the DK core cannot be used to model the age-depth relationship for samples 

older than WWII, the age dates (Table 2) yield an average peat accumulation rate from AD 1950 

to AD 1980 of 0.47 cm/y, and since AD 1980 of 0.21 cm/yr (Fig. 3d). Using these peat 

accumulation rates, the net atmospheric Hg accumulation rate can be calculated and is shown 

16

in Fig. 4c. This graph shows that the maximum net Hg accumulation rate was 184 :g/m 2 /yr in 

AD 1953 and that the flux has since gone into a strong decline. The value obtained from this peat 

core for 1994 (14 :g/m 2 /yr) is comparable to the Danish Eulerian Hemisphere model calculations 

(18 :g/m 2 /yr in 1995) published for Denmark (Christensen et al., 2002). 

The elevated ash content in the DK core which reaches a maximum at 15-16 cm (Fig. 

2b), is probably a consequence of disturbance to the peat profile; unfortunately, this layer is 

adjacent to the zone of maximum Hg concentration at 16-17 cm (Fig. 2b). It is difficult to 

determine what effect this disturbance may have had on the Hg concentration profile, the peat 

growth rate and therefore the Hg accumulation rate. However, the maximum Hg accumulation 

rates presented here (Fig. 4) are comparable to those published using other Danish peat cores by 

Pheiffer-Madsen (1981). In addition, Hg and Br concentrations were measured in a peat core, 

which we collected in the spring of 2000 from Store Vildmose, a second ombrotrophic bog in 

Denmark. In the Store Vildmose core, the “background” Hg/Br ratio is 0.32 x 10 -3 ± 0.12 x 10 -3 

(average of 41 samples below 20 cm) and the maximum Hg/Br ratio (3.05 x10 -3 ) found from 6 

to 7 cm exceeds this value by 9.6 times (data not shown). In the Storelung Mose profile 

described in this paper, the maximum Hg/Br ratio (3.73 x10 -3 ) exceeds this background ratio by 

11.7 times. Thus, the changes in Hg/Br at Storelung Mose, despite the disturbance by peat 

cutting in the past, are comparable in magnitude to the changes recorded by the peat core 

collected from Store Vildmose (Shotyk, unpublished data). In addition, the maximum Hg 

concentration in the Store Vildmose core (348 ng/g) and the maximum in Hg/Br (3.05 x10 -3 ) 

were dated to AD 1953 using 210 Pb; this agrees remarkably well with the age date of the peak in 

Hg concentration at Storelung (AD 1953) which was determined using the bomb pulse curve of 

14 C (Fig. 2b). In summary, the Hg concentration data presented here for the Storelung bog, as 

well as the Hg/Br ratios and the chronology of Hg accumulation, are comparable to our 

17

unpublished data for these parameters from the Store Vildmose bog. 

Comparison of Pb and As concentrations and enrichments, GL versus DK 

Lead and As concentrations are much higher in the DK core compared with GL (Fig. 

5a). However, the maximum Pb concentrations in the DK core are very similar to the maximum 

values found in the Draved Mose peat profile described by Aaby and Jacobsen (1978). To 

emphasize the difference in Pb and As between the GL and DK profiles, and to take into account 

the differences in abundance of mineral material, Enrichment Factors (EF) were calculated as 

EF = ([M]/[Ti]) peat / ([M]/[Ti]) Earth`s crust 

where [M] refers to the total concentration of Pb or As measured in the peat sample (µg/g) and 

[Ti] to the total concentration of Ti. Although Sc is the preferred reference element for these 

calculations (Shotyk et al., 2001), Sc concentration data is not yet available for these two peat 

cores. The EF indicates the extent of elemental enrichment in the peats, relative to the abundance 

of that metal in the Earth`s Upper Crust (Wedepohl, 1995) where Pb = 14.8, As = 1.7 and Ti = 

4010 µg/g. The calculated Enrichment Factors (Fig. 5a) show that these elements are enriched 

in the DK core up to 65x (Pb) and 85x (As). In contrast, the GL core reveals no significant 

enrichment of Pb, and only very slight enrichments of As. Given that the As concentrations in 

the GL profile are generally at or below the lower limit of detection by XRF (3 µg/g), any record 

of anthropogenic As in the GL profile cannot be discerned using the total concentrations of As 

as measured by XRF, and its ratio to Ti. Similarly, in the GL core the Pb concentrations are 

comparatively low and concentrations of lithogenic trace elements such as Ti and Zr are 

comparatively high; thus, using total metal concentrations it is not possible to discern any 

significant impact of anthropogenic Pb. In the DK core, just the opposite is true: concentrations 

of Pb and As are high but the concentrations of lithogenic elements (Ti, Zr) are comparatively 

low. Thus, in the DK core, enrichments of Pb and As, relative to their crustal abundance, are 

(i) 

18

clearly seen (Fig. 5a). 

Changing atmospheric Pb fluxes in Denmark 

In the DK core, all of the Pb was derived from the atmosphere. Given the immobility 

of Pb in ombrotrophic peat bog profiles (Shotyk et al, 1997, 1998, 2001, 2003; Weiss et al., 

1999a,b and references cited therein), the atmospheric Pb flux can be calculated (Fig. 5b) using 

the average peat accumulation rates, the Pb concentrations, and the bulk density data. This flux 

can be further separated into “lithogenic” and “anthropogenic” components using Ti as an 

indicator of the concentration of lithogenic-derived aerosols supplied by rock weathering: 

[Pb] lithogenic = [Ti] sample x [Pb/Ti] Earth`s Crust 

Notice that for most samples, the total Pb flux and anthropogenic Pb flux are nearly identical 

because the concentrations of lithogenic Pb are so low. Both the atmospheric Pb flux in DK (Fig. 

5b) and the relative importance of anthropogenic Pb(Fig. 5c) have been declining since the 

1950`s (Fig. 5b). 

Isotopic composition of Pb in peat from GL and DK 

The concentrations of Pb measured in the leachable and residual fractions, as well as 

the isotopic composition of this Pb (summarized as the 206 Pb/ 207 Pb ratio), are shown in Figure 

6. The large difference in total Pb concentrations between the DK and GL cores described earlier 

(seen in the “A” cores which were measured using XRF and presented in Fig. 5a) is also seen 

in the “B” cores where Pb concentrations were measured in selected samples using IDMS (Fig. 

6a, b). However, the leaching study undertaken using TIMS indicates two additional features. 

First, the abundance of “leachable” Pb tends generally to be more abundant than the “residual” 

Pb (except for two samples from GL). In the DK core in particular, the predominance of 

leachable Pb is most clearly associated with the samples containing the highest total Pb 

concentrations (Fig. 6a), with the highest concentration found at 15-16cm (141.4 µg/g Pb in the 

(ii) 

19

leachable fraction, and 32.1 µg/g Pb in the residual fraction). For comparison with these 

concentrations, the “natural background” concentration of Pb in pre-anthropogenic peats in 

Switzerland dating from ca. 8,000 to 5,000 14 C yrs BP is approximately 0.2 µg/g Pb, and this 

difference indicates in a general way the extent to which the DK core has been contaminated by 

industrial Pb. The isotopic composition of Pb in the two fractions, summarized here as the 

206 Pb/ 207 Pb ratio (Fig. 6b), is virtually identical and this clearly demonstrates a common origin 

of the Pb in both fractions. One possible explanation of the similarity in Pb isotopic composition 

in each of the DK peat fractions is that anthropogenic Pb supplied by various industrial 

emissions has been scavenged by soil-derived aerosols supplied by crustal weathering, such that 

the “soil dust signature” has been overprinted by anthropogenic Pb. A second possible 

explanation is that the “residual” fraction obtained by the extraction procedure has included 

some of the “leachable” Pb supplied primarily by anthropogenic sources. 

Second, the isotopic composition of Pb in the two peat cores is very different, with the 

GL samples being far more radiogenic (Fig. 6b). The pattern in 206 Pb/ 207 Pb in the DK core is 

remarkably similar to the temporal variation in 206 Pb/ 207 Pb seen in four peat profiles from 

Switzerland (Weiss et al., 1999a), as well as the record of 206 Pb/ 207 Pb reported for Sphagnum 

moss samples from the University of Geneva herbarium (Weiss et al., 1999b). The ratio 

206 Pb/ 207 Pb in the Upper Continental Crust (Kramers and Tolstikhin, 1997) as well as in pre- 

anthropogenic (dating from 8,000 to 5,300 14 C yr BP), atmospheric aerosols from Switzerland 

(Shotyk et al., 2001), is approximately 1.2. All of the measured values from the DK core are 

significantly less radiogenic than this, indicating that anthropogenic Pb has dominated the 

atmospheric Pb inputs to the DK bog. 

Predominant sources of anthropogenic, atmospheric Pb and As in Denmark 

The lithogenic Pb fraction derived from atmospheric soil dust can be estimated as the 

20

product of the Ti concentrations of the DK profile, and the Pb/Ti ratio of the Earth`s Crust 

(Wedepohl, 1995); this assumes that lithogenic Pb is derived exclusively from atmospheric soil 

dust, and that the Pb/Ti ratio of this dust is similar to that of the Earth`s Crust. Using the Ti 

concentrations which are available for the top 20 cm of the “B” cores, the “lithogenic” Pb 

component has been calculated (Fig. 6c) and in all cases, it is small even compared to the 

“residual” Pb fraction measured using TIMS. The lithogenic Pb fraction of the DK peat profile 

has also been estimated using Zr as the conservative, lithogenic reference element, but the 

outcome is much the same (Fig. 6c). 

One disadvantage of this approach is that the Pb/Ti and Pb/Zr ratios of atmospheric 

soil dust may not be identical to the corresponding values for the Earth`s Upper Crust (UC) as 

reported by Wedepohl (1995). In an earlier paper of a Swiss peat bog profile, we found that the 

“background” Pb/Sc ratio in pre-anthropogenic aerosols dating from ca. 5,000 to 8,000 14 C yr 

BP was approximately 4x the value presented by Wedepohl (1995) for the UC (Shotyk et al., 

1998). To take into account these findings, we have also calculated the concentrations of 

“lithogenic” Pb for the DK core using “background” values of Pb/Ti = 4x UC (Fig. 6c). 

However, even taking the value of Pb/Ti = 4x UC, the theoretical “lithogenic” Pb component in 

many samples is still ca. one-half of the concentrations measured in the “residual” fraction. The 

differences between the “residual” fraction measured using TIMS and the “lithogenic” fraction 

calculated using Pb and Ti (or Zr) concentrations indicate that further studies are needed to better 

quantify the isotopic composition of the “natural” Pb component of peats which are impacted 

by anthropogenic Pb. 

Using the concentrations of Ti, Pb, and As, it is possible to estimate the contribution 

of anthropogenic Pb and As to the inventories of these elements in the DK peat core, because the 

“lithogenic” Pb or As fraction is small compared to the total concentrations. Taking Ti as an 

21

indicator of the concentration of aerosols supplied by rock weathering, the concentration of Pb 

(or As) which was supplied to the bog via atmospheric deposition of soil-derived aerosols can 

be estimated using Eqn (ii). Here, we have calculated the “lithogenic” Pb component using both 

Pb/Ti = UC and Pb/Ti = 4x UC (Fig. 6c). With respect to As, we recently reported As/Sc ratios 

for pre-anthropogenic, atmospheric aerosols (for the same Swiss bog, and the same period of 

time as for Pb/Sc) which are approximately 10x the ratio for the UC (Shotyk et al., 2002b); here, 

therefore, we have calculated “lithogenic” As using both As/Ti = UC and As/Ti = 10x UC (Fig. 

6c). Once the lithogenic Pb (or As) component has been quantified, anthropogenic Pb (or As) 

is calculated as 

[Pb] anthropogenic = [Pb] total - [Pb] lithogenic 

The calculated values for lithogenic and anthropogenic Pb and As calculated in this way are 

shown in Figure 6c, along with a graph of the ratio 206 Pb/ 207 Pb, and selected 14 C age dates 

(Table 2). Notice that both sets of calculations for each element indicate that the anthropogenic 

component dominates the Pb and As inventories in this section of the DK core. 

The maximum concentration of anthropogenic Pb and As is found in peats dating from 

1954; this matches the maximum flux of atmospheric Hg (Fig. 2b and Fig. 4c), and suggests a 

common source: this is most likely coal, as coal is commonly enriched in all three of these 

elements (Bouska, 1981; Valkovic, 1983; Swaine, 1990); Zn (not shown) is also clearly enriched 

at this depth. The data presented here (Figs. 2b, 4c, and 6c) suggest that the maximum extent of 

atmospheric Pb, As, and Hg contamination in Denmark had already peaked in 1954, and has 

more or less declined ever since. It is noteworthy that the Clean Air Act was passed in the U.K. 

in 1956, primarily to reduce the industrial emission of particulates and gases from burning coal. 

Given that the industrial heartland of the U.K. is nearly due west of Denmark and that the 

predominant wind direction is from west to east, British coal burning is likely to be an important, 

(iii) 

22

if not predominant source of these contaminants. According to the compilation published by 

Farmer et al. (1999), British coal consumption peaked in the early 1950`s. Taken together, it 

appears that the changing accumulation rates of anthropogenic Hg, Pb and As revealed by the 

DK peat core may reflect to a large extent the history of the British coal industry and the 

chronology of the Second Industrial Revolution. 

To further evaluate the possible link between these contaminants and British coals, the 

ratio 208 Pb/ 206 Pb has been plotted against the 206 Pb/ 207 Pb values (Fig. 7). Clearly, the DK peat 

samples are much closer in isotopic composition to the values for British coals (Farmer et al., 

1999) than they are to the values for U.K. gasoline leads (Monna et al., 1997). Thus, the Pb 

isotope data, combined with the chronology of the enrichments in Hg, Pb and As concentrations, 

suggest that coal burning was the main source of these contaminants to the DK peat profile. 

In samples above ca. 20 cm in the DK profile, there is a sharp shift (after AD 1959) 

toward much less radiogenic values 206 Pb/ 207 Pb (Fig. 6c) which reflects the growing importance 

of leaded gasoline contributions; the ores used to synthesize alkyllead compounds for gasoline 

such as Pb from the Broken Hills mine of Australia, have had 206 Pb/ 207 Pb values as low as 1.04. 

The most recent sample in the profile (AD 1999) shows significantly more radiogenic values, 

compared to the 1970`s, clearly reflecting the reduction in gasoline Pb concentrations, and the 

eventual phasing-out of leaded gasoline in Europe; this change is also seen in four peat profiles 

from Switzerland (Weiss et al., 1999a, Shotyk et al., 2003), as well as the record of 206 Pb/ 207 Pb 

reported for Sphagnum moss samples from the University of Geneva herbarium (Weiss et al., 

1999b). It is important to emphasise that the decline in anthropogenic Pb concentrations 

(starting in 1954) which is shown in Figure 6c pre-dates by approximately 25 years the minimum 

in 206 Pb/ 207 Pb; the minimum in this Pb isotope ratio corresponds to the maximum impact in 

gasoline Pb emissions (Shotyk et al., 2002a). As noted earlier, this decline in 206 Pb/ 207 Pb was 

23

caused by the introduction of gasoline leads of Australian origin with very low 206 Pb/ 207 Pb 

values. The minimum 206 Pb/ 207 Pb value (in 1979) is consistent with production records which 

indicated leaded gasoline consumption in Europe reached its zenith in 1980 (Hagner, 2000). 

Moreover, direct air Pb measurements in Denmark document strong declines in Pb 

concentrations since the late 1970`s (Jensen and Fenger, 1994). However, there appears to be no 

published air Pb data older than ca. 1970. The Pb isotope results presented in Fig. 6c are 

significant, as they show that the greatest percentage of anthropogenic Pb recorded by the peat 

profile in DK was not caused primarily by leaded gasoline consumption, but rather from coal- 

burning and other industrial sources. Moreover, these results show that anthropogenic Pb went 

into decline well before either the introduction of maximum allowable Pb concentration in 

gasoline (in 1978 in the EU according to Hagner, 2000), or the ban on leaded gasoline sales in 

the EU in 1987 (Hagner, 2000). While leaded gasoline certainly contributed to a pronounced 

shift in the isotopic composition of Pb-bearing aerosols, other sources of anthropogenic Pb were 

quantitatively more important in DK during the first half of the 20 th century. Similar conclusions 

were drawn from a recent study of duplicate peat cores from a Swiss bog which showed a 

maximum in Pb EF pre-dating the minimum in 206 Pb/ 207 Pb (Shotyk et al., 2003). 

DISCUSSION 

Natural sources of Hg to the minerotrophic GL core 

The elevated Hg concentrations in the uppermost layers of the GL core are effectively 

contemporaneous with those of the DK core (Fig. 2) which has received Hg solely from 

atmospheric deposition. Given the fundamental differences in the hydrology and geochemistry 

of the two sites (ombrotrophic, acidic bog in DK, minerotrophic, alkaline fen in GL), and the 

differences in peat accumulation rates (Fig. 3), it is unlikely that natural processes could have 

caused such a similar chronology of atmospheric Hg. The most likely explanation for the 

24

elevated Hg concentrations in the surface layers of the GL profile, therefore, is the changing 

rates of anthropogenic emissions of Hg to the atmosphere during the past century. However, the 

possible importance of other sources of Hg to the GL core also require consideration. 

atmospheric soil dust 

The concentrations of mineral matter in peat profiles may vary because of temporal 

changes in atmospheric soil dust deposition rates, organic matter decomposition, or both. 

Titanium, Zr, and Y are conservative, lithogenic elements in the sense that their oxides and 

silicates are resistant to chemical weathering, and their distribution in the peat profiles reflects 

the abundance of mineral material (Fig. 8a). Rubidium is found primarily in K feldspar where 

it substitutes for K, and its distribution in the profile generally resembles that of Zr and Y (Fig. 

8a); Rb too, therefore, provides an indication of the abundance and distribution of mineral matter 

in the profile. In the DK profile, the measured values of these elements are in a range typical of 

ombrotrophic bogs which receive minerals exclusively from atmospheric soil dust (Shotyk et al., 

2001). In contrast, these elements are typically 5x more abundant in the minerotrophic GL 

profile (Fig. 8a) core. Assuming that the lowest concentrations of Hg in the GL profile (1.6 

ng/cm 3 at 47 cm) represent pre-Industrial, “background” values, the maximum concentration 

(24.0 ng/cm 3 at 21 cm) exceeds this value by 15x. This concentration difference greatly exceeds 

the variation in Zr concentrations over the same distance (1.6x as shown in Fig. 8a). Therefore, 

the variation in mineral matter supply (e.g. from changes in the flux of atmospheric soil dust) is 

not a viable explanation of the magnitude of the variation in Hg concentrations with depth. 

marine aerosols 

The porewaters of the GL and DK cores typically contain ca. 10 mg/l Cl - , compared 

to continental ombrotrophic peat bogs from Switzerland which average ca. 0.3 mg/l Cl - 

(Steinmann and Shotyk, 1997). While the elevated chloride concentrations in the GL and DK 

25

porewaters reveals the influence of sea salt spray at both locations, and while this may be a 

natural source of Hg to the cores, the relative importance of this source must have been relatively 

constant over time. Thus, atmospheric deposition of marine aerosols cannot explain the change 

in Hg concentrations with respect to depth and time seen in both peat profiles (Fig. 2). 

aquatic inputs via chemical weathering of local soils and rocks 

The low abundance of Ca and Sr (Fig. 8b) in the DK core (average 0.40 % and 25.7 

µg/g, respectively) is typical of ombrotrophic bogs (Shotyk et al., 2001) which receive inputs 

exclusively from the atmosphere. In contrast, the higher Ca and Sr concentrations in the GL core 

(average 2.2 % and 347 µg/g, respectively) demonstrate extensive rock-water interaction 

characteristic of minerotrophic mires (Fig. 8b). Similarly, Mn and Fe are much more abundant 

in the GL profile (Fig. 8b). The relatively high pH of the waters combined with the abundance 

of Ca and Sr in the peats indicates active dissolution of carbonate minerals, either in the 

sediments underlying the peat, in the watershed, or both (Shotyk, 2002). The elevated Mn 

concentrations in the surface layers of both cores may either be due to plant uptake and 

recycling, oxidation, or both. Iron concentrations are very high in the GL profile, and are 

strongly enriched in the uppermost layers: this is most likely a redox-related transformation 

(Steinmann and Shotyk, 1997) as the Fe concentrations are well in excess of the concentration 

required by growing plants. Like Ca and Sr, the Mn and Fe which have been supplied to the GL 

profile may originate in carbonate mineral phases in the surrounding rocks and underlying 

sediments (Andersen, 1997). However, between 47 and 21 cm where the Hg concentrations 

increase by 15x, there is no change in the Sr concentrations (Fig. 8b). Thus, weathering of local 

rocks and soils cannot explain the increasing Hg concentrations in the uppermost, recent layers 

of the GL peat profile. 

Natural geochemical processes and their effects on Hg in the minerotrophic GL profile 

26

decomposition of organic matter 

The bulk density of the samples at 47 and 21 cm in the GL profile differ only by a 

factor of 1.7x (Fig. 2a), in contrast with the differences in Hg concentrations (15x). We assume 

that the bulk density of a given type of peat is a reflection of the extent of its decomposition 

(degree of humification). Support for this assumption comes from a detailed spectroscopic 

characterisation of organic matter in the Swiss peat bog profile at Etang de la Gruère (Cocozza 

et al., 2003). Bulk density, therefore, helps to take into account any changes in metal 

concentrations which take place during the decomposition of organic matter. In a recent paper 

about Hg accumulation rates in peat cores from Patagonia, Biester et al. (2003) suggested that 

bulk density is not an adequate parameter to express changes in peat humification, and that Hg 

accumulation rates (calculated as we have described here) should be corrected for humification 

to take into account mass loss during decay. However, in the GL peat profile, the difference in 

bulk density at 47 versus 21 cm (1.6x) is similar to the difference in Zr concentrations (1.7x). 

Zirconium is a conservative element which resides almost exclusively in zircon, and this element 

(along with other conservative trace metals) should increase in concentration with increasing 

extent of peat decay. In fact, the differences in Zr concentrations are comparable to the 

differences in bulk density; taken together these data suggest that the differences in humification 

alone cannot explain differences in Hg concentrations by more than a factor of two. As a 

consequence of these data and arguments, it is unlikely that physical processes such as organic 

matter decomposition can explain the magnitude of the variation in Hg concentrations with depth 

in the GL profile. 

redox-related processes in the oxic zone 

It has been suggested that redox-related transformations of Fe and Mn in marine and 

lacustrine sediments may contribute to Hg enrichments in the surface layers of these sediments. 

27

For example, Hg may become adsorbed onto Fe and Mn oxides which are formed when Fe (II) 

and Mn (II) in the anoxic zone diffuse upward and become oxidized (Gobeil et al., 1999). 

Asmund and Nielsen (2000) have summarised some of these studies, and document very strong 

correlations between Mn and Hg or Fe and Hg in some lake sediment cores. While peatlands do 

not have an overlying water column for Mn and Fe to diffuse into, they may have a shallow oxic 

zone, depending on the depth to water table which varies seasonally (Shotyk, 1988). The shapes 

of the Mn and Fe concentration profiles in the GL core (Fig. 8b) suggest that there may be some 

oxidation of Fe and Mn in the surface and near surface layers. To help evaluate the possible 

importance of this process on the Hg concentration profile, samples from the uppermost 20 cm 

of the “B” cores (1 cm slices) were also measured for trace elements, including Mn and Fe, using 

XRF (Fig. 9). Except for the elevated Mn concentration in the top layers at DK, the DK core 

reveals no peak in either element. In contrast, the GL core shows a pronounced enrichment of 

Mn at 12-13 cm, and Fe at 15-16 cm (arrows in Fig. 9). However, these Mn and Fe peaks are 

well above the peak in volumetric Hg concentrations which is found in the GL core between 20 

and 25 cm (Fig. 2a). The elevated Hg concentrations beginning at ca. 25 cm at GL, therefore, is 

independent of those of Mn and Fe, suggesting that redox-related transformations of Mn and Fe 

have not noticeably affected Hg. While redox-related transformations of Mn and Fe are probably 

operating within the peat profile, there has been no observable effects of these changes on the 

Hg profile. The elevated Hg concentrations in the near surface layers, therefore appears to be 

mainly influenced by atmospheric deposition. Thus, even in the minerotrophic peat profile at GL, 

the Hg concentration profile appears to provide a record of the changing rates of atmospheric Hg 

deposition. 

redox-related processes in the anoxic zone 

Mercury is highly enriched in all of the peat layers below 50 cm of the GL peat profile, 

28

compared to the lowest concentrations in the middle of the core (Fig. 2a). As these samples are 

more than one thousand years old, natural geochemical processes must be invoked to explain the 

Hg enrichments. Copper concentrations are also elevated in the deeper layers of the GL core, and 

the Cu/Y ratio illustrates the pronounced enrichment of this metal with increasing depth (Fig. 

10a). Similarly, U concentrations reach more than 100 µg/g (Fig. 10a) which is ca. 50x crustal 

abundance and 2000x the concentration found in ombrotrophic peats. Again, normalising the U 

concentrations to Y, the U/Y ratio reveals the shape of the U enrichment (Fig. 10a). Both Cu and 

U are commonly enriched in anoxic, minerotrophic peats due to reductive dissolution Cu(II) to 

Cu(0) and U(VI) to U(IV), respectively (Shotyk, 1988). The zone of maximum Cu and U 

enrichment is found in the limnic peat layer made up predominately of the remains of aquatic 

plants just above the basal sediment; previous studies also revealed enrichments of Cu and U in 

this zone of some Canadian bogs (Shotyk et al., 1992). 

Bromine and Se are both supplied to these peatlands from sea salt spray, but the Se/Br 

ratio (Fig. 10b) indicates that the deeper peat layers at GL are preferentially enriched in Se. 

While the variations in Br concentrations can largely be explained in terms of the decay of 

organic matter, this process alone cannot explain the large increase in Se concentrations with 

depth through the GL profile. The general shape of Se (Fig. 10b), however, resembles that of U 

and Cu/Y (Fig.10a), suggesting that a redox-related process has contributed to the Se enrichment. 

The pronounced enrichments of Cu and U (as evident by their ratios to Y) and Se (revealed by 

comparison to Br) at the same depths as the enrichment of Hg (Fig. 2b) suggest that the natural 

Hg enrichments in the deeper peat layers of the GL profile are related to redox transformations. 

The details of this process, however, are unclear. Given that the solubility product constant of 

Hg selenide (K sp = 10 -62 ) is orders of magnitude lower than even that of Hg sulphide (K sp = 10 - 

52 ), it may be that Se has played a direct role in the natural enrichment of Hg toward the bottom 

29

of the GL profile. 

Comparison with marine and lacustrine sedimentary records of Hg accumulation 

Age dated lake sediments from across the Arctic have been measured for Hg and the 

ratio of post-industrial to pre-industrial Hg accumulation rates were found to be in the range 0.7 

to 9 (Landers et al., 1998). The authors admit, however, that their study was hampered in many 

cases by local metal sources contributing to high background values: this is a fundamental 

problem with lake sediment archives, as metals are supplied by atmospheric as well as non- 

atmospheric sources. In cases where “background” values are elevated due to natural, 

geochemical sources, the additional impact of anthropogenic Hg may be difficult to detect. 

Asmund and Nielsen (2000) studied Hg in marine sediments from Greenland, and found that Hg 

fluxes had increased only by a factor of two, compared to the values for sediments from the 

previous century. Bindler et al. (2001) examined many lake sediment profiles from Greenland, 

and reported concentration increases of only 2 to 3 x from pre-industrial to post-industrial times. 

In each of these cases, however, the measured Hg accumulation rates reflect not only 

atmospheric inputs, but also inputs to the sediments from the entire watershed (from physical and 

chemical weathering) as well as focusing processes within the sedimentary basin. In contrast, 

the GL peat profile studied here provides the first long-term record of atmospheric Hg 

accumulation in Greenland. Because the surface peat layers receive Hg only from the air, it 

provides a much more sensitive record of changes in atmospheric deposition. This sensitivity is 

revealed by the difference between the background fluxes which are as low as 0.3 to 0.5 

:g/m 2 /yr (AD 550 to 975) and the maximum flux of 164 :g/m 2 /yr (AD 1953). In sediment 

records, these extreme values are masked by the continuous input of Hg from non-atmospheric 

sources. In fact, peat bogs are probably the most sensitive continental archives of atmospheric 

Pb and Hg deposition, and may serve as archives of many of other trace elements of 

30

contemporary environmental interest. 

Implications for the global atmospheric Hg cycle 

A model of the global atmospheric Hg cycle has recently been published (Lamborg 

et al., 2002) in which an annual pre-industrial annual flux of Hg to the continents of 4 Mmoles 

was reported. If this value is taken to represent the continental land mass (total 147 x 10 6 km 2 ), 

then an average, pre-industrial flux of 5.5 :g/m 2 /yr is implied. For comparison, the minimum 

pre-industrial flux recorded by the GL core is only 0.3 :g/m 2 /yr; this is also the minimum Hg 

accumulation rate recorded by the Swiss peat bog (Roos-Barraclough et al., 2002). In fact, the 

value reported by Lamborg et al. (2002) is at the upper end of the range (0.3 to 8 :g/m 2 /yr) 

reported for the Swiss bog (Roos-Barraclough et al., 2002), with the uppermost Swiss values 

seen only during periods of volcanic activity. The modelled result is also outside of the pre- 

industrial range reported here for GL (0.3 to 3 :g/m 2 /yr). Given that the deeper GL peats 

containing elevated Hg accumulation rates are enriched in Cu, Se, and U, we assume that the 

lowest Hg accumulation rates (0.3 to 0.5 :g/m 2 /yr from AD 550 to 975) are representative of the 

pre-industrial, atmospheric Hg flux. The pre-industrial flux calculated by Lamborg et al. (2002), 

therefore, may be too large by as much as one order of magnitude. 

At this time, we cannot determine how much of the atmospheric Hg accumulation in 

GL may have been due to long range atmospheric transport, and how much due to local sources. 

Since our study began, we have learned that the U.S. military operated a secret base called 

“Bluie West One”, across the fjord from the sampling site, from 1941 to 1958 (the former airfield 

is now Narsarssuaq airport). This base is said to have housed up to 10,000 persons during its 

zenith, and had its own kilns for manufacturing bricks: this may help to explain the magnitude 

and chronology of the elevated Hg fluxes from this period. Additional studies of peat bogs from 

other locations should help improve our understanding of local and long-range Hg transport, and 

31

further improve our knowledge of pre-industrial and post-industrial Hg accumulation rates. 

Temporal changes in the isotopic composition of Pb in Greenland 

As noted earlier, Pb concentrations are far lower in the GL core compared to DK. The 

Pb EF calculated using the Pb and Ti concentrations failed to reveal a significant enrichment of 

Pb in the surface layers of the GL core (Fig. 5a). However, in the surface and near-surface peat 

layers at GL, the Pb is much less radiogenic in both the leachable and residual fractions (Fig. 6b), 

compared to deeper, older peat layers. Given that this zone dates from the past century, the shift 

in Pb isotope ratios probably reflects the input of anthropogenic Pb with lower 206 Pb/ 207 Pb ratios. 

Anthropogenic, atmospheric Pb in Greenland has previously been shown to be derived primarily 

by gasoline lead used in North America (Rosman et al., 1998). Alklyllead compounds in N. 

America were synthesized primarily using lead ores of the Mississipi Valley type deposits which 

are much more radiogenic ( 206 Pb/ 207 Pb ca. 1.2) compared to the ores used for gasoline lead in 

Europe. However, even these, relatively radiogenic leads are considerably less radiogenic than 

the rocks and minerals of Greenland ( 206 Pb/ 207 Pb up to 1.3 in Fig. 6b). Thus, addition to the 

surface peat layer at GL of Pb derived from N. American gasoline lead, could easily have caused 

the significant decrease in 206 Pb/ 207 Pb revealed by the peat core. The shift in 206 Pb/ 207 Pb to less 

radiogenic values in samples above ca. 20 cm in the GL core (Fig. 6b), therefore, most likely 

reflects the addition of Pb derived from the combustion of leaded gasoline in N. America. Thus, 

while total Pb concentrations and their ratio to Ti failed to indicate an anthropogenic impact, the 

Pb isotope data clearly does. Additional work is required with respect to leaching techniques 

such as those now being used to fractionate Pb in soils (Harlavan and Erel, 2002; Emmanuel and 

Erel, 2002), with the goal of improving the recovery of the atmospheric Pb component which has 

been supplied by anthropogenic emissions. 

The Pb concentrations in the leachable fraction generally exceed those of the residual 

32

fraction, even in the deeper peat layers dating from pre-anthropogenic times (Fig. 6b). In the 

deepest peat samples studied (88cm), the leached and residual fractions are isotopically different: 

both fractions are highly radiogenic which suggests that both sources are local. Many geological 

studies employing leaching techniques indicate that radiogenic Pb in minerals is more easily 

recovered in the “acid-leachable” fraction because Pb is found in radiation lattice defects where 

it is less strongly held, compared with lattice-bound Pb which was originally incorporated in U- 

Th-bearing minerals. In the GL samples, however, it is the residual fraction which is more 

radiogenic than the leachable fraction. Given that this section of the peat profile is strongly 

minerotrophic, the most obvious possible sources of Pb include i) physical incorporation of fine 

grained mineral matter by the plants which were growing at the time peat formation began (ca. 

3,000 14 C yr BP), and ii) chemical adsorption/complexation of dissolved Pb which had been 

released to the plants and basal peat by chemical weathering of the host minerals. While it may 

be that the former process is primarily responsible for the isotopic composition of the Pb in the 

more radiogenic “residual” fraction, and the latter for the less radiogenic Pb in the “leachable” 

fraction, there are other possibilities. For example, to some extent the “leachable” fraction is an 

artefact of leaching the samples for 2 hours in 2N HCl which would not only remove 

exchangeable Pb, but also dissolve some fine grained mineral phases hosting Pb such as micas 

and feldspars; some of these may ultimately have been supplied by atmospheric soil dust and 

would therefore be less radiogenic. 

The isotopic composition of Pb in the peat from GL is inherently more complicated 

than the DK core, as Pb is derived from both atmospheric and non-atmospheric (aquatic + 

terrestrial) sources. Starting again in the basal peat layer, the residual fraction was found to be 

significantly more radiogenic than the leachable fraction (Fig 6b). The residual fraction must 

reflect the isotopic composition of Pb-bearing minerals in the local rocks, and the sediments 

33

which are derived from them. In contrast, the leachable fraction of the peat must also include Pb 

supplied by atmospheric soil dust. Assuming that the residual fraction ( 206 Pb/ 207 Pb = 1.3) reflects 

the composition of local sediments, and that atmospheric soil dust is similar to that found in 

crustal rocks ( 206 Pb/ 207 Pb = ca. 1.2), a mixture of 65% of the former and 35% of the latter would 

account for the isotopic composition of Pb in the leachable fraction ( 206 Pb/ 207 Pb = 1.265 for the 

leachate fraction of the GL sample from 88cm+ as shown Table 3). 

SUMMARY AND CONCLUSIONS 

The similar chronology of changing Hg concentrations in the surface peat layers at GL 

and DK (Fig. 2) suggest that minerotrophic peatlands also may preserve a record of atmospheric 

Hg accumulation. Using the bomb pulse curve to accurately date the past 50 years of peat 

accumulation using 14 C, the GL peat profile provides a reconstruction of atmospheric Hg which 

is consistent with the 40 year record provided by Greenland snow (Boutron et al., 1998), and 

complements the records provided for this region by marine (Asmund and Nielsen, 2000) and 

lake sediments (Landers et al., 1998; Bindler et al., 2001). Mercury fluxes in GL peats dating 

from pre-industrial times were as low as 0.3 to 0.5 µg/m 2 /yr between AD 550 and 975 which 

provides an estimate of the “background” atmospheric Hg fluxes. The accumulation rate reached 

a maximum of 164 µg/m 2 /yr in AD 1953. The GL core indicates that the Hg flux has since 

declined but the value in 1995 (14 µg/m 2 /yr) is clearly elevated with respect to the natural range. 

The chronology of Hg accumulation rates recorded by the ombrotrophic peat bog in 

DK are similar to GL. The sample containing the greatest Hg accumulation rate is also the 

sample which is most enriched in anthropogenic Pb and As. The synchroneity of these 

enrichments suggests that the predominant source of all three elements was coal-burning, a view 

which is supported by the Pb isotope data. Moreover, the Pb isotope data show very clearly that 

the supply of anthropogenic Pb to the air went into decline well before the phase-out of unleaded 

34

gasoline. 

The ratio of modern/pre-Industrial Hg concentrations between the two sites (59x in DK 

versus 15x in GL) is only a factor of four, compared to the differences between Pb and As EF 

(approximately 60 to 80 x). The similar rates of atmospheric Hg accumulation in southern 

Greenland and Denmark show that atmospheric dispersion of Hg released by human activities 

is fundamentally different to that of As and Pb: Hg is primarily transported as Hg o in the gas 

phase with a residence time of one to two years, whereas Pb and and As are transported mainly 

as aerosols with a residence time of ca. one week (Hutchinson and Meema, 1987); this supports 

the prediction (Morel et al., 1998) that the chronology of atmospheric Hg contamination in rural 

and remote areas should be largely comparable because of long range, vapour phase transport 

of Hg. 

35

ACKNOWLEDGEMENTS 

We are grateful to A.K. Cheburkin for all of the XRF analyses, W.O. van der Knaap for 

identifying plant macrofossils, W. Rom for help with the atmospheric bomb pulse curve, and S. 

Reese for the 14 C age dates obtained by decay counting of the basal peat layers. Thanks also to 

C. Ihrig for supplementary measurements of bulk density and to G. LeRoux and N. Givelet for 

helpful suggestions which have led to considerable improvement in our peat sample handling 

and preparation protocol. For fruitful discussions, W.S. thanks A. Martinez-Cortizas, N. Givelet, 

and G. LeRoux; thanks also to H.L. Nielsen, the reviewers, and M. Novak for comments which 

helped to considerably improve the manuscript. This investigation was supported by the Swiss 

National Science Foundation (to W.S. at the University of Berne) and the Danish Cooperation 

for Environment in the Arctic (DANCEA) to M.E.G. Additional support was provided by GKSS, 

Germany (thanks to H. von Storch), and the Carlsberg Foundation (to C. Lohse). M.E.G. wishes 

to acknowledge the advice, guidance, and support of C. Lohse and T.S. Hansen for helping to 

launch the project, the expert technical assistance of T. Nørnberg and P.B. Hansen, and B. 

Odgaard and B. Aaby for helpful discussions about Danish bogs. M. E. G. was subsequently 

supported by the Danish Research Agency and The Department of Atmospheric Environment 

of Denmark (NERI) with a PhD fellowship at the Copenhagen Global Change Initiative 

(COGCI), University of Copenhagen (special thanks to Ole John Nielsen, Henrik Skov and Steve 

E. Lindberg for their supervision and guidance). Thanks also to the Danish Polar Centre, 

Greenland Homerule, Director of Environment, Greenland Office of Tourism, Greenland 

National Museum and Archive, and the Municipality and inhabitants of Narsaq. Finally, special 

thanks to B.E. Haas for improving the English, and the usual message from W.S.. 

36

REFERENCES 

Aaby, B. and Jacobsen, J. (1979) Changes in biotic conditions and metal deposition in the last millenium as 

reflected in ombrotrophic peat in Draved Mose, Denmark. Danm. geol. Unders., Årbog 1978, 

5-43. 

AMAP (1998) AMAP Assessment Report: Arctic Pollution Issues. Arctic Monitoring and Assessment 

Programme (AMAP), Oslo, Norway. xii + 859 pp. 

Andersen, T. (1997) Age and petrogenesis of the Qassiarsuk carbonatite-alkaline silicate volcanic complex 

in the Gardar rift, South Greenland. Mineralogical Magazine 61:499-513. 

Armour-Brown, A.; Steenfelt, A.; Kunzendorf, H.(1983) Uranium districts defined by reconnaissance 

geochemistry in South Greenland. Journal of Geochemical Exploration 19:127-45. 

Asbirk, S. Bertelsen, U., Engelbøøl, S.E. & Lorenzen, H.P. 1973. En naturhistorisk undersøgelse af 

høøjmoserne Holmegaards Mose, Storelung og Skidendam. -Meddelelser om danske 

naturlokaliteter nr. 6 (Udgivet af foreningne Natur og Ungdom), København, 122 s. 

Asmund, G. and Nielsen, S.P. (2000) Mercury in dated Greenland marine sediments. Science of the Total 

Environment 245:61-72. 

Benoit, J.M., Fitzgerald, W.F., and Damman, A.W.H. (1998) The biogeochemistry of an ombrotrophic bog: 

evaluation of use as an archive of atmospheric mercury deposition. Environmental Research, 

Section A 78:118-133. 

Biester, H., Kilian, R., Hertel, C., Woda, C., Mangini, A, Schöler, H.F. (2002): Elevated Mercury 

Concentrations in Peat Bogs of South Patagonia, Chile – An Anthropogenic Signal. Earth and 

Planetary Science Letters 201:609-620. 

Biester, H., Martinez-Cortizas, A., Birkenstock, S., Kilian, R. (2003). Historic Mercury Records in Peat 

Bogs. The Role of Peat Decomposition, and Mass Losses. Environmental Science and 

Technology 37:32-39. 

37

Bindler, R. (2003) Estimating the natural background atmospheric deposition rate of mercury utilizing 

ombrotrophic bogs in south Sweden. Environmental Science and Technology 37: 40-46. 

Bindler, R., Renberg, I., Appleby, P.G., Anderson, N.J., and Rose, N.L. (2001) Mercury accumulation rates 

and spatial patterns in lake sediments from West Greenland: a coast to ice margin transect. 

Environmental Science and Technology 35:1736-1741. 

Bollhöfer, A. and Rosman, K.J.R. (2001) Isotopic signatures for atmospheric lead: The Northern Hemisphere. 

Geochimica et Cosmochimica Acta 65:1727-1740. 

Bouska, V. 1981. The Geochemistry of Coal. Elsevier, Amsterdam. 

Boutron, C.F., Vandal., V.M., Fitzgerald, W.F., and Ferrari, C.P. (1998) A forty year record of mercury in 

central Greenland snow. Geophyical Research Letters 25:3315-3318. 

Boyarkina, A.P., Vasil`ev, N.V., Glukhov, G.G., Rezchikov, V.I., and Tyulyupo, E.B. (1980) Gold and 

mercury levels in Sphagnum peats. Byull. Pochv. Inst. im. V. V. Dokuchaeva. 24:24-5 

Braune, B., et al. (1999) Spatial and temporal trends of contaminants in Canadian Arctic freshwater and 

terrestrial ecosystems: a review. Science of the Total Environment 230:145-207. 

Bränvall, M.L., Bindler, R., Emteryd, O., Nilsson, M., and Renberg, I. (1997) Stable isotope and 

concentration records of atmospheric lead pollution in peat and lake sediments in Sweden. Water 

Air Soil Pollution 100:243-252. 

Christensen, J. H., M.. E. Goodsite, N. Z. Heidam, H. Skov and P. Wåhlin (2002): Atmospheric Environment. 

Chapter 1. In: Riget, F., J. Christensen & P. Johansen (eds). AMAP Greenland Environment 

(1997-2001). Ministry of Environment, Denmark Department of Atmospheric Environment, 

National Environmental Research Institute of Denmark, Frederiksborgvej 399, Box 358, DK- 

4000 Roskilde, Denmark 

Clymo, R.S. (1987) The ecology of peatlands. Science Progress (Oxford) 71:593-614. 

Ebinghaus, R.. Tripathi, R.M., Wallschlager, D., and Lindberg, S.E. (1999) Natural and anthropogenic 

38

mercury sources and their impact on the air-surface exchange of mercury on regional and global 

scales, in: R. Ebinghaus, R.R. Turner, D. Lacerda, O. Vasiliev, W. Salomons (eds.), Mercury 

Contaminated Sites, Springer Verlag, Heidelberg, 1999. 

Emmanuel, S. and Erel, Y. (2002) Implications from concentrations and isotopic data for Pb partitioning 

processes in soils. Geochimica et Cosmochimica Acta 66:2517-2527. 

Farmer, J.G., Eades, L.J., and Graham, M.C. (1999) The lead content and isotopic composition of British 

coals and their implications for past and present releases of lead to the U.K. environment. 

Environmental Geochemistry and Health 21:257-272. 

Farmer, J.G., Mackenzie, A.B., Sugden, C.L., Edgar, P.J. and Eades, L.J. (1997) A comparison of the 

historical lead pollution records in peat and freshwaer lake sediments from central Scotland. 

Water Air Soil Pollution 100:253-270. 

Flament, P, Bertho, M.-L., Deboudt, K., Veron, A. and Puskaric, E. (2002) European isotopic signatures for 

lead in atmospheric aerosols: a source apportionment based upon 206 Pb/ 207 Pb ratios. Science of 

the Total Environment 296:35-57. 

Fitzgerald, W.F. and Clarkson, T.W. (1991) Mercury and methylmercury: present and future concerns. 

Environmental Health Perspectives 96:159-166. 

Fitzgerald, W.F., Engstrom, D.R., Mason, R.P., and Nater, E.A. (1997) The case for atmospheric mercury 

contamination in remote areas. Environmental Science and Technology 32:1-7. 

Frei, R., Villa, I.M., Nägler, Th.F., Kramers, J.D., Przybylowicz, W.J., Prozesky, V.M., Hofmann, B.A., and 

Kamber, B.S., 1997, Single mineral dating by the Pb-Pb step-leaching method, assessing the 

mechanism. Geochimica et Cosmochimica Acta 61:393-414. 

Gobeil, C., Macdonald, R.W., and Smith, J.N. (1999) Mercury profiles in sediments of the Arctic Ocean 

basins. Environmental Science and Technology 33:4194-4198. 

Goodsite, M.E. (2000) Heavy metal deposition determined by correlation with 14C . M.Sc. thesis, University 

39

of Southern Denmark. 

Goodsite, M.E., Heinemeier, J., Rom, W., Lange, T., Ooi, S., Appleby, P.G., Shotyk, W., van der Knaap, 

W.O., Lohse, C. and Hansen, T.S. (2002) High resolution AMS 14 C dating of post bomb peat 

archives of atmospheric pollutants. Radiocarbon 43(2B):495-515. 

Hagner, C. (2000). European regulations to reduce lead emissions from automobiles - did they have an 

economic impact on the German gasoline and automobile markets?. Regional Environmental 

Change 1:135-151. 

Harlavan, Y. and Erel, Y. (2002) The release of Pb and REE from granitoids by the dissolution of accessory 

phases. Geochimica et Cosmochimica Acta 66:837-848. 

Hutchinson, T.C. and Meema, K.M. (eds.) 1987. Lead, Mercury, Cadmium, and Arsenic in the Environment. 

SCOPE 31. John Wiley and Sons, New York. 

Jensen, F.P. and Fenger, J. (1994) The air quality in Danish urban areas. Environmental Health Perspectives 

102:55-60. 

Jensen, A. and Jensen, A. (1991) Historical rates of mercury in Scandinavia estimated by dating and 

measurement of mercury in cores of peat bogs. Water, Air and Soil Pollution 56:769-777. 

Kempter, H., Görres, M. and Frenzel, B. (1997) Ti and Pb concentrations in rainwater-fed bogs in Europe as 

indicators of past anthropogenic activities. Water Air Soil Pollution 100:367-377. 

Kempter, H. and Frenzel, B. (1999) The local nature of anthropogenic emission sources on the elemental 

content of nearby ombrotrophic peat bogs, Vulkaneifel, Germany. Science of the Total 


Kempter, H. and Frenzel, B. (2000) The impact of early mining and smelting on the local tropospheric 

aerosol detected in ombrotrophic peat bogs in the Harz, Germany. Water Air, Soil Pollution 

121:93-108. 

Kober, B. & Wessels, M. & Bollhöfer, A. & Mangini, A. (1999) Pb isotopes in sediments of Lake Constance, 

40

Central Europe constrain the heavy metal pathways and the pollution history of the catchment, 

the lake and the regional atmosphere. Geochimica et Cosmochimica Acta 63:1293-1303. 

Kramers, J.D. and Tolstikhin, I.D. (1997) Two terrestrial lead isotope paradoxes, forward transport 

modelling, core formation, and the history of the continental crust. Chemical Geology 139: 75- 

110. 

Lamborg, C.H., Fitzgerald, W.F., O`Donnell, J., and Torgersen, T. (2002) A non-steady state compartmental 

model of global-scale mercury biogeochemistry with interhemispheric atmospheric gradients. 

Geochimica et Cosmochimica Acta 66:1105-1118. 

Landers, D.H., Gubala, C., Verta, M., Lucotte, M., Johansson, K., Vlasova, T., and Lockhart, W.L. (1998) 

Using lake sediment mercury flux ratios to evalute the regional and continental dimensions of 

mercury deposition in Arctic and Boreal ecosystems. Atmospheric Environment 32:919-928. 

Lin, C-J. and Pehkonen, S.O. (1999) The chemistry of atmospheric mercury: a review. Atmospheric 


Lindberg, S., P.M. Stokes, E. Goldberg, and C. Wren (1987). Group Report: Mercury In T.C. Hutchinson and 

K.M. Meema (eds). Lead, Mercury, Cadmium, and Arsenic in the Environment, pp. 17-33. John 

Wiley and Sons, New York. 

Lodenius, M., Seppänen, A. and Uusi-Rauva, A. (1983) Sorption and mobilization of mercury in peat soil. 

Chemosphere 12:1575-1581. 

Martinez-Cortizas, A., Pontevedra Pomba, X., Novoa Munoz, J.C. and Garcia-Rodeja, E. (1997) Four 

thousand years of atmospheric Pb, Cd, and Zn deposition recorded by the ombrotrophic peat bog 

of Penido Vello (northwestern Spain) Water Air Soil Pollution 100:387-403. 

Martinez-Cortizas, A.M., Ponteedra Pomba, X., Garcia-Rodeja, E., Novoa Munoz, J.C. and W. Shotyk (1999) 

Mercury in a Spanish peat bog: archive of climate change and atmospheric metal deposition. 

Science 284:939-942. 

41

MacKenzie AB, Farmer JG, Sugden CL. (1997) Isotopic evidence of the relative retention and mobility of 

lead and radiocaesium in Scottish ombrotrophic peats. Science of the Total Environment 

203:115-127. 

MacKenzie AB, Logan EM, Cook GT, Pulford ID. (1998) Distributions, inventories, and isotopic 

composition of lead in 210 Pb-dated peat cores from contrasting biogeochemical environments: 

Implications for lead mobility. Science of the Total Environment 223:25-35. 

Monna, F., Lancelot, J., Croudace, I.W., Cundy, A.B. and Lewis, J.T. (1997) Pb isotopic composition of 

airborne particulate material from France and the southern United Kingdom: Implications for Pb 

pollution sources in urban areas. Environmental Science and Technology 31:2277-2286. 

Morel, F.M.M., Kraepiel, A.M.L., and Amyot, M. (1998) The cycle and bioaccumulation of mercury. Annual 

Reviews of Ecological Systems 29:543-566. 

Naucke, W., Heathwaite, A.L., Eggelsmann, R. and Schuch, M. (1993). Mire chemistry. In Mires. Process, 

Exploitation, and Conservation (K. Göttlich, editor and J. Cooke, translator). John Wiley and 

Sons, New York, 1993. 

Norton, S.A., Evans, G.C. and Kahl, J.S. (1997) Comparison of Hg and Pb fluxes to hummocks and hollows 

of ombrotrophic Big Heath Bog and to nearby Sargent Mt. Pond, Maine, USA. Water Air Soil 

Pollution 100:271-286. 

Novak M., Emmanuel S., Vile M. A., Erel Y., Veron A., Paces T., Wieder R.K., Vanecek M., Stepanova M., 

Brizova E., Hovorka J. (2003). Origin of lead in eight Central European Peat Bogs determined 

from isotope ratios, strengths and operation times of regional pollution sources. Environmental 

Science and Technology 37: 437-445. 

Oechsle, D. (1982) Untersuchungen zur Mobilität von Quecksilber-Spezies in Hochmooren und 

Möglichkeiten zu ihrer analytische Trennung and Bestimmung. Ph.D.Thesis, Universität 

Stuttgart. 

42

Pheiffer-Madsen, P. (1981) Peat bog records of atmospheric mercury deposition.Nature 293:127-129. 

Puxbaum, H. (1991). Metal compounds in the atmosphere In E. Merian (ed). Metals and their Compounds 

in the Environment. VCH, Weinheim, pp. 257-286. 

Rasmussen, P.E. (1994) Current methods of estimating atmospheric mercury fluxes in remote areas. 

Environmental Science and Technology 28:2233-2241. 

Reuer M. K. and Weiss D. J. (2002) Anthropogenic lead dynamics in the terrestrial and marine environment. 

Philosophical Transactions of the Royal Society of London A 360, 2889-2904. 

Roos-Barraclough, F. and Shotyk, W. (2003) Millennial-scale records of atmospheric mercury deposition 

obtained from ombrotrophic and minerotrophic peat from the Swiss Jura Mountains. 

Environmental Science and Technology 37(2):235-244. 

Roos-Barraclough, F., Martinez-Cortizas, A., Garcia-Rodeja, E., and Shotyk, W. (2002). A 14,500 year 

record of the accumulation of atmospheric mercury in peat: volcanic signals, anthropogenic 

influences, and a correlation to bromine accumulation. Earth and Planetary Science Letters 

202(2):435-451. 

Roos-Barraclough, F., Biester, H.F., Goodsite, M.E., Martinez-Cortizas, A.. and Shotyk, W. (2002) 

Analytical protocol for measuring total Hg concentrations in peat cores. Science of the Total 


Rosman, K.J.R., Chisholm, W., Boutron, C.F., Candelone, J.P., Jaffrezo, J.-L., and Davidson, C.I. (1998) 

Seasonal variations in the origin on lead in snow at Dye 3, Greenland. EPSL 160, 383-389. 

Salvato, N. and Pirola, C. (1996) Analysis of mercury traces by means of solid sample atomic absorption 

spectrometry. Mikrochimica Acta 123:63-71. 

Schaller, M., Steiner, O., Studer, I., Frei, R., and Kramers, J.D., (1997) Pb stepwise leaching (PbSL) dating 

of garnet - addressing the inclusion problem. Schweizerische Mineralogische Petrographische 

Mitteilungen, v.77, p.113-121 

43

Schroeder, W.H. and Munthe, J. (1998) Atmospheric mercury - an overview. Atmospheric Environment 

32:809-822. 

Shotyk, W. (1988) Review of the inorganic geochemistry of peats and peatland waters. Earth-Science 

Reviews 25(2):95-176. 

Shotyk, W. (2002). The chronology of anthropogenic, atmospheric Pb deposition recorded by peat cores in 

3 minerotrophic peat deposits from Switzerland. Science of the Total Environment 292:19-31 

Shotyk, W., Nesbitt, H.W., and Fyfe, W.S. (1992) Natural and anthropogenic enrichments of trace metals in 

peat profiles. International Journal of Coal Geology 20:49-84. 

Shotyk, W., Weiss, D., Appleby, P.G., Cheburkin, A.K., Frei, R., Gloor, M., Kramers, J.D., Reese, S., and 

van der Knaap, W.O. (1998). History of atmospheric lead deposition since 12,370 14 C yr BP 

recorded in a peat bog profile, Jura Mountains, Switzerland. Science 281:1635-1640. 

Shotyk, W., Blaser, P., Grünig, A., and Cheburkin, A.K. (2000). A new approach for quantifying cumulative, 

anthropogenic, atmospheric lead deposition using peat cores from bogs: Pb in eight Swiss peat 

bog profiles. Science of the Total Environment 249:257-280. 

Shotyk, W., Weiss, D., Kramers, J.D., Frei, R., Cheburkin, A.K., Gloor, M. and Reese, S. (2001) 

Geochemistry of the peat bog at Etang de la Gruère, Jura Mountains, Switzerland, and its record 

of atmospheric Pb and lithogenic trace elements (Sc, Ti, Y, Zr, Hf and REE) since 12,370 14 C 

yr BP. Geochimica et Cosmochimica Acta 65(14) 2337-2360. 

Shotyk, W., Weiss, D., Heisterkamp, M., Cheburkin, A.K., and Adams, F.C. (2003) A new peat bog record 

of atmospheric lead pollution in Switzerland: Pb concentrations, enrichment factors, isotopic 

composition, and organolead species Environmental Science and Technology 37(2):235-244.. 

Shotyk, W., Krachler, M., Martinez-Cortizas, A., Cheburkin, A.K., and Emons, H. (2002b) A peat bog record 

of natural, pre-anthropogenic enrichments of trace elements in atmospheric aerosols since 12,370 

14 C yr BP, and their variation with Holocene climate change. Earth and Planetary Science 

44

Letters 199:21-37. 

Steinmann, P. and Shotyk, W. (1997) The pH, redox chemistry, and speciation of Fe and S in pore waters 

from two contrasting Sphagnum bogs, Jura Mountains, Switzerland. Geochimica et 

Cosmochimica Acta 61(6):1143-1163. 

Swaine, D.J. Trace Elements in Coal. Butterworth, London, 1990. 

Todt, W., Cliff, R.A., Hanser, A., Hofmann, A.W., 1993. Re-calibration of NBS lead standards using a 

202 Pb+ 205 Pb double spike. Terra Abstracts 5, Suppl. 1, 396. 

Valkovic, V. Trace Elements in Coal. CRC Press, Boca Raton, 1983 (2 Vols.). 

Vile, M.A., Novak, M.J., Brizova, E., Wieder, R.K., and Schell, W.R. (1995) Historical rates of atmospheric 

Pb deposition using 210 Pb dated peat cores: corroboration, computation, and interpretation. 

Water, Air, Soil Pollution 79:89-106. 

Vile MA, Wieder RK, Novak M. (1999) Mobility of Pb in Sphagnum-derived peat. Biogeochemistry 45:35- 

52. 

Wardenaar, E.C.P. (1987). A new hand tool for cutting peat profiles. Canadian Journal of Botany 65:1772- 

1773. 

Wedepohl, K.H. (1995) The composition of the continental crust. Geochimica et Cosmochimica Acta 

59:1217-1232. 

Weiss, D., Shotyk, W., Appleby, P.G., Cheburkin, A.K., and Kramers, J.D. (1999a). Atmospheric Pb 

deposition since the Industrial Revolution recorded by five Swiss peat profiles: enrichment 

factors, fluxes, isotopic compositon, and sources Environmental Science and Technology 

33:1340-1352. 

Weiss, D., Shotyk, W., Gloor, M. and Kramers, J.D. (1999b). Herbarium specimens of Sphagnum moss as 

archives of recent and past atmospheric Pb deposition in Switzerland: isotopic composition and 

source assessment Atmospheric Environment 33:3751-3763. 

45

Table 1. AMS 14 C dating of plant macrofossils from the peat core (GL2B) from Tasiusaq, 

Greenland. 

Lab nr 

AAR- 

Average 

depth 

(cm) 

δ 13 C 

(‰) 

Conv. 14 C age 

(BP) 

14 C content 

(pMC) 

Calibrated age 1 ) 

5620 0.5 -26.2 -840 ± 40 110.99 ± 0.57 1957;1996-1999 

5621 0.0 -28.3 -900 ± 45 111.88 ± 0.62 1957;1994-1999 

5622 -0.5 -26.8 -1075 ± 40 114.34 ± 0.57 1957-1958;1991-1994 

6861 -1.5 -26.55 -1180 ± 40 115.84 ± 0.58 1958;1989-1992 

6899 -2.5 -26.62 -1090 ± 30 114.54 ± 0.41 1957-1958;1991-1993 

6862 -3.5 -26.05 -2150 ± 35 130.65 ± 0.54 1962;1978-1979 

5623 -5.5 -26.3 -2720 ± 35 140.33 ± 0.61 1962;1973-1974 

5624 -7.5 -25.8 -2880 ± 40 143.13 ± 0.69 1962;1973 

5625 -9.5 -25.9 -3890 ± 35 162.32 ± 0.72 1963;1967 

5626 -12.5 -27.6 -4685 ± 35 179.13 ± 0.83 1963-1965 

5627 -14.5 -26.7 -1895 ± 35 126.62 ± 0.59 1961-1962;1980-1981 

5628 -16.5 -27.2 -1535 ± 35 121.07 ± 0.54 1958-1961 

5629 -18.5 -27.1 -200 ± 40 102.52 ± 0.54 1956 

6900 -26.5 -24.59 65 ± 40 99.18 ± 0.51 AD: 1700-1720; 1810-1830; 

1880-1920; 1950-55 

6901 -29.5 -25.29 130 ± 40 98.38 ± 0.52 AD: 1670-1760; 1800-1890; 

1910-1950 

6620 -36.5 -33.3 1095 ± 45 87.24 ± 0.51 AD: 890-1000 

6621 -54.5 -23.6 1115 ± 45 87.04 ± 0.50 AD: 890-985 

6622 -74.5 -26.6 2485 ± 45 73.39 ± 0.41 BC: 770-520 

6623 -80.5 -26.6 2845 ± 50 70.19 ± 0.42 BC: 1110-1100; 1080-920 

5630 -87.5 -26.7 2920 ± 50 69.52 ± 0.44 BC: 1210-1010 

1 ) Calibrated age ranges corresponding to 95% confidence interval for depths 0-18.5 cm, and 

68% confidence interval for depths 26.5-87.5 cm.

Table 2. AMS 14 C dating of plant macrofossils from the peat core (DK1B) from Storelung 

mose, Denmark. 

Lab nr 

AAR- 

Average 

depth 

(cm) 

δ 13 C 

(‰) 

Conv. 14 C 

age (yrs BP) 

14 C content 

(pMC) 

Calibrated age 1 ) 

5611 0.0 -27.5 -865 ± 40 111.36 ± 0.54 1957;1995-1999 

5612 -0.5 -26.1 -860 ± 45 111.31 ± 0.61 1957;1995-1999 

5613 -2.5 -26.8 -1180 ± 45 115.84 ± 0.65 1958; 1989-1992 

6855 -3.5 -26.88 -1830 ± 35 125.62 ± 0.53 1961; 1980-1982 

6856 -4.5 -26.25 -2210 ± 40 131.65 ± 0.65 1962; 1978-1979 

6857 -5.5 -24.92 -2520 ± 45 136.88 ± 0.75 1962; 1975-1976 

6858 -6.5 -23.93 -2785 ± 40 141.44 ± 0.70 1962; 1973-1974 

6859 -7.5 -23.24 -3465 ± 35 153.92 ± 0.69 1963; 1970 

5614 -8.5 -24.3 -3400 ± 40 152.68 ± 0.76 1963; 1970-1971 

6860 -9.5 -22.78 -4580 ± 30 176.82 ± 0.68 1963-1965 

6612 -10.5 -23.0 -2529 ± 43 136.99 ± 0.74 1962; 1975-1976 

6613 -11.5 -24.0 -1707 ± 42 123.68 ± 0.65 1959-1961; 1982-1984 

6614 -12.5 -23.6 -1926 ± 42 127.09 ± 0.67 1962; 1980-1981 

6615 -13.5 -24.3 -746 ± 47 109.74 ± 0.65 1957; 1995-1999 

5615 -14.5 -24.7 -1480 ± 35 120.19 ± 0.55 1958; 1960; 1984-1987 

5616 -15.5 -25.7 -1505 ± 35 120.58 ± 0.56 1958-1961; 1984-1987 

5617 -16.5 -27.0 -10 ± 45 100.12 ± 0.53 AD: 1693-1726; 1813-1850; 

1862-1918; 1951-1956 

5618 -18.5 -24.2 45 ± 40 99.43 ± 0.49 AD: 1895-1905; 1951-55 

6898 -19.5 -25.34 160 ± 40 98.00 ± 0.49 AD: 1660-1700; 1720-1820; 

1850-1870; 1910-1950 

6616-1 -28.5 -26.7 2395 ± 45 74.20 ± 0.43 BC: 760-720; 540-390 

6616-2 -28.5 -27.5 380 ± 45 95.41 ± 0.55 AD: 1440-1520; 1590-1630 

6617-1 -34.5 -25 3225 ± 50 66.94 ± 0.42 BC: 1600-1590; 1530-1430 

6617-2 -34.5 -28.7 2720 ± 45 71.29 ± 0.41 BC: 905-820 

6618 -46.5 -26.7 2935 ± 50 69.40 ± 0.43 BC: 1260-1240; 1220-1040 

6619 -68.5 -26.5 2970 ± 50 69.11 ± 0.45 BC: 1300-1080; 1060-1050 

5619 -78.5 -24.5 3050 ± 45 68.39 ± 0.38 BC: 1390-1260; 1230-1220 

1 ) Calibrated age ranges corresponding to 95% confidence interval for depths 0-16.5 cm, and 

68% confidence interval for depths 18.5-78.5 cm.

Table 3. Pb isotope data of leachates and residues from DK and GL peatlands 

sample depth phase acid* time Pb 206Pb/204Pb ± 2s+ 207Pb/204Pb ± 2s+ 208Pb/204Pb ± 2s+ 206Pb/207Pb ± 2s+ 208Pb/206Pb ± 2s+ 208Pb/207Pb r1** r2†† 

(ppm) 

DK 1B 0+, L, IC 0 leachate 2N HCl 1h 8,28 17,872 0,092 15,508 0,081 37,549 0,196 1,1525 0,0006 2,1010 0,0012 2,4213 0,992 0,994 

DK 1B 1-2, L, IC 2 leachate 2N HCl 1h 24,24 17,726 0,025 15,562 0,023 37,528 0,057 1,1390 0,0003 2,1172 0,0008 2,4116 0,979 0,968 









DK 1B 78-79, L, IC 79 leachate 2N HCl 1h 17,639 0,063 15,128 0,055 36,952 0,135 1,1660 0,0004 2,0949 0,0009 2,4427 0,993 0,993 

DK 1B 0+, R, IC 0 residue 8N HBr - HF1d 1,27 17,241 0,168 14,892 0,146 36,125 0,354 1,1577 0,0008 2,0953 0,0019 2,4257 0,996 0,996 

DK 1B 1-2, R, IC 2 residue 8N HBr - HF1h 10,19 17,787 0,022 15,560 0,022 37,566 0,053 1,1432 0,0005 2,1120 0,0011 2,4143 0,923 0,935 




DK 1B 15-16, R, IC 16 residue 8N HBr - HF1d 32,08 17,897 0,012 15,556 0,013 37,732 0,036 1,1505 0,0002 2,1083 0,0010 2,4255 0,947 0,886 



DK 1B 40-41, R, IC 41 residue 8N HBr - HF1d 5 18,396 0,037 15,602 0,033 38,311 0,081 1,1791 0,0003 2,0826 0,0009 2,4556 0,986 0,980 


DK 1B 78-79, R, IC 79 residue 8N HBr - HF1d 18,023 0,066 15,464 0,057 37,743 0,140 1,1655 0,0005 2,0942 0,0011 2,4407 0,988 0,991 

GL 2B 0+, L, IC 0 leachate 2N HCl 1h 1,57 18,603 0,072 15,266 0,060 37,592 0,148 1,2186 0,0005 2,0208 0,0013 2,4625 0,986 0,986 

GL 2B 5-6, L, IC 5 leachate 2N HCl 1h 1,63 19,510 0,041 15,644 0,034 38,815 0,086 1,2472 0,0004 1,9895 0,0010 2,4812 0,971 0,972 






GL 2B 88+, L, IC 88 leachate 2N HCl 1h 7,08 19,417 0,037 15,347 0,030 38,297 0,076 1,2652 0,0003 1,9724 0,0008 2,4955 0,982 0,979 

GL 2B 0+, R, IC 0 residue 8N HBr - HF1d 1,25 18,816 0,088 14,992 0,071 36,424 0,173 1,2550 0,0005 1,9358 0,0011 2,4295 0,991 0,992 

GL 2B 5-6, R, IC 5 residue 8N HBr - HF1d 1,63 19,510 0,041 15,644 0,034 38,815 0,086 1,2472 0,0004 1,9895 0,0010 2,4812 0,971 0,972 




GL 2B 26-27, R, IC 27 residue 8N HBr - HF1h 3,24 20,594 0,068 15,653 0,052 37,497 0,126 1,3157 0,0003 1,8207 0,0010 2,3956 0,992 0,988 


GL 2B 88+, R, IC 88 residue 8N HBr - HF1d 5,01 20,038 0,051 15,419 0,040 39,894 0,105 1,2996 0,0004 1,9909 0,0011 2,5873 0,976 0,979 

UGS 1878-P, L, IC leachate 2N HCl 1h 74,88 17,821 0,025 15,583 0,023 37,647 0,059 1,1436 0,0003 2,1126 0,0011 2,4159 0,972 0,947 

UGS 1878-P, R, IC residue 8N HBr - HF1d 7,21 17,684 0,071 15,489 0,063 37,412 0,155 1,1417 0,0006 2,1156 0,0016 2,4154 0,988 0,982 

GL 2B 5-6, L, IC 5 leachate 2N HCl 1h 4,45 19,158 0,053 15,508 0,044 37,339 0,107 1,2354 0,0004 1,9490 0,0010 2,4078 0,983 0,982

FIGURE CAPTIONS 

1. Location maps of the coring sites in Greenland (GL) and Denmark (DK). 

2. a) Ash contents (%), dry bulk density (g/cm 3 ), gravimetric (ng/g) and volumetric 

(ng/cm 3 ) Hg concentrations in the GL “B” cores (cut into 1 cm slices). b) Ash contents 

(%), dry bulk density (g/cm 3 ), gravimetric (ng/cm 3 ) and volumetric (ng/cm3) Hg 

concentrations in the DK “B” cores. Selected age dates (from Tables 1 and 2) are 

shown for convenience. 

3. Age depth relationship in the GL “B” core. a) all samples. b) samples dated using the 

bomb pulse curve for 14 C. Age depth relationship in the DK “B” core. a) all samples. 

b) samples dated using the bomb pulse curve for 14 C. 

4. Mercury accumulation rates (µg/m 2 /yr). a) GL, all samples. b) GL (samples dated 

using the bomb pulse curve for 14 C. The upper solid line represents the estimated flux 

+ 20%, the lower line the estimated flux - 20%. c) DK samples dated using the bomb 

pulse curve for 14 C. The upper empty symbols represent the estimated flux + 20%, the 

lower hollow symbols the estimated flux - 20%. 

5. a) Pb and As concentrations (µg/g)in the GL and DK cores, and the Pb and As EFs 

calculated as described in the text. There is no measurable enrichment of Pb or As in 

the surface layers of the GL core. Age dates of the deepest samples from these “A” 

cores obtained using 14 C (decay counting, University of Berne) which yielded GL 

(>78cm) 3540 ± 30 14 C yr BP and DK (>84cm) 2790 ± 40 14 C yr BP (conventional 

radiocarbon yrs BP). b) the atmospheric fluxes of total Pb (solid symbols) and 

anthropogenic Pb (hollow symbols) in DK, obtained as described in the text. c) the 

percentage of anthropogenic Pb and its temporal evolution in DK. 

6. a) Pb concentrations and the isotopic composition of Pb in the exchangeable and 

49

esidual fractions of the “B” cores from DK. b) Pb concentrations and the isotopic 

composition of Pb in the exchangeable and residual fractions of the “B” cores from 

GL. c) “Residual” Pb measured using the extraction procedure described in the text, 

compared with “lithogenic Pb” calculated using either Pb/Zr = UCC, Pb/Ti = UCC, 

or Pb/Ti = 4x UCC. Also shown are “lithogenic” Pb and As calculated using either 

Pb/Ti = UCC or 4x UCC, and As/Ti = UCC or 10x UCC, in relation to the 206 Pb/ 207 Pb 

ratio of selected samples, and selected age dates obtained using 14 C (atmospheric 

bomb pulse). The small arrows point to the maximum concentration of Pb and As 

which was dated to AD 1954. 

7. Plot of 208 Pb/ 206 Pb versus 206 Pb/ 207 Pb for the leached fraction of the DK peat samples 

(Table 3), U.K. coal (from Farmer et al., 1999) and U.K. leaded gasoline (from Monna 

et al., 1997). Also shown is the isotopic composition of Pb in an Oxfordian sediment 

(from Shotyk et al., 1998). The direction of the arrows illustrate the temporal 

evolution of atmospheric Pb in DK from 1950 to 1980 (towards less radiogenic values 

as gasoline lead grows in importance) and from 1980 to 2000 (toward more radiogenic 

values as gasoline lead consumption declines). 

8. a) Ti (%), Zr, Y and Rb (µg/g) concentrations in the GL (green) and DK (red) “A” 

cores (3 cm slices. b) Ca (%), Sr, Mn (µg/g), and Fe (%) concentrations in the GL and 

DK “A” cores (3 cm slices). 

9. a) Ti (%), Zr, Y and Rb (µg/g) concentrations in the GL (green) and DK (red) “B” 

cores (1 cm slices) b) Ca (%), Sr, Mn (µg/g), and Fe (%) concentrations in the GL and 

DK “B” cores. 

10. a) Cu and U concentrations (µg/g), and Cu/Y and U/Y rations in the GL (green) and 

DK (red) “A” cores. b) Br and Se concentrations (µg/g) in the GL and DK “A” cores, 

50

and the Se/Br ratio. 

51

Depth (cm) 

Depth (cm) 

a 

b 

0 

-10 

-20 

-30 

-40 

-50 

-60 

-70 

-80 

-90 

0 10 20 30 40 

0 

-10 

-20 

-30 

-40 

-50 

-60 

-70 

Ash (%) 

-80 

0 5 10 15 

Ash (%) 

0 

-10 

-20 

-30 

-40 

-50 

-60 

-70 

-80 

-90 

0 0.1 0.2 0.3 

0 

-10 

-20 

-30 

-40 

-50 

-60 

-70 

-80 

-90 

0 50 100 150 

Bulk Density (g/cm3) Hg (ng/g) 

0 

-10 

-20 

-30 

-40 

-50 

-60 

-70 

-80 

0 0.05 0.1 0.15 0.2 

0 

-10 

-20 

-30 

-40 

-50 

-60 

-70 

-80 

0 100 200 300 

Bulk Density (g/cm3) Hg (ng/g) 

Shotyk et al., "Atmospheric Hg, Pb, and As in peat from Greenland...", revised version, April 15, 2003 

0 

-10 

-20 

-30 

-40 

-50 

-60 

-70 

-80 

-90 

0 5 10 15 20 25 

0 

-10 

-20 

-30 

-40 

-50 

-60 

-70 

1973 

AD 948 +/- 52 

AD 938 +/- 48 

Hg (ng/cm3) 

1964 

1981 

1992 

1962 

-80 

0 10 20 30 40 50 

Hg (ng/cm3) 

1956 

AD 1683 to AD 1928 

2395 +/- 45 14C yr BP 

3225 +/- 50 14C yr BP 

2935 +/- 50 14C yr BP 

2970 +/- 50 14C yr BP 

GREENLAND 

2485 +/- 45 14C yr BP 

2845 +/- 45 14C yr BP 

1958 

ca.17th to 20th C. 

disturbance by peat cutting 

DENMARK 

FIGURE 2

Depth (cm) 

Depth (cm) 

0 

25 

50 

75 

100 

0 

25 

50 

75 

100 

GREENLAND GREENLAND 

0.019 cm/yr 

-2000 -1000 0 1000 2000 

Calibrated Age (Year BC/AD) Calibrated Age (Year AD) 

DENMARK DENMARK 

-2000 -1000 0 1000 2000 

Calibrated Age (Year BC/AD) Calibrated Age (Year AD) 


Depth (cm) 

Depth (cm) 

0 

10 

20 

0.68 cm/yr 

0.20 cm/yr 

a b 

30 

0 

10 

20 

30 

1900 1920 1940 1960 1980 2000 

0.47 cm/yr 

0.21 cm/yr 

c d 

1900 1920 1940 1960 1980 2000 

FIGURE 3

Hg accumulation rate (µg/m2/yr) 

1000 

300 

100 

30 

10 

3 

1 

0.3 

a 

GL 

0.1 

-1500 -1000 -500 0 500 1000 1500 2000 

Calibrated Age (Year BC/AD) 




200 

150 

100 

50 

0 

250 

200 

150 

100 

50 

b 

1950 1960 1970 1980 1990 2000 

c 

Calibrated Age (Year AD) 

0 

1950 1960 1970 1980 1990 2000 


GL 

DK 

FIGURE 4

Depth (cm) 

a 

Pb flux (mg/m2/yr) 

b 

0 

-10 

-20 

-30 

-40 

-50 

-60 

-70 

-80 

0 100 200 300 

120 

100 

80 

60 

40 

20 

Pb (µg/g) 

0 

-10 

-20 

-30 

-40 

-50 

-60 

-70 

-80 

LLD = 3 µg/g 

0 10 20 30 

As (µg/g) 

DK 

total 

anthropogenic 

0 

1950 1960 1970 1980 1990 2000 


Anthropogenic Pb (% of total) 

c 

0 

-5 

-10 

-15 

-20 

0 20 40 60 80 

100 

80 

60 

GL 

Pb EF 


DK 

0 

-5 

-10 

-15 

-20 

0 20 40 60 80 

As EF 

40 

1950 1960 1970 1980 1990 2000 


DK 

FIGURE 5

Depth (cm) 

a 

Pb/Zr = 

UCC 

Depth (cm) 

Pb/Ti = 

UCC 

c 

0 

10 

20 

30 

40 

50 

60 

70 

0 

-5 

-10 

-15 

-20 

0 5 10 15 20 25 30 35 

Pb (µg/g) 

DK DK 

GL GL 

residual 

leached 

total 

80 

80 

0 50 100 150 200 1.13 1.14 1.15 1.16 1.17 1.18 1.19 

Pb (µg/g) 

0 

10 

20 

30 

40 

50 

60 

70 

Pb/Ti = 

4x UCC 

Pb 

isotope 

"residual" 

206Pb/207Pb 

0 

-5 

-10 

-15 

lithogenic 

anthropogenic 

-20 

0 100 200 

Pb (µg/g) 


Depth (cm) 

0 

10 

20 

30 

40 

50 

60 

70 

80 

90 

0 5 10 15 

90 

b Pb (µg/g) 

0 

-5 

-10 

-15 

lithogenic 

-20 

0 10 20 30 

As (µg/g) 

anthropogenic 

0 1999 

1994 

1979 

1970 

-5 

-10 

-15 

0 

10 

20 

30 

40 

50 

60 

70 

80 

1.20 1.24 1.28 1.32 

206Pb/207Pb 

leached 

residual 

-20 

1.13 1.14 1.15 1.16 1.17 

206Pb/207Pb 

1959 

1954 

DENMARK 

FIGURE 6

208Pb/206Pb 

2.20 

2.15 

2.10 

2.05 

U.K leaded gasoline 

DK peat samples 

1980 to 2000 

1950 to 1980 

coal, England 

2.00 

1.04 1.06 1.08 1.10 1.12 1.14 1.16 1.18 1.20 1.22 

206Pb/207Pb 

coal, Scotland 


sediment 

FIGURE 7

Depth (cm) 

a 

Depth (cm) 

b 

0 

-10 

-20 

-30 

-40 

-50 

-60 

-70 

-80 

0 

-10 

-20 

-30 

-40 

-50 

-60 

-70 

-80 

0 0.1 0.2 0.3 

DK 

Ti (%) 

0 1 2 3 4 

Ca (%) 

GL 

0 

-10 

-20 

-30 

-40 

-50 

-60 

-70 

-80 

0 

-10 

-20 

-30 

-40 

-50 

-60 

-70 

-80 

DK 

GL 

0 100 200 300 

0 

-10 

-20 

-30 

-40 

-50 

-60 

-70 

-80 

0 10 20 30 

Zr (µg/g) Y (µg/g) 

0 100 200 300 400 500 

0 

-10 

-20 

-30 

-40 

-50 

-60 

-70 

-80 

0 100 200 300 

Sr (µg/g) Mn (µg/g) 


1177 

0 

-10 

-20 

-30 

-40 

-50 

-60 

-70 

-80 

0 

-10 

-20 

-30 

-40 

-50 

-60 

-70 

-80 

0 10 20 30 40 50 

Rb (µg/g) 

0.0 0.5 1.0 1.5 2.0 2.5 

Fe (%) 

FIGURE 8

Depth (cm) 

a 

Depth (cm) 

b 

0 

-5 

-10 

-15 

-20 

0 0.05 0.1 0.15 0.2 

0 

-5 

-10 

-15 

DK 

Ti (%) 

-20 

0 1 2 3 4 

Ca (%) 

GL 

0 

-5 

-10 

-15 

-20 

0 20 40 60 80 100 120 

0 

-5 

-10 

-15 

0 

-5 

-10 

-15 

-20 

0 2 4 6 8 10 

Zr (µg/g) Y (µg/g) 

-20 

0 100 200 300 

0 

-5 

-10 

-15 

-20 

-20 

0 400 800 1200 0 1 2 3 4 5 

Sr (µg/g) Mn (µg/g) 


0 

-5 

-10 

-15 

-20 

0 10 20 30 

0 

-5 

-10 

-15 

Rb (µg/g) 

Fe (%) 

FIGURE 9

Depth (cm) 

Depth (cm) a 

b 

0 

-10 

-20 

-30 

-40 

-50 

-60 

-70 

-80 

0 

-10 

-20 

-30 

-40 

-50 

-60 

-70 

-80 

0 10 20 30 40 

Cu (µg/g) 

DK 

0 50 100 150 200 

Br (µg/g) 

GL 

0 

-10 

-20 

-30 

-40 

-50 

-60 

-70 

-80 

0 

-10 

-20 

-30 

-40 

-50 

-60 

-70 

-80 

0 10 20 30 40 

Cu/Y 

0 5 10 15 20 

Se (µg/g) 

0 

-10 

-20 

-30 

-40 

-50 

-60 

-70 

-80 

0 

-10 

-20 

-30 

-40 

-50 

-60 

-70 

-80 

0 20 40 60 80 100 120 

U (µg/g) 

0.0 0.5 1.0 1.5 

Se/Br 


0 

-10 

-20 

-30 

-40 

-50 

-60 

-70 

-80 

0 2 4 6 8 10 12 

U/Y 

FIGURE 10


Paper 9: Noernberg, T., Goodsite, M.E., Shotyk, W., An Improved Motorized Corer and Sample Processing 

System for Frozen Peat. In review at Arctic.

Saturday, 28 December 2002 (Revised June 19, 2003) 

In review at Arctic 

An Improved Motorized Corer and Sample Processing System for Frozen Peat 

TOMMY NOERNBERG 1 , MICHAEL E. GOODSITE* 2,1 and WILLIAM SHOTYK 3,4 

1 University of Southern Denmark, Department of Chemistry, Campusvej 55, DK-5230 Odense M, 

Denmark 

2 (previous address) National Environmental Research Institute, Department of Atmospheric 

Environment, Frederiksborgvej 399, DK-4000 Roskilde, Denmark 

3 Institute of Environmental Geochemistry, University of Heidelberg, INF 236, D-69120 Heidelberg, 

Germany 

4 (previous address) Geological Institute, University of Berne, Baltzerstrasse 1, CH-3012 Berne, 

Switzerland 

*Corresponding author 

1

ABSTRACT. An improved corer and associated equipment for obtaining continuous samples of 

frozen peat are described. In addition, new machines for precise slicing of frozen peat cores and 

accurate sub-sampling of volumetric slices are described and illustrated. 

Key words: Arctic, frozen peat, coring equipment, permafrost, and tundra. 

2


Recent research utilizing peat deposits in the High Arctic as archives of environmental 

contaminants has required a coring system to obtain continuous, undisturbed cores from deposits 

that are typically more than 2 m deep and are frozen below 10 - 20 cm. After examining sampling 

options, we decided to build our own corer, improving upon previous, unpublished designs such as 

the SIPRE/CRREL corer (Snow, Ice and Permafrost Research Establishment, U.S. Army Corps of 

Engineers which later became Cold Regions Research and Engineering Laboratory), and a similar 

corer described by Hughes and Terasmae (1963) which was used extensively by the Geological 

Survey of Canada (GSC) in the 1960’s and 1970’s (Blake 1964, 1974, 1977; Veillette and Nixon, 

1980). The motorized corer described by Hughes and Terasmae was based on improvements made 

to a hand-operated, SIPRE-type ice corer manufactured by AB Stålsvets, Sollentuna, Sweden. The 

corer employed by Hughes and Terasmae in the Yukon in 1962 (Hughes and Terasmae, 1963, pp. 

270-272 (as cited in Blake, 1964) had teeth of tooled steel. Blake replaced these with carboloy 

teeth, noting excellent results (Blake, 1964). Carboloy is an alloy containing cobalt, tungsten, and 

carbon teeth. This alloy is harder than steel; it is commonly used to cut steel, quartz, and other 

materials. Its hardness is little affected by heat. 

Visually, the corer that we built looks similar to the Austin Kovacs Enterprises (AKE) MARK 

V Ice Corer. However, the AKE MARK V corer is designed for coring ice, and the materials used 

for constructing it (stainless steel cutting teeth, anodized aluminum cutting head, fiberglass barrel) 

render it unsuitable for coring frozen peat which might contain sand or other mineral particles 

(Austin Kovacs, personal communication). Visual inspection of the peat cores from Bathurst Island 

collected by Blake in 1963 revealed abundant grains of mineral material, clearly indicating the need 

for robust coring equipment using suitable construction materials. Because there was no 

commercially available corer for collecting undisturbed cores of frozen peat, we decided to design 

and build our own. After a prototype corer had been built, we compared it to a GSC corer as 

3

described in technical drawings which were kindly provided to us by W. Blake, Jr.. However, we 

have remained with our original design, because of the improvements described below. 

4 

The new corer (Fig. 1) is a rugged, Teflon ® coated coring system which is able to 

continuously core frozen peat up to 10 m deep in 70 cm long sections of 9.7 cm diameter. The 

complete system (minus motor and fuel) weighs 26 kg. This weight includes 10 m worth of treated 

aircraft aluminum extension rods which are connected with male and female locking flange ends. 

These locking ends transfer the torque and are secured with a locking pin, in case the flanges 

deform or break. There is a core recovery system as a back-up and extra cutting teeth. Thus, one 

person can easily carry the complete corer. The corer is powered by a 1 man 1.6 h.p. or a 2 man 4.1 

h.p. motor which allow a 70 cm core section to be recovered in approximately 15 minutes 

(including packing time). 

NEW CORER DESIGN AND FEATURES 

Our system includes a quick-release motor drive shaft to corer tube coupling system 

(Fig. 2) so that samples can be removed immediately, preventing them from freezing into the 

Teflon ® coated, stainless steel (AISI 304) coring tube. The coating helps prevent the samples or 

chips of peat from freezing to the tube. Moreover, the hardened (Rockwell 62) steel (Sverker ® 21) 

cutting teeth cut the core 5 mm smaller than the tube diameter, allowing the core to easily slide into 

commercial polyethylene (PE) stockings. These are sealed, labelled and placed in capped plastic 

tubes, whose diameter allows the cores to fit snugly; helping to protect them from deformation 

during shipping and storage. 

Our coring system also includes spring-loaded cutting blades on the inside of the 

cutting head (Fig. 3): these are designed to cut horizontally across the bottom of the core section 

when the motor is reversed. Thus, this important feature quickly and uniformly cuts off the bottom 

of the core section; moreover, these cutting blades support the peat core while it is recovered from 

the hole. Therefore, the corer provides an effectively continuous record of peat accumulation, a 

feature that is especially important in palaeoenvironmental studies.

In addition, the coring system includes a compact, manual, backup core recovery 

system, designed at the GSC (as depicted in Blake, 1978, 1982; Veillette and Nixon, 1980), which 

can be assembled in the field in case it is needed. The kit consists of rods and two circular plates 

with three grabbing teeth on the plate at the bottom, and includes a conical metal wedge, which is 

attached to the end of the extension rods. By sliding the wedge forcefully between the core and the 

borehole sidewall, the core breaks at the bottom. The wedge is removed and the extension rods are 

attached to the core recovery frame described above. The frame slides over the core, and is gently 

pulled up, catching the teeth into the core, which is then lifted to the surface. We did not experience 

our corer getting stuck or fouled with sediment in the coring process in the field. 

SAMPLE PROCESSING 

In the laboratory, a slicing and volumetric sub-sampling system was designed and 

constructed to facilitate uniform sectioning of the frozen core, with easy access for cleaning 

between slices, to reduce the risk of contamination and to allow for easy blade replacement. Our 

cores are usually cut into sections of 1 cm, but the saw system can be adjusted to any thickness 

required. Prior to developing this system, we had sliced cores with a vertical band saw, on a sliding 

board. Unfortunately, the band saw retains debris in the blade housing and there is a risk of cross 

contamination between samples. Because of the potentially large differences in contaminant 

concentrations between modern and ancient peat samples, the risk of contamination is a serious 

drawback which needs to be reduced as much as possible. To overcome this problem, a band saw 

was mounted horizontally on a hinge, and attached to a frame that firmly holds the core, with an 

adjustable backstop for the core (Fig. 4). A peat core is placed within the frame and slid to the 

backstop after each slice – thus providing uniform thickness slices. Because the saw blade is 

mounted horizontally, the debris generated during cutting will fall into the basin used to collect 

waste. As an extra precaution to further reduce the risk of contamination, the saw blade can be 

rinsed each time without removing any blade housing covers. 

5

6 

Frozen slices are sub-sampled uniformly with a hand-operated stainless steel (AISI 

304) press, which recovers volumetric plugs for further analysis (Fig. 5). These plugs are then used 

as is for measurements of physical properties such as moisture content and bulk density, and for 

recovering plant macrofossils for 14 C age dating. Plugs can also be dried and milled to provide a 

homogeneous fine powder for subsequent chemical analyses, and for 210 Pb age dating. 

FIELD WORK 

The Environmental Chemistry Research Group at the University of Southern 

Denmark, in collaboration with Institute for Environmental Geochemistry at the University of 

Heidelberg, the Danish National Environmental Research Institute, the University of Berne, and the 

Geological Survey of Denmark and Greenland (GEUS), are investigating long-term records of 

contaminants (Hg, Cd, Pb and polycyclic aromatic hydrocarbons (PAHs)) in the Arctic, using 

permanently frozen peat deposits as environmental archives. The sites chosen to date are on 

Bathurst Island, Nunavut, Canada and Nordvestø, Carey Islands, Greenland, based on previous 

investigations at the sites by the Geological Survey of Canada (GSC) undertaken approximately 30- 

40 years ago (Blake 1964, 1974, 1977,1978,1982). The verified stratigraphy, physical and 

palynological determinations in the peat deposits made the sites prudent candidates for the present 

study, as did the advice offered by Dr. Weston Blake Jr. with respect to the sites and coring 

experiences. The excellent field descriptions provided by Dr. Blake enabled us to return to the exact 

sites where he cored and actually find the still-capped boreholes left after his investigations on 

Nordvestø and near his site at Bracebridge Inlet. By sampling at or near the previous sites, we were 

able to obtain samples that provide supplementary and complementary information to those taken 

for the previous studies. 

CONCLUSION 

Our sampling system is the result of laboratory testing prior to our Bathurst Island 

campaign during the summer of 2000, field-testing on Bathurst Island, subsequent improvements

and testing prior to the Carey Islands and a satisfactory field run in the Carey Islands in 2001 (Fig. 

6). Our coring system met or exceeded the original design goals: 1. portable - can easily be carried 

by an individual; 2. recovers continuous cores as long as necessary; 3. function well in the Arctic 

environment with all equipment amenable to service in the field, while wearing gloves, including 

replacement of the cutting head teeth; 4. robust - all components should be able to survive overland 

transport, weather extremes and the wear and tear associated with coring peat containing high 

concentrations of mineral matter; 5. efficient - in addition to coring peat, this corer is equally as 

effective in recovering ice from lenses and mineral matter (sand and silt) when present. Thus, it 

should find wide applications for recovering a broad range of Quaternary materials frozen in the 

Arctic. 

Further information about the coring and preparation system described here, including 

all technical drawings, are available without cost from the senior author (tno@chem.sdu.dk). 

REFERENCES 

BLAKE, W., Jr. 1964. Preliminary account of the glacial history of Bathurst Island, Arctic 

Archipelago. Geological Survey of Canada, Paper 64-30, 8 pp. 

BLAKE, W., Jr. 1974. Periglacial features and landscape evolution, central Bathurst Island, District 

of Franklin. In Report of Activities, Geological Survey of Canada, Paper 74-1B, 235-244. 

BLAKE, W, JR, 1977. Radiocarbon Age determinations from the Carey Islands, Northwest 

Greenland; in, Report of Activities, Part A; Geological Survey of Canada, Paper, 77-1A, 1977, 445- 

454. 

BLAKE, W., Jr. 1978. Coring of Holocene pond sediments at Cape Herschel, Ellesmere Island, 

Arctic Archipelago. In Current Research, Geological Survey of Canada, Paper 78-1C, 119-122. 

BLAKE, W., Jr. 1982. Coring of frozen pond sediments, east-central Ellesmere Island: a progress 

report. In Current Research, Geological Survey of Canada, Paper 82-1C, 104-110. 

7

HUGHES, O.L., and TERASMAE, J., 1963. Sipre Ice-corer for obtaining samples from 

permanently frozen bogs. Arctic 16(4):271-272. 

VEILLETTE, J.J., and Nixon, F.M. 1980. Portable drilling equipment for shallow permafrost 

sampling. Geological Survey of Canada, Paper 79-21, 35pp. 

8

ACKNOWLEDGEMENTS 

Development of this system was made possible by financial support from the Danish 

Environmental Protection Agency as part of the environmental support program DANCEA – 

Danish Cooperation for Environment in the Arctic and The University of Southern Denmark, 

Department of Chemistry. Stihl GmbH (Switzerland and Sweden) provided us with a BT 360 two- 

man motor, extension rods for initial laboratory and field-testing of the prototype corer, as well as 

relevant technical specifications. AN IARC grant (to 3d author) provided for the fieldwork on 

Bathurst Island and for this work, and Stihl Canada loaned us a BT 360 two-man motor and 

provided technical manuals, oil, and spare parts (thanks to K. Eberle, G. Quigg and E. Zynomirski). 

Special thanks to Weston Blake Jr. for his assistance while we were preparing the sampling plans, 

testing and building the corer and for providing his original peat cores from Bathurst Island for 

testing, in addition to critically reading this manuscript. For the fieldwork phases we would like to 

thank our scientific colleagues: Dr.’s A. Cheburkin, O. Bennike and E. Warncke. The help from 

team members who did not take part in the fieldwork, especially Dr. C. Lohse, N. Givelet and G. Le 

Roux (esp. Fig. 4 and 5.) is gratefully acknowledged. M.E.G. was supported by a Ph.D. fellowship 

from NERI and the Danish Research Agency, as a graduate student at the University of 

Copenhagen (Global Change Initiative Graduate School) under the academic supervision of H. 

Skov, S. Lindberg, and O.J. Nielsen, for which he is most grateful. 

9

FIGURE CAPTIONS 

Fig. 1. Complete kit, minus the motor, fuel and core packing tubes weighs 26 kg. The kit as shown 

contains enough treated aircraft-aluminum extension rods to core a 10 m-deep, frozen peat deposit. 

Extra sets of teeth, maintenance tools, a manual core recovery system, attachments to turn the core 

manually if needed and a plastic pusher attachment (for an extension rod), to manually push the 

core out of the tube into the polyethelene (PE) sock. A one-man motor (1.6 h.p.), 150 rpm, weighs 

approximately 6.8 kg, and a two-man motor (Stihl BT 360, 4.1 h.p.) weighs 25.9 kg. The gear 

reduction unit on the Stihl motor provides 50 r.p.m. Weight of fuel is 0.5 kg, which is sufficient for 

a few hours of operation. Packing material weighs an additional 3 kg (high density plastic core 

tubes with end cap). The cores themselves (70 cm x 9.7 cm) will weigh between 5 kg (pure frozen 

peat with bulk density of 1 g/cm 3 ) to 14 kg (sediment of bulk density 2.65 g/cm 3 ). 

Fig. 2. Quick-release motor drive shaft to corer-tube coupling system. It is important to remove the 

corer as quickly as possible from the coring tube. The quick release locks and unlocks the motor to 

the coring tube with the press of one button and has just three connecting points, allowing chips to 

fall into the tube when drilling below surface levels. 

Fig. 3. Spring-loaded cutting blades on the inside of the cutting head cut through the bottom of the 

core when the motor is stopped and the direction is manually reversed. These can easily be replaced 

or sharpened in the field. They stay together and are thick enough to bear the weight of the core 

when it is lifted out of the hole. 

Fig. 4. Horizontally hinged mounted band saw, on an aluminium frame, with adjustable backstop 

for varying the slice thickness. The bottom cover of the band saw is removed during operation, so 

that debris does not accumulate and allows quick and easy cleaning between slicing. 

10

Fig. 5. The stainless steel (AISI 304) press for volumetric sub sampling of frozen slices. A sub 

sample is shown beside the slice. 

Fig. 6. Senior author standing on top of a peat mound at the Nordvestø peat site (July, 2001) with 

the assembled corer and 1 man (1.6 h.p.) motor. Although one man can operate the corer, a two- 

man team is the minimum necessary for safe field operations in Arctic conditions, and this also 

expedites the labelling and packing of the cores. 

11

Fig. 1 

12

Fig. 2. 

13

Fig. 3. 

14

Fig. 4. 

15

Fig. 5. 

16

Fig. 6. 

17

FATE OF MERCURY IN THE ARCTIC Michael Evan ... - COGCI

You also want an ePaper? Increase the reach of your titles

Delete template?

Save as template?