Pesticides in air and in precipitation and effects on ... - Miljøstyrelsen

www2.mst.dk

Pesticides in air and in precipitation and effects on ... - Miljøstyrelsen

The Danish Environmental Protection Agency will, when opportunityoffers, publish reports ong>andong> contributions relatong>inong>g to environmentalresearch ong>andong> development projects fong>inong>anced via the Danish EPA.Please note that publication does not signify that the contents of thereports necessarily reflect the views of the Danish EPA.The reports are, however, published because the Danish EPA fong>inong>ds thatthe studies represent a valuable contribution to the debate onenvironmental policy ong>inong> Denmark.


PrefaceThe admission of pesticides on plants, water ong>andong> soil surfaces may come fromeither direct application or unong>inong>tentionally through spillage, run off ong>andong> dry or wetdeposition. Increasong>inong>g efforts has been made to protect agaong>inong>st the unong>inong>tendedpollution of the aquatic ong>andong> terrestrial environments. The present project wasstarted ong>inong> 1996 as a co-ordong>inong>ated research ong>inong>itiative to elucidate pesticidedistribution via the atmosphere ong>andong> consequences to plant life.To cover this broad range of ong>inong>formation, the project consisted of four parts tocover both emissions to the atmosphere, modellong>inong>g of spreadong>inong>g ong>inong> the atmosphere,deposition ong>inong> ong>precipitationong> ong>andong> ong>effectsong> on plants. The participatong>inong>g ong>inong>stitutes wereThe Danish Technological Institute (Laboratory system to determong>inong>e pesticidevolatility), National Environmental Research Institute (Modellong>inong>g of atmospherictransport ong>andong> deposition) ong>andong> Danish Institute of Agricultural Sciences (Analysis ofpesticides ong>inong> ong>precipitationong> ong>andong> Effects of herbicides ong>inong> ong>precipitationong> on plants). Theproject was supported by The National Environmental Protection Agency(Miljøstyrelsen).Besides the expected fong>inong>dong>inong>g of pesticides ong>inong> use ong>inong> Denmark ong>inong> the ong>precipitationong>,an unexpected experience from the project was the fong>inong>dong>inong>g of high concentrationsof the herbicide DNOC ong>inong> ong>precipitationong>. This compound has not been used ong>inong>Denmark song>inong>ce the end of the eighties. DNOC appeared ong>inong> raong>inong>water all over thesamplong>inong>g period, ong>andong> it was concluded, that DNOC hardly came from pesticide use,but was due to photochemical reactions with toluene ong>andong> nitrogen oxides ong>inong> theatmosphere. The concentrations of DNOC were much higher than for the otherpesticides ong>andong> probably also a number of other chemicals are deposited on plantsfrom the atmosphere.Information gathered ong>inong> such broad co-operative projects are of great importancefor the elucidation of the spreadong>inong>g of pesticides ong>andong> to calculate the load ofpesticides on agricultural ong>andong> non-agricultural areas. Further these values areimportant for the evaluation of pesticides ong>inong> relation to their registration.The current ong>inong>terest of the project was shown by an ong>inong>quiry from members ofparliament about spreadong>inong>g of pesticides to Denmark from other Europeancountries ong>andong> by the need for a technical report on causes of DNOC-pollution ong>inong>the atmosphere.A management group for the project was established with participants fromUniversity of Odense (Christian Lohse) ong>andong> The National EnvironmentalProtection Agency (Inge Vibeke Hansen). Chong>airong>man of the Management group wasErik Kirknel (DIAS), who also took the ong>inong>itiative to the project. The managementof the project was taken over by Arne Helweg ong>inong> 1998.5


A laboratory model system fordetermong>inong>ong>inong>g the volatility of pesticidesJørgen LarsenDanish Technological Institute, Environment7


Executive summaryIn the present study the volatilisation potential of pesticides is discussed ong>andong>illustrated based on both physical/chemical properties of the pesticides ong>andong> onexperimental volatilisation experiments. A laboratory model system for the directdetermong>inong>ation of the volatilisation of pesticides from different surfaces has beendeveloped. In this system it is possible to evaluate volatilisation of pesticidesrelative to each other. Experiments have been carried out with mecoprop-P(MCPP-P), mecoprop methylester (MCPP methyl), ong>andong> long>inong>dane as test substances.The volatilisation of these pesticides has been tested ong>inong> different volatilisationchambers, at different temperatures, ong>airong>flows ong>andong> from different surfaces.Furthermore, recommendations of important aspects that should be ong>inong>cluded ong>inong> thedevelopment of a new guidelong>inong>e on assessong>inong>g pesticide volatilisation are described.The maong>inong> conclusions regardong>inong>g the experimental set-up are drawn takong>inong>g practicalaspects of testong>inong>g ong>andong> the demong>andong> for simple cost-effective tests ong>inong>to account.At the begong>inong>nong>inong>g of the experiments, MCPP-P ong>andong> long>inong>dane were applied on a filterpapercorrespondong>inong>g to an application rate of 1.5; 15; 75, ong>andong> 150 kg ai ha-1 forMCPP-P ong>andong> 1; 10; 50; ong>andong> 100 kg ai ha-1 for long>inong>dane; correspondong>inong>g to therecommended application rate; 10×; 50×; ong>andong> 100× the recommended applicationrate. Durong>inong>g the experimental duration of 24 hours only ong>inong>significant volatilisationof MCPP-P was measured (less than 1%) while the volatilisation of long>inong>dane variedfrom 3.6 to 75.6 % of the applied dose dependong>inong>g of the concentrations. However,the actual amount of volatilised long>inong>dane was relatively constant at concentrationsabove 10× the recommended doses. The results ong>inong>dicate that MCPP-P does notappreciably volatilise from an artificial surface whereas long>inong>dane volatilises. Incontrast to MCPP-P, a great deal of the applied MCPP methyl volatilised under thepresent test conditions. Thus, when MCPP methyl was applied at a rate equivalentto the recommended application rate, about 90% of the applied dose werevolatilised durong>inong>g a 24 hours period. Furthermore, the results revealed comparablevolatilisation of the test substances when the experiments were carried out ong>inong>different volatilisation chambers with varyong>inong>g design.Interpretong>inong>g the volatilisation as a function of time from various volatilisationexperiments with pesticides, it seems that the ong>airong> samplong>inong>g 1, 3, 6, ong>andong> 24 hoursafter application asked for ong>inong> the German guidelong>inong>e are reasonable. Normally, highvolatilisation rates are seen withong>inong> the first few hours after application, ong>andong> the 24hours values after application are measured at a time when the volatilisationprocess has decreased considerably.To test the ong>inong>fluence of the temperature on the volatilisation of the two pesticideslong>inong>dane ong>andong> MCPP methyl were applied to filter paper at the recommended doseong>andong> the volatilisation was tested at 15oC ong>andong> 23oC under otherwise identicalconditions. The volatilisation of long>inong>dane ong>inong>creased significantly with temperature(from about 46% to 80% of the applied dose) whereas only a slight ong>inong>crease wasnoted for MCPP methyl (from 90% at 15oC to 96% of the applied dose at 23oC).At the recommended application rate, about 80% of the applied long>inong>dane volatilised,while at 10× the recommended dose only about 20% of the applied amountvolatilised correspondong>inong>g to the actual amount of long>inong>dane volatilisong>inong>g ong>inong>creasedwith a factor of 2.4. When the amount of MCPP methyl was ong>inong>creased with a factor11


of 10, the amount of MCPP methyl volatilisong>inong>g ong>inong>creased with a factor of 8.9illustratong>inong>g the higher volatilisation potential of this compound.Volatilisation of long>inong>dane ong>andong> MCPP methyl was also tested at two differentong>airong>flows. The two ong>airong>flows were 0.17 ong>andong> 0.67 m per sec ong>andong> the application ratecorresponded to the recommended application rate ong>andong> 10 × this dose. The resultsrevealed that an ong>inong>crease of the ong>airong>flow by a factor of 4 does not significantlyong>inong>crease the amount of volatilisation of the two pesticides.MCPP methyl ong>andong> long>inong>dane were applied as active ong>inong>gredients to the surface of twoartificial surfaces, filter paper representong>inong>g an absorbent surface ong>andong> a bowl ofstaong>inong>less steel illustratong>inong>g a non-absorbent surface. Volatilisation of the twopesticides after 24 hours were comparable when applied to the two selectedsurfaces. However, it seems that the volatilisation of MCPP methyl was faster atthe non-absorbent surface, while the rate of volatilisation of long>inong>dane wascomparable when applied at the two different surfaces. In contrast to the artificialsurfaces only a small fraction (about 15% of the applied dose) of the pesticidesvolatilised when applied to a typical Danish soil.12


Dansk sammendragI denne undersøgelse diskuteres fordampnong>inong>gspotentialet af pesticider ud frapesticidernes iboende fysisk/kemiske egenskaber samt ud fra eksperimentellefordampnong>inong>gsforsøg. Der er udviklet en eksperimentel forsøgsopstillong>inong>g, hvor deter muligt at måle fordampnong>inong>gen af pesticider fra forskellige overflader. I denneforsøgsopstillong>inong>g er det muligt at sammenligne fordampnong>inong>gspotentialet afforskellige pesticider. De eksperimentelle undersøgelser er blevet udført medmecoprop-P (MCPP-P), mecoprop methylester (MCPP methyl) og long>inong>dan, ogfordampnong>inong>gen af disse pesticider er blevet undersøgt i forskelligefordampnong>inong>gskamre, ved forskellige temperatur, luftflow og fra forskelligeoverflader. Baseret på erfarong>inong>gerne fra disse og ong>andong>res undersøgelser er derudarbejdet forslag til, hvilke elementer der skal ong>inong>ddraget i forbong>inong>delse medudarbejdelsen af en ny guidelong>inong>e til vurderong>inong>g af pesticiders fordampnong>inong>g.Ved starten af forsøgsrækken blev MCPP-P tilført filtrerpapir i følgendekoncentrationer 1,5; 15; 75 og 150 kg ai per ha. og long>inong>dan i 1, 10, 50, og 100 kg aiper ha, svarende til den anbefalede doserong>inong>g samt 10×, 50×, og 100× denanbefalede doserong>inong>g. Gennem forsøgsperioden på 24 timer blev der kun registrereten ubetydeligt fordampnong>inong>g af MCPP-P (mong>inong>dre end 1% af den tilførte mængde),mens fordampnong>inong>gen af long>inong>dan varierede fra 3,6 til 75,6% af den tilførte mængde(den totale mængde af long>inong>dan, der fordampede, var dog relativt konstant vedkoncentrationerne over 10× den anbefalede dosis). Disse resultater viser, at MCPP-P ikke forventes at fordampe fra sprøjtede overflader, mens long>inong>dan forventes atfordampe i store mængder. I modsætnong>inong>g til MCPP-P blev der iagttaget en megetstor fordampnong>inong>g af MCPP metyl. Således fordamper 90% af den tilførte mængdeong>inong>denfor 24 timer, når MCPP metyl tilføres i en koncentration svarende til denanbefalede mængde. Der blev iagttaget samstemmende resultater i forsøg udført iforsøgskamre med varierende design.Ved at gennemgå kong>inong>etikken hvormed forskellige pesticider fordamper gennemforsøgsperioden, virker prøveudtagnong>inong>gsong>inong>tervallerne på 1, 3, 6 og 24 timer, somangives i den tyske guidelong>inong>e, rimelige. Normalt ses der en høj fordampnong>inong>g deførste timer efter stoftilførslen, og efter 24 timer er fordampnong>inong>gen aftaget markant.For at undersøge temperaturens ong>inong>dflydelse på fordampnong>inong>gen blev der udførtforsøg med MCPP methyl og long>inong>dan ved både 15 o C og 23 o C. Ved disse forsøg blevder iagttaget en markant forøget fordampnong>inong>g af long>inong>dan (fra ca. 46% ved 15oC tilca. 80% af den tilførte mængde ved 23 o C), mens der kun blev iagttaget en mong>inong>drestignong>inong>g i fordampnong>inong>gen for MCPP methyl (fra 90% ved 15 o C til ca. 96% ved23 o C).Ved den anbefalede doserong>inong>g fordampede ca. 80% af den tilførte mængde efter 24timer, mens der ved en stoftilførsel på 10× den anbefalede doserong>inong>g kun bleviagttaget en fordampnong>inong>g på ca. 20% af den tilførte mængde svarende til, at denabsolutte mængde af long>inong>dan, der fordampede, blev forøget med en faktor på 2,4.Når den tilførte mængde af MCPP methyl blev forøget med en faktor 10, blev denabsolutte mængde MCPP methyl, der fordampede, forøget med en faktor 8,9hvilket illustrerer, at dette stof fordamper lettere.13


Fordampnong>inong>gen af long>inong>dan og MCPP methyl blev undersøgt ved to forskelligeluftgennemstrømnong>inong>ger (ved en lufthastighed på henholdsvis 0,17 og 0,67 meter pr.sekund). Resultaterne viste, at en forøgelse af hastigheden påluftgennemstrømnong>inong>gen med en faktor 4 ikke forøgede den mængde af pesticid, derfordampede.MCPP methyl og long>inong>dan blev tilført til to forskellige kunstige overflader,filtrerpapir, der illustrerer en absorberende overflade, og en metalbakke, somillustrerer en ikke absorberende overflade. Fordampnong>inong>gen af de to pesticider fra deto forskellige overflader var sammenlignelige efter 24 timer, men der blev iagttageten hurtigere fordampnong>inong>g af MCPP methyl fra den ikke absorberende overflade.Ved tilførsel af long>inong>dan og MCPP methyl til en typisk dansk jordtype fordamper dertil forskel fra de kunstige overflader kun en lille ong>andong>el (ca. 15%) af den tilførtepesticidmængde.14


1 IntroductionSong>inong>ce the late 1960s, losses of pesticides by volatilisation ong>andong> subsequentatmospheric transport have been ong>inong>creasong>inong>gly recognised as a process, sometimesof major importance, ong>inong> the loss of pesticides from the areas where they are applied(a process limitong>inong>g their effectiveness), ong>andong> as a pathway for general environmentalcontamong>inong>ation (extensive lists of literature are provided ong>inong> Noltong>inong>g. et al. 1988;Gottschild, et al. 1995; Jansma ong>andong> Long>inong>ders 1995; Spencer et al. 1973; Hartley,1969; Plimmer 1976; Willis et al. 1983). Therefore, volatilisation is assumed to bethe reason why some pesticides are widely distributed, contributong>inong>g to pollution ofong>airong>, raong>inong>, soil, surface, ong>andong> seawater. Recently, rather high concentrations ofpesticides ong>inong> raong>inong>water have been determong>inong>ed (Buser, 1990; Gath et al. 1992; Gathet al. 1993; Siebers et al. 1994; Glotfelty et al. 1990; Schomburg et al. 1991;Nations ong>andong> Hallberg 1992; Scharf ong>andong> Bächmann 1993, Feldong>inong>g et al. 1999).Volatilisation rates from plant or moist soil surfaces can be very large for morevolatile compounds, with losses approachong>inong>g 90% withong>inong> a few days (Taylor, 1978;Jansma ong>andong> Long>inong>ders, 1995). Although pesticides range ong>inong> volatility from fumigants,such as gaseous methyl bromide, to herbicides, with vapour pressures below 10 –6Pa, the same physical/chemical prong>inong>ciples are assumed to govern their rates ofvolatilisation. However, even though factors, ong>inong>fluencong>inong>g volatilisation ong>inong> prong>inong>ciple,have been clarified ong>andong> understood for a long time (Spencer et al. 1973; Hartley,1969; Plimmer, 1976), many problems concernong>inong>g volatilisation still remaong>inong>unsolved.Song>inong>ce volatilisation is an important factor ong>inong> the fate of pesticides ong>inong> theenvironment, it is necessary to have an experimental design to measure thisprocess. In 1990 a German Guidelong>inong>e on assessong>inong>g pesticide volatilisation wasdeveloped. This BBA guidelong>inong>e (Noltong>inong>g et al. 1990) was prompted not only by thelegal demong>andong>s for protectong>inong>g the ong>airong>, but also by ong>inong>creasong>inong>g public concern forpesticide residues found ong>inong> ong>precipitationong> ong>andong> ground water. It was decided to startwith a “liberal” guidelong>inong>e with only a few specific demong>andong>s, which then later oncould be refong>inong>ed, based on the results obtaong>inong>ed. Song>inong>ce then a number of methodshave been developed, rangong>inong>g from very simple to high-tech designs.The volatilisation rate is not only determong>inong>ed by the properties of the compound,the application rate, ong>andong> the crop, but also by other factors like meteorological ong>andong>soil conditions. As these other factors are highly variable field experiments for thesame compound, application rate ong>andong> crop can show highly variable results. Thismakes it difficult to determong>inong>e whether the volatilisation rate of a particularcompound is comparable with other compounds. Moreover, field experiments arerather expensive. For that reason it would be preferable if laboratory experimentscould be undertaken that could complement field experiments ong>andong> could ong>inong>dicatethe volatilisation potential of different compounds relative to each other.For this reason a simple laboratory model system has been designed for the presentstudy, ong>inong> which a wide range of outdoor conditions can be simulated. The fong>inong>alisedmethod has then been used to measure the volatility of selected pesticides fromdifferent surfaces ong>andong> under different “climatic” conditions such as temperature,wong>inong>d speed etc.15


2 Materials ong>andong> methods2.1 Test systemA simple laboratory model system for the direct determong>inong>ation of the volatilisationof pesticides has been developed for this study. The system permits a directdetermong>inong>ation of the volatilisation of a compound from a surface. Moreover, itgives the possibility to check the mass balance at the end of the study. Furthermore,the system permits determong>inong>ation of the volatilisation potential of differentcompound added on varyong>inong>g surfaces, at different ong>airong>flows, temperatures etc.2.1.1 Volatilisation chamberTwo chambers of different sizes have been used ong>inong> the present study. Bothchambers were made of staong>inong>less steel, ong>andong> each chamber was made up of 3 parts:The wong>inong>d tunnel (1) ong>inong> which a rectangular bowl of staong>inong>less steel (2) contaong>inong>ong>inong>gthe test sample can be placed. By the open side (3), opposite to the ong>airong> outlet, thebowl can be ong>inong>troduced ong>andong> removed for changong>inong>g of the test substrate. The openside also permits the cleanong>inong>g of the chamber ong>inong>ner walls.The test sample is placed ong>inong> the rectangular metal bowl that is ong>inong>troduced ong>inong>to thevolatilisation chamber immediately after application of the pesticides to thesurface. After the bowl has been placed ong>inong> the “wong>inong>d tunnel”, the open side of thechamber is closed with a plate of staong>inong>less steel ong>inong> which the ong>airong> ong>inong>let is placed. Theong>airong> ong>inong>let is connected to an ORBO-42 (large) adsorption tube to remove potentialcontamong>inong>ation of pesticides from the ong>inong>let ong>airong>.The volatilised pesticides are collected on three ORBO-42 (large) adsorption tubesconnected ong>inong> series to the outlet of the chamber. A given ong>airong> flow ong>inong> the “wong>inong>dtunnel” is ensured by a constant flow pump, SKC model 224-17SD connected atthe end of each series of adsorption tubes. Each series of adsorption tubes can bereplaced at different samplong>inong>g ong>inong>tervals durong>inong>g an experiment.2.1.2 System set-upTest compoundsSize of the chambersThe volatilisation of mecoprop –P (MCPP-P), mecoprop methylester(MCPP methyl) ong>andong> long>inong>dane have been tested ong>inong> this study. A range fong>inong>dong>inong>g test hasbeen carried out with concentrations correspondong>inong>g to the recommended dose, 10×,50×, ong>andong> 100× the dose to determong>inong>e which dose would be useful to apply ong>inong> thefong>inong>al test. In the fong>inong>al test the pesticides were applied at an application rateequivalent to 1 ong>andong> 10 kg ai ha -1 for long>inong>dane ong>andong> 1.5 ong>andong> 15 kg ai ha -1 for MCPPmethyl correspondong>inong>g to the recommended ong>andong> 10× the recommended dose.Initially the experiments were performed ong>inong> two volatilisation chambersof different sizes ong>inong> order to test the ong>inong>fluence of the size of the test chamber on thevolatilisation of the pesticides.One set of experiments was carried out ong>inong> a ”short” volatilisation chamber ong>inong> whichthe ”wong>inong>d tunnel” was 12 cm long, 3.5 cm wide ong>andong> 1.7 cm high. Another set ofexperiments was carried out ong>inong> a ”longer ong>andong> lower” chamber ong>inong> which the ”wong>inong>d16


tunnel” was 23 cm long, 3.6 cm wide ong>andong> 1.0 cm high. The metal bowl ong>inong> the”short” chamber had the followong>inong>g ong>inong>ner dimensions: 11.1 cm long, 2.25 cm wideong>andong> a depth of 0.3 cm. The dimension of the bowl ong>inong> the ”longer chamber” was22.9 cm long, 2.4 cm wide ong>andong> 0.3 cm deep.Temperatures ong>andong> ong>airong>humidityThe volatilisation of the pesticides were tested at two different temperatures.Normally the experiments were carried out ong>inong> a climatic room at a temperature of23C o +/- 0.5 o C ong>andong> an ong>airong> humidity of 50%. The volatilisation was also tested at15 o C +/- 0.5 o C ong>andong> ong>inong> these cases the volatilisation chamber was placed ong>inong> anong>inong>cubator ong>andong> the ong>airong> was passed through an ong>airong> glass bubbler contaong>inong>ong>inong>g a calciumchloride solution. The ong>airong> arrived ong>inong> the volatilisation chamber at about 50%relative humidity.Air temperature ong>andong> moisture were monitored durong>inong>g the study period usong>inong>g a Testo600, Testoterm.AirflowDifferent ong>airong>flows ong>inong> the “wong>inong>d tunnel” can be created by openong>inong>g differentnumbers of ong>airong> outlet from the volatilisation chamber which each was connected toa set of adsorption tubes ong>andong> an ong>airong> pump. This was done to ensure therecommended flow rate through the ORBO tubes. In the present study, an ong>airong>flowrate of 0.08 m/sec ong>andong> 0.31 m/sec was tested ong>inong> the “short chamber”. In the “longerong>andong> lower chamber”, an ong>airong>flow rate of 0.17 m/sec ong>andong> 0.67 m/sec was tested.The ong>airong>flow was measured usong>inong>g a Termo-Anemometer, Alnor GGA-65P.Application on different The pesticides were applied to three different surfaces:surfaces• Filter paper (Whatman 1) illustratong>inong>g an artificial absorbent surface.• Directly to the bowl of staong>inong>less steel illustratong>inong>g a non-absorbent surface.• A coarse song>andong>y soil from the Danish National Agricultural Research station atJyndevad, typical Danish soil. The relevant soil characteristics are described ong>inong>table 2.1. The test was carried out usong>inong>g 8.5 g sieved soil less than or equal to 2mm which has been stored ong>airong>-dried ong>andong> re-equilibrated with deionised waterjust before the experiment to give an overall moisture content of about 35 percent of the dry weight.17


Table 2.1Soil characteristics for the Jyndevad soil.Jordkarakteristika for Jyndevadjorden.Clay 3.9 %Silt 4.1%Song>andong> 89.0%Humus 3.0%Soil water capacity (a) 34.0%PH 6.52.2 Performance of the test2.2.1 Chemicals ong>andong> stong>andong>ardsAll solvents used were of analytical grade. MCPP-P, MCPP methyl ong>andong> long>inong>danewere obtaong>inong>ed from Dr. Ehrenstorfer GmbH, Germany. Hexachlorbenzene- 13 C 6 ong>andong>dichlorprop- 13 C 6 were obtaong>inong>ed from Cambridge Isotope laboratories, Woburn,MA, USA. Stong>andong>ards were prepared usong>inong>g stong>andong>ard volumetric techniques.2.2.2 Preparation ong>andong> spikong>inong>g of the samplesThe test surface was placed ong>inong> the rectangular metal bowl ong>andong> spiked with theanalytes dissolved ong>inong> acetone. After evaporation of acetone at room temperature themetal bowl was placed ong>inong> the volatilisation chamber. The application time was lessthan 3 mong>inong>utes. For tests with filter paper the analytes were dissolved ong>inong> 250 ong>andong>500 µl acetone for the “short” ong>andong> “longer” chamber, respectively, ong>andong> added byseveral applications by a Hammilton syrong>inong>ge. For tests ong>inong> which the pesticides wereapplied directly to the metal bowl or to soil the analytes were dissolved ong>inong> 3 mlacetone ong>andong> applied to the metal bowl or to the soil by several applications.Pre-test experimentsTo ong>inong>vestigate whether some part of the pesticides applied to the test surface wouldevaporate together with the acetone durong>inong>g application before the test sample wasplaced ong>inong> the volatilisation chamber, spiked samples were placed at roomtemperature for 30 mong>inong>utes after which they were analysed. To test the recoveryfrom the adsorption tubes, the analytes were dissolved ong>inong> 200 µl acetone ong>andong> addedto the ORBO-42 (large) adsorption tubes after which the tubes were connected toan ong>airong> pump. Airflow of 2 l/h was maong>inong>taong>inong>ed durong>inong>g 24 hours. After this period, theabsorption tubes were analysed.2.2.3 Sample preparationSamplong>inong>g ong>andong> samplestreatmentWhen several samplong>inong>gs were takong>inong>g durong>inong>g an experiment the adsorptiontubes were immediately replaced with a new trap set-up. At the end of theexperiments each set of adsorption tubes was removed from the chamber ong>andong> eachof the three tubes ong>inong> a set was analysed, separately.At the end of an experiment the volatilisation chamber ong>andong> the bowl ong>inong>cludong>inong>g thetest surface applied with the pesticides were extracted with 200.0 ml methylenechloride ong>inong> a Pyrex bottle by shakong>inong>g for 30 mong>inong>utes. 2.0 ml of the extract wasdiluted ong>andong> spiked with ong>inong>ternal stong>andong>ard(s).The adsorption tubes were extracted with 5.0 ml methylene chloride by shakong>inong>g for5 mong>inong>utes. 2.0 ml of the extract was diluted ong>andong> spiked with ong>inong>ternal stong>andong>ard(s).18


The sample extracts to be analysed for MCPP-P ong>andong> long>inong>dane were spiked with 20µl HCB- 13 C 6 (500 ng/µl) ong>andong> 20 µl dichlorprop- 13 C 6 (500 ng/µl) as ong>inong>ternalstong>andong>ards. MCPP-P ong>andong> dichlorprop- 13 C 6 were methylated with diazomethan beforeanalysong>inong>g.The sample extracts to be analysed for MCPP methyl ong>andong> long>inong>dane were spiked with20 µl HCB- 13 C 6 (500 ng/µl) as ong>inong>ternal stong>andong>ard.The extracts were analysed by GC-MS-SIM.2.2.4 GC-MS-SIM analysisThe extracts were analysed by GC-MS-SIM ong>inong> the electron impact mode (EI, 70eV).Ions for qualification ong>andong> quantification were selected on the basis of the massspectra of the components. The ions 219 ong>andong> 221 amu were selected for long>inong>dane,169, 228 ong>andong> 230 amu for MCCP methyl, 168 ong>andong> 254 amu for dichlorprop- 13 C 6ong>andong> 290 ong>andong> 292 for HCB- 13 C 6 . The quantification was done on the basis of thearea ratio analyte/ong>inong>ternal stong>andong>ard for the selected ions.The mass scale was calibrated usong>inong>g PFTBA as a calibration gas.The calibration curves were established on solutions ong>inong> methylene chloride of theanalytes. This was done on the basis of experiments showong>inong>g that the extractionefficiency of the analytes from the filter paper was 92-96 % ong>andong> from theadsorption tubes 87-93 %.Injection technique: on-column ong>inong>jection (35°C). Precolumn: 1 meter 0.25 mm id,0.25 µm RTX-5. Column: WCOT fused silica 50m × 0.25 mm id, 0.4 µm CP Sil13CB. Temperature programmong>inong>g: 35°C (0.5 mong>inong>)-280°C, 30°C/mong>inong>. Carrier gas:Helium (25 psi). GC-MS: HP5971.19


3 ResultsThe test design ong>inong>cludes two ong>inong>dependent tests (on different days). Thus, all datarepresent the mean of two experiments. All experiments were carried out at 23°Cunless somethong>inong>g else has been stated. The detection limit for all analytes was 0.05µg.3.1 Pre-test of the experimental set-upTo ong>inong>vestigate whether some of the pesticides applied to the test surface wouldevaporate together with the acetone durong>inong>g application before the test sample wasplaced ong>inong> the volatilisation chamber, spiked samples were placed at roomtemperature for 30 mong>inong>utes after which they were analysed. The analyses showedthat the pesticides did not evaporate together with the acetone ong>andong> a recovery ofmore than 90% was found ong>inong> these experiments.The extraction efficiency from the adsorption tubes was tested ong>andong> the resultsrevealed that the recovery was more than 95%.3.2 Results of the experiments ong>inong> a short volatilisationchamber after 24 hours3.2.1 Volatilisation of MCPP-P applied to filter paper at different concentrationsThe volatilisation of MCPP-P applied to filter paper was tested at differentconcentrations: 0.015, 0.15, 0.75, ong>andong> 1.5 mg MCPP-P/cm 2 , correspondong>inong>g to therecommended doses, 10×, 50×, ong>andong> 100× the recommended dose. The resultsrevealed that less than 0.06% of the applied pesticide volatised after 24 hours ong>inong> thevolatilisation chamber with ong>airong>flow of 0.08 m/sec. At the end of the experimentbetween 89 ong>andong> 110% of the applied pesticide could be extracted from the filterpaper ong>andong> the volatilisation chamber.3.2.2 Volatilisation of long>inong>dane applied to filter paper at different concentrationsWhen long>inong>dane was applied to a filter paper at concentrations of 0.01, 0.1, 0.5, ong>andong>1.0 mg/cm 2 correspondong>inong>g to the recommended application rate, 10×, 50×, ong>andong>100× the recommended dose a significant amount of the added pesticides volatised.The ong>airong>flow was 0.08 m/sec. The results are depicted ong>inong> table 3.1.Table 3.1Volatilisation of long>inong>dane after a 24 hours period. S.D. less than 10%.Fordampnong>inong>gen af long>inong>dan efter en periode på 24 timer. S.D. mong>inong>dre end 10%Applied concentrationof long>inong>daneVolatilisation oflong>inong>dane (% ofExtraction frompaper ong>andong>Total recovery(% of appliedRecommended dose10× recomm. dose50× recomm. dose100× recomm. doseapplied dose)75.631.28.13.6chamber (ong>inong> %)24.060.080.886.8dose)99.691.288.990.420


Figure 3.1% Volatilisation80706050403020100-100 20 40 60 80 100Recommended dose XThe volatilisation of long>inong>dane relative to the applied amount (R 2 = 0.98).Fordampnong>inong>g af long>inong>dan relativt i forhold til tilført mængde (R 2 = 0,98).The volatilisation of long>inong>dane under the present experimental conditions correspondsto 7.6, 31.2, 40.5 ong>andong> 36 µg long>inong>dane/cm 2 for the recommended doses, 10×, 50× ong>andong>100× the recommended doses, respectively. Thus, the results revealed that acomparable amount of long>inong>dane is volatised ong>inong> concentrations correspondong>inong>g to 10×50× ong>andong> 100× the recommended doses. It was therefore decided to ong>inong>vestigate thevolatilisation of the pesticides at the recommended dose ong>andong> 10× the concentrationong>inong> the further experiments.3.2.3 Volatilisation of MCPP methyl ong>andong> long>inong>dane applied to different artificial surfacesong>andong> at different ong>airong>flowsVolatilisation of MCPP Song>inong>ce no volatilisation could be demonstrated ong>inong> the MCPP-P experiments itmethyl from a non- was decided to start experiments with MCPP methyl. The results from theseabsorbent surface experiments revealed that a great deal of the applied MCPP methyl volatilisedunder the present test conditions. Application of MCPP methyl at a concentrationcorrespondong>inong>g to 10× the recommended dose to a non-absorbent metal surfaceresulted ong>inong> a volatilisation of the pesticide correspondong>inong>g to 85.5% of the appliedamount after 24 hours (ong>airong> flow 0.08 m/sec). Extraction of pesticides from themetal surface ong>andong> the chamber revealed 4.8 % of the applied dose resultong>inong>g ong>inong> arecovery of 90.3%.Volatilisation of long>inong>dane The volatilisation of long>inong>dane from a non-absorbent metal surface was testedfrom a non-absorbent with ong>airong>flow of 0.08 m/sec ong>andong> at a pesticide concentration correspondong>inong>g tosurface10× the recommended application. The results revealed that after 24 hours 30% ofthe applied dose was volatilised while 61% could be extracted from the metalsurface ong>andong> the chamber. This gave a recovery of 91%.Volatilisation of long>inong>dane The volatilisation of long>inong>dane ong>andong> MCPP methyl applied to filter paper at aong>andong> MCPP methyl with concentration correspondong>inong>g to 10× the recommended dose was tested ong>inong> thean ong>inong>creased ong>airong>flow volatilisation chamber with an ong>inong>creased ong>airong>flow (0.31 m/sec). The results revealedthat after 24 hours 34.1% of the applied long>inong>dane were volatilised ong>andong> 64% could beextracted from the paper ong>andong> chamber givong>inong>g a recovery of 98%. For MCPP methyl95,8% of the applied dose was volatilised ong>andong> 0.3% was found on the filter paperong>andong> ong>inong> the chamber a recovery was 96%.21


3.3 Results of volatilisation of long>inong>dane ong>andong> MCPP methyl ong>inong> a “longer ong>andong> lower”volatilisation chamber3.3.1 Volatilisation of long>inong>dane ong>andong> MCPP methyl applied at recommended dose durong>inong>gan experimental test of 30 hoursFigure 3.2 illustrates the accumulated volatilisation of long>inong>dane ong>andong> MCPP methylfrom filter paper durong>inong>g a 30 hours period at a temperature of 23°C ong>andong> an ong>airong>humidity of 50%. The application rate was 0.01 mg/cm 2 for long>inong>dane ong>andong> 0.015mg/cm 2 for MCPP methyl, correspondong>inong>g to the recommended application rate.The ong>airong>flow ong>inong> the chamber was 0.17 m/s. About 10% ong>andong> 30% of the applied dosehad volatilised durong>inong>g the first hour after application of long>inong>dane ong>andong> MCPP methyl,respectively. After 8 hours about 50% of the applied dose of long>inong>dane wasvolatilised correspondong>inong>g to 2/3 of the total amount which volatised durong>inong>g a 30hours period. For MCPP methyl about 80% of the applied pesticide was volatisedafter 8 hours, correspondong>inong>g to about 9/10 of the total amount durong>inong>g a 30 hourstest. At the end of the experiment 14.5% ong>andong> 4.1% of the applied pesticides couldbe extracted from the filter paper ong>andong> the chamber for long>inong>dane ong>andong> for MCPPmethyl, respectively. Thus, the recovery ong>inong> these experiments was 90% for bothpesticides.% of applied dosis10090807060504030201001 h 3 h 6 h 8 h 24 h 30 hHours after treatmentLong>inong>daneMCPPFigure 3.2The figure illustrates the accumulated volatilisation of long>inong>dane ong>andong> MCPP methyl from filter paper durong>inong>g a 30hours period. The application rate was 0.01 mg/cm 2 for long>inong>dane ong>andong> 0.015 mg/cm 2 for MCPP methyl. The ong>airong>flowong>inong> the chamber was 0.17 m/s. All data represent the mean of at least two experiments (S.D. less than 5%).Figuren illustrerer den akkumulerede mængde af fordampet long>inong>dan og MCPPmethyl fra filterpapir over en periode på 30 timer. Stofferne blev tilført i enkoncentration svarende til 0,01 mg/cm 2 for long>inong>dan og 0,015 mg/cm 2 for MCPPmethyl. Luftgennemstrømnong>inong>gen i kammeret var 0,17 m/s. Alle data er middeltalletaf mong>inong>dst to forsøg (S.D. er mong>inong>dre end 5%).In table 3.2 the volatilisation is expressed as the amount of pesticide volatili-sedper hour durong>inong>g the experiment. The results revealed that the rate of volatilisation22


was significantly higher for MCPP methyl compared with the rate found forlong>inong>dane durong>inong>g the first 6 hours after application. For MCPP methyl a very high butrapid decreasong>inong>g volatilisation was seen durong>inong>g the first 8 hours. For long>inong>dane therate of volatilisation was much lower, even though the same pattern was seen, ong>inong>which a higher rate of volatilisation occurred durong>inong>g the first 8 hours afterapplication. However ong>inong> this case, the rate of decrease ong>inong> the volatilisation wasslower than the rate found for MCPP methyl ong>andong> after 8 hours the rate ofvolatilisation of long>inong>dane was higher than the one found for MCPP methyl.Table 3.2Volatilisation of long>inong>dane ong>andong> MCPP methyl per hour durong>inong>g a 30 hours period(S.D. less than 10%).Fordampnong>inong>g af long>inong>dan og MCPP methyl udtrykt pr. time i en periode på 30 timer (S.D. er mong>inong>dre end 10%).Hoursaftertreatment0-11-33-66-88-2424-30Volatilisation of long>inong>daneper hour (% of applieddose)7.87.05.73.81.60.6Volatilisation of MCPPmethyl per hour (% ofapplied dose)28.615.45.62.00.40.13.3.2 Volatilisation of long>inong>dane ong>andong> MCPP methyl at different temperaturesIn order to test the ong>inong>fluence of the temperature on the volatilisation of the twopesticides long>inong>dane ong>andong> MCPP methyl were applied to filter paper at therecommended dose ong>andong> the volatilisation was tested at 15°C ong>andong> 23°C underotherwise identical test conditions.As shown ong>inong> figure 3.3, the volatilisation of long>inong>dane decreased significantly whenthe temperature was decreased from 23°C to 15°C. Each column illustrates the totalamount of pesticide analysed after the test period. The lowest dark part of eachcolumn illustrates the amount of pesticide extracted from the filter paper ong>andong> thevolatilisation chamber, ong>andong> the upper light part illustrates the amount of pesticidesvolatilised. The application rate was 0.01 mg/cm 2 for long>inong>dane ong>andong> 0.015 mg/cm 2 forMCPP methyl. The ong>airong>flow ong>inong> the chamber was 0.17 m per sec. Thus, at 23°C about80% of the applied long>inong>dane was volatised after 24 hours whereas only 46% wasvolatised at 15°C. In contrast the amount of MCCP, that volatised durong>inong>g a 24hours period, only decreased from 96% at 23°C to 90% of the applied dose at15°C.23


100%Long>inong>daneMCPP80%60%40%20%0%15C 23C 15C 23CDegree CelsiusFigure 3.3The figure illustrates the volatilisation ong>andong> total recovery (whole column) oflong>inong>dane ong>andong> MCPP methyl at 15°C ong>andong> 23°C from filter paper after a 24 hoursperiod. Each column illustrates the total amount of pesticide analysed after the testperiod (total recovery). The upper light part illustrates the amount of pesticidesvolatilised. All data represent the mean of at least two experiments (S.D. less than10%).Figuren illustrerer fordampnong>inong>gen og den totale “recovery” af long>inong>dan og MCPP methyl ved 15°C og 23°C frafilterpapir efter en periode på 24 timer. Hver søjle illustrerer den totale mængde pesticid, der blev analyseretefter testperiodens ophør (total “recovery”). Den lyse del illustrerer den mængde pesticid, der er fordampet. Datarepræsenterer et gennemsnit af mong>inong>dst to forsøg (S.D. er mong>inong>dre end 10%).3.3.3 Volatilisation of long>inong>dane ong>andong> MCPP methyl applied at two different dosesFigure 3.4 illustrates the volatilisation ong>andong> the total recovery of long>inong>dane ong>andong> MCPPmethyl from filter paper after a 24 hours period. Each column illustrates the totalamount of pesticide analysed after the test period (total recovery). The lowest darkpart of each column illustrates the amount of pesticide extracted from the filterpaper ong>andong> the volatilisation chamber, ong>andong> the upper light part illustrates the amountof pesticides volatilised. The ong>airong>flow ong>inong> the chamber was 0.17 m per sec. Theapplication rate was 0.01 mg/cm 2 ong>andong> 0.1 mg/cm 2 for long>inong>dane ong>andong> 0.015 ong>andong> 0.15mg/cm 2 for MCPP methyl, correspondong>inong>g to the recommended application rate ong>andong>10× the recommended application rate. At the recommended application rate80.5% of the applied long>inong>dane volatilised correspondong>inong>g to 0.439 mg, while at 10×the recommended doses only about 19.6% of the applied amount volatisedcorrespondong>inong>g to 1.068 mg long>inong>dane. At the recommended application rate 95.9% ofthe applied MCPP methyl volatilised while the correspondong>inong>g values were 85.5% at10× the recommended dose.24


Long>inong>daneMCPP100%100%Distribution of pesticides80%60%40%20%Distribution of pesticides80%60%40%20%0%Recom.dosis10 x recom.dosis0%Recom.dosis10 x recom.dosisFigure 3.4The figure illustrates the volatilisation ong>andong> the total recovery of long>inong>dane ong>andong> MCPPmethyl at two different doses from filter paper after a 24 hours period. Eachcolumn illustrates the total amount of pesticide analysed after the test period (totalrecovery). The upper light part illustrates the amount of pesticides volatilised. Theapplication rate was 0.01 mg/cm 2 ong>andong> 0.1 mg/cm 2 for long>inong>dane ong>andong> 0.015 ong>andong> 0.15mg/cm 2 for MCPP methyl. All data represent the mean of at least two experiments(S.D. less than 8%).Figuren illustrerer fordampnong>inong>gen og den totale recovery af long>inong>dan og MCPP methyl ved to forskellige doserong>inong>ger24 timer efter tilførsel til filterpapir. Hver søjle illustrerer den totale mængde pesticid, der blev analyseret eftertestperiodens ophør (total “recovery”). Den lyse del illustrerer den mængde pesticid, der er fordampet.Doserong>inong>gen var 0,01 mg/cm 2 og 0,1 mg/cm 2 for long>inong>dan og 0,015 mg/cm 2 og 0,15 mg/cm 2 for MCPP methyl. Datarepræsenterer et gennemsnit af mong>inong>dst to forsøg (S.D. var mong>inong>dre end 8%).3.3.4 Volatilisation of long>inong>dane ong>andong> MCPP methyl at different ong>airong> flowFigure 3.5 illustrates the volatilisation ong>andong> the total recovery of long>inong>dane ong>andong> MCPPmethyl from filter paper after a 24 hours period at two different ong>airong>flows ong>andong> at twodifferent application rates. Each column illustrates the total amount of pesticideanalysed after the test period (total recovery). The lowest dark part of each columnillustrates the amount of pesticide extracted from the filter paper ong>andong> thevolatilisation chamber, ong>andong> the upper light part illustrates the amount of pesticidesvolatilised. The two ong>airong>flows were 0.17 ong>andong> 0.67m per sec ong>andong> the application ratewas 0.01 mg/cm 2 ong>andong> 0.1 mg/cm 2 for long>inong>dane ong>andong> 0.015 ong>andong> 0.15 mg/cm 2 forMCPP methyl, correspondong>inong>g to the recommended application rate ong>andong> 10× therecommended application rate. The results revealed that an ong>inong>crease of the ong>airong>flowby a factor of 4 does not significantly ong>inong>crease the amount of volatilisation of thetwo pesticides at the two selected concentrations.25


100%Long>inong>daneMCPP2 L 8L 2L 8 L 2 L 8L 2L 8 L100%Distribution of pesticides80%60%40%20%Distribution of pesticides80%60%40%20%0%Recom.dosisRecom.dosis10 xrecom.dosis10 xrecom.dosis0%Recom.dosisRecom.dosis10 xrecom.dosis10 xrecom.dosisFigure 3.5The figure illustrates the volatilisation ong>andong> the total recovery of differentconcentrations of long>inong>dane ong>andong> MCPP methyl at two different ong>airong>flows. Each columnillustrates the total amount of pesticide analysed after the test period (totalrecovery). The upper light part illustrates the amount of pesticides volatilised. Theapplication rate was 0.01 mg/cm 2 ong>andong> 0.1 mg/cm 2 for long>inong>dane ong>andong> 0.015 ong>andong> 0.15mg/cm 2 for MCPP methyl. The ong>airong>flow ong>inong> the chamber was 0.17 ong>andong> 0.67 m/s.,respectively (correspondong>inong>g to an ong>airong> exchange rate of 2 ong>andong> 8 l/mong>inong>). All datarepresent the mean of at least two experiments (S.D. less than 10%).Figuren illustrerer fordampnong>inong>gen og den totale recovery af forskellige koncentrationer af long>inong>dan og MCPP methylved to forskellige lufthastigheder. Hver søjle illustrerer den totale mængde pesticid, der blev analyseret eftertestperiodens ophør (total “recovery”). Den lyse del illustrerer den mængde pesticid, der er fordampet. Stofferneblev tilført i koncentrationer på 0,01 og 0,1 mg/cm 2 for long>inong>dan og 0,015 samt 0,15 mg/cm 2 for MCPP methyl. Deto undersøgte lufthastigheder var henholdsvis 0,17 og 0,67 m/s.(svarende til en luftgennemstrømnong>inong>gshastighedpå 2 og 8 l/mong>inong>). Data er et gennemsnit af mong>inong>dst to forsøg (S.D. er mong>inong>dre end 10%).3.3.5 Volatilisation of long>inong>dane ong>andong> MCPP methyl applied to different surfacesArtificial surfacesTable 3.3 illustrates the volatilisation of long>inong>dane ong>andong> MCPP methyl applied to twodifferent artificial surfaces. The pesticides were applied to a filter paper illustratong>inong>gan absorbent surface ong>andong> to a metal plate illustratong>inong>g a non-absorbent surface.Measurements were taken durong>inong>g a 24 hours period. The application rate was 0.01mg/cm 2 for long>inong>dane ong>andong> 0.015 mg/cm 2 for MCPP methyl. The ong>airong>flow ong>inong> thechamber was 0.17 m per sec. The recovery ong>inong> the experiments with filter paper was91.4% for long>inong>dane ong>andong> 90.3 for MCPP methyl. For the non-absorbent surface arecovery of 90.3 ong>andong> 91.7% was found for long>inong>dane ong>andong> MCPP methyl, respectively.All data represent the mean of at least two experiments (S.D. less than 10%).26


Table 3.3Volatilisation of long>inong>dane ong>andong> MCPP methyl applied to an absorbent surface ong>andong> a non-absorbent metal surfacedurong>inong>g a 24 hours period (S.D. less than 10%).Fordampnong>inong>g af long>inong>dan og MCPP methyl tilført en absorberende overflade og en ikke absorberendemetaloverflade i en 24 timers periode (S.D. mong>inong>dre end 10%).Hoursaftertreatment0-11-33-66-24Volatilisation of long>inong>dane(% of applied dose)Applied toabsorbentpapersurface7.814.017.233.5Applied tonon absorbentmetal surface10.117.213.033.3Volatilisation of MCPPmethyl(% of applied dose)Applied toabsorbentpapersurface28.630.716.810.3Applied tononabsorbentmetalsurface38.238.29.25.1Total 72.5 73.5 86.4 90.6The results revealed that the volatilisation of each pesticide was comparable whenapplied to the two selected surfaces. However, even though the total amount ofvolatilised pesticides was the same after 24 hours it seems that the volatilisation ofat least MCPP methyl was faster at the non-absorbent surface durong>inong>g the ong>inong>itialvolatilisation phase. Thus, 3 hours after application 76.4% of the applied MCPPmethyl was volatilised when applied to the metal surface compared with 59% whenapplied to the filter paper.Soil surfacesTable 3.4 illustrates the volatilisation of long>inong>dane ong>andong> MCPP methyl applied to atypical Danish soil. Measurements were made durong>inong>g a 24 hours period. Theapplication rate was 0.01 mg/cm 2 for long>inong>dane ong>andong> 0.015 mg/cm 2 for MCPP methyl.The ong>airong>flow ong>inong> the chamber was 0.17 m per sec.Table 3.4Volatilisation of long>inong>dane ong>andong> MCPP methyl applied to a typical Danish soil durong>inong>g a 24 hours period. All datarepresent the mean of at least two experiments (S.D. less than 10%).Fordampnong>inong>g af long>inong>dan og MCPP methyl tilført en typisk dansk jord i en 24 timersperiode. Alle data repræsenterer et gennemsnit af mong>inong>dst to forsøg (S.D. mong>inong>dreend 10%).HoursafterVolatilisation of pesticides(% of applied dose)treatment Long>inong>dane MCPP0-11-33-66-241.72.62.06.7methyl3.33.22.47.0Total 13.0 15.9In contrast to the artificial surfaces only a small fraction of the pesticides volatisedwhen applied to a soil. Furthermore, it was ong>inong>terestong>inong>g to notice that no marked27


difference was seen between the volatilisation of long>inong>dane ong>andong> MCPP methyl whenapplied to the soil. At the end of the experiment 76.6 % ong>andong> 64.9 % of the appliedpesticides could be extracted from the soil ong>andong> the chamber for long>inong>dane ong>andong> MCPPmethyl, respectively. Thus, the recovery ong>inong> these experiments was 89.6% forlong>inong>dane ong>andong> 80.8% for MCPP methyl.28


4 DiscussionVolatilisation plays an important role ong>inong> the dispersion of pesticides ong>inong> theenvironment. Volatilisation also leads to a rapid transport ong>andong> distribution ofpesticides ong>inong> the atmosphere, resultong>inong>g ong>inong> considerable wet ong>andong> dry deposition.Song>inong>ce volatilisation is an important factor ong>inong> the fate of pesticides ong>inong> theenvironment, it is necessary to have a method to evaluate the process.Accordong>inong>g to Thomas et al. (1990), volatilisation can be defong>inong>ed as the process bywhich a compound evaporates ong>inong> the vapour phase to the atmosphere from anotherenvironmental compartment. For pesticides ong>andong> other chemicals potential volatilityis related to physical/chemical parameters such as vapour pressure ong>andong> Henry´s lawconstant of the compound. However, actual volatilisation rate will also depend onenvironmental conditions such as wong>inong>d speed ong>andong> temperature among others, ong>andong>all factors that modify or attenuate the effective vapour pressure of the pesticide(Spencer, W.F. et al. 1973).In spite of all the work that has been done ong>inong> this field, up till now, measurong>inong>g orassessong>inong>g volatilisation is not a simple task. Many models simulatong>inong>g or predictong>inong>gvolatilisation have been published (e.g. Chen C et al. 1995; Long>inong>dhardt et al. 1994;Long>inong>dhardt ong>andong> Christensen, 1994; Jansma ong>andong> Long>inong>ders, 1995, Smit et al., 1997,1998). For a survey of part of the literature cf. Noltong>inong>g et al., 1988. However, allthe models can only be applied under certaong>inong>, usually very restrictive, conditions.Therefore, the search for a better understong>andong>ong>inong>g of the volatilisation process ong>andong> formethods accurately assessong>inong>g pesticide volatility still goes on.In the present study the volatilisation potential of pesticides is discussed ong>andong>illustrated based on both physical/chemical properties of the pesticides ong>andong> onexperimental volatilisation experiments with the three pesticides: MCPP-P, MCPPmethyl, ong>andong> long>inong>dane. A laboratory model system for the direct determong>inong>ation of thevolatilisation of pesticides from different surfaces has been developed. The purposeof this laboratory set-up is to evaluate volatilisation of pesticides relative to eachother.4.1 Estimation of volatility of pesticides based on physical/chemical parameters ofthe test substanceOne of the most important physical/chemical properties ong>inong>fluencong>inong>g a substance’svolatility is its vapour pressure. Every chemical has a characteristic saturationvapour pressure which varies with temperature. Volatile substances, such as water,ethanol, ong>andong> methanol, have comparatively high vapour pressures at roomtemperature (10 3 – 10 4 Pa). In contrast, pesticides have generally a comparativelylow vapour pressure (approx. 10 -7 – 10 -2 Pa), however, many still volatilise atconsiderable rates (Krasel ong>andong> Pestemer, 1993; Boehnke et al., 1990). The vapourpressure of long>inong>dane is measured to 5.6 x 10 -3 Pa ong>andong> 4.0 x 10 -4 Pa for MCPP-P (at20 o C). No measured value was found for MCPP methyl (Meylan et al. 1994).Thus, judgong>inong>g from the measured vapour pressure long>inong>dane is expected to have ahigher volatility than MCPP-P, which is ong>inong> agreement with our experimental data.To make a relative comparison of all three pesticides the calculated vapourpressure was compared usong>inong>g a model of all 3 pesticides. The calculated valueswere 1.04 x 10 –1 Pa for long>inong>dane, 1.1 x 10 -2 Pa for MCPP-P, ong>andong> 2.5 x 10 -1 Pa for29


MCPP methyl (all at 25 o C). Even though a direct comparison between themeasured ong>andong> calculated values can not be made due to the use of differenttemperatures the relative difference between the pesticides is not changedsignificantly. Thus, based on these calculated vapour pressures it is expected thatMCPP methyl has a higher volatility than long>inong>dane which has a higher volatilisationthan MCPP-P. This is ong>inong> agreement with our experimental measurements.Another property that is often used ong>inong> evaluation of the volatility of pesticides istheir ong>airong>-water partition coefficient, the Henry´s law constant. The Henry´s lawconstant is usually calculated by dividong>inong>g vapour pressure (dimension Pa) by watersolubility (dimension mol m -3 ). Consequently, uncertaong>inong>ties ong>inong> vapour pressurevalues are reflected by Henry´s law constants as well. A high Henry´s law constant(> 1 Pa m 3 mol -1 ) ong>inong>dicates a high volatility from aqueous solution (e.g., from moistsoil). However, based on how the Henry´s law constant is calculated a substancewith very low water solubility may have a rather high Henry´s law constant even ifits vapour pressure is comparatively low. For the three pesticides used ong>inong> thepresent study the calculated Henry´s law constant for long>inong>dane, mecoprop methylong>andong> mecoprop P is 4.1, 1.03, ong>andong> 2.7 ×10 – 3 (Pa m 3 /mol), respectively. Thesevalues are based on calculated vapour pressure ong>andong> a temperature of 25 o C. Basedon these values a much higher volatility for long>inong>dane ong>andong> MCPP methyl would beexpected compared with MCPP-P, which is ong>inong> agreement with the presentexperimental measurements. However, based on Henry´s law constant a highervolatility of long>inong>dane would also be expected compared with MCPP methyl that isnot ong>inong> agreement with our experimental results. It may be due to the fact that thewater solubility of long>inong>dane is decidedly lower than the solubility of MCPP methyl,contributong>inong>g to the high Henry´s law constant. A comparable phenomenon has alsobeen described by Walter et al. (1996) where the volatilisation of 3 differentpesticides were compared (the active ong>inong>gredients were not named because the studywas a part of an ong>inong>terlab comparison). In spite of a higher Henry´s law constant ofactive ong>inong>gredient this compound volatilised considerably less than the other twocompounds which both have a higher vapour pressure, suggestong>inong>g a highervolatilisation. Therefore, while these physico/chemical properties may be used toestimate relative volatility, deducong>inong>g actual volatilisation behaviour from them maybe erroneous. However, volatilisation of pesticides seems to be more related to thevapour pressure than to the Henry´s law constant.4.2 Experimental measurement of volatilisation of pesticides with special referenceto long>inong>dane, MCPP-P ong>andong> MCPP methylIt is evident that there are other factors besides vapour pressure ong>andong> Henry´s Lawconstant that ong>inong>fluence volatilisation, such as climatic parameters e.g. wong>inong>d speed,temperature ong>andong> humidity. Volatilisation of pesticides from natural surfaces is alsoong>inong>fluenced by their formulations, the application techniques ong>andong> ong>inong>teractions of thesubstance to which it is applied (e.g. adsorption, desorption).Lookong>inong>g at the practical field situation, many different factors ong>andong> their ong>inong>teractionsdetermong>inong>e volatilisation from plant ong>andong> soil surfaces. The results obtaong>inong>ed fromfield experiments can fully be transferred to practical outdoor conditions. Thedisadvantages are, however, the high variability of results for the same compounddue to many highly variable factors ong>inong> field experiments. To overcome thesedisadvantages laboratory experiments can be helpful tools. A number of methodshave been developed, rangong>inong>g from very simple to high-tech designs. However,most of the experimental volatilisation measurements of pesticides ong>inong> Europe arebased on the German BBA test guidelong>inong>e.30


In the present study a simple laboratory model system has been designed, ong>inong> whicha wide range of outdoor conditions can be simulated. The fong>inong>alised method wasthen used to measure the volatility of selected pesticides from different surfacesong>andong> under different “climatic” conditions such as temperature, wong>inong>d speed etc. Theaim of the experimental set-up was to develop a simple, cost-effective ong>andong> sensitivelaboratory test system to assess ong>andong> evaluate the volatilisation of pesticides relativeto each other under different conditions.4.2.1 The German BBA test guidelong>inong>eIn 1990 a BBA Guidelong>inong>e concernong>inong>g volatilisation of pesticides was published(Noltong>inong>g H-G et al.). This guidelong>inong>e was design as a “liberal” guidelong>inong>e with only afew specific demong>andong>s to be met by the method applied. The ong>inong>tention was that aftersome years of experience, the methods, ong>inong> use at that period, should be comparedong>andong> evaluated, ong>andong> it should then be decided whether a more specific guidelong>inong>ewould have to be issued (Walter et al., 1996).The German guidelong>inong>e requests details on the percentage of active ong>inong>gredientvolatilised 1, 3, 6, ong>andong> 24 hours after application, the 24 hours value beong>inong>g thecrucial value for the authorisation process. A few other details are requested by theBBA guidelong>inong>e such as: the relative humidity of the ong>airong> shall be about 35% ong>andong> thewong>inong>d velocity shall be > 1 m/s directly above the surface. In soil experiments astong>andong>ard 2.1 soil or a similar soil with maximum 1.5% organic bound carbon shallbe used, ong>andong> the song>andong> quota shall be at least 70%. The water level ong>inong> the soil shallbe 60% of the maximum water holdong>inong>g capacity ong>andong> must be kept at this leveldurong>inong>g the experiments.4.2.2 Experimental measurement ong>andong> evaluation of selected factors ong>inong>fluencong>inong>g onthe volatilisation, ong>andong> of pesticides with special focus on MCPP-P, MCPPmethyl, ong>andong> long>inong>daneModel system fordetermong>inong>ation ofvolatilisation ofchemical substancesA simple laboratory model system for the determong>inong>ation of the volatilisationhas been used for the present study. It permits a direct determong>inong>ation of thevolatilisation of a compound from a given surface ong>andong> a mass balance can bemade at the end of the study. Furthermore, different ong>airong>flows, temperatures, etc. canbe obtaong>inong>ed ong>inong> this experimental set-up.The recovery ong>andong> stability experiments showed that the pesticides were absorbedquantitatively by the adsorption tubes used ong>inong> this test design ong>andong> the total recoverywas sufficient. Furthermore, the test results revealed that the substances were stabledurong>inong>g the test period of up till 30 hours.Design ong>andong> dimensionof the chamberThe BBA guidelong>inong>e neither requests nor suggests a specific method forassessong>inong>g volatilisation. Various laboratory model systems for direct ong>andong> ong>inong>directdetermong>inong>ation of volatilisation of pesticides from treated surfaces are described ong>inong>the literature. Designs ong>andong> dimensions of volatilisation chambers differ widely,which may be one reason for a considerable variation among the results obtaong>inong>ed ong>inong>different studies. With respect to the volatilisation chamber properties, thedifferences ong>inong> size, experimental area, ong>andong> ong>airong> exchange rate are the most apparent.For practical reasons, the volatilisation is often studied ong>inong> small laboratoryvolatilisation chambers like that used ong>inong> the present study (chamber size of about0.00008 m 3 ) (Stork A. et al. 1994, Walter U. et al. 1996). However, some studiesare also carried out usong>inong>g a large scale wong>inong>d tunnel system with a chamber size ofseveral m 3 (Rüdel H. 1997) or under field-like conditions ong>inong> a wong>inong>d-tunnel/lysimeter system with a chamber size of 0.3-0.9 m 3 (Stork A. et al., 1994). The31


volatilisation rate may also be ong>inong>fluenced by the size of the treated area ong>inong> thevolatilisation chamber. Thus, when testong>inong>g the volatilisation of long>inong>dane from soil ong>inong>a wong>inong>d tunnel under defong>inong>ed conditions, it was observed that from a larger soilsurface (0.84m 2 ) a lower amount of long>inong>dane (23%) volatilised than from a smallersurface (31% at 0.28m 2 ) under otherwise identical conditions (Waymann B. ong>andong>Rüdel H., 1995).The volatilisation rateApplication rateVolatilisation atdifferent ong>airong>flowong>andong> humidityVolatilisation of long>inong>dane ong>andong> MCPP methyl applied at the recommended dose tofilter paper durong>inong>g a 30 hour period ong>inong>dicates that about 28% ong>andong> 8% of the applieddose was volatised durong>inong>g the first hour after application for MCPP methyl ong>andong>long>inong>dane, respectively. After 8 hours about 80% ong>andong> 50% of the applied MCPPmethyl ong>andong> long>inong>dane, were volatised. Only about 7% of the applied MCCP methylvolatised ong>inong> the period between 8 ong>andong> 30 hours after application while it was about30% ong>inong> the case of long>inong>dane. The most important objective of a volatilisationexperiment accordong>inong>g to the BBA Guidelong>inong>e is the determong>inong>ation of the volatilisedquantity of the substance withong>inong> 24 hours after application. This time ong>inong>tervalseems to be reasonable when examong>inong>g the experiments reported here. Theydocument the high volatilisation rates withong>inong> the first few hours after application,ong>andong> the 24 hours value is measured at a time when the volatilisation process hasconsiderably slowed down. The same pattern has also been seen ong>inong> other laboratoryvolatilisation experiments with other pesticides (e.g. Walter et al., 1996). Theresults from the present experiments also show that when MCCP-P was applied tofilter paper no volatilisation took place durong>inong>g a 24 hours period. These results areong>inong> agreement with previous ong>inong>vestigations of the volatilisation of MCPP-P fromplant surfaces ong>andong> soil under laboratory conditions (Mossong>inong> J. personalcommunication, Danish EPA, Division of ong>Pesticidesong>, 1999).The rate of loss by volatilisation depends on the concentration on ong>andong> ong>inong> a givenmedium e.g. a soil ong>andong> the concentration-vapour density relationships at the soilsurface. Accordong>inong>g to Letey ong>andong> Farmer (1974) there are two general mechanismswhereby pesticides move to the evaporatong>inong>g surface, i.e. diffusion ong>andong> mass flow.Diffusion is the process by which material is transported as a result of rong>andong>ommolecular motion caused by the molecule’s thermal energy. The rong>andong>om molecularmotions gradually cause the molecules to become uniformly distributed ong>inong> thesystem (Letey ong>andong> Farmer, 1974). Diffusion occurs whenever a concentrationgradient is present. In general the diffusion rate of a pesticide is ong>inong>creased withong>inong>creasong>inong>g concentration applied. In the present study the two pesticides wereapplied at a rate correspondong>inong>g to the recommended application rate ong>andong> 10× therecommended application rate. When the doses were ong>inong>creased with a factor of 10the amount of long>inong>dane volatisong>inong>g also ong>inong>creased, however, only with a factor of 2.4under the present experimental conditions. In contrast, when the amount of MCPPmethyl applied was ong>inong>creased with a factor of 10, the amount of MCPP methylvolatisong>inong>g ong>inong>creased with a factor of 8.9 illustratong>inong>g the higher volatilisationpotential for this compound. A comparable phenomenon has also been described ong>inong>soil experiments with different application doses of long>inong>dane ong>inong> a wong>inong>d tunnelexperiment (Waymann ong>andong> Rüdel, 1995). Thus, when the application dose isong>inong>creased the volatilisation is normally also ong>inong>creased, however, the actual ong>inong>creaseong>inong> volatilisation is highly depended on the physical/chemical properties of thepesticides ong>andong> of other environmental conditions.Wong>inong>d speed, turbulence, ong>andong> relative humidity play an important role ong>inong> theoverall loss of pesticides by volatilisation ong>inong> the field. The direct effect ofong>inong>creased ong>airong> movement ong>inong>volves a more rapid removal of pesticide vapours fromthe soil surface, ong>andong> results ong>inong> an ong>inong>creased movement of pesticide to the soilsurface (Guenzi ong>andong> Beard, 1974). Accordong>inong>g to Hartley (1969) the rate ofmovement away from the evaporatong>inong>g surface is a diffusion-controlled process.32


Close to the evaporatong>inong>g surface the ong>airong> is relatively still. The vaporisong>inong>g substanceis transported from the surface through this stagnant ong>airong> layer only by moleculardiffusion. Diffusion away from the surface is related to the vapour density ong>andong> themolecular weight of the pesticide. The thickness of this stagnant ong>airong> layer above theevaporatong>inong>g surface depends on the ong>airong>flow rate. The studies of Farmer et al. (1972)ong>andong> Igue et al. (1972) found more volatilisation of chlorong>inong>ated ong>inong>secticides withong>inong>creasong>inong>g flow rates. Waymann ong>andong> Rüdel (1995) also found an ong>inong>crease ong>inong> thevolatilisation rate of long>inong>dane from both soil ong>andong> plants with ong>inong>creasong>inong>g ong>airong> velocitiesong>inong> a wong>inong>d tunnel with lamong>inong>ar ong>airong>flow.In the present study the volatilisation of long>inong>dane ong>andong> MCPP methyl was tested attwo ong>airong>flows (0.17 ong>andong> 0.67 m per sec.) but no significant ong>inong>crease ong>inong> the amount ofvolatilisation of the two pesticides was found with ong>inong>creasong>inong>g ong>airong>flow. The ong>airong>flowwas measured at a distance less than 0.7 cm. from the surface which wouldcorrespond to a calculated wong>inong>d speed of more than 1 m per sec at a height of 1 mabove the surface (Asman, 1999). Furthermore, the ong>airong>flow reached thevolatilisation chamber as a turbulent flow. This ong>inong>creased the volatilisation ofpesticides compared with a lamong>inong>ar ong>airong>flow situation song>inong>ce the primary effect ofwong>inong>d on pesticide disappearance from foliage is assumed to be through turbulenttransfer of volatilised pesticide from plant surfaces to the atmosphere (Spencer etal., 1973). Song>inong>ce no significant ong>inong>crease ong>inong> the amount of volatilisation of the twopesticides was observed after an ong>inong>crease of the ong>airong>flow by a factor of 4, it isexpected that an ong>airong>flow of 0.17 m per sec. is sufficient to ensure that the rate of theong>airong>flow is not the limitong>inong>g factor for a high evaporation under the presentexperimental conditions.Influence of the relative If the relative humidity of the ong>airong> is not 100%, ong>inong>creases ong>inong> ong>airong>flow willhumidity of the ong>airong> hasten the dryong>inong>g of the surface e.g. a soil. This ong>inong>direct effect alters the soil-watercontent, which has an effect on the volatilisation (Guenzi ong>andong> Beard, 1974).Normally pesticides volatilise much more rapidly from wet than from dry soilsbecause the polar water molecules are strong competitors for adsorption sites onthe soil, especially to non-polar organic compounds (Dörfler et al. 1991, Petersenet al., 1994). Passong>inong>g moist ong>airong> over moist soil will result ong>inong> little water loss.Therefore, pesticide volatilisation will contong>inong>ue for a long time. However, with dryong>airong> (low relative humidity), the soil dries rapidly ong>andong> pesticide vapour pressure isdecreased withong>inong> a relatively short time. The dryong>inong>g effect, decreasong>inong>g thevolatilisation, is reversible, song>inong>ce remoistong>inong>g the ong>airong>-dry soil will ong>inong>crease thevapour density agaong>inong> to its origong>inong>al maximum value (Jansma ong>andong> Long>inong>ders, 1995).Grass et al. (1994) measured the ong>inong>fluence of ong>airong> humidity on the volatilisation oftriflualong>inong>. At a relative ong>airong> humidity of 31, 49, ong>andong> 78% the measured percentage ofvolatilised trifluralong>inong> over the first day was 66, 64, ong>andong> 96%, respectively(temperature 20 o C, ong>airong> velocity 1.0-1.2 m/s.). In the present study a relativehumidity of the ong>airong> was maong>inong>taong>inong>ed at 50% ong>inong> all experiments ong>andong> the test-substrate(filter paper or soil) was ong>inong> equilibrium with this humidity before the pesticideswere added just to ensure relative constant water content of the substrate.Effects of differenttemperaturesThe temperature ong>inong>fluences the characteristic saturation vapour pressure of agiven pesticide. The overall effect of an ong>inong>creasong>inong>g temperature is an ong>inong>creasong>inong>gvolatilisation of the pesticide. This is also illustrated ong>inong> the present study where anong>inong>crease ong>inong> the temperature from 15 o C to 23 o C ong>inong>creases the volatilisation ofespecially long>inong>dane.In the field the ong>effectsong> of different temperatures on volatilisation of pesticides aremore complex. Temperature affects the volatilisation of a given pesticide from soil33


y a direct ong>inong>fluence on the vapour density of the pesticide ong>andong> by temperatureong>inong>fluences on the physical ong>andong> chemical properties of the soil. Thus, temperaturemay ong>inong>fluence the volatilisation of soil-ong>inong>corporated pesticides through its effect onmovement of the pesticide to the surface by diffusion or by mass flow ong>inong> theevaporatong>inong>g water, or through its effect on the soil water adsorption-desorptionequilibrium. Ehlers et al. (1969) reported an exponential ong>inong>crease ong>inong> the apparentdiffusion coefficient ong>inong> the soil for long>inong>dane after ong>inong>creasong>inong>g the temperature from20 o C to 40 o C. However, an ong>inong>crease of temperature may also ong>inong>crease the dryong>inong>grate of the soil surface. Dependong>inong>g on the results of the temperature effect on soildryong>inong>g ong>andong> the effect on vapour pressure, volatilisation will ong>inong>crease or decrease(Demong>inong>g, 1963).Volatilisation whenapplied on differentsurfacesAdsorption reduces the chemical activity ong>andong>, thus, also the volatilisationrate of the compound. The rate of loss by volatilisation depends on theconcentration-vapour density relationships at the surface ong>andong> is highly depended onthe type of surface that the pesticide is applied to. It should be noted that thevariation ong>inong> the plants ong>andong> soils used ong>inong> different ong>inong>vestigations is significant. Thismay be one reason for a considerable variation among the results published formany substances. To overcome these disadvantages, laboratory experiments usong>inong>ga stong>andong>ardised artificial surface can help to assess ong>andong> evaluate the volatilisation ofpesticides relative to each other under different conditions.In the present volatilisation study long>inong>dane ong>andong> MCPP methyl were applied to twodifferent artificial surfaces, a filter paper illustratong>inong>g an absorbent surface ong>andong> ametal plate illustratong>inong>g a non-absorbent surface. The results revealed that thevolatilisation of each pesticide after 24 hours was comparable (about 73% forlong>inong>dane ong>andong> 90% for MCCP methyl of the applied doses). However, it seems thatthe volatilisation of at least MCPP methyl was faster at the non-absorbent surface.In contrast to the artificial surfaces only a small fraction of the applied dosesvolatised when the pesticides were applied to a typical Danish soil (13% forlong>inong>dane ong>andong> 15.9% for MCPP methyl). The results are comparable with the resultsfound for long>inong>dane tested ong>inong> a huge wong>inong>d tunnel where the soil volatilisation wasfound to vary between 12 ong>andong> 31% ong>andong> volatilisation from plant varied from 52 to62% of the ong>inong>itial dose after 24 hours, dependong>inong>g on the test conditions (Waymannong>andong> Rüdel, 1995). Furthermore, our results are also comparable to evaporativelosses of long>inong>dane from plant leaf ong>andong> soil surfaces after field application. Boehnckeet al. (1990) observed an evaporative loss of long>inong>dane from leafs of 77 to 95%(different plant leafs) ong>andong> 28% from a soil after 24 hours. Jansma ong>andong> Long>inong>ders,(1995) presented a review of the literature about volatilisation of selectedpesticides from soil ong>andong> plants after sprayong>inong>g. In this paper the volatilisation oflong>inong>dane from plants varied from 64 to 89% ong>inong> 5 different laboratory studies after a24 hours period. From the soil surface the volatilisation of long>inong>dane varied after 24hours from 3.1 to 22.6% of the applied dose ong>inong> 8 different laboratory experiments.However, ong>inong> one field experiment ong>inong> USA up till 50% of the applied dose wasvolatilised withong>inong> 24 hours.In general the volatilisation of pesticides from soil is much smaller than from plantsurfaces (e.g. Boehncke et al. 1990; Rüdel, 1997). However, significant variation isfound dependong>inong>g on the soil used for the experiments. This is ong>inong> agreement with theresults found ong>inong> the present study where the filter paper is more comparable with aplant surface. The volatilisation of pesticides from soil is very complex e.g. thediffusion rate of a pesticide from soil is controlled by soil bulk density, pesticideconcentration, organic matter, pH, the soil water ong>andong> clay contents ong>inong> addition toclimatic parameters like temperature, wong>inong>d velocity etc. Furthermore, verticaltransport is an important parameter ong>inong> the evaluation of the volatilisation ofpesticides from soil ong>andong> occurs as a result of external forces. The pesticide is34


considered to be either dissolved or suspended ong>inong> water, present ong>inong> the vapourphase, or adsorbed on solid mong>inong>eral or organic components of the soil. Mass flowof the pesticide is therefore the result of the mass flow of water ong>andong>/or soil particleswith which the pesticide molecule is associated. Although many factors contributeto mass transport of pesticides through soil by water, the most important factorappears to be adsorption between the pesticide ong>andong> soil (Letey ong>andong> Farmer 1974).Mass flow due to ong>airong> movement ong>inong> soil is considered negligible (Letey ong>andong> Farmer,1974).4.3 Conclusions ong>andong> suggestions for further validation workLookong>inong>g at the practical field situation, many different factors ong>andong> their ong>inong>teractionsdetermong>inong>e volatilisation from plant ong>andong> soil surfaces. The results obtaong>inong>ed fromfield experiments can fully be transferred to practical outdoor conditions. Thedisadvantages are, however, the lack of reproducibility. To overcome thesedisadvantages laboratory experiments can be helpful tools.In 1990 a German BBA guidelong>inong>e on assessong>inong>g pesticide volatilisation wasdeveloped. This guidelong>inong>e was design as a “liberal” guidelong>inong>e with only a fewspecific demong>andong>s. Song>inong>ce then a number of methods have been developed, rangong>inong>gfrom very simple to high-tech designs.In 1994 the volatilisation of three pesticides from plant ong>andong> soil surfaces wasassessed usong>inong>g eighteen different laboratory methods ong>andong> one field method, allfollowong>inong>g the BBA guidelong>inong>e. With respect to the volatilisation chamber properties,the differences ong>inong> size, experimental area, ong>andong> ong>airong> exchange rate are the mostapparent. Furthermore, the height at which wong>inong>d is measured varies from 0-3 cm tomore than 10 cm. The aim of the present study was to see whether the differentmethods yielded results that were consistent among the methods ong>andong> a number ofterms were agreed (dose, temperature ong>andong> humidity etc.). The results revealed thatfor all three substances tested, a considerable amount of variation among the resultsobtaong>inong>ed with the different methods was observed (Walter et al. 1996).From the scientific poong>inong>t of view, the ong>inong>ter-laboratory comparison once agaong>inong>proved that the problems of assessong>inong>g pesticide volatilisation have not been solvedyet, ong>andong> that it is necessary to be very cautious when comparong>inong>g volatilisation ratesassessed with different methods.In the present study a laboratory system for the direct determong>inong>ation of thevolatilisation of pesticides from different surfaces has been developed. With thesystem it is possible to evaluate volatilisation of pesticides relative to each other.Takong>inong>g practical aspects of testong>inong>g ong>andong> the demong>andong> for simple cost-effective testsong>inong>to account, the maong>inong> conclusions of the present study regardong>inong>g the experimentalset-up are as follows:The BBA guidelong>inong>e on assessong>inong>g pesticide volatilisation was designed as a “liberal”guidelong>inong>e with only a few specific demong>andong>s. This guidelong>inong>e is a good basis,however, it should be improved ong>andong> be more specific.Design ong>andong> dimensionof the chamberDesign ong>andong> dimension of the volatilisation chamber must be specified. Theconditions of most experiments are so different that comparative considerations aredifficult or impossible. An exact determong>inong>ation of volatilisation of pesticides can beachieved only by direct measurement of volatilisation rates usong>inong>g some means oftrappong>inong>g the vapourised pesticide. A small chamber is recommended. Such achamber can be extracted at the end of an experiment ong>andong> a mass balance can bemade.35


Several concentrations of a given pesticide should be tested ong>andong> the applica-tion technique must be more explicit. Application devices used ong>inong> differentexperiment differ greatly. Everythong>inong>g from nozzles used ong>inong> agricultural practice tomodified TLC applicators to simple Hamilton syrong>inong>ges has been used. Furthermore,experiments must be carried out with the formulated product as well as with theactive ong>inong>gredient.Samplong>inong>g ong>inong>tervalsAirflowHumidity of the ong>airong>Selected surfacesApplication rate ong>andong>techniqueTemperaturesInterpretong>inong>g the kong>inong>etics graphs from various volatilisation experiments withpesticides it can generally be said that the ong>airong> samplong>inong>g 1, 3, 6, ong>andong> 24 hours afterthe application asked for ong>inong> the German guidelong>inong>e seems to be reasonable.Normally, high volatilisation rates are seen withong>inong> the first few hours afterapplication, ong>andong> the 24 hours values after application are measured at a time whenthe volatilisation process has slowed down considerably.Accordong>inong>g to the BBA guidelong>inong>e the wong>inong>d velocity shall be > 1 m/s directly abovethe surface. However, the guidelong>inong>e does not exactly stress ong>inong> which height themeasurement should be made. When evaluatong>inong>g the volatilisation of a testsubstance from a surface the height above the surface, where the ong>airong> velocitymeasured is very important. Thus, if the assumed wong>inong>d speed at 30 cm heightabove a bare soil surface is 1 m/s the calculated wong>inong>d speed is about 0.7 ong>andong> 0.3m/s at a height of 5 ong>andong> 0.5 cm above the surface, respectively (Asman, personalcommunication, 1998, National Environmental Research Institute, Roskilde,Denmark). Furthermore, it is important to decide whether turbulent or lamong>inong>arong>airong>flow should be used. If a small chamber with a turbulent ong>airong>flow is used anong>airong>flow with a velocity of less than 1 m/s 1-5 mm above the surface might besufficient.The relative humidity of the ong>airong> shall be about 35% accordong>inong>g to the BBAguidelong>inong>e. Song>inong>ce the ong>airong> humidity value is relatively low ong>andong> not realistic for fieldsituations, this condition should be changed. Mostly, a more realistic value ofapproximately 50% relative humidity should be used. However, the most importantaspect is that the surface to which the pesticide is applied has the same humiditydurong>inong>g the whole experiment.Accordong>inong>g to the BBA guidelong>inong>e pesticides should be applied to two differentsurfaces one illustratong>inong>g a leaf surface ong>andong> one soil. It is important to specify theselected surface ong>andong> an artificial surface e.g. a filter paper may be used to illustratea leaf surface, however, some work has to be done to fong>inong>d a well-defong>inong>ed surfacethat can simulate a leaf surface. In soil experiments a German stong>andong>ard 2.1 soil or asimilar soil is recommended ong>inong> the BBA guidelong>inong>e. The water level ong>inong> the soil mustbe 60% of the maximum water holdong>inong>g capacity ong>andong> must be kept at this leveldurong>inong>g the experiments. This soil can be used as a reference soil. However, anothersoil, which is typical for the area of concern, should be ong>inong>cluded ong>inong> the test.Furthermore, the pH value of the soil should be specified song>inong>ce dependency ofvolatilisation seems to be strong for some pesticides (Walter et al., 1996).However, data from the present study ong>andong> many other studies suggest that plant (oran artificial surfaces illustratong>inong>g plant surfaces) volatility is always higher than soilvolatility. Therefore, it should be considered to omit the volatilisation tests withbare soil ong>andong> only perform plant volatilisation tests. In most cases these results willbe sufficient to assess the volatilisation behaviour of pesticides.The temperature must be kept constant e.g. at 20 o C or even better measured at twodifferent temperatures e.g. 10 o C ong>andong> 30 o C representong>inong>g different temperatures ong>inong>Europe.36


5 ReferencesBaas, J (1994). Emissie van gewasbeschermong>inong>gsmiddelen uit boomgaardennaar de lucht. TNO-report MW-R 94/040a, TNO –Milieu en Energie, TNO,Delft.Boehnke A., Siebers J., ong>andong> Noltong>inong>g H.-G. (1990). ”Investigation of theEvaporation of Selected ong>Pesticidesong> from Natural ong>andong> Model Surfaces ong>inong> Fieldong>andong> Laboratory”, Chemosphere, 21, pp. 1109-1124.Buser H.R. (1990). ”Atrazong>inong>e ong>andong> Other s-Triazong>inong>e Herbicides ong>inong> Lakes ong>andong>ong>inong> Raong>inong> ong>inong> Switzerlong>andong>”, Environmental Science & Technology, 24, pp.1049-1058.Chen C., Green R. E., Thomas, D.E., ong>andong> Knuteson J.A. (1995). “Modelong>inong>g1,3-Dichloropropene Fumigant Volatilization with Vapour-Phase Advectionong>inong> the Soil Profile”, Environmental Science & Technology, 29, pp. 1816-1821.Demong>inong>g D.G. (1963). Determong>inong>ation of volatility losses of C 14 -CDAA fromsoil samples. Weeds 11, 91-96.Dörfler U., Adler-Köhler R., Schneider P., Scheunert I., ong>andong> Korte F. (1991).A Laboratory Model System for Determong>inong>ong>inong>g the Volatility of ong>Pesticidesong>from Soil ong>andong> Plant Surfaces. Chemosphere, 23, pp 485-496.Ehlers W., Farmer W.J., Spencer W.F. ong>andong> Letey J. (1969). Long>inong>danediffusion ong>inong> soils: II. Water content, bulk density, ong>andong> temperature effetcs.In: Soil Sci. Soc. Amer. Proc. 33, 505-508.Farmer W.J, Igue K, Spencer W.F. ong>andong> Martong>inong> J.P. (1972). Volatility oforganochlorong>inong>e ong>inong>secticides from soil. I. Effect of concentration,temperature, ong>airong> flow rate, ong>andong> vapour pressure. In: Soil Sci. Soc. Amer.Proc. 36, 443-447.Feldong>inong>g G., Spliid N.H., ong>andong> Asman W.A.H. (1999). Fong>inong>dong>inong>gs ong>andong> ong>effectsong> ofpesticides ong>inong> ong>airong> ong>andong> ong>precipitationong>. ong>inong> press.Gath B., Jaeschke W., Ricker I., ong>andong> Zietz E. (1992).”Depositionsmonitorong>inong>g von Pflanzenschutzmitteln auf dem Kleong>inong>enFeldberg. Erste Ergebnisse”, Nachrichtenblatt des DeutschenPflanzenschutzdienstes, 44, pp. 57-66.Gath B., Jaeschke W.,Kubiak R., Ricker I., Schmider F., ong>andong> Zietz E. (1993).”Depositionsmonitorong>inong>g von Pflanzenschutzmitteln: Teil 2 SüddeutscherRaum”, Nachrichtenblatt des Deutschen Pflanzenschutzdienstes, 44, pp. 57-66.Glotfelty D.E., Majewski M.S., ong>andong> Seiber J.N. (1990). ”Distribution of37


Several Organophosphorus Insecticides ong>andong> Their Oxygen Analogues ong>inong> aFoggy Atmosphere”, Environmental Science & Technology, 24, pp. 353-357.Gottschild D., Siebers J., ong>andong> Noltong>inong>g H.-G. (1995). The volatilization ong>andong>deposition of plant protection products. Strasbourg Cedex: Publishong>inong>g ong>andong>Documentation Service council of Europe.Grass B., Wenclawiak B.W. ong>andong> Rüdel H. (1994). Influence of ong>airong> velocity,ong>airong> temperature, ong>andong> ong>airong> humidity on the volatilisation of trifluralong>inong> from soil.In: Chemosphere, 28, no 3, 491-499Guenzi W.D. ong>andong> Beard W.E. (1974). Volatilisation of pesticides. In:Guenzi, W.D. (Ed) ong>Pesticidesong> ong>inong> soil ong>andong> water. Chapter 6, 107-122. SoilScience Society of america, Madison, Wisconsong>inong>.Hartley G.S. (1969). “Evaporation of ong>Pesticidesong>”, Pesticidal FormulationsResearch – Physical ong>andong> Colloidal Chemical Abstracts, Miami Beach,Florida, USA.Igue K, Farmer W.J., Spencer W.F. ong>andong> Martong>inong> J.P. (1972). Volatility oforganochlorong>inong>e ong>inong>secticides from soils: II Effect of relative humidity ong>andong> soilwater content on dieldrong>inong> volatility. In: Soil Sci. Soc. Amer. Proc., 36, 447-450.Jansma J.W. ong>andong> Long>inong>ders J.B.H.J. (1995). Volatilization of pesticides fromsoil ong>andong> plants after sprayong>inong>g. RIVM Report no. 679102030Krasel G. ong>andong> Pestemer W. (1993). ”Pflanzenschutzmittel-Verflüchtigungvon Oberflächen”, Proceedong>inong>gs of the 8 th EWRS Symposium ”Quantitativeapproaches ong>inong> weed ong>andong> herbicide research ong>andong> their practical application”,Braunschweig,pp. 399-406.Letey J. ong>andong> farmer W.J. (1974). Movement of pesticides ong>inong> soil. In: Guenzi,W.D. (Ed.) ong>Pesticidesong> ong>inong> soil ong>andong> water. Chapter 4, 67-97. Soil ScienceSociety of America, Madison, Wisconsong>inong>.Long>inong>dhardt B. ong>andong> Christensen T.H. (1994). ”Measured ong>andong> EstimatedVolatilisation of Naphthalene from a Song>andong>y Soil”, Chemosphere, 29, pp.1407-1419.Long>inong>dhardt B., Christensen T.H., ong>andong> Brun A. (1994). ”Volatilisation of o-Xylene from Song>andong>y Soil”, Chemosphere, 29, pp. 2625-2637.Meylan W ong>andong> Howard P. (1994). MPBPVP: Meltong>inong>g poong>inong>t, Boilong>inong>g poong>inong>tong>andong> vapour pressure program for Microsoft Wong>inong>dows 3.1. Syracuse ResearchCorporation, Syracuse 1994.Nations B.K. ong>andong> Hallberg G.R. (1992). ”ong>Pesticidesong> ong>inong> Iowa Precipitation”,Journal of Environmental Quality, vol. 21, pp. 486-492.Noltong>inong>g H.-G., Boehnke A., ong>andong> Storzer W. (1988). ”Zum Eong>inong>trag, Verhalten38


und Verbleib von Pflanzenschutzmitteln ong>inong> Luft”, Biologische Bundesanstaltfür Long>andong>- und Forstwirtschaft, Abteilung für Pflanzenschutzmittel undAnwendungstechnik, Braunschweig, Literaturstudie.Noltong>inong>g H.-G., Siebers J., Storzer W., Wilkenong>inong>g A., Herrmann M., ong>andong>Schlüter C. (1990). ”Teil IV 6-1 Prüfung des Verflüchtigungsverhaltens unddes Verbleibs von Pflanzenschutzmitteln ong>inong> der Luft”, ong>inong> Richtlong>inong>ien für diePrüfung von Pflanzenschutzmitteln im Zulassungsverfahren, Abteilung fürPflanzenschutzmittel und Anwendungstechnik der BiologischenBundesanstalt Braunschweig, Ed. Ribbesbüttel: Saphir-Verlag.Parker. S.P. (1984). ”McGraw-Hill Dictionary of Physics”. New York:McGraw-Hill Book Company.Petersen, L.W., Rolston D.E., Moldrup P., Yamaguchi T. (1994). Volatileorganic vapour diffusion ong>andong> adsorption ong>inong> soils. In: J. Environ. Qual., 23,799-805.Plimmer J.R. (1976). ”Volatility”, ” ong>inong> Herbicides – Chemistry, Degradationong>andong> Mode of Action, vol. 2, P.C. Kearney ong>andong> D.D. Kaufman, Eds., 2 ed.New York, Basel: Marcel Dekker, Inc., pp. 891-933.Rüdel H. (1997). Volatilisation of pesticides from soil ong>andong> plant surfaces.Chemosphere, 35 pp 143-152.Scharf J. ong>andong> Bächmann K. (1993). ”Verteilung von Pflanzenschutzmittelnong>inong> der Atmosphäre. Nah- und Ferntransportmessungen”, Nachrichtenblattdes Deutschen Pflanzenschutzdienstes, vol. 45, pp. 82-87.Schomburg C.J., Glotfelty D.E., ong>andong> Seiber J.N. (1991). ”PesticideOccurence ong>andong> Distribution ong>inong> Fog Collected near Monterey, California”,Environmental Science & Technology, vol. 25, pp. 155-160.Siebers J., Gottschild D., ong>andong> Noltong>inong>g H.-G. (1994). ”ong>Pesticidesong> ong>inong>Precipitation ong>inong> Northern Germany”, vol. 28, pp. 1559-1570.Smit A.A.M.F.R., van den Berg F., ong>andong> Leistra M (1997). Estimationmethod for the volatilization of pesticides from fallow soil. Environmentalplannong>inong>g bureau series 2 Wagenong>inong>gen (The Netherlong>andong>s)Smit A.A.M.F.R., Leistra M., ong>andong> van den Berg F. (1998). Estimationmethod for the volatilization of pesticides from plants. Environmentalplannong>inong>g bureau series 4 Wagenong>inong>gen (The Netherlong>andong>s)Spencer W.F., Farmer W.J., ong>andong> Cliath M.M. (1973). ”Pesticidevolatilization”, Residue Reviews, vol. 49, pp. 1-47.Stork A., Witte R., ong>andong> Führ F. (1994). A Wong>inong>d Tunnel for Measurong>inong>g theGaseous Losses of Environmental Chemicals from the soil/Plant Systemunder Field-Like Conditions. Environ. Sci. & Polllut. Res. 1 (4) pp. 234-245Taylor A.W. (1978). Post-application volatilization of pesticides under fieldconditions. Journal of the Air Pollution Control Associations vol. 28 no. 9 pp922-927.Walter U, Krasel G, Pestemer W. (1996). Assessong>inong>g Volatilization of39


ong>Pesticidesong>: A Comparison of 18 laboratory methods ong>andong> a field method.Helft 16 from BBA, Braunschweig, GermanyWaymann B. ong>andong> Rüdel H. (1995). Influence of ong>airong> velocity, applicationdose, ong>andong> test area size on the volatilisation of long>inong>dane. Intern. J. Environ.Anal. Chem. vol. 58 pp 371-378.Willis G.H., McDowell L.L., Harper L.A., Southwick L.M., Smith S. (1983).Seasonal disappearance ong>andong> volatilization of toxaphene ong>andong> DDT from acotton field J. Environ. Qual. vol 12, no. 1, pp 80-85.40


Modellong>inong>g atmospheric transport ong>andong>deposition of pesticides up to 2 km from asourceWillem A.H. AsmanNational Environmental Research InstituteDepartment of Atmospheric Environment, Roskilde, Denmark41


Table of contentsEXECUTIVE SUMMARY 47DANSK SAMMENDRAG 501 MODELLING ATMOSPHERIC TRANSPORT AND DEPOSITION OFPESTICIDES UP TO 2 KM FROM A SOURCE 531.1 INTRODUCTION 531.2 SOME IMPORTANT PROPERTIES 551.2.1 Solubility ong>inong> water 551.2.2 Vapour pressure 551.2.3 Henry’s law coefficient 561.2.4 The adsorption coefficient Koc 561.3 TURBULENCE, ATMOSPHERIC TRANSPORT AND DIFFUSION 571.4 SURFACE EXCHANGE 681.5 INTERMEZZO: MAIN CONCLUSIONS ON METEOROLOGY AND SURFACEEXCHANGE 751.6 EMISSION 771.6.1 Emission of pesticides from fallow soil 771.6.2 Emission of pesticides from plants 801.6.3 Improvement ong>inong> the modellong>inong>g of the emission of pesticides 821.7 REMOVAL OF MATERIAL BY PRECIPITATION 821.7.1 Cloud physical processes 821.7.2 Exchange of gases between drops ong>andong> the ong>airong> 881.7.3 Scavengong>inong>g of gases 911.7.4 Scavengong>inong>g of particles 941.8 INTERMEZZO: MAIN CONCLUSIONS ON WET DEPOSITION AND ACOMPARISON WITH DRY DEPOSITION 951.9 CONVERSION FROM THE GASEOUS TO THE PARTICULATE PHASE 961.10 PHOTOCHEMICAL REACTION 971.11 SPRAY DRIFT AND OTHER FORMS OF DEPOSITION 981.12 MODELLING THE DEPOSITION CLOSE TO THE SOURCE 991.12.1 K-model 991.12.2 Verification of the vertical diffusion calculated with the model 1011.12.3 Situation modelled ong>inong> all further model calculations 1021.12.4 Influence of friction velocity/wong>inong>d speed 1031.12.5 Influence of atmospheric stability 1041.12.6 Vertical concentration profiles 1071.12.7 Reduction of the emission rate on the emission field due to upwong>inong>demissions 1081.12.8 Effect of surface roughness 1081.12.9 Effect of surface resistance 1101.12.10 Possibility of long-range transport 1111.12.11 Maximum sum of dry ong>andong> wet deposition 1121.12.12 Comparison of dry ong>andong> wet deposition with spray drift due tosedimentation 1131.13 INTERMEZZO: CONCLUSIONS ON THE MODEL RESULTS 1162 DISCUSSION AND CONCLUSIONS 1193 ACKNOWLEDGEMENT 1244 REFERENCES 125APPENDIX I 12943


APPENDIX II 130APPENDIX III 13244


Executive summaryIn the present study an overview is given of the atmospheric processes relevant topesticides. This knowledge is then ong>inong>corporated ong>inong> an atmospheric transport ong>andong>deposition model that is used to give ong>andong> upper estimate of the accumulatedfraction of the emission that can be deposited withong>inong> 2 km from a field wherepesticides are applied.ong>Pesticidesong> occur ong>inong> the atmosphere ong>inong> gaseous ong>andong> particulate form. Close to areaswhere they evaporate they will be maong>inong>ly ong>inong> gaseous form.The exchange of a gaseous pesticide between the atmosphere ong>andong> the surface (soil,vegetation, water) depends on the wong>inong>d speed (turbulence) ong>andong> the difference ong>inong>concentration between the atmosphere ong>andong> the surface. If the concentration ong>inong> thesurface is higher than ong>inong> the ong>airong>, net emission will occur. In the opposite case netdry deposition will occur.Both emission ong>andong> dry deposition depend on processes ong>inong> the surface that ong>inong>fluencethe concentration ong>inong> the surfaces. These processes can e.g. be vertical transport ong>inong>the surface or degradation.Both the emission ong>andong> dry deposition rate will ong>inong>crease with wong>inong>d speed. Processesong>inong> the surface will determong>inong>e to what extent they will ong>inong>crease. If processes ong>inong> thesurface are much slower than the transport ong>inong> the ong>airong> from/to the surface, theemission ong>andong> dry deposition rate will maong>inong>ly depend on the processes ong>inong> the surfaceong>andong> almost not ong>inong>crease with wong>inong>d speed.Dry depositionThe dry deposition velocity will be highest for highly soluble gaseous pesticides.For moderately ong>andong> slightly soluble gaseous pesticides the dry deposition velocitywill be lower ong>andong> will to a large extent depend on processes ong>inong> the surface.The dry deposition velocity for particulate pesticides depends not only on the wong>inong>dspeed but also on the particle size. Particles are ong>inong> general not removed by drydeposition at such high rates as highly soluble gases. Particles ong>inong> the size range 0.1- 1 µm are removed at the lowest rate. Smaller or larger particles are removed at ahigher rate.Dry deposition of gases ong>andong> particles is difficult to measure. High precision is oftenneeded because the measurements are usually based on smal concentrationdifferences. Only for highly soluble gases it is possible to get reasonable results.Even ong>inong> that case dry deposition is usually not monitored contong>inong>uously, becausethat is very expensive. In stead, dry deposition models ong>andong> measured meteorologyare used to ong>inong>fer dry deposition. Dry deposition of pesticides has almost never beenmeasured ong>inong> the field.Atmospheric diffusion ong>inong>creases also with wong>inong>d speed. This is also the case for themixong>inong>g height, i.e. the height of the lowest layer ong>inong> the atmosphere where mixong>inong>goccurs.The maximum rate at which vertically well-mixed ong>airong>borne gaseous pesticides willbe removed from the atmosphere under Danish conditions will be about 13% hr -1over croplong>andong> ong>andong> about 25% hr -1 over forests.45


Wet depositionThe removal rate of both gaseous ong>andong> particulate pesticides by ong>precipitationong>ong>inong>creases with the ong>precipitationong> rate.The removal rate of gaseous pesticides from the atmosphere due to ong>inong>-cloudprocesses is a function of the solubility of the pesticide ong>inong> water; it ong>inong>creases withsolubility. For that reason highly soluble gaseous pesticides are removed from theatmosphere by ong>inong>-cloud processes at a high rate (64% hr -1 at a raong>inong>fall rate of 1 mmhr -1 ong>andong> about 99.9% hr -1 at a raong>inong>fall rate of 10 mm hr -1 ). Less soluble gaseouspesticides are removed at a lower rate. Particulate pesticides will ong>inong> general beremoved from the atmosphere by ong>inong>-cloud processes at the same high rate as highlysoluble gases.Gaseous pesticides under the cloud are removed at a lower rate by raong>inong>drops than ong>inong>the cloud. For highly soluble gases the rate is 13% hr -1 at a raong>inong>fall rate of 1 mm hr -1 ong>andong> about 33% hr -1 at a raong>inong>fall rate of 10 mm hr -1 .The removal rate of the same particulate pesticides by raong>inong>drops under the cloud isong>inong> general much lower than the removal rate from the cloud. Moreover, it will alsodepend on the size of the particles.Reactions, conversionfrom gaseous toparticulate formBoth dry ong>andong> wet deposition of gaseous pesticides will ong>inong>crease with theirsolubility ong>inong> water. So there is a general tendency that their atmosphericlifetime is related to their solubility. Other factors, however, can also play a role.ong>Pesticidesong> can react with other compounds that are formed ong>inong> the atmosphere underong>inong>fluence of solar radiation. In this way they can “disappear” from the atmosphere,but then ong>inong>stead reaction products will appear. Gaseous pesticides can also beconverted to particles ong>andong> pesticide present ong>inong> particles can evaporate agaong>inong>. Thereaction product or conversion product will usually have different properties thanthe precursor ong>andong> will for that reason be removed at another rate from theatmosphere.The rate at which highly soluble ong>andong> particulate pesticides are removed from theatmosphere by ong>inong>-cloud processes is rather high compared to other removalmechanisms, ong>inong>cludong>inong>g dry deposition. It is, however, only raong>inong>ong>inong>g 5-10% of thetime, whereas dry deposition goes on all the time. As a result the yearly averageremoval rate by dry deposition can nevertheless be of the same order as the averageremoval rate by wet deposition. The ratio wet to dry removal will depend on theproperties of the compound ong>andong> on the meteorological ong>andong> surface conditions of thearea where the deposition occurs.It is relatively easy to monitor wet deposition of pesticides.Relatively close (e.g. 2 km) to a gaseous pesticide source the removal rates aredifferent from the removal rates mentioned before, because it has not yet beenmixed over the whole mixong>inong>g layer. Most pesticides are emitted from fields ontowhich pesticides are applied. i.e. from low-level sources. Close to the source is theground-level concentration relatively high, which will lead to higher dry depositionrates, which leads to a relatively high removal rate by dry deposition. Moreover,the pesticide “plume” from the field has not yet reached the clouds at short distancefrom the source ong>andong> the pesticide is not exposed to the more efficient removalprocesses ong>inong> clouds. In general it can be concluded that for gaseous pesticides closeto the source, dry deposition can be relatively more important ong>andong> wet deposition isrelatively less important, than further away from the source.46


Model developmentDeposition up to 2 kmfrom the sourceDry/wet deposition vs.deposition caused byspray drift due tosedimentationRecommendations forfurther researchWithong>inong> the framework of this project a model was developed to describe thesituation up to 2 km from a source. In this model pesticide emissions weregenerated by assumong>inong>g a concentration at the surface. It depends then on the wong>inong>dspeed ong>andong> other meteorological factors how large the emission rate will be. Themodel ong>inong>cludes also transport by wong>inong>d, vertical diffusion, dry deposition ong>andong>removal by ong>precipitationong> under the cloud.The maong>inong> conclusion from the model calculations is under average Danishconditions that less than 25% of the emission is deposited up to 2 km from a fieldwhere pesticide is applied. This holds for highly soluble gaseous pesticides. Forless soluble pesticides this number will be considerably less, more of the order of1%. This means that the largest part of the emitted pesticide will be transportedover long distances (>100-1000 km).Dry/wet deposition at distances less than 20 m from the field onto which thepesticides are applied can be of the same order as deposition caused byspray drift due to sedimentation. At larger distances dry/wet depositiondomong>inong>ates ong>inong> this case.Recommendations for further research are made. One of the most importantconclusions is that it is only possible to model the emission ong>andong> dry deposition ofpesticides if the concentration of the pesticides ong>inong> the surface (soil, vegetation,water) is modelled at the same time.47


Dansk sammendragDette afsnit af rapporten beskriver de atmosfæriske processer med relevans fortransport, sprednong>inong>g og deposition af pesticider. Endvidere angives modelresultaterfor transport og deposition af pesticider til estimerong>inong>g af den maksimale afsætnong>inong>gaf fordampede pesticider i nærmiljøet (< 2 km) fra sprøjtet mark.Pesticider eksisterer i atmosfæren i både gas- og partikelform. Tæt ved de områder,de fordamper fra, vil de hovedsagelig være i gasform. Udvekslong>inong>gen mellematmosfæren og overfladen (jord, vegetation, vong>andong>) afhænger af vong>inong>dhastigheden(turbulens) og koncentrationsforskellen mellem atmosfæren og overfladen. Nettoemissionfong>inong>der sted, hvis koncentrationen er større i overfladen end i atmosfæren.Netto-deposition fong>inong>der sted hvis det omvendte er tilfældet.Både emissions- og tørdepositionshastigheden afhænger af processer i overfladen,som styrer koncentrationen i overfladen. Det kan være processer såsom vertikaltransport i overfladen eller nedbrydnong>inong>g. Hvis processerne i overfladen foregårmeget langsommere end transporten mellem atmosfæren og overfladen, så bliverdet overflade-processerne der bestemmer emissions- og tørdepositionshastigheden.I dette tilfælde vil emissionen næsten ikke tiltage med vong>inong>dhastigheden.TørdepositionTørdepositionshastigheden er størst for letopløselige gasser. For moderat ogtungtopløselige gasser vil tørdepositionshastigheden være mong>inong>dre og den vil i højgrad afhænge af processer som foregår i overfladen.Både emission og tørdeposition tiltager med vong>inong>dhastigheden. Hvor meget detiltager afhænger af processerne i overfladen.Tørdepositionshastigheden for partikelformige pesticider afhænger ikke kun afvong>inong>dhastigheden, men også af partiklernes størrelse. Generelt set ertørdepositionshastigheden lavere for partikler end for letopløselige gasser. Der ermong>inong>imum for tørdepositionshastigheden i størrelsesong>inong>tervallet 0.1-1 µm, mens bådestørre og mong>inong>dre partikler afsættes hurtigere omend stadig langsomt sammenlignetmed letopløselige gasser.Det er svært at måle tørdeposition af gasser og partikler. Metoderne kræver enmeget stor præcision til bestemmelse af små koncentrationsforskelle. For det mesteer det kun muligt at opnå rimelige resultater for letopløselige gasser. Endvidere erdet ofte meget dyrt at monitere tørdeposition. I stedet anvendestørdepositionsmodeller og meteorologiske målong>inong>ger til at estimere tørdepositionen.Der fong>inong>des næsten ong>inong>gen feltmålong>inong>ger af tørdeposition for pesticider.Atmosfærisk diffusion tiltager med vong>inong>dhastigheden. Det samme gælder forblong>andong>ong>inong>gshøjden, som er højden af det nederste lag i atmosfæren, hvor luften erblong>andong>et godt op.Under danske forhold kan gasformige luftbårne pesticider, der er fordelt over heleblong>andong>ong>inong>gslaget, maksimalt tørdeponeres med en fjernelseshastighed på ca. 13% itimen over afgrøder og ca. 25% i timen over skov.VåddepositionHastigheden hvormed både gasformige og partikelformige pesticider fjernes fraatmosfæren med nedbør tiltager med nedbørsong>inong>tensiteten.48


Hastigheden, hvormed gasformige pesticider fjernes fra atmosfæren forårsaget afprocesser i skyer, stiger med vong>andong>opløseligheden af pesticidet. Af denne grundfjernes letopløselige gasformige pesticider med stor hastighed fra atmosfæren (64%i timen ved en nedbørsong>inong>tensitet af 1 mm i timen og 99.9% i timen ved ennedbørsong>inong>tensitet på 10 mm i timen). Gasser med en rong>inong>gere opløselighed fjernesmed en lavere hastighed. Partikelformige pesticider vil for det meste fjernes fraatmosfæren ved processer i skyerne med samme hastighed som letopløseligegasser.Gasformige pesticider som befong>inong>der sig i atmosfæren under skyerne fjernes medlavere hastighed af regndråber end i skyerne. Fjernelseshastigheden er 13% i timenved en nedbørsong>inong>tensitet på 1 mm i timen og 33% i timen ved en nedbørsong>inong>tensitetpå 10 mm i timen.Fjernelseshastigheden for den samme partikelformige pesticider er for det mestelavere under skyen end i skyen. Desuden vil den også afhænge af partiklernesstørrelse.Reaktioner, omdannelse Både tør- og våddeposition af gasformige pesticider tiltager medtil partiklervong>andong>opløseligheden af pesticidet. Der er dermed en generel tendens, til atpesticiders atmosfæriske levetid afhænger af deres vong>andong>opløselighed. Andrefaktorer kan dog også spille en rolle. Pesticider kan reagere med ong>andong>re stoffer somdannes i atmosfæren under ong>inong>dflydelse af sollys. På denne måde kan de “fjernes”fra atmosfæren, men i steder for kommer deres reaktionsprodukter. Gasformigepesticider kan også optages i partikler, eller et pesticid kan fordampe fra partikler.Reaktionsprodukterne har ong>andong>re egenskaber end udgangsstofferne og fjernes deformed en ong>andong>en hastighed fra atmosfæren.Hastigheden hvormed letopløselige gasser og partikler fjernes fra atmosfærenforårsaget af processer i skyer er forholdsvis høj sammenlignet med ong>andong>refjernelsesmekanismer, ong>inong>klusive tørdeposition. Det regner dog kun 5-10% af tiden,mens tørdeposition foregår hele tiden. Af denne grund kan den årlige tørdepositionalligevel være af samme størrelsesorden som våddepositionen. Forholdet mellemvåd- og tørdeposition afhænger af stoffernes egenskaber samt meteorologiske ogoverfladeforhold hvor depositionen fong>inong>der sted.Det er relativt enkelt at monitere våddeposition af pesticider.Forholdsvist tæt (f. eks. 2 km) ved en kilde for gasformige pesticider erfjernelseshastighederne forskellige fra de ovennævnte idet pesticidet endnu ikke erblong>andong>et op i hele blong>andong>ong>inong>slaget. De fleste pesticider fordamper fra marker, dvs. atkilderne har en lav højde. Tæt ved kilder er koncentrationen i jordhøjde relativ høj,hvilket vil give anlednong>inong>g til en forholdsvis stor fjernelse ved tørdeposition.Desuden er “pesticidfanen”, som stammer fra marken endnu ikke nået op iskyhøjde og pesticidet kan ikke udsættes for den mere effektive fjernelsesproces iskyen. Det kan konkluderes, at tæt ved kilder kan tørdeposition af letopløseligegasformige pesticider være af større betydnong>inong>g end våddeposition, mens detomvendte er tilfældet længere væk fra kilden.Modeludviklong>inong>gIndenfor rammerne af dette projekt er der udviklet en model til at beskrivesituationen op til 2 km fra en kilde. I modellen genereres emissionerne ved atantage en koncentration i overfladen. Selve emissionen vil så afhænge afvong>inong>dhastigheden og ong>andong>re meteorologiske forhold. Transport med vong>inong>den, vertikaldiffusion, tørdeposition og fjernelse af stoffer under skyerne beskrives også imodellen.49


Deposition ong>inong>denfor Den vigtigste konklusion som afledes fra modelberegnigerne er, at mong>inong>dre2 km fra en kilde end 25% af emissionen er tørdeponeret i et område op til 2 km fra en mark, der ersprøjtet med pesticid. Det gælder for letopløselige gasformige pesticider. Forpesticider, som ikke er så vong>andong>opløselige vil tallet være mong>inong>dre, mere istørrelsesorden 1%. Det betyder, at den største del af den fordampede mængdepesticid transporteres over store afstong>andong>e (> 100-1000 km).Tør/våddeposition på afstong>andong>e mong>inong>dre en 20 m nedstrøms fra en sprøjtetmark kan være af samme størrelsesorden som deposition forårsaget af af-drift pga. sedimentation. På større afstong>andong>e domong>inong>erer tør/våd deposition.Tør/vådeposition sammenlignetmed depositionforårsaget af afdriftpga. sedimentationAnbefalong>inong>ger for videre Der gives anbefalong>inong>ger for videre forsknong>inong>g. En af de vigtigste konklusionerforsknong>inong>ger, at det kun er muligt at modellere emission og tørdeposition for pesticider, hviskoncentrationen i overfladen (jord, vegetation, vong>andong>) er modelleret samtidig.50


1 Modellong>inong>g atmospheric transport ong>andong>deposition of pesticides up to 2 kmfrom a source1.1 Introductionong>Pesticidesong> can enter the atmosphere by volatilisation ong>andong> by resuspension of soilparticles where they can be attached to. After they have entered the atmospherethey can be transported over some distance before they are removed from theatmosphere.Figure 1 shows an overview of the important pesticide processes, not only ong>inong> theatmosphere, but also ong>inong> other parts of the environment. The processes that are mostrelevant to the atmospheric behaviour of pesticides ong>andong> their ong>inong>teractions will bediscussed ong>inong> the followong>inong>g sections.This report focuses on the followong>inong>g processes ong>inong> the atmosphere:a) Emission (maong>inong>ly the atmospheric aspect of emission is treated).b) Atmospheric transport ong>andong> diffusion; these processes have a great ong>inong>fluence onwhere a pesticide is depositedc) Dry deposition, which is an important removal process.d) Wet deposition, which is also an important removal process.This is a logical order goong>inong>g from the emission via transport to deposition.Limited attention will be paid to important processes as conversion of pesticidesfrom the gaseous to the particulate phase ong>andong> to photochemical processes. Thereason for this is that the scope of the report is limited to the first 2 km from asource. At such a short distance from the source these processes are not likely to bevery important ong>andong> these processes are therefore not treated extensively.A model will be presented that ong>inong>corporates emission, atmospheric transport ong>andong>diffusion ong>andong> dry ong>andong> wet deposition, so that ong>inong>teractions between these processescan be revealed. The implications of the model results will also be discussed.Order ong>inong> which the The order ong>inong> which the processes are discussed ong>inong> this report is not the sameprocesses are presented logical order as mentioned under a) to d) above. This is done for pedagogicalreasons.Knowledge of the properties of pesticides will be presented first because theydetermong>inong>e the rate at which emission ong>andong> dry ong>andong> wet deposition occur.Then turbulence, atmospheric transport ong>andong> diffusion are discussed because thisknowledge is necessary to describe surface exchange ong>inong> the followong>inong>g section. Thisong>inong>cludes the meteorological aspects of emission ong>andong> dry deposition. The emissionis not only determong>inong>ed by meteorological processes but also by processes ong>inong> thesurface (soil, plants). Although this is not withong>inong> the scope of the project, a briefsection on these processes was added. In this way the reader can see how they51


ong>inong>teract with the meteorological processes ong>andong> what the important factors are thatgovern the emission from these surfaces.Then wet deposition is treated ong>inong>cludong>inong>g removal of compounds from theatmosphere by processes ong>inong> clouds (ong>inong>-cloud scavengong>inong>g) ong>andong> below the clouds(below-cloud scavengong>inong>g).Then very briefly some ong>inong>formation is given on conversion of compounds from thegaseous to the particulate phase ong>andong> on photochemical reaction ong>andong> spray drift.Then a model is presented that ong>inong>tegrates the presented knowledge on emission,atmospheric transport ong>andong> diffusion ong>andong> dry ong>andong> wet deposition. The diffusion partof the model is tested agaong>inong>st measurements ong>andong> it is shown how the results of thevary with important parameters as the surface roughness, the dry depositionvelocity ong>andong> the ong>precipitationong> rate. Then the maong>inong> result of this project is presented:An estimate of the maximum accumulated fraction of the emission that can bedeposited withong>inong> 2 km from a field where pesticides are applied.Fong>inong>ally the results are discussed ong>andong> conclusions are drawn, ong>inong>cludong>inong>grecommendations for further research.Summarisong>inong>gconclusionsThe first ong>andong> largest part of this report discusses more general prong>inong>ciplesrelevant to emission, atmospheric transport ong>andong> dry ong>andong> wet deposition, but almostno results specific to pesticides are given. As this report is rather large, the readerwould easily miss the overview ong>andong>, moreover, the reader is ong>inong>terested ong>inong> pesticideswould be discouraged. For that reason, “ong>inong>termezzo sections” were ong>inong>cluded, thatsummarise the knowledge on important processes ong>andong> discuss their ong>inong>teractionswith reference to pesticides.Such “ong>inong>termezzo sections” are presented after the followong>inong>g subjects have beentreated:• Turbulence, atmospheric transport, diffusion ong>andong> surface exchange (ong>inong>cludong>inong>gdry deposition ong>andong> meteorological aspects of the emission) (section 1.5).• Removal of material by ong>precipitationong> (compares also dry ong>andong> wet deposition)(section 1.8).• Model results (section 1-13).In the last section on discussion ong>andong> conclusions, the ong>inong>formation from the“ong>inong>termezzo sections” is not repeated. In stead, more general conclusions arepresented ong>andong> discussed ong>andong> recommendations for further research are given.Relation to otherpesticide processesFigure 1 shows how the atmospheric processes ong>inong>volvong>inong>g pesticides arerelated to other pesticide processes, so that the reader gets an impression how theyong>inong>teract.52


Figure 1Processes for pesticides ong>inong> the environment, startong>inong>g at the left side withapplication.Processer for pesticider i miljøet, begyndende til venstre med sprøjtnong>inong>g.1.2 Some important propertiesThe atmospheric behaviour of pesticides is ong>inong>fluenced by their properties ong>andong> bymeteorological conditions. In the followong>inong>g some important properties will bediscussed that can ong>inong>fluence the atmospheric behaviour of pesticides.1.2.1 Solubility ong>inong> waterThe solubility of a substance ong>inong> water is defong>inong>ed at the maximum amount of thatsubstance that will dissolve ong>inong> pure water at a specified temperature. Above thisconcentration, two phases will exist if the pesticide is a solid or a liquid at thattemperature: a saturated aqueous solution ong>andong> a solid or liquid phase of thesubstance. The solubility is a function of the substance ong>andong> of the temperature ong>andong>can vary over many orders of magnitude.1.2.2 Vapour pressureThe vapour pressure of a pesticide is defong>inong>ed as the pressure exerted when a solidor liquid is ong>inong> equilibrium with its own vapour. It is a function of the substance ong>andong>of the temperature. The vapour pressure ong>inong>creases with temperature. The vapourpressure of pesticides can vary over many orders of magnitude ong>andong> is expressed ong>inong>mPa. Compounds with a vapour pressure larger than 0.1 mPa show a noticeablevolatilisation (Smit et al., 1998).53


1.2.3 Henry’s law coefficientThe solubility of a gas is an important parameter that ong>inong>fluences behaviour of thepesticide ong>inong> the soil, ong>inong> vegetation, water bodies ong>andong> removal of pesticides from theatmosphere by ong>precipitationong>. A measure of the solubility of a slightly soluble gas ong>inong>water is the Henry’s law coefficient, which simply states that there is a long>inong>earrelation between the concentration of a dissolved gas ong>andong> the concentration ong>inong> ong>airong>just above it. The Henry’s law coefficient can only be applied to that part of a gasthat is present as dissolved gas ong>inong> the solution. If the dissolved gas e.g. dissociatesong>inong> the solution, the Henry law coefficient describes only the relation between theconcentration of the gas ong>inong> the gas phase ong>andong> the undissociated part of the dissolvedgas. There are different defong>inong>itions of the Henry’s law coefficient, usong>inong>g differentunits ong>andong> different ratio’s (gas phase concentration/liquid phase concentration orliquid phase concentration/gas phase concentration). In this report we use thefollowong>inong>g defong>inong>ition:KHcg= (1)clWhere K H is the Henry’s law coefficient, c g is the concentration ong>inong> the gas phase(kg m -3 ) ong>andong> c l is the concentration ong>inong> the water phase. K H is a function of thecompound ong>andong> the temperature. The solubility of gases generally decreases withtemperature, i.e. K H decreases with temperature. K H of different pesticides can varyover at least 10 orders of magnitude.Relation between The partial pressure of the pesticide over a saturated pesticide solution isHenry’s law coefficient, equal to the vapour pressure above the pure pesticide. This has somesolubility ong>andong> vapour important consequences (Stork, 1995):pressure• Compounds with a low solubility like DDT have always Henry’s lawcoefficients with high values. For that reason they are already very volatile atlow concentrations ong>inong> water.• The Henry’s law coefficient be found from the solubility ong>andong> the vapourpressure above the pure compound. This method can be useful ong>inong> the case theHenry’s law coefficient has not been determong>inong>ed directly.1.2.4 The adsorption coefficient KocAdsorption:Freundlich equationong>Pesticidesong> dissolved ong>inong> the soil solution can adsorb onto soil particles. Aftersome time equilibrium will be reached where part of the pesticide is adsorbed ong>andong>another part is still ong>inong> the solution. The relation between those parts is given by theempirically derived power function known as theFreundlich equation:S=NK f C(2)where S is the sorbed concentration (kg pesticide kg -1 dry soil), C is the solutionconcentration (kg m -3 soil solution). K f ong>andong> N are empirical constants. The value ofN ong>inong> the Freundlich equation is usually less than 1 (commonly between 0.75 ong>andong>0.95; Green ong>andong> Karickhoff, 1990). It should be noted, that K f ong>inong> this equation has adimension that depends on the value of N. Although this is not strictly correct N issometimes set to 1, because then the equation is long>inong>ear, which allows simplermathematical solutions ong>inong> models:S=K C f(3)54


For many pesticide-soil combong>inong>ations K f is not known ong>andong> should either bedetermong>inong>ed or be estimated. For hydrophobic pesticides, that is non-ionic pesticideswith a water solubility less than 10 -3 mole l -1 , the adsorption to the soil is maong>inong>lydetermong>inong>ed by the adsorption to organic material ong>inong> the soil.This means that the ratio K f to the fraction of organic material ong>inong> the soil isapproximately constant for these pesticides. It is then useful to defong>inong>e an adsorptioncoefficient K oc :KocK=ffoc(4)where f oc is the mass of organic carbon per mass of dry soil. If K f (for the long>inong>earequation) is known for a soil with a certaong>inong> content of organic carbon, K f for thesame pesticide for another soil type with know organic carbon content can becalculated with (4) usong>inong>g relation from K f = f oc .K oc . But ong>inong> many cases no K oc valueis known, ong>andong> then K oc can be estimated from empirical relations of the followong>inong>gtype (Lyman et al., 1990):log Koc = alogY+b(5)where a ong>andong> b are constants ong>andong> Y is a relevant property of the pesticide, e.g. theK ow (octanol-water partition coefficient) or the water solubility.This approach to estimate K f can strictly only be applied to hydrophobic pesticidesthat show a long>inong>ear sorption isotherm ong>andong> for which sorption maong>inong>ly occurs to thesoil organic carbon. It appears, however, that this approach is also practicallysuitable for many pesticides that are slightly polar ong>andong> too water soluble to beconsidered hydrophobic (Green ong>andong> Karickhoff, 1990).1.3 Turbulence, atmospheric transport ong>andong> diffusionThis report will maong>inong>ly discuss atmospheric processes ong>inong> the planetary boundarylayer. The planetary boundary layer is defong>inong>ed as that part of the atmosphere that isdirectly ong>inong>fluenced by the presence of the earth’s surface with a time scale of aboutan hour or less. The planetary boundary layer is not constant, but varies from about100 - 3000 m, dependong>inong>g on meteorological conditions.Mechanical turbulenceAtmospheric movements are almost always turbulent. Wong>inong>d speed, wong>inong>d direction,temperature, pressure, humidity ong>andong> concentration of atmospheric constituentsshow a spatial ong>andong> temporal variability. This is caused by atmospheric whirls,called “eddies”. Large atmospheric eddies can be observed on sequences ofsatellite pictures where clouds rotate around low pressure areas. There are eddies ofall sizes ong>inong> the atmosphere, also very small ones. Near the surface they manifestthemselves through the flutter of leaves of trees, irregular movements of dustparticles, ripples ong>andong> waves on water surfaces. They cause e.g. diffusion of a plumeperpendicular to the wong>inong>d direction or exchange between the surface ong>andong> the ong>airong>.There are two different mechanisms that generate turbulence: mechanicalturbulence ong>andong> thermal turbulence. It is important to differentiate between thesetwo types of turbulence because they are associated with eddies of different sizesong>andong> lifetime, which ong>inong>fluence diffusion ong>andong> surface exchange ong>inong> a different fashion.Mechanical turbulence is generated due to friction exerted on the wong>inong>d by thesurface. This friction is caused by the roughness of the surface. As a result the wong>inong>dspeed ong>inong>creases with height. A rough surface like a forest generates moreturbulence than a smooth surface like water. Essential for this form of turbulence is55


that it is generated by the wong>inong>d. Mechanical turbulence is characterised by smalleddies, with a relatively short lifetime especially near the surface.Thermal turbulenceInfluence of turbulenceon diffusion of plumesThermal turbulence is caused by heatong>inong>g of the ong>airong> near the surface due to solarradiation. This ong>airong> is somewhat warmer than the surroundong>inong>g ong>airong>, has consequentlya lower density, ong>andong> is lifted up. Colder ong>airong> is takong>inong>g its place. Due to these ong>airong>movements larger, so called “convective”, eddies are generated. They haverelatively long lifetimes ong>andong> cause diffusion due to upward ong>andong> downward ong>airong>movements that can last up to 10-20 mong>inong>utes.Close to a poong>inong>t source the plume is narrow. In this case only eddies of asize smaller than the plume width can cause diffusion, i.e. mixong>inong>g by exchange ofthe polluted ong>airong> parcels with the clean ong>airong> parcels (Figure 2).;;;;;;;;;;;;;;;;;QQQQQQQQQQQQQQQQQ;;;;;;;;;;;;;;;;;QQQQQQQQQQQQQQQQQ;;;;;;;;;;;;;;;;;QQQQQQQQQQQQQQQQQ;;;;;;;;;;;;;;;;;QQQQQQQQQQQQQQQQQ;;;;;;;;;;;;;;;;;QQQQQQQQQQQQQQQQQ;;;;;;;;;;;;;;;;;QQQQQQQQQQQQQQQQQ;; QQ Figure 2Schematic illustration of mixong>inong>g of a plume by exchange of ong>airong> parcels between theplume ong>andong> the ong>airong> outside the plume.Skematisk illustration af blong>andong>ong>inong>g af en røgfane ved udvekslong>inong>g af luftpakkermellem fanen og luften udenfor røgfanen.Larger eddies close to the source do not cause diffusion of the plume, but lead to adisplacement (“meong>andong>erong>inong>g”) of the whole plume (Figure 3). At larger distancefrom the source, when the plume has become wider, larger ong>andong> larger eddies willalso play a role ong>inong> the diffusion.56


;;;;;;;;;;;;;;;;;QQQQQQQQQQQQQQQQQ;;;;;;;;;;;;;;;;;QQQQQQQQQQQQQQQQQ;;;;;;;;;;;;;;;;;QQQQQQQQQQQQQQQQQ;;;;;;;;;;;;;;;;;QQQQQQQQQQQQQQQQQ;;;;;;;;;;;;;;;;;QQQQQQQQQQQQQQQQQ;;;;;;;;;;;;;;;;;QQQQQQQQQQQQQQQQQ;;;;;;;;;;;;;;;;;QQQQQQQQQQQQQQQQQ;;;;;;;;;;;;;;;;;QQQQQQQQQQQQQQQQQFigure 3Effect of large eddies on the shape of a plume.Effekt af store hvirvler på røgfanens form.Vertical temperaturegradientIn the atmosphere the pressure decreases with height. Due to this pressuredecrease an ong>airong> parcel that is lifted up rapidly by e.g. atmospheric turbulence willexpong>andong>. Some energy is needed for this expansion ong>andong> this will be taken from theong>airong> parcel itself, so that the ong>airong> cools down ong>andong> consequently gets a higher density.As a result the temperature of the ong>airong> parcel will decrease with height at a rate of0.01°C m -1 if there are no other processes that ong>inong>fluence the temperature. If an ong>airong>parcel is moved downward rapidly its temperature will due to the same mechanismong>inong>crease with 0.01°C m -1 . In that case the ong>airong> parcel will get a lower density. Parcelsthat are lifted upward or downward will show this temperature change if theirmovement is relatively fast, so that no other mechanisms can ong>inong>fluence theirtemperature. Ideally one would expect a temperature gradient of -0.01°C m -1 ong>inong> theatmosphere. But over longer time periods other processes thanexpansion/compression, like solar radiation, coolong>inong>g due to long wave radiationfrom the ong>airong> (“radiative coolong>inong>g”), condensation of water vapour to clouds orevaporation of clouds may lead to vertical temperature gradients ong>inong> the realatmosphere that deviate from the theoretical gradient of -0.01°C m -1 .Stable atmosphere Is the vertical temperature gradient ong>inong> the real atmosphere less than -0.01°C m -1 ,then a risong>inong>g ong>airong> parcel (of which the temperature still changes with -0.01°C m -1 )will become colder ong>andong> hence more dense than the surroundong>inong>g ong>airong> ong>andong> will show atendency to move downward to the level where it came from. If an ong>airong> parcel isforced to move downward ong>inong> the same situation it will become warmer ong>andong> henceless dense than the surroundong>inong>g ong>airong> ong>andong> will show a tendency to move upward tothe level where it came from. In such a situation the vertical movements, e.g.generated by mechanical turbulence are suppressed ong>andong> the atmosphere is called“stable”. This situation occurs often ong>inong> a cloudless atmosphere durong>inong>g night time,when the ong>airong> close to the surface is cooled down because it looses its energy byradiation. In such an atmosphere there is not much turbulence at all.Temperature ong>inong>versionAn extreme case is where the temperature ong>inong> the real atmosphere ong>inong>creases withheight (“temperature ong>inong>version”). Vertical movements are then suppressed so muchthat there is almost no exchange across the ong>inong>version ong>andong> the wong>inong>d speed at eitherside of the ong>inong>version can differ much.57


Unstable atmosphereNeutral atmosphereEffect of stability ondiffusionDiffusion ong>inong> a stableatmosphereIs the vertical temperature gradient ong>inong> the real atmosphere more than-0.01°C m -1 , then a risong>inong>g ong>airong> parcel of which the temperature still changes with -0.01°C m -1 will become warmer ong>andong> hence less dense than the surroundong>inong>g ong>airong> ong>andong>will contong>inong>ue to rise ong>andong> even accelerate, until it reaches a part of the atmospherewhere the vertical temperature gradient is less than-0.01°C m -1 . If ong>inong> the same situation an ong>airong> parcel is forced to moved downward itwill become colder ong>andong> hence more dense than the surroundong>inong>g ong>airong> ong>andong> willcontong>inong>ue to move downward ong>andong> even accelerate, until it reaches a part of theatmosphere where the vertical temperature gradient is less than-0.01°C m -1 or it reaches the surface. In such a situation the vertical movementsgenerated by e.g. mechanical turbulence are stimulated, ong>andong> mixong>inong>g up to largerheights occurs. The atmosphere is called “unstable” ong>inong> such situations. Thissituation occurs often ong>inong> a cloudless atmosphere durong>inong>g daytime ong>inong> the summer,when the earth’s surface is warmed up by radiation ong>andong> warm “ong>airong> bubbles” risefrom the surface ong>andong> can even rise so high up that they lead to formation ofcumulus clouds. In this situation thermal turbulence is important.In a neutral atmosphere the temperature gradient is -0.01°C m -1 ong>andong> mechanicalturbulence domong>inong>ates. This situation occurs often when it is cloudy ong>andong> wong>inong>dy. InWestern Europe the atmosphere is much more frequently neutral or nearly neutralthan stable or unstable. For that reason most examples that are presented ong>inong> thisreport are for a neutral atmosphere.Atmospheric stability has an effect on diffusion. The effect of atmosphericstability on diffusion ong>inong> the vertical for an elevated poong>inong>t source is illustrated byFigure 4.In a stable atmosphere the plume is narrow ong>andong> can be observed at longdistances from the chimney, because the diffusion is reduced ong>andong> consequently theplume is not diluted much. Usually the wong>inong>d speed is relatively low ong>inong> a stableatmosphere ong>andong> the variation ong>inong> wong>inong>d direction can be relatively large. The plume issaid to be “fannong>inong>g”. In the case of a ground-level source, like a field afterapplication of pesticides, the plume is also very narrow ong>andong> the concentration isrelatively high close to the ground.Diffusion ong>inong> an unstable In an unstable atmosphere there are strong vertical movements. This doesatmospherenot only lead to faster diffusion ong>andong> dilution, but causes also to plume to reach thesurface at a relatively short distance from the chimney. The plume is said to be“loopong>inong>g” ong>inong> this case. In the case of a ground-level sources the averageconcentration at ground-level is relatively low compared to the stable case, but atsome distances durong>inong>g a short time relatively high concentrations can be observed.a. Stable atmosphere.58


. Unstable atmosphere.c. Neutral atmosphere.Figure 4The ong>inong>fluence of atmospheric stability on the vertical mixong>inong>g of a plume.Indflydelse af atmosfærisk stabilitet på vertikal opblong>andong>ong>inong>g af en røgfane.Diffusion ong>inong> a neutralatmosphereIn the neutral atmosphere the plume is somewhat wider than ong>inong> a stable atmosphere, is better mixed ong>andong> cannot be observed over such long distances becauseit is diluted more rapidly by diffusion. In this case high concentrations are notobserved close to the source as is the case ong>inong> an unstable atmosphere. The plume issaid to be “conong>inong>g” ong>inong> this case.Figure 4 illustrates that it is important to take atmospheric stability ong>inong>to accountwhen describong>inong>g atmospheric diffusion. Atmospheric stability also ong>inong>fluences theexchange between the surface ong>andong> the atmosphere. The higher up ong>inong> theatmosphere, the more important it is to take atmospheric stability ong>inong>to accountwhen describong>inong>g the exchange between the surface ong>andong> that height.Wong>inong>d speed profileWong>inong>d speed profile ong>inong>a neutral atmosphereAs mentioned previously, the wong>inong>d speed near the surface is retarded by friction atthe surface. By how much, will depend on the surface roughness. The wong>inong>d speedat above about 500 m is generally not ong>inong>fluenced by the surface, but at lowerheights it is ong>inong>fluenced. At about 60 m height the wong>inong>d speed is ong>inong>fluenced more bythe surface roughness of a larger area (about 5×5 km 2 ). At lower height the wong>inong>dspeed is more ong>inong>fluenced by the local surface roughness.Measurements of the wong>inong>d speed as a function of height have revealed thatthe wong>inong>d speed ong>inong>creases with the logarithm of the height:u*⎛ z ⎞uz ( ) = ln⎜⎟κ ⎝ z 0 m ⎠(6)where u(z) is the wong>inong>d speed (m s -1 ) at height z (m); z 0m is the surface roughnesslength (m) ong>andong> is the extrapolated height at which the wong>inong>d speed is 0, z 0m is of theorder of 1/10 th of the height of the obstacles (vegetation, trees etc.); u * is the59


friction velocity (m s -1 ) ong>andong> is a measure of mechanical turbulence; κ is the vonKarman’s constant ≈ 0.4 (dimensionless). With this equation it is possible tocalculate the wong>inong>d speed at one height from the wong>inong>d speed at another height if thesurface roughness is known. The wong>inong>d speed profile can be described with thesame type of function for stable ong>andong> unstable conditions. It has then to be correctedsomewhat so that the non-neutral situation is described correctly (Arya, 1988).Surface roughnessIn Table 1 values for the surface roughness length are presented for differentsurfaces.Table 1Surface roughness length of different surfaces (Stull, 1988).Ruhedshøjder for forskellige overflader (Stull, 1988).SurfaceSurface roughness length (m)Ice, mud flats 0.00001Open sea at wong>inong>d speed of 3 m s -1 a) 0.00005Snow covered flat or rollong>inong>g ground 0.00006Open sea at wong>inong>d speed of 10 m s -1 a) 0.0003Cut grass (∼ 0.03 m high) 0.006Long grass, crops 0.04Farmlong>andong> ong>inong>cl. some trees 0.25Forest 1.00Centres of cities 2.00a) The surface roughness of the sea depends on the wong>inong>d speed, see (7).It should be noted that the surface roughness is not constant ong>inong> agricultural areas,but depends on the heights of the crops, which vary durong>inong>g a year. The surfaceroughness length for the open sea is not constant either, but is a function of thewong>inong>d speed (Long>inong>dfors et al., 1991):z0m013 . ν 0.0144u= +u g*2*(7)where ν is the kong>inong>ematic viscosity of ong>airong> (∼ 1.5x10 -5 m 2 s -1 ) ong>andong> g is the gravitation(9.81 m s -2 ). In the wong>inong>d speed range of 0-3 m s -1 the roughness length decreaseswith wong>inong>d speed, for larger wong>inong>d speeds the roughness length ong>inong>creases, maong>inong>lybecause the wave height ong>inong>creases.Effect of surfaceroughness on wong>inong>dspeed profileThe surface roughness varies with the nature of the terraong>inong>. For that reasonthe wong>inong>d speed near the surface will be a function of the surface roughness.This means that under the same meteorological conditions (i.e. same wong>inong>d speed atgreater height) the wong>inong>d speed near the surface will be different for e.g. bare soil,crops, forest ong>andong> water. The friction velocity u * , which is a measure of themechanical turbulence, is ong>inong> that case larger for a more rough surface like a cropthan for bare soil. As a result the wong>inong>d speed near the surface will decrease withsurface roughness. Figure 5 shows some vertical wong>inong>d profiles for two differentsurfaces under neutral conditions: a crop ong>andong> an almost bare soil. The surfaces arechosen so that they represent typical situations encountered ong>inong> agricultural areas onor near fields where pesticides are applied.60


Figure 5Vertical wong>inong>d speed profile over a crop with a surface roughness length of 0.1 mong>andong> almost bare soil with a surface roughness length of 0.006 m. It is assumed thatong>inong> both cases the wong>inong>d speed at 60 m height is the same.Vertikal vong>inong>dhastighedsprofil over en afgrøde med en ruhedshøjde på 0,1 m ognæsten bar jord med en ruhedshøjde på 0,006 m. Det er antaget, at vong>inong>hastighedeni 60 m højde er det samme i begge tilfælde.Effect of wong>inong>d speedprofile on atmospherictransportEffect of surface roughnesson atmospherictransportWong>inong>d direction profileMixong>inong>g heightThe existence of a wong>inong>d speed profile ong>inong>fluences the average speed at whicha released compound is transported ong>inong> the atmosphere. At some distance fromthe source part of the released compound has been transported upward by diffusionong>andong> encounters a higher wong>inong>d speed than near the surface. This means that theaverage speed at which a compound is transported ong>inong>creases with the distance tothe source until it is mixed over the whole mixong>inong>g layer (see below).Figure 5 shows that the wong>inong>d speed is higher ong>inong> the case of the bare soil.This means that a compound released from a field with almost bare soil istransported at a greater speed, than a compound released from a field with a crop.At some distance from the released poong>inong>t, however, this difference is not anylonger so large, because the ong>airong> has been transported over areas with other surfaceroughness lengths. Moreover, the compound is then transported higher up ong>inong> theatmosphere where the wong>inong>d speed is less ong>inong>fluenced by the surface.Not only the wong>inong>d speed is ong>inong>fluenced by the presence of the surface, but also thewong>inong>d direction. Usually the wong>inong>d direction is veerong>inong>g with height ong>inong> the NorthernHemisphere. This means that the origong>inong> of the ong>airong> at greater heights is differentfrom that at ground-level.In some parts of the atmosphere stable layers may exist, where vertical ong>airong>movements are suppressed. This means that ong>airong> origong>inong>atong>inong>g from below cannot betransported across this layer ong>andong> this layer is than functionong>inong>g as a “lid” on theatmosphere below, where mixong>inong>g can occur. Air from above this layer can also notbe transported downward. In case of tall stacks that emit pollution above this lid,this is a favourable situation, because the pollution cannot reach the earth’s surfacewhere humans live. In case of emission of pollution from low or ground level61


sources, as is the case for pesticides, this is an unfavourable situation, where veryhigh concentrations can occur near the ground. This can especially be the casedurong>inong>g night time when temperature ong>inong>versions close to the ground are observedfrequently.The “mixong>inong>g layer” is the layer nearest to the earth’s surface where mixong>inong>g takesplace. It is bounded by the surface ong>andong> the first layer where vertical movements aresuppressed. The height of the mixong>inong>g layer is called “mixong>inong>g height”, It isimportant to know the mixong>inong>g height, at least when transport of pollutants at somedistance from the source has to be described, because the pollution plume may thenhave come so wide that it has reached the mixong>inong>g layer height. Further dilution ong>inong>the vertical is then not any longer possible ong>andong> this will ong>inong>fluence theconcentrations at the surface.The mixong>inong>g height shows large diurnal variations. Durong>inong>g a cloudless night,radiative coolong>inong>g may cause the temperature near the surface may to drop so muchthat a temperature ong>inong>version is observed. As we have seen before this leads to highconcentrations at ground-level near low sources. If the atmosphere is also cloudlessafter sunrise the next day, the earth’s surface will be heated ong>andong> the ong>inong>versiondisappears. Also wong>inong>d can have an effect on the mixong>inong>g height. If there is muchwong>inong>d durong>inong>g a cloudless night, the ong>airong> is well-mixed ong>andong> ong>airong> that is cooled downnear the surface is transported upward so that the mixong>inong>g height will not be close tothe surface. It must be noted here that water bodies (lakes, seas) do not show thelarge diurnal temperature variations as the upper layer of the soil. For that reasonthe mixong>inong>g height over sea will generally be different than over long>andong>.Durong>inong>g night time the mixong>inong>g height is often below 200 m, whereas it is oftenhigher than 500 m durong>inong>g day time. If the atmosphere is very unstable the mixong>inong>gheight can be ong>inong>defong>inong>ite (i.e. over 2000 m).For neutral ong>andong> stable conditions there is a relation between the friction velocityong>andong> the mixong>inong>g height:zmix= cu 1 *fcor(8)where z mix = mixong>inong>g height (m); c 1 = a constant; f cor = Coriolis parameter (s -1 ) ong>andong> isgiven by 2Ωsong>inong>(lat), where Ω = angular velocity of the earth (radians s -1 ) ong>andong> lat =latitude (radians), for a latitude of 50° N the f is 1.11x10 -4 s -1 . Van Jaarsveld (1995)uses a value of 0.08 for c 1 . This value is chosen so that reasonable results areobtaong>inong>ed for compounds that are transported over long distances. It should benoted, that other scientists use different values. In the short range dispersion modelOML developed at the National Environmental Research Institute, Roskilde,Denmark a value of 0.25 is used for c 1 (Berkowicz, National EnvironmentalInstitute, Roskilde, Denmark, personal communication). The reason for usong>inong>g thismuch higher value is that otherwise observed concentrations near power plantswith high stacks cannot be reproduced by the OML model. So there is considerableuncertaong>inong>ty ong>inong> the function for the mixong>inong>g height. Equation (8) shows that themixong>inong>g height ong>inong>creases with the friction velocity u * , i.e. with the wong>inong>d speed. As u *also depends on the surface roughness, the mixong>inong>g height is ong>inong> prong>inong>ciple a functionof the surface roughness. This is, however, not a local surface roughness, but asurface roughness of the whole long>andong>scape as “seen” from the mixong>inong>g height.The mixong>inong>g height under unstable conditions can only be found frommeasurements (vertical temperature profile) or from a more complicated model that62


describes the development of the mixong>inong>g height durong>inong>g the day as a function of e.g.solar radiation.The diffusion near a poong>inong>t source produces a normally distributed time-averaged concentration perpendicular to the plume axis. Such a Gaussiong>andong>istribution can be described with a stong>andong>ard deviation, just as the normaldistribution used ong>inong> statistics. In this case, however, the stong>andong>ard deviation is notconstant, but is a function of the distance to the source. As there is diffusion ong>inong> alldirections, the concentration can be described with a normal distribution ong>inong> alldirections. If the poong>inong>t source is contong>inong>uous, the diffusion ong>inong> the wong>inong>d direction canbe neglected, because the effect of diffusion at subsequent time steps willcompensate each other ong>inong> this case. Figure 6 illustrates how the concentrationdistribution at ground-level as a function of distance from a poong>inong>t source looks like.It can be seen that the plume becomes wider ong>andong> the maximum concentration loweras a function of the distance to the source. In this case the x direction is parallel tothe wong>inong>d direction (the wong>inong>d blows from left to right), the y direction is ong>inong> thehorizontal perpendicular to the wong>inong>d direction. On the vertical axis theconcentration is shown.Atmospheric diffusionnear a poong>inong>t sourcecyxSourceFigure 6Ground-level concentration due to a poong>inong>t source for several downwong>inong>d cross-wong>inong>dsections.Koncentration i jordniveau på tværs af vong>inong>den i forskellige afstong>andong>e fra enpunktkilde.MathematicaldescriptionThe diffusion of ong>airong>borne material from a poong>inong>t source can be describedmathematically with the Gaussian distribution:222y ⎡ ( z−h) ( z+h)⎤−Qcxyzu x x e 2−−2σ y ( x) ⎢e 2z( x) e22σ 2σz( x)⎥( , , ) = ⎢ + ⎥2π σy( ) σz( )⎢⎥⎣⎦(9)In this model it is assumed that the wong>inong>d direction is parallel to the x axis, that the yaxis is perpendicular to the plume axis ong>inong> the horizontal direction, ong>andong> that the zaxis is perpendicular to the plume axis ong>inong> the vertical direction. The coordong>inong>atesystem starts at ground level at the position of the poong>inong>t source (there x,y ong>andong> z are0); c(x,y,z) is the concentration (kg m -3 ); Q is the source strength (kg s -1 ); u is theaverage wong>inong>d speed at stack height (m s -1 ); h is the stack height (m), σ y (x) ong>andong>σ z (x) are the stong>andong>ard deviations of the concentration distribution ong>inong> the y ong>andong> z63


direction (m). The second exponential function withong>inong> the brackets results fromreflection of the plume at the earth’s surface. c(x,y,z), u , σ y (x) ong>andong> σ z (x) shouldreflect the same averagong>inong>g time. The stong>andong>ard deviations of the concentration σ y (x)ong>andong> σ y (x) are determong>inong>ed experimentally ong>andong> depend on meteorological conditions(atmospheric stability), surface roughness, averagong>inong>g time ong>andong> to some extent onthe source height.Stong>andong>ard deviations For neutral conditions σ y (x) (m) ong>andong> σ z (x) can be described by estimates offor concentration Briggs for a rural situation (equations (10) ong>andong> (11); Pasquill ong>andong> Smith, 1983).σ y( x)=008 . x−1+ 10 . × 10 4 x(10)σ z( x)=006 . x−1+ 15 . × 10 3 x(11)It should be noted here that these equations are, ong>inong> fact, only valid for elevatedreleases. The values for surface releases can be up to a factor 2 larger. σ y (x) ong>andong>σ z (x). For other than neutral conditions there are also equations forσ y (x) ong>andong> σ z (x) available (Pasquill ong>andong> Smith, 1983).Figure 7 ong>andong> equation (11) show how σ z (x) ong>inong>creases as a function of distance tothe source. Close to the source it ong>inong>creases long>inong>early with x ong>andong> at greater distanceproportionally with x . This is important to notice, because that means that thediffusion close to a source is not the same as the diffusion at some distance fromthe source.Figure 7Stong>andong>ard deviation for concentration ong>inong> the vertical (σ z (x)) accordong>inong>g to Briggs asa function of distance to a poong>inong>t source.Stong>andong>ardafvigelse for koncentrationen i vertikal retnong>inong>g (σ z (x)) i følge Briggs somen funktion af afstong>andong>en til en punktkilde.64


Area sourceK-modelA field onto which pesticides have been applied does not act as a poong>inong>t source, butas an area source because the whole area is covered with pesticides. The diffusionfrom an area source can be described ong>inong> the same way as the diffusion from a poong>inong>tsource. For that reason it can be simulated ong>inong> a model by puttong>inong>g many small poong>inong>tsources on the field. Another method to simulate an area source ong>inong> a model is byusong>inong>g one poong>inong>t source that is at such an upward distance from that σ y (x) has thesame width at the position of the field ong>inong> the model as the field itself (“virtual poong>inong>tsource”). These approaches give good results at some distance from the field, butnear the field ong>andong> especially on the field itself they do not give an accuratedescription.There exists another type of diffusion model, a so called “K-model”. In such amodel the atmosphere is divided ong>inong>to cubic ong>airong> parcels ong>andong> the diffusive transportbetween these ong>airong> parcels ong>inong> all directions is described with so called “eddydiffusivity coeffcients”. We will here only focus on the diffusion ong>inong> the vertical,which is described ong>inong> the followong>inong>g way: The atmosphere is divided ong>inong>to verticallayers. Between these layers the diffusion is described by an eddy diffusivitycoefficient (m 2 s -1 ), which for a neutral atmosphere is:KHeat ( z) =κu*z(12)From the equation it can be seen that the K Heat (z) ong>inong>creases with height. Thisassumption is, ong>inong> fact, not entirely correct as K Heat (z) decreases short before themixong>inong>g height is reached, but the above equation leads nevertheless to goodsimulations of ground-level concentrations. It can be seen from (12) that K Heat (z) isa function of the friction velocity u * . This type of model is called a K-model,because it uses eddy diffusivity coefficients K to describe the diffusion. For otherthan neutral conditions (12) can still be used, but then some corrections have to bemade.Difference between a K- It should be noted, that there is a difference ong>inong> how a K-model ong>andong> a Gausmodelong>andong> a Gaussian sian plume model describe the diffusion as a function of distance to theplume modelsource. The results of a K-model for which K Hear (z) does not vary with height as ong>inong>(12) behaves ong>inong> the same fashion a Gaussian plume model for which the σ z (x)ong>inong>creases is everywhere with x . This is different from the description of a realplume, where σ z (x) ong>inong>creases with x close to the source ong>andong> with x at greaterdistance. If a K-model is applied where K Heat (z) ong>inong>creases with height the diffusionwill ong>inong>crease with distance to a ground-level source. The reason is that anong>inong>creasong>inong>g fraction of the released compound is transported upward, where thediffusivity is greater than near the surface. This effect will be enhanced if also thewong>inong>d speed is made a function of the height. So the approach ong>inong> the two types ofmodels is different ong>andong> for that reason they will also give somewhat differentresults.A K-model can adequately describe area sources ong>andong> even concentrations withong>inong>the area source (a field onto which pesticides are applied) can be modelled well.Gaussian plume models are not able to describe concentrations withong>inong> area sourceswell. A K-model can also better hong>andong>le dry deposition, i.e. deposition that takesplace at the surface (see also the next section). As a result of dry deposition theconcentration near the surface decreases, which is more difficult to reproduce witha Gaussian plume model.Effect of surface Equation (12) shows that the diffusion depends u * . As was noted before u *roughness on diffusion ong>inong>creases with surface roughness. Consequently the diffusion ong>inong>creases also withthe surface roughness.65


1.4 Surface exchangeExchange of material between the atmosphere ong>andong> the surface is caused by eddiesong>inong> the atmosphere. The simplest way of describong>inong>g the process is usong>inong>g a “big leafmodel”. In this model the surface is treated as one big leaf on a tree ong>andong> not as avery complicated surface with differences ong>inong> surface properties as ong>inong> reality.Big leaf modelThe transport is considered to consist of three distong>inong>ct steps. These steps can bedescribed with resistances ong>inong> series by analogy with electrical circuits. The firstpart is the transport by turbulent diffusion from a certaong>inong> height ong>inong> the atmosphere(reference height) to a very thong>inong> (∼ 1 mm) layer just above the surface. Theresistance belongong>inong>g to this part is called aerodynamic resistance (r a ). The secondpart of the transport is through the thong>inong> layer. In this layer the flow is lamong>inong>ar (= notturbulent) ong>andong> the transport has to occur by non-turbulent diffusion. For gases thisis the molecular diffusion, for small particles this is the Brownian diffusion. Forlarger particles ong>inong>ertial ong>andong> gravitational ong>effectsong> become important, the lattercharacterised by the settlong>inong>g velocity. The thong>inong> layer is called lamong>inong>ar boundarylayer ong>andong> the resistance is called lamong>inong>ar boundary layer resistance (r b ). Then thematerial arrives at the surface. The properties of the surface: solubility,reactivity/degradation, transport velocity ong>inong>to the surface determong>inong>e how much ofthe material that arrives at the surface is taken up. This resistance associated withthis step is called surface resistance (r c ). It is ong>inong> fact an over-all resistance categorydescribong>inong>g many processes. The best thong>inong>g to do is, ong>inong> fact, not to use just a valuefor the surface resistance ong>inong> the big leaf model, but to model the surface resistanceby modellong>inong>g all relevant processes that take place ong>inong> the surface. In that way it isalso possible to understong>andong> the underlyong>inong>g mechanisms ong>andong> to model e.g. temporalvariations ong>inong> the surface resistance. This is especially necessary for pesticides,because there are no measurements of the exchange of pesticides between theatmosphere ong>andong> the surface ong>andong> because some pesticides can be re-emitted oncethey have been deposited. If models that describe the fate of pesticides ong>inong> soil,vegetation ong>andong> water bodies are coupled to models for the atmospheric part of theexchange, it will be possible not only to model deposition, but also emission frome.g. bare soil, crops etc.Big leaf model for gases Figure 8 shows the big leaf model for gases, which describes the dry deposition tolong>andong> surfaces. Highly soluble/reactive gases like nitric acid (HNO 3 ) areimmediately taken up by the surface, the surface resistance will be negligible ong>andong>the over-all transport is determong>inong>ed by the aerodynamic ong>andong> lamong>inong>ar boundary layerresistance. An ong>inong>ert gas like helium (He) is not at all taken up ong>inong> the surface. Thesurface resistance will ong>inong> that case be ong>inong>defong>inong>itely large ong>andong> determong>inong>es the over-alltransport: although He is transported to the surface it is also transported from thesurface so that no net transport occurs. We can conclude that the exchange velocitydepends both on the properties of the gas ong>andong> the properties of the surface ong>andong> thatdependong>inong>g on theses properties either the atmospheric processes (characterised bythe aerodynamic ong>andong> lamong>inong>ar boundary layer resistance) domong>inong>ate the velocity orthe surface processes or both.66


Figure 8“Big leaf” model for dry deposition of gases.“Big leaf” model for tørdeposition af gasser.Aerodynamic resistance The aerodynamic resistance is the same for gases ong>andong> particles. For neutralfor gases ong>andong> particles conditions it is given by:ra1 ⎛ z= ln⎜κ u * ⎝ zr0m⎞⎟⎠(13)where r a = the aerodynamic resistance (s m -1 ) ong>andong> z r is the reference height (m).Equation (13) shows that r a decreases with u * . This means that the exchangeong>inong>creases with u * ong>andong> therefore also with the wong>inong>d speed. This way of expressong>inong>gthe exchange gives exactly the same results as the eddy diffusivity conceptpresented ong>inong> the previous section. The same type of function but then withcorrection factors can be used to describe the aerodynamic resistance for other thanneutral conditions.Lamong>inong>ar boundary layer For gases the lamong>inong>ar boundary resistance is given by Hicks et al. (1987):resistance for gasesrb =⎛ Sc⎞2⎜⎟⎝ Pr ⎠κu*23(14)where r b = lamong>inong>ar boundary layer resistance (s m -1 ); Pr = Prong>andong>tl number(dimensionless = 0.72); Sc = Schmidt number (dimensionless) which can be foundfrom Sc = ν/D g , where ν = kong>inong>ematic viscosity of ong>airong> (m 2 s -1 ) ong>andong> D g is thediffusivity of the gas ong>inong> ong>airong> (m 2 s -1 ). Sc is not a function of temperature, so for ν ong>andong>D g the values for 25° C can be taken; ν at 25° C is 1.55x10 -5 m 2 s -1 ong>andong> D g dependson molecular mass of the gas. The larger the molecule the smaller the diffusivity,because large molecules do not move so fast. If D g is not known, it can beestimated by:67


Dg−105 . x10 4=Mg(15)where D g is ong>inong> m 2 s -1 ong>andong> M g is the molecular mass of the gas (g mol -1 ).This description of r b is still somewhat uncertaong>inong>. It is important to note thatpesticides have much higher molecular masses (typical 200-300 g mol -1 ) than morefrequently studied atmospheric gases. Equations (14) ong>andong> (15) show us that r bdecreases with the molecular mass of the compound that is exchanged.Gaseous fluxThe exchange of gases between the atmosphere ong>andong> the surface depends on boththe concentration ong>inong> the atmosphere ong>andong> the concentration ong>inong> the surface ong>andong> isgiven by:* 1*Fg = ve( cgong>airong> , − cgsurf, ) = − ( cgong>airong>, − csurf)r + rab(16)where F g = flux (kg m -2 s -1 ), which is by defong>inong>ition negative if the material isremoved from the atmosphere; v e is the exchange velocity (m s -1 ) ong>andong> is equal to1/(r a + r b ) ; c g,ong>airong> is the concentration of the compound ong>inong> the ong>airong> at reference height(kg m -3 ); c g,surf * is the (theoretical) gas phase concentration ong>inong> the surface. If the gasphase concentration ong>inong> the surface is not known it can be calculated from theconcentration ong>inong> the water phase ong>andong> the Henry’s law coefficient: c g,surf * = Hc w ,where c w is the water phase concentration (kg m -3 ). If c g,ong>airong> is larger than c g,surf * drydeposition will occur. If c g,surf * is larger than c g,ong>airong> emission will occur. ong>Pesticidesong>like long>inong>dane can be re-emitted after beong>inong>g dry or wet deposited. So it is importantthat both ways can be described, but then processes ong>inong> the surface that ong>inong>fluence theconcentration ong>inong> that surface should be described as well. Both r a ong>andong> r b decreasewith u * . This means that the flux ong>inong>creases with wong>inong>d speed.For compounds for which reemission is not important ong>andong> for which the surfaceresistance has been measured (or can be modelled), the flux is described by:F1=− v c =−r + r + rg d g, ong>airong>gong>airong> ,a b cc(17)In this equation v d is the dry deposition velocity (m s -1 ), which is equal to 1/(r a + r b+ r c ); r c is the surface resistance (s m -1 ). In this case not only atmospheric processesare described as with (16), but also surface processes. The flux ong>inong>creases also ong>inong>this case with the wong>inong>d speed. It should be noted here that both r a ong>andong> r b are afunction of the wong>inong>d speed, but r c is not. If the surface resistance r c is smallcompared to r a ong>andong> r b (almost all material is taken up by the surface) the fluxong>inong>creases relatively much with wong>inong>d speed. If the surface resistance r vc is relativelylarge, the flux does not ong>inong>crease much with wong>inong>d speed.Dry deposition to the sea The exchange of gases between the atmosphere ong>andong> water surfaces depends onturbulence ong>inong> the ong>airong>, solubility of the gas,ong>andong> also on the turbulence ong>inong> the water(Liss ong>andong> Slong>inong>n, 1983; Asman et al., 1994). As this deposition is beyond the scopeof this report it will not be discussed here.Maximum drydeposition velocity forgases to long>andong> surfacesAs can be seen from (17) the dry deposition velocity v d has a maximumvalue if the surface resistance r c is 0, i.e. the dry deposition velocity is thentotally determong>inong>ed by the atmospheric processes. This maximum dry depositionvelocity will be a function of the friction velocity, which can vary appreciably. Thefriction velocity depends on the wong>inong>d speed at greater height ong>andong> the surface68


oughness. It is possible, however, to calculate an average maximum dry depositionvelocity for a area ong>andong> surface roughness. For Danish conditions this averagemaximum dry deposition velocity will be of the order of 2×10 -2 m s -1 for croplong>andong>(surface roughness length of 0.1 m) ong>andong> of the order of 4×10 -2 m s -1 for forests(surface roughness length of 1 m). Assumong>inong>g that the concentration of thecompound is the same everywhere over a mixong>inong>g height of 500 m, only 13% ong>andong>25% of the compound is removed after 1 hour with dry deposition over croplong>andong>respectively forests (22). This shows that dry deposition is not a fast process, evennot for the fastest depositong>inong>g gas. The only exception is close to a low source wherethe compound is not yet mixed over the whole mixong>inong>g layer ong>andong> the concentrationis relatively high for that reason.Determong>inong>ation of theuptake rate by drydeposition of gaseouscompounds ong>inong> thelaboratoryThe dry deposition velocity for many organic compounds, like pesticides, isnot so high ong>andong> is more determong>inong>ed by the processes that play a role ong>inong> thesurface. This gives the possibility to study uptake by dry deposition ong>inong> thelaboratory, where it is difficult to generate realistic atmospheric turbulence.The only condition is then that there is enough turbulence ong>inong> the experiment,created e.g. by a ventilator, that the diffusion ong>inong> the atmosphere does not limit theuptake rate. In that case the uptake rate is only determong>inong>ed by processes ong>inong> thesurface. These processes ong>inong> the surface are diffusion ong>inong>to the surface, adsorptionong>andong> degradation. For water also mixong>inong>g ong>inong> the water itself plays a role. The uptakerate will not be constant ong>inong> such experiments, but will decrease with time becausethe surface gets saturated with the pesticide.Duyzer ong>andong> van Oss (1997) performed laboratory experiments where theymeasured the uptake rate of organic chemicals by soil, fresh water, sea water ong>andong>by grass as a function of time. They developed also simple models for uptake as afunction of time ong>andong> were ong>inong> most cases able to obtaong>inong> reasonable agreementbetween modelled ong>andong> measured uptake rates.Effective dry deposition velocities for soils for field conditions can be obtaong>inong>ed byusong>inong>g the model developed for the laboratory experiments ong>andong> then takong>inong>g ong>inong>toaccount the occurrence of ong>precipitationong>. The uptake rate by dry deposition ong>inong> thefield also decreases also with time due to saturation, just as ong>inong> the laboratory.Raong>inong>fall will wash down compounds that are for a large part present dissolved ong>inong>soil water. The surface will then become less saturated ong>andong> can then agaong>inong> take upgas at a higher rate. In some cases also re-emission will occur. A more complicatedof this type was presented by van Jaarsveld (1996). Van Pul et al. (1998) give alsosome useful parameterizations takong>inong>g the degradation rate ong>inong> the surface ong>inong>toaccount.Big leaf model forparticlesParticles of many different sizes are found ong>inong> the atmosphere. Most particlesare so small that gravitational settlong>inong>g is negligible compared to the verticaltransport caused by turbulent diffusion ong>inong> the atmosphere. As a result, mostparticles are transported ong>andong> diffused ong>inong> the atmosphere ong>inong> the same fashion asgases. Once particles are deposited, they are usually not re-emitted to theatmosphere agaong>inong>. As a result, the surface resistance can usually be omitted ong>inong>exchange models for particles, because it is zero. The dry deposition velocity is ong>inong>that case given by:F= v c = −r1+ rp d p, ong>airong>pong>airong> ,a bc(18)Particles have some properties that cause their lamong>inong>ar boundary layer resistance tobe much higher than for gases, because their transport through the lamong>inong>ar layer is69


much slower. Only very small particles with radius less than 0.1 µm, which do notcontribute much to the total atmospheric aerosol mass, have reasonable Browniong>andong>iffusivities ong>andong> are transported with a reasonable speed through the lamong>inong>arboundary layer. For particles with a radius larger than about 1 µm, transportthrough the lamong>inong>ar boundary layer is more efficient because then impaction ong>andong>ong>inong>terception at the surface are also important. But a large fraction of the particleshas a radius between 0.1 ong>andong> 1 µm ong>andong> are only transported very slowly throughthe lamong>inong>ar boundary layer. The velocity at which particles are transported throughthe lamong>inong>ar boundary layer varies highly with their size an as a result the drydeposition velocity for particles varies also highly with their size. For particles witha radius > 5 µm gravitational settlong>inong>g is important ong>andong> should be taken ong>inong>to account.As this mechanism is active at the same time as the other atmospheric transportmechanisms the big leaf model can be modified by addong>inong>g an additionalgravitational settlong>inong>g resistance r vg to transport parallel to the other resistances(Figure 9).Figure 9“Big leaf” model for dry deposition of particles over long>andong>.“Big leaf” model for tørdeposition af partikler over long>andong>.An important fraction of the atmospheric particles has a radius ong>inong> between 0.1 ong>andong>1 µm. For these particles the lamong>inong>ar boundary layer resistance r b is usually islarger than r a ong>andong> domong>inong>ates the overall resistance to transport from the atmosphereto the surface. As a result the dry deposition velocity of particles is relatively low,because they do not cross the lamong>inong>ar boundary layer fast (Ruijgrok et al., 1995).For particles of a particular size the dry deposition velocity ong>inong>creases wong>inong>d speed(u * ), maong>inong>ly because the thickness of the lamong>inong>ar boundary layer is reduced. Thedry deposition velocity for particles is also highly dependent on the properties ofthe surface. The knowledge on particle dry deposition is still ong>inong>sufficient.Primitive model for dry For particles with a radius between 0.1 ong>andong> 1 µm the estimated dry deposidepositionof particles to tion velocity for neutral atmospheric conditions is (Erisman et. al, 1994):vegetationvd = 0002 . u* (for low vegetation) (20)or70


vd = 001 . u* (for forests) (21)These equations show that the dry deposition velocity ong>inong>creases with the wong>inong>dspeed.Dry deposition ofparticles to the seaThe dry deposition velocity of particles to the sea is generally thought to beless than the dry deposition velocity of particles to long>andong> surfaces, because the sea isrelatively smooth. The deposition velocity is also ong>inong>fluence by the fact that theparticles are mostly hygroscopic, take up water ong>andong> consequently grow if theycome very near the water surface. Figure 10 shows how the dry deposition velocityof particles varies with particle size for various wong>inong>d speeds (Slong>inong>n ong>andong> Slong>inong>n,1980).Figure 10Dry deposition velocity of particles for sea areas as a function of size for differentwong>inong>d speeds.Tørdepositionshastighed for partikler for havområder som funktion af størrelse ogved forskellige vong>inong>dhastigheder.Atmospheric lifetime of As a result of the relatively low dry deposition velocity, particles have aparticlesrelative long atmospheric residence time (∼ 5 days). They are only removed wellwhen they meet ong>precipitationong> on their way. Particles can also coagulate ong>andong> as aresult larger particles are formed that are removed at another speed by dry ong>andong> wetdeposition than the particles they were formed of. Moreover, the pesticides boundong>inong> particles can volatilise if the particles are exposed to higher temperatures or tolower concentrations of pesticides ong>inong> the gas phase.Field measurements ofthe dry depositionvelocity for pesticidesThere almost no field measurements of the dry deposition velocity of gaseous of particulate pesticides. Some estimates have been made of the drydeposition velocity by usong>inong>g measured concentrations ong>inong> the ong>airong> ong>andong> exchangemodels. For particles it is generally very difficult to measure the dry depositionvelocity because it is so low. It has usually to be derived from very small71


concentration differences ong>andong> the measurements are for that reason very uncertaong>inong>.The same holds for the classical non-polar gaseous pesticides. But for the moresoluble gases it would perhaps be possible to measure the dry deposition velocity.The dry deposition velocity of gaseous pesticides is not only determong>inong>ed byproperties of the compound, the properties of the surface ong>andong> the meteorologicalconditions, but also by the concentration ong>inong> the surface. As a result the drydeposition velocity measured under the same meteorological ong>andong> surfaceconditions at one site is not necessarily the same as the dry deposition velocitymeasured somewhere else, because the concentration ong>inong> the surface can bedifferent. For that reason it is also necessary to specify the surface concentrationwhen a dry deposition velocity of a gaseous pesticide is reported. Only for highlysoluble gases it can presumably be assumed that the concentration ong>inong> the surface isso low that it has no ong>inong>fluence on the flux ong>andong> it is for that reason not necessary toreport it. But most gaseous pesticides are not highly soluble.Even ong>inong> the case of highly soluble gases it is not likely that the dry deposition rateis measured contong>inong>uously, because these measurements are extremely expensive.For such gases the dry deposition rate is usually estimated from a dry depositionvelocity measured durong>inong>g a limited number of campaigns for different atmosphericconditions ong>andong> contong>inong>uous meteorological measurements.Change ong>inong> concentration If the compound is distributed homogeneously over the whole mixong>inong>g layerong>inong> ong>airong> due to dry the change ong>inong> concentration c(t) due to dry deposition as a function of time tdeposition(s) is given by:vdt−zmixct () = c()0 e(22)where v d is the dry deposition velocity (m s -1 ) ong>andong> z mix the mixong>inong>g height (m). Closea source the compound will not yet be mixed over the whole mixong>inong>g layer. Thismeans that ong>inong> the case of a low source the concentration near the surface can bevery high ong>andong> the removal rate close to the source can for that reason be muchhigher. In that case the change ong>inong> concentration due to dry deposition will be muchgreater than that one derived from (22).1.5 Intermezzo: Maong>inong> conclusions on meteorology ong>andong> surface exchangeIn this section the ong>inong>formation ong>inong> the previous sections on meteorology ong>andong> surfaceexchange is summarised.Neutral atmosphereMechanical turbulence is generated due to friction exerted on the wong>inong>d by thesurface. This friction is caused by the roughness of the surface. A measure of thefriction is the friction velocity. The friction velocity depends on the wong>inong>d speed atgreater height ong>andong> the surface roughness. The friction velocity ong>inong>creases with thesurface roughness. Under neutral atmospheric conditions the turbulence is entirelymechanical. In northwestern Europe the atmosphere is most frequently neutral. Thefriction velocity is a very important parameter. If ong>inong>fluences the followong>inong>g:• Wong>inong>d speed. The wong>inong>d speed ong>inong>creases with the friction velocity at oneparticular height (6). The wong>inong>d speed itself ong>inong>creases with height.• Atmospheric diffusion. The atmospheric diffusion ong>inong>creases with the frictionvelocity (12) ong>andong> therefore also with the wong>inong>d speed. The atmosphericdiffusion ong>inong>creases with height ong>inong> the lower part of the atmosphere.• Surface exchange. The surface exchange ong>inong>creases also with the frictionvelocity (13,14,16,17) ong>andong> therefore also with the wong>inong>d speed. It does not onlydepend on the friction velocity, but also on processes ong>inong> the surface,characterised by the surface resistance (see section 1.4). If the surface72


esistance is low, the exchange is maong>inong>ly governed by atmospheric turbulence,characterised by the friction velocity. In this case the surface exchange willshow a relatively large ong>inong>crease with wong>inong>d speed. If the surface resistance ishigh, the surface exchange will be maong>inong>ly domong>inong>ated by processes ong>inong> thesurface an will only slow a relatively small ong>inong>crease with wong>inong>d speed. If the netflux is from the atmosphere to the surface the surface exchange is also calleddry deposition. If the net flux is from the surface to the surface the surfaceexchange is also called emission.• Mixong>inong>g height. The mixong>inong>g height ong>inong>creases with the friction velocity (8).Non-neutral atmosphere If the atmosphere is stable or unstable thermal turbulence plays also an importantrole. The wong>inong>d speed, atmospheric diffusion, surface exchange ong>andong> to some extentthe mixong>inong>g height will ong>inong> this situations also ong>inong>crease with the friction velocity ong>andong>therefore also with the wong>inong>d speed. The functions used to describe the ong>inong>crease are,however, somewhat modified. Under these conditions the heat ong>andong> moisture flux atthe surface play also a role. These heat ong>andong> moisture fluxes also ong>inong>fluence theemission of pesticides from the soil (see section 1.6). For this reason it is alsoimportant to describe these fluxes ong>inong> the same fashion ong>inong> the meteorological part ofa pesticide surface exchange model as ong>inong> the soil part of such a model.Surface gas exchange: The rate of exchange between the atmosphere ong>andong> the surface depends onboth directions possible the friction velocity ong>andong> on processes that occur ong>inong> the surface. The direction of theflux depends on the difference of the concentration ong>inong> the ong>airong> ong>andong> the gas phaseconcentration that would be ong>inong> equilibrium with the concentration ong>inong> the surface. Ifthe concentration ong>inong> the ong>airong> is higher than the equilibrium gas concentration ong>inong> thesurface, the net flux will be from the atmosphere to the surface. Is the equilibriumconcentration ong>inong> the surface higher than the concentration ong>inong> the ong>airong>, net emissionfrom the surface will occur. This is the case on a field after application ofpesticides.Highly soluble volatile pesticides can also be emitted from a field, but when theyare transported to areas where no emission has taken place the surfaceconcentration will so low that the surface can be considered as a perfect song>inong>k forthese gaseous pesticides i.e. the concentration ong>inong> the surface is rather low ong>andong> theflux is always from the atmosphere to the surface.For slightly soluble pesticides the surface is not a perfect song>inong>k. The concentration ong>inong>the surface can be relatively high. Their net deposition flux will be lower for thatreason. As a result they have a lower dry deposition velocity than highly solublegases. Their concentration ong>inong> the soil can even become so high that no netdeposition occurs, but net emission. This means that they can be re-emitted agaong>inong>after deposition. Several cycles of dry deposition ong>andong> re-emission result ong>inong> atransport over considerable distances. For these pesticides is it extremely importantto describe the processes ong>inong> the surface ong>inong> the same way as for emission (section1.6). For slightly soluble pesticides it is possible to conduct laboratorymeasurements that give ong>inong>formation on uptake by dry deposition.Dry deposition ofparticlesParticles are usually not re-emitted agaong>inong> after they have been dry deposited.The dry deposition velocity of particles depends on the friction velocity, size ong>andong>properties of the particles ong>andong> properties of the surface (presence of “hong>airong>s” onplants etc. It is usually much lower than the dry deposition velocity for highlysoluble gases. The dry deposition velocity of particles is very different for particlesof different size.Measurements of the dry It is ong>inong> prong>inong>ciple very difficult to measure the dry deposition rate (= flux) at73


deposition velocityall, both for gaseous as well as particulate pesticides. The only exception ispresumably for highly soluble gases. In that case it is also not likely that it ismonitored contong>inong>uously, but that it is estimated from a limited number ofmeasurements ong>andong> contong>inong>uous meteorological measurements.Potential for long-range The average dry deposition velocity cannot become higher for a certaong>inong>transportcombong>inong>ation of climate ong>andong> surface roughness than a certaong>inong> maximum value. Thismeans for Danish conditions that after 1 hour with dry deposition 13% of thecompound is removed from the ong>airong> ong>inong> the case of croplong>andong> ong>andong> 25% ong>inong> the case offorests. This ong>inong>dicates that there exists a general potential for long-range transportfor all released compounds ong>inong> the atmosphere as long as there is no ong>precipitationong>. Inthe case of a low source the concentration near the surface can be very high ong>andong> theremoval rate close to the source can for that reason be much higher until thecompound is fully mixed over the whole mixong>inong>g layer. But even ong>inong> this case there isa certaong>inong> limit to the removal rate by dry deposition ong>andong> a substantial part of thereleased compound will also be transported over long distances (see section 1.12).1.6 EmissionThe purpose of this report is only to describe the atmospheric processes related topesticides. As we have seen ong>inong> section 1.4, surface exchange (emission, drydeposition) of pesticides is also ong>inong>fluenced by processes ong>inong> or on the surface (soil,crops), which fall beyond the scope of this report. It was, nevertheless, decided togive some general ong>inong>formation here on emission of pesticides after application tothe soil ong>andong> crops that ong>inong>clude some processes ong>inong> the soil ong>andong> on the crops.1.6.1 Emission of pesticides from fallow soilAfter application of pesticides volatilisation from the soil starts. In the ong>inong>itial phaseit is rather important to know the ong>inong>itial penetration depth of the pesticide as thisdetermong>inong>es the gaseous pesticides concentration ong>inong> the upper layer of the soil,which drives the volatilisation (16). About this ong>inong>itial penetration depth not much isknown. Moreover, if the soil is rather dry, drops with dissolved pesticide can justlay on the surface ong>inong> stead of beong>inong>g adsorbed ong>inong> the upper soil layer. Due to theuncertaong>inong>ty ong>inong> these ong>inong>itial processes, it is difficult to model the emission ofpesticides ong>inong> this ong>inong>itial phase of the emission.To model the emission of pesticides from fallow soil requires a description of allsoil ong>andong> atmospheric processes ong>inong>volved. In section 1.4 the exchange between thesoil ong>andong> the atmosphere was already discussed. So we will focus here on thedescription of the processes ong>inong> the soil that are of importance.ong>Pesticidesong> are transported ong>inong> the soil together with the solvent they are dis-solved ong>inong>. Moreover, transport occurs by diffusion ong>inong> the soil water ong>andong> ong>inong> the gasphase ong>inong> the soil. It should be noted that although the concentration of the pesticideong>inong> the gas phase ong>inong> the soil is small, the transport by diffusion ong>inong> the gas phase cannevertheless be as important as the diffusive transport ong>inong> the water phase becausethe diffusivity of the pesticide ong>inong> the gas phase is several thousong>andong> times larger thanthat ong>inong> the water phase. The transport of pesticides ong>inong> the soil is reduced by the factthat pesticides can be adsorbed onto soil particles. ong>Pesticidesong> compete with waterfor the adsorption onto soil particles. If more water is present a substantial fractionof non-polar pesticides can desorb from the soil. For that reason it is important toknow the soil water content ong>andong> the water transport that determong>inong>es this content.The vapour pressure over an aqueous solution of pesticides is highly temperaturedependent (see section 1.2.3). Due to the fact that the transport ong>andong> adsorption ofpesticides is ong>inong>fluenced by water ong>andong> by the temperature, volatilisation of pesticidesTransport of pesticidesong>inong> the soil74


can only be modelled if the water transport (ong>inong>cludong>inong>g ong>precipitationong> ong>andong>evaporation) ong>andong> the heat transport ong>inong> the soil is modelled. When the heat transportis modelled, the temperature of the soil is known, which is not only essential tomodellong>inong>g of the volatilisation of the pesticide, but also to modellong>inong>g of theevaporation of water from the soil. As a result a model for volatilisation of apesticide contaong>inong> a description of the followong>inong>g processes:a) Transport of the pesticide dissolved ong>inong> soil water ong>inong>cludong>inong>g diffusion ong>inong> thewater phase.b) Transport of the pesticide ong>inong> gas phase ong>inong> the soil by diffusion.c) Adsorption of the pesticide to the soil particles.d) Transformation of the pesticide ong>inong> the soil.e) Equilibrium between the pesticide ong>inong> the gas phase ong>andong> the water phase(Henry’s law coefficient).f) Transport of water ong>inong> the soil.g) Transport of heat ong>inong> the soil.h) Transport of pesticide from the soil across the lamong>inong>ar boundary layer ong>inong>to theatmosphere.i) Transport of water (water vapour, raong>inong>) across the lamong>inong>ar boundary layerto/from the atmosphere.Jury et al. (1983, 1984) were among the first to model the volatilisation ofpesticides from the soil. Jury et al. (1984) ong>andong> Spencer et al. (1988) distong>inong>guishbetween three categories of pesticides with regard to volatilisation from fallow soil:highly volatile pesticides, slightly volatile pesticides ong>andong> moderately volatilepesticides. To which category a pesticide belongs depends on the Henry’s lawcoefficient K H .Highly volatile pesticides The volatilisation for this category of pesticides (K H > 2.65×10 -5 ) is controlledby diffusion ong>inong> the gas phase ong>inong> the soil. The volatilisation decreases with time,ong>inong>dependent of the evaporation of water vapour. The pesticide is not concentratedat the soil surface.The volatilisation of this category of pesticides (K H ≈ 2.65×10 -5 ) is con-trolled by diffusion ong>inong> the gas phase ong>inong> the soil if the soil humidity is low ong>andong> bydiffusion through the lamong>inong>ar boundary layer ong>inong> the atmosphere just above the soilsurface.Slightly volatilepesticidesModerately volatilepesticidesEmission of freshlyapplied pesticides fromsoil surfacesThe volatilisation of this category of pesticides (K H < 2.65×10 -5 ) is controlledby the rate at which the pesticide can diffuse through the lamong>inong>ar boundary layer ong>inong>the atmosphere just above the soil surface. Evaporation of water causesconcentration of the pesticide at the soil surface, because the pesticides are lessvolatile than water. The volatilisation can ong>inong>crease with time if evaporation ofwater contong>inong>ues. It can be expected that the emission of this category of pesticidesdepends rather much on the meteorological conditions, because they determong>inong>e thethickness of the lamong>inong>ar boundary layer.Woodrow et al. (1997) were ong>inong>terested ong>inong> the worst case fluxes of pesticides,that often occur withong>inong> a few hours after application. They were especiallyong>inong>terested ong>inong> these high values because they can be used to estimate the maximumconcentrations ong>andong> ong>effectsong>. They used published studies to derive a relationbetween the flux ong>andong> physical ong>andong> chemical properties of the compounds. Theyused compounds with vapour pressures that differed up to 10 7 . They found thefollowong>inong>g relation (Figure 11):75


⎛ VPln ( flux)= 28. 355+16158 . ln⎜⎝ K Socw⎞⎟⎠(23)where:flux = flux (µg m -2 hr -1 )VP= vapour pressure (Pa)K oc = soil adsorption coefficient (ml g -1 )S w = water solubility (mg l -1 )It should be noted here, that VP, S w ong>andong> Henry’s law coefficient K H are related (seesection 1.2.3)1. Beacon oil 6. Fonofos 11. Atrazong>inong>e2. Chevron oil 7. Long>inong>dane 12. p,p´-DDT3. Eptam 8. Dieldrong>inong> 13. Dacthal4. PCNB 9. Chlorpyrifos 14. Dacthal5. Trifluralong>inong> 10. Diazong>inong>on 15. PrometonFigure 11Correlation of pesticide flux from soil with chemical properties. Reprong>inong>ted withpermission from Woodrow et al., (1997). Copyright 1997 American ChemicalSociety.Korrelation af pesticidfluxen fra jord med kemiske egenskaber.For most experiments data were available on the application rate of the compound.So they tried to correlate the measured flux for those experiments with theproperties of the compounds, but now also takong>inong>g the application rate ong>inong>to accountbecause it is logical to expect that the flux ong>inong>creases with the application rate. Theyfound the followong>inong>g relation:⎛ VP × AR⎞ln ( flux)= 19. 35 + 10533 . ln⎜⎟⎝ K S ⎠ocw(24)76


where AR is the application rate (kg ha -1 )It should be noted here that the uncertaong>inong>ty ong>inong> the flux is still considerable(presumably a factor 2-3). They did not take any possible relationship withmeteorological conditions (temperature, wong>inong>d speed) ong>inong>to account as they wouldexpect that these conditions were similar for the different studies used. This doesnot mean that meteorological conditions do not play an important role, it illustratesjust that the flux is highly correlated e.g. the vapour pressure, because thecompounds have such an extreme wide range of vapour pressures. Nevertheless,this approach can be useful to get a first estimate of a possible maximum flux.Cumulative emission ofpesticides from the soilSmit et al. (1997) were ong>inong>terested ong>inong> the cumulative emission from the soilafter most of the pesticide had been volatilised ong>andong> not ong>inong> the worst case situationas Woodrow et al. (1997). They used also published experimental results tocorrelate the emission with physical ong>andong> chemical properties. They found thefollowong>inong>g relations:10( )CV = 71.9 + 11.6 log 100FPgasFor moistsoil. (25)CV = 42.3 + 9.010( )log 100FPgasFor dry soil.where:CV = cumulative volatilisation (% of dosage active ong>inong>gredient).FP gas = faction of the pesticide ong>inong> the gas phase ong>inong> the soil.FP gas can be found from the concentration of pesticide ong>inong> the soil, the volumefraction of gas ong>inong> the soil, the volume fraction of moisture ong>inong> the soil ong>andong> the drybulk density of the soil (see Smit et al., 1997 for further details).1.6.2 Emission of pesticides from plantsThis is a more complicated case. One reason is that the pesticide is not onlydeposited onto the plants, but also onto the underlyong>inong>g soil. Other reasons are thatdifferent plants have different surfaces (wax layer, hong>airong>s, structure), that there areleaves on various heights of the plants ong>andong> that they actively can take up thepesticide. Moreover, pesticides can be washed off by raong>inong>. As a result there are noreal mechanistic models to calculate the emission of pesticides from plants as doexist for emission from soils.Emission of freshly Woodrow et al. (1997) studied also the maximum volatilisation of pesticidesapplied pesticides from from ong>inong>ert surfaces (glass, plastic) ong>andong> from plant surfaces. It appeared thatplantsthe flux from plant surfaces durong>inong>g this ong>inong>itial phase of volatilisation correlated ong>inong>the same fashion with the vapour pressure as the fluxes from the ong>inong>ert surfaces. Itwas therefore concluded that plants were non-ong>inong>teractive, at least durong>inong>g the ong>inong>itialphase of volatilisation. This meant that they also could use the experimental data ofvolatilisation from ong>inong>ert surfaces to estimate volatilisation from plants, so that agreater data set could be used ong>inong> the correlation studies. Figure 12 shows thecorrelation found:( flux) = 11779 + 0 85543 ( VP)ln . . ln(27)It should be noted here, just as ong>inong> the case of volatilisation from the soil, that theyfong>inong>d such a good correlation because they use compounds with a very wide range77


of vapour pressure. It does not mean that meteorological conditions do not play arole, but that the variation ong>inong> the flux due to variations ong>inong> meteorological conditionsis much less than the variation ong>inong> the flux due to variations ong>inong> the vapour pressures.1. Beacon oil 7. Pendimethalong>inong>2. Chevron oil 8. 2,4-D (iso-octyl)3. Dodecane 9. Diazong>inong>on4. n-Octanol 10. Dieldrong>inong>5. Tridiphane 11. Toxaphene6. Trifluralong>inong> 12. p,p’-DDTFigure 12Correlation of pesticide flux from ong>inong>ert surfaces (plants, glass, plastic) with vapourpressure. Reprong>inong>ted with permission from Woodrow et al., (1997). Copyright 1997American Chemical Society.Korrelation af pesticidfluxen fra ong>inong>erte overflader (planter, glas, plastic) meddamptrykket.Cumulative emission of Smit et al. (1998) studied the cumulative volatilisation of pesticides at 7pesticides from plants days after application. They found the followong>inong>g relation:10log10( CV) = 1.528 + 0.466 log( VP)For dry soil. (28)In this equation is VP the vapour pressure ong>inong> mPa.They also found that the volatilisation was only to a mong>inong>or extent ong>inong>fluence bysorption processes ong>inong> ong>andong> on the leaves that are commonly represented by the K ow .The volatilisation was also not very well correlated with the K H , which ong>inong>dicatesthat the water ong>inong> the pesticides drops evaporates relatively quickly ong>andong> that the purecompound is left on the plant surfaces.78


1.6.3 Improvement ong>inong> the modellong>inong>g of the emission of pesticidesAlthough it could be satisfactory for many purposes to estimate the ong>inong>itialvolatilisation rate or the cumulative volatilisation by the statistical relations foundby Woodrow et al. (1997), Smit et al. (1997) ong>andong> Smit et al. (1998), mechanisticmodels are needed to better understong>andong> the underlyong>inong>g processes related tovolatilisation. Only ong>inong> this way it would be possible to get models to study theong>effectsong> of a strategies to reduce the evaporation of pesticides, to study the ong>inong>fluenceof meteorological conditions on the volatilisation ong>andong> to generalise the results.Measurements are needed for compounds with a wide range of properties to testthese mechanistic models, because the number of experiments currently availableis limited.1.7 Removal of material by ong>precipitationong>Airborne material can be removed from the atmosphere by ong>precipitationong> processes.This is called "ong>precipitationong> scavengong>inong>g”. It is an overall term, which covers theresult of many different processes. Some of these processes will be discussed here.1.7.1 Cloud physical processesCloud formationThe atmosphere contaong>inong>s aerosol particles, each of which consists of a variety ofcompounds, that can vary with particle size. Most aerosol particles contaong>inong>hygroscopic substances. When the relative humidity is more than 40%, aerosolscontaong>inong> at least 30% water by weight. When the relative humidity ong>inong>creases, morewater vapour condenses onto the aerosols ong>andong> cloud droplets are formed. In thisway aerosols act as “condensation nuclei” ong>andong> as a result the compoundsorigong>inong>atong>inong>g from the aerosols can be found ong>inong> the cloud droplets. Cloud droplets areso small that they have small fall velocities compared to the vertical wong>inong>d speed ong>inong>the cloud. As a result many of them remaong>inong> ong>airong>borne. If they come outside a cloud,where the relative humidity is less than 100% they will evaporate ong>inong> a few seconds.Cloud droplets can also take up gases. Table 2 gives some characteristic propertiesassociated with different cloud types (Cotton ong>andong> Anthes, 1989). It should be notedthat the values can vary much withong>inong> a cloud type ong>andong> that there are not onlydroplet of one size ong>inong> a cloud, but droplets of a whole range of sizes.Table 2Some characteristic properties of different cloud types.Nogle karakteristiske egenskaber for forskellige skytyper.Cloud typeVertical wong>inong>d speed ong>inong>the cloud(m s -1 )Liquid water content ofthe cloud(10 -6 m 3 water/m 3 cloud)Fog 0.01 0.2Stratus/stratocumulus 0.1 0.05-0.25Cumulus3 0.3(humilis/mediocris)Cumulus congestus 10 0.5-2.5Cumulonimbus 30 1.5-4.5Clouds as reactorsMost clouds will never give any ong>precipitationong>. Their droplets will evaporate ong>andong> thecompounds present ong>inong> the drops will remaong>inong> as aerosols ong>andong> gases. It is estimatedthat this cycle is repeated on the average about ten times before the content of thedrops reaches the earth’s surface ong>inong> the form of ong>precipitationong>. Compounds presentong>inong> the cloud droplets can react with each other. In that sense are clouds chemicalreactors.79


Precipitation formation Cloud droplets are very small, they have radii of 1-100 µm. Raong>inong>drops have a radiiof 100-2500 µm. This means that about a million cloud droplets need to fong>inong>d eachother to form one raong>inong>drop. Precipitation can already be formed relatively shortly(0.3-1 hours) after the first cloud has been formed. So there must be a veryeffective process leadong>inong>g to ong>precipitationong> formation.Warm cloudsCloud droplets of different sizes, have different ong>inong>ertia ong>andong> may collide due toturbulence ong>inong> the cloud to form a larger droplet. In the tropics this process can leadto ong>precipitationong>, because the number of droplets is larger ong>inong> these areas is largerthan ong>inong> the midlatitudes. This process is, however, not so effective that it canexplaong>inong> formation of ong>precipitationong> ong>inong> the midlatitudes. It can maybe lead to somedrizzle, but not to significant amounts of ong>precipitationong>.Cold cloudsIn-cloud scavengong>inong>gUptake by snowflakesong>andong> sleetMost clouds that give ong>precipitationong> ong>inong> the midlatitudes extend to heights where thetemperature is below 0°C. At these heights water is “undercooled”, i.e. it is stillliquid ong>andong> not frozen. This occurs at temperatures > - 15 °C. It is apparentlydifficult to start formation of ice crystals ong>inong> the very clean water ong>inong> the clouds.Some types of aerosols consistong>inong>g of unsoluble material (soil particles or particlesthat origong>inong>ate from plants) can act as “ice nucleii”, i.e. that the formation of icecrystals starts on this material. This because they contaong>inong> molecules or crystals witha similar structure as ice crystals. At the same temperature the water vapourpressure over ice is lower than the water vapour pressure over liquid water. As aresult the water from the undercooled cloud droplets will evaporate ong>andong> bedeposited on the ice crystals. In this way ice crystals will become so large that theyfall fast enough to pick up other crystals ong>andong>/or undercooled cloud droplets whichwill then freeze. In this way snowflakes are formed which are transformed ong>inong>toraong>inong>drops if they durong>inong>g their fall pass through a part of the atmosphere with atemperature above 0°C. This chaong>inong> of processes is much more effective thancollision of cloud droplets ong>andong> can explaong>inong> the formation of ong>precipitationong> withong>inong> arelative short time. As a result almost all raong>inong> ong>inong> the midlatitudes has once beensnow.The ong>precipitationong> formation process leads to the removal of the compoundsdissolved ong>inong> the cloud water. This is the way compounds associated withcondensation nucleii ong>andong> gases dissolved ong>inong> cloud ong>andong> raong>inong>drops are removed fromthe part of the atmosphere where clouds are. Removal from this part of theatmosphere is called ong>inong>-cloud scavengong>inong>g.Not much is known about removal by snowflakes ong>andong> sleet under the cloud.Snowflakes ong>andong> sleet can maybe take up aerosols more efficiently than raong>inong>dropsbecause they have a large surface area combong>inong>ed with a relatively low fall velocity.Sleet consists of meltong>inong>g snowflakes which are covered by a layer of water ong>andong> cantake up gasses more efficiently than raong>inong>drops for the same reasons why it is amore efficient scavenger of aerosols. About 10% of the ong>precipitationong> ong>inong> Denmarkarrives ong>inong> the form of snow ong>andong> sleet at the earth’s surface, but a somewhat largerfraction of the ong>precipitationong> has been scavengong>inong>g the below-cloud part of theatmosphere partly ong>inong> the form of snowflakes ong>andong> sleet. Due to its relativeunimportance ong>andong> the fact that not much is known, uptake of ong>airong>borne material bysnowflakes ong>andong> sleet is not treated ong>inong> this report.80


Precipitation ong>inong>DenmarkFall velocity of cloudong>andong> raong>inong>dropsThe average annual amount of ong>precipitationong> ong>inong> Denmark is about 700 mmong>andong> varies from about 550 to 900 mm. The variation from year to year can be high.When a drop starts to fall it will take a few seconds to accelerate to theirtermong>inong>al velocity. At this velocity the friction forces exerted by the surroundong>inong>g areequals the gravitation forces (Pruppacher ong>andong> Klett, 1997). Figure 13 shows thetermong>inong>al velocity of water drops as a function of their size. (Raong>inong>)drops with aradius greater than about 3500 µm (3.5 mm) are unstable ong>andong> break up ong>inong>to smallerdroplets. Figure 13 shows that the termong>inong>al velocity is about 0.01 m s -1 for clouddroplets, about 1 m s -1 for the smallest raong>inong>drops ong>andong> about 9 m s -1 for the largestones. The termong>inong>al velocity ong>inong>fluences the residence time of raong>inong>drops under thecloud. It ong>inong>fluences the rate at which water evaporates from drops or the rate atwhich gases can be taken up. The smallest raong>inong>drops will evaporate before theyreach the ground ong>andong> will not give a contribution to ong>precipitationong>. The larger onesfall so fast that they will only be able take up a very small amount of gas beforethey reach the ground. Droplets of different sizes can unite durong>inong>g their fall,because they have different termong>inong>al velocities. In the discussions ong>inong> this report wewill not take this process ong>inong>to account because we will illustrate the prong>inong>ciples ofthe uptake of gases ong>andong> particles by drops, rather than to give a completedescription.Figure 13Termong>inong>al velocity of raong>inong>drops as a function of their radius.Regndråbens faldhastighed som funktion af radius.Size distribution ofraong>inong>dropsThere is a relation between the raong>inong>fall rate ong>andong> the size distribution ofraong>inong>drops. At higher raong>inong>fall rates the average radius of the drops is larger, aseverybody knows by experience. Figure 14 shows number of raong>inong>drops as afunction of their size for two common raong>inong>fall rates usong>inong>g an equation given byMarshall ong>andong> Palmer (1948). Figure 15 shows the distribution of the liquid watercontent of ong>airong> belongong>inong>g to these size distributions. Both Figure 14 ong>andong> Figure 15are given for an arbitrary raong>inong>drop radius ong>inong>terval. This radius ong>inong>terval is notimportant here. What is important to note is the relative difference between Figure14 ong>andong> 15. This difference reflects the fact that the smallest drops are mostabundant, but that their contribution to the liquid water content of the cloud is notlarge. The liquid water content of the cloud a maximum about a drop radius of 0.5-1 mm. Such raong>inong>drop distributions as show ong>inong> Figure 14 are more representative for81


an average situation. The size distributions durong>inong>g events may deviate from theequation given by Marshall ong>andong> Palmer. Moreover, different authors use differenttypes of equations, which give somewhat different results. Most distributions of theraong>inong>drop size have been measured to relate the radar reflection of ong>precipitationong>drops to measured size distributions at ground level. They give a good descriptionof the larger raong>inong>drops, which are reflect the radar waves most ong>andong> contribute mostto the amount of ong>precipitationong>, but the description of the smaller drops ong>inong> thedistribution, that are more important for the uptake of gases, is somewhat moreuncertaong>inong>. The resultong>inong>g uncertaong>inong>ty ong>inong> the removal rate of gases by raong>inong>drops willbe discussed later. It should be noted that the liquid water volume ofraong>inong>drops/snowflakes ong>inong> cloud is much less than the liquid water volume of cloudwater.Figure 14Number of raong>inong>drops per volume of ong>airong> as a function of their radius for a dropradius ong>inong>terval of 1.25×10 -5 m.Antallet af regndråber per luftvolumen som funktion af deres radius med etdråberadiusong>inong>terval på 1,25×10 -5 m.82


Figure 15Liquid water volume of raong>inong>drops per volume of ong>airong> as a function of their radius fora drop radius ong>inong>terval of 1.25×10 -5 m.Flydendevong>andong>ong>inong>dhold af regndråber per luftvolumen som funktion af deres radiusmed et dråberadiusong>inong>terval på 1,25×10 -5 m.Cloud base heightEstimated cloud baseheightDurong>inong>g unstable conditions (convection) an ong>airong> parcel which is unsaturated withwater vapour can from near the earth’s surface will be transported upward ong>andong> willbe cooled down with -0.01°C m -1 . As the saturation pressure (Pa) over liquid waterdecreases with temperature, the ong>airong> parcel will become more ong>andong> more saturatedwith height due to coolong>inong>g ong>andong> if the saturation pressure is reached, a substantialcondensation of water vapour on aerosols will occur ong>andong> a cloud is formed. Thecloud base is ong>inong> this case at the height, where the first cloud droplets form. Durong>inong>gthese unstable conditions, there is a relation between the relative humidity ong>andong> thetemperature at the ground ong>andong> the cloud base height. Durong>inong>g neutral ong>andong> stableconditions there is no such relation. The layer with the clouds may have a totallydifferent origong>inong> than the ong>airong> at ground level. This is especially the case near warmong>andong> cold fronts, where warm ong>andong> cold ong>airong> meet. It is important to know the cloudbase height, because ong>airong>borne material is removed by ong>precipitationong> by differentprocesses ong>inong> ong>andong> below the cloud ong>andong> what is even more important material isremoved at different rates.The cloud base height is either measured (usual at large ong>airong>ports) or estimatedby meteorological observers. Figure 16 shows the frequency distribution of thecloud base height durong>inong>g ong>precipitationong> ong>inong> Denmark for 5 stations (Asman ong>andong>Jensen, 1993). The cloud base height is given as a code, which refers to heightong>inong>tervals (Table 3). It should be noted here that the height ong>inong>tervals are not the samefor all codes. On the basis of the frequency distribution an average cloud baseheight was calculated for each of the five stations (Table 4). The uncertaong>inong>ty ong>inong>these values is rather large due to the procedure ong>andong> because the estimated cloudbase height depends also to some extent on the observer.Figure 16Frequency distribution of the cloud base height durong>inong>g ong>precipitationong> for 1979-1988. Data from the Danish Meteorological Institute (DMI).83


Station code: Kas = Kastrup, Kar = Karup, Skr = Skrydstrup, Bel = Beldrong>inong>ge, Aal= Aalborg. The cloud base height classes are explaong>inong>ed ong>inong> Table 3.Frekvensfordelong>inong>g af skybasehøjden under nedbørhændelser for 1979-1988. Datafra Danmarks Meteorologiske Institut (DMI).Stationskode: Kas = Kastrup, Kar = Karup, Skr = Skrydstrup, Bel = Beldrong>inong>ge,Aal = Aalborg. Skybasehøjdeklasserne er forklaret i Tabel 3.Table 3Explanation of the cloud base height code ong>andong> the height ong>inong>terval used ong>inong> Figure16.Forklarong>inong>g af koden for skybasehøjde og højdeong>inong>terval anvendt i Figur 16.CodeHeight ong>inong>terval (m)0 0-491 50-992 100-1993 200-2994 300-5995 600-9996 1000-14997 1500-19998 2000-24999 2500 or more84


Table 4Average cloud base height durong>inong>g ong>precipitationong> for five Danish meteorologicalstations.Gennemsnitlige skybasehøjde under nedbørsperioder for fem danskemeteorologiske stationer.StationCloud base height (m)Kastrup 365Karup 349Skrydstrup 373Beldrong>inong>ge 412Aalborg 4201.7.2 Exchange of gases between drops ong>andong> the ong>airong>The uptake of gases by cloud ong>andong> raong>inong> drops can be described by a sequence ofsteps that must be taken (Seong>inong>feld, 1986):1. Diffusion of the gas from some distance of the drop to the drop surface.2. Transfer of gas across the gas-water ong>inong>terface.3. Extremely fast reactions ong>inong> the water phase near the surface of the drop. Thisreaction should be so fast, that it is completed before the diffusong>inong>g gas has beenmixed throughout the whole drop. This is usually only the case for ionisationreactions. If this occurs, not the dissolved gas will diffuse ong>inong>to the water phase,but the reaction products4. Diffusion of the dissolved gas or extremely rapidly formed reaction products ong>inong>the water phase throughout the drop.5. Other, not extremely fast reactions ong>inong> the water phase.In this report we will only focus on uptake of gases where no reaction occurs, asthis seems to be more relevant for pesticides.After less than 0.1 seconds a steady state situation is achieved for cloud ong>andong>raong>inong>drops that are exposed to a gas. The steady state situation is defong>inong>ed as thesituation where the flux ong>inong> the gas phase from some distance to the drop equals theflux through the gas-water ong>inong>terface, which equals the flux ong>inong>to the water phase.The overall rate at which a gas can be taken up depends on the resistance totransport ong>inong> the gas phase, across the ong>inong>terface ong>andong> ong>inong> the water phase. Although itcannot be excluded that ong>inong> some cases the resistance ong>inong> the water phase isconsiderable, this seems not to be the case ong>inong> general. One of the reasons is thatdrops with a radius > 100 µm develop durong>inong>g their fall an ong>inong>ternal circulationleadong>inong>g to mixong>inong>g withong>inong> the drops ong>andong> as a result the resistance ong>inong> the water phaseis small compared to the resistance ong>inong> the water phase.Flux ong>inong> the gas phaseThe transport of the gas ong>inong> the gas phase ong>andong> transfer across the gas-water ong>inong>terfaceis given by (Pruppacher ong>andong> Klett, 1997):Fg*fgDg= ( cg−cwH)r(29)where:Fg =flux ong>inong> the gas phase to/from the droplet surface (kg m-2 s-1).85


D g*=f g =apparent diffusivity of the gas ong>inong> the gas phase (m 2 s -1 ). D * g dependson the diffusivity of the gas ong>inong> the gas phase, but ong>inong>cludes also a factorthat depends on the probability that a gas molecule that hits the watersurface is absorbed by the drop (see Appendix 1). If the probability islarge enough D * g will become approximately equal to D g , the realdiffusivity of the gas ong>inong> the gas phase (m 2 s -1 ). This factor is unknownfor pesticides ong>andong> for that reason D * g is set D g ong>inong> the calculations. Inthis way a maximum value is obtaong>inong>ed of D gventilation coefficient for gases (dimensionless). This is a correctionfactor. It is the ratio between the flux for a stagnant drop ong>andong> the fluxfor a drop that moves relative to the ong>airong> at its termong>inong>al velocity. Theexchange is greater for movong>inong>g drops, because they all the time willmeet “fresh”, i.e. ong>airong> that has not yet been depleted with gas.The ventilation coefficient is approximately given by:1fg = 1+03 . Re2Sc13(30)where Re is the Reynolds number: Re =2rv t /ν (dimensionless), where r = radius ofthe drop (m), v t = termong>inong>al velocity of the drop (m s -1 ) , ν = kong>inong>ematic viscosity ofthe ong>airong> (m 2 s -1 ) ong>andong> Sc is the Schmidt number (dimensionless). The Schmidt numberis ν/D g , where D g is the diffusivity of the gas ong>inong> the gas phase (m 2 s -1 ).The ventilation coefficient f g is 1 for small, almost not movong>inong>g cloud droplets to upto about 20-30 for the largest fallong>inong>g raong>inong>drop ong>andong> gases with molecular weightsvaryong>inong>g from 200-400. It depends maong>inong>ly the radius of the drop (which alsodetermong>inong>es the termong>inong>al velocity) ong>andong> the diffusivity of the gas.c gcwH= gas phase concentration ong>inong> the bulk phase, i.e. at some distancefrom the drop ong>inong>terface (kg m -3 ).= dissolved gas concentration ong>inong> the bulk water phase (kg m-3).= Henry’s law coefficient (kg m-3 gas/kg m-3 water)F g is the flux, i.e. kg of the gas that passes through 1 m 2 of drop surface per second.It is more useful to have a change ong>inong> concentration per second ong>inong> the drop ong>inong> stead.This can be done by multiplyong>inong>g the flux with the surface of the drop (4πr 2 ) ong>andong>dividong>inong>g by the volume of the drop ((4/3)πr 3 ):dcdt3f D= −2rw g g( cgcwH)(31)It is ong>inong>terestong>inong>g to note that c w H is just equal to the gas phase concentration thatwould be ong>inong> equilibrium with the concentration ong>inong> the bulk water phase ong>inong> the drop.This equation is, ong>inong> fact, much alike the equation for the dry deposition flux (16),which also depends on a concentration difference.If we ong>inong>tegrate (31) ong>andong> assume that c w = 0 at t = 0, the followong>inong>g equation isobtaong>inong>ed:cwkt 1( 1 e )cg= − −H(32)where:86


k3fgDgH=r1 2=2rfDHτ abs3 g g(33)(34)The characteristic time constant of absorption is τ abs = 1/k 1 . If t = τ abs about 63% ofthe gas that can potentially be absorbed has been absorbed. The value of τ abs fordrops of different size can be calculated from Table A2-1 ong>inong> Appendix 2.For raong>inong>drops, which have a noticeable termong>inong>al velocity it is more convenient toexpress the concentration ong>inong> the drop as a function of the fall distance. Realisong>inong>gthat the fall distance (m) is z f = v t *t the followong>inong>g is found from (32):dcdzwf3fgDg= −2r vt( cgcwH)(35)If we ong>inong>tegrate (35) ong>andong> assume that c w = 0 at z f = 0, the followong>inong>g equation isobtaong>inong>ed:cwk2z( 1 e )c g= − −H(36)where:k3fDHg g=rv2 2t(37)∆ absrvt=3fDHg2g(38)The characteristic distance of absorption is ∆ abs = 1/k 2 . If the drop has fallen adistance z f = ∆ abs about 63% of the gas that can potentially be absorbed has beenabsorbed. The value of ∆ abs for drops of different size can be calculated from TableA2-1 ong>inong> Appendix 2.It can be seen from (36) that the concentration approaches the equilibriumconcentration c g /H (saturation) if t is relatively large. It can also be seen from k 1ong>andong> k 2 that if H is larger (the gas is more volatile, i.e. less soluble) it takes lesstime/fall distance to get saturation of the drop. The ventilation coefficient f g ong>inong> (33)ong>inong>creases with r, but much less than with r 2 , so k 1 decreases with the drop radius.This means, that it takes more time for larger drops to get saturated. The value of k 2decreases even faster with the drop radius, because here v t is ong>inong> the denomong>inong>atorong>andong> v t ong>inong>creases with the drop radius (see Figure 13).87


1.7.3 Scavengong>inong>g of gasesGeneral situationFigure 17 illustrates the differences between uptake of highly ong>andong> slightly solublegases ong>inong> clouds (ong>inong>-cloud scavengong>inong>g) ong>andong> below-clouds (below-cloud scavengong>inong>g).In each “comic” three situations are shown:1) The situation before formation of the cloud.2) Effect of ong>inong>-cloud scavengong>inong>g.3) Effect of below-cloud scavengong>inong>g.Figure 17Differences ong>inong> ong>inong>-cloud ong>andong> below-cloud scavengong>inong>g of highly (upper figures) ong>andong>slightly soluble gases (lower figures).Forskelle mellem ong>inong>-cloud scavengong>inong>g og below-cloud scavengong>inong>g af letopløselige(øverste figurer) og tungtopløselige gasser (nederste figurer).1. Before the formation Before the formation of the cloud the concentrations ong>inong> the ong>airong> are the sameof the cloud everywhere for both gases ong>inong> this example (situation 1).2. Uptake of gases ong>inong> the After formation of the cloud (situation 2) equilibrium is reached betweencloud water after concentrations ong>inong> the gas ong>andong> water phase ong>inong> the cloud droplets. This holdscloud formation for both the highly ong>andong> the slightly soluble gas. Most of the highly soluble gas canbe found ong>inong> the droplets (one droplet is shown) ong>andong> the concentration ong>inong> the ong>airong>withong>inong> the cloud is very low, i.e. much lower than the concentration ong>inong> the ong>airong>outside the cloud. It will take somewhat more time to reach equilibrium for highlysoluble gases than for slightly soluble gases, because a larger amount of gas has tobe transported ong>inong>to the droplets. But even then it will take less than a mong>inong>ute beforethese small cloud droplets have reached equilibrium. Their lifetime is much morethan a mong>inong>ute (see Table A2-1 ong>inong> Appendix 2), so equilibrium is reached rather fastcompared to their lifetime. The situation for the slightly soluble gas is somewhatdifferent. The concentration ong>inong> the cloud droplets is rather low ong>andong> as a result theconcentration ong>inong> the ong>airong> withong>inong> the cloud has not changed much due to the uptake ofthe gas ong>andong> is almost equal to the concentration of the gas ong>inong> the ong>airong> outside thecloud.88


3. Uptake of gases below After some time ong>precipitationong> is formed ong>andong> the raong>inong>drops that contaong>inong> aboutthe cloudthe same concentration as the cloud droplets, they will start to fall ong>andong> take up gason their way (situation 3). In the case of a highly soluble gas the drops are notsaturated at all compared to the surroundong>inong>g ong>airong> ong>andong> they will take up a additionalgas. Due to the relatively short time drops are under the cloud base (1-6 mong>inong>utes fora cloud base of 400 m; see Table A2-1 ong>inong> Appendix 2) the drops do not getsaturated as there is a physical limit as to how much gas can be taken up ong>inong> thisrelatively short time. Drops may get more ong>andong> more saturated after they have beencollected ong>inong> a ong>precipitationong> sampler if they still are ong>inong> contact with the surroundong>inong>gong>airong>. The situation for slightly soluble gases is quite different. The raong>inong>drops arealmost saturated when they leave the cloud ong>andong> if the concentration ong>inong> the ong>airong> underthe cloud is not much different from the concentration ong>inong> the ong>airong> ong>inong> the cloud, theywill almost be saturated when they reach the surface. So for those drops it isreasonable to adopt equilibrium. For both highly ong>andong> slightly soluble gases holdsthat if the concentration ong>inong> the ong>airong> is the same at ground-level ong>andong> higher up ong>inong> theatmosphere where clouds are formed, the contribution from ong>inong>-cloud scavengong>inong>g islarger than the contribution from below-cloud scavengong>inong>g. Near relatively strongsources, the concentration ong>inong> the ong>airong> below the cloud can become much higher thanthe background concentration ong>inong> the atmosphere. Only ong>inong> this situation thecontribution from below-cloud scavengong>inong>g can become larger than that from ong>inong>cloudscavengong>inong>g.The over-all effect of ong>inong>- ong>andong> below-cloud scavengong>inong>g can be described empiricallyby a an overall scavengong>inong>g ratio:Soverallc=cprg(39)where:c pr = concentration of the compound ong>inong> ong>precipitationong> (kg m -3 )c g = concentration of the compound ong>inong> the ong>airong> (kg m -3 )The philosophy behong>inong>d this empirical approach is that it is logical to adopt a long>inong>earrelation between the concentration ong>inong> ong>precipitationong> ong>andong> the concentration ong>inong> ong>airong>.S overall is usually found from measured concentrations ong>inong> ong>precipitationong> ong>andong> ong>airong> atground-level. As can be see from Figure 17 this is a reasonable assumption forslightly soluble gases if the concentration ong>inong> ong>airong> at ground-level is about the same asong>inong> ong>airong> at cloud height. In that case equilibrium can be assumed to exist between thegas ong>andong> the raong>inong> ong>andong> S overall can be found from the Henry’s law coefficient:Soverall= 1 KH(40)For highly soluble gases this approach is more tricky. Usually there is noequilibrium between the concentration ong>inong> the ong>airong> ong>andong> ong>precipitationong> at ground-level,or between the concentration ong>inong> the ong>airong> outside the cloud at cloud height ong>andong> theconcentration ong>inong> ong>precipitationong>. For highly soluble gases the S overall is often set toabout 1×10 6 , an upper limit, which depends more on the removal rate ofong>precipitationong> itself than on the Henry’s law coefficient of the gas. Sometimes S overallis also made slightly dependent on the raong>inong>fall rate. The rate at which a compoundis removed from the atmosphere by scavengong>inong>g is called scavengong>inong>g coefficient ong>andong>is defong>inong>ed by:89


Λ overallS=zoverallmixI(41)where:Λ overall = scavengong>inong>g coefficient (s -1 )I = raong>inong>fall rate (m s -1 )z mix = mixong>inong>g height (m)For highly soluble gases ong>andong> adoptong>inong>g a mixong>inong>g height of 1000 m this givesremoval rates of 2.8×10 -4 s -1 at 1 mm hr -1 to 2.8×10 -3 s -1 at 10 mm hr -1 . This meansthat at a ong>precipitationong> rate of 1 mm hr -1 64% of the gas is removed from the ong>airong> after1 hour with ong>precipitationong>. At a ong>precipitationong> rate of 10 mm hr -1 more than 99.9% isremoved from the ong>airong>. (This can be calculated with (43)).The situation close to important low-level sources is different form the gen-eral situation ong>inong> that the plume comong>inong>g from the field on which the pesticide hasbeen applied usually reaches the cloud base height at a distance between 2-10 kmfrom the source. In that case all scavengong>inong>g is below-cloud scavengong>inong>g which is notso effective for highly ong>andong> slightly soluble gases as ong>inong>-cloud scavengong>inong>g. Usuallyno equilibrium will be reached, neither for the highly nor for the slightly solublegas. There is a limitation to the uptake of a gas: no more gas can be taken up than istransported to the raong>inong>drop. This sets an upper limit to the below-cloud scavengong>inong>gwhich is reached for highly soluble gases. For slightly soluble gases the belowcloudscavengong>inong>g is less than this value. The upper limit for convective conditionsis given by (Asman, 1995):Scavengong>inong>g close toimportant sourcesΛ bbav= aI(42)In this equation a ong>andong> b depend on the relative humidity ong>andong> temperature at groundlevel ong>andong> the diffusivity of the gas at 25°C (see Appendix 3). In this case I is theraong>inong>fall rate ong>inong> mm hr -1 . If the diffusivity of the gas is not known it can be estimatedfrom (15). For non-convective conditions one could e.g. calculate a scavengong>inong>gcoefficient by assumong>inong>g a temperature of 10°C ong>andong> a relative humidity of 85%.Figure 18 shows this maximum below-cloud scavengong>inong>g coefficient of a gaseouspesticide with a molecular weight of 300 as a function of the raong>inong>fall rate. Due tothe uncertaong>inong>ty ong>inong> the drop size distribution the uncertaong>inong>ty ong>inong> the scavengong>inong>gcoefficient is at least a factor of two. For highly soluble gases this gives removalrates of 3.8×10 -5 s -1 at 1 mm hr -1 to 1.6×10 -4 s -1 at 10 mm hr -1 . After one hour withong>precipitationong> 13 ong>andong> 44% will be removed at a ong>precipitationong> rate of 1 ong>andong> 10 mm hr -1 respectively. (see (43)). These rates that are about a factor of 10 lower than for ong>inong>cloudscavengong>inong>g. The fact that the drops that reach the ground are not ong>inong>equilibrium with the concentration ong>inong> the ong>airong> for highly soluble gases can lead toartefacts durong>inong>g samplong>inong>g of ong>precipitationong>, because the collected drops on thefunnel or ong>inong> the bottle will still be able to take up gas after the ong>precipitationong> hasstopped.90


Figure 18Below-cloud scavengong>inong>g coefficient for a highly soluble gas with molecular weightof 300 as a function of the raong>inong>fall rate at ground level. At ground-level thetemperature is 10°C ong>andong> the relative humidity is 85%.Below-cloud scavengong>inong>gkoefficienten for letopløselig gas med en molekylemasse på300 som funktion af regnong>inong>tensiteten i jordniveau. Temperaturen og relativfugtigheden i jordniveau er henholdsvis 10°C og 85%.1.7.4 Scavengong>inong>g of particlesParticles contaong>inong>ong>inong>g pesticides will usually be removed rather efficiently by ong>inong>cloudscavengong>inong>g. Below-cloud scavengong>inong>g of particles is less efficient ong>andong>depends on the size distributions of the raong>inong>drops ong>andong> the particles (Slong>inong>n, 1983). Ascavengong>inong>g ratio of 1×10 6 can be adopted for an overall-scavengong>inong>g of particlesbecause scavengong>inong>g is ong>inong> general domong>inong>ated by ong>inong>-cloud scavengong>inong>g. Close tosources only below-cloud scavengong>inong>g is important, the scavengong>inong>g ratio will oftenbe considerably less ong>andong> has to be calculated with the appropriate size distributionsof raong>inong>drops ong>andong> particles. Raong>inong> contaong>inong>s also particles that are only partlydissolved. ong>Pesticidesong> with a low solubility may therefore not only exist ong>inong> raong>inong> as adissolved gas but also as be part of particles.Change ong>inong> concentration The concentration left ong>inong> the ong>airong> after a ong>precipitationong> period with length t (s)ong>inong> the ong>airong> due to wet can be found from:deposition−Λct ( ) = c( 0 ) e t(43)where Λ can be the ong>inong>-cloud scavengong>inong>g coefficient, below-cloud scavengong>inong>gcoefficient or the overall scavengong>inong>g coefficient.Measurements of thewet deposition rateThe wet deposition of pesticides can be measured. It is ong>inong> prong>inong>ciple alsopossible to derive an overall scavengong>inong>g coefficient for pesticides from measuredconcentrations ong>inong> ong>airong> ong>andong> ong>precipitationong> usong>inong>g (39) ong>andong> (41). In this case themeasured ong>airong> concentration should be representative for the whole ong>airong> column, i.e. itshould not be measured close to sources.91


1.8 Intermezzo: Maong>inong> conclusions on wet deposition ong>andong> a comparison with dry depositionIn-cloud scavengong>inong>gong>andong> properties ofthe compoundHighly soluble ong>andong> reactive gases are removed by ong>inong>-cloud scavengong>inong>g at ahigh rate. Almost all gas will be dissolved ong>inong> the cloud droplet ong>andong> theconcentration ong>inong> the ong>inong>terstitial ong>airong> ong>inong> the cloud will be low. For that reason thescavengong>inong>g rate does not depend on the Henry’s law coefficient, but on the raong>inong>fallrate, as almost all gas is dissolved ong>inong> the cloud drops.For slightly soluble gases the situation is different: a substantial fraction of thecompound is found ong>inong> the ong>inong>terstitial ong>airong> ong>inong> the cloud. The removal rate for thesegases will be a function of the Henry’s law coefficient ong>andong> the ong>precipitationong> rate.Most particles act as condensation nucleii ong>andong> will for that reason be removed at ahigh rate that is determong>inong>ed by the ong>precipitationong> rate. The removal rate will notdepends much on the size of the particles as long as they are hygroscopic.Below-cloud scavengong>inong>g Highly soluble ong>andong> reactive gases are removed by below-cloud scavengong>inong>gong>andong> properties of at a higher rate than slightly soluble gases, but the removal rate by belowthecompound cloud scavengong>inong>g is less than by ong>inong>-cloud scavengong>inong>g (see below). The removal ratefor highly soluble gases does ong>inong> general not depend on the Henry’s law coefficientbut on the speed at which the gas can diffuse ong>inong>to the drop, which is e.g. a functionof the diffusivity of the gas ong>andong> the raong>inong>drop size, which ong>inong> turn is a function of theong>precipitationong> rate. For slightly soluble gases the raong>inong>drop is almost saturated withthe gas by ong>inong>-cloud scavengong>inong>g when the drop reached the ong>airong> below the cloud. Ifthe ong>airong> below the cloud has the same concentration no more gas will be taken up. Ifthe ong>airong> below the cloud has a higher concentration more gas can be taken up. If theong>airong> below the cloud has a lower concentration that ong>inong> the cloud the drops will degas,i.e. dissolved gas will evaporate ong>inong>to the atmosphere.Particles will ong>inong> general also be removed by below-cloud scavengong>inong>g at a lower ratethan by ong>inong>-cloud scavengong>inong>g. Their removal rate will be a function of their size ong>andong>the size of the raong>inong>drops which is a function of the ong>precipitationong> rate.In-cloud scavengong>inong>g vs. In case of ong>inong>-cloud scavengong>inong>g at a ong>precipitationong> rate of 1 mm hr -1 64% of abelow-cloud scavengong>inong>g highly soluble gas is removed after 1 hour with ong>precipitationong>. At a ong>precipitationong> rateof 10 mm hr -1 more than 99.9% is removed from the ong>airong> after 1 hour.For highly soluble gases the removal rate by below-cloud scavengong>inong>g is highest.This gives removal rates of 3.8×10 -5 s -1 at 1 mm hr -1 to 1.6×10 -4 s -1 at 10 mm hr -1 .After one hour with ong>precipitationong> 13% ong>andong> 33% are removed from the ong>airong> at aong>precipitationong> rate of 1 ong>andong> 10 mm hr -1 respectively. In case only below-cloudscavengong>inong>g is occurrong>inong>g, e.g. close to a low level source, still a considerablefraction of the highly soluble gas remaong>inong>s ong>airong>borne. The removal rate for slightlysoluble gases will be less than for highly soluble gases.The wet removal rate for particles will also be domong>inong>ated by ong>inong>-cloud scavengong>inong>gong>andong> will be the same as for highly soluble gases, at least if the concentration atcloud level is about the same as below the cloud.Close to low sources, like fields after application of pesticides, the plumehas not yet reached the cloud ong>andong> for that reason only below-cloud scav-engong>inong>g occurs. As a result the wet deposition near a low source will only be arelatively small fraction of the amount of pesticide volatilised, even ong>inong> the case of ahighly soluble pesticide. It should be noted that ong>inong> general a backgroundconcentration of pesticides will be present ong>inong> the atmosphere. This backgroundconcentration will also exist at cloud level. As ong>inong>-cloud scavengong>inong>g is much moreefficient than below-cloud scavengong>inong>g, a larger fraction of the background will beOrigong>inong> of pesticide ong>inong>ong>precipitationong> close toan important source92


scavenged near the low source than of the released compound from the low source.Only ong>inong> the case where the low source emits at a relatively high rate or theong>precipitationong> period last for a long time, the concentration ong>inong> ong>precipitationong> can bedomong>inong>ated by a nearby low source.Non-soluble particlesWet deposition vs.dry depositionRaong>inong> contaong>inong>s also particles that are only partly dissolved. ong>Pesticidesong> with a lowsolubility may therefore not only exist ong>inong> raong>inong> as a dissolved gas but also as be partof particles.The dry deposition velocity of gases ong>inong>creases ong>inong> general with the solubilityof the gas. The same is the case for removal rates of gases by ong>inong>-cloud ong>andong>below-cloud scavengong>inong>g. For particles the situation is different. Particles with aradius between 0.1 ong>andong> 1 µm are not removed well from the atmosphere by drydeposition, but are removed well by ong>inong>-cloud scavengong>inong>g because they can act ascondensation nucleii. As it does not raong>inong> that often, their atmospheric lifetime israther long, of the order of 5 days or longer.The maximum removal rate by dry deposition is much lower than the maximumremoval rate by wet deposition. After 1 hour under Danish conditions only 13% ofthe a highly soluble gas is removed ong>inong> the case of croplong>andong> ong>andong> 25% ong>inong> the case offorests.The maximum removal rate by ong>precipitationong> is determong>inong>ed by ong>inong>-cloud scavengong>inong>g,if the concentration of the compound ong>inong> the ong>airong> at cloud level ong>andong> below the cloudare the same. It is then 64% hr -1 at a ong>precipitationong> rate of 1 mm hr -1 or more than99.9% hr -1 at a ong>precipitationong> rate of 10 mm hr -1 .So it looks as if wet deposition will be the domong>inong>atong>inong>g removal process. This is,however, not necessarily true. The reason for this is that dry deposition occurs allthe time, even durong>inong>g ong>precipitationong>, whereas wet deposition only occurs durong>inong>gong>precipitationong>, i.e. durong>inong>g 5-10% of the time. Consequently the average amount drydeposited over a longer period can be of the same order of magnitude as theamount wet deposited, despite the fact that the process is less efficient than drydeposition.It is very difficult to measure the dry deposition rate ong>andong> even ong>inong> the case it can bemeasured like for highly soluble gases, it is not likely that it will be donecontong>inong>uously. On the contrary it is possible to measure the wet deposition rate.1.9 Conversion from the gaseous to the particulate phaseAs we have seen ong>inong> section 1.4, dry deposition of particles proceeds at anotherspeed than dry deposition of gases. ong>Pesticidesong> released as gases can be converted toparticles ong>andong> as a result the pesticide will be removed from the atmosphere atanother rate than if it were ong>inong> the gas phase. This process can potentially beimportant for slightly soluble gases that are not well removed by dry deposition orong>precipitationong> scavengong>inong>g.It is therefore crucial to know if a pesticide exists ong>inong> the gas phase or ong>inong> theparticulate phase. Moreover, it is necessary to know at which rate a gaseouspesticide can be converted to its particulate form, so that we e.g. whether thegaseous pesticide is already converted to the particle phase to a substantial extentdurong>inong>g the transport over the first 2 km discussed ong>inong> this reportRatio particle-gas forIf gaseous pesticides are transformed to particles, they do not form pure93


pesticidespesticide particles. For thermodynamical reasons it is favourable for gases tocondense on existong>inong>g particles. Junge (1977) found the followong>inong>g relation for theratio R pg of the concentration of a compound ong>inong> the particulate phase over the totalconcentration of the compound ong>inong> the particulate ong>andong> gaseous phase:Rpgczr,p= =czr,totPcϕ0L+ cϕ(44)where:c zr,p = concentration of the compound ong>inong> the particulate phase (µg m -3 ).C zr,tot = concentration of the compound ong>inong> the gaseous ong>andong> particulate phase(µg m -3 ).c = a constant, Junge assumed a value of 0.17 (Pa m) for high molecularweight organic compounds.ϕ = available particle surface (m 2 m -3 of ong>airong>).P 0 L = sub-cooled liquid phase vapour pressure of the compound (Pa).The typical average background value for ϕ is 1.5×10 -4 m 2 m -3 (Whitby, 1978).This means that if P 0 L > 2×10 -4 Pa, over 90% of the compound is ong>inong> the gaseousphase, whilst at P 0 L < 2×10 -6 Pa, over 90% of the compound is ong>inong> the particulatephase (van Pul et al., 1998).1.10 Photochemical reactionGaseous pesticides can be transformed ong>inong> the atmosphere by photolysis(degradation under ong>inong>fluence of sunlight) or by reaction with the followong>inong>gphotochemically formed reactive compounds: OH-radical, NO 3 -radical ong>andong> O 3 . Therate at which a gaseous pesticide is transformed by the reactive compoundsdepends on the concentrations of these reactive compounds ong>andong> the reaction rate ofthe pesticides with these compounds. The reaction rate of pesticides depends ontheir chemical structure ong>andong> will be different for different pesticides. For manypesticides the reaction rates will not be known, but sometimes they can beestimated usong>inong>g some empirical relationships between the structure of the pesticideong>andong> its reactivity (Atkong>inong>son, 1987, 1988; Wong>inong>er ong>andong> Atkong>inong>son, 1990). It should benoted here that if a pesticide reacts, reaction products are formed. These reactionproducts can also be toxic ong>andong> are ong>inong> some cases even more toxic than thepesticides themselves. One should ong>inong> fact also ong>inong>vestigate the fate of these reactionproducts ong>inong> the atmosphere. This can be rather complicated because usually morethan one reaction product is formed. The reaction products can be removed fromthe atmosphere at different rates than their precursors, because they have differentproperties.1.11 Spray drift ong>andong> other forms of depositionDifferent forms ofdepositionPesticide application can lead to different forms for deposition:a) Deposition of the larger sprayed drops due to sedimentation caused bygravitation. This form of deposition is maong>inong>ly ong>inong>fluenced by the physicalproperties of the drops (e.g. ong>inong>ertia, which is a function of the size of thedroplets) ong>andong> not by the chemical properties of the pesticide. Moreover, it isong>inong>fluenced by other factors like wong>inong>d speed, boom height ong>andong> othermeteorological factors.94


) Deposition of sprayed droplets that are so tong>inong>y that their movement is notong>inong>fluenced by gravitation, but only by atmospheric turbulence. This is a form ofdry deposition. The deposition mechanism is different than for the larger dropspreviously mentioned, that deposit due to sedimentation. However, ong>inong> bothcases the physical properties of the drops ong>andong> not the chemical properties of thedrops determong>inong>e the deposition flux to a substantial extent. This form ofdeposition can be more important than the deposition of sprayed drops due tosedimentation (Peter Kryger Jensen, Danish Institute of Agricultural Sciences,Flakkebjerg, Slagelse, personal communication). For more ong>inong>formation on thisprocess see under dry deposition.c) Durong>inong>g sprayong>inong>g, part of the pesticide can evaporate from the droplets beforethey reach the ground. How much evaporates will depend on the size of thedroplets (physical factor), the Henry’s law coefficient of the pesticide,properties of the solvent ong>andong> other related chemical properties. Moreover, it willbe ong>inong>fluence by meteorological factors like temperature, wong>inong>d speed ong>andong> by theheight of the boom sprayer over the surface ong>andong> the type of surface. Theevaporated pesticide can later reach the surface ong>inong> the form of dry ong>andong> wetdeposition.d) After the spray drops have hit the surface, pesticide can evaporate, ong>andong> thegaseous pesticide can then later be deposited ong>inong> the form of dry or wetdeposition.e) Dry deposition. This is deposition of pesticides ong>inong> gaseous form, or afterconversion to particles ong>inong> particulate form to the surface under ong>inong>fluence ofatmospheric turbulence. For gases the chemical properties, the meteorologicalconditions, the concentration ong>inong> ong>andong> properties of the surface they deposit ontothat determong>inong>e the deposition flux. For particles (solid or liquid, i.e. alsoong>inong>cludong>inong>g tong>inong>y spray droplets) physical properties, properties of the surfacewhere they deposit on ong>andong> meteorological processes determong>inong>e the drydeposition flux.f) Wet deposition. This is deposition of pesticides that are removed from theatmosphere ong>inong> gaseous or particulate form by ong>precipitationong>. For pesticides ong>inong>gaseous form the chemical properties (Henry’s law coefficient) can have a largeong>inong>fluence on the wet deposition flux. For pesticides ong>inong> particulate form thephysical properties (which is a function of the size of the particles) has a largeong>inong>fluence on the wet deposition flux.Deposition of sprayed drops due to sedimentation occurs only on the field wherethe pesticide is applied or up to about 20 m outside this area. All other forms ofdeposition occur also at longer distances from the field (0 to greater than 500 km).ong>Pesticidesong> are either deposited on the field where they are applied or outside thefield. The physical ong>andong> chemical processes are the same whether they are depositedon the field or not, but for other reasons (agricultural, environmental, political)differentiation between deposition on the field ong>andong> outside the field is required.Different defong>inong>itions ofspray driftSpray drift due tosedimentation:boom sprayersThere exist apparently different (implicit) defong>inong>itions of spray drift. Onedefong>inong>ition of spray drift is that all pesticide deposited outside the target area isspray drift. This is not a very useful defong>inong>ition because it ong>inong>cludes the results ofmany different processes operatong>inong>g on various scales. In the followong>inong>g only spraydrift due to sedimentation will be discussed, but it should be kept ong>inong> mong>inong>d that thisis not the only form of deposition of drops.In this report we will only compare the dry ong>andong> wet deposition of pesticidesas a function of distance to the field onto which pesticides are applied with thespray drift due to sedimentation. Table 5 gives an impression of the depositioncaused by spray drift due to sedimentation onto crops caused by conventionalboom sprayers (Ganzelmeier, 1995). This results were based on experiments where95


the fluorescence on crops was measured of a dye that was added to the sprayeddrops.Table 5Deposition caused by spray drift due to sedimentation as a function of distance tothe downward wong>inong>d edge of a field onto which pesticides are applied (Ganzelmeier,1995).Deposition forårsaget af afdrift på grund af sedimentation som funktion afafstong>andong>en til kanten af en mark nedstrøms, hvor pesticider er sprøjtet(Ganzelmeier, 1995).Distance from upwong>inong>d edge of the field(m)Deposition caused by spraydrift due to sedimentation(% of the applied dose)1)1 5.02 1.83 1.44 1.05 0.77.5 0.510 0.415 0.220 0.11) The applied dose is the amount per surface area applied (e.g. kg ha -1 ).1.12 Modellong>inong>g the deposition close to the source1.12.1 K-modelA two-dimensional K-model (x ong>andong> z direction) was developed to calculate thediffusion, surface exchange ong>andong> below-cloud scavengong>inong>g of a gaseous non-reactong>inong>gpesticide over a distance of about 2000 m (Figure 19). This model gives thecrosswong>inong>d-ong>inong>tegrated concentration/deposition perpendicular to the wong>inong>d direction,i.e. no diffusion ong>inong> the y-direction is taken ong>inong>to account. Close to a field whereemission occurs the crosswong>inong>d-ong>inong>tegrated concentration is almost equal to theconcentration because the spreadong>inong>g of the plume ong>inong> the y-direction is not importantcompared to the width of the field ong>inong> the y-direction. At larger distance from thesource this is not any longer the case, unless the width of the field ong>inong> the y-directionis extremely large. The reason to use a two-dimensional K-model is threefold:a) Withong>inong> this project we are ong>inong>terested ong>inong> generalised ong>inong>formation, i.e. not for aparticular situation. It is a large, maybe even impossible task, to generalisemodel results for all possible combong>inong>ations of shape ong>andong> size of fields ong>andong>meteorological conditions (wong>inong>d direction, wong>inong>d speed, atmospheric stabilityetc.). For policy makers it is crucial to know at which distance from the sourceong>effectsong> can occur. For that reason it was decided to choose a two-dimensionalK-model, which gives the crosswong>inong>d-ong>inong>tegrated concentrations ong>andong> depositionsas a function of the distance to the source.b) The description of the diffusion of the plume ong>inong> the y-direction is somewhatmore uncertaong>inong> than the diffusion of the plume ong>inong> the z-direction.c) Calculations with a two-dimensional model do not take so much cpu-time ascalculations with a three-dimensional model.96


Figure 19Overview of processes ong>inong>corporated ong>inong> the K-model.Oversigt over de processer, der er ong>inong>dbygget i K-modellen.Model descriptionThe K-model developed is a steady-state model. The emission ong>inong> the model forboth poong>inong>t sources ong>andong> area sources can be given by the user, but for area sourcesthere is the possibility that the user supplies a surface concentration ong>andong> that themodel generates then the emission from the surface concentration ong>andong> themeteorological conditions. As a result the model can the emission of pesticides ong>inong> amore realistic way. In the work reported here a poong>inong>t source was only used tocompare model results with measurements. All other runs were made withemissions that are generated from a surface concentration. The surface exchangewas described with an exchange velocity (16), the vertical diffusion with a an eddydiffusivity coefficient (12), the wong>inong>d speed was made a function of height (6).Scavengong>inong>g was only taken ong>inong>to account ong>inong> one case, where a maximum belowcloudscavengong>inong>g coefficient for highly soluble gases with a molecular weight of300 g mol -1 was applied (42). In-cloud scavengong>inong>g was not taken ong>inong>to account,because the plume will ong>inong> general not have reached the cloud level at 2 km from thesource. Photochemical reaction ong>andong> conversion from the gaseous phase to theparticulate phase can also be taken ong>inong>to account. For all calculations 40logarithmically spaced vertical layers were used, except for the calculation of thelong-range transport for which 10 layers were used. A more detailed description ofthis type of model can be found ong>inong> Asman (1998).1.12.2 Verification of the vertical diffusion calculated with the modelTracer experimentThe vertical diffusion part of the model was tested agaong>inong>st the results of atracer experiment ong>inong> the U.S.A., where the tracer was also release from alow-level source, just as pesticides are. In this experiment sulphur dioxide wasreleased from a 0.46 m high poong>inong>t source (van Ulden, 1978). The concentrationswere measured at a height of 1.5 m at distances of 50, 200 ong>andong> 800 from the source.The surface roughness length for the observations was 0.008 m. It appeared that themodel overestimated the crosswong>inong>d-ong>inong>tegrated concentrations at 800 m from thesource, especially for stable ong>andong> unstable atmospheres. This phenomenon was alsoobserved by other modellers (Grynong>inong>g et al., 1983; Brown et al., 1993). They guessthat part of the observed differences can be explaong>inong>ed by the fact that sulphurdioxide is dry depositong>inong>g to a mong>inong>or extent. But they mention also other reasons.97


Such experiments are very expensive ong>andong> for that reason there are not manyexperimental data to verify the diffusion.Adjustong>inong>g the It was decided to ong>inong>crease the vertical eddy diffusivity ong>inong> the model by 30%vertical eddy diffusivity to get better results. In Figure 20 the concentrations modelled ong>inong> this way werecompare with measurements. At 50 m from the source the model underpredicts theconcentrations by 9%, at 200 m it overpredicts the. concentrations by 7% ong>andong> at800 m it overpredicts the concentrations by 23%. Takong>inong>g ong>inong>to account theuncertaong>inong>ty ong>inong> the measurements this is a very reasonable result. In all followong>inong>gcalculations the eddy diffusivity was ong>inong>creased by 30%.Figure 20Modelled vs. measured crosswong>inong>d-ong>inong>tegrated concentration divided by the sourcestrength for three downwong>inong>d distances: 50, 200 ong>andong> 800 m.Modelleret vs. målt koncentration på tværs af vong>inong>den divideret med kildestyrken fortre nedstrøms afstong>andong>e: 50, 200 ong>andong> 800 m.1.12.3 Situation modelled ong>inong> all further model calculationsModel areaAll the calculations made ong>inong> the followong>inong>g are made for a field of 250 m length ong>inong>the x-direction (wong>inong>d direction) where pesticide has been applied ong>andong> emissionoccurs (“emission field”) followed by an area where only deposition can occur of1750 m ong>inong> the x-direction (“deposition area”). The only exception is thecalculations of the possible long-range transport where the deposition up to about1000 km from the field are calculated (see section 1.12.10).Emission modelled usong>inong>g The emission of the field is calculated assumong>inong>g a constant crosswong>inong>d-98


a surface concentration ong>inong>tegrated gas concentration ong>inong> the surface (here arbitrarily set to 1 kg m -2 ). It doesnot matter that the surface concentration is arbitrary ong>inong> this case, because ong>inong> thefollowong>inong>g the deposition is always expressed as an accumulated fraction of theemission, or cases are compared where one parameter is varied, but where thesurface concentration is the same. In the present model concept it is only possibleto use sources at the ground when usong>inong>g the surface concentrations to describe theemission. There is, however, not a large difference ong>inong> the results between situationswhere a ground source or a source at e.g. 1 m height is used. This was verified forpoong>inong>t sources, for which the model can take sources at all heights ong>inong>to account. Themodel does not take ong>inong>to account that this surface concentration usually willdecrease as a function of time due to loss processes (volatilisation, leachong>inong>g,degradation, uptake by plants etc.). The reason is that this was beyond the scope ofthis project. Moreover, by not varyong>inong>g the surface concentration, the results can begeneralised.Relative changesimportantThe absolute concentration or flux is not important ong>inong> the calculations shownong>inong> the followong>inong>g, only relative differences between runs for different conditions areimportant. The maong>inong> purpose of the followong>inong>g sections is to show how the emissionong>andong> deposition can vary as a function of e.g. the friction velocity/wong>inong>d speed, theatmospheric stability or the surface roughness length. For the results where theaccumulated fraction of the emission deposited is shown, the absolute numbers areimportant.Values model parameters Variations of the conditions with respect to a “base case” are studied. In theusedbase case both the field ong>andong> the deposition area have a surface roughness length of0.1 m (crops of about 1 m high), a friction velocity of 0.3 m s -1 ong>andong> a neutralatmosphere (characterised by a Monong>inong>-Obukhov length of 2000 m) ong>andong> a mixong>inong>gheight of 400 m. The atmosphere is most often neutral ong>inong> Denmark, but theatmosphere can also be unstable or stable. It should be noted that the mixong>inong>g heightonly ong>inong>fluences the calculated concentrations ong>andong> depositions if pesticides havebeen mixed up to the mixong>inong>g height. This is not the case for transport up to about2000 m from the low source. But the mixong>inong>g height has an ong>inong>fluence on thecalculated concentrations ong>andong> depositions durong>inong>g long-range transport that is shownong>inong> one case. The wong>inong>d speed at 10 m height adopted ong>inong> the calculations is 3.97 m s -1 , which is about the average wong>inong>d speed ong>inong> Denmark. It is assumed that thepesticide is gaseous ong>andong> has a molecular mass of 300 g mol -1 . It is assumed that noatmospheric reaction occurs or conversion from the gaseous to the particulatephase. In all cases, except the case where the surface resistance has been varied, asurface resistance for both the emission ong>andong> deposition area of 0 s m -1 has beenadopted. This gives both a maximum emission ong>andong> dry deposition rate. This is doneto calculate the “worst case situation”. In reality it is likely that at least thedeposition rate is much lower than adopted ong>inong> the calculations here. Apart from theresults presented ong>inong> one section only dry deposition is taken ong>inong>to account ong>andong> notwet deposition.1.12.4 Influence of friction velocity/wong>inong>d speedFigure 21 shows the ong>inong>fluence of the variation of the friction velocity (whichong>inong>fluences the wong>inong>d speed) on the horizontal flux of the pesticide verticallyong>inong>tegrated over the whole mixong>inong>g layer as function of the distance to the upwong>inong>dedge of the emission field. This gives ong>inong>formation on much of the pesticide isong>airong>borne. The case with a friction velocity of 0.30 m s -1 is the base case. The wong>inong>dspeed at 10 m height is 1.98, 3.97 ong>andong> 7.93 m s -1 for friction velocities of 0.15, 0.30ong>andong> 0.60 m s -1 .99


The emission (occurrong>inong>g ong>inong> the first 250 m) ong>inong>creases long>inong>early with wong>inong>d speed. Theaccumulated fraction of the emission dry deposited as a function of distance fromthe downwong>inong>d edge of the emission field is, however, the same for all cases (theseresults are the same as presented for the neutral case ong>inong> Figure 23 ong>inong> the nextsection). As mentioned before the calculations are made with a surface resistanceof zero. If the surface resistance is not 0, i.e. e.g. for moderately ong>andong> slightly theaccumulated fraction deposited as a function of distance to the downwong>inong>d edge ofthe emission area will decrease the friction velocity/wong>inong>d speed if the surfaceconcentration is 0 (Asman, 1998).Figure 21Modelled horizontal flux vertically ong>inong>tegrated over the whole mixong>inong>g layer fordifferent friction velocities as a function of distance from the upwong>inong>d edge of theemission field. This flux ong>inong>dicates how much ong>airong>borne material ong>inong>tegrated over thewhole mixong>inong>g height passes by per unit of time.Modelleret horisontal flux vertikalt ong>inong>tegreret over hele blong>andong>ong>inong>gshøjden forforskellige friktionshastigheder som funktion af afstong>andong>en til kanten af markenopstrøms, hvor fordampnong>inong>g fong>inong>der sted. Denne flux er et mål for hvor megetmateriale der passerer forbi pr. tidsenhed, ong>inong>tegreret over alle højder.1.12.5 Influence of atmospheric stabilityIn this section the ong>inong>fluence of the atmospheric stability on the model results isdiscussed. For the stable atmosphere (characterised by a Monong>inong>-Obukhov length of20 m) a friction velocity of 0.08 m s -1 was chosen ong>andong> a mixong>inong>g height of 400 m.The wong>inong>d speed at 10 m height is 1.58 m s -1 ong>inong> this case. For the unstableatmosphere (characterised by a Monong>inong>-Obukhov length of -17 m) a frictionvelocity of 0.28 m s -1 was chosen ong>andong> a mixong>inong>g height of 400 m. The wong>inong>d speed at10 m height is 3.03 m s -1 ong>inong> this case. All these situations ong>inong>cludong>inong>g the conditionsdurong>inong>g the neutral atmosphere are chosen to be “typical” of such conditions. Inreality the meteorological conditions show a wide variation ong>andong> durong>inong>g a particularevent where pesticides are applied none of those “typical” conditions may occur.Figure 22 shows the crosswong>inong>d-ong>inong>tegrated concentration as a function of distancefrom the source for different atmospheric conditions. The concentrations are highereverywhere under stable conditions than under neutral or unstable conditions. The100


eason for this is twofold. In stable conditions the friction velocity is less whichleads to both a reduced wong>inong>d speed ong>andong> a reduced vertical diffusion.Figure 22Modelled crosswong>inong>d-ong>inong>tegrated concentration for different atmospheric stabilitiesas a function of the distance from the upwong>inong>d edge of the emission field.Modelleret koncentration på tværs af vong>inong>den for forskellig atmosfærisk stabilitetsom funktion af afstong>andong>en fra kanten af marken opstrøms, hvor fordampnong>inong>g fong>inong>dersted.Figure 23Modelled horizontal flux vertically ong>inong>tegrated over the whole mixong>inong>g layer fordifferent atmospheric stabilities as a function of distance from the upwong>inong>d edge ofthe emission field. This flux ong>inong>dicates how much ong>airong>borne material ong>inong>tegrated overthe whole mixong>inong>g height passes by per unit of time.Modelleret horisontal flux vertikal ong>inong>tegreret over hele blong>andong>ong>inong>gshøjden forforskellig atmosfærisk stabilitet som funktion af afstong>andong>en til kanten af marken101


opstrøms, hvor fordampnong>inong>g fong>inong>der sted. Denne flux er et mål for hvor megetmateriale der passerer forbi pr. tidsenhed, ong>inong>tegreret over alle højder.Emission rateVariation ong>inong> theemission rate withatmospheric stabilityFigure 23 shows the horizontal flux (vertically ong>inong>tegrated over the wholemixong>inong>g height) as a function of the distance from the upwong>inong>d edge of the fieldwhere pesticides are applied. The horizontal flux ong>inong>dicates how much pesticide isong>airong>borne. This gives ong>inong>formation on how much pesticide is ong>airong>borne. Durong>inong>g thefirst 250 m the horizontal flux ong>inong>creases because the ong>airong> is transported over theemittong>inong>g field. Then the horizontal flux decreases slowly due to dry deposition ong>andong>that the largest part of the emitted pesticide is still ong>airong>borne at 2000 m from theupwong>inong>d edge of the field. It should be stressed here, that the dry deposition is set toits maximum value ong>inong> these calculations by assumong>inong>g a surface resistance of 0 s m -1 .The emission rate varies with the atmospheric conditions. Durong>inong>g neutralong>andong> unstable conditions the emission rate is relatively large due to the largerturbulence. Durong>inong>g stable conditions the emission rate is much lower due to thelower turbulence. Figure 24 shows the accumulated fraction of the emission (i.e.fraction of the horizontal flux at the end of the emission field) that is deposited as afunction of the distance from the downwong>inong>d edge of the emission field. This figureshows clearly that the deposition is largest for stable conditions where theconcentration near the surface is relatively large due to reduced vertical mixong>inong>g ong>andong>wong>inong>d speed. For neutral ong>andong> stable conditions the fraction deposited is smaller.Figure 24Modelled accumulated fraction of the emission dry deposited for differentatmospheric stabilities as a function of distance from the downwong>inong>d edge of theemission field.Modelleret akkumuleret fraktion af emissionen, som er tørdeponeret ved forskelligatmosfærisk stabilitet som funktion af afstong>andong>en til kanten af en mark nedstrøms,hvor fordampnong>inong>g fong>inong>der sted.Crosswong>inong>d-ong>inong>tegrated Figure 24 shows the crosswong>inong>d-ong>inong>tegrated concentration at 1 m height as aconcentration at function of the distance from the upwong>inong>d edge of the field. This clearly shows1 m height that the concentration under stable conditions is larger than under neutral ong>andong>unstable conditions. This is, as mentioned before, a result of the reduced vertical102


mixong>inong>g ong>andong> wong>inong>d speed. The concentration decreases very sharply after thedeposition area has been reached. This is maong>inong>ly due to vertical mixong>inong>g ong>andong> not todry deposition.1.12.6 Vertical concentration profilesAt 500 m from the source, i.e. 250 m ong>inong> the deposition area the concentra-tion decreases with height ong>inong> the lowest metres. This is caused by dry deposition,that removes material faster than turbulence can replenish from higher up ong>inong>the atmosphere. At the 2000 m poong>inong>t the concentration is still lower ong>inong> the lowestmetres, but all concentrations are lower than at the 500 m poong>inong>t, except at 40-50 mheight. This is maong>inong>ly caused by the fact that the pesticide has been mixed over agreater height. At 40-50 m height the concentration is higher at 2000 m from theupwong>inong>d edge of the emission field than at 500 m from that edge. The reason for thatis that at 500 m, mixong>inong>g has not yet been able to transport so much material upwardas at 2000 m.Vertical profiles on theemission fieldVertical profiles ong>inong> thedeposition areaFigure 25 shows the vertical concentration profiles ong>inong> the lowest 50 m of theatmosphere at different distances from the upwong>inong>d edge of the emission field. Theprofiles at 12.5 ong>andong> 250 m belong to the emission field. The concentrations at 250m from the upwong>inong>d edge of the emission field are at all heights higher than at 12.5m from the upwong>inong>d edge of the emission field. This is caused by the fact that theconcentration at 250 m distance is the result of a much larger emittong>inong>g area upwong>inong>dthan the concentration at 12.5 m distance.Figure 25Modelled vertical concentration profiles as a function of the distance to the upwong>inong>dedge of the emission field.Modellerede vertikale koncentrationsprofiler som funktion af afstong>andong>en til kantenaf den mark nedstrøms, hvor fordampnong>inong>gen fong>inong>der sted.103


1.12.7 Reduction of the emission rate on the emission field due to upwong>inong>d emissionsFigure 26 shows how the net emission rate (resultong>inong>g from the upward emissionflux ong>andong> the downward deposition flux) on the emission field decreases as afunction of distance from the edge of the emission field for neutral atmosphericconditions (base case). This is caused by the fact that the net emission rate dependson the difference (c g,surf - c g,ong>airong> ); c g,surf is take constant ong>inong> the model, but the ong>airong>concentration c g,ong>airong> ong>inong>creases with distance to the upwong>inong>d edge of the emission fielddue to the emission ong>inong> from the upwong>inong>d area. The ong>inong>creased c g,ong>airong> concentrationreduces the net emission rate. This effect is not very substantial, but noticeable ong>andong>will ong>inong>crease with the size of the emission field ong>inong> the wong>inong>d direction (x-direction).In the model c g,surf is taken constant. In reality this is not the case ong>andong> c g,surf willmaong>inong>ly decrease as a function of time as a result of different loss processes(volatilisation, leachong>inong>g, degradation, uptake by plants etc.). But the maong>inong> messageof the results presented ong>inong> Figure 26 is that the average net emission rate of a fieldalso to a mong>inong>or extent will depend on the size of the field ong>inong> the wong>inong>d direction forthe same pesticide application rate (kg ha -1 ).Figure 26Modelled change ong>inong> net emission flux on the emission field a function of thedistance to the upwong>inong>d edge of the field.Modelleret ændrong>inong>g i netto emissionsfluxen på den mark hvorpå fordampnong>inong>genfong>inong>der sted som funktion af kanten af marken opstrøms.1.12.8 Effect of surface roughnessIn the base case a surface roughness length (z 0m ) of 0.1 m is adopted, beong>inong>grepresentative of crops of about 1 m high. Sometimes, however, pesticides areapplied when there are almost no crops. The surface roughness is then about 0.006m. The friction velocity itself is also ong>inong>fluenced by the surface roughness. Thismeans that we cannot use the base case friction velocity of 0.3 m s -1 ong>inong> thecalculation for the case with the lower surface roughness.104


Method to estimate thefriction velocity foranother surfaceOne way to tackle this problem is to fong>inong>d the wong>inong>d speed at 60 m height forthe base case usong>inong>g (6) ong>andong> then to fong>inong>d friction velocity for the low surfaceroughness case from this wong>inong>d speed usong>inong>g the same equation. This can be donebecause the wong>inong>d speed at 60 m does not any longer depend on the local surfaceroughness, but on the surface roughness of an area of about 5x5 km 2 (Wierong>inong>ga ong>andong>Rijkoort, 1983). By doong>inong>g this it is assumed that the average surface roughness ofthis area does not vary ong>andong> that we only go down from 60 m height to another partof the area with a lower surface roughness. This method is to some extent arbitrary,but at least the friction velocity becomes more realistic ong>inong> this way. With thismethod a friction velocity of 0.21 m s -1 is found.Figure 27 shows the horizontal flux (vertically ong>inong>tegrated over the whole mixong>inong>gheight). This is just the amount of pesticide that is ong>airong>borne. Figure 27 shows thatthe emission rate is much less for the lower surface roughness. This is caused byreduced turbulence (characterised by a lower friction velocity). Figure 28 showsthat the accumulated fraction of the emission that is dry deposited does not varymuch with surface roughness. The reason for this is presumably that ong>inong> the case ofthe lower surface roughness not only the deposition velocity is reduced, but alsothe vertical diffusion. A reduced vertical diffusion leads to higher concentrationsnear the surface ong>andong> this may compensate for the lower dry deposition rate. Itshould be noted here that it is more difficult to model the case where the emissionarea has a different surface roughness than the deposition area because an ong>inong>ternalboundary layer is formed at the boundary between the two roughnesses.Figure 27Modelled horizontal flux for different surface roughnesses as a function of distancefrom the upwong>inong>d edge of the emission field. This flux ong>inong>dicates how much ong>airong>bornematerial ong>inong>tegrated over the whole mixong>inong>g height passes by per unit of time.Modelleret horisontal flux vertikal ong>inong>tegreret over hele blong>andong>ong>inong>gshøjden forforskellige ruhedshøjder som funktion af afstong>andong>en til kanten af marken opstøms,hvor fordampnong>inong>g fong>inong>der sted. Denne flux er et mål for hvor meget materiale derpasserer forbi pr. tidsenhed, ong>inong>tegreret over alle højder.105


Figure 28Modelled accumulated fraction of the emission dry deposited for different surfaceroughnesses as a function of distance from the downwong>inong>d edge of the emissionfield.Modelleret akkumuleret fraktion af emissionen som er tørdeponeret ved forskelligeruhedshøjder som funktion af afstong>andong>en til kanten af en mark nedstrøms, hvorfordampnong>inong>g fong>inong>der sted.1.12.9 Effect of surface resistanceUntil now we have only discussed the extreme case where the surface resistance(r c ) is 0 ong>andong> a maximum exchange/dry deposition velocity occurs. This is ong>inong> realityonly the case for high soluble ong>andong> very reactive compounds like gaseous nitric acid(HNO 3 ). Without modellong>inong>g the processes ong>inong> the soil ong>andong> ong>inong> ong>andong> on the crops it isdifficult to know what the surface resistance is for pesticides, but some examplesfor well known ong>airong> pollutants can maybe give an impression of possibleimplications of a non-zero surface resistance.Figure 29 shows the effect of variations ong>inong> the surface resistance. The followong>inong>gsurface resistances r c were chosen: 0, 100 ong>andong> 1000 s m -1 . This reflects the situationfor a highly soluble gas like HNO 3 (r c = 0), a moderately soluble gas like SO 2 (r c =100) ong>andong> a slightly soluble gas like NO (r c = 1000) For a highly dry deposited gasthe accumulated fraction of the emission deposited at the end of the deposition areais 0.1775. For a moderately soluble gas this is 0.0741 ong>andong> for a slightly soluble gasthis is 0.0118. This means that moderately ong>andong> slightly soluble gases are beong>inong>gtransported over much longer distances than highly soluble gases, if they are notdepleted relatively fast by other processes.106


Figure 29Modelled accumulated fraction of the emission dry deposited for different surfaceresistances as a function of distance from the downwong>inong>d edge of the emission field.Modelleret akkumuleret fraktion af emissionen som er tørdeponeret for forskelligeoverflademodstong>andong>e som funktion af afstong>andong>en til kanten af en mark nedstrøms,hvor fordampnong>inong>g fong>inong>der sted.1.12.10 Possibility of long-range transportFigure 30 gives the accumulated fraction of the emission dry deposited as afunction of distance from the upwong>inong>d edge of the emission field for a compoundwith zero surface resistance, i.e. with the maximum possible dry depositionvelocity like for gaseous HNO 3 . The calculations are ong>inong> this case made fordistances up to 1000 km from the downwong>inong>d edge of the emission field.Even ong>inong> this extreme case about half of the released compound is still ong>airong>borne after100 km. As pesticides generally are less soluble ong>andong> reactive as HNO 3 , it is likelythat they can be transported over considerable distances before they are deposited.As it is only raong>inong>ong>inong>g 5-10% of the time, ong>inong>corporation of raong>inong> events will not changethe overall picture much. Some pesticides that are not very water soluble ong>andong> donot react can have atmospheric lifetimes of at least several years (e.g. long>inong>dane) ong>andong>are then found everywhere (Wania ong>andong> Mackay, 1995).107


Figure 30Modelled accumulated fraction of the emission dry deposited as a function ofdistance from the downwong>inong>d edge of the emission field.Modelleret akkumuleret fraktion af emissionen som er tørdeponeret som funktionaf afstong>andong>en til kanten af en mark nedstrøms, hvor fordampnong>inong>g fong>inong>der sted.1.12.11 Maximum sum of dry ong>andong> wet depositionUsually pesticides are not applied durong>inong>g raong>inong> events or if raong>inong> is expected to occurong>inong> the near future. It is, however, useful to know how much wet deposition couldmaximally be compared to the maximum dry deposition (i.e for the case the surfaceresistance is 0). In all cases it is assumed that it is only raong>inong>ong>inong>g ong>inong> the depositionarea ong>andong> not ong>inong> the emission field. It is also assumed that only below-cloudscavengong>inong>g occurs, which is a reasonable assumption because the pesticide plumewill not have been diffused to a height where there can be clouds (350-400 m onthe average, see Table 4).In Figure 31 three cases are presented. The first case is the base case where onlydry deposition occurs. In the second case not only dry deposition occurs, but alsowet deposition at a raong>inong>fall rate of 1 mm hr -1 ong>inong> the deposition area only. In the thirdcase the raong>inong> fall rate is ong>inong>creased to the rather high value of 10 mm hr -1 . In thesecalculations it is assumed that the surface resistance is zero (maximum drydeposition velocity) ong>andong> that the raong>inong>drops cannot be saturated, i.e. that all gaseouspesticide reachong>inong>g the raong>inong>drop surface is absorbed (maximum below-cloudscavengong>inong>g rate).Figure 31 shows that the contribution from wet deposition, also for the rela-tively high raong>inong>fall rate of 10 mm hr -1 is much less than the contribution from drydeposition. This is the case for highly soluble gases. For other gases the ratio wetvs. dry deposition can be different.Maximum sum of dryong>andong> wet depositionWet deposition vs.dry depositionIn this situation less than 25% of the emitted pesticide is deposited withong>inong>1750 m from the downwong>inong>d edge of the emission field. This percentage could behigher ong>inong> case the wong>inong>d speed is higher than about 4 m s -1 (used ong>inong> the base case),which is about the average wong>inong>d speed ong>inong> Denmark. It is, however, unlikely thatfarmers will apply pesticides at high wong>inong>d speeds.108


Figure 31Modelled accumulated fraction of the emission dry deposited for different raong>inong>fallrates as a function of distance from the downwong>inong>d edge of the emission field.Modelleret akkumuleret fraktion af emissionen som er tørdeponeret for forskelligeregn-ong>inong>tensiteter som funktion af afstong>andong>en til kanten af en mark nedstrøms, hvorfordampnong>inong>g fong>inong>der sted.1.12.12 Comparison of dry ong>andong> wet deposition with spray drift due to sedimentationFigure 26 shows the dry ong>andong> wet deposition fluxes for each downwong>inong>d dis-tance expressed as a percentage of the average emission flux on the fieldonto which pesticides are applied. The emission flux, however, is not thesame everywhere on the field. For that reason the dry ong>andong> wet deposition fluxeswere expressed as a function of the average emission flux on the field. Thecalculations were made for the case with the maximum sum of dry ong>andong> wetdeposition (surface resistance r c = 0 ong>andong> a high raong>inong>fall rate of 10 mm hr -1 )presented ong>inong> the previous section. Such high deposition fluxes occur only for highlysoluble gases (such as HNO 3 ). Gaseous pesticides will usually not have suchextremely high solubilities. For that reason also results are presented formoderately soluble (r c = 100) ong>andong> slightly soluble (r c = 1000) gaseous pesticides.The problem is that it was not possible to develop a model for below-cloudscavengong>inong>g of moderately or slightly soluble gases withong>inong> this project. So onlyong>inong>formation is available on the dry deposition of these compounds ong>andong> not on theDepositionSpray drift due tosedimentationDry ong>andong> wet depositionflux expressed as percentageof the emissionfluxSo far this report has been focused on how much of the emitted pesticide can bedry ong>andong> wet deposited ong>inong> the deposition area, i.e. the accumulated fraction of theemission deposited ong>inong> the area of 0-1750 m from a field onto which pesticides areapplied. The amount of pesticide emitted is only a fraction of the total amount ofpesticide applied per unit area (dose).Deposition caused by spray drift due to sedimentation is usually expressedas a percentage of the dose ong>andong> not the accumulated deposition due to spray drift asa function of distance to the downwong>inong>d edge of the field is given (which wasreported previously), but the deposition at each distance. This means that the dryong>andong> wet deposition results presented previously have to be expressed as thedeposition for each distance ong>andong> that assumptions have to be made on the fractionof the dose that volatilises ong>inong> order to compare the dry ong>andong> wet deposition with thedeposition caused by spray drift due to sedimentation.109


wet deposition. Moreover, if it had been possible to model the below-cloudscavengong>inong>g of these compounds, another problem would arise: the scavengong>inong>g ofsuch compounds depends also on the concentration ong>inong> raong>inong>drops before they reachthe ong>airong> below the cloud ong>andong> this concentration should then be set arbitrarily. Forthat reason it was decided only to ong>inong>clude dry deposition ong>inong> the results presentedhere. Figure 32 shows that the flux is highest for the highly soluble gas ong>andong> can be38% of the emission flux at a distance of 1 m from the field. For other gases, theflux is lower, but can still be 1% of the emission flux close to the field.Figure 32Modelled deposition flux expressed as percentage of the average emission flux onthe field as a function of distance from the downwong>inong>d edge of the emission field.The results are given for 3 cases: a) a surface resistance of 0 s m -1 ong>andong> a raong>inong>fallrate of 10 mm hr -1 , b) a surface resistance of 100 s m -1 ong>andong> no raong>inong> ong>andong> c) a surfaceresistance of 1000 s m -1 ong>andong> no raong>inong>.Modelleret depositionsflux udtrykt som procentdel af den gennemsnitlige emissionflux på marken, som funktion af afstong>andong>en til kanten af en mark nedstrøms, hvorfordampnong>inong>g fong>inong>der sted. Resultaterne er givet for 3 situationer. a) enoverflademodstong>andong> på 0 s m -1 og en regnong>inong>tensitet på 10 mm i timen, b) enoverflademodstong>andong> på 100 s m -1 og c) en overflademodstong>andong> på 1000 s m -1 .Emission factor of As has been mentioned previously, not all pesticide that has been applied20% of the applied volatilises. In order to be able to compare the dry/wet deposition with thedose assumed spray drift an emission factor has to be assumed, i.e. the percentage of the appliedpesticide that volatilises. ong>Pesticidesong> with an emission factor of 20% are notuncommon. For that reason such a factor was chosen for the example below. Theestimated deposition caused by spray drift due to sedimentation was taken fromGanzelmeier (1995).110


Figure 33Modelled dry/wet deposition flux ong>andong> estimated deposition flux caused by spraydrift due to sedimentation expressed as percentage of the applied dose as afunction of distance from the downwong>inong>d edge of the field onto which the pesticide isapplied. The results are given for 4 situations: a) dry ong>andong> wet deposition with asurface resistance of 0 s m -1 ong>andong> a raong>inong>fall rate of 10 mm hr -1 , b) dry depositionwith a surface resistance of 100 s m -1 ong>andong> no raong>inong>, c) dry deposition with a surfaceresistance of 1000 s m -1 ong>andong> no raong>inong> ong>andong> d) deposition caused by spray drift due tosedimentation.Modelleret tør/våddepositionsflux og den estimerede afdriftsflux pga.sedimentation udtrykt som procentdel af den anvendte dosis på marken, somfunktion af afstong>andong>en til kanten af en mark nedstrøms, hvor pesticidet er sprøjtet.Resultaterne er givet for 4 situationer. a) tør- og våddeposition med enoverflademodstong>andong> på 0 s m -1 og en regnong>inong>tensitet på 10 mm i timen, b)tørdeposition med en overflademodstong>andong> på 100 s m -1 c) tørdeposition med en enoverflademodstong>andong> på 1000 s m -1 og d) deposition forårsaget af afdrift pga.sedimentation.Importance of dry ong>andong> Figure 33 shows the deposition (derived from Figure 32 by assumong>inong>g 20%wet deposition of s volatilisation) ong>andong> spraydrift flux expressed as a percentage of the dose.pesticidecompared to Figure 34 is ong>inong> prong>inong>ciple the same as Figure 33, but shows only results at thedeposition due to first 50 m. The maong>inong> conclusion is that for highly soluble gases the deposispraydrifttion is everywhere larger than the spray drift (for an emission factor of 20%). Alsofor moderately ong>andong> slightly soluble gases the deposition is appreciable ong>andong> is largerthan the spray drift at distances larger than 2 m (moderately soluble gases) or 15 m(slightly soluble gases) from the downwong>inong>d edge of the field. From Figure 33 it canbe noted that deposition due to spray drift decreases much faster with distance thanthe dry/wet deposition. This is maong>inong>ly due to the fact that the relatively large spraydroplets have fall velocities that are relatively large ong>andong> hence cannot betransported over large distances.111


Figure 34Modelled dry/wet deposition flux ong>andong> estimated deposition flux due to spray driftexpressed as percentage of the applied dose as a function of distance from thedownwong>inong>d edge of the field onto which the pesticide is applied. The results aregiven for 4 situations: a) dry ong>andong> wet deposition with a surface resistance of 0 s m -1ong>andong> a raong>inong>fall rate of 10 mm hr -1 , b) dry deposition with a surface resistance of 100s m -1 ong>andong> no raong>inong>, c) dry deposition with a surface resistance of 1000 s m -1 ong>andong> noraong>inong> ong>andong> d) deposition due to spray drift.Modelleret tør/våddepositionsflux og den estimerede afdriftsflux udtrykt somprocentdel af den anvendte dosis på marken, som funktion af afstong>andong>en til kanten afen mark nedstrøms, hvor pesticidet er sprøjtet. Resultaterne er givet for 4situationer. a) tør- og våddeposition med en overflademodstong>andong> på 0 s m -1 og enregnong>inong>tensitet på 10 mm i timen, b) tørdeposition med en overflademodstong>andong> på100 s m -1 c) tørdeposition med en en overflademodstong>andong> på 1000 s m -1 og d)deposition forårsaget af afdrift.1.13 Intermezzo: conclusions on the model resultsOnly the ong>inong>fluence ofmeteorological factorson the emission ratereportedIn the followong>inong>g the emission rate is discussed. In prong>inong>ciple this report dealsonly with the meteorological factors ong>inong>fluencong>inong>g the emission. Other factorswill, however, also ong>inong>fluence the emission rate. These factors are e.g. theproperties of the surface (plant, soil) ong>andong> the processes goong>inong>g on ong>inong> ong>andong> on thesurface (see section 1.6. for a brief discussion). A basic conclusion is, that itdepends on the properties of the compounds ong>andong> the surface how important themeteorological ong>inong>fluence on the emission rate is. Here only the effect ofmeteorological factors on the emission rate are discussed.Results are maong>inong>ly given Unless ong>inong>dicated otherwise the results presented are for a highly soluble gas.for a highly soluble gas: This is done because the deposition is highest ong>inong> that case. In this way angives maximum estimate is obtaong>inong>ed on the maximum possible deposition close to a field ontodepositionwhich pesticides are applied. This was one of the maong>inong> objectives of the project.Emission rate as afunction of wong>inong>d speedDry deposition as aThe emission rate of pesticides applied to a field ong>inong>creases with wong>inong>d speed.In case of highly soluble gases, when the surface resistance is 0, the accu-112


function of the wong>inong>dspeedmulated fraction of the emission dry deposited as a function of distance to thedownwong>inong>d edge of the field will be ong>inong>dependent of the wong>inong>d speed. For moderatelyor slightly soluble a surface resistance has a higher value than 0. In that case theaccumulated fraction of the emission dry deposited will decrease with wong>inong>d speed,at least if the surface concentration is 0. In other words: moderately ong>andong> slightlysoluble gases will ong>inong> that case be transported over longer distances at higher wong>inong>dspeed.Emission as a function of Calculations were made for typical atmospheric conditions for differentatmospheric stability stability classes. The emission rate is higher durong>inong>g neutral ong>andong> stable atmosphericconditions than durong>inong>g stable conditions.Dry deposition as a The accumulated fraction of the emission dry deposited is larger ong>inong> the folfunctionof atmospheric lowong>inong>g order: for stable < neutral < unstable atmospheric conditions.stabilityEmission rate as afunction of field sizeEmission rate as afunction of the surfaceroughnessThe (net) emission rate (kg m -2 s -1 ) decreases with the size of the field.This is, however, a mong>inong>or effect. In the calculations a constant surfaceconcentration was adopted, that was the same everywhere. The surfaceconcentration will ong>inong> reality decrease with time due to depletion caused by theemission ong>andong> for that reason it is likely that the observed effect is negligible ong>inong>reality.The emission rate will ong>inong>crease with the surface roughness, if the wong>inong>dspeed is the same at greater height (60 m). This is due to ong>inong>creasedturbulence. For that reason the emission rate form a crop will be higher than frombare soil (for the same compound under the same conditions).Deposition as a function The accumulated fraction of the emission dry deposited ong>inong>creases onlyof the surface roughness slightly with surface roughness. It should be noted, however, that the emissionitself ong>inong>creases much with surface roughness.Variation ong>inong> the The variation ong>inong> the emission rate due to variations ong>inong> the friction wong>inong>demission rate due to speed ong>andong> the atmospheric stability can easily be more than a factor 4 forvariations ong>inong> the those pesticides for which the emission rate is governed by meteorologicalmeteorological conditions processes. In practise the variation can be even greater because the gas phaseconcentration ong>inong> the soil is highly temperature dependent, because the Henry’s lawcoefficient is a function of temperature.Mong>inong>imum long-rangetransportResults for a highly soluble gas ong>inong>dicate that about half of the emittedamount will still be ong>airong>borne after a transport distance of 100 km. For less solublegases ong>andong> for particles contaong>inong>ong>inong>g pesticides much more than half of the emittedamount will be ong>airong>borne after 100 km. Often half of it will still be ong>airong>borne after1000 km.Rapidly decreasong>inong>g ong>airong> Measurements close to sources will show that the concentration ong>inong> the ong>airong> atconcentrations with ground-level decreases rapidly with distance to the source. It is temptong>inong>g todistance does not conclude then that the pesticide is not travellong>inong>g over long distances. In fact,necessarily mean a short the opposite is true. The concentration at ground-level decreases rapidly withtransport distance distance because the compound is beong>inong>g mixed rapidly to greater heights, where itis not subject to removal by dry deposition. As a result it can be transported overlong distances, at least when it is not removed by raong>inong>.Influence of ong>precipitationong> Withong>inong> 2 km from a source most of the emitted pesticide has not yet reachedthe clouds ong>andong> only wet removal by the less efficient below-cloud scavengong>inong>gprocess occurs. For a highly soluble gas the wet deposition ong>inong> this area is much less113


than the dry deposition. The ratio wet to dry deposition depends e.g. on themeteorological conditions ong>inong>cludong>inong>g the ong>precipitationong> rate. The ratio wet to drydeposition will ong>inong> general be different for different compounds. For gases both thedry ong>andong> the wet removal rate ong>inong>crease with solubility, ong>inong>dicatong>inong>g lower dry ong>andong> wetremoval rates for these gases. For most particulate pesticides the wet removal ratewill be much larger than the dry removal rate, because particles are not dry depositat a high rate, but are removed very efficiently by ong>precipitationong> (see section 1.5).Maximum accumulated Under average Danish conditions less than 25% of the emission of a highlyfraction of the emission soluble gas will be removed by dry ong>andong> wet deposition withong>inong> 2 km from thedeposited withong>inong> 2 km source area. For less soluble gases ong>andong> for particulate pesticides much lessfrom the source than 25% will be deposited withong>inong> 2 km from the source. If only dry depositionoccurs, a very rough estimate will be that 7% of a moderately soluble gas ong>andong> 1%of a slightly soluble gas will be deposited withong>inong> 2 km from the source. These veryrough estimates are highly uncertaong>inong> ong>andong> ong>inong>dicate merely the right order ofmagnitude. For less soluble gases processes ong>inong> the surface (soil, plant) are veryimportant ong>andong> no good model results can be obtaong>inong>ed for these gases unless theconcentrations ong>inong> ong>andong> on the surface (soil, plant) is modelled as well.Deposition vs. spraydrift due to sedimentationDry ong>andong> wet deposition of highly soluble gaseous pesticides at distances of1-20 m from the field onto which pesticides are applied can be moreimportant than deposition caused by spray drift due to sedimentation for pesticidesof which 20% or more volatilises. For moderately soluble or slightly solublegaseous pesticides the dry ong>andong> wet deposition can be as important as depositioncaused by spray drift due to sedimentation for these pesticides at distances greaterthan 2 m (moderately soluble gases) or 15 m (slightly soluble gases). Another formof spray drift, i.e. the spray drift caused by small droplets is not ong>inong>vestigated much,but can be potentially more important than spray drift due to sedimentation.114


2 Discussion ong>andong> conclusionsIn section 1.5 the maong>inong> conclusions on meteorology ong>andong> surface exchange arepresented. In section 1.8 the maong>inong> conclusions on wet deposition are presentedong>inong>cludong>inong>g a comparison with dry deposition. In section 1.13 the maong>inong> conclusionson the model results are presented. These conclusions will not be repeated ong>inong> thissection, which maong>inong>ly deals with overall conclusions on the atmospheric behaviourof pesticides ong>andong> not at least with recommendations for future research.Important pesticidepropertiesUncertaong>inong>ty ong>inong> crucialpropertiesDifficult to generaliseThe atmospheric behaviour of pesticides is to a large extent governed bytheir properties, amongst which the solubility ong>inong> water (Henry’s law coefficient)ong>andong> the vapour pressure are most important. For possible reactions ong>inong> theatmosphere their chemical structure, which also determong>inong>es their solubility ong>andong>vapour pressure, are important.Modellong>inong>g the atmospheric behaviour of pesticides is hong>andong>icapped by thefact that ong>inong>formation on crucial properties of pesticides is not available or isuncertaong>inong>. E.g. a factor of 10 uncertaong>inong>ty ong>inong> a crucial variables as the Henry’s lawcoefficient or the vapour pressure is not uncommon.ong>Pesticidesong> are compounds that have ong>inong> common that they are biologicallyactive ong>andong> that they often are organic compounds. That is about the only propertiesthey have ong>inong> common. The properties that are of crucial importance to theirbehaviour ong>inong> the environment, ong>inong> particular ong>inong> the atmosphere, can vary up to 5orders ong>inong> magnitude. This ong>inong> combong>inong>ation with a large variety of possible surfaceswith agaong>inong> different properties, gives a huge number of different possiblecombong>inong>ations. For that reason it is difficult to generalise. It is even more difficult togeneralise the atmospheric behaviour because we have not enough knowledgepresently. This ong>inong> turn is also the challenge. Summarisong>inong>g it could be stated thatpesticides have one more property ong>inong> common: they are different.It should be noted here, that many hundreds of pesticides are used or have beenused ong>inong> the past. Some pesticides have been abong>andong>oned, but are still ong>inong> theenvironment, others are novel. They have all ong>inong> common that the ong>inong>formationprovided by the manufactures for the official approval procedure is not enough topredict their atmospheric behaviour. For that reason improvement of the approvalprocedure is needed ong>andong> research should be conducted to fong>inong>d easy but effectivemethods to screen the potential of pesticides to cause any harm via the atmosphericpathway.Processes related to the Surface exchange (emission, dry deposition) of gaseous pesticides dependssurface should be not only on atmospheric processes, but also on the properties of the surfaceong>inong>cluded(soil, plant) ong>andong> on the processes that take place ong>inong> the surface. The atmosphericbehaviour of pesticides cannot be modelled without takong>inong>g these processes ong>inong>toaccount. Important processes with that respect are transport ong>inong>to the surface ong>andong>degradation ong>inong> the surface. For many gaseous pesticides dry deposition will bedetermong>inong>ed more by these processes than by atmospheric processes. Emission ofpesticides from soil depends also to a large extent on soil processes. It should benoted here, that even if the pesticides are applied onto plants, part of the appliedamount reaches the soil, either directly or after havong>inong>g been washed down byong>precipitationong>.115


Moreover, heatong>inong>g of the surface ong>andong> evaporation of water from the surfacedetermong>inong>es to some extent the friction velocity ong>andong> the atmospheric stability ong>andong>has therefore ong>inong>fluence on vertical diffusion ong>andong> surface exchange. Water ong>andong> heatare also important for the surface exchange (emission, dry deposition) of pesticidesfrom the soil. Water is important for soil processes because it can replace pesticidesadsorbed onto soil particles ong>andong> can be carrier of dissolved pesticides. Heat isimportant, because the evaporation of water depends on it ong>andong> because the vapourpressure ong>andong> Henry’s law coefficient are temperature dependent. In that wayatmospheric processes ong>andong> processes ong>inong> the soil ong>andong> on plants (evaporation ofwater) are ong>inong>terrelated. In stead of treatong>inong>g atmospheric ong>andong> surface processesseparately ong>inong> models they should be ong>inong>tegrated, otherwise model results will be lessrealistic.Stable atmosphericconditionsDurong>inong>g the development ong>andong> testong>inong>g of the model it was noted that thediffusion under very stable conditions modelled with a K-model can be much lessthan the diffusion modelled with the OML-model, a Gaussian plume model. Thisillustrates probably the uncertaong>inong>ty ong>inong> the diffusion durong>inong>g these conditions. The drydeposition close to the source modelled with a K-model can then becomeextremely high. In such model situations the wong>inong>d speed near the ground isextremely low. Moreover, the vertical mixong>inong>g is highly reduced. As a result thepesticide is “hangong>inong>g around”. In such situations with a very low wong>inong>d speed mostdiffusion models are not any longer correct, because the ong>airong> flow is ong>inong> such casesmore determong>inong>ed by height differences ong>inong> the terraong>inong>, spatial differences ong>inong> the heatflux etc. than by the wong>inong>d. These situations occur, however, quite frequently ong>inong> theevenong>inong>g when the surface cools down. Under these conditions the emission is alsovery much reduced. It should for that reason be ong>inong>vestigated what this lack ofknowledge means for conclusions on the atmospheric behaviour of pesticides, e.g.can it be assumed that the emission ong>andong> consequently the dry deposition isnegligible durong>inong>g these conditions.Dry deposition of For slightly soluble gaseous pesticides is it possible to obtaong>inong> ong>inong>formation onslightly soluble gaseous the uptake rate by dry deposition from laboratory experiments. It would bepesticides from useful to consider the possibility to obtaong>inong> ong>inong>formation from laboratorylaboratory experiments measurements for these gaseous pesticides ong>inong> the future.Dry deposition close to a It was shown, that at maximum less than 25% of the emission of pesticidessource has never been can be dry deposited withong>inong> 2 km from the source. This ong>inong> case of highlymeasured soluble gases. In case of moderately soluble gases this will be of the order of 7%ong>andong> for slightly soluble gases this will be of the order of 1%. In these calculationsknowledge on atmospheric diffusion ong>andong> dry deposition of gases are combong>inong>ed.The model for atmospheric diffusion has been verified with measurements ong>andong> themodel used here for dry deposition has been verified at some distance from sourcesfor other compounds than pesticides. The modelled dry deposition as a function ofdistance from the source relatively close to a source has never been measured,although ong>inong> extreme cases about 20% of the emission could be dry deposited withong>inong>a few hundred metres from the source. It would therefore be useful to measure thisfor a highly soluble gas that is easy to measure. This could be used to obtaong>inong> anupper estimate of the deposition close to a source.Scavengong>inong>g ratio canbe measuredBy measurong>inong>g concentrations ong>inong> ong>airong> ong>andong> ong>precipitationong> simultaneously thescavengong>inong>g ratio for both gaseous ong>andong> particulate pesticides can be determong>inong>ed.This can only be done at such a distance from important sources that the pesticidehas been mixed ong>inong> such a way that the concentration measured at ground-level isrepresentative of the concentration of the ong>airong> that enters the cloud. The scavengong>inong>gcoefficient, i.e. the rate at which material is removed from the atmosphere can thenbe calculated if the mixong>inong>g height is known. For slightly soluble gases the116


scavengong>inong>g ratio should be the same as calculated from the Henry’s law coefficientwith (40).Partitionong>inong>gIt is important to determong>inong>e how much of a pesticide is ong>inong> the gas phase ong>andong> howmuch ong>inong> the particulate phase ong>andong> to obtaong>inong> ong>inong>formation on the size of the particles.The reason for this is that the rate at which pesticides are removed from theatmosphere depend to a large extent on the form or size (particles) ong>inong> which theyare present. It would be useful to try to measure this partitionong>inong>g. Moreover, itwould be useful to have more ong>inong>formation on the conversion rate from the gaseousto the particulate phase ong>andong> the factors that have an ong>inong>fluence on it.Photochemical reactions In this report it is assumed, that photochemical atmospheric reactions ofong>andong> conversion from the pesticides ong>andong> conversion of pesticides from the gaseous to the particulategaseous to the particulate phase are not important withong>inong> 2 km from a source. This is likely, but by nophase are important means proven.ong>Pesticidesong> are removed from the atmosphere by dry ong>andong> wet deposition. Both thedry ong>andong> wet removal rate of gaseous pesticides ong>inong>crease with the solubility ong>inong> water.For not very soluble gases this means that they can be transported over very longdistances, if there are no other processes that can contribute to their removal. It ishere that photochemical reactions ong>andong> conversion from the gaseous to theparticulate phase play a role. These processes do not remove compounds from theatmosphere. They do, however, lead to other products, that may be removed moreefficiently from the atmosphere. E.g. slightly soluble pesticides associated withparticles are generally removed rather efficiently from the atmosphere byong>precipitationong> than the same pesticides ong>inong> the gas phase.This illustrates that it is very important to have ong>inong>formation on the photochemicalreactions ong>andong> the conversion from the gaseous to the particulate phase (ong>andong> viceversa), because these processes limit the long-range transport of many pesticides.Need for mechanisticemission modelsDifference ong>inong> ong>inong>formationneeded forscreenong>inong>g ong>andong> atmo-Some prelimong>inong>ary methods exist to estimate the maximum or cumulativeemission of pesticides (see section 1.6). These methods are based on a statisticalcorrelation of measured emission fluxes with properties of the pesticide ong>andong> not ona mechanistic description of the processes that are goong>inong>g on. Although theseprelimong>inong>ary methods can be useful to estimate emissions ong>inong> screenong>inong>g proceduresfor pesticides, it is less useful to apply them to generate emissions ong>inong> atmospherictransport ong>andong> deposition models used to calculate the deposition as a function ofdistance to the emission field. The prong>inong>cipal reason for this is that emissions,diffusion ong>andong> dry deposition depend on the same meteorological factors(turbulence, heat flux, water vapour flux). And this is not taken ong>inong>to account ifemissions, diffusion ong>andong> dry deposition are calculated separately, e.g. if theemission is calculated for average atmospheric conditions ong>andong> the atmosphericconditions are far from average. Recent calculations for ammonia have shown thatif the meteorological conditions are such that they lead to ong>inong>creased emission, theyalso favour long-range transport (Asman et al., 1998). It is only possible to revealthis type of ong>inong>teractions if mechanistic models are applied with the samemeteorology to describe both the emission, diffusion ong>andong> dry deposition processes.More mechanistic models should be developed to describe the emission ofpesticides ong>andong> their results should be verified with already existong>inong>g or newlaboratory ong>andong> field experiments.The ong>inong>formation presented ong>inong> the previous section illustrates the fact that ong>inong>general ong>inong>formation that is sufficient for screenong>inong>g procedures ong>inong> connectionwith the approval of a pesticide not necessarily is sufficient to describe the117


spheric transport models processes needed to quantify fluxes with atmospheric transport ong>andong> depositionmodels.Measurements fordifferent purposesThere are different reasons to measure pesticides. The first reason is tomonitor, e.g. to get an impression of the compounds present ong>andong> to knowwhether the concentration is so high that it could lead to ong>effectsong>. Moreover, thistype of measurements can be used to verify the results of atmospheric transport ong>andong>deposition models. This type of measurements is necessary, but it is a kong>inong>d ofreaction on the action performed by the pesticide manufacturers ong>andong> farmers.Another reason to measure pesticides is to obtaong>inong> ong>inong>formation on processes thatoccur, e.g. processes that could be part of a model. This model could be anemission model, a deposition model or even an ong>inong>tegrated atmospheric transportong>andong> deposition model. Such models can be used for various purposes for which it isdifficult or impossible to use monitorong>inong>g:• Prediction of the behaviour of compounds that are not yet on the marked.• Estimation of import ong>andong> export of pesticides.• Interpolation of measurements ong>inong> space ong>andong> time.• Description of historical situations for which no monitorong>inong>g data are available.• Design of experiments to obtaong>inong> ong>inong>formation on processes or to design of thelocations of stations ong>inong> a monitorong>inong>g network.• Study of the effect of different scenarios, e.g. different possibilities to applypesticides, so that the best pesticide or conditions can be chosen for a givenpurpose.Field experiments, It should be noted here that models never can replace measurements enmodelexperiments ong>andong> tirely, but can provide best estimates. A good way to apply models is tolaboratory experiments design experiments. The results of the experiments can then lead to animprovement of the model, which ong>inong> turn can lead an improved measurementstrategy.The analysis of pesticides is time consumong>inong>g ong>andong> difficult, partly becauseconcentrations are so low. For that reason they are also very expensive. This meansthat experiments should be planned with much more care than usually is done ong>inong>atmospheric science. It is here that models can play a crucial role. Also laboratoryexperiments can be a relative good ong>inong>vestment, e.g. for emission, uptake by plants,photochemical reactions etc. Laboratory experiments are useful, because it ispossible to study processes under controlled conditions ong>andong> it gives the possibilityof just varyong>inong>g one factor, whereas ong>inong> the real world many factors vary at the sametime, so that it is difficult to get ong>inong>sight ong>inong> different processes that occur at the sametime. It is, however, never possible to rely fully on laboratory experiments. Onereason is that it is very difficult to create the same turbulence ong>inong> the laboratory as ong>inong>the atmosphere as ong>inong> the field. An good research strategy should have an optimalbalance between field experiments, laboratory experiments ong>andong> model developmentong>andong> calculations.Process researchong>Pesticidesong> come ong>inong> the atmosphere by emission, are diffused, react ong>andong> deposit (ong>inong>that order). There is not so much known on the atmospheric behaviour ofpesticides. It would for that reason be advisable to start explorong>inong>g their behaviourby studyong>inong>g separate processes ong>andong> not the result of more than one process. Itwould also be advisable to pay much attention to the emission process, becausestudies of the other processes would benefit from ong>inong>formation on the first processong>inong> the chaong>inong>. Moreover, it is useful to know whether or not the pesticide is emittedto the atmosphere. If not, further atmospheric research is not necessary.118


Spray driftThe deposition of tong>inong>y spray droplets that do not deposit due to sedimentationshould be ong>inong>vestigated.Time perspective It took about 20 years before atmospheric scientists hadenough knowledge to understong>andong> the atmospheric behaviour of sulphur dioxide. Ittook then only about 10 years before the more complicated behaviour of nitrogenoxides were understood, because knowledge gaong>inong>ed durong>inong>g research on sulphurdioxide could be used. These compounds occur ong>inong> concentrations that are a factorof 1000 higher than pesticides. Much attention was paid to just a few compounds.There are hundreds of pesticides around ong>inong> the atmosphere, they have very differentproperties ong>andong> they occur ong>inong> very low concentrations. For that reason it is a far fromeasy task to study them, ong>andong> the study will therefore take much time.119


3 AcknowledgementI thank Helle Vibeke Andersen (National Environmental Research Institute,Roskilde, Denmark) ong>andong> Peter Kryger Jensen (Danish Institute of AgriculturalSciences, Flakkebjerg, Denmark) for their comments on the draft version of thisreport. I am grateful to Addo van Pul (National Institute of Public Health ong>andong> theEnvironment, RIVM, Bilthoven, The Netherlong>andong>s) ong>andong> Erik van den Berg (DLOWong>inong>ong>andong> Starong>inong>g Centre for Integrated Long>andong>, Soil ong>andong> Water Research, Wagenong>inong>gen,The Netherlong>andong>s) for the discussions on the behaviour of pesticides ong>inong> theatmosphere ong>andong> ong>inong> the soil durong>inong>g my stay ong>inong> The Netherlong>andong>s ong>inong> 1998.120


4 ReferencesArya, S.P. (1988). Introduction to micrometeorology. Academic Press, San Diego,U.S.A.Asman, W.A.H. (1995). Parameterization of below-cloud scavengong>inong>g of highlysoluble gases under convective conditions. Atmospheric Environment 29, 1359-1368.Asman, W.A.H. (1998). Factors ong>inong>fluencong>inong>g local dry deposition of gases withspecial reference to ammonia. Atmospheric Environment 32, 415-421.Asman, W.A.H., Cellier, P., Génermont, S., Hutchong>inong>gs, N.J., Sommer, S.G. (1998).Ammonia emission research: from emission factors to process descriptions.EUROTRAC Newsletter 20, 2-10.Asman, W.A.H., Jensen, P.K. (1993). Processer for våddeposition (Wet depositionprocesses, ong>inong> Danish). Report No. 26, Havforsknong>inong>g fra Miljøstyrelsen. DanishEnvironmental Protection Agency, Copenhagen, Denmark.Asman, W.A.H., Sørensen, L.-L., Berkowicz, R., Granby, K., Nielsen, H., Jensen,B., Runger, E.H., Lykkelund, C., Grynong>inong>g, S.-E., Sempreviva, A.M. (1994).Processer for tørdeposition (Dry deposition processes, ong>inong> Danish). Report 35,Havforskong>inong>g fra Miljøstyrelsen, Danish Environmental Protection Agency,Copenhagen, Denmark.Atkong>inong>son, R. (1987). A structure reactivity relationship for the estimation of rateconstants for the gas phase reactions of OH radicals with organic compounds. Int.J. Chem. Kong>inong>et. 19, 799-828.Atkong>inong>son, R. (1988). Estimation of gas-phase hydroxyl radical rate constants fororganic chemicals. Environ. Toxicol. Chem. 7, 435-442.Best, A.C. (1950). The size distribution of raong>inong>drops. Quart. J. R. Met. Soc. 76, 16-36.Brown M.J., Arya, S.P., Schnyder, W.H. (1993). Vertical dispersion from surfaceong>andong> elevated releases: An ong>inong>vestigation of a non-Gaussian plume model. J. Appl.Met. 32, 490-505.Cotton, W.R., Anthes, R.A. (1989). Storm ong>andong> cloud dynamics. Academic Press,San Diego, U.S.A.Duyzer, J.H., van Oss, R.F. (1997). Determong>inong>ation of deposition parameters of anumber of persistent organic pollutants by laboratory experiments. Report R97/150, TNO-MEP, Apeldoorn, The Netherlong>andong>s.Erisman J.W., van Pul, A., Wyers, P. (1994). Parameterization of surfaceresistance for the quantification of atmospheric deposition ong>andong> ozone. AtmosphericEnvironment 28, 2595-1607.Ganzelmeier, H. , Rautmann, D., Spangenberg, R., Streloke, M., Hermann, M.,Wenzelburger, H.J., Walter, H.-F. (1995). Untersuchungen zur Abtrift von121


Pflanzenschutzmitteln im Freilong>andong>, Richtlong>inong>ien für die Prüfung von Pflanzenschutz.Mitteilungen aus der Biologischen Bundesanstalt f¨r Long>andong>. und Forstwirtschaft. Heft304, Berlong>inong>, Germany.Green, R.E., Karickhoff, S.W. (1990). Sorption estimates for modelong>inong>g. In: ChengH.H. (editor) ong>Pesticidesong> ong>inong> the soil environment: processes, impacts, ong>andong> modelong>inong>g.Soil Science Society of America, Madison, Wisconsong>inong>, U.S.A., 79-101.Grynong>inong>g, S.E., van Ulden, A.P., Larsen, S.E. (1983). Dispersion from a contong>inong>uousground-level source ong>inong>vestigated by a K model. Quart. J. R. Met. Soc. 109, 355-364.Hicks, B.B., Baldocchi, D.D., Meyers, T.P., Hosker, R.P., Matt, D.R. (1987). Aprelimong>inong>ary multiple resistance routong>inong>e for derivong>inong>g dry deposition velocities frommeasured quantities. Wat. Air soil Pollut. 36, 311-330.Junge, C.E. (1977). Basic considerations about trace constituents ong>inong> theatmosphere. In: Suffet, I.H. (Ed.) Fate of pollutants ong>inong> the ong>airong> ong>andong> waterenvironment. Part 1. Advances ong>inong> Environmental Science ong>andong> Technology, vol 8.Wiley-Interscience, New York, U.S.A., 7-25.Jury, W.A., Spencer, W.F., Farmer, W.J. (1983). Behavior assessment models fortrace organics ong>inong> soil: I. Model description. J. Environ. Qual. 12, 558-564. (Erratasee: J. Environ. Qual. 16, 448).Jury, W.A., Farmer, W.J., Spencer, W.F. (1984). Behavior assessment models fortrace organics ong>inong> soil: II.Chemical classification ong>andong> parameter sensitivity. J.Environ. Qual. 12, 558-564.Long>inong>dfors, V., Joffre, S.M., Damski, J. (1991). Determong>inong>ation of the wet ong>andong> drydeposition of sulphur ong>andong> nitrogen compounds over the Baltic Sea usong>inong>g actualmeteorological data. FMI Contr. No. 4, Fong>inong>nish Meteorological Institute, Helsong>inong>ki,Fong>inong>long>andong>.Liss, P.S., Slong>inong>n, W.G.N. (Eds.) Air-sea exchange of gases ong>andong> particles. Reidel,Dordrecht, The Netherlong>andong>s.Lyman, W.J., Reehl,W.F., Rosenblatt,D.H. (1990). Hong>andong>book of chemical propertyestimation methods. American Chemical Society, Washong>inong>gton DC, U.S.A.Marshall, J.S., Palmer, W.M. (1948). The distribution of raong>inong>drops with size. J.Meteor. 5, 165-166.Pasquill, F., Smith, F.B. (1983). Atmospheric diffusion. Ellis Horwood, Chichester,U.K.Pruppacher, H.R., Klett, J.D. (1997). Microphysics of clouds ong>andong> ong>precipitationong>.Kluwer, Dordrecht, The Netherlong>andong>s.Ruijgrok, W., Davidson, C.I., Nicholson, K.W. (1995). Dry deposition of particles.Tellus 47B, 587-601.Seong>inong>feld, J.H (1986). Atmospheric chemistry ong>andong> physics of ong>airong> pollution. JohnWiley, New York, USA.122


Slong>inong>n, W.G.N. (1983). Air-to-sea transfer of particles. In: Liss, P.S., Slong>inong>n, W.G.N.(Eds.):Air-sea exchange of gases ong>andong> particles. Reidel, Dordrecht, TheNetherlong>andong>s, 299-405.Slong>inong>n, S.A., Slong>inong>n, W.G.N. (1980). Predictions for particle deposition on naturalwaters. Atmospheric Environment 15, 123-129.Smit, A.A.M.F.R., Leistra, M., van den Berg, F. (1998). Estimation method for thevolatilization of pesticides from plants. Report 4, Environmental Plannong>inong>g Bureauseries 2, DLO Wong>inong>ong>andong> Starong>inong>g Centre, Wagenong>inong>gen, The Netherlong>andong>s.Smit, A.A.M.F.R., van den Berg, F., Leistra, M. (1997). Estimation method for thevolatilization of pesticides from fallow soil. Report 2, Environmental Plannong>inong>gBureau series 2, DLO Wong>inong>ong>andong> Starong>inong>g Centre, Wagenong>inong>gen, The Netherlong>andong>s.Spencer, W.F., Cliath, M.M., Jury, W.A., Zhang, L.-Z. (1988). Volatilization oforganic chemicals from soil as related to their Henry’s law coefficients. J. Environ.Qual. 17, 504-509.Stork, A. (1995). Wong>inong>dkanalanlage zur Bestimmung gasförmiger Verluste vonUmweltchemikalien aus dem System Boden/Pflanze unter praxisgerechtenBedong>inong>gungen mit direkten luftanalytischen Methoden unter Nutzung der 14 C-Traceranalytik. Ph. D. Thesis Rheong>inong>ischen Friedrich-Wilhelms-Universität, Bonn,Germany.Stull, R.B. (1988). An ong>inong>troduction to boundary layer meteorology. Kluwer,Dordrecht, The Netherlong>andong>s.Van Jaarsveld, J.A. (1995). Modellong>inong>g the long-term atmospheric behaviour ofpollutants on various spatial scales. Ph.D. thesis, University of Utrecht, TheNetherlong>andong>s.Van Jaarsveld, J.A. (1996).The dynamic exchange of pollutants at the ong>airong>-soilong>inong>terface ong>andong> its impact on long range transport. In: Grynong>inong>g, S.E., Schiermeier,F.A. (1996) Air pollution modelong>inong>g ong>andong> its application XI, Plenum Press, NewYork, U.S.A., 333-344.Van Pul, W.A.J., de Leeuw, F.A.A.M., van Jaarsveld, J.A., van der Gaag, M.A.,Sliggers, C.J. (1998). The potential for long-range transboundary atmospherictransport. Chemosphere 37, 113-141.Van Ulden, A.P. (1978). Simple estimates for vertical diffusion from sources nearthe ground. Atmospheric Environment 12, 2125-2129.Wania, F. Mackay, D. (1995). A global distribution model for persistent organicchemicals, Sci. Total Environ. 160/161, 211-232.Whitby, K.T. (1978). The physical characteristics of sulphur aerosols.Atmospheric Environment 12, 135-159.Wierong>inong>ga, J., Rijkoort, P.J. (1983). Wong>inong>dklimaat van Nederlong>andong> (Wong>inong>d climate ong>inong>The Netherlong>andong>s, ong>inong> Dutch, with summary ong>inong> English). Staatsuitgeverij, Den Haag,The Netherlong>andong>s.123


Wong>inong>er, A.M., Atkong>inong>son, R. (1990). Atmospheric reaction pathways ong>andong> lifetimesfor organophosphorus compounds. In: Kurtz, D.A. (Ed.) Long range transport ofpesticides. Lewis Publishers, Chelsea, Michigan, U.S.A., 115-126.Woodrow, J.E., Seiber, J.N., Baker, L.W. (1997). Correlation techniques forestimatong>inong>g pesticide volatilization flux ong>andong> downwong>inong>d concentrations. Environ. Sci.Technol. 31, 523-529.124


Appendix ID g * : the apparent diffusivity of the gas ong>inong> the gas phaseD* g =Dg4Dg1+rαvggwhere:*D g = apparent diffusivity of the gas ong>inong>the gas phase (m 2 s -1 ).D g= diffusivity of the gas ong>inong> the gasphase (m 2 s -1 ).α g= accomodation coefficient(dimensionless) which gives the probability that the gas molecule that hits thesurface of the drop is absorbed ong>andong> v gis the average speed of the gas molecules(m s -1 ) defong>inong>ed by:vg=8RTπMgwhere:R = gas constant ( 8.317 J °K -1 mol -1 )T= temperature (°K)M g = molecular mass of the gas (kg mol -1 )If α g is larger than about 0.01, D g * is approximately equal to D g .125


Appendix IICharacteristic time ong>andong> distance of absorptionThe characteristic time ong>andong> distance of absorption can be calculated from TableA2-1 ong>andong> the Henry’s law coefficient usong>inong>g equations (34) or (38).τ abs can be found from:τ absfactor1=KH∆ abs can be found from:∆ absfactor2=KHwhere K H = Henry’s law coeffcient (c gas /c water ) (dimensionless).Table A2-1Properties of drops ong>andong> factors that can be used to fong>inong>d the characteristic timeconstant of absorption τ abs ong>andong> the characteristic distance of absorption ∆ abs . Themolecular mass of the gas is 300, which gives an estimated diffusivity ong>inong> ong>airong> of6.062×10 -6 m 2 s -1 .Egenskaber for dråber og faktorer som kan anvendes til at fong>inong>de den karakterisketidskonstant for absorption τ abs og den karakteristiske afstong>andong> for absorption ∆ abs .Gassens molekylmasse er 300, som giver en estimeret diffusivitet i luft på6.062×10 -6 m 2 s -1 .RadiusTermong>inong>alvelocity v t(m s -1 )f g(dim.less)factor1 =r 2 /(3f g D g )(s)factor2 =(r 2 v t /(3f g D g )(m)(m)1×10 -6 1.328×10 -4 1.000 5.499×10 -8 7.303×10 -122×10 -6 5.118×10 -4 1.000 2.199×10 -7 1.126×10 -103×10 -6 1.137×10 -3 1.000 4.948×10 -7 5.626×10 -104×10 -6 2.008×10 -3 1.000 8.796×10 -7 1.766×10 -95×10 -6 3.126×10 -3 1.000 1.374×10 -6 4.295×10 -96×10 -6 4.489×10 -3 1.001 1.978×10 -6 8.880×10 -97×10 -6 6.099×10 -3 1.001 2.691×10 -6 1.641×10 -88×10 -6 7.955×10 -3 1.002 3.513×10 -6 2.795×10 -89×10 -6 1.006×10 -2 1.002 4.443×10 -6 4.469×10 -81×10 -5 1.237×10 -2 1.003 5.480×10 -6 6.778×10 -82×10 -5 4.834×10 -2 1.026 2.144×10 -5 1.036×10 -63×10 -5 1.043×10 -1 1.084 4.565×10 -5 4.763×10 -64×10 -5 1.745×10 -1 1.187 7.409×10 -5 1.293×10 -55×10 -5 2.539×10 -1 1.327 1.036×10 -4 2.630×10 -56×10 -5 3.390×10 -1 1.473 1.344×10 -4 4.557×10 -57×10 -5 4.273×10 -1 1.620 1.663×10 -4 7.107×10 -5126


8×10 -5 5.173×10 -1 1.768 1.991×10 -4 1.030×10 -49×10 -5 6.081×10 -1 1.916 2.325×10 -4 1.414×10 -41×10 -4 6.991×10 -1 2.064 2.664×10 -4 1.862×10 -42×10 -4 1.587 3.516 6.255×10 -4 9.930×10 -43×10 -4 2.438 4.933 1.003×10 -3 2.446×10 -34×10 -4 3.247 6.314 1.393×10 -3 4.524×10 -35×10 -4 3.983 7.633 1.801×10 -3 7.173×10 -36×10 -4 4.608 8.854 2.236×10 -3 1.030×10 -27×10 -4 5.135 9.987 2.698×10 -3 1.385×10 -28×10 -4 5.616 1.107×10 1 3.178×10 -3 1.785×10 -29×10 -4 6.060 1.212×10 1 3.675×10 -3 2.227×10 -21×10 -3 6.468 1.313×10 1 4.188×10 -3 2.709×10 -21×10 -3 7.982 1.758×10 1 7.036×10 -3 5.616×10 -22×10 -3 8.723 2.106×10 1 1.044×10 -2 9.109×10 -22×10 -3 8.976 2.378×10 1 1.445×10 -2 1.297×10 -13×10 -3 9.010 2.603×10 1 1.901×10 -2 1.713×10 -13×10 -3 9.006 2.804×10 1 2.402×10 -2 2.163×10 -1Note: This example is for a gas with a molecular mass of 300. If a gas with amolecular mass of 200 were chosen f g would have been the same for small dropsong>andong> about 7% less for the larger drops; factor1 ong>andong> factor2 would have been 13-19% less. If a gas with a molecular mass of 400 were chosen, f g would have beenthe same for small drops ong>andong> about 5% larger for the larger drops; factor1 ong>andong>factor2 would have been 10-15% larger. The conclusion is that these factors do notdepend so much on the diffusivity D g .127


Appendix IIIParametrization of the below-cloud scavengong>inong>g coefficient for highly solublegasesAsman (1995) made a detailed model for below-cloud scavengong>inong>g underconvective conditions that ong>inong>cludes the evaporation of the droplets. The model canhong>andong>le different descriptions of the raong>inong>drop size distribution. The results of thismodel for the raong>inong>drop size distribution of Best (1950) were parameterized with thedrop size distribution ong>inong> such a way that they easily could be ong>inong>corporated ong>inong>atmospheric transport models. The only ong>inong>put parameters necessary are the raong>inong>fallrate, temperature ong>andong> relative humidity at ground-level ong>andong> the diffusivity of thegas at 25°C. The parameterization of the below-cloud scavenong>inong>g coefficient is:Λ bbav= aIwhere:Λ b = below cloud scavengong>inong>g coefficient (s -1 )I mm = raong>inong>fall rate at ground level (mm hr -1 )In this equation a ong>andong> bav are coefficients that are functions of the relative humidity atground level (rh(0), ong>inong> %) ong>andong> the temperature at ground level (T(0), ong>inong> °K) ong>andong> of thediffusivity of the gas at 25°C ong>andong> 1 atm (D g ong>inong> m 2 s -1 ). If D g is not known it can beestimated from the molecular weight of the gas ong>andong> (15).The value of a is found from the followong>inong>g set of equations:a = aa + bb D gwhere:aa = a0 + a1 rh(0)bb = b0 + b1 rh(0)with the followong>inong>g values of the coefficients:a0 = 4.476x10 -5 - 1.347x10 -7 T(0)a1 = -3.004x10 -7 + 1.498x10 -9 T(0)b0 = 8.717 - 2.787x10 -2 T(0)b1 = -5.074x10 -2 + 2.894x10 -4 T(0)The value of bav is found from the followong>inong>g set of equation:bav = bav0 + bav1 rh(0)where:bav0 = 9.016x10 -2 + 2.315x10 -3 T(0)bav1 = 4.458x10 -3 - 2.115x10 -5 T(0)The functions are here given with greater accuracy than actually known, to avoid anyroundong>inong>g off. This parameterization is made for below-cloud scavengong>inong>g underconvective conditions. This parameterization can, however, also be used to get anestimate of the below-cloud scavengong>inong>g coefficient under non-convective conditions.128


ong>Pesticidesong> ong>inong> ong>precipitationong>Gitte Feldong>inong>gGeological Survey of Denmark ong>andong> Greenlong>andong>Department of GeochemistryNiels Henrik SpliidDanish Institute of Agricultural SciencesDepartment of Crop ProtectionWillem A.H. AsmanNational Environmental Research InstituteDepartment of Atmospheric Environment129


130


Table of contentsSUMMARY 137DANSK SAMMENDRAG 1381 INTRODUCTION AND BACKGROUND 1392 INTERNATIONAL FINDINGS 1403 MATERIALS AND METHODS 1423.1 SAMPLING AND ANALYSIS OF RAINWATER 1423.2 THE PESTICIDES 1434 RESULTS 1445 DISCUSSION AND CONCLUSION 1606 ACKNOWLEDGEMENTS 1627 REFERENCES 163APPENDIX I 165APPENDIX II 166APPENDIX III 167APPENDIX IV 169131


132


Summaryong>Pesticidesong> ong>inong>ong>precipitationong>The concentration of pesticides ong>inong> ong>precipitationong> was ong>inong>vestigated durong>inong>g theperiod May 1996 - December 1998 at three sites on Zealong>andong> ong>inong> Denmark(Gadevang, Gisselfeld ong>andong> Lorup). These sites had not been directly affected bylocal emissions. The samples were analysed for isoproturon (ong>inong>cludong>inong>g 2metabolites), the phenoxyalkanoic acid herbicides MCPA, mecoprop, ong>andong>dichlorprop plus bentazone (ong>inong>cludong>inong>g 2 metabolites), ong>andong> DNOC.Effects of pesticidesong>Pesticidesong> ong>inong> ong>precipitationong>when applicatedThis project was part of a larger project, which studied the ong>effectsong> of pesticides ong>inong>ong>precipitationong> on plants ong>andong> plant ecosystems. The highest measured concentrationswere 0.9µg/L for isoproturon ong>andong> 0.6µg/L for phenoxyalkanoic acid herbicides. Inmost cases the concentrations ong>inong>ong>precipitationong> were found at times when the pesticides were known to beapplied to crops. Combong>inong>ed samples for these three sites for the period September1996 - November 1997 were analysed for 44 compounds. Concentrations over thedetection limit were only found for isoproturon, metamitron, DNOC (2-methyl-4,6-dong>inong>itrophenol), mecoprop, methabenzthiazuron, 2-hydroxyterbuthylazong>inong>e,terbuthylazong>inong>e ong>andong> 2,4-D.DNOC ong>inong> ong>precipitationong> Unexpectedly high concentrations of DNOC (0.38 - 4.5µg/L) were foundcaused by photochemical durong>inong>g the whole samplong>inong>g period. Although DNOC has not been applied ong>inong>reactionsDenmark song>inong>ce 1986, it has been detected ong>inong> other ong>inong>vestigations ong>inong> the top layersof the ground water ong>andong> ong>inong> streams. Current literature ong>inong>dicates that DNOC is likelyto be formed by photochemical reactions of toluene ong>andong> nitrogen oxides ong>inong> theatmosphere probably from traffic pollution. The atmospheric deposition of DNOC,mecoprop ong>andong> isoproturon is respectively 7.5-, 0.3- ong>andong> 0.3g/ha/year.Generally, pesticides will be transported over distances of more than severalhundred kilometers before they are deposited, unless they are (photochemically)degraded ong>inong> the atmosphere at a high rate.133


Dansk sammendragDen danske nedbørs eventuelle ong>inong>dhold af pesticider er undersøgt ved opsamlong>inong>g afregnvong>andong>sprøver fra lokaliteter, som er beliggende således, at et lokalt bidrag ermong>inong>imeret.Pesticider i nedbør fraskovlysnong>inong>gerPesticider fundet isprøjtesæsonenDNOC i nedbør stammerfra atmosfæriskeprocesserDer er udtaget nedbørsprøver fra lysnong>inong>ger i skove på 3 sjællong>andong>skelokaliteter Gadevang, Gisselfeld og Lorup, som i perioden 1996-1997 er blevetanalyseret for phenoxysyrerne: MCPA, mechlorprop og dichlorprop samt forisoproturon. I 1998 er analyseprogrammet udvidet, idet bentazon og 2 metabolitterheraf, DNOC samt 2 metabolitter af isoproturon er medtaget.Den højeste koncentration af phenoxysyrerne var på 0,6µg/L. Forisoproturons vedkommende var den maksimale koncentration 0,9µg/L. I langt defleste tilfælde var der sammenfald mellem det tidspunkt, hvor herbiciderne blevpåvist i nedbøren og anvendelsestidspunktet. DNOC er påvist i alle de vong>andong>prøver,der er udtaget i 1998 i koncentrationsområdet fra 0,04µg/L til0,87µg/L.Phenoxysyrerne forekommer på alle tre lokaliteter forår og efterår 1996. I1997 er der kun fundet mechlorprop i en enkelt prøve i sprøjteperioden omefteråret og i 1998 er der ikke påvist phenoxysyrer. Isoproturon er påvist efterår1996 og 1997 på alle tre lokaliteter og forår og efterår 1998 på alle tre lokaliteter.Det er ikke tilladt at anvende isoproturon om foråret i Danmark, så der kan væretale om langtransport fra long>andong>e, hvor sprøjtnong>inong>g er tilladt eller eventuelt sprøjtnong>inong>g iDanmark uden for efterårsperioden, hvor det har været tilladt.Derudover er der i perioden september 1996 til november 1997 udtaget 13 prøvertil analyse for de 44 stoffer, som ong>inong>dgår i DMU´s analysemetode, heraf er de 8påvist. Det er følgende pesticidkemikalier: isoproturon,metamitron, DNOC (2-methyl-4,6-dong>inong>itrophenol), mechlorprop,methabenzthiazuron, 2-hydroxyterbuthylazong>inong>, terbuthylazong>inong> og 2,4-D. Det,der især overraskede, var ong>inong>dholdet af DNOC, som blev fundet igennem heleperioden og i et forholdsvis højt koncentrationsområde fra 0,38µg/l til 4,5µg/l.Stoffet har ikke været anvendt i Danmark de sidste 10 år, så her er formentlig taleom en mere global forurenong>inong>g, song>andong>synligvis stammende fra atmosfærekemiskeprocesser forårsaget af traffikkens forurenong>inong>g. Beregnes belastnong>inong>gen i g pr. ha iovennævnte periode, er belastnong>inong>gen af DNOC 7,5 g pr. ha, mechlorprop ogisoproturon tilfører begge ca. 0,3 g pr. ha svarende til henholdsvis 88% og 8% afden totale belastnong>inong>g.134


1 Introduction ong>andong> backgroundTransport of pesticidesong>inong> the atmosphere dependson atmosphericlifetimeong>Pesticidesong> that evaporate can be transported ong>inong> the atmosphere before theyare deposited on the surface agaong>inong>. How far a substance can be transporteddepends on its atmospheric lifetime, which is determong>inong>ed by how fast thecompound is removed from the atmosphere caused by reaction, dry deposition, ong>andong>wet deposition.In a co-operation project withong>inong> the framework of Nordic Council of Mong>inong>istersfrom 1992 to 1994, raong>inong>water samples from 2 places ong>inong> Denmark, respectively fromUlborg plantation 10 km from the Jutlong>andong>ic west coast (56 o 17’ N , 8 o 26’ E) ong>andong>from Gadevang ong>inong> Gribskov were gathered. The content of the samples wasanalysed for the followong>inong>g 10 pesticides: propiconazole, prochloraz, lambdacyhalothrong>inong>,cypermethrong>inong>, esfenvalerate, deltamethrong>inong>, atrazong>inong>e, mecoprop,dichlorprop ong>andong> MCPA. Only phenoxyalkanoic acid herbicides were found. Themaximum concentrations were 0.4µg/L (Kirknel ong>andong> Feldong>inong>g, 1995).Long>inong>dane ong>inong> ong>precipitationong> In 1990-1991, the National Environmental Institute ong>inong> Denmark measured thecontents of α-HCH ong>andong> γ-HCH (long>inong>dane) ong>inong> raong>inong>water on 2 localities ong>inong> DenmarkHusby (56 o 17’ N , 8 o 8’ E) ong>andong> Ulborg ong>inong> West Jutlong>andong> respectively. Threelocalities, Ulborg, Bagenkop ong>andong> Anholt, respectively were tested ong>inong> 1992. Themaximum concentration found was 0.1µg/L, The conclusion was, that fong>inong>dong>inong>gs oflong>inong>dane was due to their use ong>inong> countries South ong>andong> West of Denmark (Cleemann etal., 1995).Wet deposition of pesti- In the present study, wet deposition of pesticides was ong>inong>vestigated.cides ong>inong>vestigated Precipitation samples have been gathered over a period of 3 years from 1996-1998from 3 localities on Zeelong>andong>: Gadevang, Gisselfeld ong>andong> Lorup, which have beenanalysed for the phenoxyalkanoic acid herbicides: MCPA, mecoprop ong>andong>dichlorprop together with isoproturon. In 1998 the program was expong>andong>ed, whenbentazone ong>andong> 2 of its metabolites, DNOC, ong>andong> 2 metabolites of isoproturon wereong>inong>cluded.44 pesticides ong>inong> analyti- 13 water samples, taken ong>inong> the period from September 1996 to Novembercal programme 1997, from the 3 above mentioned localities, were analysed for 44 differentpesticide compounds.135


2 International fong>inong>dong>inong>gsA short summary about the fong>inong>dong>inong>gs of specific pesticides ong>inong> ong>precipitationong> ong>inong>different countries is given below.USA (U.S. Geological Survey, 1995). U.S. Geological Survey has gathered theresults from 132 studies of pesticide fong>inong>dong>inong>gs ong>inong> ong>airong> ong>andong> ong>inong> ong>precipitationong> ong>inong>Chlorong>inong>ated ong>inong>secticides the United States durong>inong>g 30 years ong>inong> a summary report. The studies show:ong>inong> USA- that most of the pesticides beong>inong>g analysed for were found ong>andong> that the fong>inong>dong>inong>gsrepresent many different chemical groups of pesticides,- that pesticides have been found ong>inong> the atmosphere ong>inong> all regions ong>inong> the UnitedStates,- that the highest concentrations of pesticides are found ong>inong> the sprayong>inong>g season,- ong>andong> that certaong>inong> slowly degradable pesticides are found throughout the year ong>inong>low concentrations ong>inong> the atmosphere. The most frequently found pesticidesare the ong>inong>secticides; DDT, HCH, heptachlor ong>andong> dieldrong>inong>e. The annual wetdeposition of pesticides is generally found to be less than 1% of the amount ofpesticides used ong>inong> the region concerned.Phenoxyacid herbicides Norway (Lode et al., 1995). In Norway durong>inong>g 1992-1993, 520 tests ofong>inong> Norway, Sweden raong>inong>water were conducted. Phenoxyalkanoic acid herbicides MCPA ong>andong>ong>andong> Fong>inong>long>andong>dichlorprop were analysed among others. The maximum concentrations found wererespectively 0.32 µg/L ong>andong> 0.25 µg/L. ong>Pesticidesong> have been found ong>inong> every tenthsample. Regardong>inong>g phenoxyalkanoic acid herbicides a correlation was foundbetween the fong>inong>dong>inong>gs of the pesticides ong>inong> the ong>precipitationong> ong>andong> the sprayong>inong>g season.Sweden (Kreuger et al., 1995). From 1990 to 1992, 18 pesticides were detected.Phenoxyalkanoic acid herbicides were found most frequently (MCPA ong>inong> 35% ofthe samples, max. concentration 0.24 µg/L). For MCPA the load ong>inong> g/ha/year wasfrom 0.0001- 0.092.Fong>inong>long>andong> (Hirvi et al., 1995). From 1991-1992, the largest measured concentrationof phenoxyalkanoic acid herbicides ong>inong> 22 samples, were 0.19µg/L for dichlorprop.Isoproturon, atrazong>inong>e, Germany (Hüskes ong>andong> Levsen, 1997). In 1992, 40 samples were gatheredlong>inong>dane terbuthylazong>inong>e from the area around Hannover, ong>andong> they were analysed for 59 pesticides, 11ong>andong> others ong>inong> Germany pesticides were found ong>inong> more than 10 of the samples. The highest concentrationswere due to terbuthylazong>inong>e. There is a correlation between the fong>inong>dong>inong>gs ong>inong> thesamples ong>andong> the sprayong>inong>g season as demonstrated ong>inong> other projects. In 1995, Besteret al. also detected terbuthylazong>inong>e. In 1994, Siebers et al., published a study fromNorth-Germany, where they analysed for 11 pesticides: isoproturon, atrazong>inong>e,long>inong>dane ong>andong> terbuthylazong>inong>e, ong>andong> others. The pesticides were maong>inong>ly detected ong>inong> thesprayong>inong>g season. The highest concentrations found were about 0.7µg/L. The load ofisoproturon on the soil amounts to about 0.1g/ha/år. Jeaschke et al. also found IPUong>inong> samples taken ong>inong> the area around Frankfurt ong>inong> 1995.MCPA ong>inong> ItalyItaly (Trevisan et al.,1993). In 1988, 166 ong>airong> ong>andong> ong>precipitationong> samples weregathered ong>andong> analysed for 11 pesticides; MCPA among others, (max. concentration0.3µg/L), ong>inong> 49 of the 166 samples at least one of the 11 pesticides was present.136


Long>inong>dane ong>inong> IndiaIndia (Dua et al., 1994). In the period from January to September 1992, sampleswere analysed for HCH, which constitutes 55% of the pesticide consumption (1984figures). The average concentration was 0.077µg/L. The largest concentrationswere measured ong>inong> the sprayong>inong>g season.Japan (Suzuki 1996). From 1989-1992, samples were analysed for 9 pesticides,most of them were only found ong>inong> the sprayong>inong>g season.137


3 Materials ong>andong> methodsThe localities were chosen so they were at some distances from agricultural areas.They are all situated ong>inong> a clearong>inong>g ong>inong> a forest.Three localities awayfrom agricultural areasThe Gadevang locality (12°16.29´Ø, 55°58.01´N) is placed about 1 kmWest, 5 km South, 5 km East ong>andong> more than 10 km North of the nearest agriculturalarea. The Lorup locality (11°29.86´Ø, 55°23.24´N) is situated about 0.5 km ong>inong> alldirections from smaller fields, but 2-3 km from larger connected agricultural areasong>inong> all directions. The Gisselfeld locality (11°55.20´Ø,55°15.50´N) is placed about 1km West, 3 km East ong>andong> 2,5 km North of sprayed areas. Samples were taken ong>inong>March/April, when the first sprayong>inong>g is conducted ong>andong> until November/December,when the last sprayong>inong>g is done before it begong>inong>s to freeze.3.1 Samplong>inong>g ong>andong> analysis of raong>inong>waterThe raong>inong>water is sampled through a glass funnel (20 cm ong>inong> diameter), which isplaced about 2 meters above ground. From the funnel, the water runs through aTeflon tube ong>inong>to a 2 liter glass bottle, which is ong>inong>sulated to prevent fluctuations ong>inong>the temperature ong>andong> photochemical reactions. Four bottles were placed at eachlocality, each collects water from two funnels to get a sample that is large enoughfor analysis. The bottles were acidified to2 weeks collection periods avoid/delay microbiological degradation ofwhatever pesticides, collected with the raong>inong>water. The samples were normallycollected after 2 weeks. In the laboratory the samples were kept at approximately –18 o C until sample preparation ong>andong> analysis.With the described experimental design it was not possible to avoid degradation ofthe pesticides from the time they were collected ong>inong> the bottle to the time theyarrived ong>inong> the laboratory. Pesticide residues due to raong>inong>fall ong>inong> the begong>inong>nong>inong>g of thecollection period will result ong>inong> lower concentrations due to degradation comparedto pesticides from raong>inong>fall happenong>inong>g just before the samples are collected ong>andong>brought to the laboratory.Mong>inong>imum concentrations Therefore the concentrations listed ong>inong> the tables ong>inong> the result section, aredependong>inong>g on storage mong>inong>imum concentrations. Stability tests were performed to study the keepong>inong>gperiodes on locations qualities of the raong>inong>water samples when they were frozen before extraction. Theamount of pesticides beong>inong>g degraded at -18°C, is negligible compared to theamount degraded ong>inong> the collection period.Solid phase extraction The method for analysong>inong>g isoproturon ong>inong> the samples is based on a prefollowedby gas- or concentration by solid phase extraction. The phenoxyalkanoic acidliquid chromatography herbicides ong>inong> the samples are also determong>inong>ed by a concentration step withwith mass spectrometric solid phase extraction followed by derivatization withdetectionpentafluorobenzylbromide, which produces the correspondong>inong>gpentafluorobenzylesters. The samples are analysed with GC-MS ong>inong> SIM ong>andong> SCANmode (see appendix I-III). The water samples collected ong>inong> 1998 are analysed withthe use of LC-MS (see appendix IV). With this method it is not necessary toderivatize the phenoxyalkanoic acid herbicides. The detection limit ong>inong> usong>inong>g theLC-MS has been 0.01µg/L. Together with the raong>inong>water samples, samples spikedwith the pesticides concerned, were analysed.138


For the GC-MS-analysis, several water samples were spiked ong>inong> the concentrationrange 0.05µg/L-1.00µg/L, r 2 values were ≥0.95 ong>andong> the detection limits differedfrom 0.01µg/L ong>andong> 0.09µg/L. The detection limits were set (Miller ong>andong> Miller,1988), every time a set of samples was analysed by quantifyong>inong>g the spiked samplestogether with the collected raong>inong> water samples. The samples were scanned toidentify, if it was isoproturon, mecoprop, MCPA or dichlorprop etc. The resultsfrom the tests were transferred to SAS, which calculates the detection limits ong>andong>the concentrations of pesticides ong>inong> the samples based on the calibration data.3.2 The pesticidesMCPA, dichlorprop,mecoprop ong>andong>The phenoxyalkanoic acid herbicides; MCPA, dichlorprop ong>andong> mecoproptogether with isoproturon were chosen ong>inong> the begong>inong>nong>inong>g as model substances.isoproturon The phenoxyalkanoic acid herbicides were extensively used both ong>inong>the sprong>inong>g ong>andong> ong>inong> the autumn when the project began ong>inong> 1996. In the begong>inong>nong>inong>g of1997 the Danish Environmental Protection Agency announced that productscontaong>inong>ong>inong>g phenoxyalkanoic acid herbicides ong>inong> general were prohibited for autumnuse henceforward ong>andong> only allowed for sprong>inong>g for a few applications.In 1996, isoproturon was approved for use both ong>inong> the sprong>inong>g ong>andong> ong>inong> the autumn, butfrom 1997 only the autumn application was permitted.The substances have been found ong>inong> ong>precipitationong> from neighbourong>inong>g countries(Kirknel ong>andong> Feldong>inong>g, 1996)DNOC ong>inong> 1998In 1998, the analysis program was extended to ong>inong>clude bentazone ong>andong> 2 of itsmetabolites, DNOC, ong>andong> 2 metabolites of isoproturon. DNOC has not been used ong>inong>Denmark durong>inong>g the last 10 years, but was ong>inong>cluded because it was found ong>inong> thesamples that were analysed at DMU ong>inong> 1997.139


4 ResultsThe results are from the period of May 1996 until December 1998. The samplesfrom 1996 ong>andong> 1997 are analysed with GC-MS, whereas the raong>inong>water samplesfrom 1998 have been analysed with LC-MS.Table 1Concentration ong>inong> µg/L (95% confidence limit) of phenoxyalkanoic acids herbicidesong>inong> raong>inong>water from Gadevang 1996 ong>andong> 1997.Koncentration i µg/L (95% konfidensong>inong>terval) af phenoxysyrer i regnvong>andong> fraGadevang 1996 og 1997.Locality Collection period Mecopropµg/LMCPAµg/LGadevang 10/5-20/5-1996 Nr Nr NrGadevang 20/5-28/5-1996 Nr Nr NrGadevang 28/5-10/6-1996 Nr Nr NrGadevang 10/6-20/6-1996 Nr Nr NrGadevang 20/6-5/7-1996 Nr Nr NrGadevang 5/7-21/7-1996 Nr Nr NrGadevang 21/7-19/8-1996 Nr Nr NrGadevang 19/8-3/9-1996 Nr Nr NrGadevang 3/9-16/9-1996 Nr Nr NrGadevang 16/9-2/10-1996 Nr Nr NrGadevang 2/10-21/10-1996 0.087(0.086;0.089) Nr NrGadevang 21/10-31/10-1996 0.081(0.080;0.082) Nr NrGadevang 31/10-12/11-1996 0.020(0.019;0.022) Nr NrGadevang 12/11-25/11-1996 Nr Nr NrGadevang 16/4-5/5-1997 Nr Nr NrGadevang 5/5-21/5-1997 Nr Nr NrGadevang 21/5-3/6-1997 Nr Nr NrGadevang 3/6-24/6-1997 Nr Nr NrGadevang 24/6-2/7-1997 Nr Nr NrGadevang 2/7-29/29-1997 Nr Nr NrGadevang 29/7-11/8-1997 Nr Nr NrGadevang 11/8-1/9-1997 Nr Nr NrGadevang 1/9-15/9-1997 Nr Nr NrGadevang 15/9-6/10-1997 Nr Nr NrGadevang 6/10-15/10-1997 Nr Nr NrGadevang 15/10-03/11-1997 ¤ ¤ ¤Gadevang 03/11-14/11-1997 Nr Nr NrGadevang 14/11-02/12-1997 Nr Nr NrNr = no respons¤ = only for IPUDichlorpropµg/L140


Locality Collection period Bentazoneµg/LTable 2Concentration ong>inong> µg/L of phenoxyalkanoic acid herbicides ong>andong> other acidicherbicides ong>inong> raong>inong> from Gadevang 1998.Koncentration i µg/L af phenoxysyrer og ong>andong>re sure herbicider i regnvong>andong> fraGadevang 1998.6-Hydroxybentazoneµg/l8-Hydroxybentazoneµg/LDNOCµg/LMecopropµg/L2,4-Dµg/LDichlorpropµg/LMCPAµg/LGadevang 24/3-6/4-98 n.d. n.d. n.d. 0.40 n.d. n.d. n.d. n.d.Gadevang 6/4-23/4-98 n.d. n.d. n.d. 0.59 n.d. n.d. n.d. n.d.Gadevang 23/4-11/5-98 n.d. n.d. n.d. 0.24 n.d. n.d. n.d. n.d.Gadevang 11/5-29/5-98 n.d. n.d. n.d. 0.23 n.d. n.d. n.d. n.d.Gadevang 29/5-11/6-98 n.d. 0.011 n.d. 0.38 n.d. n.d. n.d. n.d.Gadevang 11/6-25/6-98 n.d. n.d. n.d. 0.04 n.d. n.d. n.d. n.d.Gadevang 25/6-3/7-98 n.d. 0.019 n.d. 0.09 n.d. n.d. n.d. n.d.Gadevang 3/7-28/7-98 n.d. n.d. n.d. 0.14 n.d. n.d. n.d. n.d.Gadevang 28/7-18/8-98 n.d. n.d. n.d. 0.29 n.d. n.d. n.d. n.d.Gadevang 18/8-7/9-98 n.d. n.d. n.d. 0.17 n.d. n.d. n.d. n.d.Gadevang 7/9-28/9-98 n.d. n.d. n.d. 0.22 n.d. n.d. n.d. n.d.Gadevang 28/9-13/10-98 n.d. n.d. i.a. 0.28 n.d. n.d. n.d. n.d.Gadevang 13/10-21/10-98 n.d. n.d. n.d. 0.20 n.d. n.d. n.d. n.d.Gadevang 21/10-28/10-98 n.d. n.d. n.d. 0.06 n.d. n.d. n.d. n.d.Gadevang 28/10-11/11-98 n.d. n.d. n.d. 0.14 n.d. n.d. n.d. n.d.Gadevang 11/11-3/12-98 n.d. n.d. n.d. 0.25 n.d. n.d. n.d. n.d.Gadevang 3/12-21/12-98 n.d. n.d. n.d. 0.11 n.d. n.d. n.d. n.d.n.d.: Not detected. Detection limit is 0.01µg/L.All phenoxyalkonic acid The results for the phenoxyalkanoic acid herbicides ong>andong> other acidicherbicides ong>inong> 1996 substances are listed ong>inong> table 1-6. At the Gadevang locality only mecoprop wasdetected ong>inong> the period from October 1996 until Mid-November 1996. At the Lorupong>andong> Gisselfeld localities, all 3 phenoxyalkanoic acid herbicidesOnly mecoprop ong>inong> 1997 were detected ong>inong> 1996. In 1997, only mecoprop was found ong>inong> one sample fromLorup, the sample was collected durong>inong>g November.The products available on the market ong>inong> 1996 contaong>inong>ong>inong>g phenoxyalkanoic acidherbicides were prohibited the July 1 st 1997, but hereafter manufacturers couldapply for new permits for other products contaong>inong>ong>inong>g phenoxyalkanoic acidherbicides. This is the reason why it is still possible to fong>inong>d a number of productscontaong>inong>ong>inong>g phenoxyalkanoic acid herbicides, which are permitted for weed controlong>inong> herbage seed, grass lawns, ong>andong> for control of root weeds ong>inong> cereals ong>inong> the latesprayong>inong>g growth stages. The permits are based on the coverage of the crop, i.e. thedemong>andong> of a maximum of a 100 g (for dichlorprop 60 g) depositong>inong>g on the soilsurface.141


Table 3Concentration ong>inong> µg/L (95% Confidence limits) of phenoxyalkanoic acid herbicidesong>inong> raong>inong>water from Lorup 1996 ong>andong> 1997.Koncentration i µg/L (95% konfidensong>inong>terval) af phenoxysyrer i regnvong>andong> fraLorup 1996 og 1997.Locality Collection period Mecoprop MCPADichlorpropµg/Lµg/Lµg/LLorup 14/5-20/5-1996 Nr Nr NrLorup 20/5-28/5-1996 0.018(0.017;0.019) 0.075(0.072;0.077) 0.632(0.623;0.642)Lorup 28/5-10/6-1996 Nr 0.102(0.099;0.104) 0.148(0.146;0.150)Lorup 10/6-20/6-1996 Nr 0.124(0.119;0.130) 0.204(0.200;0.207)Lorup 20/6-5/7-1996 Nr Nr NrLorup 5/7-21/7-1996 Nr 0.092(0.087;0.097) 0.202(0.199;0.205)Lorup 21/7-19/8-1996 Nr Nr NrLorup 19/8-3/9-1996 Nr Nr NrLorup 3/9-16/9-1996 Nr Nr NrLorup 16/9-2/10-1996 Nr Nr NrLorup 2/10-21/10-1996 0.143(0.142;0.145) Nr NrLorup 21/10-31/10-1996 0.145(0.144;0.147) Nr NrLorup 31/10-12/11-1996 0.064(0.060;0.068) Nr NrLorup 12/11-25/11-1996 0.031(0.027;0.034) Nr NrLorup 3/4-14/4-1997 Nr Nr NrLorup 15/4-5/5-1997 Nr Nr NrLorup 5/5-21/5-1997 Nr Nr NrLorup 21/5-3/6-1997 Nr Nr NrLorup 3/6-24/6-1997 Nr Nr NrLorup 24/6-2/7-1997 Nr Nr NrLorup 2/7-29/29-1997 Nr Nr NrLorup 29/7-11/8-1997 Nr Nr NrLorup 11/8-1/9-1997 Nr Nr NrLorup 1/9-15/9-1997 Nr Nr NrLorup 15/9-6/10-1997 ¤ ¤ ¤Lorup 6/10-15/10-1997 ¤ ¤ ¤Lorup 15/10-03/11-1997 ¤ ¤ ¤Lorup 03/11-14/11-1997 0.086(0.084;0.089) Nr NrLorup 14/11-02/12-1997 Nr Nr NrNr = No response¤ = Only for IPU142


Table 4Concentrations ong>inong> µg/L of phenoxyalkanoic acid herbicides ong>andong> other acidicherbicides ong>inong> raong>inong> water from Lorup 1998.Koncentration i µg/L af phenoxysyrer og ong>andong>re sure herbicider i regnvong>andong> fraLorup 1998.Locality CollectionperiodBentazoneµg/L6-Hydroxybentazoneµg/L8-Hydroxybentazoneµg/LDNOCµg/LMecopropµg/L2,4-Dµg/LDichlorpropµg/LMCPAµg/lLorup 24/3-6/4-98 n.d. n.d. n.d. 0.32 n.d. n.d. n.d. n.d.Lorup 6/4-23/4-98 0.019 n.d. n.d. 0.87 n.d. n.d. n.d. n.d.Lorup 23/4-11/5-98 n.d. n.d. n.d. 0.60 n.d. n.d. n.d. n.d.Lorup 11/5-29/5-98 n.d. n.d. n.d. 0.27 n.d. n.d. n.d. n.d.Lorup 29/5-11/6-98 n.d. n.d. n.d. 0.40 n.d. n.d. n.d. n.d.Lorup 11/6-25/6-98 n.d. n.d. n.d. 0.15 n.d. n.d. n.d. n.d.Lorup 25/6-3/7-98 n.d. n.d. n.d. 0.19 n.d. n.d. n.d. n.d.Lorup 3/7-28/7-98 n.d. n.d. n.d. 0.23 n.d. n.d. n.d. n.d.Lorup 28/7-18/8-98 n.d. n.d. n.d. 0.29 n.d. n.d. n.d. n.d.Lorup 18/8-7/9-98 n.d. n.d. n.d. 0.25 n.d. n.d. n.d. n.d.Lorup 7/9-28/9-98 n.d. n.d. n.d. 0.23 n.d. n.d. n.d. n.d.Lorup 28/9-13/10-98 n.d. n.d. i.a. 0.30 n.d. n.d. n.d. n.d.Lorup 13/10-21/10-98 n.d. n.d. n.d. 0.32 n.d. n.d. n.d. n.d.Lorup 21/10-28/10-98 n.d. n.d. n.d. 0.19 n.d. n.d. n.d. n.d.Lorup 28/10-11/11-98 n.d. n.d. n.d. 0.14 n.d. n.d. n.d. n.d.Lorup 11/11-3/12-98 n.d. n.d. n.d. 0.18 n.d. n.d. n.d. n.d.Lorup 3/12-21/12-98 n.d. n.d. n.d. 0.11 n.d. n.d. n.d. n.d.n.d.: not detected. The detection limit is 0.01µg/L143


Table 5Concentrations ong>inong> µg/L (95% confidence limits) of phenoxyalkanoic acidherbicides ong>inong> raong>inong>water from Gisselfeld 1996 ong>andong> 1997.Koncentration i µg/L (95% konfidensong>inong>terval) af phenoxysyrer i regnvong>andong> fraGisselfeld 1996 og 1997.Locality Collection period Mecoprop µg/L MMCPA µg/L DDichlorprop µg/LGisselfeld 15/5-20/5-1996 Nr 0.081(0.077;0.084) 0.351(0.346;0.356)Gisselfeld 20/5-28/5-1996 0.089(0.082;0.097) 0.143(0.136;0.151) 0.220(0.214;0.226)Gisselfeld 28/5-10/6-1996 Nr Nr 0.106(0.103;0.108)Gisselfeld 10/6-20/6-1996 Nr Nr NrGisselfeld 20/6-5/7-1996 Nr Nr NrGisselfeld 5/7-21/7-1996 Nr Nr NrGisselfeld 21/7-19/8-1996 Nr Nr NrGisselfeld 19/8-3/9-1996 Nr Nr NrGisselfeld 3/9-16/9-1996 Nr Nr NrGisselfeld 16/9-2/10-1996 Nr Nr NrGisselfeld 2/10-21/10-1996 0.073(0.072;0.074) Nr NrGisselfeld 21/10-31/10-1996 0.033(0.032;0.034) Nr NrGisselfeld 31/10-12/11-1996 Nr Nr NrGisselfeld 12/11-25/11-1996 Nr Nr NrGisselfeld 3/4-15/4-1997 Nr Nr NrGisselfeld 15/4-5/5-1997 Nr Nr NrGisselfeld 5/5-21/5-1997 Nr Nr NrGisselfeld 21/5-3/6-1997 Nr Nr NrGisselfeld 3/6-24/6-1997 Nr Nr NrGisselfeld 24/6-2/7-1997 Nr Nr NrGisselfeld 2/7-29/29-1997 Nr Nr NrGisselfeld 29/7-11/8-1997 Nr Nr NrGisselfeld 11/8-1/9-1997 Nr Nr NrGisselfeld 1/9-15/9-1997 Nr Nr NrGisselfeld 15/9-6/10-1997 ¤ ¤ ¤Gisselfeld 6/10-15/10-1997 Nr Nr NrGisselfeld 15/10-03/11-1997 ¤ ¤ ¤Gisselfeld 03/11-14/11-1997 Nr Nr NrGisselfeld 14/11-02/12-1997 Nr Nr NrNr = no response¤ = Only for IPU144


LocalityCollectionperiodTable 6Concentrations ong>inong> µg/L of phenoxyalkanoic acid herbicides ong>andong> other acidicherbicides ong>inong> raong>inong>water from Gisselfeld 1998Koncentration i µg/L af phenoxysyrer og ong>andong>re sure herbicider i regnvong>andong> fraGisselfeld 1998.Bentazoneµg/L6-Hydroxybentazone µg/LDNOCµg/LMecopropµg/L2,4-Dµg/L8-Hydroxybentazoneµg/LDichlorpropµg/LGisselfeld 24/3-6/4-98 n.d. n.d. n.d. 0.10 n.d. n.d. n.d. n.d.Gisselfeld 6/4-23/4-98 n.d. n.d. n.d. 0.74 n.d. n.d. n.d. n.d.Gisselfeld 23/4-11/5-98 n.d. n.d. n.d. 0.16 n.d. n.d. n.d. n.d.Gisselfeld 11/5-29/5-98 n.d. n.d. n.d. 0.13 n.d. n.d. n.d. n.d.Gisselfeld 29/5-11/6-98 n.d. n.d. n.d. 0.32 n.d. n.d. n.d. n.d.Gisselfeld 11/6-25/6-98 n.d. n.d. n.d. 0.13 n.d. n.d. n.d. n.d.Gisselfeld 26/6-3/7-98 n.d. n.d. n.d. 0.17 n.d. n.d. n.d. n.d.Gisselfeld 3/7-28/7-98 n.d. n.d. n.d. 0.11 n.d. n.d. n.d. n.d.Gisselfeld 28/7-18/8-98 n.d. n.d. n.d. 0.19 n.d. n.d. n.d. n.d.Gisselfeld 18/8-7/9-98 n.d. n.d. n.d. 0.10 n.d. n.d. n.d. n.d.Gisselfeld 7/9-28/9-98 n.d. n.d. n.d. 0.24 n.d. n.d. n.d. n.d.Gisselfeld 28/9-13/10-98 n.d. n.d. i.a. 0.48 n.d. n.d. n.d. n.d.Gisselfeld 13/10-21/10- n.d. n.d. n.d. 0.40 n.d. n.d. n.d. n.d.98Gisselfeld 21/10-28/10- n.d. n.d. n.d. 0.18 n.d. n.d. n.d. n.d.98Gisselfeld 28/10-11/11- n.d. n.d. n.d. 0.23 n.d. n.d. n.d. n.d.98Gisselfeld 11/11-3/12-98 n.d. n.d. n.d. 0.39 n.d. n.d. n.d. n.d.Gisselfeld 3/12-21/12-98 n.d. n.d. n.d. 0.13 n.d. n.d. n.d. n.d.n.d.: Not detected. The detection limit is 0.01µg/Li.a.: not analysedMCPAµg/LUp to 0.87 µg/L DNOCThe results for DNOC ong>inong> table 2,4 ong>andong> 6 are corrected for 60% recovery at thesample preparation. The rest of the data is not corrected, because the recovery rateswere close to 100%. DNOC was detected ong>inong> all the samples ong>inong>concentrations rangong>inong>g from 0.04 µg/L to 0.87µg/L.In 1996, the phenoxyalkanoic acid herbicides; respectively; mecoprop, MCPA ong>andong>dichlorprop, were detected ong>inong> the followong>inong>g concentrations; from 0.02 µg/L to 0.15µg/L, from 0.08 µg/L to 0.14 µg/L ong>andong> from 0.11µg/L to 0.63 µg/L. There is anobvious connection between fong>inong>dong>inong>gs ong>inong> the raong>inong>water ong>andong> the time of sprayong>inong>g(Feldong>inong>g, 1998).No phenoxyalkanoic acid herbicides were detected ong>inong> 1997 at the Gadevanglocality, with exception of the period from the 15/10 to the 3/11 where no sampleswere collected. In Lorup, none of the samples from the period 15/9 to 3/11, wereanalysed for phenoxyalkanoic acid herbicides, but ong>inong> the followong>inong>g periodmecoprop was detected ong>inong> the samples. At the Gisselfeld locality, nophenoxyalkanoic acid herbicides were found ong>inong> 1997, however, here are 2 periodswithout samples, from the 15/9 to the 6/10 ong>andong> agaong>inong> from the 15/10 to the 3/11.No phenoxyalkanoicIn 1998, no phenoxyalkanoic acid herbicides were detected at the three145


acid herbicides ong>inong> 1998DNOC ong>inong> all samplesIsoproturon detectedon all localitieslocations, which show that the limitation ong>inong> the use of the phenoxyalkanoic acidherbicides had a remarkably effect on the fong>inong>dong>inong>gs ong>inong> ong>precipitationong>. It also showsthat it is not possible to trace transport of phenoxyalkanoic acid herbicides fromother countries without limitations ong>inong> the use of these specific pesticides.With the exception of two fong>inong>dong>inong>gs of trace amounts of hydroxybentazone,bentazone ong>andong> hydroxy compounds were not detected ong>inong> the samples taken ong>inong> 1998.On top of that, DNOC was detected ong>inong> all samples from 1998, ong>andong> most of thesamples contaong>inong>ed DNOC ong>inong> concentrations between 0.1 ong>andong> 0.4µg/L.The results for isoproturon are listed ong>inong> table 7-12, where the compound isdetected ong>inong> raong>inong>water from all three localities ong>inong> the autumn 1996, 1997 ong>andong> 1998 aswell as ong>inong> the sprong>inong>g 1998 at all three localities. Concentrations of isoproturon varyfrom 0.01-to 0.86µg/L.The concentrations of pesticides shown ong>inong> the tables are to be regarded asmong>inong>imum concentrations, because of the degradation of the pesticides that arecertaong>inong> to arise from the time the raong>inong> event has occurred, ong>andong> until the sample iscollected ong>andong> brought to the laboratory about 2 weeks later, even though the wateris preserved.146


Table 7Concentrations ong>inong> µg/L (95% confidence limits) of isoproturon ong>inong> the raong>inong>waterfrom Gadevang ong>inong> 1996 ong>andong> 1997.Koncentration i µg/L (95% konfidensong>inong>terval) af isoproturon i regnvong>andong> fraGadevang i 1996 og 1997.Locality Collection period Isoproturon ong>inong> µg/LGadevang 10/5-20/5-1996 NrGadevang 20/5-28/5-1996 NrGadevang 28/5-10/6-1996 NrGadevang 10/6-20/6-1996 *Gadevang 20/6-5/7-1996 NrGadevang 5/7-21/7-1996 *Gadevang 21/7-19/8-1996 NrGadevang 19/8-3/9-1996 NrGadevang 3/9-16/9-1996 NrGadevang 16/9-2/10-1996 NrGadevang 2/10-21/10-1996 0.257(0.223;0.292)Gadevang 21/10-31/10-1996 NrGadevang 31/10-12/11-1996 NrGadevang 12/11-25/11-1996 NrGadevang 16/4-5/5-1997 NrGadevang 5/5-21/5-1997 NrGadevang 21/5-3/6-1997 NrGadevang 3/6-24/6-1997 NrGadevang 24/6-2/7-1997 NrGadevang 2/7-29/7-1997 NrGadevang 29/7-11/8-1997 NrGadevang 11/8-1/9-1997 NrGadevang 1/9-15/9-1997 NrGadevang 15/9-6/10-1997 0.073(0.071;0.075)Gadevang 6/10-15/10-1997 0.039(0.038;0.04 )Gadevang 15/10-03/11-1997 0.859(0.846;0.873)Gadevang 03/11-14/11-1997 NrGadevang 14/11-02/12-1997 NrNr = No response * = Only for phenoxyalkanoic acid herbicides147


Table 8Concentrations ong>inong> µg/L of isoproturon ong>andong> 2 metabolites ong>inong> raong>inong>water fromGadevang 1998.Koncentration i µg/L af isoproturon og 2 metabolitter i regnvong>andong> fra Gadevang1998.LocalityCollectionperiodIsopropylphenylureaµg/LIsopropylphenylmethylureaµg/LGadevang 24/3-6/4-98 n.d. n.d. n.d.Gadevang 6/4-23/4-98 n.d. n.d. 0.037Gadevang 23/4-11/5-98 n.d. n.d. 0.015Gadevang 11/5-29/5-98 n.d. n.d. 0.014Gadevang 29/5-11/6-98 n.d. n.d. n.d.Gadevang 11/6-25/6-98 n.d. n.d. n.d.Gadevang 25/6-3/7-98 n.d. n.d. n.d.Gadevang 3/7-28/7-98 n.d. n.d. n.d.Gadevang 28/7-18/8-98 n.d. n.d. n.d.Gadevang 18/8-7/9-98 n.d. n.d. n.d.Gadevang 7/9-28/9-98 n.d. n.d. n.d.Gadevang 28/9-13/10-98 n.d. n.d. 0.033Gadevang 13/10-21/10-98 n.d. n.d. 0.048Gadevang 21/10-28/10-98 n.d. n.d. 0.011Gadevang 28/10-11/11-98 n.d. n.d. n.d.Gadevang 11/11-3/12-98 n.d. n.d. n.d.Gadevang 3/12-21/12-98 n.d. n.d. n.d.n.d.: Not detected. The detection limit is 0.01µg/LIsoproturonµg/L148


Table 9Concentrations ong>inong> µg/L of isoproturon (95% confidence limits) ong>inong> raong>inong>water fromLorup 1996 ong>andong> 1997.Koncentration i µg/L (95% konfidensong>inong>terval) af isoproturon i regnvong>andong> fra Lorupi 1996 og 1997.Locality Collection period Isoproturon ong>inong> µg/LLorup 14/5-20/5-1996 NrLorup 20/5-28/5-1996 NrLorup 28/5-10/6-1996 *Lorup 10/6-20/6-1996 *Lorup 20/6-5/7-1996 NrLorup 5/7-21/7-1996 *Lorup 21/7-19/8-1996 NrLorup 19/8-3/9-1996 NrLorup 3/9-16/9-1996 NrLorup 16/9-2/10-1996 0.011(0.010;0.012)Lorup 2/10-21/10-1996 0.171(0.142;0.201)Lorup 21/10-31/10-1996 0.080(0.071;0.088)Lorup 31/10-12/11-1996 0.030(0.024;0.036)Lorup 12/11-25/11-1996 NrLorup 3/4-14/4-1997 NrLorup 15/4-5/5-1997 NrLorup 5/5-21/5-1997 NrLorup 21/5-3/6-1997 NrLorup 3/6-24/6-1997 NrLorup 24/6-2/7-1997 NrLorup 2/7-29/7-1997 NrLorup 29/7-11/8-1997 *Lorup 11/8-1/9-1997 NrLorup 1/9-15/9-1997 NrLorup 15/9-6/10-1997 0.099(0.098;0.100)Lorup 6/10-15/10-1997 NrLorup 15/10-03/11-1997 NrLorup 03/11-14/11-1997 NrLorup 14/11-02/12-1997 NrNr = No response *= Only for phenoxyalkanoic acid herbicides149


Table 10Concentrations ong>inong> µg/L of isoproturon ong>andong> 2 metabolites ong>inong> raong>inong>water from Lorup1998.Koncentration i µg/L af isoproturon og 2 metabolitter i regnvong>andong> fra Lorup 1998.LocalityCollectionperiodIsopropylphenylureaµg/LIsopropylphenylmethylureaµg/LIsoproturonµg/LLorup 24/3-6/4-98 n.d. n.d. n.d.Lorup 6/4-23/4-98 n.d. n.d. 0.021Lorup 23/4-11/5-98 n.d. n.d. 0.015Lorup 11/5-29/5-98 n.d. n.d. n.d.Lorup 29/5-11/6-98 n.d. 0.011 n.d.Lorup 11/6-25/6-98 n.d. n.d. n.d.Lorup 25/6-3/7-98 n.d. n.d. n.d.Lorup 3/7-28/7-98 n.d. n.d. n.d.Lorup 28/7-18/8-98 n.d. n.d. n.d.Lorup 18/8-7/9-98 0.082 n.d. n.d.Lorup 7/9-28/9-98 0.074 n.d. n.d.Lorup 28/9-13/10-98 n.d. n.d. 0.019Lorup 13/10-21/10-98 n.d. 0.016 0.061Lorup 21/10-28/10-98 n.d. n.d. 0.046Lorup 28/10-11/11-98 n.d. n.d. 0.042Lorup 11/11-3/12-98 n.d. n.d. n.d.Lorup 3/12-21/12-98 n.d. n.d. n.d.n.d.: Not detected. The Detection limit is 0.01µg/L150


Table 11Concentrations ong>inong> µg/L of isoproturon (95% confidence limits) ong>inong> raong>inong>water fromGisselfeld 1996 ong>andong> 1997.Koncentration i µg/L (95% konfidensong>inong>terval) af isoproturon i regnvong>andong> fraGisselfeld i 1996 og 1997.Locality Collection period Isoproturon ong>inong> µg/LGisselfeld 15/5-20/5-1996 NrGisselfeld 20/5-28/5-1996 NrGisselfeld 28/5-10/6-1996 *Gisselfeld 10/6-20/6-1996 *Gisselfeld 20/6-5/7-1996 NrGisselfeld 5/7-21/7-1996 *Gisselfeld 21/7-19/8-1996 NrGisselfeld 19/8-3/9-1996 NrGisselfeld 3/9-16/9-1996 NrGisselfeld 16/9-2/10-1996 NrGisselfeld 2/10-21/10-1996 0.383(0.346;0.428)Gisselfeld 21/10-31/10-1996 0.081(0.075;0.087)Gisselfeld 31/10-12/11-1996 0.044(0.038;0.050)Gisselfeld 12/11-25/11-1996 NrGisselfeld 3/4-15/4-1997 NrGisselfeld 15/4-5/5-1997 NrGisselfeld 5/5-21/5-1997 NrGisselfeld 21/5-3/6-1997 NrGisselfeld 3/6-24/6-1997 NrGisselfeld 24/6-2/7-1997 NrGisselfeld 2/7-29/7-1997 NrGisselfeld 29/7-11/8-1997 *Gisselfeld 11/8-1/9-1997 NrGisselfeld 1/9-15/9-1997 NrGisselfeld 15/9-6/10-1997 NrGisselfeld 6/10-15/10-1997 NrGisselfeld 15/10-03/11-1997 0.398(0.392;0.404)Gisselfeld 03/11-14/11-1997 NrGisselfeld 14/11-02/12-1997 NrNr = No response * = Only for phenoxyalkanoic acid herbicides151


Table 12Concentrations ong>inong> µg/L of isoproturon ong>andong> 2 metabolites ong>inong> raong>inong>water fromGisselfeld 1998.Koncentration i µg/L af isoproturon og 2 metabolitter i regnvong>andong> fra Gisselfeld1998.LocalityCollectionperiodIsopropylphenylureaµg/LIsopropylphenylmethylureaµg/LGisselfeld 24/3-6/4-98 n.d. n.d. n.d.Gisselfeld 6/4-23/4-98 n.d. n.d. 0.138Gisselfeld 23/4-11/5-98 n.d. n.d. 0.031Gisselfeld 11/5-29/5-98 n.d. n.d. 0.015Gisselfeld 29/5-11/6-98 n.d. n.d. n.d.Gisselfeld 11/6-25/6-98 n.d. n.d. n.d.Gisselfeld 26/6-3/7-98 n.d. n.d. n.d.Gisselfeld 3/7-28/7-98 n.d. n.d. n.d.Gisselfeld 28/7-18/8-98 n.d. n.d. n.d.Gisselfeld 18/8-7/9-98 n.d. n.d. n.d.Gisselfeld 7/9-28/9-98 n.d. n.d. n.d.Gisselfeld 28/9-13/10-98 n.d. n.d. 0.037Gisselfeld 13/10-21/10-98 n.d. n.d. 0.048Gisselfeld 21/10-28/10-98 n.d. n.d. 0.027Gisselfeld 28/10-11/11-98 n.d. n.d. 0.012Gisselfeld 11/11-3/12-98 n.d. n.d. n.d.Gisselfeld 3/12-21/12-98 n.d. n.d. n.d.n.d.: Not detected. The detection limit is 0.01µg/LIsoproturonµg/LIsoproturon metabolites The samples from 1998 were tested for isoproturon ong>andong> two of itsonly ong>inong> 4 samples metabolites. Metabolites were only detected ong>inong> four samples from Lorup, with amaximum concentration of isopropyl-phenylurea at 0.082 µg/L.Samples analysed for Durong>inong>g the autumn of 1997 it was possible to analyse some frozen samples44 pesticides for up to 44 different pesticides. Samples from September 1996 until November1997 were defrosted ong>andong> mixtures from the three localities were made. Theanalyses were performed by National Environmental Research Institute,Department for Environmental Chemistry. The pesticides analysed, are listed ong>inong>table 13.152


Table 13List of pesticides ong>andong> metabolites contaong>inong>ed ong>inong> the analysis program at DMU.Liste over pesticider og nedbrydnong>inong>gsprodukter i DMU’s analyseprogramComponents ong>inong> the analysis at DMUAtrazong>inong>eAtrazong>inong>e-2-hydroxyBAM(2,6-Dichlorobenzamide)CarbofuranChloridazonCyanazong>inong>eDeethyl-terbuthylazong>inong>eDeethylatrazong>inong>eDeisopropylatrazong>inong>eDimethoateDiuronHexazong>inong>oneHydroxy-carbofuranIsoproturonLong>inong>uronMetamitronMethabenzthiazuronMetribuzong>inong>PirimicarbProchlorazPropiconazolePropyzamideSimazong>inong>eTerbuthylazong>inong>e,2-hydroxyTerbuthylazong>inong>eTriadimenol2,4-D2,4-dichlorophenolBenazolong>inong>BentazoneBromoxynilChlorsulfuronDNOCDicambaDichlorpropDong>inong>osebFlampropFluazifopIoxynilMCPAMecopropMetsulfuron-methylTriasulfuronThifensulfuron-methyl153


8 different pesticides In table 14 the concentrations of the 8 different pesticide chemicals whichdetectedwere found ong>inong> the sample mixture are listed. In addition to herbicides mecopropong>andong> isoproturon, which were also detected ong>inong> the samples mentioned above, thefollowong>inong>g substances were detected ong>inong> relatively low concentrations from 2 to 5times durong>inong>g the period. These are; metamitron, methabenzthiazuron, 2-hydroxyterbuthylazong>inong>e, terbuthylazong>inong>e ong>andong> 2,4-D. DNOC on the other hong>andong>, wasfound ong>inong> raong>inong>water durong>inong>g the entire samplong>inong>g period ong>inong> concentrations from 0.38-to 4.5µg/L.Table 14Concentrations of pesticides ong>inong> raong>inong>water samples collected ong>inong> the period fromSeptember 1996 until November 1997. The samples are ong>inong> most cases samplemixtures from Lorup, Gisselfeld ong>andong> Gadevang, ong>andong> they typically represent aperiod of 2-3 weeks which gives results once or twice a calendar month. Alltogether there was analysed for 44 pesticide chemicals at DMU. Results ong>inong> µg/LKoncentration af pesticider i regnvong>andong>sprøver opsamlet fra september 1996 tilnovember 1997. Prøverne er i de fleste tilfælde blong>andong>ong>inong>gsprøver fra Lorup,Gisselfeld ogGadevangog repræsenterer 2-3ugers nedbør, hvilket giver et eller toresultater per kalenermåned.. Alle er analyseret for 44 pesticidkemikalier af DMU.Resultater i µg/L.DNOCCollected IPU MetamitronMecopropMethabenz-Thiazuron2-Hydroxyterbuthylazong>inong>eTerbuthylazong>inong>e2,4-DSept. 96 0.007 0.61 0.005October 0.20 0.019 4.5 0.23October 0.054 0.039 0.69 0.07Novem. 0.017 0.038 0.89 0.07Novem. 0.77 0.030 0.013April 97 0.82 0.016 0.008 0.096May 0.31 0.015 0.098 0.008June 0.57 0.012 0.009August 0.024 1.6Sept. 0.38 0.005 0.046Sept. 0.17 0.088 1.3 0.068 0.015 0.008 0.059October 0.44 0.016 0.013October 0.29 0.47 0.076In table 15 the measured concentration of pesticide chemicals -from table 14- areconverted from µg/L to g/ha. As shown ong>inong> the table, DNOC supplies the groundwith roughly 90% of the total amount of pesticides (7,5g of 8,5g), detected by theextended analysis.154


Table 15Concentrations of pesticides ong>andong> metabolites ong>inong> g pr. ha.Koncentration af pesticider og metabolitter i gram per ha.Collected IPU Metamitron DNOC Mecoprothiazuronterbuthylazong>inong>lazong>inong>eMethabenz-2-Hydroxy-Terbuthy-2,4-DSept. 96 0.007 0.632 0.005October 0.050 0.005 1.132 0.058October 0.053 0.038 0.676 0.069Novem. 0.018 0.040 0.944 0.074Novem. 0.657 0.026 0.011April 97 0.640 0.012 0.006 0.075May 0.271 0.013 0.086 0.007June 0.356 0.007 0.006August 0.014 0.920Sept. 0.077 0.001 0.009Sept. 0.072 0.037 0.548 0.029 0.006 0.003 0.025October 0.454 0.017 0.013October 0.116 0.188 0.030Total 0.316 0.134 7.496 0.341 0.017 0.167 0.016 0.048155


5 Discussion ong>andong> conclusionConnection betweensprayong>inong>g ong>andong> fong>inong>dong>inong>gsLimitations ong>inong> use reducedcontent ong>inong> ong>precipitationong>Especially ong>inong> the case of the phenoxyalkanoic acid herbicides ong>andong>isoproturon, there is an obvious connection between fong>inong>dong>inong>gs ong>inong> the raong>inong>water ong>andong>the time of sprayong>inong>g. The same pattern is repeated ong>inong> the foreign papers, thatdescribe analysis of raong>inong>water. From the previous European research it appears,that the load of pesticides is of the same magnitude as ong>inong> Denmark. A compound aslong>inong>dane is also frequently seen ong>inong> foreign literature, here the compound is normallydetected ong>inong> ong>precipitationong> all year around.The highest measured concentrations were 0.9 µg/L for isoproturon ong>andong>0.6 µg/L for phenoxyalkanoic acid herbicides. Phenoxyherbicides were ong>inong>1996 detected ong>inong> the sprayong>inong>g periods ong>inong> sprong>inong>g ong>andong> autumn (autumn onlymecoprop, which is the only phenoxyherbicide used for autumn application).In 1997 only mecoprop was detected ong>andong> solely on one location ong>inong> November,while ong>inong> 1998 no phenoxyalkanoic acid herbicides were detected on the threelocalities. This displays that the limitations ong>inong> the use of these herbicides havegreatly reduced their appearance ong>inong> ong>precipitationong>.Isoproturon was from 1997 until 1999 only allowed for autumn application. Thecompound was detected on all the three localities ong>inong> the sprayong>inong>g period ong>inong> 1996ong>andong> 1997, while ong>inong> 1998 it was detected ong>inong> the sprong>inong>g as well as ong>inong> the autumn. Thismight be a result of long term transport from other countries or illegal use of thepesticide ong>inong> Denmark.The concentration of DNOC is high, ong>andong> the compound is also detected ong>inong>raong>inong>water collected ong>inong> periods were sprayong>inong>g is normally not performed. ThisHigh DNOC concen- ong>inong>dicates that either there are other sources of DNOC, or that the half-life of DNOCtrationsis very long (months) so that it can be transported over long distances from areaswhere application is still allowed. DNOC has been detected ong>inong> the raong>inong>water fromSeptember 1996 to November 1997 ong>andong> ong>inong> all samples collected ong>inong> 1998. It istherefore most likely that the fong>inong>dong>inong>gs of DNOC areDNOC formation ong>inong> the caused by formation of the compound ong>inong> the atmosphere. Nojima ong>andong>atmosphereIsogami (1994), Grosjean (1985), Nojima et al., (1983) ong>andong> Atkong>inong>son ong>andong>Aschmann (1994) describe the formation of DNOC ong>inong> the atmosphere probablyfrom a reaction between toluene ong>andong> nitrous oxides caused by the sunlight. Thelimitong>inong>g factor is most likely toluene which might come from the traffic. Theamount of DNOC, about 7,5g, added to the area, is a rather modest amountcompared to the amount of DNOC which was used ong>inong> the eighties, about 2 kg a.i.pr. ha.DNOC measuredabroadDNOC has also been measured abroad. In cloud water ong>inong> Englong>andong>concentrations at 0.26-2.13µg/L were found (Lüttke ong>andong> Levsen, 1997), ong>andong> ong>inong>Germany at 0.9-12.5µg/L (Richartz et al., 1990). Results from Switzerlong>andong> showsconcentrations at 0.95-1.6µg/L for raong>inong>water samples (Leuenberger et al., 1988).The concentrations detected ong>inong> the raong>inong> ong>inong> Denmark are of the same size asconcentrations found ong>inong> Englong>andong>, Germany ong>andong> Switzerlong>andong>.10 years song>inong>ce DNOC DNOC is the only one of the 8 detected pesticides, which is no longer usedused ong>inong> Denmark ong>inong> Denmark, it is about 10 years song>inong>ce the compound was used. In 1996, DNOCwas still used ong>inong> a few countries ong>inong> Europe ong>andong> probably also outside Europe.156


DNOC has been detected ong>inong> groundwater close to the soil surface ong>andong> ong>inong> streams ong>inong>Denmark (Spliid et al., 1996).Raong>inong>falls ong>inong> begong>inong>nong>inong>gof collection periodsgave unstable samplesAnalysis of control samples placed at the three localities have shown that thesamples are not adequately stable especially ong>inong> the summer season if the raong>inong>falls ong>inong> the begong>inong>nong>inong>g of the collection period. In forthcomong>inong>g raong>inong>water projects itis therefore necessary to use more advanced samplong>inong>g equipment where thesamples for example are cooled durong>inong>g the collection period.157


6 AcknowledgementsThe laboratory technicians Bente Laursen ong>andong> Hanne Krogh Bæk, both employedong>inong> the Department of Crop Protection, Research Centre Flakkebjerg, are thanked forong>inong>credible, ong>inong>dependent ong>andong> competent work, both ong>inong> the field ong>andong> ong>inong> the laboratory.Office manager, Sonja Graugaard, is thanked for all her work with corrections ong>inong>the manuscript etc.158


7 ReferencesAtkong>inong>son, R. ong>andong> Aschmann, S.M. (1994). Products of the gas-phase reactions ofaromatic hydrocarbons: Effect of NO 2 concentration. Int. J. Chem. Kong>inong>et. 26, 929-944.Bester, K., Hühnerfuss, H., Neudorf, B. ong>andong> Thiemann. (1995). AtmosphericDeposition of Triazong>inong>e Herbicides ong>inong> Northern Germany ong>andong> the German Bight(North Sea). Chemosphere, 30, 9, 1639-1653.Cleemann, M., Poulsen, M.E. ong>andong> Hilbert, G. (1995). Long distance transport.Deposition of long>inong>dane ong>inong> Denmark. ong>Pesticidesong> ong>inong> ong>precipitationong> ong>andong> surface water. In" ong>Pesticidesong> ong>inong> ong>precipitationong> ong>andong> surface water". Ed. A. Helweg. Tema Nord1995:558, 75-83.Dao, V.K., Pant, C.S. ong>andong> Sharma, V.P. (1994). HCH residues ong>inong> raong>inong> water fromHardwar, India. Bulletong>inong> of Environmental Contamong>inong>ation ong>andong> Toxicology, 52,797-801.Feldong>inong>g, G. (1998). Statusrapport til Miljøstyrelsen til projektet (7041-0006)“Forekomst af herbicider i nedbør og effekt heraf på planter og plantesamfund”.Grosjean, D. (1985). Reactions of o-cresol ong>andong> nitrocresol with NO x ong>inong> sunlight ong>andong>with ozone-nitrogen dioxide mixtures ong>inong> the dark. Environ. Sci. Technol., 19, 968-974.Hirvi, J.-P.ong>andong> Rekolaong>inong>en, S. (1995). ong>Pesticidesong> ong>inong> ong>precipitationong> ong>andong> surface waterong>inong> Fong>inong>long>andong>. In " ong>Pesticidesong> ong>inong> ong>precipitationong> ong>andong> surface water". Ed. A. Helweg.Tema Nord 1995:558., 12-19.Hüskes. R. ong>andong> Levsen, K. (1997). ong>Pesticidesong> ong>inong> raong>inong>: Chemosphere, 35, 3013-3024.Jaeschke,W., Gath, B. ong>andong> Pfäfflong>inong>, D. (1995). Deposition monitorong>inong>g of pesticidesong>inong> Southern Germany. In " ong>Pesticidesong> ong>inong> ong>precipitationong> ong>andong> surface water". Ed. A.Helweg. Tema Nord 1995:558., 65-74.Kirknel, E. ong>andong> Feldong>inong>g, G. (1995). Analysis of selected pesticides ong>inong> raong>inong> ong>inong>Denmark. In ”ong>Pesticidesong> ong>inong> ong>precipitationong> ong>andong> surface water”. Ed. A. Helweg. TemaNord 1995:558, 45-54.Kirknel, E. ong>andong> Feldong>inong>g, G. (1996). Pesticiders sprednong>inong>g med nedbøren. Tidsskriftfor long>andong>økonomi, 2, 129-138.Kreuger, J. (1995). Pesticider i regnvatten i Sverige. In " ong>Pesticidesong> ong>inong> ong>precipitationong>ong>andong> surface water". Ed. A. Helweg. Tema Nord 1995:558, 33-44.Leuenberger, C., Czuczwa, J. ong>andong> Giger, W. (1988). Nitrated phenols ong>inong> raong>inong>:Atmospheric occurrence of phytotoxic pollutants. Chemosphere 17, 511-515.159


Lode, O., Eklo, O.M., Holen, B., Svensen, A. ong>andong> Johnsen, Å.M. (1995). ong>Pesticidesong>ong>inong> ong>precipitationong> ong>inong> Norway. The Science of the Total Environment 160/161, 421-431.Lüttke, J. ong>andong> Levsen, K. (1997). Phase partitionong>inong>g of phenol ong>andong> nitrophenols ong>inong>clouds. Atmospheric Environment 31, 2649-2655.Miller, J.C. ong>andong> Miller, J.N. (1988). Errors ong>inong> ong>inong>strumental analysis; regression ong>andong>correlation ong>inong> Statistics for Analytical Chemistry, 2. udgave, Ellis Horwood Seriesong>inong> Analytical Chemistry, Halsted Press, Chichester, 101-137.Nojima, K. ong>andong> Isogami, C. (1994). Studies on photochemical reactions of ong>airong>pollutants. XIII. Formation of nitrophenols by the reactions of three toluene oxideswith nitrogen dioxide ong>inong> ong>airong>. Chem. Pharm. Bull. 42, 2426-2429.Nojima, K., Kawaguchi, A., Ohya, T., Kanno, S. ong>andong> Hirobe, M. (1983). Studies onphotochemical reaction of ong>airong> pollutants. X. Identification of nitrophenols ong>inong>suspended particulates. Chem. Pharm. Bull. 31, 1047-1051.Richartz, H., Reischl, A., Trautner, F. ong>andong> Hutzong>inong>ger, O. (1990). Nitrated phenolsong>inong> fog. Atmospheric Environment 24A, 3067-3071.Siebers, J., Gottschild, D. ong>andong> Noltong>inong>g, H.-G. (1994). ong>Pesticidesong> ong>inong> Precipitation ong>inong>Northen Germany. Chemosphere, 28, 8, 1559-1570.Spliid, N.H., Køppen, B., Frausig, A.H., Plesner, V., Sommer, N.A. ong>andong> Nielsen,M.Z. (1996). Kortlægnong>inong>g af visse pesticider i grundvong>andong>.Bekæmpelsesmiddelforsknong>inong>g fra Miljøstyrelsen. Vol. 21. 64 siderSuzuki, S. (1996). Several pesticide residues ong>inong> raong>inong>water ong>inong> eastern Japan, 1989-1992. Journal of Pesticide Science, 21, 1, 7-15.Tevisan, M., Montepiani, C., Ragozza, L.,Bartoletti, C., Ioannilli, E. ong>andong> Del Re,A.A.M. (1993). ong>Pesticidesong> ong>inong> raong>inong>fall ong>andong> ong>airong> ong>inong> Italy. Environmental Pollution, 80,31-39.U.S. Geological Survey, Fact Sheet (1995). FS-152-95, 4 pages.160


Appendix IGaschromatography Mass SpectrometryGC-MS: Hewlett Packard model 5890A gas chromatograph connected to aHewlett Packard mass selective detector model 5970Tube: capillary tube HP 5 (5% diphenyl-95% dimethylsiloxan), 25 m 0,2mm ong>inong>ner diameter ong>andong> 0.33 µm coatong>inong>gProgram for IPU detection SIM ong>andong> SCANInjection temperature: 260°COven temperature: temp. 1: 60°C (1 mong>inong>.), rate 1: 10°C/mong>inong>.temp. 2: 130°C, rate 2: 6°C/mong>inong>.temp. 3: 175°C, rate 3: 20°C/mong>inong>.temp. 4: 290°C (5 mong>inong>.)Transfer long>inong>e temperature: 290 °CCarrier gas: helium, 15 psiInjection: 1 µl splitless (1.0 mong>inong>. purge)Solvent delay 8 mong>inong>., dwell time 100 ms mass -1 , electron multiplier -2200VSIM ion 146Program for detection of phenoxyalkanoic acid herbicides SIM ong>andong> SCANInjection temperature: 280°COven temperature: temp. 1: 60°C (1 mong>inong>.), rate 1: 10°C/mong>inong>.temp. 2: 130°C, rate 2: 6°C/mong>inong>.temp. 3: 260°C, rate 3: 10°C/mong>inong>.temp. 4: 290°C (5 mong>inong>.)Transfer long>inong>e temperature: 290 °CCarrier gas: helium, 15 psiInjection: 1 µl splitless (1.0 mong>inong>. purge)Solvent delay 20 mong>inong>., dwell time 100 ms mass -1 , electron multiplier -2200VSIM ion 181161


Appendix IIMethod for analysis of isoproturon (IPU)50 ml preservation solution is added to the raong>inong>water pr. 2 L bottle before it is placedat the collection site. When the water samples arrive at the laboratory the pH valueis checked. It is supposed to be 1.The raong>inong>water is filtered via büchner funnels through a glass micro fibre filter, thisis done before a possible freezong>inong>g of the water, which is parted ong>inong>to 1 L pr. bottle.Extraction:5 ml methanol / L water is added.The RDX-tubes are pre-treated with:1. 10 ml acetonitrile2. 10 ml methanol3. 20 ml milli-Q waterThe fluid is sucked slowly through every time.It is very important that the tubes are not runnong>inong>g dry.The water samples are fitted with a vacuum suction at about 8 mm Hg.1 L sample is sucked through each column, 2 L from each locality is extracted ong>andong>the two fractions are treated together (same registry number ).The columns are ong>airong> dried with maximum suction ong>inong> further 20 mong>inong>utes.Each column is eluted with 5 ml methanol.First 1 ml is eluted. Then 2 mong>inong>utes stong>andong>ong>inong>g ong>andong> the rest of the methanol is passedthrough the column. The elution is repeated with further 5 ml methanol.Preconcentration:The methanol fraction is evaporated to dryness under a gentle stream of nitrogen atambient temperature.The residue is redissolved ong>inong> 1 ml ethylacetate ong>inong> an ultra sonication bath for 5mong>inong>utes.The ethylacetate fraction is filtered through 0,2 µm filter ong>andong> evaporated to exactly300 µl ong>andong> is now ready for analysis. The vial is kept at –18 o C until analysis.162


Appendix IIIAnalysis of phenoxyalkanoic acid herbicides:50 ml preservation solution is added to the raong>inong>water pr. 2 L bottle before it is placedat the collection site. When the water samples arrive at the laboratory the pH valueis checked. It is supposed to be 1.The raong>inong>water is filtered via büchner funnels through a glass micro fibrefilter, this is done before a possible freezong>inong>g of the water, which is partedong>inong>to 1 L pr. bottle.Extraction:A C8- Empore-disc is placed on the glass filter for millipore filtration.The filter is treated with 10 mL CH 2 Cl 2 . After 3 mong>inong>utes stong>andong>ong>inong>g thesolvent is slowly sucked through the filter. The filter is dried by furthersuction for one mong>inong>ute. The filter is then treated with 10 mL methanol. Thefilter must not go dry before the water sample is passed through the filter.The filter is ong>airong> dried by further suction durong>inong>g 15 mong>inong>utes. The C8-filter iseluted with 10 mL CH 2 Cl 2 , ong>andong> the elution is repeated.Further 10 ml CH 2 Cl 2 is used for transfer of the solutes ong>andong> rong>inong>song>inong>g of theflask. The extracts are transferred to a separation funnel.50 mL 1% NaOH is added ong>andong> the funnel is shaken for 2 mong>inong>utes. After 10mong>inong>utes stong>andong>ong>inong>g the CH 2 Cl 2 fraction is discarded. 3 mL 24 % H 2 SO 4 isadded to the separation funnel, pH less than 1. The funnel is shaken 2 mong>inong>with 20 mL CH 2 Cl 2 . The dichloromethane-fraction is after 10 mong>inong>utesstong>andong>ong>inong>g transferred to a round bottom flask equipped with a funnel withanhydrous sodium sulphate on glass wool. The extraction is repeated twice.Na 2 SO 4 is washed with 5-10 mL dichloromethane. The dichloromethane isevaporated on a rotoevaporator at 38-40 o C at -0,4 bar. The evaporation isstopped when about 500µL is left ong>andong> 2 ml acetone is added. The acetonefraction is filtered through a 0.2 µm filter ong>inong>to a 20 mL test tube. Further 2x2mL acetone is used for rong>inong>song>inong>g of the flask ong>andong> the filter.Derivatisation:25 µL 1% pentafluorbenzylbromide solution ong>andong> 20 mg ong>inong>cong>inong>erated K 2 CO 3is added to the acetone fraction.The test tube is closed an shaken for 5 sec. on a Whirley mixer ong>andong> is thenstong>andong>ong>inong>g over night at ambient temperature. The liquid phase is transferredto a new test tube while the precipitate is cleaned with 1 mL acetone whichis transferred to the test tube.2 mL isooctane is added to the test tube ong>andong> the volume is reduced to about1 ml under a gentle stream of nitrogen. This step is repeated twice ong>andong> theresidual volume is transferred to an amber glass vial passong>inong>g a 0,2 µm filter.163


The test tube is rong>inong>sed twice with max. 6 drops of isooctane which istransferred to the vial. The volume is reduced to exactly 300 µL under agentle stream of nitrogen. The vial is closed with a teflon coated cap ong>andong> iskept at –18 o C until analysis.164


Appendix IVProng>inong>ciples for analysis of water samples collected ong>inong> 1998:The water sample is passed through a solid phase extraction column (SPEcolumn)ong>andong> the retaong>inong>ed compounds are eluted with an organic solvent. Thesolvent is evaporated ong>inong> a vacuum centrifuge ong>andong> the residue is redissolvedong>inong> a suitable solvent for HPLC. The compounds are separated on a HPLCcolumn ong>andong> detected ong>inong> a mass spectrometer after electrospray ionisation(ESI)Chemicals, solutions ong>andong> stong>andong>ards:Methanol, HPLC grade ( Rathburn )Acetonitril, gradient grade (Merck)Acetic acid 100 %, p.a. (Merck),1,2-propanediol (propylene glycol), (Sigma ),Konc. HCl rauchend, 37 %, (Merck),Precervation solution for raong>inong> water: 60,0 ml konc. HCl – op to 1000 mlwith milli-Q-treated water.Deionised water from central osmose plant further cleaned ong>inong> Milli-Qsystem (Millipore). Ammonium acetate, p.a. (Merck).Stong>andong>ards from Dr Ehrendorfer or Riedel de Häen.Stock solutions, stong>andong>ards ong>andong> ong>inong>ternal stong>andong>ards:Stock solutions are made by weighong>inong>g of 50 mg stong>andong>ard, which is dissolvedong>inong> 50 ml acetonitril. Stock solutions are stored ong>inong> refrigerator ong>andong> areregistered ong>inong> a stock solution record.Stong>andong>ards are made by dilution ong>inong> acetonitril ong>andong> fong>inong>al dilution ong>inong> A-eluentor ong>inong> a solvent comparable to HPLC-start conditions. Stong>andong>ardconcentrations should cover the concentration range for the samples afterpreconcentration, for ong>inong>stance 10, 50 ,100 ong>andong> 200 µg/L.13 C- or 2 H-labelled ong>inong>ternal stong>andong>ards are recommended for addong>inong>g to thesamples before preconcentration to checkong>inong>g recovery.HPLC-Eluents:For analysis of isoproturon the followong>inong>g gradient system is used:A1-eluent: methanol/10 mM Ammonium acetate ong>inong> milliQ-water 10/990,B1-eluent: methanol/10 mM ammonium acetate 90/10,For phenoxyalkanoic acid herbicides ong>andong> other acidic compounds:A1-eluent: methanol/ 20 mM acetic acid 10/90,B1-eluent: 20 mM acetic acid ong>inong> methanol.The eluents are filtered through 0.2 µm millipore filter (type Fluropore, FG),Millipore.LC-conditions165


The LC-MS system was composed by a Hewlett-Packard LC-MSD 1100system with a bong>inong>ary gradient pump system.The HPLC-column was a Hypersil BDS 250 x 2.1 mm column.The same long>inong>ear gradient profile were used for both acidic compounds ong>andong>isoproturon compounds with different eluents, as described above:Time, mong>inong> B-solvent, %0 03 5030 10033 10036 045 0Injection volume: 50 µl.Flow: 0,2 ml/mong>inong>Column temperature: 30 o C.Mass spectrometry:Phenoxyalkanoic acid herbicides, bentazon ong>andong> – metabolites together withDNOC was detected by MS with electrospray ong>inong>let (ESI), negative modeong>andong> selected ion monitorong>inong>g (SIM):Dryong>inong>g gas temperature: 350 o CDryong>inong>g gas flow: 10,0 L/mong>inong>.Nebulizer pressure: 40 psigCapillary voltage: 4000 V.The followong>inong>g masses were measured (m/z): 197, 199, 201, 213, 215, 233,235, 239, 255, 380,Dwell time 146 ms.Fragmentor voltage 80 V.Isoproturon ong>andong> degradation products were detected by MS withelectrospray ong>inong>let (ESI),positive mode ong>andong> selected ion monitorong>inong>g (SIM):Dryong>inong>g gas temperature: 350 o CDryong>inong>g gas flow: 10,0 l/mong>inong>.Nebulizer pressure: 40 psigCapillary voltage: 4000 V.The followong>inong>g masses were measured (m/z): 136, 179, 193, 207.Dwell time 146 ms.Fragmentor voltage 80 V.Sample preparation:50 ml preservation solution is added to the sample flasks before they areplaced on the locations.The samples are stored at 4 o C if sample preparation takes places withong>inong> afew days. Otherwise the samples are frozen at –18 o C until samplepreparation will take place.166


The sample volume is registered ong>andong> eventually ong>inong>ternal stong>andong>ard is addedThe sample is passed through a glass fibre prefilter ong>andong> a glass fibre filterong>andong> the filters are rong>inong>sed with 5 ml methanol.The filtrate or 250 ml water sample ( + 1,25 ml Methanol ) (+ 6,25 mlpreservation solution to recovery ong>andong> blanks) are added to Porapak Rdxsolid phase extraction columns, which are preconditioned with 10 mlacetonitril, 10 ml Methanol ong>andong> 20 ml Milli-Q water. The water sample isapplied under vacuum pressure with a flow about 10-20 ml/mong>inong>. Thecolumns must never run dry durong>inong>g the application.Afterwards the columns are ong>airong> dried with vacuum suction ong>inong> further 20 mong>inong>.The columns are eluted with 5 ml methanol/acetonitril 1/1 ong>inong> the followong>inong>gway: 1 ml is passed through the column, 2 mong>inong>utes stong>andong>ong>inong>g ong>andong> the rest ofthe solvent is passong>inong>g without vacuum ong>inong>to a vial with 50 µL propyleneglycol as keeper.The sample volume is reduced to 50 µl ong>inong> a vacuum centrifuge. Thepropylene glycol residue is diluted with 500 µl 10% methanol ong>andong> is readyfor HPLC-analysis.167


168


Effects of herbicides ong>inong> ong>precipitationong> onplants ong>andong> plant communitiesSolvejg Kopp Mathiassen & Per KudskDanish Institute of Agricultural SciencesDepartment of Crop Protection169


170


IndholdSUMMARY 177SAMMENDRAG 1781 INTRODUCTION 1792 MATERIALS AND METHODS 1802.1 EXPERIMENTS FOR DETERMINING ED 10 DOSES 1802.2 COMPETITION EXPERIMENTS 1813 RESULTS AND DISCUSSION 1833.1 DETERMINING ED 10 DOSES 1831.1.1 Song>inong>apis alba 1831.1.2 Thlaspi arvense 1831.1.3 Stellaria media 1833.2 COMPETITION EXPERIMENTS 1844 CONCLUSIONS 1895 ACKNOWLEDGEMENT 1906 REFERENCES 191171


172


SummaryIn several European studies pesticides have been found ong>inong> ong>precipitationong>. Thepotential risk of adverse ong>effectsong> of herbicides ong>inong> ong>precipitationong> on plants ong>andong> plantcommunities ong>inong> the agro-ecosystem ong>andong> adjacent ecosystem can be assessed if theNo Observable Effect Level (NOEL) is known. In this project NOEL was defong>inong>edas the dose required to reduce fresh or dry weight by 10% (ED 10 ). Mecoprop-P wasselected as model pesticide ong>andong> the ED 10 dose was determong>inong>ed on a number ofsusceptible plant species at different growth stages. In addition, the competitionong>inong>deces between selected species grown ong>inong> bong>inong>ary mixtures were compared withoutherbicide treatment ong>andong> followong>inong>g application of mecoprop-P doses close to ED 10.The ED 10 dose of the most susceptible plant species ong>inong>cluded ong>inong> this study wasmore than 3 times higher than the maximum yearly deposition of mecoprop ong>inong>Danish raong>inong> water. In the study, different methods were applied to assess ong>effectsong> ofthe low doses of mecoprop-P. The results showed that biomass production was assusceptible a parameter as seed production. The substitution rate between plantspecies with similar competitive ability but different susceptibility to mecopropwas not significantly ong>inong>fluenced of doses close to ED 10 .173


SammendragI såvel ong>inong>ternationale som danske undersøgelser er der påvist forekomst afpesticider i nedbør. For at kunne vurdere hvorvidt disse forekomster udgør enrisiko i økotoksikologisk sammenhæng, er der behov for viden om hvor lavekoncentrationer af pesticider, der kan forventes at resultere i målbare effekter ong>inong>aturen. Af hensyn til den statistiske bearbejdnong>inong>g af resultaterne har vi valgt atdefong>inong>ere NOEL (No Observable Effect Level), som den doserong>inong>g, der resulterer i10% reduktion af frisk- eller tørvægten på de behong>andong>lede planter (=ED 10 ).Mechlorprop-P er anvendt som modelstof ,og ED 10 doserong>inong>gen af dette herbicid erbestemt overfor en række følsomme plantearter (Agersennep, Pengeurt,Hyrdetaske, Agertidsel, Haremad, Hvidmelet gåsefod og Fuglegræs).Bestemmelsen er foretaget ved behong>andong>long>inong>g på forskellige udviklong>inong>gstrong>inong>, vedforskellige høsttidspunkter og ved målong>inong>g af biomasse og frøproduktion. Desudener det undersøgt, om konkurrenceevnen af en følsom art overfor en mong>inong>dre følsomart påvirkes ved behong>andong>long>inong>g med doserong>inong>ger af mechlorprop-P omkrong>inong>g ED 10 .Resultaterne viste, at der ikke er målbare effekter af de mængder af mechlorprop-P,som er fundet i opsamlede regnvong>andong>sprøver , idet ED 10 doserong>inong>gen på den mestfølsomme planteart var ca. 10 gange højere end det maksimale fund i en enkeltudtagnong>inong>g og 3 gange højere end den gennemsnitlige årlige deponerong>inong>g pr.arealenhed. Undersøgelsen har desuden vist, at hverken frøproduktion ellerkonkurrenceevnen er mere følsomme parametre end biomasseproduktion.174


1 IntroductionSeveral European studies have revealed that pesticides can occur ong>inong> ong>precipitationong>(Sibers et al., 1995; Jaeschke et al., 1995, Lode et al., 1995; Kreuger, 1995). In theyears 1992-94 raong>inong> water collected at two locations ong>inong> Denmark was analysed forthe content of 10 pesticides but only phenoxyalkanoic acid herbicides were foundong>andong> this group of herbicides occured at concentrations up to 0.4 mg/l (Kirknel &Feldong>inong>g, 1995). For this study we therefore selected mecoprop as model herbicide.Mecoprop was extensively used ong>inong> Denmark ong>inong> 1996 when the project wasong>inong>itiated, however ong>inong> the followong>inong>g year the usage of phenoxyalkanoic acidherbicides was severely restricted resultong>inong>g ong>inong> a 80-85% reduction ong>inong> theconsumption from 1996 to 1997.It seems probable that herbicides ong>inong> precipation can affect the growth ong>andong>reproductivity of terrestrial plants. However no attempts have been made so far toassess the possible ong>effectsong> of herbicides ong>inong> ong>precipitationong> on plants ong>andong> plantcommunities ong>inong> the agro-ecosystems ong>andong> adjacent ecosystems.The objective of this part of the project was to assess the potential risk of adverseong>effectsong> of mecoprop-P present ong>inong> ong>precipitationong> on plants ong>andong> plant communities ong>inong>the agro-ecosystem.The No Observable Effect Level (NOEL) is a recognized ecotoxicological termdesignatong>inong>g the highest dose havong>inong>g no ong>effectsong> on the growth of the test organism.As the experimental design ong>andong> number of replications can have a markedong>inong>fluence on the estimation of NOEL it is more convenient to defong>inong>e NOEL as thedose required to reduce growth by 5 or 10%. In this project the dose resultong>inong>g ong>inong> a10% reduction ong>inong> growth of the test plants (ED10) has been used as a measure forNOEL.175


2 Materials ong>andong> methodsThe ong>effectsong> of low doses of mecoprop were examong>inong>ed ong>inong> two different types ofexperiments:1. The ED 10 dose of mecoprop-P was determong>inong>ed on a number of susceptible plantspecies2. The competition ong>inong>deces between selected species grown ong>inong> bong>inong>ary mixtureswere compared followong>inong>g no application of herbicide ong>andong> application ofmecoprop-P doses close to the ED 10 .2.1 Experiments for determong>inong>ong>inong>g ED 10 dosesTest plantsThe ED 10 doses of mecoprop-P were determong>inong>ed on Capsella bursa-pastoris,Thlaspi arvense, Song>inong>apis alba, Circium arvense, Lapsana communis Chenopodiumalbum ong>andong> Stellaria media. All species are known to be very susceptible tophenoxyalkanoic acid herbicides.The experiments were carried out ong>inong> a glasshouse ong>inong> the period from April 15 toOctober 15. Seeds were sown ong>inong> 2 L pots ong>inong> a soil/ song>andong>/ peat mixture (2:1:1 w/w)contaong>inong>ong>inong>g all necessary macro ong>andong> micro nutrients. After emergence the number ofplants per pot was reduced to a similar number.Application ong>andong> dosesThe treatments with mecoprop were carried out on different growth stages of theplants varyong>inong>g from two to eight leaves. In one experiment Thlaspi arvense wasalso exposed to mecoprop-P at the flowerong>inong>g stage. Mecoprop-P was applied usong>inong>ga laboratory pot sprayer equipped with two Hardi 4110-14 nozzles ong>inong> a sprayvolume varyong>inong>g between 130 ong>andong> 165 L ha -1 . Each plant species was treated with 7to 10 doses of mecoprop-P (Duplosan MP, 600 g L -1 mecoprop-P). In order toobtaong>inong> responses from 0 to 100% effect the doses were varied between plantspecies ong>andong> growth stages. Each treatment was replicated three times usong>inong>g acompletely rong>andong>omised layoutPlants were harvested 3-4 weeks after treatment ong>andong> fresh ong>andong> dry weights werecollected. In one experiments the ong>inong>fluence of time of harvest was examong>inong>ed byharvestong>inong>g at two different time ong>inong>tervals (2 respectively 4 weeks) after treatment.This experiment ong>inong>cluded also a comparison of the effect of mecoprop-P on seedversus biomass production.Model used for For each plant species ong>andong> growth stage dose response curves were estimatedestimatong>inong>g dose usong>inong>g non-long>inong>ear regressions. A four parameter logistic model was used toresponse curves describe the relationship between plant weight ong>andong> dose :D − CU =1+exp(2b(log(ED10)−1.099 / b − log( z)))In this model U is the fresh- or dry weight of the plants, D denotes the upper limitof plant biomass at zero dose, C the lower limit at high doses, b is the slope, z thedose ong>andong> ED 10 is the dose resultong>inong>g ong>inong> 10% reduction of growth.The regression model was evaluated by a test for lack of fit comparong>inong>g the sum ofresiduals from the regression ong>andong> variance analyses.176


2.2 Competition experimentsTest plants In the competition experiments different densities of C. bursa-pastoris ong>andong> G.dissectum were grown ong>inong> monoculture ong>andong> ong>inong> bong>inong>ary mixtures varyong>inong>g the ratio ong>andong>density of the two species.At the 1-2 leaf stage plants of the two species were transplanted to 40 by 40 cmpolystyren boxes (growth area 35 by 35 cm) filled with a soil/song>andong>/peat mixture(2:1:1 w/w) contaong>inong>ong>inong>g all necessary micro- ong>andong> macro nutrients.Experimental designThe experiments were based on a complete additive design as proposed by Cousens(1991). Plants were transplanted ong>inong> a geometric series of plant densities usong>inong>gregular patterns (Figure 1). The density of each plant species varied between 1 ong>andong>32 plants per box coverong>inong>g scenarios from no competition to high competitionong>inong>tensity. For each plant species, 5 different densities were used givong>inong>g 25 differentcombong>inong>ations. In addition a high density (64 plants per box) of each species alonewas ong>inong>cluded ong>inong> order to assess the ong>inong>traspecific competition at a high density.The plants were placed ong>inong> a glasshouse ong>andong> watered daily. The boxes were placedong>inong> blocks accordong>inong>g to the total number of plants ong>inong> the boxes. One block ong>inong>cludedboxes with less than 20 plants, another ong>inong>cluded boxes with 20 to 40 plants ong>andong> ong>inong>the last block boxes with more than 40 plants were placed.7264No. of plants /boxCapsella bursa-pastoris564840322416800 8 16 24 32 40 48 56 64 72No. of plants /boxGeranium dissectumFigure 1Composition of two component mixtures of G. dissectum ong>andong> C. bursa-pastoris ong>inong> the experiments. The axesshow the number of plants per box (growth area 35 by 35 cm).Sammensætnong>inong>g af to-komponent blong>andong>ong>inong>ger af G. dissectum og C. bursa-pastoris i forsøgene. Akserne viserantallet af hver planteart pr. kasse (vækstareal 35 x 35 cm).Application ong>andong> dosesMecoprop-P was applied to the plants at the 6-8 leaves stage. Each combong>inong>ation ofplant densities ong>andong> ratios were represented by an untreated ong>andong> one or two doses ofmecoprop-P close to the expected ED 10 dose of C. bursa-pastoris. Each treatmentwas replicated twice. Mecoprop-P was applied usong>inong>g a laboratory pot sprayerequipped with two Hardi 4110-14 flat fan nozzles producong>inong>g a spray volume of 150L ha -1 .177


Concurrently with the competition experiments the ED 10 doses of each of the twoplant species were estimated ong>inong> pot experiments.The plants were harvested 3 weeks after treatment. The outer 5 cm of the boxeswas not harvested. The number of plants of each species was counted ong>andong>harvested separately. Fresh- ong>andong> dry weights were collected.The biomass per plant y 1 ong>andong> y 2 of species 1 (C. bursa-pastoris) ong>andong> species 2 (G.dissectum) can be calculated when the number of plants n 1 ong>andong> n 2 ong>andong> the totalplant biomass x 1 ong>andong> x 2 are known.y 1 =xn11y 2 =xn22Competition modelIf however we wish to describe the biomass y 1 as a function of n 1 ong>andong> n 2 the ong>inong>verslong>inong>ear model suggested by Spitters (1983) can be used:1y1= B0+B1n1+ B2n2In this model B 0 , B 1 ong>andong> B 2 are unknown parameters. As y 1 ong>andong> y 2 are the mostrelevant factors ong>inong> this context the equation was rewritten as:y1=B0+1+ B2nB1n12Madsen et al. (1995) reparameterized this equation ong>inong> order to give all parameters abiological meanong>inong>g:Ay1=A1 + ( n1+ Cn2− n0)By 1 is now the biomass per plant of species 1 when grown ong>inong> a bong>inong>ary mixturecontaong>inong>ong>inong>g n 1 plants of species 1 ong>andong> n 2 plants of species 2 per box. A is thebiomass per plant at an arbitrar zero density (n 0 ), which ong>inong> the analyses was set to8. B denotes the maximum yield per box at ong>inong>fong>inong>ite plant densities ong>andong> C is thesubstitution rate. If plants compete for the same resources C can be used to decribethe competition rate of species 2 agaong>inong>st species 1. If C >1 species 2 is morecompetitive than species 1 ong>andong> vise versa. The arbitrary value for n o is used toavoid extrapolation to non-observed densities as well as devotong>inong>g a welldefong>inong>edbiological explanation for B (Fredshavn, 1994).When the substitution rate C is known the effective density n e can be calculated:ne = n1 + Cn2178


3 Results ong>andong> discussion3.1 Determong>inong>ong>inong>g ED 10 dosesSusceptibility of different Table 1 shows the estimated ED 10 doses. The most susceptible speciesplant species were S. alba ong>andong> C. bursa-pastoris with ED 10 doses lower than 1 g a.i. ha -1while the ED 10 doses of T. arvense, C. album ong>andong> C. arvense were 3 to 4 ga.i. ha -1 . L. communis ong>andong> S. media were the most tolerant species with ED 10 doseshigher than 5 g a.i. ha -1 . One would expect the ED 10 dose to ong>inong>crease with growthstage, however as the experiments were carried out at different times of the yearfactors other than growth stage could have an ong>inong>fluence on the susceptibility of theplants.Table 1Estimated ED 10 doses of mecoprop on different species ong>andong> growth stages. Thefigures ong>inong> parentheses are mong>inong>imum ong>andong> maximum of the estimated mean values.Estimerede ED10 doserong>inong>ger af mechlorprop på forskellige arter og udviklong>inong>gstrong>inong>.Tallene i parentes er mong>inong>imum og maximum af de estimerede værdier.Plant species Growth stage ED 10(g a.i.ha -1 )Numberof trials1.1.1 Song>inong>apis alba 2-3 leaves 0.6 (0.5; 0.7) 34-5 leaves 0.8 11.1.2 Thlaspi arvense 2-3 leaves 3.0 14-5 leaves 2.8 16-8 leaves 3.2 (0.9;7.3) 39-10 leaves 3.3 (1.8; 4.3) 3Capsella bursa- 4-5 leaves 1.2 (1.1; 1.3) 3pastoris 6-8 leaves 0.6 (0.4; 0.8) 29-10 leaves 0.5 1Circium arvense 2-3 leaves 5.0 (1.2; 9.3) 5Lapsana communis 2-3 leaves 12.6 (8.5; 16.6) 24-5 leaves 5.0 (2.4; 7.5) 2Chenopodium album 4-5 leaves 2.6 (0.5; 5.3) 31.1.3 Stellaria media 4-5 leaves 6.5 (5.5; 7.6) 26-8 leaves 5.6 (2.6; 8.5) 2Deposition of mecoprop The maximum yearly deposition of mecoprop calculated on basis of theby raong>inong>analysis of the raong>inong> samples was 0.2 g mecoprop ha -1 (Table 3.15). Consequently,the ED 10 dose of the most susceptible species was more than 3 times higher than179


the maximum yearly deposition by raong>inong> ong>andong> 10 times the maximum concentrationfound ong>inong> one sample.Influence of time ofharvestIn the experiments shown ong>inong> Table 1 the plants were harvestedapproximately 3 weeks after treatment. In practise plants will often grow for alonger period after exposure to herbicides which raises the question whether theywill gradually recover. Here the ong>inong>fluence of the time of harvest on the fong>inong>al effectwas determong>inong>ed ong>inong> one experiment where plants were harvested at different timeong>inong>tervals after treatment. In addition the susceptibility to mecoprop-P at theflowerong>inong>g stage as well as the effect on seed production was examong>inong>ed. The resultsare shown ong>inong> Table 2.Table 2The ong>inong>fluence of growth stage at treatment ong>andong> time of harvest on ED 10 doses ofThlaspi arvense estimated on biomass ong>andong> seed production, respectively. Figures ong>inong>parentheses are stong>andong>ard errors.Betydnong>inong>g af udviklong>inong>gstrong>inong> på behong>andong>long>inong>gstidspunktet og høsttidspunkt for e ED 10doserong>inong>ger på Thlaspi arvense beregnet udfra henholdsvis biomasse ogfrøproduktion. Tallene i parentes er stong>andong>ardafvigelser.Growth stageInterval betweentreatment ong>andong>harvestBiomassED 10 dose (g.a.i. ha -1 )Seed production8-10 leaves 2 weeks 3.9 (2.5)4 weeks 1.8 (1.4)> 15 weeks 10.7 (4.6)Flowerong>inong>g 2 weeks 1.3 (0.6)> 15 weeks 3.5 (2.3)When treated at the 8 to 10 leaves stage no significant differences were foundbetween harvest 2 ong>andong> 4 weeks after treatment. Surprisong>inong>gly biomass productionwas ong>inong>hibited more when plants were exposed to mecoprop-P at the flowerong>inong>gstage compared to the 8 to 10 leaves stage. At both growth stages the ED 10 dosesestimated on basis of seed production were significantly higher than thoseestimated on basis of biomass production.3.2 Competition experimentsCharacteristics of testplantsSusceptibility toThree competition experiments were carried out. The purpose of the firstexperiment was to determong>inong>e plant species ong>andong> densities to be used ong>inong> thesubsequent experiment (results not shown). C. bursa-pastoris ong>andong> G. dissectumwere selected as two plant species possessong>inong>g the same competitive ability butdifferong>inong>g ong>inong> susceptibility to mecoprop. The growth habit of these two plant speciesdiffer widely. C. bursa-pastoris is a relative small plant which elongate at an earlystage where as G. dissectum forms a vigorous roset at an early growth stage. C.bursa-pastoris is known to be much more susceptible to mecoprop thanG.dissectum.In subsequent experiments, the competition between the two species without180


mecopropSusceptibility of A ong>andong>B parametersherbicide ong>andong> after treatment with 1 or 2 doses of mecoprop-P was determong>inong>ed. Inone trial a dose of 3 g a.i. ha -1 of mecoprop-P was applied while ong>inong> the second trial0.5 ong>andong> 2.0 g a.i. ha -1 were applied. The ED 10 doses estimated on the pot-grownplants grown ong>inong> monoculture revealed that the ED 10 doses of mecoprop-P on C.bursa-pastoris ong>andong> G. dissectum were respectively 0.4 g a.i. ha -1 ong>andong> 8.2 g a.i. ha -1 .Table 3 shows the estimated parameters ong>inong> the two competition experiments.In both experiments the A parameter (weight per plant at n 0 =8 plants per box) ofC. bursa-pastoris was significantly higher than the A parameter of G. dissectum. Inexperiment 1, the applied dose of mecoprop-P was much higher than the ED 10 doseof C. bursa-pastoris ong>andong> the A parameter of C. bursa-pastoris was significantlyreduced while no effect was found on G. dissectum. A tendency to a reduction ofthe B parameter of C. bursa-pastoris (the maximum production per box) was alsoseen.In experiment 2 the A ong>andong> B parameters were unaffected of the lowest dose ofmecoprop-P while on C. bursa pastoris both parameters tended to decrease withong>inong>creasong>inong>g dose. The B parameter of G. dissectum also tended to decrease withong>inong>creased dose.Susceptibility of CparameterThe ong>inong>fluence of herbicide treatments on the ong>inong>terspecific competition can beassessed by the C parameter. No significant differences ong>inong> C parameters werefound between the species ong>andong> doses ong>inong>dicatong>inong>g that the ong>inong>terspecific competitionwas equal. In experiment 1, the estimated C parameters on G. dissectum was 1.7for untreated ong>andong> 0.9 after treatment with 3 g a.i. ha -1 . Consequently, 1 plant of C.bursa-pastoris can be replaced by 1.7 ong>andong> 0.9 plants of G. dissectum respectively.Based on the C. bursa-pastoris data the results showed that when untreated 1 plantof G.dissectum could be replaced by 1.4 plants of C. bursa-pastoris while thesubstitution rate after treatment with 3 g a.i. ha -1 of mecoprop was 1. Herbicidetreatment halves the substitution rate of G. dissectum while the change of thesubstitution rate of C.bursa-pastoris was lower ong>inong>dicatong>inong>g that the appliedmecoprop dose had more ong>inong>fluence on the growth of C. bursa-pastoris than of G.dissectum. However, as none of the estimated C parameters differed significantlyfrom 1 ong>andong> as they were not significantly affected by the applied doses it can beconcluded that the competitiveness of the species was not significantly ong>inong>fluencedby the applied mecoprop-P dose.In experiment 2 the C parameter of G. dissectum tended to decrease when 0.5 g a.i.mecoprop was applied confirmong>inong>g that C. bursa-pastoris was more susceptible thanG. dissectum. Similarly the number of G. dissectum plants necessary to substituteone C. bursa-pastoris plant was lower after herbicide treatment. However when thedose was ong>inong>creased to 5 times the ED 10 dose the C parameter tended to ong>inong>crease,which we can not explaong>inong>. None of the differences were significant.Figure 2 shows the observed ong>andong> predicted A values (biomass per plant) as afunction of n e (effective plant density). The curves of C. bursa-pastoris are moresteep than the correspondong>inong>g curves of G. dissectum, ong>inong>dicatong>inong>g that C. bursapastoris was more susceptible to ong>inong>creasong>inong>g plant density than G. dissectum. Thiswas not reflected as a difference ong>inong> the competition ability because ong>inong>traspecificcompetition played a major role with C. bursa-pastoris.In the competition experiments, the two selected plant species, although differong>inong>gwidely ong>inong> growth habits, were expected to possess the same competition abilitywhereas the susceptibility to mecoprop differed. Followong>inong>g application of amecoprop-P dose that would ong>inong>fluence the growth of the most susceptible species,one would expect a change ong>inong> the competition ong>inong>dex favourizong>inong>g the most tolerant181


species. However, no significant differences ong>inong> the competitive ability of C. bursapastorisong>andong> G. dissectum were observed when applyong>inong>g a mecoprop-P dose closeto the ED 10 dose of C. bursa-pastoris or a dose 5 times higher. Therefore it can beconcluded that when usong>inong>g the experimental design adopted ong>inong> this studycompetition was not a more susceptible factor to study than biomass production.Table 3Estimated parameters ong>inong> the competition experiments (fresh weight). The figures ong>inong>parentheses are 95% confidence ong>inong>tervals.Estimerede parametre i konkurrenceforsøgene (friskvægt). Tallene i parentes er95% konfidens ong>inong>tervaller.Exp. Plant species Dose A (g/box) B (max. Prod) C1 G. dissectum 0 7.5(6.6-8.4)G. dissectum 3 6.6(5.8-7.5)C. bursa-pastoris 0 57.0(47.3 – 66.6)C. bursa-pastoris 3 39.0(31.6-46.3)344.4(205.5-483.4)429.0(225.3-632.5)693.7(538.3 – 849.1)444.6(337.6 –551.6 )1.7(0.8-2.6)0.9(0.3-1.5)1.4(0.8-1.9)1.0(0.5-1.5)2 G. dissectum 0 10.9(9.8-11.9)G. dissectum 0.5 10.5(8.6-12.3)G. dissectum 2.0 9.6(8.0 –11.2)C. bursa-pastoris 0 31.2(25.7-36.8)C. bursa-pastoris 0.5 31.0(25.8-36.2)C bursa-pastoris 2.0 27.5(22.5-32.5)328.0(199.6 –456.9 )299.3(177.7 –420.9 )362.1(194.5- 529.7 )345.4(256.2– 434.7 )342.1(256.6-427.6)301.5(222.4–380.6 )1.9(0.8-2.9)1.6(0.7-2.5)2.3(0.9-3.6)1.2(0.7-1.8)1.3(0.8-1.9)1.4(0.8-2.0)Fresh weightg/plant Control a182


Effective densityFresh weightg/plant 0.5 g/ha mecoprop-P bEffective densityFresh weightg/plant 2.0 g/ha mecoprop-P cEffective densityFigure 2The ong>inong>fluence of effective plant number on production of biomass per plant (A) of C. bursa-pastoris (a-c) ong>andong> G.dissectum (d-f) without herbicide ong>andong> after treatment with 0.5 or 2 g a.i. ha -1 of mecoprop. Experiment 2.Indflydelse af effektivt plantetal på biomasseproduktionen pr. plante af C. bursa-pastoris (a-c) og G. dissectum(d-f) uden herbicid og efter behong>andong>long>inong>g med 0.5 og 2 g ha-1 mechlorpropFresh weightg/plant Control d183


Fresh weightg/plant 0.5 g/ha mecoprop-P eEffective densityEffective densityFresh weightg/plant 2.0 g/ha mecoprop-P fEffective densityFigure 2 (contong>inong>ued)The ong>inong>fluence of effective plant number on production of biomass per plant (A) of C. bursa-pastoris (a-c) ong>andong> G.dissectum (d-f) without herbicide ong>andong> after treatment with 0.5 or 2 g a.i. ha -1 of mecoprop. Experiment 2.Indflydelse af effektivt plantetal på biomasseproduktionen pr. plante af C. bursa-pastoris (a-c) og G. dissectum(d-f) uden herbicid og efter behong>andong>long>inong>g med 0.5 og 2 g ha -1 mechlorprop.184


4 ConclusionsThe potential risk of adverse ong>effectsong> of mecoprop ong>inong> ong>precipitationong> has been assessedby applyong>inong>g different methods. The results have revealed that biomass is assusceptible a parameter as seed production. The substitution rate between specieswith similar competitive ability but different susceptibility to the herbicide was notsignificantly ong>inong>fluenced by doses close to ED 10 .The results have shown that the No Observable Effect Level (ong>inong> this study defong>inong>edas the ED 10 dose) of mecoprop-P on the most susceptible plant species ong>inong>cluded ong>inong>this study was more than 3 times higher than the maximum yearly deposition ofmecoprop ong>inong> Danish raong>inong> water ong>andong> 10 times higher than the maximum depositionwithong>inong> a two weeks period. Consequently, ong>effectsong> of mecoprop ong>inong> ong>precipitationong> isnot very likely.In this project we only studied the ong>inong>fluence of a song>inong>gle herbicide, however theanalyses of raong>inong> water have revealed the presence of several herbicides. A questionto address ong>inong> a future project is the possible ong>effectsong> of such ‘pesticide cocktails’ onterrestrial plants.185


5 AcknowledgementWe wish to thank Dr Jens Erik Jensen at the Royal Veterong>inong>ary ong>andong> AgriculturalUniversity for statistical assistance ong>inong> analyzong>inong>g the competition experiments.186


6 ReferencesCousens R. (1991). Aspects of the design ong>andong> ong>inong>terpretation of competition(ong>inong>terference) experiments. Weed Technology, 5 (3), 664-673.Fredshavn J.R. (1994). Use of substitution rates to describe competition ong>inong> mixedplant populations. Acta Agricultura Scong>andong>ong>inong>avica, Section B, Soil ong>andong> PlantScience, 44 (1), 47-54.Jaeschke W., Gath B. & Pfäfflong>inong> D. (1995). Deposition of pesticides ong>inong> SouthernGermany. In: ong>Pesticidesong> ong>inong> ong>precipitationong> ong>andong> surface water, Tema Nord, 1995:558.Ed. A. Helweg, 65-74.Kirknel E. & Feldong>inong>g G. (1995). Analysis of selected pesticides ong>inong> raong>inong> ong>inong> Denmark.In: ong>Pesticidesong> ong>inong> ong>precipitationong> ong>andong> surface water, Tema Nord, 1995:558. Ed. A.Helweg, 45-54.Kreuger J. (1995). Pesticider i regnvatten i Sverige. In: ong>Pesticidesong> ong>inong> ong>precipitationong>ong>andong> surface water, Tema Nord, 1995:558. Ed. A. Helweg, 33-44.Lode O., Eklo O.M., Holen B., Svensen A & Johnson Å.M. (1995). ong>Pesticidesong> ong>inong>ong>precipitationong> ong>inong> Norway. In: ong>Pesticidesong> ong>inong> ong>precipitationong> ong>andong> surface water, TemaNord, 1995:558. Ed. A. Helweg, 19-32Madsen K.H., Poulsen G.S., Fredshavn J.R., Jensen J.E. & Steen P. (1995). Amethod to study competitive ability of seabeet hybrids (Beta vulgaris ssp.maritima) ong>andong> glyphosate tolerant sugarbeet (Beta vulgaris ssp. vulgaris). ActaAgricultura Scong>andong>ong>inong>avica, Section B, vol. 48, no. 3, 170-174.Sibers J., Gottschild D. & Noltong>inong>g H.-G. (1995). Deposition of pesticides ong>inong>Nothers Germany. In: ong>Pesticidesong> ong>inong> ong>precipitationong> ong>andong> surface water, Tema Nord,1995:558. Ed. A. Helweg, 55-64.Spitters C.J.T. (1983). An alternative approach to the analysis of mixed croppong>inong>gexperiments. 1. Estimation of competition ong>effectsong>. Netherlong>andong>s Journal ofAgricultural Sciences, 31, 1-11.187


188


Fong>inong>al conclusionThe report describes how volatility of pesticides can be determong>inong>ed by a laboratorymodel system; further models for pesticide diffusion ong>andong> transport ong>inong> theatmosphere as well as deposition are described ong>andong> evaluated. Precipitation wascollected ong>andong> analysed for selected pesticides over a couple of years ong>andong> further theeffect of the herbicide mecoprop was determong>inong>ed on a number of sensitive testplants to see whether the found pesticide concentrations ong>inong> raong>inong>water possess athreat to plant communities.The laboratory experiments have elucidated a great need for stong>andong>ardisedlaboratory systems to determong>inong>e the volatility of pesticides. Especially the testconditions need to be described carefully.Atmospheric transport ong>andong> deposition models for pesticides have been developed,but ong>inong>formation is lackong>inong>g on some of the basis processes as emission, drydeposition, atmospheric reactions ong>andong> conversion from the gaseous to theparticulate phase. Information on transport, diffusion ong>andong> wet deposition processesfor pesticides is sufficiently known. The description of emission ong>andong> dry depositionong>inong> such models can be improved if the processes that determong>inong>e the concentrationong>inong> the surface (soil, vegetation, water) are modelled at the same time.The concentrations of pesticides found ong>inong> raong>inong>water seemed generally to be of thesame order of magnitude as ong>inong> other European countries. The pesticides appearedong>inong> the highest concentrations ong>inong> the sprayong>inong>g season, ong>andong> when the use of phenoxyherbicideswas limited ong>inong> Denmark, they were only rarely found ong>inong> theong>precipitationong>. This is known not to be true for the chlorong>inong>ated hydrocarbonong>inong>secticides, which may be transported over long distances. The herbicide DNOCwas found ong>inong> the highest concentrations, but this compound seems to be formed ong>inong>the atmosphere.When the ong>effectsong> of pesticide concentrations found ong>inong> raong>inong> water were determong>inong>edon susceptible plant species, the experiments showed, that the NOEL (NoObservable Effect Level) for mecoprop was more than three times higher than themaximum yearly deposition of mecoprop ong>inong> Denmark. Tests have only beenperformed with song>inong>gle pesticides ong>andong> need to be performed with the manycombong>inong>ations of pesticides that are normally found ong>inong> raong>inong>water.Determong>inong>ation ofvolatilisationVolatilisation of pesticides are ong>inong>fluenced by many different factors underfield conditions. It is difficult to study the ong>inong>fluence of ong>inong>dividual parameters (e.g.temperature, wong>inong>d speed) on the emission with field experiments. Thereforestong>andong>ardised laboratory experiments can be a helpful tools. Song>inong>ce the GermanBBA-guidelong>inong>es appeared ong>inong> 1990 different methods have been developed.In a test of the different methods with three pesticides, it was found, that chamberproperties ong>andong> size ong>andong> experimental area ong>andong> ong>airong> exchange rates are the mostimportant for the determong>inong>ation of volatilisation. Great variations between thedifferent methods were found, which showed that it is very difficult to compareresults from different test methods.189


In this project is developed a laboratory system which makes it possible todetermong>inong>e evaporation from different surfaces ong>andong> to compare different pesticides.Based on the fong>inong>dong>inong>gs from the experiments it is concluded, that:• The BBA guidelong>inong>e should be improved ong>andong> the description of the chamber ong>andong>the conditions should be more specific• Volatilisation chamber must be specified (a small chamber is recommended)• Samplong>inong>g after 1, 3, 6 ong>andong> 24 hours seems reasonable• It is important to measure the height over the surface ong>andong> whether turbulent orlamong>inong>ar flow is used when ong>airong> velocity is determong>inong>ed. A velocity of less than 1m per sec. 1 to 5 mm above the surface might be sufficient.• Approximately 50% relative humidity should be used ong>andong> the surface to whichthe pesticide is applied should have the same humidity durong>inong>g the wholeexperiment• Filter paper may be used to mimic leaf surface• Stong>andong>ard soil with specified pH should be kept at 60% of maximum waterholdong>inong>g capacity durong>inong>g the whole experiment• It should be considered not to perform experiments with soil but only fromartificial surface song>inong>ce volatility is always highest from the artificial surfaces.In this way a maximum emission rate can be determong>inong>ed.• Several concentrations of a given pesticide should be tested ong>andong> experimentmust be carried out with both formulated product ong>andong> active ong>inong>gredient• Preferably the test should be carried out at 10 ong>andong> 30°CModellong>inong>g atmospheric The behaviour of pesticides ong>inong> the atmosphere is very much dependent upontransport ong>andong> deposition their solubility ong>inong> water ong>andong> their vapour pressure. Modellong>inong>g of the fate ofpesticides ong>inong> the atmosphere is difficult, because ong>inong>formation about importantproperties of pesticides is not available or is uncertaong>inong>. Further it is very difficult togeneralise their behaviour, because the properties of pesticides are so different,there is a large variety of surfaces with different properties. For the approval ofpesticides general ong>inong>formation is provided by the manufactures, but this is notsufficient to predict their behaviour ong>inong> the atmosphere.Emission ong>andong> dry deposition depend on atmospheric turbulence, temperature ong>andong>properties of the surface of soil ong>andong> plants ong>andong> the processes that take place ong>inong> thesurface (degradation ong>andong> sorption). Temperature is especially important becausethe evaporation of water ong>andong> the vapour pressure ong>andong> water solubility of thepesticide depend on it.It has been shown that for highly soluble gaseous pesticides a maximum of lessthan 25% of the emission can be deposited withong>inong> 2 km from the source. Formoderately soluble gaseous pesticides it ong>inong> the order of 7% ong>andong> for slightly solublegases about 1% only. The dry deposition relatively close to the source has not beenmeasured although models have shown that ong>inong> extreme cases about 20% of theemission could be dry deposited withong>inong> a few hundred metres from the source.If the concentrations ong>inong> ong>airong> ong>andong> ong>inong> ong>precipitationong> are measured simultaneously, therate at which pesticides are removed from the atmosphere by wet deposition can becalculated, if the mixong>inong>g height is known. For gaseous pesticides the removal rateby dry ong>andong> wet deposition ong>inong>creases with their solubility ong>inong> water. For pesticides ong>inong>particulate form the removal rate by dry ong>andong> wet deposition depends on the sizedistribution of the particles. In general particles are removed rather efficiently bywet removal processes where clouds are ong>inong>volved.Photochemical atmospheric reactions of pesticides are not thought to be importantwithong>inong> 2 km from a source but can play an important role durong>inong>g transport over190


longer distances. They will limit the distance over which gaseous pesticides can betransported if dry or wet deposition is not sufficient.Conversion of pesticides from the gaseous to the particulate phase is thought not tobe important withong>inong> 2 km from the source, but is important durong>inong>g transport overlonger distances. The reason is that dry ong>andong> wet removal rates for gaseous ong>andong>particulate pesticides are different. A phase change will therefore ong>inong>fluence theremoval rate.Models can never replace measurements, but they can provide a best estimate, ong>andong>especially ong>inong> designong>inong>g experiments models are useful, song>inong>ce experiments forpesticides should be planned with great care because chemical analysis areexpensive. Laboratory experiments are useful, because they make it possible tostudy processes under controlled conditions, with only one factor varied at a time.A good research strategy should ong>inong>clude both field ong>andong> laboratory experiments ong>andong>model development.Compared to the ong>airong> pollutants sulphur dioxide ong>andong> nitrogen oxides, modellong>inong>g ofpesticides ong>inong> the atmosphere is far from easy, they occur ong>inong> much lowerconcentrations, ong>andong> there are hundreds of different pesticides around ong>inong> theatmosphere. On the other hong>andong>, it needs less knowledge to come to sound policydecisions than to get a quantitative scientific description of the whole system.ong>Pesticidesong> ong>inong> ong>airong> ong>andong> ong>inong>ong>precipitationong>Raong>inong>water has been collected on three locations on the islong>andong> of Zealong>andong>.The analysis of phenoxyalkanoic acids (mecoprop, dichlorprop ong>andong> MCPA) ong>andong> ofisoproturon showed that there was an obvious connection between the fong>inong>dong>inong>gs ofpesticides ong>inong> raong>inong>water ong>andong> the time of sprayong>inong>g with the herbicides. Concentrationsong>andong> deposition were of the same order of magnitude as found ong>inong> other Europeancountries.Phenoxyalkanoic herbicides were detected durong>inong>g the first years of the experiment,but ong>inong> 1998 no phenoxyalkanoic acid were detected on the three locations. Thisong>inong>dicates that the limitations ong>inong> the use of these herbicides have greatly reducedtheir appearance ong>inong> ong>precipitationong>.The herbicide DNOC was found ong>inong> high concentrations both durong>inong>g ong>andong> outside thesprayong>inong>g season, even though this compound has not been allowed ong>inong> Denmark forthe last 10 years. The concentrations detected ong>inong> the raong>inong> ong>inong> Denmark are of thesame order of magnitude as found ong>inong> Englong>andong>, Germany ong>andong> Switzerlong>andong>. Thefong>inong>dong>inong>g of DNOC ong>inong>dicates that either the compound can be transported ong>inong> theatmosphere over long distances or other sources of DNOC than pesticide use areimportant. Most likely, the fong>inong>dong>inong>g of DNOC is caused by formation of thecompound ong>inong> the atmosphere, probably from reaction between toluene ong>andong> nitrousoxides under ong>inong>fluence of sunlight.The pesticide concentrations found are regarded as mong>inong>imum concentrations song>inong>cethe samples are not adequately stable especially ong>inong> the summer season if the raong>inong>falls ong>inong> the begong>inong>nong>inong>g of the collection period.Effects on plants ong>andong>plant communitiesThe herbicide mecoprop-P was used as model pesticide. The ong>inong>fluence oflow doses of this herbicide was determong>inong>ed on the followong>inong>g testplants, which areall very susceptible to phenoxyalkanoic acid herbicides (Capsella bursa-pastoris,Thlaspi arvense, Song>inong>apis alba, Cirsium arvense, Lapsana communis, Chenopodiumalbum ong>andong> Stellaria media. Further the ong>effectsong> were determong>inong>ed on the competitionbetween Capsella bursa-pastoris ong>andong> Geranium dissectum.191


The No observable Effect Level (NOEL) was defong>inong>ed as the ED 10 value (thecalculated dose that caused a 10% reduction ong>inong> the growth of the test plant). Theresults show that the NOEL of mecoprop on the most susceptible plant species ong>inong>the study was more than 3 times higher than the maximum yearly deposition ofmecoprop ong>inong> Denmark.Plant biomass was found to be as susceptible as seed production to assess thepesticide ong>inong>fluence. Further competitive ability between plant species was notfound to be a more susceptible test method. Effects of the combong>inong>ations ofpesticides that have been found ong>inong> raong>inong>water still needs to be determong>inong>ed.192

More magazines by this user
Similar magazines