keck geology consortium proceedings of the twenty-fourth annual ...
keck geology consortium proceedings of the twenty-fourth annual ...
keck geology consortium proceedings of the twenty-fourth annual ...
Create successful ePaper yourself
Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.
KECK GEOLOGY CONSORTIUM<br />
PROCEEDINGS OF THE TWENTY-FOURTH<br />
ANNUAL KECK RESEARCH SYMPOSIUM IN<br />
GEOLOGY<br />
April 2011<br />
Union College, Schenectady, NY<br />
Dr. Robert J. Varga, Editor<br />
Director, Keck Geology Consortium<br />
Pomona College<br />
Dr. Holli Frey<br />
Symposium Convenor<br />
Union College<br />
Carol Morgan<br />
Keck Geology Consortium Administrative Assistant<br />
Diane Kadyk<br />
Symposium Proceedings Layout & Design<br />
Department <strong>of</strong> Earth & Environment<br />
Franklin & Marshall College<br />
Keck Geology Consortium<br />
Geology Department, Pomona College<br />
185 E. 6 th St., Claremont, CA 91711<br />
(909) 607-0651, <strong>keck</strong><strong>geology</strong>@pomona.edu, <strong>keck</strong><strong>geology</strong>.org<br />
ISSN# 1528-7491<br />
The Consortium Colleges The National Science Foundation ExxonMobil Corporation
KECK GEOLOGY CONSORTIUM<br />
PROCEEDINGS OF THE TWENTY-FOURTH ANNUAL KECK<br />
RESEARCH SYMPOSIUM IN GEOLOGY<br />
ISSN# 1528-7491<br />
Robert J. Varga<br />
Editor and Keck Director<br />
Pomona College<br />
April 2011<br />
Keck Geology Consortium<br />
Pomona College<br />
185 E 6 th St., Claremont, CA<br />
91711<br />
Diane Kadyk<br />
Proceedings Layout & Design<br />
Franklin & Marshall College<br />
Keck Geology Consortium Member Institutions:<br />
Amherst College, Beloit College, Carleton College, Colgate University, The College <strong>of</strong> Wooster,<br />
The Colorado College, Franklin & Marshall College, Macalester College, Mt Holyoke College,<br />
Oberlin College, Pomona College, Smith College, Trinity University, Union College,<br />
Washington & Lee University, Wesleyan University, Whitman College, Williams College<br />
2010-2011 PROJECTS<br />
FORMATION OF BASEMENT-INVOLVED FORELAND ARCHES: INTEGRATED STRUCTURAL AND<br />
SEISMOLOGICAL RESEARCH IN THE BIGHORN MOUNTAINS, WYOMING<br />
Faculty: CHRISTINE SIDDOWAY, MEGAN ANDERSON, Colorado College, ERIC ERSLEV, University <strong>of</strong><br />
Wyoming<br />
Students: MOLLY CHAMBERLIN, Texas A&M University, ELIZABETH DALLEY, Oberlin College, JOHN<br />
SPENCE HORNBUCKLE III, Washington and Lee University, BRYAN MCATEE, Lafayette College, DAVID<br />
OAKLEY, Williams College, DREW C. THAYER, Colorado College, CHAD TREXLER, Whitman College, TRIANA<br />
N. UFRET, University <strong>of</strong> Puerto Rico, BRENNAN YOUNG, Utah State University.<br />
EXPLORING THE PROTEROZOIC BIG SKY OROGENY IN SOUTHWEST MONTANA<br />
Faculty: TEKLA A. HARMS, JOHN T. CHENEY, Amherst College, JOHN BRADY, Smith College<br />
Students: JESSE DAVENPORT, College <strong>of</strong> Wooster, KRISTINA DOYLE, Amherst College, B. PARKER HAYNES,<br />
University <strong>of</strong> North Carolina - Chapel Hill, DANIELLE LERNER, Mount Holyoke College, CALEB O. LUCY,<br />
Williams College, ALIANORA WALKER, Smith College.<br />
INTERDISCIPLINARY STUDIES IN THE CRITICAL ZONE, BOULDER CREEK CATCHMENT,<br />
FRONT RANGE, COLORADO<br />
Faculty: DAVID P. DETHIER, Williams College, WILL OUIMET. University <strong>of</strong> Connecticut<br />
Students: ERIN CAMP, Amherst College, EVAN N. DETHIER, Williams College, HAYLEY CORSON-RIKERT,<br />
Wesleyan University, KEITH M. KANTACK, Williams College, ELLEN M. MALEY, Smith College, JAMES A.<br />
MCCARTHY, Williams College, COREY SHIRCLIFF, Beloit College, KATHLEEN WARRELL, Georgia Tech<br />
University, CIANNA E. WYSHNYSZKY, Amherst College.<br />
SEDIMENT DYNAMICS & ENVIRONMENTS IN THE LOWER CONNECTICUT RIVER<br />
Faculty: SUZANNE O’CONNELL, Wesleyan University<br />
Students: LYNN M. GEIGER, Wellesley College, KARA JACOBACCI, University <strong>of</strong> Massachusetts (Amherst),<br />
GABRIEL ROMERO, Pomona College.<br />
GEOMORPHIC AND PALEOENVIRONMENTAL CHANGE IN GLACIER NATIONAL PARK,<br />
MONTANA, U.S.A.<br />
Faculty: KELLY MACGREGOR, Macalester College, CATHERINE RIIHIMAKI, Drew University, AMY MYRBO,<br />
LacCore Lab, University <strong>of</strong> Minnesota, KRISTINA BRADY, LacCore Lab, University <strong>of</strong> Minnesota
Students: HANNAH BOURNE, Wesleyan University, JONATHAN GRIFFITH, Union College, JACQUELINE<br />
KUTVIRT, Macalester College, EMMA LOCATELLI, Macalester College, SARAH MATTESON, Bryn Mawr<br />
College, PERRY ODDO, Franklin and Marshall College, CLARK BRUNSON SIMCOE, Washington and Lee<br />
University.<br />
GEOLOGIC, GEOMORPHIC, AND ENVIRONMENTAL CHANGE AT THE NORTHERN<br />
TERMINATION OF THE LAKE HÖVSGÖL RIFT, MONGOLIA<br />
Faculty: KARL W. WEGMANN, North Carolina State University, TSALMAN AMGAA, Mongolian University <strong>of</strong><br />
Science and Technology, KURT L. FRANKEL, Georgia Institute <strong>of</strong> Technology, ANDREW P. deWET, Franklin &<br />
Marshall College, AMGALAN BAYASAGALN, Mongolian University <strong>of</strong> Science and Technology.<br />
Students: BRIANA BERKOWITZ, Beloit College, DAENA CHARLES, Union College, MELLISSA CROSS, Colgate<br />
University, JOHN MICHAELS, North Carolina State University, ERDENEBAYAR TSAGAANNARAN, Mongolian<br />
University <strong>of</strong> Science and Technology, BATTOGTOH DAMDINSUREN, Mongolian University <strong>of</strong> Science and<br />
Technology, DANIEL ROTHBERG, Colorado College, ESUGEI GANBOLD, ARANZAL ERDENE, Mongolian<br />
University <strong>of</strong> Science and Technology, AFSHAN SHAIKH, Georgia Institute <strong>of</strong> Technology, KRISTIN TADDEI,<br />
Franklin and Marshall College, GABRIELLE VANCE, Whitman College, ANDREW ZUZA, Cornell University.<br />
LATE PLEISTOCENE EDIFICE FAILURE AND SECTOR COLLAPSE OF VOLCÁN BARÚ, PANAMA<br />
Faculty: THOMAS GARDNER, Trinity University, KRISTIN MORELL, Penn State University<br />
Students: SHANNON BRADY, Union College. LOGAN SCHUMACHER, Pomona College, HANNAH ZELLNER,<br />
Trinity University.<br />
KECK SIERRA: MAGMA-WALLROCK INTERACTIONS IN THE SEQUOIA REGION<br />
Faculty: JADE STAR LACKEY, Pomona College, STACI L. LOEWY, California State University-Bakersfield<br />
Students: MARY BADAME, Oberlin College, MEGAN D’ERRICO, Trinity University, STANLEY HENSLEY,<br />
California State University, Bakersfield, JULIA HOLLAND, Trinity University, JESSLYN STARNES, Denison<br />
University, JULIANNE M. WALLAN, Colgate University.<br />
EOCENE TECTONIC EVOLUTION OF THE TETONS-ABSAROKA RANGES, WYOMING<br />
Faculty: JOHN CRADDOCK, Macalester College, DAVE MALONE, Illinois State University<br />
Students: JESSE GEARY, Macalester College, KATHERINE KRAVITZ, Smith College, RAY MCGAUGHEY,<br />
Carleton College.<br />
Funding Provided by:<br />
Keck Geology Consortium Member Institutions<br />
The National Science Foundation Grant NSF-REU 1005122<br />
ExxonMobil Corporation
Keck Geology Consortium: Projects 2010-2011<br />
Short Contributions— Front Range, CO<br />
INTERDISCIPLINARY STUDIES IN THE CRITICAL ZONE, BOULDER CREEK CATCHMENT,<br />
FRONT RANGE, COLORADO<br />
Project Faculty: DAVID P. DETHIER: Williams College, WILL OUIMET: University <strong>of</strong> Connecticut<br />
CORING A 12KYR SPHAGNUM PEAT BOG: A SEARCH FOR MERCURY AND ITS IMPLICATIONS<br />
ERIN CAMP, Amherst College<br />
Research Advisor: Anna Martini<br />
EXAMINING KNICKPOINTS IN THE BOULDER CREEK CATCHMENT, COLORADO<br />
EVAN N. DETHIER, Williams College<br />
Research Advisor: David P. Dethier<br />
THE DISTRIBUTION OF PHOSPHORUS IN ALPINE AND UPLAND SOILS OF THE BOULDER<br />
CREEK, COLORADO CATCHMENT<br />
HAYLEY CORSON-RIKERT, Wesleyan University<br />
Research Advisor: Timothy Ku<br />
RECONSTRUCTING THE PINEDALE GLACIATION, GREEN LAKES VALLEY, COLORADO<br />
KEITH M. KANTACK, Williams College<br />
Research Advisor: David P. Dethier<br />
CHARACTERIZATION OF TRACE METAL CONCENTRATIONS AND MINING LEGACY IN SOILS,<br />
BOULDER COUNTY, COLORADO<br />
ELLEN M. MALEY, Smith College<br />
Research Advisor: Amy L. Rhodes<br />
ASSESSING EOLIAN CONTRIBUTIONS TO SOILS IN THE BOULDER CREEK CATCHMENT,<br />
COLORADO<br />
JAMES A. MCCARTHY, Williams College<br />
Research Advisor: David P. Dethier<br />
USING POLLEN TO UNDERSTAND QUATERNARY PALEOENVIRONMENTS IN BETASSO GULCH,<br />
COLORADO<br />
COREY SHIRCLIFF, Beloit College<br />
Research Advisor: Carl Mendelson<br />
STREAM TERRACES IN THE CRITICAL ZONE – LOWER GORDON GULCH, COLORADO<br />
KATHLEEN WARRELL, Georgia Tech<br />
Research Advisor: Kurt Frankel<br />
METEORIC 10 BE IN GORDON GULCH SOILS: IMPLICATIONS FOR HILLSLOPE PROCESSES AND<br />
DEVELOPMENT<br />
CIANNA E. WYSHNYSZKY, Amherst College<br />
Research Advisor: Will Ouimet and Peter Crowley<br />
Keck Geology Consortium<br />
Pomona College<br />
185 E. 6 th St., Claremont, CA 91711<br />
Keck<strong>geology</strong>.org
24th Annual Keck Symposium: 2011 Union College, Schenectady, NY<br />
INTERDISCIPLINARY STUDIES IN THE CRITICAL ZONE,<br />
BOULDER CREEK CATCHMENT, FRONT RANGE, COLORADO<br />
DAVID P. DETHIER, Williams College<br />
WILL OUIMET, University <strong>of</strong> Connecticut<br />
INTRODUCTION<br />
Processes in <strong>the</strong> critical zone, <strong>the</strong> life-sustaining surficial<br />
mantle <strong>of</strong> <strong>the</strong> earth, involve wea<strong>the</strong>red geologic<br />
materials, water, and <strong>the</strong> biosphere, mediated by<br />
atmospheric processes that are controlled by changing<br />
climate. Field and laboratory studies that investigate<br />
geologic, hydrologic and geochemical components <strong>of</strong><br />
<strong>the</strong> critical zone provide valuable data about processes<br />
and <strong>the</strong> physical basis for <strong>the</strong>ir integration into<br />
models <strong>of</strong> short and long-term geomorphic, hydrologic<br />
and biochemical response. The Keck Colorado<br />
Project is working in cooperation with a large<br />
interdisciplinary study <strong>of</strong> <strong>the</strong> critical zone (Boulder<br />
Creek Critical Zone Observatory: Wea<strong>the</strong>red pr<strong>of</strong>ile<br />
development in a rocky environment and its influence<br />
on watershed hydrology and biogeochemistry— Suzanne<br />
Anderson, PI, Institute for Arctic and Alpine<br />
Studies, University <strong>of</strong> Colorado). The observatory<br />
(CZO) consists <strong>of</strong> 3 small, instrumented catchments<br />
in <strong>the</strong> Boulder Creek basin, Colorado Front Range:<br />
(1) Green Lakes Valley (GLV; el. 3400 m)--a steep,<br />
glacially scoured alpine area in <strong>the</strong> City <strong>of</strong> Boulder<br />
watershed; (2) Gordon Gulch (el. 2600 m)--a forested,<br />
mid-elevation catchment that exposes isolated bedrock<br />
remnants (tors) developed on a surface <strong>of</strong> low<br />
relief; and (3) Betasso gulch (el. 1950 m)--a steep,<br />
thinly forested basin that preserves thick regolith in<br />
<strong>the</strong> upper catchment and exposes extensive bedrock<br />
outcrops at lower elevations (Fig. 1).<br />
The glaciated GLV, low relief surface, and bedrock<br />
canyons are developed in granitic or gneissic rocks<br />
and are influenced by <strong>the</strong> strong gradient in elevation,<br />
climate and vegetation from west to east. Variation<br />
in critical-zone development in <strong>the</strong>se different environments<br />
allows us to test models <strong>of</strong> wea<strong>the</strong>ring and<br />
regolith generation, elemental cycling, slope evolution<br />
and sediment transport in an accessible field<br />
setting. Land-use, vegetation and hydrologic response<br />
93<br />
in each CZO catchment also reflect changes produced<br />
by anthropogenic activities such as mining and timber<br />
harvest over <strong>the</strong> past 150 years. Keck Colorado field<br />
studies focus on using a variety <strong>of</strong> techniques to map<br />
and characterize <strong>the</strong> geologic history, near-surface<br />
geologic materials and geochemical properties for<br />
each <strong>of</strong> <strong>the</strong> study catchments.<br />
Figure 1. Perspective view looking west across <strong>the</strong> Front<br />
Range from Boulder, Colorado, showing Middle Boulder<br />
Creek and location <strong>of</strong> Betasso, Gordon Gulch and Green<br />
Lakes Valley catchments, Boulder Creek Critical Zone<br />
Observatory. White filled area shows approximate extent<br />
<strong>of</strong> latest Pleistocene glaciers (after Madole et al., 1999).<br />
SETTING<br />
The middle Boulder Creek catchment (Fig. 1) extends<br />
from <strong>the</strong> glaciated alpine zone <strong>of</strong> <strong>the</strong> Continental<br />
Divide east to <strong>the</strong> semi-arid western edge <strong>of</strong> <strong>the</strong> Great<br />
Plains. The high-relief zone <strong>of</strong> cirques and deep, Ushaped<br />
valleys in <strong>the</strong> glaciated area become shallower<br />
eastward through a zone <strong>of</strong> low relief and relatively<br />
low slopes. To <strong>the</strong> east, valleys deepen into steep,<br />
narrow bedrock canyons as <strong>the</strong>y pass knickzones, and<br />
flatten to lower channel slopes near <strong>the</strong> piedmont margin.<br />
Small glaciers and late-persisting snowfields dot<br />
<strong>the</strong> alpine zone, which exposes bedrock and relatively<br />
thin deposits related to <strong>the</strong> latest Pleistocene Pinedale<br />
glaciation and to Holocene erosion. The forested zone
24th Annual Keck Symposium: 2011 Union College, Schenectady, NY<br />
<strong>of</strong> low relief exposes local areas <strong>of</strong> thick (characteristically<br />
3 to 8 m) regolith, saprolite and oxidized bedrock,<br />
but <strong>the</strong> wea<strong>the</strong>red mantle is thin in o<strong>the</strong>r areas.<br />
Low terraces and alluvial fans as thick as 4 m line<br />
channels locally. In <strong>the</strong> vicinity <strong>of</strong> knickzones and<br />
in downstream areas such as Betasso gulch, slopes<br />
near channels are steep and fresh bedrock is exposed,<br />
whereas areas more distant from channels retain a<br />
thicker wea<strong>the</strong>red mantle.<br />
APPROACH<br />
In our third project year, we used field mapping and<br />
sampling in all three CZO catchments, supplemented<br />
by geophysical measurements, in order to provide<br />
basic data about soils and <strong>the</strong>ir geochemistry, shallow<br />
subsurface <strong>geology</strong> and erosional history <strong>of</strong><br />
<strong>the</strong> critical zone. Students supported by <strong>the</strong> Keck<br />
Geology <strong>consortium</strong> and by NSF learned geophysical<br />
techniques and initial data reduction, processing<br />
and visualization methods in <strong>the</strong>se settings. Students<br />
chose from a variety <strong>of</strong> potential projects in <strong>the</strong> study<br />
catchments; 2010 project topical areas included:<br />
1. Trace-metal studies <strong>of</strong> soils and bog sediment,<br />
Boulder Creek catchment<br />
2. Soil chemistry (Fe and P) in CZO catchments<br />
and meteoric 10Be studies <strong>of</strong> soils, focused on<br />
Gordon Gulch<br />
3. Stratigraphy and palynology <strong>of</strong> valley-fill<br />
sediment, Gordon and Betasso catchments<br />
4. Geomorphic research: Ice erosion and geo<br />
morphic evolution <strong>of</strong> <strong>the</strong> Green Lakes Valley<br />
and <strong>the</strong> evolution <strong>of</strong> knickpoints along channels<br />
in <strong>the</strong> Boulder Creek catchment<br />
We ran resistivity lines in Gordon Gulch, resistivity<br />
and ground-penetrating radar on Niwot Ridge and<br />
ground-penetrating radar in <strong>the</strong> vicinity <strong>of</strong> <strong>the</strong> bogs<br />
near <strong>the</strong> N. Branch Boulder Creek. In Gordon Gulch,<br />
we worked on regolith studies in cooperation with<br />
investigators from <strong>the</strong> University <strong>of</strong> Colorado and <strong>the</strong><br />
US Geological Survey.<br />
STUDENT PROJECTS<br />
Six Keck students joined Williams students Evan<br />
Dethier, Keith Kantack and James McCarthy and<br />
94<br />
Erin Camp (Amherst), who were supported directly<br />
by NSF funding. David Dethier and Will Ouimet<br />
supervised students on a daily basis and field teams<br />
frequently joined investigators and graduate students<br />
from <strong>the</strong> University <strong>of</strong> Colorado. Matthias Leopold<br />
(Technical University <strong>of</strong> Munich) worked with Keck<br />
students for two weeks in <strong>the</strong> field. Keck students<br />
were among <strong>the</strong> only undergraduates to present preliminary<br />
results <strong>of</strong> <strong>the</strong>ir field research at <strong>the</strong> 3rd <strong>annual</strong><br />
Boulder Creek CZO meeting on 10 August 2010.<br />
Keck Colorado students worked in pairs on a daily<br />
basis and sometimes as geophysical support teams.<br />
Geophysical data provided important background data<br />
for extensive studies in Gordon Gulch and for coring<br />
<strong>of</strong> bogs (Fig. 2).<br />
Figure 2. Coring a bog in a late Pinedale moraine complex<br />
along <strong>the</strong> N. Fork Boulder Creek, supervised by<br />
Robert Nelson, Colby College.<br />
Short papers elsewhere in this volume report results<br />
<strong>of</strong> <strong>the</strong> field and laboratory studies in some detail.<br />
We summarize and provide brief comments on this<br />
research here.<br />
Studies <strong>of</strong> trace metals in organic sediment, soil<br />
and regolith in Gordon Gulch and adjacent areas<br />
Erin Camp (Amherst) reports “Coring a 12 kyr sphagnum<br />
bog in <strong>the</strong> N. Boulder Creek valley—a search<br />
for mercury and its implications” and Ellie Maley<br />
(Smith) worked on “Characterization <strong>of</strong> trace metal<br />
concentrations and mining legacy in soils, Boulder<br />
County, Colorado”. Both <strong>of</strong> <strong>the</strong>se studies demonstrate<br />
that Hg is enriched in recent organic-rich
24th Annual Keck Symposium: 2011 Union College, Schenectady, NY<br />
sediment <strong>of</strong> <strong>the</strong> montane zone when compared to<br />
older bog sediment or to soil C-horizons. The bog Hg<br />
pr<strong>of</strong>ile and 14 C ages show a strong correspondence<br />
between elevated Hg levels and large, silicic volcanic<br />
eruptions in Holocene time. Peak Hg levels produced<br />
by eruptions decline over narrow intervals. Mercury<br />
concentrations in bog sediments increase to a broad<br />
peak coincident with exploitation <strong>of</strong> precious metals<br />
in <strong>the</strong> western hemisphere beginning in <strong>the</strong> 16th century<br />
and peaking in <strong>the</strong> late 19th century in Colorado.<br />
Ellie’s work shows that Hg, As and possibly Pb are<br />
enriched in organic-rich O and A-horizons compared<br />
to soil parent material; o<strong>the</strong>r minor elements such as<br />
Cu and Zn do not show enrichment. Ellie’s research<br />
suggests that metal enrichment likely represents <strong>the</strong><br />
legacy <strong>of</strong> local metal mining, milling and smelting,<br />
and <strong>the</strong> affinity <strong>of</strong> organic matter for <strong>the</strong>se relatively<br />
volatile trace metals.<br />
Studies <strong>of</strong> soil and regolith geochemistry and age<br />
in Gordon Gulch and adjacent areas<br />
Four students, including Ellie Maley, studied soils<br />
and regolith exposed in pits that were dug in Gordon<br />
Gulch and at o<strong>the</strong>r nearby exposures. The “pit<br />
crew”, aided by <strong>the</strong>ir advisors and o<strong>the</strong>r CZO investigators,<br />
collected from many <strong>of</strong> <strong>the</strong> same sites and<br />
worked on separate, but complimentary topics. Cianna<br />
Wyshnytzky (Amherst) documented meteoric<br />
10 Be accumulation in soils from Gordon Gulch and<br />
at Silver Lake—“Erosion, particle paths and deposition—meteoric<br />
10 Be in Gordon Gulch”. James<br />
McCarthy (Williams) studied <strong>the</strong> texture and Fed<br />
(dithionite-extractable iron) accumulation in soils<br />
from Gordon Gulch and adjacent subalpine and alpine<br />
areas: “Assessing eolian contributions to soils in <strong>the</strong><br />
Boulder Creek catchment”, whereas Hayley Corson-<br />
Rikert (Wesleyan) studied “Extractable P in soils <strong>of</strong><br />
<strong>the</strong> Boulder Creek catchment, Colorado”.<br />
Data from <strong>the</strong>se studies demonstrate that <strong>the</strong>re are<br />
fundamental differences between soils developed on<br />
stable sites and those that formed on regolith-covered<br />
hillslopes and that dustfall and chemical wea<strong>the</strong>ring<br />
influence <strong>the</strong> partitioning <strong>of</strong> Fed and P in soils. Both<br />
erosion and wea<strong>the</strong>ring rates are related to aspect in<br />
Gordon Gulch; regolith is thicker, more wea<strong>the</strong>red<br />
and contains more meteoric 10 Be on <strong>the</strong> north-facing<br />
95<br />
Figure 3. Soil pits in Gordon Gulch. A. James McCarthy,<br />
Cianna Wyshnytzky, Hayley Corson-Rikert and Ellie<br />
Maley in a 1.8 m-deep pit in regolith, north-facing slope.<br />
B. Soil-sampling tools and thin regolith over saprolite,<br />
south-facing slope. Cianna Wyshnytzky sampling for meteoric<br />
10Be analysis.<br />
slope. Fe and P mobility also are influenced by <strong>the</strong><br />
moisture and temperature gradient between soils exposed<br />
in Gordon Gulch and soils <strong>of</strong> similar age in <strong>the</strong><br />
alpine and subalpine portions <strong>of</strong> <strong>the</strong> Boulder Creek<br />
CZO. Extractable Fe, clay and 10 Be reach peak values<br />
in <strong>the</strong> B-horizon <strong>of</strong> <strong>the</strong> till-derived soil at Silver Lake<br />
(Fig. 4), whereas P is depleted in <strong>the</strong> soil compared to<br />
unwea<strong>the</strong>red till; soils in Gordon Gulch and Betasso<br />
do not display comparable patterns.<br />
The inventory <strong>of</strong> meteoric 10 Be in Gordon Gulch is<br />
consistent with model predictions for deposition rates<br />
(Graly et al., 2011), but <strong>the</strong> inventory at Silver Lake is<br />
too high, suggesting that at this site dustfall rates or
24th Annual Keck Symposium: 2011 Union College, Schenectady, NY<br />
Figure 4. Plot showing relationship <strong>of</strong> Fed, clay and<br />
meteoric 10 Be to depth below <strong>the</strong> surface and soil horizons<br />
at Silver Lake, Green Lakes Valley. P (not plotted) is depleted<br />
and relatively labile in <strong>the</strong> upper soil and relatively<br />
enriched and associated with inorganic material in <strong>the</strong><br />
unoxidized till.<br />
late Pleistocene precipitation may have been substantially<br />
higher than at present. Cianna’s meteoric 10 Be<br />
research represents <strong>the</strong> first application <strong>of</strong> this technique<br />
in <strong>the</strong> Boulder Creek CZO catchments.<br />
Stratigraphy and palynology <strong>of</strong> valley-fill sediment,<br />
Gordon and Betasso catchments<br />
In Gordon Gulch and in upper Betasso Gulch, colluvium<br />
and deposits beneath terraces as much as 4<br />
m above <strong>the</strong> channel comprise local valley fills <strong>of</strong><br />
Holocene and latest Pleistocene (?) age. Kathleen<br />
Warrell (Georgia Tech) studied terrace morphology<br />
and sampled deposits exposed beneath low terraces<br />
in Gordon Gulch-- “Stream terraces in <strong>the</strong> critical<br />
zone-- lower Gordon Gulch, Colorado”. Her work<br />
96<br />
shows that at least 75,000 m 3 <strong>of</strong> sediment is stored in<br />
near <strong>the</strong> channel in lower Gordon Gulch. Sediment<br />
beneath <strong>the</strong> 1-m terrace is less than 1500 years old<br />
(Fig. 5);<br />
Figure 5. Graphic log <strong>of</strong> sediment (description from C.<br />
Shircliff) exposed beneath K. Warrell’s 1-m terrace. Basal<br />
sediment is approximately 1500 years old.<br />
higher terraces are cut into middle and early Holocene<br />
alluvial and colluvial deposits. Corey Shircliff<br />
(Beloit) studied organic material in Holocene terrace<br />
deposits in Gordon Gulch and Betasso--“Using pollen<br />
to understand Quaternary paleoenvironments in<br />
Betasso Gulch, Colorado”. She was able to separate<br />
pollen from a buried soil developed on colluvium and<br />
to identify many <strong>of</strong> <strong>the</strong> pollen grains. Corey’s work<br />
indicates that early Holocene pollen at Betasso is<br />
richer in Picea (spruce) than a modern pollen sample<br />
and records a climate that was likely wetter and perhaps<br />
slightly cooler than that at present.<br />
Ice erosion and geomorphic evolution <strong>of</strong> <strong>the</strong> Green<br />
Lakes Valley and <strong>the</strong> evolution <strong>of</strong> knickpoints<br />
along channels in <strong>the</strong> Boulder Creek catchment<br />
Keith Kantack (Williams) used field measurements
24th Annual Keck Symposium: 2011 Union College, Schenectady, NY<br />
in GLV and interpretation <strong>of</strong> DEMs derived from<br />
Lidar flown in August 2011 to map <strong>the</strong> extent <strong>of</strong> late<br />
Pinedale ice and glacial moraines in <strong>the</strong> upper North<br />
Boulder Creek catchment (Fig. 6).<br />
Figure 6. Keith Kantack, Evan Dethier and James Mc-<br />
Carthy stand on glacially sculpted and smoo<strong>the</strong>d bedrock<br />
knob, upper Green Lakes Valley.<br />
His work “Reconstructing<strong>the</strong> Pinedale glaciation in<br />
<strong>the</strong> Green Lakes valley, Colorado” shows that latest<br />
Pleistocene ice in <strong>the</strong> GLV was thin (mainly less<br />
than 150 m), but extended out <strong>of</strong> <strong>the</strong> cirques and<br />
more than 10 km to <strong>the</strong> east to elevations as low as<br />
2650 m. Measured moraine volumes suggest that <strong>the</strong><br />
average erosion rate by North Boulder cirque glaciers<br />
was about 1mm yr -1 during <strong>the</strong> maximum late Pinedale<br />
glaciation, which lasted from about 21 to 15 ka.<br />
Those rates are similar to estimates used by Ward et<br />
al. (2009) for modeling ice flow in Front Range catchments.<br />
Evan Dethier studied <strong>the</strong> evolution <strong>of</strong> channels and<br />
slopes at different scales near knickpoints in Betasso,<br />
Gordon Gulch, and along Middle Boulder Creek--<br />
“Knickpoints—a study <strong>of</strong> channels in <strong>the</strong> Boulder<br />
Creek catchment”. His field studies, combined with<br />
RiverTools interpretation <strong>of</strong> DEMs derived from August,<br />
2010 Lidar, suggest that channels and adjacent<br />
hillslopes reflect <strong>the</strong> slow migration <strong>of</strong> knickpoints<br />
in <strong>the</strong> Boulder Creek catchment, moderated by local<br />
rock strength. Betasso gulch, <strong>the</strong> smallest <strong>of</strong> <strong>the</strong><br />
Boulder Creek CZO catchments, has a channel that is<br />
97<br />
steep and rough throughout and is flanked by steep,<br />
smooth slopes that expose bedrock and thin regolith<br />
near Boulder Creek. In <strong>the</strong> upper part <strong>of</strong> <strong>the</strong> catchment,<br />
however, <strong>the</strong> channel only locally exposes<br />
bedrock, and is cut mainly in s<strong>of</strong>t, deeply wea<strong>the</strong>red<br />
saprolite and through thick colluvium that was deposited<br />
in latest Pleistocene time. At Gordon Gulch, rock<br />
strength appears to control <strong>the</strong> location <strong>of</strong> <strong>the</strong> knickzone<br />
that separates <strong>the</strong> upper and lower basin; morphology<br />
<strong>of</strong> adjacent hillslopes reflect local channel<br />
slope. At <strong>the</strong> scale <strong>of</strong> Boulder Creek, hillslope evolution<br />
appears to “lag” knickpoint migration because<br />
local rocks are strong and hillslope erosion requires<br />
removal <strong>of</strong> large volumes <strong>of</strong> rock.<br />
Figure 7. Sculpted bedrock and boulder-rich channel<br />
in <strong>the</strong> knickzone along N. Boulder Creek above Boulder<br />
Falls.<br />
CONCLUSIONS<br />
“Piggybacking” <strong>the</strong> Keck Colorado Geology Project<br />
on <strong>the</strong> NSF-Boulder Creek Critical Zone Observatory<br />
has allowed Keck undergraduates to integrate <strong>the</strong>ir<br />
projects with <strong>the</strong> research <strong>of</strong> graduate and postdoctoral<br />
students from <strong>the</strong> University <strong>of</strong> Colorado and<br />
o<strong>the</strong>r research universities. Keck student research<br />
has benefitted from <strong>the</strong> personnel, monitoring efforts,<br />
and general level <strong>of</strong> scientific interest associated with<br />
<strong>the</strong> NSF project. The Boulder Creek CZO has gained<br />
from <strong>the</strong> focused field and laboratory research <strong>of</strong> <strong>the</strong><br />
Keck students, <strong>the</strong>ir energy, and <strong>the</strong>ir collective demonstration<br />
<strong>of</strong> what can be accomplished by <strong>the</strong> best
24th Annual Keck Symposium: 2011 Union College, Schenectady, NY<br />
undergraduates.<br />
ACKNOWLEDGMENTS<br />
Field studies and measurements in <strong>the</strong> Boulder Creek<br />
area were performed in cooperation with <strong>the</strong> Boulder<br />
Creek CZO Project (National Science Foundation),<br />
<strong>the</strong> USDA Forest Service, and <strong>the</strong> City <strong>of</strong> Boulder<br />
Watershed and Parks and Recreation Departments.<br />
Nel Caine (University <strong>of</strong> Colorado) and Craig Skeie<br />
(City Watershed Manager) guided work in <strong>the</strong> Green<br />
Lakes basin, Pete Birkeland (University <strong>of</strong> Colorado)<br />
taught us about soils. Greg Tucker, Cam Wobus and<br />
Abby Langston (all affiliated with <strong>the</strong> University <strong>of</strong><br />
Colorado) shared <strong>the</strong>ir knowledge <strong>of</strong> <strong>the</strong> Critical Zone<br />
and how to study it. Suzanne Anderson and Bob Anderson<br />
joined us for many field “teaching moments”.<br />
We gratefully acknowledge <strong>the</strong> field and laboratory<br />
skills <strong>of</strong> Bob Nelson (Colby College), and <strong>the</strong> ongoing<br />
cooperation, digging ability and cogent advice <strong>of</strong><br />
Joerg Voelkel and Matthias Leopold (Technical University<br />
<strong>of</strong> Munich). The hospitality <strong>of</strong> <strong>the</strong> Mountain<br />
Research Station made this project possible.<br />
REFERENCES<br />
Graly, J. A., Reusser, L. J., and Bierman, P. R., 2011,<br />
Short and long-term delivery rates <strong>of</strong> meteoric<br />
10Be to terrestrial soils: Earth and Planetary<br />
Science Letters 302, p. 329-336.<br />
Madole, R.F., VanSistine, D.P., and Michael, J. A.,<br />
1999, Pleistocene glaciation in <strong>the</strong> upper Platte<br />
River drainage basin, Colorado. U.S. Geol. Surv.<br />
Geol. Invest. Series I-2644.<br />
Ward, D. J., R. S. Anderson, Z. S. Guido, and J. P.<br />
Briner (2009), Numerical modeling <strong>of</strong> cosmogenic<br />
deglaciation records, Front Range and San<br />
Juan mountains, Colorado, J. Geophys. Res.,<br />
114, F01026, doi:10.1029/2008JF001057.<br />
98
24th Annual Keck Symposium: 2011 Union College, Schenectady, NY<br />
CORING A 12KYR OMBROTROPHIC SPHAGNUM PEAT BOG:<br />
A HISTORY OF ATMOSPHERIC MERCURY<br />
ERIN CAMP, Amherst College<br />
Research Advisor: Anna Martini<br />
INTRODUCTION<br />
Elemental mercury (Hg 0 ) is primarily transported<br />
through <strong>the</strong> atmosphere, where it has an average<br />
residence time <strong>of</strong> one year and can be deposited<br />
worldwide (Bindler, 2003). Hg 0 is deposited both<br />
naturally and anthropogenically in <strong>the</strong> environment,<br />
where it can chemically transform into a highly toxic<br />
methylated form <strong>of</strong> mercury (Vandal et al., 1993).<br />
Mercury is introduced naturally into <strong>the</strong> environment<br />
through volcanism, geo<strong>the</strong>rmal activity, and emission<br />
from <strong>the</strong> biosphere and water bodies, and anthropogenically<br />
through coal combustion, waste incineration,<br />
and metal ore processing (Bindler, 2003). Additionally,<br />
mercury retention is known to increase in<br />
colder temperatures, thus can be used as a paleotemperature<br />
proxy (Martínez-Cortizas et al., 1999).<br />
Ombrotrophic peat bogs topped by Sphagnum moss<br />
are excellent archives <strong>of</strong> elemental Hg deposition<br />
because <strong>the</strong>y receive all <strong>the</strong>ir nutrients from <strong>the</strong><br />
atmosphere and allow little vertical mixing (Madsen,<br />
1981; Lodenius et al., 1983). The Colorado Front<br />
Range has a rich history <strong>of</strong> gold and silver mining,<br />
smelting and mercury amalgamation, thus it is an<br />
ideal location for mercury studies (Nriagu, 1994).<br />
This project has measured <strong>the</strong> amount <strong>of</strong> Hg deposition<br />
in North Boulder Creek Bog, CO in order to 1)<br />
identify <strong>the</strong> natural background Hg deposition for<br />
this location, 2) correlate concentrations with natural<br />
and anthropogenic historical events, and 3) calculate<br />
<strong>the</strong> amount <strong>of</strong> anthropogenically deposited mercury<br />
in this location.<br />
PROJECT LOCATION<br />
North Boulder Creek Bog (40.007349º,-105.560421º;<br />
Fig. 1) is a 3-meter deep subalpine Sphagnum mosscoated<br />
ombrotrophic bog that began to accumulate<br />
organic sediment at 12,000 cal. 14 C years BP (yBP;<br />
99<br />
Figure 1: Google Earth image <strong>of</strong> project location,<br />
illustrating <strong>the</strong> perimeter <strong>of</strong> <strong>the</strong> bog and <strong>the</strong> surrounding<br />
glacial moraine.<br />
present refers to 1950) (Leopold and Dethier, 2007;<br />
Leopold, 2010 personal communication). The bog<br />
is located in <strong>the</strong> Front Range <strong>of</strong> Colorado, situated<br />
within a kettle hole associated with <strong>the</strong> Pinedale<br />
glaciation, and flanked by a glacial moraine on its<br />
eastern side (Richmond, 1960; Fig. 1). North Boulder<br />
Creek Bog is believed to have formed after <strong>the</strong><br />
rapid drainage <strong>of</strong> Lake Devlin about 13,000 years<br />
ago (Leopold pers. comm. 2010; Madole, 1985).<br />
METHODS<br />
The coring, performed with a modified Livingstone<br />
piston corer (Livingstone, 1955), produced a complete<br />
1.8m core that had been compacted by an average<br />
<strong>of</strong> 50%—slightly less near <strong>the</strong> top and more<br />
near <strong>the</strong> bottom—representing a 3.65m core. Sampling<br />
<strong>of</strong> <strong>the</strong> core was performed at 2.5cm intervals,<br />
representing ‘expanded’ intervals <strong>of</strong> 5cm. All depths<br />
referred to in this paper are ‘expanded’ depths. All<br />
samples were dried overnight—maximum <strong>of</strong> 12<br />
hours—at 105ºC. All seventy-four samples were
24th Annual Keck Symposium: 2011 Union College, Schenectady, NY<br />
ground with mortar and pestle, and run individually<br />
in a direct combustion cold vapor AA instrument<br />
(Hydra-C, Leeman Labs) for total mercury concentration.<br />
The machine was calibrated using a marine<br />
sediment standard with a precision <strong>of</strong> ±6.03% and an<br />
accuracy (recovery) <strong>of</strong> 103.6%. Following <strong>the</strong>se runs,<br />
higher resolution sampling was conducted at 2cm<br />
‘expanded’ intervals near <strong>the</strong> uppermost section <strong>of</strong> <strong>the</strong><br />
core, in order to obtain more precise data during<br />
modern times. Six samples were taken near 35cm<br />
depth, and five additional samples were taken from<br />
<strong>the</strong> top 5cm <strong>of</strong> <strong>the</strong> core. Additionally, five samples<br />
were sent to <strong>the</strong> Woods Hole NOSAMS Facility for<br />
AMS radiocarbon analysis. These samples were extracted<br />
from various locations along <strong>the</strong> core at 100,<br />
160, 265, 315, and 365cm depths. Radiocarbon age is<br />
calculated from <strong>the</strong> δ13C-corrected Fraction Modern<br />
(Fm) according to <strong>the</strong> following formula:<br />
Age = -8033ln (Fm).<br />
Figure 2: Photos <strong>of</strong> top two sections <strong>of</strong> <strong>the</strong> core. Top<br />
photo represents <strong>the</strong> first meter <strong>of</strong> core, compacted to<br />
48cm. Bottom picture represents <strong>the</strong> second meter <strong>of</strong> core,<br />
compacted to 46cm.<br />
RESULTS<br />
The radiocarbon data yield a typical, slightly curved<br />
age vs. depth trend. With a linear regression analysis,<br />
<strong>the</strong> bog has a deposition rate <strong>of</strong> 0.36 mm yr -1 . A poly-<br />
100<br />
Figure 3: Age-depth graphs for C-14 data <strong>of</strong> <strong>the</strong> five bog<br />
samples, calibrated at Woods Hole HOSAMS facility.<br />
Top graph is a polynomial regression; bottom graph is<br />
a linear regression. Reporting <strong>of</strong> ages and/or activities<br />
follows <strong>the</strong> convention outlined by Stuiver and Polach<br />
(1977) and Stuiver (1980). Ages are calculated using<br />
5568 years as <strong>the</strong> half-life <strong>of</strong> radiocarbon and are reported<br />
without reservoir corrections or calibration to calendar<br />
years. Boxed inset is illustrated in Figure 5.<br />
nomial regression was used in order to fit <strong>the</strong> data at<br />
<strong>the</strong> shallowest and deepest parts <strong>of</strong> <strong>the</strong> core, where<br />
<strong>the</strong> best fit line tapers <strong>of</strong>f in a convex fashion (Fig.<br />
3). The polynomial regression suggests a calibrated<br />
age <strong>of</strong> 9105 yBP at <strong>the</strong> deepest part <strong>of</strong> <strong>the</strong> core.<br />
The data from our mercury analysis, shown in Figure<br />
4, range from 5.2 ppb to 201.2 ppb (ng/g). The highest<br />
values are recorded at shallow depths in <strong>the</strong> core,<br />
at recent times with greater anthropogenic influence,<br />
while <strong>the</strong> lowest values are recorded deep within<br />
<strong>the</strong> core, during historic times. The natural mercury<br />
background concentration for <strong>the</strong> North Boulder<br />
Creek Bog was calculated at approximately 23.2 ppb.<br />
Discernable peaks in mercury concentration occur at<br />
0.5cm (158.9 ppb), 20cm (201.2 ppb), 42cm (125.8<br />
ppb), 70cm (61.9 ppb), 80cm (49.4 ppb), 105cm<br />
(58.6 ppb), 135cm (54.6 ppb), and 230cm (72.5 ppb).
24th Annual Keck Symposium: 2011 Union College, Schenectady, NY<br />
DISCUSSION<br />
CARBON DATING<br />
The oldest date calculated using 14 C was 8,800 BP at<br />
365cm. Radiocarbon dating <strong>of</strong> this bog from previous<br />
studies demonstrated a maximum age <strong>of</strong> approximate<br />
ly 12,000 yBP (Leopold and Dethier, 2007), which<br />
suggests that our core location may not have been at<br />
<strong>the</strong> deepest or oldest part <strong>of</strong> <strong>the</strong> bog, or <strong>the</strong> core may<br />
not have reached <strong>the</strong> bottom <strong>of</strong> <strong>the</strong> bog.<br />
Accumulation rates in peat bogs from o<strong>the</strong>r studies<br />
range anywhere between 0.015-0.920 mm yr -1<br />
(Bindler, 2003; Biester et al., 2002). The average peat<br />
accumulation rate in North Boulder Creek Bog was<br />
calculated at 0.359 mm yr -1 , indicating a slow to<br />
intermediate deposition rate. This value is likely slow<br />
due to <strong>the</strong> subalpine location <strong>of</strong> <strong>the</strong> bog, where <strong>the</strong>re<br />
are minimal organic inputs.<br />
MERCURY ANALYSIS<br />
Each sample represent approximately 28 years, yet<br />
<strong>the</strong>re is a large gap <strong>of</strong> approximately 140 years be<br />
tween each sample due to <strong>the</strong> 5cm interval, due to <strong>the</strong><br />
difficulty <strong>of</strong> high-resolution sampling. Additionally,<br />
<strong>the</strong>re is typically much more compaction at depth<br />
within bogs. Thus, <strong>the</strong>re may be events within <strong>the</strong>se<br />
gaps <strong>of</strong> time that are not recorded by our mercury<br />
analysis. At <strong>the</strong> top <strong>of</strong> <strong>the</strong> core, however, sampling<br />
was conducted at tighter intervals and <strong>the</strong>re is likely<br />
to be much less compaction.<br />
The mercury concentration data from <strong>the</strong> core show<br />
reliable signals at appropriate depths, matching<br />
closely with results from o<strong>the</strong>r mercury studies <strong>of</strong><br />
peat bogs (Martínez-Cortizas et al., 1999; Biester et<br />
al., 2002; Givelet et al., 2003; Schuster et al., 2002,<br />
Bindler, 2002). Results from this core demonstrate<br />
a set <strong>of</strong> peaks in mercury concentrations in <strong>the</strong> top<br />
75cm depth, with a major drop in mercury concentration<br />
approaching <strong>the</strong> top <strong>of</strong> <strong>the</strong> core above 20cm.<br />
These peaks are all well above 50 ppb, indicating an<br />
additional source <strong>of</strong> mercury in addition to <strong>the</strong> natural<br />
deposition.<br />
From 75-70cm depth, mercury concentration rises<br />
101<br />
Natural Background Level--23.2 ppb<br />
135cm: Aniakchak, AK eruption 3435 BP<br />
Unknown<br />
See Figure 5<br />
75cm: Silver mining in <strong>the</strong> Americas 400 BP<br />
80cm: Ceboruco eruption, 1020 BP<br />
105cm: Okmok, AK eruption 2050 BP<br />
180cm: Indonesian eruptions, 5500-5300 BP<br />
305-345cm: Nor<strong>the</strong>rn Hemisphere climatic shift<br />
}<br />
Figure 4: Mercury concentrations with depth in <strong>the</strong> complete<br />
North Boulder Creek Bog core, correlated to 14 CyBP.<br />
Red dots represent individual samples, and <strong>the</strong> dashed<br />
black line is <strong>the</strong> interpreted flux in concentration.<br />
sharply to 61.9 ppb. According to <strong>the</strong> age-depth<br />
model, 75cm corresponds to an age <strong>of</strong> 509 yBP,<br />
closely following <strong>the</strong> onset <strong>of</strong> silver mining in <strong>the</strong><br />
Americas in <strong>the</strong> 1550s when mercury was used to<br />
amalgamate <strong>the</strong> silver from its natural compounds.<br />
The sharp peak above 75cm depth likely represents<br />
<strong>the</strong> first clear distinction between natural and anthropogenic<br />
mercury deposition. Below this depth, it can<br />
be inferred that <strong>the</strong> majority <strong>of</strong> mercury deposition<br />
was due to natural causes, including dust loads,<br />
volcanic events, and climatic fluxes (Pirrone et al.,<br />
2010; Martínez-Cortizas et al., 1999). Alternatively,<br />
above 75cm it can be inferred that anthropogenicallyinduced<br />
mercury deposition is a significant contribu<br />
tor in addition to <strong>the</strong> natural mercury deposition.<br />
Thus, according to historic data, <strong>the</strong> mercury depos<br />
ited above 70cm would have originated from modern<br />
mining, industrial pollution, waste incineration,<br />
WWII, coal burning, volcanic events, and climatic<br />
shifts (Hylander and Meili, 2003; Pirrone et al.,<br />
2010).<br />
In addition to <strong>the</strong> onset <strong>of</strong> ore mining and processing,<br />
<strong>the</strong> climate was also changing rapidly at this time in<br />
<strong>the</strong> Nor<strong>the</strong>rn Hemisphere. Colder temperatures are<br />
known to sequester higher quantities <strong>of</strong> elemental<br />
–60<br />
0<br />
180<br />
1,800<br />
4,000<br />
6,000<br />
7,300<br />
8,400<br />
9,000<br />
Age (cal. years BP)
24th Annual Keck Symposium: 2011 Union College, Schenectady, NY<br />
mercury, thus shifts to a colder climate may have<br />
caused a higher retention <strong>of</strong> mercury in <strong>the</strong> bog near<br />
70 cm (Martínez-Cortizas et al., 1999). The Little<br />
Ice Age took place from 500-250 yBP (Mann et al.,<br />
2009), and could have contributed to <strong>the</strong> higher retention<br />
<strong>of</strong> mercury.<br />
PRE-ANTHROPOGENIC<br />
Below 70cm, anomalous mercury peaks occur at<br />
80cm, 105cm, 135cm, 180cm, and 230cm in <strong>the</strong> core.<br />
From 85 to 80cm depth, mercury concentration rises<br />
to a small peak <strong>of</strong> approximately 50 ppb. At 85cm,<br />
<strong>the</strong> age-depth model estimates an age <strong>of</strong> 1039 yBP,<br />
and at 80cm an estimate <strong>of</strong> 776 BP. At 1020 ± 200<br />
BP, <strong>the</strong> Mexican volcano Ceboruco erupted, releasing<br />
about 1.1 ± 0.08 x 10 10 m 3 <strong>of</strong> volcanic material and<br />
producing an explosion rated as a 6 on <strong>the</strong> Volcanic<br />
Explosivity Index (Smithsonian Institution, Global<br />
Volcanism Program). Given a time window <strong>of</strong> 1039-<br />
776 yBP, <strong>the</strong> mercury peak at 80cm likely resulted,<br />
in part, from <strong>the</strong> 1020 ± 200 BP eruption, which took<br />
place 2,090km away from <strong>the</strong> bog.<br />
The next peak <strong>of</strong> 58.6 ppb Hg is recorded at 105cm,<br />
corresponding to an age <strong>of</strong> 2048 yBP. This peak may<br />
in part be explained by <strong>the</strong> Okmok eruption in <strong>the</strong><br />
Aleutian Islands in 100 ± 50 BC (2050 BP). This<br />
eruption was a VEI 6 and ejected 5.0 ± 1.0 x 10 10 m 3<br />
<strong>of</strong> tephra (Smithsonian Institution). The Okmok Caldera<br />
is located just 4,828km from North Boulder<br />
Creek Bog and may be a contributor to <strong>the</strong> anomalous<br />
mercury peak that occurs just two years after its alculated<br />
eruption date.<br />
The 54.6 ppb Hg peak at 135cm corresponds to an<br />
age <strong>of</strong> 3437 yBP, just following <strong>the</strong> Aniakchak eruption<br />
in Alaska, US. This eruption was rated a VEI 6<br />
and released over 5 x 10 10 m 3 <strong>of</strong> tephra (Smithsonian<br />
Institution). The extreme magnitude <strong>of</strong> <strong>the</strong> Aniakchak<br />
eruption and its close proximity to <strong>the</strong> deposition site<br />
(6,700km) make it a likely candidate for <strong>the</strong> anomalous<br />
peak at 135cm.<br />
Climate may also have an effect on <strong>the</strong> amount <strong>of</strong><br />
mercury deposited in <strong>the</strong> bog between 3438 and 2048<br />
yBP, when <strong>the</strong> two previously mentioned mercury<br />
peaks were likely deposited. The cooler time period<br />
102<br />
between 3,500-2,500 yBP in <strong>the</strong> Nor<strong>the</strong>rn Hemisphere<br />
was characterized by ice rafting in <strong>the</strong> North<br />
Atlantic, high latitude cooling, and alpine glacier<br />
retreat, and is believed to be a period <strong>of</strong> Rapid Climate<br />
Change (RCC) resulting from a decline in solar<br />
output (Mayewski et al., 2004).<br />
At 180cm depth mercury rises to 58.8 ppb, but re<br />
mains at high values between 195-175cm (5932-5056<br />
yBP). This lengthy increase in mercury concentration<br />
may be partly due to <strong>the</strong> combined effects <strong>of</strong> <strong>the</strong> two<br />
eruptions <strong>of</strong> Luzon in Indonesia, which occurred at<br />
5530 and 5500 yBP at Taal and Pinatubo, respectively.<br />
Toge<strong>the</strong>r <strong>the</strong> VEI 6 eruptions released a total<br />
<strong>of</strong> 6.3 x 10 10 m 3 <strong>of</strong> tephra into <strong>the</strong> atmosphere (Smithsonian<br />
Institution). Alternatively, <strong>the</strong> sustained rise<br />
in mercury concentration in this part <strong>of</strong> <strong>the</strong> core may<br />
also be due to a shift to colder temperatures between<br />
6000 and 5000 BP—similar to <strong>the</strong> climate shift that<br />
also took place between 3,500-2,500 yBP. The colder<br />
climates in <strong>the</strong> Nor<strong>the</strong>rn Hemisphere during this<br />
time are similarly attributed to solar variability<br />
(Mayewski et al., 2004).<br />
From 245cm to 230cm depth, ano<strong>the</strong>r peak climbs<br />
from 22.4 ppb Hg to a maximum <strong>of</strong> 72.5 ppb Hg. At<br />
245cm, our age-depth model approximates an age<br />
<strong>of</strong> 7118 yBP (5,168 BC). Sufficient volcanic or climatic<br />
events cannot be correlated with <strong>the</strong> timing <strong>of</strong><br />
this peak; <strong>the</strong>refore <strong>the</strong> mercury concentration at this<br />
depth cannot be attributed to a single point source.<br />
Higher resolution mercury analyses at this depth may<br />
yield more informative data.<br />
Finally, <strong>the</strong>re is an anomalous increase in mercury<br />
concentration between 345 and 305 cm in <strong>the</strong> core,<br />
representing a plateau just slightly above <strong>the</strong> background<br />
concentration. This plateau seems to linger<br />
for approximately 500 years, between 8962 and 8477<br />
years BP, with a slight drop at 325 cm. Nor<strong>the</strong>rn<br />
Hemispheric climate experienced a rapid shift to<br />
colder temperatures at 8,200 years BP, when <strong>the</strong><br />
North Atlantic region received a large meltwater burst<br />
from proglacial lakes, causing both deepwater circulation<br />
and Nor<strong>the</strong>rn Hemisphere temperature regulation<br />
to weaken (Born and Levermann, 2010). Due to <strong>the</strong><br />
improbability <strong>of</strong> sustained volcanic influence during<br />
<strong>the</strong> time at which this plateau appears, it is highly
24th Annual Keck Symposium: 2011 Union College, Schenectady, NY<br />
likely that Holocene climate shifts played a large role<br />
in <strong>the</strong> retention <strong>of</strong> mercury in North Boulder Creek<br />
Bog.<br />
ANTHROPOGENIC<br />
Assuming <strong>the</strong> natural background level <strong>of</strong> mercury<br />
deposition has remained constant at this location, <strong>the</strong><br />
anthropogenic input <strong>of</strong> mercury has ranged from<br />
about 13-136 ppb since <strong>the</strong> 1550s. Above 75cm in<br />
<strong>the</strong> core, <strong>the</strong> most prominent peak resides at 20cm<br />
(Fig. 5), which has been fit to o<strong>the</strong>r mercury curves in<br />
order to obtain an accurate date <strong>of</strong> 67 yBP, when<br />
Mount Krakatau erupted in Indonesia on August<br />
26, 1883. The massive eruption was rated a VEI 6<br />
and ejected 2.0 ± 0.2 x 10 10 m 3 <strong>of</strong> tephra (Smithsonian<br />
Institution), and is <strong>the</strong>refore a reliable marker with<br />
which to pinpoint <strong>the</strong> date <strong>of</strong> that peak. There is a<br />
smaller peak just below Krakatau at 42cm depth,<br />
which is interpreted as <strong>the</strong> Tambora eruption <strong>of</strong><br />
1815 in Indonesia—a VEI 7 that erupted 1.6 x 10 11<br />
m 3 <strong>of</strong> tephra. Between <strong>the</strong>se two peaks, <strong>the</strong> mercury<br />
concentration drops significantly before<br />
3.5cm: WWII<br />
20cm: Krakatau<br />
Mt. St. Helens 1980 – -30 BP<br />
38-32cm: Gold Rush<br />
42cm: Tambora<br />
– 5 BP<br />
–67 BP<br />
– 85 BP<br />
– 100 BP<br />
– 135 BP<br />
Figure 5: Zoomed inset from Figure 4. Mercury concentrations<br />
<strong>of</strong> high-resolution samples above 45cm in <strong>the</strong> core,<br />
calibrated to yBP. Red dots represent individual samples,<br />
and <strong>the</strong> dashed black line is <strong>the</strong> interpreted flux in concentration.<br />
Age (yrs BP)<br />
103<br />
Krakatau, and remains at a small plateau between 38<br />
and 32cm (137.2-120.4 ppb). This small plateau,<br />
less concentrated in mercury than <strong>the</strong> peak <strong>of</strong><br />
Krakatau but slightly more than that <strong>of</strong> Tambora, is<br />
interpreted as <strong>the</strong> signal <strong>of</strong> mercury deposited by <strong>the</strong><br />
American Gold Rush from 1850-1865, when mercury<br />
was used as an amalgamator for gold and silver ore<br />
processing.<br />
Above <strong>the</strong> 20cm peak, ano<strong>the</strong>r significant peak<br />
<strong>of</strong> 158.9 ppb resides at 0.5cm. Due to its shallow<br />
position and extremely high mercury signal, this peak<br />
is interpreted as <strong>the</strong> mercury released and deposited<br />
during <strong>the</strong> Mount St. Helens eruption in March <strong>of</strong><br />
1980. The VEI 5 eruption, just 1510km away from<br />
<strong>the</strong> deposition site, released 7.4 x 10 7 m 3 <strong>of</strong> lava and<br />
1.2 x 10 9 m 3 <strong>of</strong> tephra into <strong>the</strong> atmosphere (Smithsonian<br />
Institution). Below <strong>the</strong> Mt. St. Helens peak,<br />
<strong>the</strong>re is a smaller increase in mercury concentration at<br />
3.5cm (126.6 ppb). This small jump is likely a mercury<br />
signal deposited during WWII, when <strong>the</strong> defense<br />
industry was utilizing mercury to manufacture explosives.<br />
Above <strong>the</strong> Mt. St. Helens peak, our data record <strong>the</strong><br />
drop in atmospheric mercury concentration during<br />
<strong>the</strong> past few decades, marked by a total concentration<br />
<strong>of</strong> 86.6 ppb at 0cm, down from a peak <strong>of</strong> 158.9 ppb.<br />
This result is congruent with modern measurements<br />
that have shown a decrease in atmospheric mercury<br />
contributions from anthropogenic sources<br />
(Hylander and Meili, 2003). Given a natural background<br />
deposition <strong>of</strong> 23.2 ppb Hg, our most recent<br />
sample indicates an input <strong>of</strong> 63.4 ppb Hg from human<br />
activity at present, which includes coal combustion,<br />
industrial processes, and waste incineration.<br />
CONCLUSION<br />
As an ombrotrophic peat bog, North Boulder Creek<br />
Bog holds a well-recorded history <strong>of</strong> mercury deposition<br />
since approximately 9,000 yBP. The core used in<br />
this project did not reach <strong>the</strong> oldest portion <strong>of</strong> <strong>the</strong><br />
bog, thus analyses on an additional core in a deeper<br />
location are recommended. A search for tephra using<br />
SEM analysis is recommended at <strong>the</strong> depths<br />
where we believe volcanic signals are located.
24th Annual Keck Symposium: 2011 Union College, Schenectady, NY<br />
ACKNOWLEDGEMENTS<br />
This project was made possible by funding from <strong>the</strong><br />
National Science Foundation. I would also like to extend<br />
<strong>the</strong> utmost appreciation to Bob Nelson and Matthias<br />
Leopold for <strong>the</strong>ir assistance in tackling <strong>the</strong> bog,<br />
and to David Dethier and Will Ouimet for <strong>the</strong>ir instruction<br />
and advising in <strong>the</strong> field.<br />
REFERENCES<br />
Bindler, R., 2003, Estimating <strong>the</strong> Natural Background<br />
Atmospheric Deposition Rate <strong>of</strong> Mercury Utilizing<br />
Ombrotrophic Bogs in Sou<strong>the</strong>rn Sweden:<br />
Environ. Sci. Technol., 37 (1), p. 40–46.<br />
Biester, H., Kilian, R., Franzen, C., Woda, C., Mangini,<br />
A., and Schöler, H.F., 2002, Elevated mercury<br />
accumulation in a peat bog <strong>of</strong> <strong>the</strong> Magellanic<br />
Moorlands, Chile (53ºS) – an anthropogenic<br />
signal from <strong>the</strong> Sou<strong>the</strong>rn Hemisphere: Earth and<br />
Planetary Science Letters no. 201, p. 609-620.<br />
Born, A., Levermann, A., 2010, The 8.2 ka event:<br />
Abrupt transition <strong>of</strong> <strong>the</strong> subpolar gyre toward a<br />
modern North Atlantic circulation: Geochemistry<br />
Geophysics Geosystems, v. 11, no. 6, p. 1-8.<br />
Givelet, N., Roos-Barraclough, F., and Shotyk, W.,<br />
2003, Predominant anthropogenic sources and<br />
rates <strong>of</strong> atmospheric mercury accumulation in<br />
sou<strong>the</strong>rn Ontario recorded by peat cores from<br />
three bogs: comparison with natural “background”<br />
values (past 8000) years: J. Environ.<br />
Monit., v. 5, p. 939-949.<br />
Hylander, L.D., and Meili, M., 2003, 500 years <strong>of</strong><br />
mercury production: global <strong>annual</strong> inventory by<br />
region until 2000 and associated emissions: The<br />
Science <strong>of</strong> <strong>the</strong> Total Environment, v. 304, p. 13-<br />
27.<br />
Leopold, M., and Dethier, D., 2007, Near surface geophysics<br />
and sediment analysis to precisely date<br />
<strong>the</strong> outbreak <strong>of</strong> glacial Lake Devlin, Front Range<br />
Colorado: Eos Trans. AGU, 88(52), Fall Meet.<br />
Suppl., Abstract H51E– 0809.<br />
104<br />
Livingstone, D.A., 1955, A lightweight piston sampler<br />
for lake deposits: Ecology, v. 36, p. 137-139.<br />
Lodenius, M., Seppänen, A., and Uusi-Rauva, A.,<br />
1983, Sorption and mobilization <strong>of</strong> mercury in<br />
peat soil: Chemosphere, v. 12, p. 1575–1581.<br />
Madole, R.F., 1985, Lake Devlin and Pinedale glacial<br />
history, Front Range, Colorado: Quaternary<br />
Research, v. 25, p. 43-54.<br />
Madsen, P.P., 1981, Peat bog records <strong>of</strong> atmospheric<br />
mercury deposition: Nature, v. 293, p. 127-130.<br />
Mann, M.E., Zhang, Z., Ru<strong>the</strong>rford, S., Bradley,<br />
R.S., Hughes, M.K., Shindell, D., Ammann, C.,<br />
Faluvegi, G., and Ni, F., 2009, Global Signatures<br />
and Dynamical Origins <strong>of</strong> <strong>the</strong> Little Ice Age and<br />
Medieval Climate Anomaly: Science, v. 326, no.<br />
5957, p. 1256-1260.<br />
Martínez-Cortizas, A., Pontevedra-Pombal, X., García-Rodeja,<br />
E., Nóvoa-Muñoz, J.C., and Shotyk,<br />
W., 1999, Mercury in a Spanish peat bog: Archive<br />
<strong>of</strong> climate change and atmospheric metal<br />
deposition: Science, v. 284, p. 939-942.<br />
Mayewski, P.A., Rohling, E.E., Stager, J.C., Karlén,<br />
W., Maasch, K.A., Meeker, L.D., Meyerson,<br />
E.A., Gasse, F., van Kreveld, S., Holmgren, K.,<br />
Lee-Thorp, J., Rosqvist, G., Rack, F., Staubwasser,<br />
M., Schneider, R.R., and Steig, E.J., 2004,<br />
Holocene climate variability: Quaternary Research,<br />
v. 62 (3), p. 243-255.<br />
Nriagu, J.O., 1994, Mercury pollution from <strong>the</strong> past<br />
mining <strong>of</strong> gold and silver in <strong>the</strong> Americas: Science<br />
<strong>of</strong> <strong>the</strong> Total Environment, v. 149, p. 167-<br />
181.<br />
Pirrone, N., Cinnirella, S., Feng, X., Finkelman, R.B.,<br />
Friedli, H.R., Leaner, J., Mason, R., Mukherjee,<br />
A.B., Stracher, G.B., Streets, D.G., and Telmer,<br />
K., 2010, Global mercury emissions to <strong>the</strong> atmosphere<br />
from anthropogenic and natural sources:<br />
Atmos. Chem. Phys., v.10, p. 5951-5964.<br />
Schuster, P.F., Krabbenh<strong>of</strong>t, D.P., Naftz, D.L., Cecil,
24th Annual Keck Symposium: 2011 Union College, Schenectady, NY<br />
L.D., Olson, M.L., Dewild, J.F., Susong, D.D.,<br />
Green, J.R., and Abbott, M.L., 2002, Atmospheric<br />
mercury depostion during <strong>the</strong> last 270 years:<br />
a glacial ice core record <strong>of</strong> natural and anthropogenic<br />
sources: Environ. Sci. Technol., v. 36, no.<br />
11, p. 2303-2310.<br />
Smithsonian Institution, National Museum <strong>of</strong> Natural<br />
History, Global Volcanism Program, “Volcanoes<br />
<strong>of</strong> <strong>the</strong> World: Large Eruptions,” http://www.<br />
volcano.si.edu/world/largeeruptions.cfm.<br />
Stuiver, M. and Polach, H. A., 1977, Discussion:<br />
Reporting <strong>of</strong> 14C data: Radiocarbon, v. 19, p.<br />
355-363.<br />
Stuiver, M., 1980. Workshop on 14C data reporting:<br />
Radiocarbon, v. 22, p. 964-966.<br />
Vandal, G.M., Fitzgerald, W.F., Boutron, C.F., and<br />
Candelone, J., 1993, Variations in mercury deposition<br />
to Antarctica over <strong>the</strong> past 34,000 years:<br />
Nature, v. 362, p. 621-623.<br />
105
24th Annual Keck Symposium: 2011 Union College, Schenectady, NY<br />
THE DISTRIBUTION OF PHOSPHORUS IN ALPINE AND<br />
UPLAND SOILS OF THE BOULDER CREEK, COLORADO<br />
CATCHMENT<br />
HAYLEY CORSON-RIKERT, Wesleyan University<br />
Research Advisor: Timothy Ku<br />
INTRODUCTION<br />
Biological growth within terrestrial ecosystems is<br />
generally limited by <strong>the</strong> concentration <strong>of</strong> nitrogen,<br />
phosphorus, or both (Sato et al., 2009). Investigation<br />
into <strong>the</strong> availability <strong>of</strong> both <strong>the</strong>se macronutrients<br />
in modern day alpine environments is important as<br />
N + P availability determines how such ecosystems<br />
will respond to climatic changes and anthropogenic<br />
alterations <strong>of</strong> soil chemistry (Wu et al., 2006). Recent<br />
studies have shown that enhanced rates <strong>of</strong> nitrogen<br />
deposition can force alpine systems that are typically<br />
N-limited to become P-limited, especially when P is<br />
efficiently cycled, making investigation into soil P<br />
dynamics yet more important (Sievering et al., 1996;<br />
Hedin et al., 2003; Vitousek et al., 2010).<br />
Distribution <strong>of</strong> soil P occurs through geochemical<br />
and biochemical pathways, and is controlled by <strong>the</strong><br />
demand for and supply <strong>of</strong> P in soil horizons (McGill<br />
and Cole, 1981). Crystalline or primary mineral P in<br />
deeper soils represents a long-term soil P reservoir,<br />
whereas secondary mineral forms and in particular<br />
labile forms are cycled more rapidly in upper and/or<br />
surface horizons (Walker and Syers, 1976). On short<br />
timescales, <strong>the</strong> availability <strong>of</strong> labile soil P to plants is<br />
dependent on a number <strong>of</strong> factors, including temperature,<br />
moisture, aeration, and soil microorganism activity<br />
(Tate & Salcedo, 1988). In <strong>the</strong> long term, labile P<br />
availability is dependent on <strong>the</strong> state <strong>of</strong> soil development,<br />
which in turn is determined by soil residence<br />
time and <strong>the</strong> rate <strong>of</strong> chemical and physical wea<strong>the</strong>ring<br />
(Walker and Syers, 1976; Porder et al., 2007).<br />
In this study, I examine <strong>the</strong> soil P reservoirs <strong>of</strong> four<br />
soil pr<strong>of</strong>iles across an elevation gradient in Boulder<br />
County, Colorado, in order to better understand<br />
<strong>the</strong> patterns <strong>of</strong> and controls on soil P distribution in<br />
alpine environments. The four selected pr<strong>of</strong>iles are<br />
a subset <strong>of</strong> a broader set <strong>of</strong> studied soils in <strong>the</strong> Boulder<br />
Creek NCZO, and represent a range <strong>of</strong> elevations and<br />
climatic conditions. From greatest to least elevation,<br />
<strong>the</strong> sites are GLV, in <strong>the</strong> Green Lakes Valley; SLM, at<br />
<strong>the</strong> moraine below Silver Lake; UGG, in upper Gordon<br />
Gulch; and Betasso, in <strong>the</strong> Betasso Preserve (Table 1).<br />
The soils at GLV and SLM are relatively stable, with<br />
minimal soil movement, while <strong>the</strong> UGG and Betasso<br />
pr<strong>of</strong>iles are marked by buried horizons, which represent<br />
discontinuities in <strong>the</strong> soil sequence. Mean <strong>annual</strong><br />
temperature near <strong>the</strong> highest site averages -3.7ºC, while<br />
average <strong>annual</strong> temperature at Betasso are about 10ºC<br />
(Niwot Ridge LTER; NOAA). Annual precipitation at<br />
this lower altitude is about 40 cm, while precipitation at<br />
<strong>the</strong> continental divide above GLV can amount to more<br />
than 100 cm <strong>annual</strong>ly (Table 1; Birkeland et al., 2003).<br />
METHODS<br />
In July and August 2010, soils from 31 sites were collected<br />
from newly scraped exposures or fresh soil pits.<br />
Collected samples were stored in plastic bags vacated<br />
<strong>of</strong> air in order to best preserve field moisture. Soil pH<br />
in water and soil moisture were determined by standard<br />
methods (Carter and Gregorich, 2008). Total carbon<br />
and nitrogen concentrations were determined on a<br />
Thermo Flash 1112 Elemental Analyzer. Given <strong>the</strong> lack<br />
<strong>of</strong> carbonate minerals in <strong>the</strong>se soils, total carbon (TC)<br />
is assumed to equal total organic carbon (TOC). Bulk<br />
chemistry analysis for metals was determined by ICP-<br />
OES techniques at SGS Mineral Services, after dissolving<br />
soils in a four-acid digest (HCl/HNO 3 /HF/HClO 4 ).<br />
The digestion may not have completely dissolved very<br />
recalcitrant mineral phases. On 21 samples, soil P<br />
pools were determined by a modified Hedley sequential<br />
extraction procedure (Figure 1; Hedley et al., 1982;<br />
Ruttenberg, 1992; Tiessen and Moir, 1993). Inorganic<br />
(Pi) and total phosphorus (Pt) concentrations were<br />
determined by spectrophotometry methods <strong>of</strong> Murphy<br />
112
24th Annual Keck Symposium: 2011 Union College, Schenectady, NY<br />
and Riley (1962) using a Beckman Coulter DU5300<br />
at a wavelength <strong>of</strong> 885 nm. Organic phosphorus (Po)<br />
concentrations were determined by subtraction <strong>of</strong> Pi<br />
from Pt.<br />
Figure 1. Flowchart <strong>of</strong> sequential phosphorus extraction<br />
methodology. The procedure is a modification <strong>of</strong> Tiessen<br />
and Moir (1993), with <strong>the</strong> ashing step <strong>of</strong> Ruttenberg<br />
(1992). Pi = inorganic phosphorus, and Pt = total phosphorus.<br />
Autoclaving conditions were 121ºC, 17 psi, for 50<br />
minutes.<br />
Measured extractable P fractions were grouped to<br />
obtain operationally-defined soil pools: Exchangeable<br />
P (NaHCO3 Pi); Organic P (NaHCO 3 Po + NaOH Po<br />
+ C. HCl Po); Fe-bound P (NaOH Pi); Ca-bound P<br />
(1M HCL Pi + 1M HCl Po); Recalcitrant P (C. HCl<br />
Pi); and Highly Recalcitrant P (Ashed Pi + Ashed Po)<br />
(Tiessen and Moir, 1993). The percentage <strong>of</strong> initial<br />
total P remaining in individual horizons was calculated<br />
relative to Al as follows, where X = sample<br />
horizon and Y = parent material (or deepest available<br />
horizon): Initial total P concentration equals [Al]X<br />
x ([Pt]Y/[Al]Y). The % <strong>of</strong> initial total P remaining<br />
113<br />
equals 100 x ([Pt]X/[Pt]Y) (Vitousek et al., 2004,<br />
Supp. Mat.). The percent <strong>of</strong> initial Ca-bound Pi was<br />
calculated in <strong>the</strong> same manner.<br />
RESULTS AND DISCUSSION<br />
General soil properties<br />
A summary <strong>of</strong> results is shown in Table 1. In <strong>the</strong>se<br />
four soil pr<strong>of</strong>iles, soil pH increases with depth, with<br />
surface horizons displaying pH ranging from 4.43 to<br />
5.61 and base horizons pH ranging from 5.51 to 6.05.<br />
The greater acidity <strong>of</strong> surface horizons is typically <strong>the</strong><br />
result <strong>of</strong> organic matter decay, which lowers <strong>the</strong> pH<br />
<strong>of</strong> soil pore waters (Twidale, 1990). This assumption<br />
is supported by <strong>the</strong> consistently high TOC concentrations<br />
in surface horizons (Figure 2). As expected,<br />
TOC is correlated with organic N throughout all horizons,<br />
and C:N, C:P 0 , N:P 0 , and soil moisture values<br />
decrease with depth (Table 1; Figure 2). P organic:P<br />
inorganic decreases with depth, demonstrating <strong>the</strong><br />
transition from surface horizons rich in organic and<br />
plant available P to deeper horizons dominated by<br />
primary and secondary mineral P (Figure 2). Soil<br />
concentrations <strong>of</strong> Al increase with depth at all sites<br />
(Table 1).<br />
Table 1. Table <strong>of</strong> study site information and soil properties.<br />
Soil ages from Dethier et al., unpublished data -- *<br />
denotes exposure ages measured by OSL techniques, †<br />
denotes age based on CRN techniques.<br />
Total Soil P<br />
Total P concentration pr<strong>of</strong>iles are presented in <strong>the</strong><br />
right-hand column <strong>of</strong> Figures 3 and 4. Total soil P<br />
concentration varies from 198 to 2853 ug/g across all<br />
horizons. These values are comparable to those <strong>of</strong><br />
o<strong>the</strong>r alpine soil studies, which were generally be
24th Annual Keck Symposium: 2011 Union College, Schenectady, NY<br />
Figure 2. Soil depth vs. TOC, C:N. C:P 0 , N:P 0 , and P<br />
organic:P inorganic. Note log scale on x-axis.<br />
tween 121 to 2540 ug/g (Table 1; e.g. Makarov et al.,<br />
1997). Total soil phosphorus concentrations show no<br />
discernable correlation with elevation. It is assumed<br />
that <strong>the</strong> variation in moisture regimes, parent material,<br />
soil residence time, and wea<strong>the</strong>ring rates is too great<br />
for a clear pattern to emerge.<br />
The Development <strong>of</strong> Soil P Reservoirs<br />
Soil P is widely held to exist in three main pools:<br />
primary inorganic P, secondary inorganic P, and organic<br />
P. Walker and Syers (1976) presented a model<br />
in which primary inorganic P (mineral P) is abundant<br />
in early stages <strong>of</strong> soil development and, as wea<strong>the</strong>ring<br />
progresses, is transformed into organic forms<br />
and sorbed to secondary minerals. Thus, <strong>the</strong> primary<br />
mineral P fraction becomes depleted as <strong>the</strong> organic P<br />
and secondary mineral P fractions are enriched. At<br />
first, a portion <strong>of</strong> this sorbed inorganic secondary<br />
mineral P is exchangeable, or plant available, but this<br />
labile fraction is later diminished with <strong>the</strong> exhaustion<br />
<strong>of</strong> <strong>the</strong> primary mineral P reservoir. In more highly<br />
wea<strong>the</strong>red pr<strong>of</strong>iles, labile P is fur<strong>the</strong>r depleted due<br />
to <strong>the</strong> progressive transformation <strong>of</strong> <strong>the</strong> secondary<br />
mineral P into occluded, or recalcitrant, forms that are<br />
biologically unavailable. Within soil pr<strong>of</strong>iles, horizon<br />
development progresses vertically, with upper horizons<br />
originating from parent material. Each horizon,<br />
<strong>the</strong>refore, represents a stage and/or type <strong>of</strong> soil development,<br />
and soil P distribution within <strong>the</strong>se horizons<br />
114<br />
should reflect <strong>the</strong>ir position on <strong>the</strong> continuum. In<br />
effect, primary mineral P is expected to decrease from<br />
deeper to surface horizons, while organic, secondary<br />
mineral, and labile P are expected to increase<br />
(Walker and Syers, 1976; Stewart and Tiessen, 1987;<br />
Crews, 1995; Porder et al., 2007). In stable soils, this<br />
increase in plant-available P and decrease in primary<br />
mineral P occurs as total P is diminished due to net P<br />
removal by wea<strong>the</strong>ring. This pattern is most visible<br />
in sites that experience high <strong>annual</strong> levels <strong>of</strong> precipitation,<br />
due to <strong>the</strong> enhancement <strong>of</strong> soil redox processes<br />
and thus <strong>the</strong> quickening <strong>of</strong> mineral P dissolution and<br />
removal (Miller et al., 2001; Hedin et al., 2003).<br />
Distribution <strong>of</strong> Soil P Pools in <strong>the</strong> Boulder Creek<br />
Catchment<br />
Figures 3 and 4 show <strong>the</strong> distribution <strong>of</strong> P fractions,<br />
total P, and calculated values <strong>of</strong> % P remaining at <strong>the</strong><br />
four study sites. The importance <strong>of</strong> various soil P<br />
transformation processes is reflected in <strong>the</strong> distribution<br />
<strong>of</strong> <strong>the</strong>se soil P pools, and is impacted by <strong>the</strong> extent<br />
to which <strong>the</strong> soil pr<strong>of</strong>ile has been disturbed during<br />
development (Beck and Elsenbeer, 1999). Buried<br />
horizons at UGG and Betasso indicate that <strong>the</strong>se soils<br />
have experienced more soil movement than <strong>the</strong> more<br />
stationary pr<strong>of</strong>iles <strong>of</strong> GLV and SLM. At GLV and<br />
SLM, Ca-bound Pi generally increases with depth and<br />
exchangeable P decreases with depth. The O and A<br />
horizons at GLV are <strong>the</strong> exception to this pattern, as<br />
<strong>the</strong>y have proportionally greater concentrations <strong>of</strong> Cabound<br />
P than <strong>the</strong> horizons immediately below. These<br />
elevated levels <strong>of</strong> primary mineral P, coupled with <strong>the</strong><br />
concomitant rise <strong>of</strong> remaining initial total P to percentages<br />
greater than 100, points to an external input<br />
<strong>of</strong> comparatively unwea<strong>the</strong>red hillslope colluvium or<br />
eolian material. This is supported by <strong>the</strong> P organic:P<br />
inorganic ratios <strong>of</strong> <strong>the</strong> O and A horizons (Figure 2),<br />
which are slightly lower than in <strong>the</strong> horizons immediately<br />
below, suggesting that <strong>the</strong>se upper horizons are<br />
less wea<strong>the</strong>red than <strong>the</strong> B horizons below (Tate and<br />
Salcedo, 1988). Soil P distribution within <strong>the</strong>se GLV<br />
B horizons instead appears to be <strong>the</strong> result <strong>of</strong> continued<br />
wea<strong>the</strong>ring, as a net loss <strong>of</strong> total P due to mineral<br />
dissolution is apparent from <strong>the</strong> Cu to Bw1 horizon.<br />
Site SLM, due to its stability, has <strong>the</strong> most standard<br />
distribution <strong>of</strong> soil P fractions, with a clear inverse
24th Annual Keck Symposium: 2011 Union College, Schenectady, NY<br />
Figure 3. GLV and SLM Soil P reservoirs, with <strong>the</strong><br />
left-hand chart depicting <strong>the</strong> relative percentage <strong>of</strong> each<br />
P-reservoir per soil horizons, and <strong>the</strong> right-hand chart<br />
illustrating total P (bars, lower x-axis) and <strong>the</strong> % <strong>of</strong> initial<br />
total P and Ca-bound Pi remaining in each horizon, relative<br />
to Al (upper x-axis).<br />
relationship between <strong>the</strong> organic P and Ca-bound P<br />
fractions. This clean pattern indicates that soil surface<br />
horizons are developed almost entirely from <strong>the</strong><br />
parent material, though some eolian deposition may<br />
have altered surface horizon composition. Here, like<br />
in <strong>the</strong> lower GLV horizons, total soil P is highest at<br />
depth and decreases above <strong>the</strong> parent material horizon.<br />
Measurements <strong>of</strong> % initial total P remaining<br />
also decrease, suggesting that P is consistently being<br />
removed from <strong>the</strong> soil system throughout all horizons,<br />
leading to a decrease in Ca-bound P and a subsequent<br />
enrichment <strong>of</strong> <strong>the</strong> organic P fraction.<br />
The lower two studies sites, UGG and Betasso, have<br />
a more complex history than GLV and SLM, as both<br />
contain buried horizons. At UGG, though organic P<br />
content decreases with depth, as expected, and Cabound<br />
P concomitantly increases, <strong>the</strong>re is a clear<br />
difference between buried and current soil horizons.<br />
Total P decreases sharply between <strong>the</strong> lower bBt2 and<br />
<strong>the</strong> bBt above it. Similarly, Ca-bound P is at least<br />
70% <strong>of</strong> total P in <strong>the</strong> lowermost horizons, but only<br />
~7% <strong>of</strong> total P in <strong>the</strong> Bt1, Cox, and Bw horizons.<br />
This low total P content in <strong>the</strong>se middle three horizons,<br />
coupled with <strong>the</strong>ir corresponding low percentage<br />
<strong>of</strong> primary mineral P, and high percentage <strong>of</strong> organic<br />
P and recalcitrant P indicate that <strong>the</strong>se horizons<br />
115<br />
are highly wea<strong>the</strong>red, despite <strong>the</strong>ir relatively young<br />
exposure age (Table 1). This conclusion, in turn, suggests<br />
that that <strong>the</strong>se upper horizons have ei<strong>the</strong>r experienced<br />
intense and rapid wea<strong>the</strong>ring, or formed from<br />
already wea<strong>the</strong>red material that was transported to<br />
this site, burying <strong>the</strong> lowermost bBt2 and CRt horizons<br />
that had formed in situ. Soil exposure ages support<br />
this last conclusion, as <strong>the</strong> two bottom horizons<br />
were last exposed 20,000 years ago, while <strong>the</strong> upper<br />
‘moved’ horizons were exposed much more recently,<br />
roughly 2,000 years ago (Table 1). Importantly,<br />
within <strong>the</strong> two soil brackets above, {CRt-bBt2} and<br />
{bBt-A}, <strong>the</strong> total soil P and % P remaining do not<br />
diminish with decreasing depth, indicating no net P<br />
loss. This is in contrast to <strong>the</strong> net P loss at <strong>the</strong> higher<br />
GLV and SLM pr<strong>of</strong>iles. The relatively larger fraction<br />
<strong>of</strong> Ca-bound P in <strong>the</strong> surface A horizon suggests that<br />
soil P distribution in <strong>the</strong> near surface environment is<br />
skewed by an external influx <strong>of</strong> unwea<strong>the</strong>red material<br />
rich in primary mineral P (Figure 4).<br />
Figure 4. UGG and Betasso Soil P reservoirs, with <strong>the</strong><br />
left-hand chart depicting <strong>the</strong> relative percentage <strong>of</strong> each<br />
P-reservoir per soil horizons, and <strong>the</strong> right-hand chart<br />
illustrating total P (bars, lower x-axis) and <strong>the</strong> % <strong>of</strong> initial<br />
total P and Ca-bound Pi remaining in each horizon, relative<br />
to Al.<br />
The Betasso soil pr<strong>of</strong>ile shows no similar enrichment<br />
<strong>of</strong> Ca-bound P in surface horizons, but contains a thin<br />
O horizon that is heavily enriched in organic P. This<br />
horizon is primarily composed <strong>of</strong> fresh and decaying<br />
plant litter and needles. Below <strong>the</strong> O horizon, organic
24th Annual Keck Symposium: 2011 Union College, Schenectady, NY<br />
P levels do not increase greatly, total P increases only<br />
slightly, and Ca-bound P remains a dominant portion<br />
<strong>of</strong> total soil P -- suggesting that <strong>the</strong> lower A, B, bA,<br />
and bBt horizons have experienced little wea<strong>the</strong>ring<br />
despite <strong>the</strong>ir age <strong>of</strong> 5 to 12 kyr. Given <strong>the</strong> comparatively<br />
low <strong>annual</strong> levels <strong>of</strong> rainfall at this site, ~40<br />
cm, this slow soil P development is likely due to <strong>the</strong><br />
limited percolation <strong>of</strong> moisture to deeper horizons,<br />
which would limit both chemical wea<strong>the</strong>ring and soil<br />
microbial activity.<br />
CONCLUSIONS<br />
Soil P pools at <strong>the</strong> four sites can be explained by<br />
continued wea<strong>the</strong>ring and patterns <strong>of</strong> soil movement.<br />
SLM shows <strong>the</strong> most consistent trend in soil development,<br />
with surface horizons enriched in exchangeable<br />
and organic P and deeper portions <strong>of</strong> <strong>the</strong> pr<strong>of</strong>ile<br />
enriched in Ca-bound P. GLV has a similar pr<strong>of</strong>ile,<br />
except that <strong>the</strong> upper layer likely contains relocated<br />
primary mineral P. This addition <strong>of</strong> outside material<br />
is also evident in <strong>the</strong> A and Bw horizons <strong>of</strong> <strong>the</strong> UGG<br />
pr<strong>of</strong>ile, suggesting that <strong>the</strong> relocation <strong>of</strong> primary mineral<br />
P by ei<strong>the</strong>r hillslope removal or eolian deposition<br />
may be an important factor in soil P distribution and<br />
development in surface soil environments across <strong>the</strong><br />
Front Range gradient. The Betasso site is relatively<br />
unwea<strong>the</strong>red, with little accumulation <strong>of</strong> organic P in<br />
<strong>the</strong> A horizon. This may be <strong>the</strong> result <strong>of</strong> a low degree<br />
<strong>of</strong> wea<strong>the</strong>ring experienced by soils at this altitude.<br />
Overall, wea<strong>the</strong>ring appears to be more intense at <strong>the</strong><br />
higher, wetter alpine sites <strong>of</strong> GLV and SLM, where a<br />
considerable fraction <strong>of</strong> soil P has been lost, than at<br />
UGG and Betasso, though <strong>the</strong> relocation <strong>of</strong> unwea<strong>the</strong>red<br />
and wea<strong>the</strong>red soil material, as seen at both SLM<br />
and UGG, serves to complicate this trend.<br />
REFERENCES<br />
Beck, M., and Elsenbeer, H., 1999, Biogeochemical<br />
cycles <strong>of</strong> soil phosphorus in sou<strong>the</strong>rn Alpine<br />
spodosols: Geoderma, vol. 91, p. 249-260<br />
Birkeland, P., Shroba, R., Burns, S., Price, A., and<br />
Tonkin, P., 2003, Integrating soils and geomorphology<br />
in mountains—an example from <strong>the</strong><br />
Front Range <strong>of</strong> Colorado: Geomorphology, v. 55,<br />
p. 329-344<br />
116<br />
Carter, M., and Gregorich, E., editors, 2008, Soil<br />
Sampling and Methods <strong>of</strong> Analysis: CRC Press,<br />
1224 p.<br />
Crews, T., Kitayama, K., Fownes, J., Riley, R., Herbert,<br />
D., Mueller-Dombois, R., and Vitousek, P.,<br />
1994, Changes in Soil Phosphorus Fractions and<br />
Ecosystem Dynamics Across and Long Chronosequence<br />
in Hawaii: Ecology, vol. 76, no. 5,<br />
p.1407-1424<br />
Hedin, L., Vitousek, P., and Matson, P., 2003, Nutrient<br />
Losses over Four Million Years <strong>of</strong> Tropical<br />
Forest Development: Ecology, vol. 84, no. 9, p.<br />
2231-2255<br />
Hedley, M., Stewart, J., and Chauhan, B., 1982,<br />
Changes in inorganic and organic soil phosphorus<br />
fractions induced by cultivation practices and<br />
laboratory incubations: Soil Sci. Soc. Am. J., vol.<br />
46, p. 970-976<br />
McGill, W., and Cole, C., 1981, Comparative Aspects<br />
<strong>of</strong> Cycling <strong>of</strong> Organic C, N, S and P through Soil<br />
Organic Matter: Geoderma, vol. 26, p. 267-286<br />
Makarov, M., Malysheva, T., Haumaier, L., Alt, H.,<br />
and Zech, W., 1997, The forms <strong>of</strong> phosphorus<br />
in humic and fulvic acids <strong>of</strong> a toposequence <strong>of</strong><br />
alpine soils in <strong>the</strong> nor<strong>the</strong>rn Caucasus: Geoderma,<br />
vol, 80, p. 61-73<br />
McGill, W., and Cole, C., 1981, Comparative Aspects<br />
<strong>of</strong> Cycling <strong>of</strong> Organic C, N, S and P through Soil<br />
Organic Matter: Geoderma, vol. 26, p. 267-286<br />
Miller, A., Schuur, E., and Chadwick, O., 2001,<br />
Redox control <strong>of</strong> phosphorus pools in Hawaiian<br />
montane forest soils: Geoderma, vol. 102, p.<br />
219-237<br />
Murphy, J., Riley, J., 1962, A modified single solution<br />
method for <strong>the</strong> determination <strong>of</strong> phosphate<br />
in natural waters: Anal. Chim. Acta., vol. 27, p.<br />
31-36<br />
Niwot Ridge LTER, “Site Information.” Niwot Ridge<br />
LTER. Web. 12 Dec. 2010.
24th Annual Keck Symposium: 2011 Union College, Schenectady, NY<br />
.<br />
NOAA, “Boulder Monthly Mean Temperature<br />
1897-present.” NOAA Earth System Research<br />
Laboratory. Web. 13 Dec. 2010. .<br />
Porder, S., Vitousek, P., Chadwick, O., Chamberlain,<br />
C., and Hilley, G., 2007, Uplift, Erosion, and<br />
Phosphorus Limitation in Terrestrial Ecosystems:<br />
Ecosystems, vol. 10, p. 158-170<br />
Ruttenberg, K., 1992, Development <strong>of</strong> a Sequential<br />
Extraction Method for Different Forms <strong>of</strong> Phosphorus<br />
in Marine Sediments: Limnology and<br />
Oceanography, vol. 37, no. 7, p. 1460-1482<br />
Sato, S., Neves, E., Solomon, D., Liang, B., and<br />
Lehmann, J., 2009, Biogenic calcium phosphate<br />
transformation in soils over millennial time<br />
scales: J Soils Sediments, vol. 9, p. 194-205<br />
Sievering, H., Rusch, D., and Marquez, L., 1996,<br />
Nitric acid, particulate nitrate, and ammonium in<br />
<strong>the</strong> continental free troposphere: Nitrogen deposition<br />
to an alpine tundra ecosystem: Atmospheric<br />
Environment, vol. 30, no. 14, p. 2527-2537<br />
Stewart, J., and Tiessen, H., 1987, Dynamics <strong>of</strong> soil<br />
organic phosphorus: Biogeochemistry, vol. 4, p.<br />
41-60<br />
Tate, K., and Salcedo, I., 1988, Phosphorus control<br />
<strong>of</strong> soil organic matter accumulation and cycling:<br />
Biogeochemistry, vol. 5, p. 99-107<br />
Tiessen, H., and Moir, J., 1993, Characterization <strong>of</strong><br />
available phosphorus by sequential extraction, in<br />
Carter, M. (Ed.), Soil Sampling and Method <strong>of</strong><br />
Analysis: Lewis, Chelsea, MI, p. 75-86<br />
Twidale, C., 1990, Wea<strong>the</strong>ring, soil development, and<br />
landforms: GSA Special Paper 252, p. 29-50<br />
Vitousek., P., Ladeforged, T., Kirch, P., Hartshorn, A.,<br />
Graves, M., Hotchkis, S., Tuljapurkar, S., and<br />
Chadwick, O., 2004, Supplementary Material<br />
117<br />
to Soils, Agriculture, and Society in Precontact<br />
Hawaii: Science, vol. 304, p. 1665-1669<br />
Vitousek, P., Porder, S., Houlton, B., and Chadwick,<br />
O., 2010, Terrestrial phosphorus limitation:<br />
mechanisms, implications, and nitrogen-phosphorus<br />
interactions: Ecological Applications,<br />
Vol. 20, No. 1, p. 5-15<br />
Walker, T., and Syers, J., 1976, The Fate <strong>of</strong> Phosphorus<br />
During Pedogenesis: Geoderma, vol. 15, p.<br />
1-19<br />
Wu, G., Wei, J., Deng, H., and Zhao, J, 2006, Nutrient<br />
cycling in an Alpine tundra ecosystem on<br />
Changbai Mountain, Nor<strong>the</strong>ast China: Applied<br />
Soil Ecology, vol. 32, p. 199-209
24th Annual Keck Symposium: 2011 Union College, Schenectady, NY<br />
EXAMINING KNICKPOINTS IN THE BOULDER CREEK<br />
CATCHMENT, COLORADO<br />
EVAN N. DETHIER, Williams College<br />
Research Advisor: David P. Dethier<br />
INTRODUCTION<br />
The apparent stasis <strong>of</strong> our current landscape belies <strong>the</strong><br />
constant change it has undergone for millions <strong>of</strong> years<br />
before <strong>the</strong> present. The landscape continues to transform<br />
as erosional denudation balances slow uplift <strong>of</strong><br />
rock material. Interaction between tectonics, climate,<br />
and surface processes are linked by channels, and<br />
<strong>the</strong>ir transmission <strong>of</strong> signals through <strong>the</strong> landscape.<br />
(Whipple and Tucker, 1999; Zaprowski et al, 2005;<br />
Wobus et al, 2010).<br />
In rapidly eroding landscapes, a knickpoint—a steep<br />
reach bounded on both sides by relatively shallower<br />
reaches—is a physical indicator <strong>of</strong> transient channel<br />
response to climate and tectonic signals. In postorogenic<br />
landscapes, knickpoints may also reflect<br />
rock strength or slow, complex response to external<br />
forcing. Explanations for <strong>the</strong> existence <strong>of</strong> knickpoints<br />
are varied, but consensus holds that <strong>the</strong>se features<br />
are transitory, migrating in a front from an initial<br />
source at <strong>the</strong> head or foot <strong>of</strong> <strong>the</strong> channel (Crosby and<br />
Whipple, 2006; Wobus et al, 2010). If <strong>the</strong> knickpoint<br />
is migrating upstream, <strong>the</strong> topography below will<br />
have undergone greater adjustment than <strong>the</strong> topography<br />
upstream <strong>of</strong> <strong>the</strong> knickpoint, and vice versa if<br />
<strong>the</strong> knickpoint is migrating downstream (Crosby and<br />
Whipple, 2006; Wobus et al, 2010). Studying <strong>the</strong><br />
location and characteristics <strong>of</strong> knickpoints can help us<br />
understand <strong>the</strong> dynamics <strong>of</strong> topographic response to<br />
different forcings (Crosby and Whipple, 2006; Wobus<br />
et al, 2010).<br />
Most knickpoint research has focused on regions<br />
with high uplift rates, weak rock, and rapid landscape<br />
evolution. In contrast, <strong>the</strong> Front Range in Colorado—<br />
in <strong>the</strong> interior <strong>of</strong> <strong>the</strong> North American continent—is<br />
a region with low uplift and low precipitation, relatively<br />
strong rock, and thus comparatively low rates<br />
<strong>of</strong> incision.<br />
Studying streams in <strong>the</strong> Front Range can provide<br />
insight into <strong>the</strong> behavior <strong>of</strong> a slowly evolving environment.<br />
Pr<strong>of</strong>iles <strong>of</strong> steady-state channels and hillslopes<br />
have different concavity. Steady-state channels are<br />
concave up, and steady-state hillslopes are concave<br />
down in <strong>the</strong> upper reach and concave up near <strong>the</strong> channel<br />
(Anderson, 2008). These shapes are disturbed by <strong>the</strong><br />
introduction <strong>of</strong> a knickpoint to <strong>the</strong> system, or prevented<br />
from occurring by a permanent knickpoint. As knickpoints<br />
travel through <strong>the</strong> river system, <strong>the</strong> long pr<strong>of</strong>ile<br />
<strong>of</strong> <strong>the</strong> river becomes locally convex, with shallow<br />
reaches bounding a steep section. Adjacent hillslopes<br />
respond to <strong>the</strong> resulting changes in boundary conditions,<br />
<strong>of</strong>ten exhibiting greater concavity and roughness<br />
as a result <strong>of</strong> increased incision. Examining <strong>the</strong>se features<br />
around knickpoints allows us to characterize <strong>the</strong><br />
ways a landscape moves back towards equilibrium.<br />
The thrust <strong>of</strong> this project is to identify knickpoints,<br />
characterize <strong>the</strong> morphology <strong>of</strong> <strong>the</strong> channels that include<br />
<strong>the</strong> knickpoints, and describe <strong>the</strong> nature <strong>of</strong> <strong>the</strong><br />
adjacent hillslopes. After this identification, I hope to<br />
draw conclusions by contrasting <strong>the</strong> basin area above<br />
<strong>the</strong> knickpoint with <strong>the</strong> area below, and comparing <strong>the</strong><br />
area within <strong>the</strong> knickpoint with <strong>the</strong> area that bounds it.<br />
RESEARCH AREA<br />
I conducted my field research in <strong>the</strong> Middle Boulder<br />
Creek watershed (Fig. 1). I focused on Middle Boulder<br />
Creek and its tributaries, particularly two small channels:<br />
Gordon Gulch and Betasso Gulch.<br />
The Colorado Front Range, which formed and evolved<br />
during <strong>the</strong> Laramide orogeny from 65 to 40 Ma, extends<br />
westward from <strong>the</strong> piedmont at Boulder to <strong>the</strong><br />
Continental Divide. Initially formed by rapid uplift,<br />
since 40 Ma <strong>the</strong> Front Range has been tectonically inac-<br />
106
24th Annual Keck Symposium: 2011 Union College, Schenectady, NY<br />
Figure 1. Map showing <strong>the</strong> Boulder Creek Catchment,<br />
outlined in black. Betasso Gulch and Gordon Gulch are<br />
shown in orange. Middle Boulder Creek is traced to its<br />
headwaters at <strong>the</strong> Continental Divide, and to its outlet in<br />
<strong>the</strong> South Platte River on <strong>the</strong> plains.<br />
tive, with isostatic response driving rock uplift in <strong>the</strong><br />
region (Kellogg et al, 2008). Melt from small glaciers<br />
and seasonal snowpack in upland areas runs down<br />
numerous tributaries <strong>of</strong> Middle Boulder Creek and<br />
eventually empties onto <strong>the</strong> Great Plains. The relatively<br />
high relief alpine and subalpine zone has been<br />
sculpted by glacial and periglacial activity. Below<br />
~2500 m, a rolling upland landscape is deeply incised<br />
by narrow canyons that extend up from <strong>the</strong> piedmont.<br />
Due to high evapotranspiration rates during <strong>the</strong> summer<br />
months, many small drainages east <strong>of</strong> <strong>the</strong> glacial<br />
limit are ephemeral. The vegetated low-relief surface<br />
has not been affected by glaciation, and away from<br />
<strong>the</strong> deep canyons, locally thick wea<strong>the</strong>red deposits<br />
overlie fresh bedrock (Birkeland et al 2003). Hillslope<br />
and channel processes govern landscape evolution in<br />
this montane zone, which includes <strong>the</strong> Gordon Gulch<br />
and Betasso Gulch catchments.<br />
Climate in <strong>the</strong> Front Range is largely dependent on<br />
elevation and proximity to <strong>the</strong> continental divide,<br />
which runs North-South roughly 30 km west <strong>of</strong> Boulder.<br />
Annual precipitation is highest near <strong>the</strong> divide<br />
and low on <strong>the</strong> plains (PRISM Climate Group). The<br />
orographic effect is strong in <strong>the</strong> winter, with significant<br />
snow accumulation near <strong>the</strong> continental divide<br />
and lighter precipitation down low. Frequent, low<br />
elevation thunderstorms during <strong>the</strong> summer months<br />
mitigate this orographic effect.<br />
METHODS<br />
I surveyed channel longitudinal pr<strong>of</strong>iles in <strong>the</strong> field,<br />
while measuring numerous river parameters, hillslope<br />
character, and rock strength, and used DEMs based on<br />
Lidar (Gordon Gulch and Betasso Gulch) and USGS<br />
maps (Boulder Creek) to characterize channels and<br />
hillslopes.<br />
At <strong>the</strong> Gordon Gulch and Betasso catchments, we<br />
surveyed <strong>the</strong> longitudinal pr<strong>of</strong>ile <strong>of</strong> <strong>the</strong> stream, using<br />
a tripod-mounted Tru-Pulse 360 Laser Rangefinder.<br />
All measurements were parallel to <strong>the</strong> stream channel.<br />
I recorded <strong>the</strong> vertical distance, horizontal distance,<br />
and azimuth for each section <strong>of</strong> stream. I also measured<br />
channel width and estimated bankfull width<br />
using a tape measure I estimated d 50 and dmax grain<br />
sizes, and boulder percentage for <strong>the</strong> reach. Prominent<br />
tributaries <strong>of</strong> <strong>the</strong> main stream were also surveyed using<br />
<strong>the</strong> same process.<br />
We also conducted a series <strong>of</strong> cross-valley surveys using<br />
<strong>the</strong> channel as a base and measuring perpendicular<br />
to <strong>the</strong> channel, using a GPS point for location. At<br />
each survey point on <strong>the</strong> hillslope, I recorded bedrock<br />
and boulder percentages and noted bedrock lithology,<br />
<strong>the</strong> presence <strong>of</strong> vegetation, and local slope characteristics.<br />
We completed five cross-valley pr<strong>of</strong>iles for<br />
each main stream, and three for each <strong>of</strong> <strong>the</strong> tributaries.<br />
As we surveyed we measured rock strength with<br />
a Schmidt hammer on outcrops in <strong>the</strong> channel and at<br />
selected outcrops on <strong>the</strong> hillslopes. A Schmidt hammer<br />
measures rock strength by applying a specific<br />
amount <strong>of</strong> force to a rock with a spring loaded piston,<br />
<strong>the</strong>n recording <strong>the</strong> force <strong>of</strong> <strong>the</strong> piston as it rebounds<br />
Figure 2. Schmidt hammer measurements being taken on<br />
Betasso Gulch hillslopes.<br />
107
24th Annual Keck Symposium: 2011 Union College, Schenectady, NY<br />
<strong>of</strong>f <strong>the</strong> outcrop.<br />
The scale <strong>of</strong> Middle Boulder Creek and North Boulder<br />
Creek precluded <strong>the</strong> comprehensive field surveys<br />
that we carried out in Gordon Gulch and Betasso<br />
Gulch. The level <strong>of</strong> detail provided by Digital Elevation<br />
Models (DEMs) is sufficient for analysis <strong>of</strong> <strong>the</strong>se<br />
channels. To supplement <strong>the</strong> digital analysis we surveyed<br />
individual, representative reaches distributed<br />
along <strong>the</strong> Boulder Creek channels. Beginning by taking<br />
a GPS point in <strong>the</strong> middle <strong>of</strong> a reach, we surveyed<br />
slope, width, and reach length with <strong>the</strong> rangefinder.<br />
We described each reach and photographed <strong>the</strong> channel<br />
and <strong>the</strong> surrounding hillslopes. We measured river<br />
and valley width with <strong>the</strong> rangefinder, estimated d 50<br />
where cobbles could be seen through <strong>the</strong> water, and<br />
estimated dmax by identifying <strong>the</strong> largest boulder in<br />
<strong>the</strong> reach. We noted <strong>the</strong> presence <strong>of</strong> hillslope features<br />
such as rock slides and falls, tributary entrances, and<br />
cliffs. We also applied <strong>the</strong> same Schmidt hammer process<br />
described above to each section that we recorded,<br />
measuring bedrock if it was present at <strong>the</strong> water level<br />
or stationary boulders where bedrock was not exposed<br />
in <strong>the</strong> channel. In sections with cliffs adjacent<br />
to <strong>the</strong> channel, we searched for evidence <strong>of</strong> sculpting,<br />
potholes, or polish on <strong>the</strong> bedrock, and used <strong>the</strong><br />
rangefinder to measure <strong>the</strong> vertical distance above <strong>the</strong><br />
current channel.<br />
I supplemented my field data with extensive remote<br />
sensing analysis. Using 1m-resolution LIDAR data<br />
for Gordon Gulch and Betasso Gulch, and 10m resolution<br />
DEMs for Middle Boulder Creek, I calculated<br />
additional cross-valley pr<strong>of</strong>iles and basin slopes. I<br />
corroborated my surveyed long pr<strong>of</strong>iles with <strong>the</strong> LI-<br />
DAR and DEMs, and calculated power relationships<br />
with drainage area, downstream distance, and slope.<br />
To ease comparison between basins <strong>of</strong> different magnitudes,<br />
I normalized longitudinal pr<strong>of</strong>iles and mean<br />
basin slope pr<strong>of</strong>iles, setting <strong>the</strong> lowest distance and<br />
elevation values equal to zero, and <strong>the</strong> highest values<br />
equal to one.<br />
RESULTS<br />
The project focuses on relationships between knickpoints,<br />
mean basin slope, channel slope, and rock<br />
strength, so I will focus on data in those areas.<br />
Distance from <strong>the</strong> headwaters is correlated with drain-<br />
Figure 3. A comparison <strong>of</strong> four basins, showing remarkable<br />
similarity in catchment shape despite a large disparity<br />
in magnitude.<br />
Figure 4. Normalized longitudinal pr<strong>of</strong>iles for <strong>the</strong> channels<br />
<strong>of</strong> Betasso Gulch, Gordon Gulch, and Middle Boulder<br />
Creek. The pr<strong>of</strong>iles are made by setting <strong>the</strong> minimum elevation<br />
and distance equal to 1 and <strong>the</strong> maximum elevation<br />
and distance equal to 0, <strong>the</strong>n adjusting each intermediate<br />
point by dividing its value by <strong>the</strong> maximum value.<br />
108
24th Annual Keck Symposium: 2011 Union College, Schenectady, NY<br />
Figure 5. Plots <strong>of</strong> channel slope and mean basin slope orthogonal<br />
to <strong>the</strong> channel. The x-axis represents <strong>the</strong> distance<br />
in kilometers from <strong>the</strong> headwaters. Mean basin slopes are<br />
calculated by averaging <strong>the</strong> slopes <strong>of</strong> hillslope transects.<br />
The black line on <strong>the</strong> channel slopes is a moving average.<br />
109<br />
age area in each <strong>of</strong> <strong>the</strong> catchments I worked in (Fig.<br />
3).<br />
I found knickpoints in each channel that I surveyed. A<br />
graph <strong>of</strong> normalized longitudinal pr<strong>of</strong>iles shows <strong>the</strong>se<br />
knickpoints on a standard scale (Fig. 4).<br />
The largest channel, Middle Boulder Creek, contains<br />
a prominent knickpoint located on a reach ~30.5-<br />
33 kilometers from <strong>the</strong> headwaters. This knickpoint<br />
has an average slope <strong>of</strong> 0.074, higher than <strong>the</strong> 0.042<br />
channel average. Basin slopes are steepest at <strong>the</strong><br />
knickpoint and just downstream (Fig. 5). There is a<br />
second minor knickpoint near <strong>the</strong> top <strong>of</strong> <strong>the</strong> reach we<br />
surveyed, though slopes in that reach are only locally<br />
as high as 0.04.<br />
In Gordon Gulch, two knickpoints—located ~2.5 and<br />
2.8 kilometers from <strong>the</strong> headwaters and separated by<br />
300 meters <strong>of</strong> relatively low channel slope— are distinguished<br />
by an average slope <strong>of</strong> 0.153. These slopes<br />
are considerably higher than <strong>the</strong> mean reach slope <strong>of</strong><br />
0.108, and more than twice <strong>the</strong> average slope <strong>of</strong> <strong>the</strong><br />
non-knickpoint reaches (0.067) (Fig. 5). Basin slope<br />
analysis shows that <strong>the</strong> average slope on <strong>the</strong> Gordon<br />
Gulch hillsides is higher normal to <strong>the</strong> knickpoints<br />
than o<strong>the</strong>r locations on <strong>the</strong> channel. Schmidt hammer<br />
rock strength values are highest on <strong>the</strong> hillslopes and<br />
in <strong>the</strong> channel at <strong>the</strong> knickpoints: mean values at <strong>the</strong><br />
knickpoints are 45-55, as opposed to mean values <strong>of</strong><br />
30-40 elsewhere in <strong>the</strong> basin.<br />
Betasso Gulch contains three small knickpoints (Fig.<br />
5). Each knickpoint in <strong>the</strong> channel is a short, steep<br />
outcrop covered by little or no sediment. Ranging<br />
in height from 1.5-3 meters, <strong>the</strong>se bedrock steps<br />
have generally higher Schmidt values than bedrock<br />
elsewhere in <strong>the</strong> channel. Hillslope rock strength is<br />
highest between <strong>the</strong> lowest and middle knickpoint. At<br />
<strong>the</strong> two lower knickpoints, Schmidt values range from<br />
40-50, higher than <strong>the</strong> values <strong>of</strong> 25-25 elsewhere. At<br />
<strong>the</strong> upper knickpoint, <strong>the</strong> values are between 25-35,<br />
in contrast to values <strong>of</strong> 0-15 that pervade in <strong>the</strong> upper<br />
zone <strong>of</strong> Betasso Gulch. This uppermost knickpoint is<br />
near <strong>the</strong> low margin <strong>of</strong> a saprolite and colluvial zone:<br />
<strong>the</strong>se disintegrated materials provide a thick cover for<br />
solid bedrock, which rarely crops out at <strong>the</strong> surface in<br />
<strong>the</strong> upper half <strong>of</strong> <strong>the</strong> basin. In this upper section, both
24th Annual Keck Symposium: 2011 Union College, Schenectady, NY<br />
channel slopes and hillslopes are shallowest.<br />
DISCUSSION<br />
Though knickpoints take different forms in <strong>the</strong> three<br />
study areas, <strong>the</strong>y mark <strong>the</strong> boundary between steadystate<br />
and adjusting landscapes in all catchments.<br />
Above <strong>the</strong> knickpoint, basin slopes adjust steadily to<br />
accommodate slow downcutting, and are shallower<br />
and smoo<strong>the</strong>r than <strong>the</strong> hillslopes below. The steeper<br />
channel slopes within a knickpoint focus higher<br />
stream power and greater incision rates on that reach<br />
in <strong>the</strong> channel. In <strong>the</strong> catchments in this study, <strong>the</strong> influence<br />
<strong>of</strong> a knickpoint on a longitudinal pr<strong>of</strong>ile scales<br />
with <strong>the</strong> size <strong>of</strong> <strong>the</strong> basin in which it is located (Fig.<br />
4). Middle Boulder Creek has <strong>the</strong> most prominent<br />
knickpoint: <strong>the</strong> zone <strong>of</strong> higher slopes is more than two<br />
kilometers long, and is almost twice as steep as <strong>the</strong><br />
channel average. The knickpoint significantly disrupts<br />
<strong>the</strong> concave-up shape expected for a steady-state<br />
channel. Gordon Gulch, with intermediate drainage<br />
size, includes two closely spaced knickpoints with a<br />
slope disparity comparable to that <strong>of</strong> Middle Boulder<br />
Creek. But <strong>the</strong>se knickpoints are less dominant<br />
features: <strong>the</strong>y appear as large lumps within a generally<br />
smooth pr<strong>of</strong>ile. The knickpoints in Betasso Gulch<br />
are less remarkable: <strong>the</strong>y appear as small steps in <strong>the</strong><br />
longitudinal pr<strong>of</strong>ile but do not affect <strong>the</strong> concavity <strong>of</strong><br />
<strong>the</strong> channel. The bedrock steps that mark knickpoint<br />
location are dramatic in <strong>the</strong> field but are less convincing<br />
when plotted (Fig. 4).<br />
Rapid channel lowering in <strong>the</strong> knickpoint increases<br />
adjacent basin slope: as <strong>the</strong> knickpoint moves up <strong>the</strong><br />
channel it leaves higher basin slopes in its wake. The<br />
rate at which hillslopes adjust to <strong>the</strong> new boundary<br />
conditions depends on <strong>the</strong> mobility <strong>of</strong> constituent<br />
material and <strong>the</strong> capacity <strong>of</strong> <strong>the</strong> channel to move <strong>the</strong><br />
material downstream. Stronger rock resists wea<strong>the</strong>ring<br />
and preserves steeper hillslopes, whereas weak<br />
rock and colluvium are susceptible to wea<strong>the</strong>ring and<br />
mass movements. A stream with high competence can<br />
transport wea<strong>the</strong>red debris and allow fur<strong>the</strong>r wea<strong>the</strong>ring<br />
to occur, but for streams with low stream power<br />
hillslopes only slowly return to a steady state.<br />
The most compelling relationship <strong>of</strong> hillslope to<br />
channel slope is in Middle Boulder Creek (Fig. 5).<br />
110<br />
High rock strength has produced a lag in slope response<br />
to <strong>the</strong> migration <strong>of</strong> <strong>the</strong> knickpoint. Steep basin<br />
slopes persist below <strong>the</strong> main knickpoint, despite high<br />
stream power during snowmelt and summer thunderstorms.<br />
The hillslopes above <strong>the</strong> knickpoint are<br />
smoo<strong>the</strong>r and shallower: <strong>the</strong>y are in relative steady<br />
state, having adjusted to <strong>the</strong> passage <strong>of</strong> a previous,<br />
smaller knickpoint through <strong>the</strong> system.<br />
Hillslopes in Gordon Gulch have responded similarly<br />
to <strong>the</strong> presence <strong>of</strong> knickpoints. A trend <strong>of</strong> steepening<br />
hillslopes with distance from <strong>the</strong> headwaters is broken<br />
only by <strong>the</strong> shallow section between knickpoints<br />
(Fig. 5). These low slopes can be explained by weak<br />
rock—reflected in low Schmidt hammer measurements—that<br />
underlies <strong>the</strong> area between knickpoints.<br />
Additionally, <strong>the</strong> proximity <strong>of</strong> <strong>the</strong> lower knickpoint—<br />
with its associated higher stream power—may help<br />
to efficiently transport material and accelerate <strong>the</strong><br />
hillslope adjustment to <strong>the</strong> new baselevel. Steep<br />
hillslopes below <strong>the</strong> lower knickpoint suggest that <strong>the</strong><br />
landscape <strong>the</strong>re continues to adjust.<br />
Betasso Gulch displays <strong>the</strong> most dramatic basin slope<br />
increase with distance down <strong>the</strong> channel (Fig. 5).<br />
The shallow sloped colluvial and saprolite hillslopes<br />
above <strong>the</strong> knickpoints stand in stark contrast to <strong>the</strong><br />
steep, outcrop-dominated hillslopes that flank <strong>the</strong><br />
knickpoints and below. Several factors can account<br />
for <strong>the</strong> steep lower slopes. The relative strength <strong>of</strong> <strong>the</strong><br />
rock below <strong>the</strong> knickpoints may have prevented <strong>the</strong><br />
lower hillslopes from readjusting. Even if material<br />
was available to move, <strong>the</strong> tiny channel in Betasso<br />
Gulch has a low capacity for transport, and hillslope<br />
evolution would be slowed by that limiting factor.<br />
These characteristics <strong>of</strong> Betasso Gulch have prevented<br />
its basin from adjusting to <strong>the</strong> passage <strong>of</strong> several<br />
knickpoints.<br />
CONCLUSION<br />
Catchments <strong>of</strong> dramatically different sizes can be<br />
compared effectively: general knickpoint mechanics<br />
are similar in each basin we examined. Knickpoints<br />
mark <strong>the</strong> boundary between hillslopes in steady<br />
state and hillslopes that are struggling to adjust to<br />
new boundary conditions. The disruption <strong>of</strong> previous<br />
steady state is driven by increased incision rates<br />
associated with knickpoints. Hillslopes at and below
24th Annual Keck Symposium: 2011 Union College, Schenectady, NY<br />
knickpoints become steeper with a rapid baselevel<br />
fall. Rock strength, stream power, and continued<br />
disruption limit <strong>the</strong> ability <strong>of</strong> a hillslope to approach<br />
a new steady state after a knickpoint passes. In each<br />
study area, we found steep, rough hillslopes downstream<br />
from knickpoints, as opposed to comparatively<br />
smoo<strong>the</strong>r and flatter hillslopes above. We found high<br />
rock strength within and below knickpoints, perhaps<br />
<strong>the</strong> result <strong>of</strong> recent exposure with enhanced denudation<br />
following <strong>the</strong> knickpoint-driven baselevel lowering.<br />
The parameters involved in channel and hillslope<br />
evolution—channel slope, rock strength, hillslope<br />
shape, sediment transport, and wea<strong>the</strong>ring—all are<br />
inherently linked in <strong>the</strong>se systems to <strong>the</strong> presence <strong>of</strong><br />
knickpoints. The catchments in my study area have<br />
not fully adjusted to <strong>the</strong> passage <strong>of</strong> knickpoints. This<br />
lack <strong>of</strong> response suggests a slowly evolving landscape<br />
dominated by strong rock, subdued wea<strong>the</strong>ring, and<br />
low stream power.<br />
ACKNOWLEDGMENTS<br />
I would like to thank my advisor Dr. David P. Dethier<br />
for his guidance, advice, and helpful questions. Thank<br />
you also to Dr. William Ouimet for his assistance<br />
in <strong>the</strong> field and as an advisor through <strong>the</strong> process. I<br />
would like to thank Keith Kantack for his instrumental<br />
work assisting me in <strong>the</strong> field. I would also like to<br />
thank <strong>the</strong> Williams College Geosciences Department,<br />
<strong>the</strong> National Science Foundation, and <strong>the</strong> KECK Geology<br />
Consortium.<br />
REFERENCES<br />
Anderson, R. S. (2008), The Little Book <strong>of</strong> Geomorphology:<br />
Exercising <strong>the</strong> Principle <strong>of</strong> Conservation.<br />
Anderson, R. S. and Anderson, S. P. (2010), Geomorphology:<br />
The Mechanics and Chemistry <strong>of</strong> Landscapes<br />
(Cambridge University Press) textbook,<br />
640 pp., published June 2010.<br />
Birkeland, P.W., Shroba, R.R., Burns, S.F., Price, A.B.<br />
and Tonkin, P.J. (2003), Integrating soils and<br />
geomorphology in mountains - an example from<br />
<strong>the</strong> Front Range <strong>of</strong> Colorado, Geomorphology<br />
55, p. 329-344.<br />
111<br />
Crosby BT, Whipple KX. 2006. Knickpoint initiation<br />
and distribution within fluvial networks: 236 waterfalls<br />
in <strong>the</strong> Waipaoa River, North Island, New<br />
Zealand. Geomorphology 82: 16–38.<br />
Kellogg, K.S., Shroba, R.R., Bryant, Bruce, and<br />
Premo, W.R. (2008), Geologic map <strong>of</strong> <strong>the</strong> Denver<br />
West 30’ x 60’ quadrangle, north-central<br />
Colorado: U.S. Geological Survey Scientific<br />
Investigations Map 3000, scale 1:100,000, 48-p.<br />
pamphlet.<br />
PRISM Climate Group, Oregon State University,<br />
http://prism.oregonstate.edu, created 2011.<br />
Whipple, K. X., and G. E. Tucker (1999), Dynamics<br />
<strong>of</strong> <strong>the</strong> stream power river incision model: Implications<br />
for height limits <strong>of</strong> mountain ranges,<br />
landscape response time scales, and research<br />
needs, J. Geophys. Res., 104, 17,661–17,674.<br />
Wobus, C. W., G. E. Tucker, and R. S. Anderson<br />
(2010), Does climate change create distinctive<br />
patterns <strong>of</strong> landscape incision?, J. Geophys. Res.,<br />
115, F04008, doi:10.1029/2009JF001562.<br />
Zachos, J., Pagani, M., Sloan, L., Thomas, E., and<br />
Billups, K. (2001), Trends, rhythms, and aberrations<br />
in global climate 65 Ma to present. Science,<br />
292, p. 686-693.<br />
Zaprowski, B. J., et al. (2005), Climatic influences on<br />
pr<strong>of</strong>ile concavity and river incision, J. Geophys.<br />
Res., 110, F03004, doi:10.1029/2004JF000138.<br />
Zhang, P., P. Molnar, and W. R. Downs (2001), Increased<br />
sedimentation rates and grain sizes 2 – 4<br />
Myr ago due to <strong>the</strong> influence <strong>of</strong> climate change<br />
on erosion rates, Nature, 410, 891–897.
24th Annual Keck Symposium: 2011 Union College, Schenectady, NY<br />
RECONSTRUCTING THE PINEDALE GLACIATION, GREEN<br />
LAKES VALLEY, COLORADO<br />
KEITH M. KANTACK, Williams College<br />
Research Advisor: David P. Dethier<br />
INTRODUCTION<br />
During <strong>the</strong> Pinedale Glaciation (~32-15 kya) alpine<br />
glaciers covered much Colorado’s Front Range<br />
(Madole 1998). While <strong>the</strong>se glaciers are small and<br />
scarce today, <strong>the</strong>ir work still dominates <strong>the</strong> landscape.<br />
As powerful erosive agents, glaciers remove massive<br />
volumes <strong>of</strong> material from <strong>the</strong>ir beds and leave<br />
equally impressive volumes <strong>of</strong> sediment in <strong>the</strong> form<br />
<strong>of</strong> moraines and outwash plains. By using field and<br />
modeling techniques based on glacial evidence, we<br />
can determine glacier size and ice flow direction, as<br />
well as how much sediment <strong>the</strong>y deposited, which<br />
ultimately helps to constrain how quickly glaciers cut<br />
<strong>the</strong>ir cirques and valleys.<br />
While <strong>the</strong> scoured bedrock surface <strong>of</strong> <strong>the</strong> deglaciated<br />
alpine zone is not typical <strong>of</strong> <strong>the</strong> critical zone, glaciers<br />
are sediment factories. Much <strong>of</strong> <strong>the</strong> sediment that<br />
ends up covering <strong>the</strong> lower reaches <strong>of</strong> alpine and<br />
subalpine basins and flanking channels downstream is<br />
generated by glacial erosion <strong>of</strong> bedrock. In this way,<br />
glaciers are not just icy blocks high in <strong>the</strong> mountains,<br />
but significant players in <strong>the</strong> all important critical<br />
zone.<br />
In this paper, I use field and LIDAR evidence to model<br />
<strong>the</strong> extent <strong>of</strong> <strong>the</strong> Pinedale glaciation in <strong>the</strong> GLV, as<br />
well as its contributions to <strong>the</strong> critical zone.<br />
SETTING<br />
The Green Lakes Valley (GLV) <strong>of</strong> <strong>the</strong> Colorado Front<br />
Range is about 13 kilometers northwest <strong>of</strong> <strong>the</strong> town <strong>of</strong><br />
Nederland in <strong>the</strong> headwater area <strong>of</strong> <strong>the</strong> North Branch<br />
<strong>of</strong> Boulder Creek (Fig. 1). The valley floor runs<br />
extends from 3300 to 3900 meters in elevation and<br />
is walled by 3900 to 4100 meter peaks that form <strong>the</strong><br />
continental divide. The valley, which is part <strong>of</strong> <strong>the</strong><br />
City <strong>of</strong> Boulder watershed, is protected and access is<br />
limited to researchers working on specific topics.<br />
Figure 1. Colorado, Boulder County, and a panoramic<br />
view <strong>of</strong> <strong>the</strong> GLV, looking west toward <strong>the</strong> continental divide<br />
over <strong>the</strong> step between Green Lake 3 (left) and 4.<br />
During <strong>the</strong> peak <strong>of</strong> <strong>the</strong> Pinedale glaciation, glaciers<br />
flowed out <strong>of</strong> small cirques in <strong>the</strong> Front Range,<br />
through valleys, and joined forces, reaching lengths<br />
<strong>of</strong> 15 kilometers and elevations as low as 2500 meters<br />
(Madole et al. 1998). The GLV bears many marks<br />
<strong>of</strong> glaciation. In favorable sites, <strong>the</strong> bedrock holds<br />
polish and striations, and moraines are preserved in<br />
downvalley locations. On <strong>the</strong> valley floor, bedrock is<br />
ubiquitously smoo<strong>the</strong>d, and bears evidence <strong>of</strong> plucking.<br />
In places, <strong>the</strong> valley walls are oversteepened,<br />
reflecting undercutting by debris sheared along by <strong>the</strong><br />
moving ice. Post-glacial talus fields cover <strong>the</strong> base <strong>of</strong><br />
<strong>the</strong>se cliffs and cover large tracts <strong>of</strong> <strong>the</strong> lower walls.<br />
The valley is U-shaped, with six lakes strung along<br />
118
24th Annual Keck Symposium: 2011 Union College, Schenectady, NY<br />
<strong>the</strong> floor. The lakes fill a series <strong>of</strong> topographic steps,<br />
<strong>the</strong> most notable dropping 125 meters from Green<br />
Lake 4 to Green Lake 3.<br />
Just south on <strong>the</strong> continental divide is <strong>the</strong> Arapaho<br />
Valley, which I did not map. Like <strong>the</strong> GLV, <strong>the</strong> floor<br />
<strong>of</strong> <strong>the</strong> Arapaho Valley is dotted with six lakes. But<br />
<strong>the</strong> valley is roughly double <strong>the</strong> area <strong>of</strong> <strong>the</strong> GLV, and<br />
ra<strong>the</strong>r than <strong>the</strong> many-stepped pr<strong>of</strong>ile <strong>of</strong> <strong>the</strong> GLV, <strong>the</strong><br />
Arapahoe Valley pr<strong>of</strong>ile has one large (500 meter)<br />
step 1.5 kilometers from <strong>the</strong> divide.<br />
Fur<strong>the</strong>r south are <strong>the</strong> Rainbow and Horseshoe cirques.<br />
These are small features, with areas < 0.5km 2 . However,<br />
<strong>the</strong>re are well-defined moraine complexes below<br />
both cirques. (Gable and Madole, 1976)<br />
The bedrock in underlying <strong>the</strong> study area is comprised<br />
<strong>of</strong> 4 main rock types: (1) cordierite and magnetitebearing<br />
sillimanite-biotite gneiss commonly referred<br />
to as Metasediments; (2) <strong>the</strong> Boulder Creek Granodiorite;<br />
(3) Silver Plume Quartz Monzonite; and (4)<br />
Monzonite .<br />
Climate in this area is cool and continental with<br />
strong seasonal variation and relatively low precipitation.<br />
Presently, mean <strong>annual</strong> temperature is -3.5ºC<br />
and mean <strong>annual</strong> precipitation is 763 mm at <strong>the</strong> Niwot<br />
Ridge D1 monitoring station, which sits atop <strong>the</strong><br />
south-facing slope <strong>of</strong> <strong>the</strong> GLV. Annual snow accumulation<br />
reaches 20 meters at <strong>the</strong> head <strong>of</strong> <strong>the</strong> valley<br />
(Caine 1995).<br />
METHODS<br />
FIELD METHODS<br />
In order to reconstruct <strong>the</strong> extent <strong>of</strong> Pinedale ice in<br />
<strong>the</strong> GLV, I examined bedrock surfaces on <strong>the</strong> floor<br />
and walls <strong>of</strong> <strong>the</strong> valley for evidence <strong>of</strong> recent glacial<br />
smoothing: primarily polish and striations. I traversed<br />
<strong>the</strong> entire valley with a Garmin Rino120 GPS and<br />
mapped evidence <strong>of</strong> glaciation and local flow directions.<br />
Trimline zones were identified on valley walls<br />
as <strong>the</strong> point where polish ceased, and rough, unglaciated<br />
outcrop began. On <strong>the</strong> south-facing slope <strong>of</strong> <strong>the</strong><br />
valley, I also mapped polished boulders.<br />
MODELING AND LABORATORY METHODS<br />
Reconstruction <strong>of</strong> ice is an important aspect <strong>of</strong> this<br />
119<br />
study, as modern glaciers in <strong>the</strong> Front Range are mere<br />
shadows <strong>of</strong> <strong>the</strong>ir former extent. Modeling an ice<br />
surface is essentially determining ice thickness at any<br />
point on <strong>the</strong> glacier and begins with concepts <strong>of</strong> ice<br />
deformation and resulting motion. I used a modeling<br />
program developed by Hulton and Benn (2010) that<br />
incorporates cross-valley pr<strong>of</strong>ile or “shape factor”,<br />
bed pr<strong>of</strong>ile, yield stress, and field observed trimline<br />
elevations.<br />
All GIS studies were done using ArcMap 10. Using<br />
<strong>the</strong> trimline points collected in <strong>the</strong> field, I made a<br />
complete trimline map <strong>of</strong> <strong>the</strong> glacier in <strong>the</strong> GLV. The<br />
Arapaho Valley glacier used Madole’s upper Platte<br />
River glaciation modeling (citation). The greatest<br />
extent <strong>of</strong> <strong>the</strong> glacier was modeled using an equilibrium<br />
line altitude (ELA- <strong>the</strong> altitude at which ablation<br />
equals accumulation) <strong>of</strong> 3350 meters (Ward et al,<br />
2009) As I moved ELA higher, I maintained an accumulation<br />
area ratio (AAR) <strong>of</strong> 0.65, which is to say <strong>the</strong><br />
area above <strong>the</strong> ELA comprised 65% <strong>of</strong> <strong>the</strong> total ice<br />
surface. Below <strong>the</strong> ELA, I modeled <strong>the</strong> glacier with a<br />
high centerline surface topography, where <strong>the</strong> surface<br />
<strong>of</strong> <strong>the</strong> glacier bowed upwards as much as 20 meters.<br />
At <strong>the</strong> ELA, <strong>the</strong> surface was flat. Above <strong>the</strong> ELA, I<br />
modeled <strong>the</strong> glacier in <strong>the</strong> accumulation zone using<br />
low centerline topography where <strong>the</strong> surface bowed<br />
down as much as 20 meters.<br />
To map glacial deposits in <strong>the</strong> Pinedale ablation zone,<br />
I used LIDAR data, which shows moraines in high<br />
detail. Light Detection and Ranging (LIDAR) data<br />
produces a digital elevation model (DEM) with a<br />
unique elevation value for every pixel. The LIDAR<br />
layer I used was flown with snow-<strong>of</strong>f in August, 2010<br />
by <strong>the</strong> National Center for Airborne Laser Mapping,<br />
and is comprised <strong>of</strong> 1-meter pixels. To map <strong>the</strong><br />
moraines, I used a hillshade layer, which accentuates<br />
terrain relief.<br />
To reconstruct <strong>the</strong> Pinedale glacier surface, I used ArcMap’s<br />
Kriging tool, which produced a layer based on<br />
<strong>the</strong> elevation <strong>of</strong> <strong>the</strong> trimline and interpolated surface<br />
points. With <strong>the</strong> layer produced by Kriging, I could<br />
reconstruct elevation contours for <strong>the</strong> Pinedale ice, as<br />
well as calculate <strong>the</strong> volume <strong>of</strong> ice held in <strong>the</strong> valley.<br />
The first step for calculating moraine volume involved<br />
mapping <strong>the</strong> moraine belts using DEMs
24th Annual Keck Symposium: 2011 Union College, Schenectady, NY<br />
derived from <strong>the</strong> LIDAR data. Within <strong>the</strong> belts, I<br />
created a new elevation layer (again by Kriging) with<br />
a 1 m 2 pixel based on <strong>the</strong> lowest points around <strong>the</strong><br />
moraines, including kettles within <strong>the</strong> moraine belt.<br />
The points were selected so as to produce a layer that<br />
would approximate <strong>the</strong> surface elevation prior to <strong>the</strong><br />
late-glacial maximum <strong>of</strong> <strong>the</strong> Pinedale glaciation. This<br />
layer was subtracted from <strong>the</strong> LIDAR values, which<br />
(because it is a 1 m 2 pixel) yields moraine volume.<br />
Because making an elevation layer for an entire<br />
moraine belt introduces large local error, each moraine<br />
belt was divided into 4 to 7 sections. In some<br />
places, this interpolated low layer was higher than <strong>the</strong><br />
LIDAR, which produced negative volume values for<br />
those pixels. These pixels were omitted from calculations,<br />
but <strong>the</strong>ir presence indicates that this volume<br />
is a lower limit. An upper limit was established by<br />
lowering <strong>the</strong> interpolated layer 4 meters, to ensure<br />
it includes <strong>the</strong> entire moraine belt. Compared to <strong>the</strong><br />
lower limit, this equates to adding 4 m 3 to each pixel<br />
value.<br />
Erosion rate E was calculated using <strong>the</strong> following<br />
equation.<br />
Where MV is calculated moraine volume, A is area <strong>of</strong><br />
<strong>the</strong> cirque or catchment in which <strong>the</strong> glacial sediment<br />
originated, and t is time. In this case, time was 6000<br />
years, reflecting <strong>the</strong> duration <strong>of</strong> <strong>the</strong> Pinedale glacial<br />
maximum, which lasted from ~21 to 15 kya (Ward et<br />
al, 2009).<br />
RESULTS<br />
Pinedale ice reached its maximum extent at 21 kya,<br />
when <strong>the</strong> ELA stood at 3350 m in <strong>the</strong> Green Lakes<br />
area <strong>of</strong> <strong>the</strong> Front Range (Ward, 2009). At this maximum,<br />
<strong>the</strong> glacier flowing out <strong>of</strong> <strong>the</strong> Green Lakes and<br />
Arapaho Valleys covered an area <strong>of</strong> 22.3 km 2 with a<br />
volume 1.86 km 3 . By 16 kya, <strong>the</strong> ELA had risen to<br />
3650 m (Ward, 2009). With this retreat, <strong>the</strong> Green<br />
Lakes and Arapaho glaciers separated. Because I<br />
did not map <strong>the</strong> Arapaho valley in <strong>the</strong> field, I did not<br />
model its glacier at this ELA, where a more intimate<br />
knowledge <strong>of</strong> <strong>the</strong> area is required. The GLV model,<br />
however, shows a reduction in glaciated area to 2.43<br />
km 2 and volume to 0.23 km 3 . (Table 1)<br />
120<br />
Figure 2. (A) Map showing extent and thickness <strong>of</strong> Pinedale<br />
ice in Green Lakes and Arapaho Valley. Blue is modeled<br />
with an ELA <strong>of</strong> 3350 m, or <strong>the</strong> Pinedale glacial maximum.<br />
Orange is modeled with an ELA <strong>of</strong> 3650 m. Darker<br />
colors indicate greater ice thickness. 10m contours represent<br />
surface elevation <strong>of</strong> <strong>the</strong> ELA 3350 m glacier. ELAs<br />
are shown in green. Black arrows are ice flow direction<br />
indicated by striation measurements. (B) Pr<strong>of</strong>ile <strong>of</strong> glacier<br />
at ELA 3350 m (blue from “A”). Red (Arapaho) and blue<br />
(Green Lakes) are pr<strong>of</strong>iles above <strong>the</strong> confluence.<br />
Ice Thickness (m)<br />
Date ELA (m) Ice Area (km 2 ) Average Maximum Ice Volume (km 3 )<br />
21kya 3350 22.27 83.32 239.58 1.85<br />
16kya 3650 2.43 93.90 207.25 0.23<br />
Table 1. ELA and size statistics for <strong>the</strong> glaciers occupying<br />
<strong>the</strong> Green Lakes and Arapaho Valleys. Note that for 16<br />
kya, only <strong>the</strong> glacier in GLV is considered.
24th Annual Keck Symposium: 2011 Union College, Schenectady, NY<br />
Area (km 2 ) Moraine volume (10 6 m 3 ) Lowering rate, in mm yr -1<br />
Cirque Catchment Moraine Minimum a Maximum b Ca d Min Ca d Max C c Min C c Max<br />
Green Lakes<br />
and Arapahoe 11.19 22.50 5.65 67.06 89.57 0.50 0.66 1.00 1.33<br />
Rainbow Lakes 0.45 0.91 0.71 0.85 3.67 0.15 0.67 0.31 1.37<br />
Horseshoe 0.30 1.45 0.77 2.20 5.28 0.25 0.61 1.23 2.98<br />
a<br />
Volume calculated using subtraction <strong>of</strong> a modeled elevation layer from lidar.<br />
b<br />
Volume with an additional 4 meters <strong>of</strong> sediment lying beneath <strong>the</strong> entire moraine<br />
complex.<br />
c<br />
Lowering rate if morainal material came only from within <strong>the</strong> cirque<br />
d<br />
Lowering rate if morainal material includes sediment from <strong>the</strong> catchment above <strong>the</strong> moraines.<br />
Table 2. Calculated moraine volume and lowering rate<br />
Figure 3. Green Lakes and Arapahoe Valleys and Rainbow and Horseshoe Cirques with associated moraine complexes.<br />
Inset shows example <strong>of</strong> hillshade view <strong>of</strong> moraines. Darker colors in <strong>the</strong> moraines reflect thicker morainal deposits. The<br />
deepest parts <strong>of</strong> <strong>the</strong> valleys and cirques are indicated by darker shades.<br />
121
24th Annual Keck Symposium: 2011 Union College, Schenectady, NY<br />
I calculated volumes for <strong>the</strong> moraine belts left by <strong>the</strong><br />
Pinedale glaciers that flowed from <strong>the</strong> Arapaho and<br />
GLVs, along with <strong>the</strong> Rainbow Lakes and Horseshoe<br />
Cirques.<br />
The Arapaho/Green Lakes glacier deposited between<br />
67.06x10 6 and 89.57x10 6 m 3 <strong>of</strong> debris. The Horseshoe<br />
glacier deposited between 2.20 and 5.28x10 6<br />
m3. The Rainbow Lakes moraines occupied between<br />
8.5x10 5 and 3.67x10 6 m 3 . (Table 2)<br />
Erosion or lowering rates during <strong>the</strong> late Pinedale<br />
maximum were calculated for each glacier based on<br />
moraine volumes. These values are likely minima,<br />
since <strong>the</strong> glaciers also released sediment as glacial<br />
outwash, which was mainly transported downstream.<br />
I calculated rates based on <strong>the</strong> entire catchment above<br />
each moraine belt, as well as by assuming that morainal<br />
debris originated from just <strong>the</strong> cirque that held<br />
<strong>the</strong> glacier. Results suggest that bed lowering was<br />
0.25 to 2.98 mm per year for Horseshoe, 0.5 to 1.33<br />
mm per year for Green Lakes and Arapaho, and 0.15<br />
to 1.37 mm per year for Rainbow (Table 2).<br />
DISCUSSION<br />
The modeled glacier shows that while it covered an<br />
impressive area, <strong>the</strong> GLV/Arapaho ice at maximum<br />
was only 83 meters thick on average, and 240 meters<br />
at its thickest. The glacier’s thinness likely reflects<br />
relatively dry climate. Thinness may also suggest a<br />
high ice velocity, but this is not likely given <strong>the</strong> cool<br />
climate and low precipitation. High velocity could<br />
also explain <strong>the</strong> area <strong>of</strong> ice below <strong>the</strong> ELA, which is<br />
strikingly large considering <strong>the</strong> small accumulation<br />
zone in <strong>the</strong> Arapaho and Green Lakes Valleys. Given<br />
<strong>the</strong> stepped nature <strong>of</strong> <strong>the</strong> bed (slope ranges from flat<br />
to near vertical), flow velocity was likely highly varied.<br />
The erosion rates calculated here are reasonable for<br />
a small glacier in a relatively dry climate flowing<br />
over hard rocks. In <strong>the</strong>ir glacial-valley pr<strong>of</strong>ile model,<br />
MacGregor and Anderson (2000) use an erosion rate<br />
<strong>of</strong> 1 mm per year. Ward et al. (2009) use erosion<br />
rates <strong>of</strong> 0.1, 1.0, and 10 mm per year in <strong>the</strong>ir modeling<br />
<strong>of</strong> <strong>the</strong> Middle Boulder Creek Valley glaciation. It<br />
is important to realize that <strong>the</strong> rates calculated here<br />
122<br />
provide provisional upper and lower limits based on<br />
field evidence. Lidar data provide a precise DEM<br />
and GIS-based spatial analysis (kriging in this case),<br />
which is a rational method for calculating <strong>the</strong> surface<br />
beneath <strong>the</strong> late Pinedale morainal debris. We have<br />
measured <strong>the</strong> depth <strong>of</strong> sediment and water in several<br />
kettles and believe that 4 m is a reasonable estimate<br />
for an average depth and thus a thickness to add to<br />
<strong>the</strong> morainal belt. Additional field measurements and<br />
shallow geophysics would allow us to better constrain<br />
<strong>the</strong>se values. Additionally we have not estimated <strong>the</strong><br />
volume <strong>of</strong> sediment that was carried away by glaci<strong>of</strong>luvial<br />
and fluvial processes. Finally, we have not estimated<br />
what fraction <strong>of</strong> <strong>the</strong> sediment in <strong>the</strong> moraines<br />
was eroded by glaciers from <strong>the</strong> cirques and which<br />
portion comes from slope processes in <strong>the</strong> rest <strong>of</strong> <strong>the</strong><br />
catchment. However, while <strong>the</strong>se unestimated values<br />
mainly prevent me from measuring <strong>the</strong> total glacial<br />
erosion rate, <strong>the</strong> range <strong>of</strong> general lowering rates I<br />
present for each location are well constrained.<br />
One would expect <strong>the</strong> Rainbow and Horseshoe glaciers<br />
to have nearly identical erosion rates, as <strong>the</strong>y<br />
are similar size cirques located less than 1 kilometer<br />
apart on <strong>the</strong> same slope. However, Horseshoe gives<br />
a greater rate <strong>of</strong> incision. This is likely <strong>the</strong> result <strong>of</strong><br />
bedrock differences and catchment size. While Rainbow<br />
is underlain with primarily metaseds, Horseshoe<br />
is cutting into <strong>the</strong> coarser Boulder Creek Granodiorite.<br />
Additionally, Horseshoe sits in a catchment 60%<br />
larger than Rainbow’s. This suggests that <strong>the</strong> periglacial<br />
and o<strong>the</strong>r processes in <strong>the</strong> catchment contributed<br />
an important fraction <strong>of</strong> <strong>the</strong> sediment in <strong>the</strong> moraine<br />
belt.<br />
When you consider cirque volume, <strong>the</strong>se values indicate<br />
that cutting <strong>the</strong> Green Lakes and Arapaho Valleys<br />
to <strong>the</strong>ir present depth would require 11 glacier periods<br />
<strong>of</strong> <strong>the</strong> same size as <strong>the</strong> 6000 years <strong>of</strong> <strong>the</strong> late Pinedale.<br />
Cutting <strong>the</strong> Rainbow and Horseshoe cirques would<br />
require 6.5 and 3.6 such glaciations, respectively. The<br />
calculated cirque-cutting time values suggest that <strong>the</strong><br />
late Pinedale was a significant event in <strong>the</strong> morphologic<br />
evolution <strong>of</strong> <strong>the</strong>se features. These calculations<br />
assume that <strong>the</strong> erosion rate would be <strong>the</strong> same in<br />
previous glaciations.
24th Annual Keck Symposium: 2011 Union College, Schenectady, NY<br />
CONCLUSIONS<br />
Using a combination <strong>of</strong> field evidence, numerical<br />
modeling, and GIS data I was able to reconstruct <strong>the</strong><br />
Pinedale glaciation <strong>of</strong> <strong>the</strong> GLV region <strong>of</strong> <strong>the</strong> Colorado<br />
Front Range.<br />
Modeling suggests that <strong>the</strong> glacier reached an areal<br />
extent <strong>of</strong> 22 km 2 , but that <strong>the</strong> ice was generally thin,<br />
with an average thickness <strong>of</strong> 83 meters. This could<br />
indicate high flow rate, but more likely reflects <strong>the</strong><br />
dry climate.<br />
The calculated moraine volumes suggest that <strong>the</strong><br />
Pinedale represents a significant episode in <strong>the</strong> evolution<br />
<strong>of</strong> Front Range morphology. The calculated<br />
erosion rates, which are on <strong>the</strong> order <strong>of</strong> 1 mm per<br />
year, represent a substantial movement <strong>of</strong> sediment<br />
within <strong>the</strong> critical zone and exposure <strong>of</strong> large volumes<br />
<strong>of</strong> fresh surfaces.<br />
These results represent a work in progress. Fur<strong>the</strong>r<br />
work will include more detailed modeling <strong>of</strong> <strong>the</strong><br />
retreat <strong>of</strong> Pinedale ice, studying <strong>the</strong> development <strong>of</strong><br />
steps within <strong>the</strong> GLV, and determining <strong>the</strong> origin <strong>of</strong><br />
<strong>the</strong> polished boulders that sit 50 m above trimline on<br />
<strong>the</strong> south-facing slope <strong>of</strong> <strong>the</strong> valley.<br />
ACKNOWLEDGEMENTS<br />
Thanks to my advisor David Dethier for getting me to<br />
this point and somehow keeping it fun. And thanks to<br />
Evan Dethier and James McCarthy for <strong>the</strong>ir help and<br />
companionship in <strong>the</strong> field and classroom.<br />
REFERENCES<br />
Caine, N. (1995). Snowpack Influences on Geomorphic<br />
Processes in Green Lakes Valley, Colorado<br />
Front Range. The Geographic Journal 161(1):<br />
55-68.<br />
Gable, F. J. and R. F. Madole (1976). Geologic Map<br />
<strong>of</strong> <strong>the</strong> Ward Quadrangle, Boulder County, Colorado,<br />
USGS.<br />
Hulton, N. R. J. and D. I. Benn (2010). An Excel<br />
spreadsheet program for reconstructing <strong>the</strong> sur-<br />
123<br />
face pr<strong>of</strong>ile <strong>of</strong> former mountain glaciers and ice<br />
caps. Computers and Geosciences 36: 605-610.<br />
MacGregor, K.R., Anderson, R.S., Anderson, S.P.,<br />
and Waddington, E.D. Numerical simiulation<br />
<strong>of</strong> glacial-valley longitudinal pr<strong>of</strong>ile evolution.<br />
Geology. 2000.<br />
Madole, R. F., D. VanSistine, and Michael, J.A.<br />
(1998). Glaciation in <strong>the</strong> upper Platte River<br />
drainage basin. Geologic Investigations Series.<br />
USGS.<br />
Ward, D. J., Anderson, Robert S., Guido, Zackry S.,<br />
and Briner, Jason P. (2009). Numerical Modeling<br />
<strong>of</strong> Cosmogenic Deglaciation Records, Front<br />
Range and San Juan Mountains, Colorado. Journal<br />
<strong>of</strong> Geophysical Research 114.
24th Annual Keck Symposium: 2011 Union College, Schenectady, NY<br />
CHARACTERIZATION OF TRACE METAL CONCENTRATIONS<br />
AND MINING LEGACY IN SOILS, BOULDER COUNTY,<br />
COLORADO<br />
ELLEN M. MALEY, Smith College<br />
Research Advisor: Amy L. Rhodes<br />
INTRODUCTION<br />
The Boulder Creek Critical Zone Observatory (CZO)<br />
is located in an area steeped in mineral exploitation<br />
dating back to <strong>the</strong> 1880s. The Colorado Mineral Belt<br />
is comprised <strong>of</strong> igneous intrusions and ore deposits<br />
from <strong>the</strong> Laramide orogeny. This band strikes nor<strong>the</strong>ast/southwest<br />
through <strong>the</strong> region, stretching 250<br />
miles (Tweto and Sims, 1963). Beginning in <strong>the</strong> late<br />
1800s, this region was mined for gold, tungsten, lead,<br />
silver, and zinc (Fig. 1). Along with prospecting,<br />
mining, and smelting, this region was subject to deforestation<br />
and development with <strong>the</strong> flux <strong>of</strong> settlers<br />
in <strong>the</strong> late 1800s (Veblen and Lorenz, 1991). The<br />
surge <strong>of</strong> European settlers in <strong>the</strong> mining boom made<br />
this region an industrial hub, utilizing processes now<br />
known to have negative impacts on ecosystems.<br />
The Boulder Creek CZO project focuses on <strong>the</strong> balance<br />
between natural wea<strong>the</strong>ring and erosion processes<br />
in <strong>the</strong> critical zone. The purpose <strong>of</strong> this investigation<br />
is to quantify heavy metal concentrations<br />
in regional soils to identify possible contamination<br />
associated with mineral exploitation. In investigation<br />
<strong>of</strong> soil processes in <strong>the</strong> critical zone, environmental<br />
impacts associated with <strong>the</strong> flux <strong>of</strong> human settlers,<br />
such as deforestation, road construction, and mining,<br />
should not be overlooked. In <strong>the</strong> case <strong>of</strong> mining,<br />
Diawara et al. (2006) report surface soil contamination<br />
<strong>of</strong> lead (300 ppm), arsenic (30 ppm), cadmium (5<br />
ppm), and mercury (200 ppb) around major smelting<br />
sites in sou<strong>the</strong>rn Colorado. While certain elements,<br />
such as Pb, have been detected in Boulder County<br />
soils, <strong>the</strong> heavy metal signature associated with this<br />
region has not been investigated in detail (Dethier,<br />
personal communication). This investigation characterizes<br />
<strong>the</strong> regional extent <strong>of</strong> metal contamination<br />
across a range <strong>of</strong> sites in Boulder County.<br />
This investigation employs <strong>the</strong> analysis <strong>of</strong> spatial and<br />
geochemical data to determine whe<strong>the</strong>r surface soils<br />
exhibit trace metal enrichment and, if so, to characterize<br />
<strong>the</strong> nature <strong>of</strong> contamination across southwestern<br />
Boulder County. This study investigates lead, arsenic,<br />
cadmium, chromium, barium, manganese, mercury,<br />
copper, iron, and aluminum. These 10 metals include<br />
major soil constituents (Fe, Al, Mn) and common<br />
environmental contaminants (Pb, Cd, As and Hg).<br />
Trace metal enrichment in <strong>the</strong> A horizon compared to<br />
material from depth (C horizon) is expected to be a<br />
function <strong>of</strong> two variables: natural accumulation and<br />
external accumulation. Pedogenesis releases mineralbound<br />
elements to <strong>the</strong> soil through wea<strong>the</strong>ring. Biochemical<br />
processes draw elements, such as manganese,<br />
from depth and concentrate <strong>the</strong>m in surface soils<br />
via decomposition. Natural accumulation depends on<br />
pr<strong>of</strong>ile wea<strong>the</strong>ring and organic enrichment.<br />
Alternatively, external accumulation reflects anthropogenic<br />
deposition. The main control on deposition<br />
is <strong>the</strong> distance between a location and its potential<br />
contamination source, in this case, mining/smelting<br />
sites. The likelihood <strong>of</strong> contamination surrounding<br />
historic mining and smelting sites is based on <strong>the</strong><br />
observation by Kabata-Pendias (2001) that metal<br />
halos surround locations such as <strong>the</strong>se. With no o<strong>the</strong>r<br />
factors considered, metals associated with mining<br />
should be highest in concentration in soils surrounding<br />
<strong>the</strong> source <strong>of</strong> enrichment, and will decrease along<br />
a gradient with distance from <strong>the</strong> source, mostly in<br />
<strong>the</strong> direction <strong>of</strong> transport.<br />
Controls on accumulation pertain to <strong>the</strong> affinity<br />
each soil has for an element. Kabata-Pendias (2001)<br />
suggests that organic and clay content are <strong>the</strong> main<br />
controls for metal adsorption in soils, as oxide minerals<br />
and organics have partial negative-charged surfaces,<br />
which tend to form weak bonds with positively<br />
charged metals. Soil pH affects metal solubility and<br />
124
24th Annual Keck Symposium: 2011 Union College, Schenectady, NY<br />
is a function <strong>of</strong> mineralogy and organic content. Eh/<br />
pH plots for elements demonstrate increased mobility<br />
in acidic, reducing soils. To characterize <strong>the</strong><br />
distribution <strong>of</strong> metals in Boulder County soils in this<br />
investigation, trace metal concentrations were evaluated<br />
in relation to <strong>the</strong> soil properties <strong>of</strong> pH, % organic<br />
content, and % fine-grained material. Possible spatial<br />
relationships between trace metal concentrations and<br />
locations <strong>of</strong> mining and smelting locations were also<br />
investigated.<br />
METHODS<br />
To obtain a regional spectrum <strong>of</strong> soils, samples were<br />
collected along an east-trending elevation gradient<br />
from alpine (3450 m) to lower montane (1850 m)<br />
elevations; samples included three Boulder CZO<br />
sites: Green Lakes Basin, Gordon Gulch, and Betasso<br />
Preserve. Accessible sites along major roads between<br />
CZO sites also were sampled to characterize <strong>the</strong><br />
region in higher detail (Fig. 1). Surface soils, consisting<br />
<strong>of</strong> organic and A horizons, were sampled to ~20<br />
cm depth. Unaltered parent material (C horizon) was<br />
sampled at >1 meter depth to compare surface soils<br />
with unwea<strong>the</strong>red parent material geochemically.<br />
Thirty-three air-dried samples were sieved to
24th Annual Keck Symposium: 2011 Union College, Schenectady, NY<br />
and Hydra-C were normalized by soil digest mass.<br />
Results for As, Al, Ba, Cd, Cr, Cu, Fe, Mn, and Pb<br />
in mg/kg soil (ppm), and Hg in µg/kg soil (ppb) are<br />
reported in Table 1. Sample concentrations were used<br />
to explore how metals fractionate, how metal concentrations<br />
vary by horizon, and how concentrations <strong>of</strong><br />
one element relate to o<strong>the</strong>rs within horizons.<br />
Correlation ellipsoid matrices were used to generalize<br />
relationships between elements for each horizon.<br />
Corresponding elemental A and C horizon concentrations<br />
were compared to evaluate whe<strong>the</strong>r processes<br />
are occurring through <strong>the</strong> entire soil pr<strong>of</strong>ile or are<br />
localized within a given horizon. Surface soil concentrations<br />
were analyzed spatially in relation to mining<br />
site locations using ArcGIS. Spatial parameters<br />
<strong>of</strong> distance to nearest mine site and number <strong>of</strong> mines<br />
within 1.5 kilometer radius for each sample site were<br />
created for statistical analysis, which is not addressed<br />
in this submission.<br />
Examination <strong>of</strong> correlations between different trace<br />
metals and between trace metals and soil properties<br />
may provide insight regarding metal origin. For<br />
example, a correlation between two metals within <strong>the</strong><br />
C horizon for all samples could suggest that <strong>the</strong> metal<br />
origin is similar and may be associated with parent<br />
material breakdown. Likewise, a strong correlation<br />
between two metals in surface soils and not at depth<br />
could suggest that <strong>the</strong> metals are derived from <strong>the</strong><br />
same source, such as atmospheric deposition.<br />
Fractionation <strong>of</strong> Elements<br />
The finest material in soils is comprised <strong>of</strong> silt- and<br />
clay-sized mineral and organic particles, which bear a<br />
partial-negative surface charge and have high relative<br />
surface area (Kabata-Pendias, 2001). The charge and<br />
area draw metal cations to clay surfaces, which results<br />
in higher metal content in fine-grained fractions. In<br />
keeping with literature predictions about <strong>the</strong> behavior<br />
<strong>of</strong> trace metals in soils (Kabata-Pendias, 2001),<br />
fraction analysis <strong>of</strong> three samples showed higher trace<br />
metal concentration in fine-grained fractions (Fig. 2).<br />
Soil chemistry was comparable between 125µm.<br />
126<br />
For example, Pb and As show a negative correlation<br />
between grain-size and concentration (Fig. 2). Concentrations<br />
are higher in finer fractions than in coarse<br />
material for all elements. Analyses suggest that aluminum<br />
is lower in concentration at Magnolia Road, a<br />
21 kyr moraine soil, compared to Gordon Gulch and<br />
Betasso sites. As this soil does not vary significantly<br />
in age with o<strong>the</strong>r sites (see Wyshnytsky, this volume),<br />
this difference is likely due to variations in host rock<br />
mineralogy between sites.<br />
Figure 2. Example plots <strong>of</strong> Al, Pb, and As concentration<br />
distribution by size fraction (in µm) from three sample<br />
locations, compared with bulk concentrations in ppm (mg/<br />
kg soil) for each site.<br />
Elemental Trends by Soil Horizon<br />
Boulder County A horizons are enriched in Mn, Pb,<br />
Ba, Hg, and As. Surface soil enrichment results from<br />
accumulation and/or atmospheric deposition. Anthropogenic<br />
addition to soils from mining activity<br />
has occurred via atmospheric deposition. However,<br />
external dust input and bioaccumulation, <strong>the</strong> tendency<br />
for plants to draw elements from depth and concentrate<br />
<strong>the</strong>m in surface soils, may also be occurring.<br />
Therefore, elements enriched at <strong>the</strong> surface may be<br />
<strong>of</strong> anthropogenic or natural origin. In contrast, C<br />
horizons are enriched in Cu and Cr. Iron and Al do<br />
not vary consistently between horizons. Elements<br />
without surface enrichment are likely derived from<br />
original host rock composition.<br />
The A to C relationship across sites was evaluated by<br />
comparing elemental concentrations between A and
24th Annual Keck Symposium: 2011 Union College, Schenectady, NY<br />
Pb (ppm)<br />
Hg (ppb)<br />
100<br />
80<br />
60<br />
40<br />
20<br />
Pb<br />
Hg<br />
As<br />
Figure 3. Mercury, As, Pb and Mn, Ba as functions <strong>of</strong> % organic content with regression line equations and associated R 2<br />
values displayed. These correlations suggest that elements that are enriched in surface soils are associated with organic<br />
content.<br />
C horizons collected at <strong>the</strong> same sample site. Linear<br />
trends between A and C horizons for Fe, Cu and Cr<br />
suggest that variations by site are associated with<br />
pedogenic breakdown <strong>of</strong> host rock. Enrichment in <strong>the</strong><br />
A horizon and <strong>the</strong> relationships between A-enriched<br />
elements and soil properties, like % organics, warranted<br />
fur<strong>the</strong>r investigation.<br />
Elemental Relationships<br />
y = 7.0426 + 2.2878x R 2 = 0.51071<br />
y = 2.1103 + 2.9885x R 2 = 0.73348<br />
y = 1.1015 + 0.14777x R 2 = 0.61033<br />
0<br />
0<br />
0 5 10 15 20 25<br />
% organic content<br />
In surface soils, enrichment <strong>of</strong> elements can be<br />
derived from natural and anthropogenic processes.<br />
Elevated concentrations <strong>of</strong> Mn and Ba may result<br />
from biological surficial concentration <strong>of</strong> elements<br />
from depth or from atmospheric dust input, while <strong>the</strong><br />
source <strong>of</strong> Pb, As and Hg is anthropogenic (Kabata-<br />
Pendias, 2001). As Mn is extracted from depth over<br />
time by plants, decomposing organic matter may concentrate<br />
Mn in surface soils (Kabata-Pendias, 2001).<br />
There is a positive correlation between Mn and Ba,<br />
which suggests a common source. Mercury, Pb, As,<br />
Mn, and Ba are enriched in surface soils, and each has<br />
a positive correlation with % organic content (Fig. 3).<br />
5<br />
4<br />
3<br />
2<br />
1<br />
As (ppm)<br />
As (ppm)<br />
Mn (ppm)<br />
6<br />
5<br />
4<br />
3<br />
2<br />
1<br />
500<br />
400<br />
300<br />
200<br />
100<br />
As<br />
Hg<br />
Mn<br />
Ba<br />
y = 41.748 + 11.948x R 2 = 0.43988<br />
y = 26.236 + 1.4091x R 2 = 0.2561<br />
0<br />
0<br />
0 5 10 15 20 25<br />
% organic content<br />
y = 1.1148 + 0.044219x R 2 = 0.55056<br />
y = 2.5292 + 0.88298x R 2 = 0.67269<br />
0<br />
0<br />
0 20 40 60 80100 Pb (ppm)<br />
Figure 4. Mercury (ppb) and As (ppm) across all horizons<br />
and sites with corresponding Pb (ppm) concentrations,<br />
showing <strong>the</strong> relationship between Hg/Pb and As/Pb across<br />
southwestern Boulder County. R2 values (Hg/Pb R2 =<br />
0.67, As/Pb R2 = 0.55) suggest correlation between variables.<br />
127<br />
100<br />
80<br />
60<br />
40<br />
20<br />
Ba (ppm)<br />
120<br />
100<br />
80<br />
60<br />
40<br />
20<br />
Hg (ppb)
24th Annual Keck Symposium: 2011 Union College, Schenectady, NY<br />
The most striking trend in trace metals occurs between<br />
Pb, As, and Hg in surface soils (Fig. 4). Strong<br />
correlation between Pb, As, and Hg (Fig. 4) in surface<br />
soils suggests that <strong>the</strong>se metals were deposited via a<br />
common atmospheric anthropogenic source.<br />
In <strong>the</strong> C horizon, Fe and Mn, Fe and Al, and Fe and<br />
Cr are positively correlated, suggesting that <strong>the</strong>se elements<br />
are associated with host rock breakdown. No<br />
relationship between Cu and Cr exists, though both<br />
<strong>of</strong> <strong>the</strong>se elements are enriched at depth. The positive<br />
correlations between Cr and Fe and Cr and Al suggest<br />
that chromium is present in host rock, and it was<br />
incorporated into soils by pedogenesis. Iron and Mn<br />
are positively correlated at depth, but not in surface<br />
soils where Mn is enriched. This difference could be<br />
attributed to Mn-rich dust inputs or bioaccumulation<br />
in organic matter (Kabata-Pendias, 2001).<br />
Mining Signal and Implications<br />
Surface soils are enriched with respect to Mn, Ba, As,<br />
and Pb, but not all A horizon enrichment can be attributed<br />
to anthropogenic activity. Barium and Mn have<br />
a weak association with organic content, and <strong>the</strong>se<br />
elements may be enriched due to bioaccumulation<br />
(Kabata-Pendias, 2001). The correlation between As,<br />
Hg, and Pb in surface soils suggests that <strong>the</strong>se metals<br />
are affiliated with atmospheric deposition. These<br />
metals are associated with ore smelting processes that<br />
may have been used in <strong>the</strong> Boulder County region<br />
(Kabata-Pendias, 2001). These trace metal signatures<br />
are likely evidence <strong>of</strong> <strong>the</strong> mining legacy in <strong>the</strong> region.<br />
CONCLUSIONS<br />
The enrichment <strong>of</strong> surface soils in elements associated<br />
with anthropogenic activity suggests that Boulder<br />
County soils bear a signature associated with historic<br />
mining activity in <strong>the</strong> region. Surface soils in sou<strong>the</strong>rn<br />
Boulder County are enriched in Hg, Mn, Ba, As,<br />
and Pb. Lead, arsenic, and mercury enrichment likely<br />
results from anthropogenic atmospheric inputs from<br />
mining activity, whereas <strong>the</strong> origin <strong>of</strong> manganese and<br />
barium enrichment is likely from bioaccumulation<br />
or dust inputs. Measured concentrations <strong>of</strong> metals<br />
reflect mechanisms <strong>of</strong> deposition and mobilization.<br />
While o<strong>the</strong>r elements may have been associated with<br />
deposition, Pb, As, and Hg are potential contaminants<br />
with a related source. This investigation <strong>of</strong> potential<br />
contamination <strong>of</strong> toxic elements bears significance for<br />
<strong>the</strong> Boulder Creek watershed, as <strong>the</strong> heavy metals in<br />
surface soils are bioavailable (U.S. EPA, 1996) and<br />
could mobilize. Fur<strong>the</strong>r exploration will quantify <strong>the</strong><br />
mass <strong>of</strong> each element by area to characterize contamination<br />
potential <strong>of</strong> <strong>the</strong> watershed.<br />
Within a soil pr<strong>of</strong>ile, trace metals form weak bonds<br />
with fine-grained (1000 ppm<br />
sample excluded from range (n=6); B: O (n=4), A (n=14),<br />
C (n=3). C: compared with mean values in U.S. soils, as<br />
reported by Kabata-Pendias (2001).<br />
128
24th Annual Keck Symposium: 2011 Union College, Schenectady, NY<br />
this project. I also thank Dr. Robert Newton, for his<br />
encouragement and wisdom, and Dr. Anna Martini<br />
and Erin Camp at Amherst College for <strong>the</strong>ir assistance<br />
with mercury analysis.<br />
REFERENCES<br />
Causey, J. Douglas. 1998. MAS/MILS Arc/Info Point<br />
Coverage for <strong>the</strong> Western United States (Excluding<br />
Hawaii): United States Geological Survey<br />
Open-File Report 98-512, Spokane, WA.<br />
Diawara, M., J. Litt, D. Unis, N. Alfonso, L. Martinez,<br />
J. Crock, D. Smith, and J. Carsella. 2006. “Arsenic,<br />
Cadmium, Lead, and Mercury in Surface Soils,<br />
Pueblo, Colorado: Implications for Population<br />
Health Risk.” Environmental Geochemistry and<br />
Health 28 (4): 297-315.<br />
Kabata-Pendias, A. and H. Pendias. 2001. Trace Elements<br />
in Soils and Plants. 3rd ed. Boca Raton,<br />
FL: CRC Press: 331 pp.<br />
Kimbrough, David E. and Janice R. Wakakuwa. 1989.<br />
“Acid Digestion for Sediments, Sludges, Soils,<br />
and Solid Wastes. A Proposed Alternative to EPA<br />
SW 846 Method 3050.” Environmental Science &<br />
Technology 23 (7): 898-900.<br />
Robertson, G. Philip, David C. Coleman, Caroline S.<br />
Bledsoe, and Phillip Sollins, eds. 1999. Standard<br />
Soil Methods for Long-Term Ecological Research.<br />
New York: Oxford University Press.<br />
Rowell, D. L. 1994. “Section 3.3: The Determination<br />
<strong>of</strong> Water Content and Loss on Ignition.” In Soil<br />
Science: Methods & Applications, 48. Reading,<br />
Massachusetts: Prentice Hall.<br />
Tweto, O. and P. K. Sims. 1963. “Precambrian Ancestry<br />
<strong>of</strong> <strong>the</strong> Colorado Mineral Belt.” Geological<br />
Society <strong>of</strong> America Bulletin 74 (8): 991-1014.<br />
Veblen, Thomas T. and Diane C. Lorenz. 1991. The<br />
Colorado Front Range: A Century <strong>of</strong> Change.<br />
Salt Lake City, Utah: University <strong>of</strong> Utah Press.<br />
United States Environmental Protection Agency.<br />
129<br />
1996. Method 3050B: Acid Digestion <strong>of</strong> Sediments,<br />
Sludges, and Soils–Revision 2. United<br />
States EPA SW-846: Test Methods for Evaluating<br />
Solid Waste, Physical/Chemical Methods.
24th Annual Keck Symposium: 2011 Union College, Schenectady, NY<br />
INTRODUCTION<br />
The Critical Zone<br />
The Critical Zone extends from fresh rock to <strong>the</strong> top<br />
<strong>of</strong> <strong>the</strong> vegetation canopy and is where <strong>the</strong> biosphere,<br />
atmosphere, hydrosphere, and rock materials interact<br />
(Anderson et al., 2007). The complex processes<br />
occurring in <strong>the</strong> Critical Zone release raw materials<br />
from minerals and create substrates for terrestrial life,<br />
supporting microbial, plant, and faunal activity. Rates<br />
<strong>of</strong> chemical and mechanical processes vary temporally<br />
and spatially due to gradients in climate, rock<br />
materials, and slope. Soil is <strong>the</strong> highly wea<strong>the</strong>red,<br />
top-most part <strong>of</strong> <strong>the</strong> regolith, but is distinct because<br />
<strong>of</strong> its unique layered habit. These layers (horizons),<br />
record more intense wea<strong>the</strong>ring conditions at <strong>the</strong><br />
surface, and record downward transport <strong>of</strong> chemical<br />
wea<strong>the</strong>ring products and organic additions from<br />
<strong>the</strong> biosphere (Anderson and Anderson, 2010). Field<br />
and laboratory studies <strong>of</strong> soils enrich models <strong>of</strong> rock<br />
wea<strong>the</strong>ring, chemical and mechanical transport, and<br />
regolith formation. Results help to highlight <strong>the</strong> role<br />
<strong>of</strong> soil-forming processes as indicators <strong>of</strong> <strong>the</strong> geomorphic<br />
controls on Critical Zone processes.<br />
The Critical Zone can be thought <strong>of</strong> as a bottom-up<br />
feed-through reactor in which physical and chemical<br />
wea<strong>the</strong>ring processes alter fresh rock material being<br />
supplied by uplift and erosion (Anderson et al., 2007).<br />
Simultaneously, physical erosion and chemical denudation<br />
processes transport mass out <strong>of</strong> <strong>the</strong> system.<br />
Thus, <strong>the</strong> rates <strong>of</strong> wea<strong>the</strong>ring and denudation toge<strong>the</strong>r<br />
determine <strong>the</strong> thickness <strong>of</strong> <strong>the</strong> Critical Zone (Anderson<br />
et al., 2007). The balance <strong>of</strong> transport-limited and<br />
wea<strong>the</strong>ring-limited environments is largely determined<br />
by topographic and climatic factors.<br />
Catenas<br />
ASSESSING EOLIAN CONTRIBUTIONS TO SOILS IN THE<br />
BOULDER CREEK CATCHMENT, COLORADO<br />
JAMES A. MCCARTHY, Williams College<br />
Research Advisor: David P. Dethier<br />
The bottom-up reactor model for regolith formation<br />
130<br />
assumes that parent materials in <strong>the</strong> system derive<br />
only from <strong>the</strong> wea<strong>the</strong>ring <strong>of</strong> underlying bedrock or<br />
sediment. However, soil pr<strong>of</strong>iles on slopes are distinctly<br />
related to <strong>the</strong> soils above and below because<br />
<strong>of</strong> <strong>the</strong> influence <strong>of</strong> slope-controlled transport mechanisms.<br />
The term catena describes a sequence <strong>of</strong> soils<br />
on a slope, emphasizing that <strong>the</strong>ir variation is due to<br />
changes in both slope gradient and position (Birkeland,<br />
1999). In this model, regolith and soils are<br />
thinnest at <strong>the</strong> shoulder and backslope, and gradually<br />
thicken, reaching a maximum at <strong>the</strong> toeslope. There<br />
is also a chemical gradient along slopes, driven partly<br />
by <strong>the</strong> physical transport <strong>of</strong> <strong>the</strong> mobile regolith, and<br />
enhanced by hydrologic factors; clay minerals and<br />
dissolved cations in a soil column may be transported<br />
down slope by throughflow, accumulating at <strong>the</strong> base<br />
<strong>of</strong> <strong>the</strong> slope. Climatic conditions determine both <strong>the</strong><br />
mobility <strong>of</strong> soil materials and chemical constituents,<br />
and thus <strong>the</strong> effects <strong>of</strong> throughflow on pedogenesis<br />
vary spatially. Thicker soils and better-developed<br />
horizons may occur at <strong>the</strong> base <strong>of</strong> slopes due to <strong>the</strong><br />
influx <strong>of</strong> wea<strong>the</strong>rable materials and increased soilmoisture<br />
status via throughflow water.<br />
Downslope transport <strong>of</strong> <strong>the</strong> mobile layer may be episodic,<br />
mediated by climate, and <strong>the</strong> regolith that eventually<br />
arrives at <strong>the</strong> toeslope includes a mixture <strong>of</strong><br />
soil and saprolite. Models <strong>of</strong> sediment flux generally<br />
assume that hillslope processes are constant through<br />
time, but episodic transport suggests more stochastic<br />
conditions (Anderson and Anderson, 2010). Regolith<br />
moves downslope and episodic transport results in <strong>the</strong><br />
burial <strong>of</strong> soils at <strong>the</strong> base <strong>of</strong> <strong>the</strong> slope; current soils<br />
in <strong>the</strong>se positions form from materials derived from<br />
upslope ra<strong>the</strong>r than from bedrock. Thus, morphological<br />
differences along hillslopes may represent changes<br />
in <strong>the</strong> strength <strong>of</strong> pedogenesis and changes in parent<br />
material.
24th Annual Keck Symposium: 2011 Union College, Schenectady, NY<br />
Significance <strong>of</strong> dustfall<br />
Dustfall is not included explicitly in <strong>the</strong> Critical Zone<br />
reactor model; however, <strong>the</strong> enrichment <strong>of</strong> eolian silt<br />
and clay affects <strong>the</strong> “mineralogy, chemistry, nutrient<br />
status, and moisture-holding capacity <strong>of</strong> soils (Muhs<br />
and Benedict, 2006, p. 120),” and thus “may control<br />
<strong>the</strong> rate and direction <strong>of</strong> pedogenesis (Mason and Jacobs,<br />
1998, p. 1135).” For this reason, characterizing<br />
<strong>the</strong> rate <strong>of</strong> eolian inputs is important for Critical Zone<br />
studies.<br />
Transport, deposition, and subsequent wea<strong>the</strong>ring <strong>of</strong><br />
eolian materials are controlled by climate. In areas<br />
distal from eolian source materials, fine sediments are<br />
deposited at relatively low rates; soil formation exceeds<br />
deposition, and <strong>the</strong> original parent material (e.g.<br />
crystalline bedrock) primarily influences pedogenesis<br />
(Birkeland, 1999; Muhs et al., 2008). Dust inputs<br />
are rapidly wea<strong>the</strong>red at <strong>the</strong> surface, and potentially<br />
<strong>of</strong>fset losses from <strong>the</strong> wea<strong>the</strong>ring <strong>of</strong> <strong>the</strong> original parent<br />
material (Mason and Jacobs, 1998). The volume<br />
and geochemistry <strong>of</strong> <strong>the</strong> dust determine its effect on<br />
pedogenesis.<br />
Based on field observations and laboratory analysis,<br />
this study seeks to: (1) track <strong>the</strong> concentration <strong>of</strong><br />
pedogenic iron within and between soil pr<strong>of</strong>iles to<br />
characterize wea<strong>the</strong>ring patterns in high-relief environments;<br />
and (2) assess <strong>the</strong> contribution and effects<br />
<strong>of</strong> eolian materials on montane soils.<br />
METHODS<br />
Field<br />
I selected 24 sites in <strong>the</strong> Boulder Creek catchment<br />
for field description and sampling (Fig. 1). The sites<br />
toge<strong>the</strong>r represent varying elevation, slope, aspect,<br />
parent material, moisture regime, and inferred age.<br />
Sites include upper montane Gordon Gulch (14), <strong>the</strong><br />
alpine and subalpine Green Lakes basin (5), a highway<br />
road cut near <strong>the</strong> town <strong>of</strong> Ward (2), and Betasso<br />
Gulch (2). Ten <strong>of</strong> <strong>the</strong> lower Gordon Gulch sites comprise<br />
two catenas (5 sites in each) that face north and<br />
south. Lower Gordon Gulch is oriented so that slopes<br />
have near north-south aspect directions. Sites on both<br />
transects stretch from just below <strong>the</strong> catchment crest<br />
to just above <strong>the</strong> break in slope with valley fans and<br />
131<br />
terrace deposits. At each site, I dug soil pits with my<br />
Keck colleagues until excessive depth or hard saprolite<br />
prevented fur<strong>the</strong>r excavation. I followed standard<br />
soil description procedures and used a Garmin GPS<br />
device to record <strong>the</strong> site location and elevation. I collected<br />
approximately 750g <strong>of</strong> soil from all horizons<br />
(98 total samples), with <strong>the</strong> exception <strong>of</strong> O and L<br />
horizons when present.<br />
Figure 1. Locations <strong>of</strong> soil pits that were dug, described<br />
and sampled in <strong>the</strong> Boulder Creek watershed.<br />
Laboratory<br />
In <strong>the</strong> laboratory I dry-sieved all samples using<br />
USDA standard soil sieves. I <strong>the</strong>n subdivided <strong>the</strong><br />
24th Annual Keck Symposium: 2011 Union College, Schenectady, NY<br />
obtained from D.P. Dethier), I established background<br />
values for Fe 2 O 3 for each site. I <strong>the</strong>n calculated <strong>the</strong><br />
mass <strong>of</strong> accumulated Fe 2 O 3 at each site by subtracting<br />
<strong>the</strong> background from <strong>the</strong> total mass <strong>of</strong> Fe 2 O 3 to<br />
yield total accumulated Fe 2 O 3 (in g cm -2 ). I calculated<br />
total accumulated clay by <strong>the</strong> same method. I also<br />
ran samples from previous field studies in <strong>the</strong> Front<br />
Range for comparison, including a pr<strong>of</strong>ile formed in<br />
Bull Lake (130ka) till.<br />
RESULTS<br />
Field<br />
Soil morphology varies widely across <strong>the</strong> study area,<br />
reflecting <strong>the</strong> differences in parent material, relief,<br />
and climate within <strong>the</strong> Boulder Creek CZO (Fig. 2).<br />
Soils in <strong>the</strong> Green Lakes basin formed from glacial<br />
till and periglacial debris and, at stable sites, typically<br />
exhibit A/Ej/Bw/Cox/Cu pr<strong>of</strong>iles. Soils in <strong>the</strong> Betasso<br />
Gulch derive from ei<strong>the</strong>r wea<strong>the</strong>red colluvium<br />
or highly wea<strong>the</strong>red saprolite, and sample sites show<br />
thick (>50cm) Bt horizons and, locally, a thick buried<br />
soil.<br />
In Gordon Gulch <strong>the</strong> two catenas are broadly similar,<br />
but show marked differences near <strong>the</strong> footslope.<br />
Soils on <strong>the</strong> south-facing slope transect are thin (
24th Annual Keck Symposium: 2011 Union College, Schenectady, NY<br />
age percent clay <strong>of</strong> <strong>the</strong> surface horizon and subsurface<br />
horizons from all 24 sites is 6.17% (1 sigma = 1.71)<br />
and 5.02% (1 sigma = 3.21), respectively. This indicates<br />
slight enrichment <strong>of</strong> <strong>the</strong> surface horizons in clay.<br />
Along <strong>the</strong> Gordon Gulch transects (9 sample sites),<br />
<strong>the</strong> average percent clay <strong>of</strong> surface and subsurface horizons<br />
is 5.07% (1 sigma = 1.08) and 3.12% (1 sigma<br />
= 1.07), respectively. Differences in silt concentration<br />
between surface and subsurface horizons are<br />
insignificant across <strong>the</strong> entire catchment. However,<br />
along <strong>the</strong> two Gordon Gulch transects, surface horizons<br />
may be slightly enriched in silt.<br />
Dithionite-extractable iron (Fe d ), as %Fe 2 O 3 , is<br />
slightly lower in surface horizons than subsurface<br />
horizons; average %Fe 2 O 3 <strong>of</strong> <strong>the</strong> surface horizon and<br />
subsurface horizons from 24 sites is 1.52% (1 sigma<br />
= 0.44) and 1.81% (1 sigma = 0.65), respectively. In<br />
Gordon Gulch, average %Fe 2 O 3 <strong>of</strong> <strong>the</strong> surface and<br />
subsurface horizons is 1.43% (1 sigma = 0.22) and<br />
1.74% (1 sigma = 0.46), respectively. Total pr<strong>of</strong>ile<br />
accumulated Fe 2 O 3 values are extremely variable<br />
across <strong>the</strong> catchment, with an average value <strong>of</strong> 1.84<br />
g cm-2 and a standard deviation <strong>of</strong> 1.51 g cm -2 . The<br />
maximum value is 5.1 g cm -2 at <strong>the</strong> site “WRC-01”;<br />
<strong>the</strong> minimum accumulated Fe 2 O 3 is 0.19 g cm-2 at <strong>the</strong><br />
site “SFT-03.”<br />
DISCUSSION<br />
Data collected in this study permit evaluation <strong>of</strong><br />
wea<strong>the</strong>ring rates and eolian sedimentation in <strong>the</strong><br />
Boulder Creek catchment. Dithionite-extractable iron<br />
(Fe d ) represents <strong>the</strong> concentration <strong>of</strong> secondary iron<br />
oxides and organically bound iron complexes in a soil<br />
(McFadden and Hendricks, 1985). Higher concentrations<br />
<strong>of</strong> Fe d in a sample thus indicate higher amounts<br />
<strong>of</strong> “free” iron that have been released by wea<strong>the</strong>ring<br />
and total pr<strong>of</strong>ile Fe d content typically increases<br />
with soil age and may correlate with total pr<strong>of</strong>ile clay<br />
content (McFadden and Hendricks, 1985). This<br />
relationship is expected in soils, as clay minerals and<br />
iron oxides are both products <strong>of</strong> wea<strong>the</strong>ring, and time<br />
positively influences <strong>the</strong> degree <strong>of</strong> wea<strong>the</strong>ring (Birkeland,<br />
1999). Therefore, variation in total pr<strong>of</strong>ile Fe d<br />
may indicate relative ages <strong>of</strong> soils across a study area.<br />
133<br />
Figure 3. Fe d accumulation rate, determined by using<br />
total pr<strong>of</strong>ile Fe d content <strong>of</strong> four soils <strong>of</strong> known age. The<br />
soil from Betasso Gulch (red bullet) formed from deep,<br />
wea<strong>the</strong>red colluvium and an uncertain portion <strong>of</strong> <strong>the</strong> Fe 2 O 3<br />
content is inherited. For this reason, <strong>the</strong> pr<strong>of</strong>ile was not<br />
used to fit <strong>the</strong> curve.<br />
Total pr<strong>of</strong>ile Fe d increases with age at stable sites in<br />
<strong>the</strong> Boulder Creek catchment (Fig. 3). The strong<br />
positive correlation <strong>of</strong> Fe d with age suggests that<br />
pedogenic clay should also increase. However,<br />
textural data show that surface horizons, relative to<br />
subsurface horizons, are commonly enriched in clay<br />
but depleted in Fe d (Table 1). The enrichment <strong>of</strong> clay<br />
in surface horizons, coupled with relatively low Fe d<br />
values, suggests that clay content indicates eolian<br />
enrichment <strong>of</strong> soils in <strong>the</strong> catchment. Clay-sized sediment<br />
suggests a distal source.<br />
Total pr<strong>of</strong>ile Fe d and clay content suggest both wea<strong>the</strong>ring<br />
and transport mechanics on slopes in Gordon<br />
Gulch (Fig. 4). On both north-facing and south-facing<br />
slopes, Fe d and clay content in soil and regolith<br />
increases downslope.
24th Annual Keck Symposium: 2011 Union College, Schenectady, NY<br />
Figure 4. Total pr<strong>of</strong>ile accumulated Fed content and total<br />
pr<strong>of</strong>ile clay content along <strong>the</strong> two Gordon Gulch transects.<br />
The north-facing slope is in black and <strong>the</strong> south-facing<br />
slope in red. The filled markers and solid lines represent<br />
Fed content, and <strong>the</strong> hollow markers and dashed lines<br />
represent clay content.<br />
The correlation <strong>of</strong> clay and Fe d content on both slopes<br />
indicates that regolith is “older” downslope. Fed<br />
and clay accumulate more rapidly on <strong>the</strong> north-facing<br />
slope than on <strong>the</strong> south-facing slope. Therefore,<br />
wea<strong>the</strong>ring rates may be higher on <strong>the</strong> north-facing<br />
slope, perhaps reflecting differences in temperature<br />
and moisture content due to aspect.<br />
CONCLUSIONS<br />
Pedogenic iron in <strong>the</strong> Boulder Creek catchment increases<br />
with age. Total pr<strong>of</strong>ile Fe d content correlates<br />
with total pr<strong>of</strong>ile clay content; this suggests that clays<br />
in sampled pr<strong>of</strong>iles are primarily pedogenic. However,<br />
clay content is slightly higher in surface horizons<br />
than in subsurface horizons and Fe d content is slightly<br />
reduced, indicating that <strong>the</strong> surface horizons are<br />
enriched in clay. High clay and low iron content in<br />
surface horizons suggests an eolian source ra<strong>the</strong>r than<br />
intensified in-situ wea<strong>the</strong>ring.<br />
Future work will focus on (1) determining <strong>the</strong> significant<br />
factors affecting eolian deposition (e.g. elevation,<br />
MAP, aspect, etc.); (2) incorporating trace-element<br />
134<br />
data in surface and parent material horizons to determine<br />
local or distant provenance <strong>of</strong> eolian clays<br />
and (3) using Fe d accumulation rate to infer ages <strong>of</strong><br />
sample sites in <strong>the</strong> catchment.<br />
ACKNOWLEDGEMENTS<br />
I’d like to thank my <strong>the</strong>sis advisor, David P. Dethier<br />
for his advice, support, and enthusiasm. Secondly,<br />
I’d like to thank my fellow Keck colleagues for <strong>the</strong>ir<br />
efforts in <strong>the</strong> field. I’d also like to thank all <strong>of</strong> those<br />
working in <strong>the</strong> Boulder Creek CZO who shared <strong>the</strong>ir<br />
knowledge, <strong>the</strong> Williams College Geosciences Department,<br />
<strong>the</strong> Sperry Fund, and <strong>the</strong> Keck Geology<br />
Consortium.<br />
REFERENCES<br />
Anderson, S.P., von Blanckenburg, F., and White,<br />
A.F., 2007, Physical and chemical controls on <strong>the</strong><br />
Critical Zone: Elements, v. 3, p. 315-319.<br />
Anderson, R.S., and Anderson, S.P., 2010, Geomorphology:<br />
<strong>the</strong> mechanics and chemistry <strong>of</strong> landscapes:<br />
Cambridge, Cambridge University Press,<br />
637 p.<br />
Birkeland, P.W., 1999, Soils and Geomorphology: New<br />
York, Oxford University Press, 430 p.<br />
Gee, G.W., and J.W. Bauder. 1986. Particle-size analysis,<br />
p. 383-411. In A. Klute (ed.) Methods <strong>of</strong> soil<br />
analysis. Part 1. 2nd ed. Agron. Monogr. 9. ASA<br />
and SSSA, Madison, WI.<br />
Mason, J.A., and Jacobs, P.M., 1998, Chemical and<br />
particle-size evidence for addition <strong>of</strong> fine dust to<br />
soils <strong>of</strong> <strong>the</strong> Midwestern United States: Geology, v.<br />
26, p. 1135-1138.<br />
McFadden, L.D., and Hendricks, D.M., 1985, Changes<br />
in <strong>the</strong> content and composition <strong>of</strong> pedogenic<br />
iron oxyhydroxides in a chronosequence <strong>of</strong> soils<br />
in Sou<strong>the</strong>rn California: Quaternary Research, v.<br />
23, p. 189-204.<br />
Muhs, D.R., and Benedict, J.B., 2006, Eolian additions<br />
to Late Quaternary alpine soils, Indian Peaks
24th Annual Keck Symposium: 2011 Union College, Schenectady, NY<br />
Wilderness Area, Colorado Front Range: Artic,<br />
Antartic, and Alpince Research, v. 38, p. 120-130.<br />
Muhs, D.R., Budahn, J.R., Johnson, D.L., Reheis, M.,<br />
Beann, J., Skipp, G., Fisher, E., and Jones, J.A.,<br />
2008, geochemical evidence for airborne dust additions<br />
to soils in Channel Islands National Park,<br />
California: Geological Society <strong>of</strong> America Bulletin,<br />
v. January/February.<br />
135
24th Annual Keck Symposium: 2011 Union College, Schenectady, NY<br />
USING POLLEN TO UNDERSTAND QUATERNARY<br />
PALEOENVIRONMENTS IN BETASSO GULCH, COLORADO<br />
COREY SHIRCLIFF, Beloit College<br />
Research Advisor: Carl Mendelson<br />
INTRODUCTION<br />
The mountain ranges <strong>of</strong> <strong>the</strong> western United States<br />
differ in climate and precipitation, and thus plant<br />
life varies as well. During <strong>the</strong> Pleistocene, oscillating<br />
climatic trends created sequences <strong>of</strong> glacial and<br />
interglacial times (Charlesworth, 1957); <strong>the</strong> Holocene<br />
is considered a modern interglacial epoch (Traverse,<br />
2007). By using palynological techniques, it is possible<br />
to understand how glacial climates have affected<br />
<strong>the</strong> environment. By documenting <strong>the</strong> current geographical<br />
trends <strong>of</strong> plant species, and <strong>the</strong>n applying<br />
that information to understanding Quaternary pollen<br />
samples, it is possible to reconstruct local environments<br />
over time.<br />
In <strong>the</strong> Rocky Mountains, <strong>the</strong> most recent glaciation is<br />
known as <strong>the</strong> Pinedale. In <strong>the</strong> Colorado Front Range,<br />
it occurred from about 30,000 YBP (years before<br />
present) to 15,000 to 12,000 years YBP (Legg and<br />
Baker, 1980). As <strong>the</strong> deglaciation occurred at higher<br />
elevations, <strong>the</strong> flora <strong>of</strong> <strong>the</strong> entire Front Range area<br />
responded. Certain floral taxa are important in determining<br />
<strong>the</strong> post-glacial environment, as discussed in<br />
detail later. One way to understand <strong>the</strong> paleoclimate<br />
<strong>of</strong> this dynamic time is by using palynological methods<br />
to examine <strong>the</strong> pollen in sediments that are known<br />
to be younger than 12,000 years in age.<br />
Betasso Gulch, located in <strong>the</strong> north-central Front<br />
Range <strong>of</strong> Colorado (Fig. 1), is a good place to collect<br />
samples for a palynological analysis due to <strong>the</strong> development<br />
<strong>of</strong> an organic-rich A soil horizon, which is<br />
buried under a modern soil pr<strong>of</strong>ile. Both radiocarbon<br />
and optically stimulated luminescence (OSL) dates<br />
are available for <strong>the</strong> soil horizons exposed in Betasso.<br />
A regional laterally continuous buried A horizon corresponds<br />
to a time between 9,000 and 6,000 years ago<br />
(see Fig. 2). This is an interesting interval to investigate<br />
palynologically, because <strong>the</strong> response <strong>of</strong><br />
<strong>the</strong> flora at this time and elevation (~ 2,000 m) is not<br />
well known in this area. Samples were taken from<br />
this buried A horizon and from modern forest litter<br />
to better understand how <strong>the</strong> climate has changed at<br />
Betasso Gulch.<br />
Figure 1. Betasso Gulch sample locality (red triangle).<br />
Previously studied sites: Devlins Park (Legg and Baker,<br />
1980) and Redrock Lake (Maher, 1972). Modified from<br />
Marr (1961) in Legg and Baker (1980, fig. 1).<br />
Although a wealth <strong>of</strong> pollen studies have been conducted<br />
in <strong>the</strong> Rocky Mountains and surrounding<br />
mountain ranges, two particularly relevant papers<br />
focus on pollen data from <strong>the</strong> Front Range, close to<br />
my sample site. Unfortunately, both localities are<br />
significantly higher in elevation (Fig. 1), but comparison<br />
with Betasso may still yield important information.<br />
Legg and Baker (1980) studied Pinedale-age<br />
lacustrine sediments from Lake Devlin; <strong>the</strong> sediments<br />
range in age from about 22,000 to 12,000 radiocarbon<br />
years before present (RCYBP). The authors<br />
found significant pollen contributions from Artemisia<br />
(sagebrush and relatives) at 42%, Poaceae (grasses)<br />
at 13%, and Cyperaceae (sedges) at 4%; Alnus (alder)<br />
and Betula (birch) appeared in small quantities at <strong>the</strong><br />
top <strong>of</strong> <strong>the</strong> section.<br />
136
24th Annual Keck Symposium: 2011 Union College, Schenectady, NY<br />
Figure 2. Dated stratigraphic section 50 m downstream<br />
from <strong>the</strong> Betasso Gulch locality. The buried soil A horizon<br />
is labeled IIIAb. Radiocarbon dates and image from<br />
Dethier (written commun., 2011).<br />
The top <strong>of</strong> <strong>the</strong> Lake Devlin section is dated at<br />
~12,000 RCYBP, or about 14,600 calibrated years,<br />
which is slightly older than <strong>the</strong> bottom <strong>of</strong> <strong>the</strong> stratigraphic<br />
section at Betasso Gulch. The o<strong>the</strong>r relevant<br />
study is <strong>of</strong> Redrock Lake, about nine kilometers north<br />
<strong>of</strong> Devlins Park and at a similar elevation (Maher,<br />
1972). Maher investigated <strong>the</strong> time from 10,000<br />
RCYBP to <strong>the</strong> modern. He found an increase <strong>of</strong><br />
Pinus (pine) and a decrease <strong>of</strong> Picea (spruce) upsection.<br />
Additionally, maxima for deposition <strong>of</strong> pollen<br />
grains occurred from 7,500 to 3,500 RCYBP and<br />
Picea deposition peaked around 8,500 RCYBP. He<br />
concluded that <strong>the</strong> interval from 6,700 to 7,600 YBP<br />
was cooler and/or wetter than earlier post-Pinedale<br />
times.<br />
MODERN VEGETATION AND CLIMATE<br />
It is possible to classify vegetation zones on <strong>the</strong><br />
eastern slope <strong>of</strong> <strong>the</strong> Front Range (Legg and Baker,<br />
1980; see Fig. 1). Beginning on <strong>the</strong> eastern plains<br />
is <strong>the</strong> grassland zone, which is below 1,800 meters<br />
137<br />
in elevation. Moving westward, on <strong>the</strong> lower slopes<br />
<strong>of</strong> <strong>the</strong> mountains, is <strong>the</strong> montane zone, which ranges<br />
from 1,800 to 2,800 meters; Betasso Gulch is located<br />
in this zone at an elevation <strong>of</strong> about 2000 m.<br />
Higher yet are <strong>the</strong> subalpine (2,800 to 3,350 m) and<br />
alpine (above 3,350 m) zones. The montane zone is<br />
characterized by an open forest populated by Pinus<br />
ponderosa (Ponderosa Pine), Pinus contorta (Lodgepole<br />
Pine), and Pseudotsuga menziesii (Douglas Fir),<br />
along with many shrubs, grasses, and herbs (Legg<br />
and Baker, 1980). This description is reflected in <strong>the</strong><br />
modern sample I collected (see Results).<br />
METHODS<br />
Field Methods<br />
Samples were collected at Gordon Gulch and Betasso<br />
Gulch. The Gordon Gulch samples, while rich in<br />
pollen, were not included in this paper due to time<br />
constraints in pollen identification and <strong>the</strong> fact that<br />
Gordon Gulch occurs at a higher elevation, and thus<br />
may have been more affected by recent glaciation.<br />
Samples at Betasso Gulch were collected at five-cm<br />
intervals from <strong>the</strong> base (1.2 m) to <strong>the</strong> top (1.05 m) <strong>of</strong><br />
<strong>the</strong> buried A horizon (Fig. 2, level IIIAb). The succeeding<br />
meter <strong>of</strong> sediment was not sampled due to<br />
bioturbation by roots, resulting in a mix <strong>of</strong> pollen<br />
grains from different levels. The sampled horizon<br />
was located in a channel, which had been eroded by a<br />
flood from pipes for a nearby water-treatment facility.<br />
A modern sample was taken at <strong>the</strong> surface, in an area<br />
at least two meters from any tree, and without significant<br />
undergrowth. Several handfuls <strong>of</strong> surface sediment<br />
were collected in an area <strong>of</strong> about one square<br />
meter, to help rid <strong>the</strong> sample <strong>of</strong> a bias toward grasses<br />
and shrubs growing on <strong>the</strong> surface. Therefore, five<br />
total samples were collected: four from <strong>the</strong> buried A<br />
horizon and one modern sample. All <strong>of</strong> <strong>the</strong>se samples<br />
were composed <strong>of</strong> dry sediment and organic debris.<br />
Laboratory Methods<br />
Pollen was extracted from <strong>the</strong> sediment and concentrated<br />
using a variety <strong>of</strong> chemical methods (see<br />
Traverse, 2007). After washing about 2 ml <strong>of</strong> sample<br />
in 5% KOH, <strong>the</strong> residue was filtered using a 250-mm<br />
screen to exclude rock fragments and large organic<br />
debris. The sample was <strong>the</strong>n washed in HCl to dis-
24th Annual Keck Symposium: 2011 Union College, Schenectady, NY<br />
solve carbonates, and washed again in hydr<strong>of</strong>luoric<br />
acid to dissolve silicates. After rinsing in acetic acid,<br />
residues were acetolyzed (nine parts acetic anhydride<br />
to one part sulfuric acid) in order to dissolve <strong>the</strong> part<br />
<strong>of</strong> <strong>the</strong> exine <strong>of</strong> <strong>the</strong> pollen grain that does not consist<br />
<strong>of</strong> <strong>the</strong> highly resistant sporopollenin. After diluting<br />
thoroughly in water, residues were washed through a<br />
10-mm screen. Remaining residues were successively<br />
dewatered with 95% ethanol and tert-butyl alcohol<br />
(TBA). Storage in TBA prevented <strong>the</strong> pollen grains<br />
from deflating or degrading. Slides were made using<br />
silicon oil as a mounting medium, and cover glasses<br />
were sealed with colored nail polish.<br />
Identification Methods<br />
It is necessary to identify 300 pollen grains from each<br />
sample in order to reach <strong>the</strong> 95% confidence interval<br />
for this kind <strong>of</strong> pollen analysis (Traverse, 2007).<br />
Pollen grains were identified using a Zeiss compound<br />
microscope under 40x and 100x (with oil) objectives.<br />
Identification was confirmed by referring to Kapp<br />
(1990) and to reference slides <strong>of</strong> modern pollen grains<br />
(courtesy <strong>of</strong> Robert Nelson, Colby College). Some<br />
grains were impossible to identify due to folding and<br />
exine degradation.<br />
RESULTS<br />
Modern<br />
Figure 3 shows selected pollen grains. High amounts<br />
<strong>of</strong> Pinus grains were identifiable; <strong>the</strong>y represented at<br />
least 67% <strong>of</strong> <strong>the</strong> Betasso sample. Almost all <strong>of</strong> <strong>the</strong><br />
unidentified pollen from <strong>the</strong> modern sample (9%) is<br />
arboreal; much <strong>of</strong> it probably belongs to Pinus, so<br />
67% is a minimum. Pseudotsuga menziesii is also an<br />
important arboreal component, at 3% <strong>of</strong> <strong>the</strong> modern<br />
sample. Juniperus (juniper) and Artemisia (sagebrush)<br />
toge<strong>the</strong>r represent <strong>the</strong> majority <strong>of</strong> <strong>the</strong> shrubs<br />
found in <strong>the</strong> Front Range area, and constitute 21% <strong>of</strong><br />
<strong>the</strong> Betasso modern sample. Grasses (Poaceae) are at<br />
a fairly average level (5%).<br />
Early Holocene<br />
Pinus, Artemisia, Poaceae, and Asteraceae (composites)<br />
contributed <strong>the</strong> most pollen grains to <strong>the</strong> early<br />
Holocene samples (Fig. 4). Progressing upward at<br />
138<br />
Figure 3. Light-microscope images <strong>of</strong> Betasso spores and<br />
pollen grains. A) unidentified spore found throughout <strong>the</strong><br />
buried A horizon. B) degraded Pseudotsuga menziesii<br />
(Douglas Fir) pollen grain. C) Pinus ponderosa pollen<br />
grain. D) Ambrosia artemesiifolia (common ragweed)<br />
pollen grain. (Ambrosia is a composite, and thus a member<br />
<strong>of</strong> <strong>the</strong> Asteraceae.)<br />
Betasso, <strong>the</strong>re is a slight increase <strong>of</strong> Pinus; in <strong>the</strong><br />
modern sample, Pinus is most common by far (67%).<br />
Artemisia (sagebrush and wormwood) pollen also<br />
represented a significant percent, but in <strong>the</strong> modern<br />
sample it dwindled to 18%. It was rarely possible to<br />
identify Artemisia grains to <strong>the</strong> species level. Poaceae<br />
and Asteraceae both showed decreasing trends<br />
(Fig. 4).<br />
Pseudotsuga menziesii (Douglas Fir) represented<br />
less than 1%, and in some cases 0%, in <strong>the</strong> buried A<br />
samples, but had a 3% representation in <strong>the</strong> modern<br />
sample. Because P. menziesii rarely rises above a 5%<br />
pollen representation in modern-day forests where it<br />
is known to grow in large quantities (Whitlock, 1993),<br />
I consider this result to be important.<br />
In <strong>the</strong> samples from <strong>the</strong> buried A horizon, many<br />
grains, particularly those <strong>of</strong> arboreal pollen types,<br />
were folded in a way that made identification impos-
24th Annual Keck Symposium: 2011 Union College, Schenectady, NY<br />
sible. They were counted, and a percent <strong>of</strong> unidentifiable<br />
grains was found. The most unidentified grains<br />
were in <strong>the</strong> middle <strong>of</strong> <strong>the</strong> buried A section, rising to<br />
almost 20% in one sample.<br />
Figure 4. Pollen percentage diagram (includes identified<br />
pollen only). Note <strong>the</strong> scale break between 0 and 1.05<br />
m. Dashed lines represent unsampled interval and are an<br />
estimate only <strong>of</strong> <strong>the</strong> trend <strong>of</strong> <strong>the</strong> pollen percents. The modern<br />
sample is plotted at 0 m. The percents <strong>of</strong> total pollen<br />
grains that could not be identified were as follows: 1.20 m<br />
(13%); 1.15 m (19%); 1.10 m (17%); 1.05 m (15%); 0 m<br />
(9%).<br />
DISCUSSION<br />
Pollen percentages are typically used to indicate<br />
changes in climate (Table 1). In modern-day forests,<br />
pine trees grow in relatively dry areas. Where precipitation<br />
increases, o<strong>the</strong>r arboreal species tend to<br />
win out in <strong>the</strong> battle for forest dominance. Pines in<br />
particular flourish in dry conditions, as seen in current-day<br />
Betasso Gulch, which hosts a pine-enriched<br />
forest; significant species include Pinus ponderosa<br />
and P. contorta. Because <strong>the</strong> percentages <strong>of</strong> P. ponderosa<br />
were lower in <strong>the</strong> buried A horizon, I infer that<br />
<strong>the</strong> climate at 9 ka was probably wetter and/or cooler<br />
than today’s climate. Today in <strong>the</strong> Front Range, P.<br />
ponderosa does not grow above <strong>the</strong> montane zone<br />
due to its inability to thrive in <strong>the</strong> cold temperatures<br />
experienced in <strong>the</strong> subalpine and alpine zones (Weber,<br />
1976). Therefore, because <strong>the</strong> P. ponderosa percent<br />
increases from <strong>the</strong> buried A section to <strong>the</strong> modern, <strong>the</strong><br />
environment may have become more favorable for<br />
Ponderosa Pines. Hence, it was becoming warmer<br />
139<br />
and drier from <strong>the</strong> buried A to present. The pollen<br />
percent <strong>of</strong> total Pinus species almost doubles in <strong>the</strong><br />
modern sample; this substantial leap suggests that <strong>the</strong><br />
climate has become considerably warmer and drier<br />
than <strong>the</strong> time represented at <strong>the</strong> top <strong>of</strong> <strong>the</strong> buried A<br />
section (1.05 m below <strong>the</strong> surface), that is, since <strong>the</strong><br />
early Holocene.<br />
Taxon Common Name Climate Implications<br />
(If pollen levels increase)<br />
Picea Spruce Moister<br />
Pseudotsuga menziesii Douglas Fir Drier<br />
Poaceae Grasses Drier<br />
Chenopodiaceae Herbs and Subshrubs Drier<br />
Pinus Pine Drier and warmer<br />
Artemisia Sagebrush Drier and warmer<br />
Table 1. Climate trends associated with increases <strong>of</strong> certain<br />
taxa in pollen samples.<br />
Whitlock (1993) concluded that a decrease in Artemisia<br />
indicated a wetter, cooler climate. She also<br />
argued that an increase in Picea suggested a wetter,<br />
cooler climate. In <strong>the</strong> surface sample, Picea decreases<br />
from 10% to 2% (indicating warming) and<br />
Artemisia levels decrease by 11% (indicating cooling).<br />
Although <strong>the</strong> decrease in Artemisia indicates a<br />
wetter and cooler environment, <strong>the</strong> decrease in Picea<br />
and significant increase in Pinus suggest a warmer<br />
environment. Moreover, although <strong>the</strong> Amaranthaceae<br />
+ Chenopodiaceae pollen group stays fairly constant<br />
in <strong>the</strong> buried A horizon, its percentage rises by 5% in<br />
<strong>the</strong> modern sample. Bright (1966) and Davis et al.<br />
(1986) inferred a drier climate when Amaranthaceae +<br />
Chenopodiaceae levels increased. Lastly, an increase<br />
in Pseudotsuga menziesii (Douglas Fir) has been<br />
shown to indicate warming (Whitlock, 1993). The<br />
increase from 0% to 3% from <strong>the</strong> buried A to modern<br />
gives fur<strong>the</strong>r evidence for warming, keeping in mind<br />
that 5% is <strong>the</strong> maximum percent P. menziesii reaches<br />
in any forest (Whitlock, 1993).<br />
I conclude that <strong>the</strong> environment at Betasso Gulch<br />
changed from a wetter, cooler climate (soil horizon<br />
A) to <strong>the</strong> current warmer and more arid conditions.<br />
Maher (1972) argued that <strong>the</strong> time from about 6,700<br />
to 7,600 YBP was cool and wet, compared to <strong>the</strong> preceding<br />
~3,000 years. Although evidence for cooling<br />
within <strong>the</strong> buried A horizon is weak, <strong>the</strong>re is a pos
24th Annual Keck Symposium: 2011 Union College, Schenectady, NY<br />
sibility that this section fits in slightly before 7,600<br />
YBP, or between <strong>the</strong>n and 6,700 YBP, and may thus<br />
be consistent with <strong>the</strong> findings <strong>of</strong> Maher (1972). Fur<strong>the</strong>r<br />
studies <strong>of</strong> this buried soil horizon should be undertaken<br />
before such a conclusion can be confirmed.<br />
My tentative conclusions are not statistically robust.<br />
For example, <strong>the</strong> percents reported for <strong>the</strong> buried<br />
A horizon are at best estimates because 15-20% <strong>of</strong><br />
<strong>the</strong> pollen grains could not be identified due to folding<br />
and degradation. In <strong>the</strong> modern sample, most<br />
unidentified pollen grains are bisaccate, so are probably<br />
Pinus species. Most significantly, <strong>the</strong> samples<br />
from <strong>the</strong> buried A horizon might not yield reliable<br />
information regarding trends in pollen percents—this<br />
horizon, like <strong>the</strong> succeeding meter <strong>of</strong> sediment, may<br />
have been bioturbated, resulting in homogenization<br />
<strong>of</strong> <strong>the</strong> micr<strong>of</strong>lora. For a more accurate understanding<br />
<strong>of</strong> climate change in this area, more samples should<br />
be taken from <strong>the</strong> buried A horizon, and additional<br />
surface samples need to be collected to confirm <strong>the</strong><br />
pollen percents representative <strong>of</strong> <strong>the</strong> current climate.<br />
It would also be beneficial to collect older samples in<br />
<strong>the</strong> section. Finally, <strong>the</strong> discrepancy between <strong>the</strong> radiocarbon<br />
and OSL ages (Fig. 2) needs to be resolved.<br />
ACKNOWLEDGMENTS<br />
I would like to pr<strong>of</strong>usely thank Dr. Bob Nelson <strong>of</strong><br />
Colby College for his tremendous assistance in this<br />
project. I would also like to thank my advisor at Beloit<br />
College, Dr. Carl Mendelson, for his mentorship<br />
and bravery in <strong>the</strong> face <strong>of</strong> a student using HF, and<br />
my advisor Dr. Jim Rougvie, for taking over advising<br />
when Carl went on sabbatical. Thanks are also in order<br />
to Dr. Carol Mankiewicz for assistance with figure<br />
preparation. Lastly, I thank both Dr. David Dethier<br />
(Williams College) and Dr. Will Ouimet (University<br />
<strong>of</strong> Connecticut) for <strong>the</strong>ir mentorship this summer and<br />
throughout <strong>the</strong> year.<br />
REFERENCES<br />
Bright, R.C., 1966, Pollen and seed stratigraphy <strong>of</strong><br />
Swan Lake, south-eastern Idaho; its relation to<br />
regional vegetational history and to Lake Bonn-<br />
140<br />
eville history: Tebiwa, v. 9, p. 1-47.<br />
Charlesworth, J.K., 1957, The Quaternary Era: with<br />
special reference to its glaciation: London, E.<br />
Arnold, 1601 p.<br />
Davis, O. K., Sheppard, J. C., and Robertson,<br />
S., 1986, Contrasting climatic histories<br />
for <strong>the</strong> Snake River Plain, Idaho, resulting from<br />
multiple <strong>the</strong>rmal maxima: Quaternary Research,<br />
v. 26, p. 321-339.<br />
Kapp, R. O., 2000, Pollen and spores (2d ed.): College<br />
Station, Texas, American Association <strong>of</strong><br />
Stratigraphic Palynologists, 279 p.<br />
Legg, T. E., and Baker, R. G., 1980, Palynology<br />
<strong>of</strong> Pinedale sediments, Devlins Park, Boulder<br />
County, Colorado: Arctic and Alpine Research,<br />
v. 12, no. 3, p. 319-333.<br />
Maher, L.J., Jr., 1972, Absolute pollen diagram <strong>of</strong> Redrock<br />
Lake, Boulder County, Colorado: Quaternary<br />
Research, v. 2, no. 4, p. 531-553.<br />
Traverse, A., 2007, Paleopalynology (2d ed.): Boston,<br />
Unwin Hyman, 600 p.<br />
Weber, W. A., 1976, Rocky Mountain flora (5th ed.):<br />
Niwot, University Press <strong>of</strong> Colorado, 479 p.<br />
Whitlock, C., 1993, Postglacial vegetation and climate<br />
<strong>of</strong> Grand Teton and sou<strong>the</strong>rn Yellowstone<br />
National Parks: Ecological Monographs, v. 63,<br />
p. 173-198.
24th Annual Keck Symposium: 2011 Union College, Schenectady, NY<br />
STREAM TERRACES IN THE CRITICAL ZONE – LOWER<br />
GORDON GULCH, COLORADO<br />
KATHLEEN WARRELL, Georgia Tech<br />
Research Advisor: Kurt Frankel<br />
INTRODUCTION<br />
Systems <strong>of</strong> stream terraces provide insight into <strong>the</strong><br />
history <strong>of</strong> a stream and how <strong>the</strong> surrounding landscape<br />
has changed throughout geologic history.<br />
Stream terraces are an integral part <strong>of</strong> <strong>the</strong> Critical<br />
Zone (CZ), which is defined as <strong>the</strong> boundary layer<br />
that extends from <strong>the</strong> buried, unwea<strong>the</strong>red bedrock up<br />
through wea<strong>the</strong>red rock and regolith to <strong>the</strong> soil where<br />
terrestrial life thrives (Anderson et al., 2007). The<br />
CZ is <strong>the</strong> vital place on Earth’s surface where rocks,<br />
soil, atmospheric gasses, and meteoric water interact.<br />
Anderson et al. (2007) described <strong>the</strong> CZ as a “feedthrough<br />
reactor” that transforms solid bedrock into<br />
soil and sediment, which is <strong>the</strong>n transported downslope<br />
into a stream channel.<br />
The morphology <strong>of</strong> a stream and its floodplain is <strong>the</strong><br />
result <strong>of</strong> a delicate balance <strong>of</strong> driving and resisting<br />
forces. Excess erosion on surrounding hillslopes can<br />
cause aggradation and increase <strong>the</strong> stream elevation.<br />
Aggraded sediment is removed when it is entrained<br />
by <strong>the</strong> stream. Sediment entrainment and deposition<br />
by a stream are driven by <strong>the</strong> depth and slope <strong>of</strong> that<br />
stream; <strong>the</strong>y are resisted by channel configuration,<br />
sediment size and sediment concentration (Ritter et<br />
al., 2002).<br />
Fill terraces are especially important in <strong>the</strong> CZ because<br />
<strong>the</strong>y store sediment and biomass eroded from<br />
surrounding hillslopes. Fill terraces are extremely<br />
productive areas in a stream valley, as <strong>the</strong>y provide<br />
a stable, flat environment with organic-rich soil on<br />
which plants and animals thrive. However, <strong>the</strong>se<br />
terraces are only temporary features in many landscapes,<br />
as stream incision and sediment entrainment<br />
are constantly removing sediment from <strong>the</strong> terraces.<br />
This study uses terrace morphology <strong>of</strong> Lower Gordon<br />
Gulch to estimate <strong>the</strong> volume <strong>of</strong> sediment stored in<br />
<strong>the</strong>se terraces and to model <strong>the</strong> timescale to remove<br />
141<br />
all <strong>of</strong> this sediment from <strong>the</strong> Gulch.<br />
GEOGRAPHIC AND GEOLOGIC SETTING<br />
The study area for this project is <strong>the</strong> 3.76 square kilometer<br />
Gordon Gulch catchment in Boulder County,<br />
Colorado. Gordon Gulch is a tributary <strong>of</strong> North<br />
Boulder Creek; it joins North Boulder Creek about 16<br />
kilometers from its headwaters. Elevations in Gordon<br />
Gulch range from 2,400 meters to 2,700 meters.<br />
Gordon Gulch is separated informally into two sections<br />
– Lower Gordon Gulch and a large tributary that<br />
constitutes Upper Gordon Gulch. A large knickpoint<br />
lies between Lower and Upper Gordon Gulch (Fig. 1).<br />
The stream in Upper Gordon Gulch is intermittent;<br />
however <strong>the</strong> majority <strong>of</strong> <strong>the</strong> stream in Lower Gordon<br />
Gulch contains water in most years.<br />
Figure 1. Map <strong>of</strong> Gordon Gulch showing location in<br />
Boulder County and Colorado. Map <strong>of</strong> Gordon Gulch is<br />
a hillshade derived from lidar flown in August 2010 with<br />
a pixel size <strong>of</strong> 1 m 2 . The start and end <strong>of</strong> <strong>the</strong> terrace map<br />
section are noted.
24th Annual Keck Symposium: 2011 Union College, Schenectady, NY<br />
Stream aggradation is sensitive to local changes in<br />
land use. During <strong>the</strong> late 1800s to early 1900s as<br />
miners began working in <strong>the</strong> area surrounding Gordon<br />
Gulch, land use changed drastically. The introduction<br />
<strong>of</strong> prospecting and small-scale mining may have<br />
generated a large amount <strong>of</strong> sediment that aggraded in<br />
Gordon Gulch. The frequency <strong>of</strong> fires also increased<br />
at this time, resulting in an increase in erosion in <strong>the</strong><br />
catchment (Goldblum and Veblen, 1992).<br />
METHODS<br />
A Laser Technology Tru-Pulse 360 laser rangefinder<br />
was used to produce a detailed base map <strong>of</strong> <strong>the</strong><br />
stream. The rangefinder has an accuracy <strong>of</strong> ±0.20 meters<br />
slope distance, ±0.25 degrees slope angle, and ±1<br />
degree azimuthal angle. Azimuthal angle was used in<br />
conjunction with horizontal distance measurements<br />
to produce x and y coordinates. The z coordinate was<br />
calculated using a base-level measurement from a<br />
GPS and cumulative vertical distance measurements.<br />
These coordinates were graphed in Matlab with equal<br />
axes to produce a base map for mapping terraces.<br />
Stream morphology and a series <strong>of</strong> flags placed along<br />
<strong>the</strong> stream were used to mark terraces on <strong>the</strong> map<br />
relative to <strong>the</strong>ir location along <strong>the</strong> stream. Terraces<br />
were differentiated based upon <strong>the</strong>ir morphology and<br />
height relative to surrounding terraces and hillslopes.<br />
The rangefinder was used to measure <strong>the</strong> height <strong>of</strong><br />
each terrace above <strong>the</strong> stream channel. Seventy-five<br />
tree core ages were collected from trees growing<br />
on <strong>the</strong> terraces to approximate <strong>the</strong> age each terrace<br />
stabilized. Two samples <strong>of</strong> buried wood were also<br />
collected from <strong>the</strong> terraces for 14 C dating.<br />
A series <strong>of</strong> eight detailed cross sections were measured<br />
along <strong>the</strong> stream using <strong>the</strong> rangefinder (Fig.<br />
2C). Valley-wide cross sections were extracted from<br />
a high resolution digital elevation model to estimate<br />
<strong>the</strong> slope <strong>of</strong> <strong>the</strong> bedrock in surrounding hillslopes<br />
(Fig. 2B). Riemann sums were used to calculate cross<br />
sectional area <strong>of</strong> sediment between <strong>the</strong> bedrock slope<br />
and terrace cross section (Fig. 2C). The area was<br />
multiplied by <strong>the</strong> distance upstream to <strong>the</strong> next cross<br />
section, and all volumes were summed to obtain <strong>the</strong><br />
total volume <strong>of</strong> sediment stored in <strong>the</strong> terraces (V s ).<br />
142<br />
Figure 2. (A) Map view <strong>of</strong> terraces at KW-ST-10 with<br />
cross section X-X’ marked. Tors (Qt) are shown. (B)<br />
Valley-wide cross section derived from lidar showing<br />
estimated slope angles <strong>of</strong> <strong>the</strong> bedrock boundary. (C) Cross<br />
section and map view <strong>of</strong> stream terrace map at location<br />
KW-ST-10 showing terraces Qt4 and Qt5 and <strong>the</strong> cross<br />
sectional area <strong>of</strong> sediment.<br />
The cross sections were also used to determine bankfull<br />
width (B) and hydraulic radius (R) <strong>of</strong> <strong>the</strong> active<br />
stream. Hydraulic radius is:<br />
where:<br />
R = HB[2H + B] -1 ,<br />
H = Q w B -1 v -1 ,<br />
where H is bankfull depth, Q w is maximum discharge<br />
<strong>of</strong> water in <strong>the</strong> stream, and v is velocity <strong>of</strong> <strong>the</strong> stream.<br />
Bankfull velocity was approximated at 0.5 m/s.<br />
Maximum discharge <strong>of</strong> water from Gordon Gulch<br />
was calculated as <strong>the</strong> 90 th percentile <strong>of</strong> daily discharge<br />
data from <strong>the</strong> stream gauge over <strong>the</strong> year 2009. At<br />
each cross section location a sample <strong>of</strong> stream sediment<br />
was collected. The 84 th percentile grain size
24th Annual Keck Symposium: 2011 Union College, Schenectady, NY<br />
(D 84 , in meters) <strong>of</strong> each sample was determined by<br />
sieving <strong>the</strong> samples.<br />
These data were used to determine <strong>the</strong> maximum<br />
sediment flux (Q s ) <strong>of</strong> <strong>the</strong> stream, which is adapted<br />
from Mueller and Pitlick (2005) as:<br />
Q = 11.2(θ-θ ) s c 4.5θ -3 3 0.5 [(s-1)gD ] B 84<br />
where θ is Shields stress, θ c is critical Shields stress<br />
(approximated at 0.03), s is specific gravity <strong>of</strong> sediment,<br />
and g is gravity. Shields stress is defined as:<br />
θ = τ[(ρ s -ρ w )gD 84 ] -1 ,<br />
where τ is shear stress (τ = ρ w gRS), ρ s and ρ w are sediment<br />
and water densities, g is gravity, and S is decimal<br />
slope (Cronin et al., 2007). Slope was measured<br />
from <strong>the</strong> base map at each sample location. When θ<br />
is equal to 0.03, <strong>the</strong> stream is capable <strong>of</strong> transporting<br />
all <strong>the</strong> sediment in its bed. Below this value, <strong>the</strong><br />
stream cannot transport all <strong>of</strong> its sediment. When θ<br />
is above 0.07 <strong>the</strong> stream is capable <strong>of</strong> carrying more<br />
sediment than its bed contains.<br />
The sediment removal time-scale (T s ) for <strong>the</strong> valley is:<br />
143<br />
T s = VsQs -1 t,<br />
Sheet1<br />
Terrace h min h max area n units age min n cores<br />
Qt1 2.2 3.3 97 2 83 1<br />
Qt2 1.2 2.1 908 8 134 6<br />
Qt3 0.9 1.7 2751 33 158 18<br />
Qt4 0.4 1.2 3043 92 162 33<br />
Qt5 0.1 0.7 1465 164 120 7<br />
Table 1. Characterization <strong>of</strong> Gordon Gulch stream terraces,<br />
with Qt1 being <strong>the</strong> oldest and Qt5 being <strong>the</strong> current<br />
floodplain. h min and h max are minimum and maximum<br />
heights <strong>of</strong> terraces above <strong>the</strong> stream channel in meters,<br />
area is total area <strong>of</strong> all units in square meters, n units is number<br />
<strong>of</strong> units mapped for each terrace, age min is minimum<br />
age obtained from tree coring in years, n cores is number <strong>of</strong><br />
tree cores obtained for each terrace.<br />
where t is a unitless time interval, defined as:<br />
t = [total years <strong>of</strong> Q w data] / [total years <strong>of</strong> Q w<br />
exceeding 90th percentile].<br />
This calculation assumes stream flow patterns remain<br />
constant over thousand year timescales.<br />
Figure 3. Stream terrace map near location KW-ST-05 (red dot). Terraces range from Qt1 (oldest) to Qt5 (youngest). Alluvial<br />
fan units (Qfa) are visible. Tor deposits (Qt) are not visible.<br />
Page 1
24th Annual Keck Symposium: 2011 Union College, Schenectady, NY<br />
RESULTS<br />
Sheet1<br />
Sample ID Dist D 84 S H R B τ θ Q s<br />
KW-ST-04 0.00 0.002 0.08 0.09 0.075 0.9 59 1.80 710<br />
KW-ST-06 0.19 0.020 0.08 0.10 0.081 0.8 63 0.20 360<br />
KW-ST-05 0.386 0.002 0.08 0.20 0.100 0.4 79 2.40 500<br />
KW-ST-01 0.64 0.020 0.09 0.27 0.096 0.3 85 0.26 260<br />
KW-ST-07 0.771 0.010 0.09 0.09 0.075 0.9 66 0.41 650<br />
KW-ST-08 0.993 0.020 0.09 0.17 0.098 0.5 87 0.27 450<br />
KW-ST-09 1.266 0.020 0.11 0.27 0.096 0.3 100 0.32 390<br />
KW-ST-10 1.414 0.020 0.10 0.07 0.061 1.2 60 0.18 470<br />
Median 0.29 460<br />
Table 2. Shields stress and sediment flux for Gordon<br />
Gulch. Dist. is distance upstream from <strong>the</strong> beginning <strong>of</strong><br />
<strong>the</strong> mapped section in meters, D 84 is grain size in meters,<br />
S is decimal slope, H is bank-full depth in meters, R is<br />
hydraulic radius in meters, B is bank-full width in meters,<br />
τ is shear stress in Newtons per square meter, θ is Shields<br />
stress (unitless), Q s is modeled sediment flux in cubic meters<br />
per 12 year cycle.<br />
Terraces along 1.6 km <strong>of</strong> Lower Gordon Gulch were<br />
characterized into five distinct levels, which are listed<br />
in Table 1. Terrace Qt5 is <strong>the</strong> current floodplain and<br />
was vegetated by mostly grasses and young plants.<br />
There was no discernible difference in vegetation on<br />
terraces Qt1 through Qt4. Figure 3 shows a section <strong>of</strong><br />
Lower Gordon Gulch in which all five terrace levels<br />
interact with alluvial fans (Qa).<br />
Page 1<br />
Morphology <strong>of</strong> terraces in Gordon Gulch varies along<br />
<strong>the</strong> stream. Downstream, <strong>the</strong>re are more terraces<br />
flanking <strong>the</strong> stream in complex patterns. The majority<br />
<strong>of</strong> terraces are not paired. The north bank <strong>of</strong> <strong>the</strong><br />
stream <strong>of</strong>ten contains few or no terraces. Terraces on<br />
<strong>the</strong> south bank are more extensive. In some locations<br />
(Fig. 3) it is possible to find all five terraces in<br />
one location. Upstream <strong>the</strong>re may be only one or two<br />
terraces flanking <strong>the</strong> stream (Fig. 2A). No bedrock is<br />
visible in <strong>the</strong> mapped stream channel. Tor deposits<br />
are more common upstream. The overall width <strong>of</strong><br />
upstream terraces is half that <strong>of</strong> downstream terraces.<br />
The total volume <strong>of</strong> sediment stored in <strong>the</strong> terraces <strong>of</strong><br />
lower Gordon Gulch (V s ) was calculated to be 50,000<br />
m 3 in <strong>the</strong> mapped 1.6 km <strong>of</strong> <strong>the</strong> stream (Table 2).<br />
The time interval t was calculated using discharge<br />
data from Boulder Creek over <strong>the</strong> past 24 years provided<br />
by <strong>the</strong> US Geological Survey. Of <strong>the</strong> 24 years<br />
<strong>of</strong> data, two years <strong>of</strong> maximum discharge values exceeded<br />
<strong>the</strong> 90th percentile <strong>of</strong> <strong>the</strong> Boulder Creek data.<br />
144<br />
Thus, <strong>the</strong> time interval between maximum discharge<br />
events in <strong>the</strong> catchment is 12 years.<br />
Parameters for <strong>the</strong> calculation <strong>of</strong> Q s are listed in<br />
Table 2. For <strong>the</strong> 90th percentile discharge (3,500 m 3 /<br />
day from Boulder Creek CZO stream gauge data),<br />
<strong>the</strong> median θ value was 0.29, more than sufficient to<br />
mobilize terrace sediment. The θ values were used to<br />
calculate sediment flux at each sample location, with<br />
<strong>the</strong> median sediment discharge value being 460 m 3<br />
<strong>of</strong> sediment transported from Gordon Gulch every 12<br />
years. Median values were used to avoid sensitivity<br />
to outliers.<br />
At <strong>the</strong> current rates <strong>of</strong> water and sediment discharge,<br />
this model estimates that it would take 1,300 years to<br />
evacuate <strong>the</strong> sediment currently in <strong>the</strong> basin.<br />
Two radiometric 14 C dates were obtained from buried<br />
wood in terrace sediments. The first sample was 30<br />
cm above <strong>the</strong> current stream channel and was dated<br />
1,110 ± 50 years before present. The second sample<br />
was 10 cm above <strong>the</strong> current stream channel and was<br />
dated 1,520 ± 40 years before present.<br />
DISCUSSION<br />
As streams go through periods <strong>of</strong> aggradation and<br />
degradation a complex system <strong>of</strong> terraces may form.<br />
In Gordon Gulch, five terraces have formed from this<br />
process.<br />
Terrace morphology <strong>of</strong> Gordon Gulch<br />
Variations in terrace morphology along Gordon Gulch<br />
can be attributed to valley morphology. Water downstream<br />
carries more sediment, as drainage area is directly<br />
related to distance from <strong>the</strong> headwaters. Thus,<br />
more sediment is carried into <strong>the</strong> stream by erosional<br />
processes. Increased sediment is counteracted by<br />
decreased slope <strong>of</strong> <strong>the</strong> stream. The combined effect<br />
<strong>of</strong> <strong>the</strong>se factors is that a larger amount <strong>of</strong> sediment<br />
accumulated in downstream terraces versus upstream<br />
terraces. Terrace sediments in Qt1 through Qt4 were<br />
accumulated in <strong>the</strong> past 2,000 years and are currently<br />
being incised into. Qt5 may be <strong>the</strong> result <strong>of</strong> a combination<br />
<strong>of</strong> current accumulation and incision into past<br />
accumulation.
24th Annual Keck Symposium: 2011 Union College, Schenectady, NY<br />
Ages obtained from tree coring are largely varied and<br />
do not accurately reflect terrace ages. This may be<br />
<strong>the</strong> result <strong>of</strong> logging and forest fires that cleared many<br />
<strong>of</strong> <strong>the</strong> trees in <strong>the</strong> past 200 years. The oldest tree was<br />
a 162 year old Ponderosa Pine on Qt4. Thus, terraces<br />
Qt1 through Qt4 stabilized at least 162 years ago.<br />
Sediment removal from Gordon Gulch<br />
Shields stress (θ) values for maximum stream discharge<br />
in Gordon Gulch have a large variation. Maximum<br />
stream discharge is more than sufficient (θ ><br />
0.07) in all sample locations to transport all sediment<br />
in <strong>the</strong> stream. In two locations (KW-ST-04 and KW-<br />
ST-05), <strong>the</strong> θ value for maximum stream discharge is<br />
very high (θ > 1.5) due to decreased grain size. These<br />
locations also have a high Q s value. In locations<br />
with a large D 84 grain size (D 84 ≥ 20 mm), θ values<br />
were below 0.40 and Q s values were below or near<br />
<strong>the</strong> median Q s value. Increased slope also resulted in<br />
increased θ and Q s values. Grain size appears to have<br />
<strong>the</strong> largest control on θ and Q s values<br />
The sediment removal timescale for terraces in Gordon<br />
Gulch calculated by this model is 1,300 years.<br />
Evacuating <strong>the</strong> sediment in this timescale would be<br />
unlikely. The model does not take into account forces<br />
holding sediments toge<strong>the</strong>r, which include roots, buried<br />
logs and o<strong>the</strong>r biologic factors, as well as compaction<br />
forces <strong>of</strong> buried sediments. The model also does<br />
not account for sediment currently being added to <strong>the</strong><br />
stream by erosion on hillslopes and from addition <strong>of</strong><br />
sediment upstream <strong>of</strong> <strong>the</strong> mapped area. Incorporating<br />
<strong>the</strong>se factors into <strong>the</strong> model would likely increase <strong>the</strong><br />
sediment removal timescale.<br />
CONCLUSION<br />
The Gordon Gulch terrace system includes five complex<br />
terrace levels that are closely related to valley<br />
morphology. Downstream terraces are wider and<br />
more complex due to aggradation from increased<br />
sediment concentration and decreased slope. Sediment<br />
stored in terraces has been accumulating for<br />
over 2,000 years. Total volume <strong>of</strong> sediment stored<br />
in <strong>the</strong> terraces was approximated to be 50,000 cubic<br />
meters. Hydrologic models applied to calculate sedi-<br />
145<br />
ment flux estimate that it would take 1,300 years to<br />
evacuate terrace sediment from Gordon Gulch. This<br />
value underestimates <strong>the</strong> time it will take to remove<br />
sediment stored in <strong>the</strong> terraces, largely because <strong>the</strong><br />
model does not take into account biologic factors and<br />
erosional input from <strong>the</strong> headwaters and hillslopes.<br />
Future research should focus on quantifying inputs<br />
<strong>of</strong> sediment into <strong>the</strong> stream by erosion on hillslopes<br />
and upstream <strong>of</strong> <strong>the</strong> mapped area. Incorporating<br />
<strong>the</strong>se factors into <strong>the</strong> model would provide a closer<br />
approximation <strong>of</strong> <strong>the</strong> sediment removal timescale.<br />
Future research should also quantify <strong>the</strong> effects <strong>of</strong><br />
biologic factors and compaction on erosion <strong>of</strong> terrace<br />
sediments. Understanding <strong>the</strong>se factors would also<br />
provide better understanding <strong>of</strong> how <strong>the</strong> complex relationships<br />
<strong>of</strong> <strong>the</strong> CZ affect sediment flux. Volume <strong>of</strong><br />
sediment should be better estimated using geophysical<br />
methods (ground penetrating radar) to measure <strong>the</strong><br />
depth to bedrock below <strong>the</strong> terraces.<br />
ACKNOWLEDGMENTS<br />
I thank David Dethier, Will Ouimet, Kurt Frankel,<br />
Corey Shircliff, Erin Camp, Reece Lyerly, Cianna<br />
Wyshnytzky, Hayley Corson-Rikert, and Ellie Maley<br />
for <strong>the</strong>ir help.<br />
REFERENCES<br />
Anderson, S. P., F. von Blanckenburg, and A. F. White<br />
(2007), Physical and chemical controls on <strong>the</strong> Critical<br />
Zone, Elements, 3(5), 315-319.<br />
Cronin, G., J. H. McCutchan, J. Pitlick, and W. M.<br />
Lewis (2007), Use <strong>of</strong> Shields stress to reconstruct and<br />
forecast changes in river metabolism, Freshw. Biol.,<br />
52(8), 1587-1601.<br />
Goldblum, D., and T. T. Veblen (1992), Fire history <strong>of</strong><br />
a Ponderosa pine Douglas-fir forest in <strong>the</strong> Colorado<br />
Front Range, Physical Geography, 13(2), 133-148.<br />
Mueller, E. R., and J. Pitlick (2005), Morphologically<br />
based model <strong>of</strong> bed load transport capacity in a headwater<br />
stream, Journal <strong>of</strong> Geophysical Research-Earth<br />
Surface, 110(F2).
24th Annual Keck Symposium: 2011 Union College, Schenectady, NY<br />
METEORIC 10 BE IN GORDON GULCH SOILS: IMPLICATIONS<br />
FOR HILLSLOPE PROCESSES AND DEVELOPMENT<br />
CIANNA E. WYSHNYTZKY, Amherst College<br />
Research Advisors: Will Ouimet and Peter Crowley<br />
INTRODUCTION<br />
Is some agent <strong>of</strong> tectonism necessary to explain <strong>the</strong><br />
dramatic relief <strong>of</strong> <strong>the</strong> Front Range or are isostatic<br />
responses to erosion induced by climate change sufficient?<br />
Could a post-Eocene change in climate and<br />
integration <strong>of</strong> channel systems alone allow for increased<br />
incision to dominate this landscape and create<br />
<strong>the</strong> high relief <strong>of</strong> deeply incised canyons and large<br />
stream valleys, or is some tectonism necessary to account<br />
for <strong>the</strong> observed amount <strong>of</strong> warping, tilting, and<br />
erosion? Understanding <strong>the</strong> evolution <strong>of</strong> Front Range<br />
hillslopes in relation to late Cenozoic climatic and<br />
tectonic evolution, hillslope processes, and fundamental<br />
critical zone principles could provide a more<br />
thorough understanding <strong>of</strong> <strong>the</strong> modern geomorphology<br />
<strong>of</strong> <strong>the</strong> region.<br />
This research used <strong>the</strong> accumulation <strong>of</strong> meteoric 10 Be<br />
to determine <strong>the</strong> age <strong>of</strong> soils on hillslopes in Gordon<br />
Gulch and helps constrain interpretations <strong>of</strong> regolith<br />
transport, extending current information about <strong>the</strong><br />
recent evolution <strong>of</strong> Colorado’s Front Range. This<br />
research contributes to research done by <strong>the</strong> Boulder<br />
Creek Critical Zone Observatory (BC-CZO) within<br />
<strong>the</strong> extents <strong>of</strong> <strong>the</strong>ir focus area in <strong>the</strong> Front Range and<br />
complements ongoing research in Gordon Gulch.<br />
This is <strong>the</strong> first project in <strong>the</strong> region using LiDAR<br />
analysis and meteoric 10 Be as a tracer <strong>of</strong> modern hillslope<br />
evolution.<br />
Due to its adherence to sediment within soils and its<br />
constant rate <strong>of</strong> production in <strong>the</strong> atmosphere, soil<br />
age can be constrained by using <strong>the</strong> inventory <strong>of</strong> total<br />
meteoric 10 Be <strong>of</strong> a soil pr<strong>of</strong>ile. Erosion rates and<br />
estimated soil transport rates can <strong>the</strong>n be quantified<br />
(Jungers et al., 2009; Graly et al., 2010; Willenbring<br />
and von Blackenburg, 2010). Studies in North Carolina<br />
(Jungers, et al., 2009) and Australia (Fifield, et al.,<br />
2010) have traced hillslope sediment production and<br />
146<br />
transport using meteoric 10 Be, after which <strong>the</strong> <strong>the</strong>oretical<br />
framework <strong>of</strong> this research has been modeled.<br />
Graly et al. (2010) have shown that <strong>the</strong> concentration<br />
<strong>of</strong> meteoric 10 Be in soil pr<strong>of</strong>iles typically conforms<br />
to one <strong>of</strong> three general pr<strong>of</strong>ile shapes: exponentially<br />
declining, bulge, and bulge/declining (small bulge<br />
towards <strong>the</strong> top <strong>of</strong> <strong>the</strong> pr<strong>of</strong>ile). As a soil pr<strong>of</strong>ile<br />
evolves, so does its meteoric 10 Be inventory due to<br />
soil formation and mixing processes. Given a steady<br />
state hillslope, <strong>the</strong> peak concentration <strong>of</strong> meteoric<br />
10 Be is expected in one horizon (Jungers et al., 2009).<br />
Concentration <strong>the</strong>n decreases with depth, and <strong>the</strong><br />
inventory is expected to increase downslope, creating<br />
a bulge pr<strong>of</strong>ile. Given a young and eroding hillslope<br />
pr<strong>of</strong>ile, <strong>the</strong> highest concentration <strong>of</strong> meteoric 10 Be will<br />
still be in a single layer, but erosion prevents this concentration<br />
from moving to depths beyond near-surface<br />
(Graly et al., 2010).<br />
GEOGRAPHIC AND GEOLOGIC SETTING<br />
Gordon Gulch is a focus area <strong>of</strong> <strong>the</strong> BC-CZO located<br />
below and east <strong>of</strong> <strong>the</strong> modern alpine environment and<br />
late Pleistocene glacial limit and generally above and<br />
west <strong>of</strong> <strong>the</strong> deeply incised landscape that characterizes<br />
<strong>the</strong> lower portion <strong>of</strong> Front Range rivers. The<br />
degree to which <strong>the</strong> drainage basin may be affected<br />
by upstream-migrating rejuvenation from <strong>the</strong> lower<br />
portion <strong>of</strong> <strong>the</strong> range and/or alpine environmental processes<br />
(i.e. periglacial activity) is debated (Anderson<br />
et al., 2006). Gordon Gulch is a 2.75 km 2 catchment<br />
with exposed bedrock in various places. It can be<br />
subdivided into two primary floral and spatial environments:<br />
<strong>the</strong> north- and south-facing hillslopes (Fig.<br />
1).<br />
METHODS<br />
Sample Collection and Transect Selection<br />
Hillslope transects were chosen using a combination
24th Annual Keck Symposium: 2011 Union College, Schenectady, NY<br />
Figure 1. Gordon Gulch. a) air photo (1m B&W DOQ,<br />
1999), b) shaded relief map, c) slope map. (b and c derived<br />
from ~1m resolution LiDAR data). Blue dots indicates<br />
sample pit locations.<br />
<strong>of</strong> topographic pr<strong>of</strong>ile and field observations <strong>of</strong> <strong>the</strong><br />
character <strong>of</strong> hillslopes at proposed soil pit locations.<br />
Areas containing evidence <strong>of</strong> recent fire were avoided<br />
to not incorporate fire as an agent <strong>of</strong> erosion, and<br />
gullies, bedrock outcrops, trails/roads, and miners’<br />
pits were avoided to obtain <strong>the</strong> most representative<br />
and continuous downslope path from ridge to stream.<br />
These criteria affected transect selection so that <strong>the</strong><br />
smoo<strong>the</strong>st pr<strong>of</strong>ile on each hillslope was found to act<br />
as a proxy for overall hillslopes <strong>of</strong> <strong>the</strong> area. Samples<br />
from 9 hillslope pits (5 on <strong>the</strong> north-facing hillslope,<br />
4 on <strong>the</strong> south-facing hillslope) were collected in 10<br />
147<br />
cm intervals from 1-2 cm thick slots, beginning below<br />
<strong>the</strong> O-horizon (Fig. 1). Each soil pr<strong>of</strong>ile was photographed<br />
and individual horizons were described in<br />
detail.<br />
Meteoric 10 Be<br />
Samples were dried to remove excess moisture and<br />
hand sieved through a wire-mesh 2 mm sieve to<br />
remove coarse particles, since meteoric 10 Be binds<br />
to particles with high surface area (Fifield, 2010).<br />
Sieved samples were ground into fine powders using<br />
a tungsten carbide shatterbox triplet.<br />
Beryllium was extracted from samples at <strong>the</strong> Cosmogenic<br />
Nuclide Laboratory at <strong>the</strong> University <strong>of</strong><br />
Vermont from ~0.5 g aliquots by ion exchange acid<br />
elutions in a method adapted from <strong>the</strong> flux fusion<br />
methods originally presented by Stone (1998). Atoms<br />
<strong>of</strong> 10 Be were counted using an accelerator mass<br />
spectrometer (AMS) by <strong>the</strong> GeoCAMS group at <strong>the</strong><br />
Center for Accelerator Mass Spectromery (CAMS) at<br />
<strong>the</strong> Livermore National Laboratory in California.<br />
In addition to <strong>the</strong> analysis <strong>of</strong> 10 Be on hillslopes <strong>of</strong><br />
unknown age, 10 Be analysis <strong>of</strong> samples from stable<br />
reference sites (with known ages) were run to quantify<br />
how much 10 Be has been delivered to <strong>the</strong> region by<br />
precipitation and dustfall during <strong>the</strong> past ~15,000 to<br />
25,000 years. The analysis <strong>of</strong> meteoric 10 Be in <strong>the</strong>se<br />
stable reference sites also improves <strong>the</strong> understanding<br />
<strong>of</strong> correlations between o<strong>the</strong>r methods <strong>of</strong> dating,<br />
such as optically stimulated luminescence (OSL), 14 C<br />
<strong>of</strong> charcoal layers, and in situ 10 Be, something which<br />
could expand this method <strong>of</strong> dating throughout <strong>the</strong><br />
geologic research community. The reference sites are<br />
a known Pinedale moraine (~15 ka) at Silver Lake<br />
and a soil pit in Upper Gordon Gulch recently dated<br />
using OSL (~26 ka) (Völkel et al., 2010).<br />
Digital Elevation Map (DEM) Analysis<br />
Snow-<strong>of</strong>f Light Detection and Ranging (LIDAR) data<br />
(National Center for Airborne Laser Mapping, August<br />
2010) were used to produce digital elevation maps<br />
(DEM) for use in ArcMap. Pr<strong>of</strong>iles <strong>of</strong> <strong>the</strong> north- and<br />
south-facing hillslopes <strong>of</strong> Gordon Gulch were made<br />
in ArcMap using Spatial Analyst to create pr<strong>of</strong>iles
24th Annual Keck Symposium: 2011 Union College, Schenectady, NY<br />
(Fig. 2). For each hillslope, <strong>the</strong> transect with pits<br />
was <strong>the</strong> original pr<strong>of</strong>ile produced. Transects parallel<br />
to <strong>the</strong>se were spaced approximately 120 m apart and<br />
extended to <strong>the</strong> borders <strong>of</strong> <strong>the</strong> drainage basin. All<br />
transects began in <strong>the</strong> channel and ended at <strong>the</strong> ridge<br />
and were normal to <strong>the</strong> channel. Spatial Analyst was<br />
also used to produce digital elevation, hillshade, and<br />
slope maps using LiDAR data. These two maps and<br />
<strong>the</strong> additional air photo provide three different visual<br />
190<br />
180 representations <strong>of</strong> Gordon Gulch drainage basin as a<br />
170<br />
160 whole (Fig. 2).<br />
Distance (m)<br />
150<br />
140<br />
130<br />
120<br />
110 a. North Facing Hillslope<br />
100<br />
90<br />
80<br />
70<br />
60<br />
50<br />
40<br />
30<br />
20<br />
10<br />
F<br />
0<br />
180<br />
160<br />
140<br />
120<br />
100<br />
40<br />
20<br />
190<br />
180<br />
170<br />
160<br />
150<br />
140<br />
130<br />
120<br />
110<br />
100<br />
90<br />
80<br />
70<br />
60<br />
50<br />
40<br />
30<br />
20<br />
10<br />
0<br />
0<br />
0<br />
D<br />
E<br />
G<br />
Figure 2. Pr<strong>of</strong>iles <strong>of</strong> <strong>the</strong> north- and south-facing hillslopes<br />
<strong>of</strong> Gordon Gulch extracted from ~1m resolution LiDAR<br />
digital elevation model. Pr<strong>of</strong>iles are labeled from west<br />
to east (downstream) beginning with “A” or “a”. Transects<br />
in bold are <strong>the</strong> focus transects <strong>of</strong> this study. Open<br />
circles indicate locations <strong>of</strong> pits sampled for meteoric 10 Be.<br />
Shaded relief map (see Fig 1. for legend) shows approximate<br />
location <strong>of</strong> transects with black lines. Dots indicate<br />
sample pit locations. Dashed line indicates approximate<br />
boundary between <strong>the</strong> upper and lower basin.<br />
RESULTS<br />
Map Analysis<br />
NGG-05<br />
NGG-04<br />
M<br />
O<br />
N R<br />
P<br />
NGG-03<br />
100 200 300 400 500<br />
NGG-02<br />
NGG-01<br />
600 700 800 900 1000<br />
Q S<br />
L<br />
K<br />
K<br />
J<br />
I<br />
H<br />
L<br />
M<br />
A B<br />
E<br />
D<br />
G<br />
F<br />
H<br />
J<br />
I<br />
C<br />
N Distance (m)<br />
S<br />
q pr<br />
s<br />
80<br />
m<br />
l<br />
g<br />
60 f<br />
t<br />
o u<br />
v<br />
n k w j<br />
e<br />
h<br />
100<br />
SGG-03<br />
i<br />
200<br />
SGG-02<br />
d<br />
300<br />
b. South Facing Hillslope<br />
c<br />
b<br />
SGG-01<br />
400<br />
a<br />
SGG-00<br />
500<br />
The air photo clearly depicts differences in present<br />
1000100 200 300 400 500 600<br />
600<br />
A BC<br />
DEF<br />
700<br />
G<br />
e<br />
f g<br />
a b c d<br />
h<br />
i<br />
j<br />
k<br />
H<br />
I<br />
J<br />
K<br />
L M N<br />
800<br />
900<br />
l<br />
V<br />
U W<br />
T<br />
m<br />
n op q r s t u v w x<br />
O P<br />
Q<br />
R<br />
S T<br />
U V<br />
W<br />
148<br />
vegetation on <strong>the</strong> north- and south-facing hillslopes <strong>of</strong><br />
lower Gordon Gulch. Vegetation is less dense on <strong>the</strong><br />
south-facing hillslope, whereas little except vegetation<br />
is visible on <strong>the</strong> north-facing hillslope.<br />
With <strong>the</strong> addition <strong>of</strong> a hillshade layer, <strong>the</strong> DEM depicts<br />
both elevation and topographic features. Bedrock<br />
outcrops are seen extruding 1-10 m above <strong>the</strong><br />
dominant soil mantled hillslopes. These bedrock outcrops<br />
are characterized by shallow slope upslope and<br />
close to vertical slope on <strong>the</strong> downslope sides. The<br />
density <strong>of</strong> bedrock outcrops between <strong>the</strong> north- and<br />
south-facing hillslopes appears similar in upper Gordon<br />
Gulch. However, <strong>the</strong>re is a disparity <strong>of</strong> bedrock<br />
outcrop density between <strong>the</strong> north- and south-facing<br />
hillslopes <strong>of</strong> lower Gordon Gulch (Trotta, 2010).<br />
More bedrock outcrops are present on <strong>the</strong> southfacing<br />
hillslope, and <strong>the</strong>se are larger than those on <strong>the</strong><br />
north-facing hillslope. The north-facing hillslope is<br />
dissected by numerous gullies, whereas few gullies<br />
are visible on <strong>the</strong> south-facing hillslope.<br />
The slope map shows that many <strong>of</strong> <strong>the</strong> north-facing<br />
hillslopes have a parabolic shape in which <strong>the</strong> highest<br />
slopes are in <strong>the</strong> middle-range <strong>of</strong> <strong>the</strong> transect and <strong>the</strong><br />
lowest slopes at <strong>the</strong> bottom <strong>of</strong> <strong>the</strong> gulch and <strong>the</strong> drainage<br />
divide. In contrast, <strong>the</strong> slope is fairly constant<br />
on <strong>the</strong> south-facing hillslope <strong>of</strong> lower Gordon Gulch,<br />
except where <strong>the</strong> slope increases in <strong>the</strong> downslope direction<br />
<strong>of</strong> bedrock outcrops and as <strong>the</strong> slope shallows<br />
near <strong>the</strong> ridge <strong>of</strong> <strong>the</strong> drainage basin.<br />
Hillslope Pr<strong>of</strong>iles<br />
The north-facing hillslope <strong>of</strong> Gordon Gulch is shorter<br />
and shallower (Fig. 2a) than <strong>the</strong> south-facing hillslope,<br />
which is longer and steeper overall (Fig. 2b).<br />
Using data drawn from ArcMap hillslope pr<strong>of</strong>iles<br />
(Fig. 2) beginning at <strong>the</strong> first stream encountered<br />
and ending at <strong>the</strong> drainage divide, <strong>the</strong> average slope<br />
<strong>of</strong> all north-facing hillslopes is 9.6º and <strong>of</strong> all southfacing<br />
hillslopes is 12.2º. Taking <strong>the</strong> average from<br />
<strong>the</strong> sampled pit transect and <strong>the</strong> two transects parallel<br />
on each side yields slopes <strong>of</strong> 15.0º on <strong>the</strong> north-facing<br />
hillslope and 19.6º on <strong>the</strong> south-facing hillslope. The<br />
steepest local (and non-bedrock outcrop) slopes are<br />
found in <strong>the</strong> lower half <strong>of</strong> <strong>the</strong> north-facing hillslope,<br />
as seen in <strong>the</strong> slope map, but not reflected in pr<strong>of</strong>ile
24th Annual Keck Symposium: 2011 Union College, Schenectady, NY<br />
data due to pr<strong>of</strong>ile length averages. In lower Gordon<br />
Gulch, <strong>the</strong> north-facing hillslope pr<strong>of</strong>iles are characterized<br />
by a ridge creating a break in slope followed<br />
by a second slope ending at <strong>the</strong> drainage basin boundary.<br />
The first slope (before <strong>the</strong> ridge break) is parabolic<br />
in nature. Bedrock outcrops are found almost<br />
solely at <strong>the</strong> top <strong>of</strong> <strong>the</strong> pr<strong>of</strong>ile, but above <strong>the</strong> break in<br />
slope. In contrast, <strong>the</strong> south-facing hillslope pr<strong>of</strong>iles<br />
are more linear than parabolic and have a more consistent<br />
slope from ridge to stream, except where broken<br />
by bedrock outcrops. Bedrock outcrops are more<br />
abundant on this slope than <strong>the</strong> north-facing hillslope,<br />
and are seen as <strong>the</strong> many small spikes on <strong>the</strong> pr<strong>of</strong>iles.<br />
Meteoric 10 Be<br />
Lower Gordon Gulch<br />
North-facing hillslope soils have 10 Be concentrations<br />
<strong>of</strong> 5-8 x 10 8 atoms/g in <strong>the</strong> uppermost samples and<br />
decline to < 2 x 10 8 atoms/g at depths >60 cm (Table<br />
1). Pits NGG-01 and NGG-05 show exponentially<br />
declining pr<strong>of</strong>iles with some variability and small<br />
bulges, corresponding to <strong>the</strong> highest concentration<br />
<strong>of</strong> 10 Be in <strong>the</strong> B-horizon and <strong>the</strong> Cox-horizon respectively.<br />
Pits NGG-02, NGG-03, and NGG-04 show<br />
exponentially declining pr<strong>of</strong>iles (with some variability)<br />
with peak meteoric 10 Be concentrations towards<br />
<strong>the</strong> top <strong>of</strong> <strong>the</strong> pr<strong>of</strong>iles. The lowest concentration for<br />
each pit is at depth, except for NGG-01 in which <strong>the</strong><br />
lowest concentration is in <strong>the</strong> Cox2-horizon (Fig. 3a).<br />
Pits NGG-01 and NGG-02 decline to a steady concentration<br />
at depth, whereas <strong>the</strong> o<strong>the</strong>r three pits on<br />
this hillslope show no clear indication <strong>of</strong> approaching<br />
a steady concentration.<br />
South-facing hillslope soils have 10 Be concentrations<br />
<strong>of</strong> 2.5-5 x 10 8 atoms/g near <strong>the</strong> surface and decline to<br />
< 2.7 x 10 8 atoms/g at depths >32 cm (Table 1). Pits<br />
SGG-00 and SGG-02 show approximately linear declining<br />
pr<strong>of</strong>iles. Pit SGG-02 also shows a bulge, corresponding<br />
to <strong>the</strong> highest concentration <strong>of</strong> 10 Be in <strong>the</strong><br />
SGG-02 A/Cox1-horizon border. Pit SGG-03 shows<br />
an exponentially declining pr<strong>of</strong>ile with a small bulge,<br />
corresponding to <strong>the</strong> highest concentration <strong>of</strong> 10 Be in<br />
<strong>the</strong> Cox1-horizon. Pit SGG-01 shows no clear declining<br />
trend with a bulge corresponding to <strong>the</strong> highest<br />
concentrations <strong>of</strong> 10 Be in <strong>the</strong> SGG-01A-horizon. The<br />
lowest concentration for each pit is at depth (Fig. 3b).<br />
149<br />
Pits SGG-01 and SGG-03 appear to be approaching a<br />
steady concentration at depth, whereas pits SGG-00<br />
and SGG-02 show no clear indication <strong>of</strong> approaching<br />
a steady concentration.<br />
Sample<br />
10Be Bulk Density Thickness 10Be inventory Total Pit Inventory Soil Age<br />
(atoms/g) (g/cm^3) (cm) (atoms/cm^2) (atoms/cm^2) (years)<br />
NGG-01-8 8.06E+08 1.2 5 3.69E+09<br />
NGG-01-18 8.96E+08 1.4 10 9.77E+09<br />
NGG-01-28 7.54E+08 1.4 25 1.95E+10<br />
NGG-01-58 2.27E+08 1.6 35 2.26E+09<br />
NGG-01-98 2.02E+08 1.6 40 1.06E+09<br />
NGG-01-138 1.47E+08 1.6 45 -2.73E+09<br />
NGG-01-188 2.06E+08 1.6 50 1.50E+10<br />
NGG-02-5 5.73E+08 1.6 10 8.09E+09<br />
NGG-02-25 2.74E+08 1.6 25 7.97E+09<br />
NGG-02-55 1.64E+08 1.6 35 4.88E+09<br />
NGG-02-95 1.17E+08 1.6 55 3.41E+09<br />
NGG-02-165 4.07E+07 1.6 70 -4.38E+09<br />
NGG-03-10 5.92E+08 1.5 5 3.42E+09<br />
NGG-03-20 3.66E+08 1.5 15 5.23E+09<br />
NGG-03-40 2.61E+08 1.5 20 3.85E+09<br />
NGG-03-60 1.85E+08 1.5 20 1.58E+09<br />
NGG-04-5 5.48E+08 1.4 5 2.87E+09<br />
NGG-04-15 4.29E+08 1.4 10 4.10E+09<br />
NGG-04-25 4.50E+08 1.6 10 5.09E+09<br />
NGG-04-35 2.49E+08 1.6 15 2.81E+09<br />
NGG-04-55 1.80E+08 1.6 20 1.53E+09<br />
NGG-05-0 5.56E+08 1.4 5 3.00E+09<br />
NGG-05-10 6.62E+08 1.5 10 7.95E+09<br />
NGG-05-20 4.41E+08 1.5 10 4.63E+09<br />
NGG-05-30 5.31E+08 1.6 15 9.57E+09<br />
NGG-05-50 1.98E+08 1.6 20 2.10E+09<br />
SGG-00-0 4.59E+08 1.2 5 1.88E+09<br />
SGG-00-10 3.99E+08 1.2 10 3.07E+09<br />
SGG-00-20 3.95E+08 1.6 10 4.21E+09<br />
SGG-00-30 3.12E+08 1.6 15 4.32E+09<br />
SGG-00-50 1.88E+08 1.6 20 1.80E+09<br />
SGG-01-0 2.85E+08 1.2 5 8.82E+08<br />
SGG-01-10 3.51E+08 1.2 10 2.52E+09<br />
SGG-01-20 4.28E+08 1.6 15 7.11E+09<br />
SGG-01-40 3.85E+08 1.6 20 8.10E+09<br />
SGG-01-60 1.29E+08 1.6 20 -1.04E+08<br />
SGG-02-2 3.51E+08 1.5 5 1.64E+09<br />
SGG-02-12 3.69E+08 1.5 10 3.55E+09<br />
SGG-02-22 2.79E+08 1.5 10 2.20E+09<br />
SGG-02-32 2.67E+08 1.5 10 2.02E+09<br />
SGG-03-0 2.93E+08 1.3 5 1.05E+09<br />
SGG-03-10 3.56E+08 1.5 10 3.38E+09<br />
SGG-03-20 2.96E+08 1.5 15 3.72E+09<br />
SGG-03-40 2.08E+08 1.5 25 2.88E+09<br />
Table 1. 10 Be concentration data for samples chosen for<br />
this analysis. Preliminary 10 Be inventory and soil ages<br />
were calculated using equations presented in <strong>the</strong> text.<br />
Silver Lake Moraine<br />
3.53E+10 32,300<br />
2.00E+10 18,300<br />
1.41E+10 12,800<br />
1.64E+10 15,000<br />
2.73E+10 24,900<br />
1.53E+10 13,900<br />
1.85E+10 16,900<br />
9.42E+09 8,580<br />
1.19E+10 10,900<br />
The Silver Lake moraine section shows a bulge pr<strong>of</strong>ile<br />
with <strong>the</strong> highest concentration in <strong>the</strong> Bw-horizon<br />
(Fig. 4a). In <strong>the</strong> Bw-horizon at 25 cm and 35 cm<br />
deep, 10 Be concentrations (1.00 x 109 and 9.34 x 108<br />
atoms/g) are greater than twice <strong>the</strong> next highest concentration<br />
(4.35 x 10 8 atoms/g) in <strong>the</strong> Ej-horizon at 15<br />
cm deep (Tab. 1). Concentrations decrease to as low<br />
as 2.28 x 10 7 atoms/g in <strong>the</strong> Cu-horizon and appear to<br />
be approaching a steady concentration.
24th Annual Keck Symposium: 2011 Union College, Schenectady, NY<br />
Depth (cm)<br />
0.00E+00 !(!!)*!!" 1.00E+08 '(!!)*!&" 2.00E+08 #(!!)*!&" 3.00E+08 +(!!)*!&" 4.00E+08 $(!!)*!&" 5.00E+08 ,(!!)*!&" 6.00E+08 %(!!)*!&" 7.00E+08 -(!!)*!&" 8.00E+08 &(!!)*!&" 9.00E+08 .(!!)*!&" 1.00E+09 '(!!)*!."<br />
0<br />
0!"<br />
20<br />
#!"<br />
40<br />
$!"<br />
60<br />
%!"<br />
60<br />
80<br />
&!"<br />
100<br />
'!!"<br />
100<br />
120<br />
'#!"<br />
140<br />
140 '$!"<br />
160<br />
'%!"<br />
180<br />
180 '&!"<br />
200<br />
#!!"<br />
0<br />
40<br />
80<br />
Figure 3. Meteoric 10 100<br />
Be from soils on <strong>the</strong> north (a.) and<br />
south (b.) facing hillslopes <strong>of</strong> lower Gordon Gulch.<br />
120<br />
140<br />
160<br />
180<br />
Upper Gordon Gulch<br />
200<br />
0<br />
20<br />
60<br />
NGG-01<br />
NGG-02<br />
NGG-03<br />
NGG-04<br />
NGG-05<br />
0 100000000 200000000 300000000 400000000 500000000 600000000 700000000 800000000 900000000 1E+09<br />
SGG-00<br />
SGG-01<br />
SGG-02<br />
SGG-03<br />
Meteoric ¹⁰Be Concentration(atoms/g)<br />
2.00 × 10⁸ 4.00 × 10⁸ 6.00 × 10⁸ 8.00 × 10⁸<br />
2.00 × 10⁸ 4.00 × 10⁸<br />
b. South Facing Hillslope<br />
The soil originally sampled for OSL dating<br />
by Jörg Völkel (2010) has a maximum 10 Be concentration<br />
<strong>of</strong> 1.66 x 10 10 atoms/g near <strong>the</strong> surface. The<br />
concentration decreases to 3.13 x 10 9 atoms/g at depth<br />
(Tab. 1) and creates an exponentially declining pr<strong>of</strong>ile<br />
(Fig. 4b).<br />
Inventory and Soil Age Calculations<br />
Meteoric 10 Be inventories I Be (atoms/cm2) were calculated<br />
using an equation taken from Jungers et al.<br />
(2009):<br />
I Be = ∑(C Be -C inh )ρ s h<br />
a. North Facing Hillslope<br />
where C Be is depth-integrated 10 Be concentration, C inh<br />
is <strong>the</strong> inherited component <strong>of</strong> meteoric 10 Be , ρ s is<br />
depth-integrated soil bulk density, and h is soil thickness<br />
for each depth subsample.<br />
Inheritance for NGG-01 was assumed as <strong>the</strong> average<br />
concentration <strong>of</strong> <strong>the</strong> three deepest samples (1.85 x 10 8<br />
atoms/g) and for NGG-02 as <strong>the</strong> average <strong>of</strong> <strong>the</strong> two<br />
150<br />
Depth (cm)<br />
NGG-01<br />
/001!'"<br />
NGG-02<br />
/001!#"<br />
NGG-03<br />
/001!+" NGG-04<br />
/001!$" NGG-05<br />
/001!,"<br />
SGG-00<br />
SGG-01<br />
SGG-02<br />
SGG-03<br />
Depth (cm)<br />
0<br />
350<br />
0<br />
0<br />
50<br />
100<br />
150<br />
200<br />
250<br />
300<br />
0<br />
20<br />
40<br />
60<br />
80<br />
100<br />
Meteoric ¹⁰Be Concentration(atoms/g)<br />
2.00 × 10⁸ 4.00 × 10⁸ 6.00 × 10⁸ 8.00 × 10⁸ 1.00 × 10⁹<br />
0 200000000 400000000 600000000 800000000 1E+09 1.2E+09<br />
Meteoric ¹⁰Be Concentration(atoms/g)<br />
a. Silver Lake Moraine<br />
1.00 × 10⁸ 2.00 × 10⁸ 3.00 × 10⁸ 4.00 × 10⁸ 5.00 × 10⁸<br />
0 100000000 200000000 300000000 400000000 500000000 600000000<br />
b. Upper Gordon Gulch OSL<br />
Dated Pit<br />
Figure 4. Meteoric 10 Be concentration plots for <strong>the</strong> Silver<br />
120<br />
Lake moraine (a.) and upper Gordon Gulch pit dated with<br />
OSL (b.).<br />
deepest samples (7.87 x 10 7 atoms/g). The average <strong>of</strong><br />
<strong>the</strong> NGG-01 and NGG-02 inheritance values (1.32 x<br />
10 8 atoms/g) was used as <strong>the</strong> value for <strong>the</strong> remaining<br />
NGG pits and <strong>the</strong> SGG pits. The deepest concentration<br />
(9.21 x 10 7 atoms/g) was used for <strong>the</strong> OSL-dated<br />
pit. Inheritance has not yet been taken into account<br />
for <strong>the</strong> Silver Lake moraine, because it is assumed<br />
to be minimal. Bulk soil density was measured in<br />
<strong>the</strong> field for several sample locations and o<strong>the</strong>r bulk<br />
density values were assumed as standard values for<br />
soil and till density. The soil thickness for each depth<br />
subsample was calculated by establishing a mid-point<br />
between sample locations. The midpoint value was<br />
added to bottom samples to account for un-collected<br />
soil.
24th Annual Keck Symposium: 2011 Union College, Schenectady, NY<br />
Soil ages (t, in years) were calculated for each pit using<br />
an equation adapted from Graly et al. (2010):<br />
t = (-1/λ)ln(1-λI Be /q)<br />
where λ is <strong>the</strong> 10 Be disintegration constant (5.1 x 10 -7 ),<br />
I Be is <strong>the</strong> total inventory <strong>of</strong> 10 Be (atoms/cm 2 ), and q is<br />
<strong>the</strong> local <strong>annual</strong> meteoric 10 Be flux (atoms/cm 2 ). The<br />
value <strong>of</strong> q was estimated from Figure 4 <strong>of</strong> Graly et al.<br />
and according to an <strong>annual</strong> precipitation rate <strong>of</strong> ~45<br />
cm in Gordon Gulch (1.1 x 10 6 atoms/cm 2 ) and ~90<br />
cm at Silver Lake (1.8 x 10 6 atoms/cm 2 ), both at ~40º<br />
latitude, (Graly et al., 2010).<br />
DISCUSSION<br />
Meteoric 10 Be<br />
Lower Gordon Gulch<br />
Meteoric 10 Be data suggest that <strong>the</strong> two opposite facing<br />
hillslopes <strong>of</strong> lower Gordon Gulch are evolving<br />
differently (Fig. 3). The three soil pits in <strong>the</strong> middle<br />
<strong>of</strong> <strong>the</strong> transect on <strong>the</strong> north-facing hillslope all show<br />
declining pr<strong>of</strong>iles <strong>of</strong> meteoric 10 Be concentration, suggesting<br />
an eroding slope because upper soil has been<br />
stripped from <strong>the</strong> column and 10 Be has not had <strong>the</strong><br />
time to concentrate in a certain layer. Potential explanations<br />
for small bulges in <strong>the</strong> lowest (NGG-01) and<br />
highest (NGG-05) pr<strong>of</strong>iles include <strong>the</strong>ir locations and<br />
horizon pr<strong>of</strong>iles. Pit NGG-05 is located above <strong>the</strong><br />
major break in slope on <strong>the</strong> hillslope, and <strong>the</strong>refore<br />
represents a different environment than <strong>the</strong> four lower<br />
pits. The highest concentrations <strong>of</strong> meteoric 10 Be in<br />
pits NGG-01 and NGG-03 are in <strong>the</strong> B-horizons.<br />
Meteoric 10 Be concentration data obtained from <strong>the</strong><br />
south-facing slope differs from <strong>the</strong> concentrations<br />
from <strong>the</strong> north-facing slope. The only declining pr<strong>of</strong>ile<br />
<strong>of</strong> <strong>the</strong> four sampled pits is <strong>the</strong> lowest (SGG-00),<br />
whereas <strong>the</strong> o<strong>the</strong>r pr<strong>of</strong>iles all contain a small bulge.<br />
Preliminary analysis suggests that perhaps <strong>the</strong> upper<br />
portion <strong>of</strong> <strong>the</strong> soil column on <strong>the</strong> south-facing hillslope<br />
has been stripped and evacuated. Had this not<br />
occurred, <strong>the</strong> pr<strong>of</strong>ile shape on <strong>the</strong> south-facing hillslope<br />
may have resembled those on <strong>the</strong> north-facing<br />
hillslope with small bulges but overall decline pr<strong>of</strong>iles<br />
151<br />
and greater inventories. Therefore consistent meteoric<br />
10 Be concentration pr<strong>of</strong>ile shapes across <strong>the</strong> two<br />
opposite facing hillslopes <strong>of</strong> Gordon Gulch would<br />
have been seen.<br />
Preliminary 10 Be inventory calculations show differences<br />
between <strong>the</strong> two hillslopes (Table 1). The<br />
north-facing hillslope has a greater inventory than <strong>the</strong><br />
south-facing hillslope, particularly in <strong>the</strong> upper ~30<br />
cm <strong>of</strong> <strong>the</strong> pr<strong>of</strong>iles. Preliminary soil age calculations<br />
(Table 1) correlate to inventory differences and suggest<br />
that <strong>the</strong> south-facing hillslope <strong>of</strong> lower Gordon<br />
Gulch is eroding faster than <strong>the</strong> north-facing hillslope.<br />
Upper Gordon Gulch<br />
The 10 Be pr<strong>of</strong>ile shows a declining pr<strong>of</strong>ile (Fig. 4b),<br />
suggesting a young, eroding landscape. The preliminary<br />
soil age calculation drawn from an inventory <strong>of</strong><br />
2.43 x 10 10 atoms/cm2 suggests an age <strong>of</strong> ~22,200<br />
years (Table 1), matching <strong>the</strong> age obtained from<br />
OSL dating <strong>of</strong> <strong>the</strong> same soil pr<strong>of</strong>ile (~26,500 years)<br />
(Völkel et al., 2010).<br />
Silver Lake<br />
The 10 Be pr<strong>of</strong>ile shows a bulge pr<strong>of</strong>ile (Fig. 4a) with<br />
<strong>the</strong> highest 10 Be concentration in <strong>the</strong> Bw-horizon, fur<strong>the</strong>r<br />
emphasizing a correlation between 10 Be and Ferich<br />
B-horizons from previous studies. Preliminary<br />
soil age calculations suggest an age <strong>of</strong> ~54 ka (Table<br />
1), older than <strong>the</strong> known age <strong>of</strong> this moraine at ~15ka.<br />
Sources <strong>of</strong> Error<br />
Potential sources <strong>of</strong> error in all inventory and soil age<br />
calculations include lack <strong>of</strong> soil bulk density measurements,<br />
estimated <strong>annual</strong> 10 Be flux rates, and estimated<br />
10 Be inheritance. Inheritance also comes into question<br />
when contemplating what this may mean for hillslope<br />
evolution. Additionally, some pits may not have been<br />
dug deep enough to sample <strong>the</strong> entire meteoric 10 Be<br />
inventory, due to regolith and potential bedrock composition.
24th Annual Keck Symposium: 2011 Union College, Schenectady, NY<br />
CONCLUSION<br />
More detailed analysis <strong>of</strong> processes, soil ages, meteoric<br />
10 Be flux, and 10 Be inventories will lead to fur<strong>the</strong>r<br />
analysis and discussion <strong>of</strong> <strong>the</strong> differences between <strong>the</strong><br />
north-and south-facing hillslopes <strong>of</strong> Gordon Gulch<br />
and more precise dating <strong>of</strong> Gordon Gulch hillslopes<br />
and <strong>the</strong> Silver Lake moraine. Preliminary age calculations<br />
show that Gordon Gulch regolith is latest<br />
Pleistocene or Holocene in age or younger, not evolving<br />
throughout <strong>the</strong> Cenozoic, and that <strong>the</strong> soil flux<br />
here is rapid.<br />
ACKNOWLEDGEMENTS<br />
Many thanks to <strong>the</strong> KECK <strong>consortium</strong> and Amherst<br />
College for funding and support. Thank you Will<br />
Ouimet and David Dethier for continuing excitement<br />
and advice. Thank you to Peter Crowley for learning<br />
a completely new subset <strong>of</strong> <strong>geology</strong> to become an<br />
effective replacement advisor. Thank you to all my<br />
fellow CO-KECK students, particularly <strong>the</strong> awesome<br />
soils group. Lastly, thank you to <strong>the</strong> many people<br />
from CU-Boulder, UMass, UVM, and <strong>the</strong> GeoCAMS<br />
group at <strong>the</strong> Livermore National Laboratory who<br />
helped in many different steps <strong>of</strong> field and laboratory<br />
work.<br />
REFERENCES<br />
Anderson, R. S., Riihimaki, C. A., Safran, E. B.,<br />
MacGregor, K. R., 2006, Facing reality: Late<br />
Cenozoic evolution <strong>of</strong> smooth peaks, glacially<br />
ornamented valleys and deep river gorges <strong>of</strong><br />
Colorado’s Front Range: in Willett, S.D., Hovius,<br />
N., Brandon, M.T., and Fisher, D.M., eds.,<br />
Tectonics, climate and landscape evolution: Geological<br />
Society <strong>of</strong> America Special Paper 398, p.<br />
397-418.<br />
Fifield L. K., Wasson R. J., Pillans B., Stone J. O. H.,<br />
2010, The longevity <strong>of</strong> hillslope soil in SE and<br />
NW Australia, CATENA, Volume 81, Issue 1, p.<br />
32-42.<br />
Graly, J. A., Bierman P. R., Reusser, L. J., Pavich,<br />
M. J., 2010, Meteoric 10Be in soil pr<strong>of</strong>iles - A<br />
152<br />
global meta-analysis, Geochimica et Cosmochimica<br />
Acta, v. 74, Issue 23, p. 6814-6829.<br />
Jungers, M. C., Bierman, P. R., Matmon, A., Nichols,<br />
K., Larsen, J., and Finkel, R., 2009, Tracing<br />
hillslope sediment production and transport with<br />
in situ and meteoric 10Be, J. Geophys. Res., v.<br />
114.<br />
Stone, J., 1998, A rapid fusion method for separation<br />
<strong>of</strong> beryllium-10 from soils and silicates: Geochimica<br />
et Cosmochimica Acta, v. 62, p. 555–561.<br />
Trotta, J. R., 2010, The Distribution <strong>of</strong> Tors in Gordon<br />
Gulch, Front Range, Colorado. Thesis submitted<br />
in partial fulfillment <strong>of</strong> <strong>the</strong> requirements for<br />
<strong>the</strong> Degree <strong>of</strong> Bachelor <strong>of</strong> Arts with Honors in<br />
Geosciences, Williams College.<br />
Völkel, J., Huber, J., Leopold, M. and Dethier, D.,<br />
2010, Young Quaternary slope sediments and<br />
paleosoils in <strong>the</strong> Colorado Front Range – process<br />
and age, Geological Society <strong>of</strong> America Abstracts<br />
with Programs, 42, no. 5, p. 469.<br />
Willenbring, J. K. and von Blanckenburg, F., 2010,<br />
Meteoric cosmogenic Beryllium-10 adsorbed to<br />
river sediment and soil: Applications for Earthsurface<br />
dynamics, Earth Science Reviews v. 98,<br />
p. 105-122.