GLACIAL OHIO 2001: THE LAST GLACIAL TRANSITIONRECORDED IN LAKES AND BOGS, SOUTHWESTERN

OHIO

FACULTYGreg Wiles, The College of Wooster

Tom Lowell, University of Cincinnati

Don Pair, University of Dayton

STUDENTSScott Bagocius, The College of Wooster

Louisa Bradtmiller, Smith College

Joel Byersdorfer, Whitman College

Monica Kaitz, Amherst College

Lisa King, University of Cincinatti

Jessica McDonough, Universoty of Dayton

Margretta Meyer, Beloit College

Kimberly Sunderlin, Franklin & Marshall F&M

VISITING and PARTNERSCIENTISTS

Katie Shoenenberger, TA University of Dayton

John Ritter, Wittenberg University

Jessica McDonough, Figue 1. Laboratory results and graphic log of Newell Lake basin (site 0102) sediments.

GLACIAL OHIO 2001: THE LAST GLACIAL TRANSITIONRECORDED IN LAKES AND BOGS, SOUTHWESTERN

OHIO

GREG WILES, The College of WoosterTHOMAS V. LOWELL, University of Cincinnati

DONALD PAIR, University of Dayton

PROJECT OVERVIEWThe major objectives of Keck Ohio-2001 wereto document the retreat of the Laurentide icesheet across western Ohio and to analyzesediments laid down since the Last GlacialMaximum (LGM) as a record ofenvironmental change. This was accomplishedby dating organic material in, and examiningthe characteristics of sediments cored fromsmall basins located on drift deposited duringthe last Ice Age. Ten sediment cores wereextracted from bogs and lakes along an 80-kilometer, north-south transect from theclassic interlobate region centered onBellefontaine, Ohio (Figure 1).

Participants of Keck Ohio-2001 processed thepaleoenvironmental records from eight cores,which include parts of the last glacial-interglacial transition. This work improves thechronologic resolution of the deglacial historyand subsequent climate events in Ohio. Thesedata contribute to a better understanding ofclimate variability in the Great Lakes region ofNorth America.

Students assembled in Dayton, Ohio and thentraveled to the Canadian Rockies to studymodern glacial and lacustrine systems. We

visited Angel, Athabasca, Hilda, Peyto andSaskachewan Glaciers in Banff and JasperProvincial Parks over the five-day trip. Ateach field stop students described and builtfacies models, which included many of thefeatures that they would subsequentlyencounter on a continental ice sheet scale inOhio. Observing the evolution of a series oflakes from kettles to bogs in the CanadanRockies provided the basis for interpretingsediment cores in Ohio.

The group returned to Ohio to extract cores inthe field. Students began their laboratoryanalyses at the Keck EnvironmentalLaboratory housed in the Department ofGeology at the University of Dayton.

BACKGROUND ANDGLACIAL GEOLOGYIce sheet advance and retreat patterns are animportant record of global environmentalchange (Broecker and Denton, 1989). Thecharacteristics of sediments laid down duringand since the last deglaciation is a record ofenvironmental change. These records includeabrupt climate oscillations such as theYounger

Dryas climate reversal at 10,900-10,000 14C yrBP.

The late Wisconsin glacial history insouthwestern Ohio is one of the best-datedrecords from North America, and hascontributed to our understanding of the timingof global interhemispheric glaciation (Lowellet al., 1990, Broecker and Denton, 1989).Stratigraphic studies and hundreds ofradiocarbon dates show that the history of thesouthern margin of the Laurentide Ice Sheetconsisted of several pulses of advance andreadvance between 22,000 and 18,000 BP(Ekberg et al., 1993). In contrast, thedeglacial sequence and the character of theglacial transition across Ohio particularly forthe interval older than 14,000 BP is relativelyunknown.

The study region (Figure 1) includes acomplex series of recessional moraines andice-contact topography. The low regional

topography in this area of southwestern Ohiocontributes to the preservation of a markedlobate geometry of moraines, and crosscuttingrelationships of landforms resulting from themultiple fluctuations of the Lake Erie Lobe.Shane (1987) has provided a generalunderstanding of the deglacial and vegetationhistory of the region based on pollen analysisfrom four basins in Ohio. Our work builds onthis background.

METHODSThis project adopts a strategy successfullyemployed in the Lake District of Chile and inthe Southern Alps of New Zealand to improveour understanding of climate change (Morenoet al., 2001). In Ohio, a series of sedimentcores were extracted along a transect normalto the latest retreat of the Laurentide Icesheet.

Each Keck participant experienced all aspectsand methods of field mapping and coring, andthe laboratory methods involved in core

14.6

14.4

15.4

16.2

16.1

15.8

15.6

20 km

OHIO

Sitelocation

TABLE 1. Radiocarbon dates (Keck Ohio - 2001).

Core #/Interval Lab #* Radiocarbon Age

0101 B3 95098 Beta-158291 11,860+/-270 BP0102 B4 17-20 Beta-158292 12,130+/-200 BP0102 A6 4-10 AA45068 35,120+-820 BP0103 A4 17.5-19.5 AA45069 15,810+-140 BP0103 B1 33-34 AA45070 14,285+-92 BP0103A268-71** Beta-163007 13,150+/-70 BP0104 B3 54-55 AA45071 12,324+-80 BP0104 B5 68-73 AA45072 15,350+-100 BP0105 A6 67-70 + 64-73 Beta – 158293 9,960 +/- 100 BP0105 B8 74--75 AA45073 15,563+-91 BP0106 A5 58-60 AA45074 14,986+-98 BP0106 B2 46-48 AA45075 16,090+-100 BP0107 A3 8-16 AA45076 13,490+-110 BP0107 B3 48-52 AA45077 14,360+-120 BP0109 A7 74-78 AA45078 14,600+-91 BP0109 B4 78-82 Beta-158294 10,000+/-19 BP0110 B7 29-31 AA45079 16,170+-97 BP0110 A4 8-10 Beta – 158295 13,370 +/- 280 BP

(*Labs: Beta = Beta Analytic, AA = U. Arizona AMS lab.), * *C-14analysis funded from the Copeland Research Funds, College of Wooster.

Figure 1. Core site locations for KeckOhio 2001 project. The samplingtransect is centered on the inter-lobate region of western Ohio.Ice retreat dates are plotted onthe detailed map. Dates are in1000s years before present.

analysis and archiving. Each lake and bogbasin was surveyed using either GPS orelectronic total station. Sediment thicknesswas measured by tile probing along a grid. AGIS data set was prepared for each basin andlocation, and isopach maps were generated.Based on this information, coring sites werelocated at the deepest, accessible part of thebasin. Two side-by-side cores offset by onemeter were extracted with a modifiedLivingstone corer (Figure 2).

The range of sediment types encountered atthe sites was remarkably rich. The depth ofmud in each of the basins ranged from four toseventeen meters and included peats, marls,laminated and massive silts and clays, andgyttja (amorphous organic material). The onlyobvious similarities in all the cores was theglacial to interglacial transition characterizedby either a gradual or abrupt increase inorganic content and a decrease in grain size(Figure 3). The landforms on which the closedbasins were sampled ranged from ice-contactsediments of kames and eskers to morainescomposed of glacial till.

Each core was described and photographed,sampled for radiocarbon dating, and measuredfor magnetic susceptibility, total organic andcarbonate content, and grain size (Figure 3).

Two radiocarbon ages on each core wereobtained from either bulk samples of peat andgyttja submitted to Beta Analytic Inc. forconventional radiocarbon dating or fromdetrital organic material submitted forAccelerator Mass Spectromety (AMS). AMSdating was performed through theLimnological Research Center (LRC) at theUniversity of Minnesota, where the sampleswere processed prior to submission at theaccelerator facilities at the University ofArizona (Table 1).

Toward the end of the project each participantwrote a proposal to analyze further one of thecores at his or her institution. Studentsperformed additional analyses according tointerests and available facilities. Theseanalyses included pollen, diatom,geochemical, environmental magnetism, anddetailed grain size analyses.

STUDENT PROJECTSIndividual student projects includedpaleoecological studies using, diatoms byMargaretta Meyer (Beloit College) and pollenby Monica Kaitz (Amherst College).Geochemical studies by Louisa Bradtmiller(Smith College) and Kim Sunderlin, (Franklinand Marshall College) explored cation

gyttja

mottledsilt

laminatedsilt

sandy silt

Graphic Log

Mag.Susc.

Grain Size

Figure 3. Suite of core analyses that served as a startingpoint for student projects. Note the overall increase inorganic content, and decreases in grain size and suscepti-bility through the glacial transition. This core (0101) wasnot chosen for further analyses.

*

*

CORE 0101

peat

a. b.

Figure 2. Schematics of the modifiedLivingstone piston corer. a - corer ispushed into the sediment one meter at atime. b - the core barrel is removed fromthe hole and sediment is extruded. Abrochure describing the coring tech-niques and the project can be accessed athttp://tvl1.geo.uc.edu/ak/keck1.pdfhttp://tvl1.geo.uc.edu/ak/keck2.pdf

organics

carbonate

Loss onIgnition (LOI )

0% 100% clay sand

11,860+/-270 (Beta-158291)

exchange capacities and clay mineralogy, anddiagenetic changes within the sediment,respectively. Scott Bagocius (The College ofWooster) focused on the record ofenvironmental magnetism using a suite ofmagnetic analyses in conjunction withphysical sediment parameters. Lisa King(University of Cincinnati) evaluated therythmic sedimentation as a record of annual orepisodic sedimentation. Jessica McDonough(University of Dayton) focused on anextensive marl lake deposit and itsgeochemistry possibly related to lake levelfluctuations. Joel Byersdorfer (WhittmanCollege) took a sedimentologic approach tohis core analyses, identifying a sand anomaly,which he interprets as having a fluvial origin.

At the close of the project each student will beexpected to contribute their data to theNOAA/NGDC Paleoclimatology Program,Boulder Colorado, USA database for use byother climate researchers worldwide. Studentsare also required to supply the landowner attheir site with the results of their projects.

SUMMARY OF RESULTSA synthesis of the bog AMS radiocarbon datesshows that ice withdrew from the interlobateregion about 16,000 years ago from thesouthernmost sites to 14,500 BP from thenorthern sites (Figure 1). This additionalspatial and temporal coverage of western Ohioimproves our understanding of the deglacialhistory of this area.

Some participants were able to suggest thatchanges in their cores may be linked to someof the well-documented, abrupt climatereversals that occurred during the glacialtransition. These climate-driven changesinclude the Bøllering – Allerød and YoungerDryas events (Yu and Wright, 2001). Otherchanges noted in the cores were attributed tothe more local evolution of a particular basin.The differentiation of the local versus larger-scale (global), climate-driven origin in thecores is fundamental to further studiesconcerned with reconstructing theenvironmental history of the Midwest duringthe last glacial transition.

ACKNOWLEDGEMENTSWe thank the W.M. Keck Foundation and theNational Science Foundation for support ofthe Keck Ohio 2001 project. Eric Leonard(Colorado College) generously loaned coringequipment. Brian Haskell of the LimnologicalResearch Center provided AMS-datingexpertise. Katie Shoenenberger providedsupport in the GIS lab and coordinatedlogistics for the trip to Canada. We appreciatethe help of Beth Palmer and StephaniePennington in navigating us through theproject.

REFERENCES CITEDBroecker, W.S. and Denton, G.H., 1989, The

role of ocean-atmosphere reorganizations inglacial cycles: Geochemica etCosmochimica Acta, 53, 2465-2501.

Ekberg, M.P, Lowell, T.V., and Stuckenrath,R., 1993, Late Wisconsin glacial advanceand retreat patterns in southwestern Ohio,U.S.A., Boreas, v. 22, p.289-304.

Lowell, T.V., Savage, K.M., Brockman, C. S.,and Stuckenrath, R., 1990, Radiocarbonanalyses from Cincinnati, Ohio and theirimplications for glacial stratigraphicinterpretations, Quaternary Research, v. 34,p. 1-11.

Moreno, P.I., Jacobson, G.L. Jr., Lowell, T.V.,and Denton, G.H., 2001, Interhemisphericclimate links revealed from a late-glacialcooling episode in southern Chile, Nature,v. 409, p. 804-808.

Shane, L.C.K., 1987, Late-glacial,vegetational and climatic history of theAllegheny Plateau and the Till Plains ofOhio and Indiana, USA: Boreas 16, 1-20.

Yu, Zicheng and Wright, H.E., 2001,Response of interior North America toabrupt climate oscillations in the NorthAtlantic region during the last deglaciation:Earth Science Reviews, v. 52, p. 333-369.

THE ENVIRONMENTAL MAGNETIC RECORD OF THELAST GLACIAL-INTERGLACIAL TRANSITION FROM

ECKURD’S POND, URBANA, OHIO

SCOTT BAGOCIUSDepartment of Geology, The College of Wooster

Faculty Sponsors: Robert Varga and Gregory Wiles

INTRODUCTIONThe Laurentide Icesheet is believed to haveinvaded Ohio roughly 25,000 BP reaching itsmaximum extent near Cincinnati atapproximately 19,000 BP (Lowell, 1995). Theadvance history of the Laurentide Icesheet inOhio is relatively well documented however,the retreat history is less well-constrained. Thefocus of the Keck Ohio Project 2001 was todetermine the age and rate of the retreat of thelast glacial movement from Ohio. This wasaccomplished by acquiring numerous bog basinsediment cores from a north/south transect andobtaining a radiocarbon date from each bogbottom. A greater understanding of theLaurentide Icesheet retreat pattern will followas more transects are completed.

The focus of this study is using magneticproperties of sediment cores as proxies for bothenvironment and climate change.Environmental magnetism is a relatively newfield and has gained increasing use andrecognition as a field of environmental study(Oldfield, 1991). By analyzing the magneticproperties of samples, information concerningthe grain size and magnetic mineralogy can begathered for interpretation of the environmentalconditions in which the sediment was deposited.This paper is concerned with the magneticproperties of sediment retrieved from a bogbasin in southwestern Ohio. The sedimentscollected span from the Late Pleistocenethrough much of the Holocene.

By conducting a series of tests involvinganhysteretic remant magnetism (ARM),isothermal remant magnetism (IRM), saturationisothermal remant magnetism (SIRM) andmagnetic susceptibility tests on the 3.7 msediment sequence, information pertaining tothe sediments’ magnetic particle grain size,magnetic mineralogy, and the magnetic mineralconcentration within a sample will giveinformation reflecting on previousenvironmental conditions (Verosub andRoberts, 1995). Fluctuations in erosion of soilswithin the basin as well as variations in eolianinput external to the site will be interpreted inthe context of the changing dynamics of theretreating Laurentide Icesheet, global climatefluctuations, and internal changes in thegeologic and ecologic evolution of the basin.

LOCATIONEkurd’s Pond is a 1.5 square kilometer closedbasin formed on an ice-contact ridge located inthe interlobate morainal area of western Ohio(83° 40’ 05’’ E longitude and 40° 05’ 00’’ Nlatitude) between the cities of Dayton andBellefontaine (Fig. 1). The basin is thought tohave been created by surrounding eskersresulting in an enclosed catchment area.

METHODSThree overlapping cores were extracted fromthe basin and together yield a 3.7 metersequence of near continuous sediment. After

extraction, each core was physically describedand preliminary lab tests (LOI, grain size andmagnetic susceptibility) were conducted oneither one or all of the cores. Material for AMS

dating was also collected and sent to BetaAnalytic Inc. and the University of Arizona forC-14 dating analysis. Further magneticresearch was conducted using two of the threecores provided from Eckurd’s Pond.

Sediment was sampled every four cm from the3.7 m core for a total of 88 samples. Thesediment was then packed into 5.83 cm3 plasticboxes and a series of magnetic tests wereconducted on each of the samples. All testswere conducted at room temperature. The testsconsisted of magnetic susceptibility (using theBartington MS2 system), which is a basicmeasurement often providing insight regardingthe magnetic concentration within a sample.SIRM (2.5 Telsas) may be used to estimatesediment flux into a basin and also providemagnetic mineralogy information. IRM (at –20,-40, -100 and –300 mT) also has the ability tohelp distinguish the magnetic mineralogy of asample. ARM (peak frequency of 0.1 T and abias field of .05 mT) provides information

Fig.1. A relief map of the upper mid-western UnitedStates with a close-up view of the coring sites inOhio. Eckurd’s Pond (site 0103) is shown by thelarge white arrow.

Fig. 2. Stratigraphic log of the core sediments along with AMS dates (Conventional radiocarbon dates) as well asmagnetic parameters used to interpret the core’s magnetic components

pertaining to concentrations of magnetic grainsize within a sample. S-ratios (IRM/SIRM) areoften calculated to interpret the ratio betweenferromagnetic and anti-ferrimagnetic materialwithin a sample (Verosub and Roberts, 1995).

By combining two or more of the aboveprimary magnetic tests, many parameters can beconstructed allowing for a greater array ofenvironmental interpretation. Such parametersinclude ARM/Xlf, which generally increases asthe ratio of fine grain magnetic material withina sample increases and ARM/SIRM, which mayact similar to ARM/Xlf reaching high valueswith increase in fine grain concentration(Gillian, 1997). Graphs were constructed of thenumerous magnetic parameters used to interpretpast environmental conditions (Fig. 3).

FINDINGSRadiocarbon analyses, total organics, totalcarbonate, grain size and a suite ofenvironmental magnetic analyses define thisrecord of environmental change through theglacial-interglacial transition and much of theHolocene. Three distinct units are recognizedbased on the lithologic character, grain size,organic content and magnetic susceptibility.The lowest unit represents the changinglandscape from the ice-contact, immediatepostglacial setting to lacustrine deposition. Abasal AMS radiocarbon dating from this siteand four adjoining bogs suggest general iceretreat before 15,810 +/- 140 BP (AA45069). A2-meter sequence of lake silts was thendeposited within a 2,000 year period, accordingto AMS radiocarbon dates of 13,150 +/- 70 BP(Beta-163007) and 14,285 +/- 92 BP (AA5070)

at 2.34 m and 2.83 m respectively below groundsurface. These lower two units are capped by amore organic-rich silt recording the transitioninto the Holocene.

A distinct sudden change in magnetic properties(shown in ARM/SIRM, Xlf, SIRM and ARM)occurs about 13,150 BP directly correlating to alarge influx of organic rich silt. This point intime simultaneously shows an increase in grainsize, shown in high values of ARM/Xlf, to thecatchment. The input of organic matter has alsobeen increasing into the basin from 13,150 BPto the present day (Fig.3). The distinctmagnetic change at 13, 150 BP may possibly bea result of the low sedimentation rate incomparison to that of the basin’s earlier years,particularly between 15,810 BP and 14,285 BP(Fig. 2). Eolian activities are not thought to belarge contributors to the magnetic record before13,150 BP, but due to the lowering of S-ratioafter that date, wind blow silt may have playeda significant role throughout the Holocene.

REFERENCES CITEDGillian, T. M. 1997. Environmental magnetism

and magnetic correlation of high resolutionlake sediment records form NorthernHawkes Bay, New Zealand. New ZealandJournal of Geology and Geophysics. 40:287-298.

Lowell,T. V. 1995. The Application ofRadiocarbon Age Estimates to the Dating ofGlacial Sequences: An Example From theMiami Sublobe, Ohio, U.S.A. QuaternaryScience Review. 14: 85-99.

Oldfield, F. 1991. Environmental magnetism; apersonal perspective. Quaternary ScienceReviews. 10: 73-85.

Verosub, K. L., and Roberts, A. P. 1995.Environmetnal magnetism: past, present,and future. Journal of GeophysicalResearch. 100: 2175-2192.

Fig. 3. Graph showing percentage of organic andcarbonate material found in sediment. Organicmaterial increases throughout the Holocene.

GEOCHEMISTRY OF LAURENTIDE GLACIALSEDIMENT AND IMPLICATIONS FOR CLIMATE

CHANGE

LOUISA BRADTMILLERGeology Dept., Smith College

Sponsor: Robert Newton, Smith College

INTRODUCTIONRecords of large scale climate change havebeen studied in relation to glacial advance andretreat patterns in a number of ways. TheLaurentide Ice Sheet of the late Wisconsinglaciation is known to have reached as farsouth as southern Ohio, Indiana, and Illinois atits maximum, and this study continuesprevious work in southwestern Ohio on theScioto sublobe of the Lake Erie lobe. Thesecond major advance around 19,900 BPextended the farthest, and then retreated in aseries of less well documented, smallerreadvances and interstades. The betterunderstood periods include the Bølling-Allerød warming from 13000-11000 BP, theYounger Dryas cooling from 11000-10000 BPand subsequent Holocene warming.

Previous studies have focused on radiocarbonstratigraphy and stable Oxygen isotopes todocument patterns of advance and retreat, andto link those patterns to global climate events.The radiocarbon studies have been useful inproviding not only patterns of climate change,but an accurate chronology to accompanythem. Using two radiocarbon dates forcorrelation, this study focuses on the geo-chemistry of two cores from a small peat bogin the Mechanicsburg, Ohio area. Specificallyit hopes to show how cation exchange capacityand other analyses might be useful indetermining or correlating global climatechange events.

GEOLOGIC SETTINGThe study area (site 0104) sits within amoraine sequence at the southern margin ofthe Laurentide Ice Sheet. While the terminalmoraine lies approximately 50km south of thesite, this area is still rich in glacial depositsdue to a resistant bedrock ridge which split thesheet into two sublobes. Site 0104 lies on thewestern edge of the Scioto sublobe- on theeastern edge of the London moraine, and eastof the older Bloomingburg moraine (Figure 1).

The basin itself is a small (50x80x10m) peatbog vegetated with shrubs and grasses, nostanding water, and no streams flowing in orout. A man-made drainage gully surrounds

the northwest part of the basin, and there are a

Figure 1. Regional map of the southern margin ofthe Laurentide Ice sheet. From Lowell, 1995.

few other depressions in the area. The area,like most of Ohio, is underlain by limestone.

METHODS

Field MethodsSamples were collected using a piston corer.Two separate cores were taken one meterapart, each core sampling a different interval-starting at 150cm for Core A and 200cm forCore B. All 17 meter-long samples wereextracted into PVC tubes with plastic lining,measured, briefly described and photographed.In addition, depth to refusal was probed at 9other points. Locations for all sample sitesand depth points were determined using aTrimble GeoExplorerIII GPS receiver.

Laboratory Methods

Radiocarbon dating, Loss on Ignition

Samples of the core were extracted in the laband sent to the Limnological Research Centerof the University of Minnesota for preparation,and sent to the University of Arizona to beradiocarbon dated using the accelerator massspectrometer (AMS) method.

1cm3 samples were taken at 4cm intervalsfrom the B core. Each sample was weighed(±.005g), placed in a crucible and heated in afurnace for one hour at 500°C to removeorganics. Upon removal it was weighed,placed back into the furnace for one hour at1000°C to remove carbonates, removed andweighed again.

Description, Photos, M.S.

Immediately after sampling, each core wassplit in half, described in detail, andphotographed at close range in smallincrements. Three magnetic susceptibilitymeasurements were made every 4cm, and theresults were averaged for each interval.

Exchangeable Cations

Samples approximately 10g each (dry weight)were taken from the cores at 40cm intervals.Samples were dried for at least 24 hours in an

oven at 40°C before being weighed, andblanks were used for comparison.

Exchangeable Acids

Samples were placed into a flask with 25ml1N KCl, and set on a slow shaker table for 30minutes. This solution was vacuum-filtered,rinsing with KCl to bring total volume to150ml. Total exchangeable acidity wasdetermined by titration with 0.1N NaOH to thefirst permanent pink endpoint. ExchangeableH+ acidity was determined by adding 10ml 1NKF and titrating with 0.1N HCl until colorless.

Exchangeable Bases

Samples were placed into flasks with 40ml of1N NH4Cl, and left on a shaker table for onehour. The solution was then vacuum-filtered,rinsing with an additional 60ml of NH4Cl intoa 100ml volumetric flask.

Samples were analyzed for Ca, Mg, Na, and Kusing a Perkin Elmer Model 3030 AtomicAdsorption Spectrometer. Standards of 1, 2, 5and 10 mg/L were made up for Na,K andCa,Mg analysis, with 20ml 5% lanthanumchloride. Dilutions of each solution to .8, .1and .01 were also made for use in the Ca/Mganalysis, again using 2ml 5% lanthanumchloride. Each sample was then analyzed byeither absorption (Ca,Mg) or emission (Na,K).

Grain Size and Clay mineralogy

1cm3 samples were taken from the B coreevery 4cm, suspended in deionized water, andanalyzed with a PC-2000 Spectrex LaserParticle Counter to record mean grain size.

Slides were prepared by separating the

Extractable Metals

Nitric acid extraction was used to preparesamples to be analyzed in an ICP-MS. Eachsample (approximately 5g dry weight) wasdried for 24 hours in a 50°C oven, weighed,and placed in a beaker on a slow shaker tablewith 50ml 0.1N HNO3 for 24 hours. Sampleswere then vacuum filtered and analyzed for Li,Mn, Sr, Mg, Na, Si, Fe, Al, Zn, B, and Ba.

RESULTS

Description, Grain Size, Loss onIgnition, Magnetic Susceptibility, andClay MineralogyA graphical summary of results from theseanalyses is presented in Figure 2. The X-raydiffraction data shows evidence for thepresence of Illite, Chlorite, and Kaolinite.

Exchangeable CationsWhile the exchangeable acidity waseffectively zero, exchangeable bases weremore interesting. K and Na did not fluctuatemuch, Ca and Mg showed definitecorresponding peaks. When added together torepresent total cation exchange capacity(CEC), the pattern becomes even more clear(see Figure 3). The top of the core shows thehighest capacity (just over 30meq/100g).From the high values, it drops gradually untiljust after the 400cm mark, and then peaksagain at 550cm. CEC then dropsprogressively before steadying out around15meq/100g for the lower 3m of core.

Radiocarbon datingCalibration sets for the time periodsurrounding the first sample are less welldeveloped than others, and so the calibratedages must be taken with this in mind.

Lab # Depth(cm)

YearsBP

CalibratedYears BP

AA45071 461-462 12,324±80

13,087-12,262

AA45072 671-676 15,350±100

16,687-16,101

Extractable Metals

DISCUSSIONA few trends become clear by comparing thisstudy with previously dated trends in climate.Based on two radiocarbon dates asbenchmarks for timing, most of the data fitspatterns which can be correlated withpreviously identified climate events. Grainsize consistently increases with depth, with afew prominent spikes in the bottom twometers. This sediment was most likelydeposited near the time of the maximumadvance, which could account for its coarsergrain size. In addition, the presence of anyloess sequences (perhaps the peaks at thosedepths) would also be represented by areas ofcoarser grained material.

Magnetic susceptibility and loss on ignitionare good illustrators of the younger half of thecore. The most noticeable increase in MSoccurs at approximately 5.75m. In previousstudies, such increases in MS have been linked

Figure 2. Comparison of Loss on Ignition,Magnetic Susceptibility, Grain Size, and descriptionby depth.

Figure 3. Total Cation Exchange Capacity by depth,including radiocarbon dates.

with well developed B soil horizons, andtherefore a warmer climate. Based on ourdates, this is consistent with a warm period asdescribed by Yu and Wright (2001).Sometime after this (slightly above the firstradiocarbon date at 4.6m), the organic contentof the section begins to increase, alsosuggesting a warming of the climate toproduce more intra-basin organic materialinstead of the primary signal coming fromclastic extra-basin sediment. This trendcontinues until the top of the core, where thereis a particularly high amount of organiccontent associated with the peat present there.

Finally, cation exchange capacity supportsdata from the other methods of analysis. Upuntil the lower radiocarbon date, totalexchange capacity remains relatively low andsteady. Shortly after that mark at 15,350 BP,however, there is a dramatic increase incapacity, which can be associated with thegeneral warming and glacial retreat at thetime. After peaking at about 5.5m, capacitydrops again just after the higher date (12,324BP), which could represent cooling at the endof the Bølling-Allerød into the YoungerDryas. Subsequent Holocene warming couldalso account for the following rise in capacitythroughout the rest of the core.

CONCLUSION

In conclusion, it is clear that there is a basisfor the use of geochemistry as an indicator ofglobal scale climate change when applied tolocal scale sediment samples. Using cationexchange capacity, loss on ignition and grainsize, previously determined climate eventsassociated with the Laurentide Ice Sheet andlate Wisconsin glaciation could be potentiallyidentified and correlated with two sedimentcores from a small basin. Changes ingeochemistry throughout the cores suggestwarm and cold intervals that may correspondwith the Younger Dryas and the Bølling-Allerød. Although this study is not used forthat purpose, this data suggests that geo-chemistry could be used along withradiocarbon dating to create a much higherresolution record of climate oscillations duringthe late Pleistocene and early Holocene.

REFERENCESBeierle, B and Smith, D.G., 1998. Severe

drought in the early Holocene (10,000-6,800 BP) interpreted from lake sedimentcores, southwestern Alberta, Canada.Palaeogeography, Palaeoclimatology,Palaeoecology 140, 75-83.

Ding, Z.L. et al, 2000. Geochemistry of thePliocene red clay formation in the ChineseLoess Plateau and implications for itsorigin, source provenance andpaleoclimate change. Geochimica etCosmochimica Acta, 65, 6, 901-913.

Ekberg, M.P. et al, 1993. Late Wisconsinglacial advance and retreat patterns insouthwestern Ohio, USA. Boreas 22, 189-204.

Lowell, T.V., 1995. The Application ofRadiocarbon Dates to the Dating ofGlacial Sequences: An Example from theMiami Sublobe, Ohio, USA. QuaternaryScience Reviews, 14, 85-99.

Yu, Z. and Wright, H.E. Jr., 2001. Responseof interior North America to abruptclimate oscillations in the North Atlanticregion during the last deglaciation. Earth-Science Reviews 52, 333-369.

Figure 4. Extractable metals; Mg and Feroughly follow the pattern seen in cationexchange capacity.

THE HISTORY OF THE LAURENTIDE ICE SHEET: ADETAILED SEDIMENTARY ANALYSIS OF A BOG, SW

OHIO

JOEL BYERSDORFERGeology Dept., Whitman College

Sponsor: Patrick Spencer

INTRODUCTIONNumerous studies have examined the history ofLaurentide ice sheet (Lowell, 1995; Shane1987;and Ekberg et al., 1993). During the last glacialmaximum, the ice sheet covered much of NorthAmerica, advancing as far south as the vicinityof Cincinnati, Ohio.

Much recent study has focused on paleoclimatereconstruction and the relationship betweenclimate and ice sheet dynamics (Broecker andDenton, 1989 and Lowell et al., 1995).

The goal of this study is to establish a record ofpaleoclimate and ice sheet behavior since thelast glacial maximum approximately 20,000years ago by extracting paleoenvironmentalinformation from a sediment core.

METHODSCoring sites were selected based on theirlocation relative to the margin of maximum icesheet advance as well as their suitability forpreserving sedimentary records. Ten sets ofcores were gathered from ten sites in the areaSE of Dayton, Ohio (Fig. 1).

Coring was accomplished using a 2-inchdiameter manual piston-coring device. Thisdevice allowed the retrieval of fine sedimentsonly; it was not possible to core material coarserthan medium to coarse grained sand. Coreswere retrieved in approximately 1 metersegments with each thrust.

Often, less than one meter was retrieved perthrust. Possibly, sediment from the bottom ofthe thrust fell into the hole. The mostuncertainty resides at the top and bottom ofindividual core sections. For this reason, twoadjacent holes were cored at each site in thehope that by staggering thrusts in each hole, acontinuous stratigraphic record could beretrieved. One hole was begun 1 meter belowthe surface and thrusts taken at successive onemeter intervals. The adjacent hole would thenbegin at 1.5 meters and thrust at 1 meterintervals as well. In theory, cores from eachhole fill the gaps from the other.

Detailed descriptions were made for eachsection, including sedimentary structures,contacts, and color. Magnetic susceptibilityreadings, grain size measurements and loss onignition tests measuring carbonate and organiccontent were taken every 4 cm.

Grain size was calculated by measuringrefraction patterns of a laser directed through asediment suspension of predetermined dilution.

Macrofossil analysis from two core locationsper site allowed retrieval of organic material insufficient quantities to make radiocarbon datingpossible.

In order to accurately interpret data, the coresfrom each site must be correlated with eachother in order to obtain absolute stratigraphicdepths. To do this one must arbitrarily assignan absolute depth to one section first, and then

correlate cores matching contacts andsedimentary structures.

RESULTS AND DISCUSSION

Age ConstraintsThe two radiocarbon dates in this study arefrom stratigraphic depths of 4.5 m and justunder 8 m. The 8 m date is 16,170±97 yrs BP(19,286±310 yrs BP calibrated, Lab #AA45079), and the 4.5 m date is 13,370±280yrs BP (16,066±390 yrs BP calibrated, Lab #Beta-158295). The upper date corresponds tothe very distinct contact at 4.5 m whereas thebottom date corresponds to date closest to thebog bottom (Fig. 2).

Time represented by this 3.5 m package ofsediment is approximately 3000 years. Eachcentimeter of sediment represents a little lessthan 10 years. The four centimeter samplinginterval used for grain size, loss on ignition, andmagnetic susceptibility cannot be expected toyield yearly data trends, but broader trends onthe order of fifty, or hundreds of years.

Sedimentology and Sand AnomalyThe top meter of the core, beginning 1.5 mbelow ground surface, is a dark brown peat unitcontaining seeds and other organics (Fig. 2).Below, for the next 0.5 m is a banded marl unitcontaining snail shells. From ~3.3 m to 4.5 m isa banded silt and organic layer, capped at thebottom by a 0.25 m gelatinous gyttja layer.From 4.5 m to 7.5 m is a massive silt layer withoccasional darker mottling. The last 0.75 mcontains a silt layer above sand layer (more onthat later). The bottom 0.1 m of the core is amix of shell fragments, sand, silt and organics.

This core, 0110, was unique with respect tocores from other sites on this project in that itwas the only one containing a coherent layer ofsand.

This layer of sand fines upward into gravel justabove 8 m. It contains shell fragments and isabove a silty layer with organics. Assumingthat regional sedimentary processes aredominated by Laurentide Ice Sheet influence,the age of the sand layer is about 20,000 years,which puts it at the same time as maximumglacial advance. At this time, the depositional

environment was probably higher energy,compared to silts and clay that dominate mostof the sedimentary record.

One possible explanation for the sand layer isthat the basin represents an eddy off of a higherenergy meltwater channel. Water andsediments were diverted into the basin duringhigh flow where a sudden loss of competencycaused deposition. Another possibility is thatthe basin was initially isolated from meltwatersources. As the outwash level rose and/or theedges of the basin eroded, the basin wasexposed suddenly to an influx of moving waterand sediment. After the sudden initial influx,the depositional environment transitioned into alower energy environment. Due to the sharpnature of the contact, the isolated basin seemslike the most likely scenario.

Grain sizeThe sediment column displays an upward finingtrend. This makes sense in terms of a retreatingice sheet. There are large mean grain sizefluctuations between 4.5 m and 8 m which showa decreasing minimum grain size. Thesefluctuations could be due to the dynamic natureof braided outwash streams associated withglacial drainage systems. As the ice sheetretreats, the stream becomes less braided andmore stable. The grain size trend fluctuatesless.

Standard deviation of grain size measurementsreflects the sorting of sediment. Higherstandard deviation in grain sizes signifies poorersorting. Thus, a greater degree of reworkingwould be shown by better-sorted sediment and alower standard deviation. One might expect tosee a pattern of decreasing standard deviationthrough time as the outwash sediments werereworked and the ice sheet retreated. From 6.5m and 4 m is a trend of lowering standarddeviation. It is not clear whether the largevariations in standard deviation are due toexperimental uncertainty or environmentalconditions.

Magnetic SusceptibilityMagnetic susceptibility measurements (Fig. 2)show high values and large fluctuations at thebase of the column, correlating with the

sand/gravel layer and proximity to the sedimentsource. After 7.4 m magnetic susceptibilityreadings gradually decline up through thesediment column until the present.

Magnetic minerals are generally heavier thannon-magnetic. Following this line of reasoning,one would expect magnetic mineral content todecrease with reduced competence and distancefrom source. The pattern observed in thesecores agrees with a signal that might beexpected from a retreating source withaccompanying lowering competency.

Loss on IgnitionOrganic and carbonate content measurementsare stable up to approximately 16,000 yearsago. At this point, organic content rapidly risesto a higher level accompanied with greaterfluctuations (Fig. 2). The most intriguingaspect of the loss on ignition measurements isthe manner in which the organic and carbonategraphs are near mirrors of each otherthroughout the entire sediment column.

A possible explanation for this mirroringinvolves oxidation associated with increasedorganic content. As organics increase in thebasin, possibly as algae or water plants, oxygenlevels decrease via eutrophication. Decay andoxidation of deceased plant matter removesoxygen from the environment. More plantscreate an oxygen poor environment inhospitableto animals. At many sites, some snail shellsmade of calcium carbonate are preserved. LOIdata from these sites exhibit various degrees ofmirroring in carbonate versus organic graphs.

It is reasonable to assume that amount of plantmaterial a given environment supports is relatedto temperature and aridity. It is likely that whenthe ice sheet reached its maximum extent,climate was at its coldest. As climate warms,the ice sheet retreats and higher organic andcarbonate contents would be expected.However, this trend does not appear in the data.In a regional picture, this trend probably waspresent, but in an outwash/glaciolacustrinesetting, cold water and sediment will continueflowing even after the ice sheet retreats. Thismight be why the ice sheet retreat signal is fardelayed in the carbonate/organic record

compared to grain size and magneticsusceptibility measurements.

CONCLUSIONThe measurements in this core point to aretreating ice sheet by 16,200 yrs BP. This daterepresents a minimum age for ice retreat. In alllikelihood, retreat began before this date. Also,the severe sedimentologic contrasts at 16,200and 13,400 yrs BP indicate significantenvironmental changes. Although the patternsin grain size, loss on ignition, and magneticsusceptibility data display a signature consistentwith ice retreat, their variable nature may beindicative of an ice sheet retreating in surges,maybe even with brief periods of advance.

The sand layer probably represents a suddeninflux of meltwater into an isolated basinfollowed by regular sedimentation. This studyhighlights the importance of differentiatingbetween local environmental changes (sandanomaly) larger scale regional environmentalchanges, such as those that drive continentalglaciations.

REFERENCES CITEDBroecker, W.S. and Denton, G.H., 1989, The

role of ocean-atmosphere reorganizations inglacial cycles: Geochemica etCosmochimica Acta, v. 53, p. 2465-2501.

Ekberg, M.P., Lowell, T.V., and RobertStuckenrath, 1993, Late Wisconsin glacialadvance and retreat patterns in southwesternOhio, USA: Boreas, v. 22, p. 189-204.

Lowell, T.V., Heusser, C.J., Andersen, B.G.,Moreno, P.I., A. Hauser, L.E. Heusser, C.Schlüchter, Marchant, D.R., and Denton,G.H., 1995, Interhemisperic correlation oflate Pleistocene glacial events: Science, v.269, p. 1541-1549.

Lowell, T.V., and Savage, K.M., 1990,Radiocarbon analyses from Cincinnati, Ohioand their implications for glacialstratigraphic interpretations: QuaternaryResearch, v. 34, p. 1-11.

Shane, L.C.K., 1987, Late-glacial vegetationaland climatic history of the AlleghenyPlateau and the Till Plains of Ohio andIndiana, U.S.A: Boreas, v. 16, p. 1-19.

Fig. 1 Location of core sites, 2001 Ohio Keck. Study site indicated by bullseye symbol. Moraine complex ofLaurentide Ice Sheet is visible enlarged portion.

Fig. 2. Data for core 0110. All depths are correlated stratigraphic depths. Left column identifies composition ofsediment column. Note mirroring in organic and carbonate content.

THE POLLEN RECORD OF THE GLACIAL-INTERGLACIAL TRANSITION IN STEVENSON’S BOG,

OHIO

MONICA KAITZGeology Dept., Amherst College

Advisor: Ed Belt, Amherst College

INTRODUCTIONThe Laurentide ice sheet began its retreat fromsouthern Ohio approximately 18,000 years bp.Associated with this retreat are a series ofconspicuous recessional moraines, outwashdeposits and topographic depressions.Stevenson’s bog, a closed basin of 0.5 km2

located on the western side of the Great Lakesoutwash plain was selected to examine theclimate record during the last deglacial toHolocene transition. Grain size, magneticsusceptibility, loss-on-ignition, radiocarbonand pollen content were used to assess pastenvironmental changes recorded in a sedimentrecord extracted from Stevenson Bog. Thebog is located in the valley south of the St.Johns and north of the Union moraines amidclassic outwash features on the Indian lakefloodplain (Fig. 1). Indian Lake is a vestigialproglacial ice-contact lake on the Erie Sublobeoutwash plain. It is underlain by ice-recessional sediment over argillose glacial tillthat covers Silurian limestone bedrock. Thecalcareous bedrock results in a surface andground water classified as very hard, calcium-magnesium-bicarbonate waters (Debrewer etal., 2000). Typical soil characteristics for thisregion are permeable calcareous silt loams.

Description of the Core Two cores were taken near the center of thebasin 3 meters apart using a modifiedLivingston corer. Three units have been identified. They correlate well across the two

Fig. 1. Surficial Map of Scioto and Miami sublobesof the Erie glacial lobe. Stevenson Bog designatedby a point marked STB (modified from Leverett,1902)

cores, having the same lithological units andthicknesses. The base of unit 3 at 541cm depthconsists of massive mottled clay with sandylayers and large pebbles coarsening in grainsize as one moves up core. There are a feworganic stringers (Figure 2). At 382 cm depth,unit 3 turns into a massive dark green organicsilt which grades upward into the bandedgreen grey silts of unit 2 at 379 cm depth(Figure 2). The bands become more red andbrown and grade up into the dark fibrouscompact peat of unit 1 at approximately 248cm. These units can be interpreted as atransitional sequence corresponding to glacialoutwash deposits at the bottom transitioninginto lake deposits and finally a modern peatbog.

AnalysesTwo AMS radiocarbon dates were obtainedfrom the Arizona AMS laboratory: oneconstrains the depth interval 358-374 cm to13,490+/-110 yrs bp (AA40576) and the otherconstrained the depth interval of 448-452 cmto 14,360+/-120 yrs bp (AA40577). The latterAMS date provides the best estimate of whenthe ice left the site or the time of deglaciationin the region.

The grain size analysis was done using a laserparticle counter at evenly spaced intervals of 4centimeters. Magnetic susceptibility (MS) wasmeasured before the core was split anddescribed using a field susceptibility meterconnected to an ‘F-probe’. The core was againmeasured at regular 4 cm intervals, three

readings per depth. The susceptibility graphplots the mean of the three measurements andreinforces the grain size data. Loss-on-Ignition(LOI) was done at the same 4 cm intervals.Total organic carbon was obtained frompyrolysizing the sample at 550°C for an hourand carbonate content was determined after a1000°C one hour burn.

Samples for palynological analysis were takenat 20 cm intervals to achieve a coarse broadspectrum of pollen fluctuations. Subsequentsamples were then selected for the purpose ofconstraining warming and cooling periodsreflected in the pollen percentage shifts.Although samples required different treatment,all were treated with hydrofluoric acid andglacial acetic acid, in order to dissolve theclastic and organic materials. The fossil grainswere then mounted on slides and identifiedbased on surface sculpturing and number ofapertures.

Interpretation of SedimentologicalAnalysesThe data show a de-glaciation sequence,generally fining up core in average mean grainsize (Figure 2). The coarser layers (fine sand)interspersed with mottled clay of unit 3probably correspond to glacial deposits.Average mean grain size shows much morevariation from 541cm to 370 cm than above370 cm due to changes in lithology: the unithas a matrix of massive silt but containsinclined fine sand layers indicating higherenergy in transport. The general fining

upwards sequence the length of the coreindicates an overall reduction in energy.

Fig. 2. Graphical representation of core units andgraphs of LOI, MS, and mean grain size (average ofthree measurements per depth). Radiocarbon datestaken from Limnological Lab. RCD 1: 13, 490+-110C-14 yrs BP. RCD 2:14, 360 +-120 C-14 yrs BP.

Fig. 4. Pyle Site summary pollen diagram. Note Piceaincrease climate reversal event ca.320cm (modifiedfrom Yu and Wright Jr., 2001 and reprinted fromShane 1987)

The fluctuations in magnetic susceptibility(MS) are even more pronounced andcorrespond to approximately the same depthinterval as the grain size changes. Thedecrease in MS up core demonstrates a declinein clastic sediment influx (with iron andmagnesium bearing minerals) and an increasein organic input into the basin as the glaciermigrates north.

The LOI analysis follows a similar trend aswell with low organics during the glacialperiod and a steady accumulation as the lakeshallows into the relatively dry peat bog it is

today. At a depth of approximately 390 cm arapid increase in organics reflects thetransition from outwash plain to accumulatinglake.

There are three possible explanations for thesedata (1) a transport-related shift in lake levelshoreline changed the distance sedimenttravels to be deposited at the core site (2)variation in sediment load into the lake shiftthe source or the location of the ice sheet, and(3) wind patterns influenced by climate andthe behavior of the glacier.

There is a general progression from fine sandat shorelines to fine muds in the center of mostwater bodies. Thus, grain size fluctuations

could be explained by shifts in the watershoreline based on meltwater input and theevaporation regime both influenced by climate(Menking, 1997). A study of the StevensonBog fluctuations provides an indirect study oftransport distance of sediment into the basinand the rate of glacial retreat. A shift inkatabatic winds off the ice sheet is anotherpotential influence. As the ice sheet recedes,the change in depth causes a tunneling effectwhich changes the wind patterns and increasesvelocity. Greater winds are capable or carryingheavier minerals into downwind basins.

Interpretation of Palynology ResultsThe use of pollen data as a proxy for climateassumes the following: (1) for pollen to beuseful in climate reconstructions, vegetationmust generally be in equilibrium spatially andtemporally with climate and (2) fossil dataoscillations are only attributable to climatechange with extraneous factors accounted for(Bartlein et al., 1984). The use of largespecial and temporal scales satisfies these twoconditions by reducing the impact of non-climatic variables (Bartlein et al., 1984).Spatially, Stevenson’s Bog is rather small aswater bodies go, and pollen signals are thusmore susceptible therefore to fluctuation fromlocal stimuli. However, using data from othercores in the region would allow forgeneralizations of my results on a regionalscale as. The Stevenson site pollen profilecorrelates well with other sites in the vicinitysuch as the Pyle site, so chronologies fromthese sites can be used in conjunction with thetwo AMS dates obtained in this study toestablish a local chronology.

A network of radiocarbon dated pollenprofiles is available for both the Mid-West andthe Northeast (Bartlein et al., 1984). Fewersites have been investigated in Ohio. Broadregional trends are: (1) a spruce-dominatedboreal forest or woodland moving north in theearly-Holocene and retreating south after 3000yr B.P and (2) a prairie-forest border movingeast into Wisconsin by 8000 yr B.P. and westagain after 6000 yr B.P. (Bartlein et al., 1984).

Preliminary pollen frequencies show thegeneral trend of a spruce-dominatedassemblage in older sediments with increasing

Fig. 3. Histogram of selected preliminary pollen datafrom Stevenson Bog core B to illustrate Picea reversal.AMS date (AA40576) noted.

amounts of hardwood and pine (Figure 3) asone moves upcore. Spruce (Picea) is a conifertypical of a cooler northern climates while ash(Fraxinus) is a deciduous tree characteristic ofsomewhat warmer temperatures. It is one ofthe first plants to react to the change inclimates. There is a reversal in the decline inspruce by 310 cm reflecting a cooler climate-probably the Younger Dryas event. A similarreversal at the Pyle site is interpreted as thislate glacial climatic event. The pollenstratigraphy for this portion of the StevensonBog core is consistent with the ‘A’ (spruce)zone of Davis (1960). Continued warmingeventually led to the oak/hickory dominateddeciduous forest (Davis ‘C’ zone). There areindications of an initial decline in spruce andincrease in ash at about 360 cm; then a sharpincrease in ash and decrease in spruce at 340cm. This probably corresponds to what is nowrecognized in Northeastern and Europeanpollen profiles as associated with the Allerodwarming event.

Davis' A-zone is characterized internally by aninitial decline in spruce and rise in ash.Samples higher and lower in the core than thedepths listed in Figure 3 are currently beingprocessed, but one would expect T-zone, agrass, Artemisia, sedge and low arborial treesdominated tundra assemblage in the oldestdeposits and B zone with oak and hickoryspecies above.

The pollen evidence is conclusive indemonstrating a cold reversal, however,although it corresponds to the Younger Dryasinterval, there multiple possibilities as to theorigins of this event. The cold reversal may bedirectly related to a climate shift or it maystem from a localized lake-induced coolingeffect. There is isotopic evidence of a shift inthe drainage patterns of Lake Agassiz duringthat time interval (10,000 to 12,000 yrs bp)from discharging into the Mississippi todischarging into the Great Lakes causing alocalized atmspheric cooling (Lewis andAnderson, 1991).

CONCLUSIONS:The analyses demonstrate a de-glaciationsequence which supports data previouslygathered at sites in the vicinity. A climatic

cooling event is expressed in the pollen profileand further pending AMS dates instrumentalin constraining the depth interval, will becompared to other sites.

REFERENCES CITED:Bartlein, P., Webb III, T., Fleri, E., 1984,

Holocene climatic change in the NorthernMidwest: pollen-derived estimates:Quaternary Research, v. 22, p. 361-374.

Davis, M., 1960, A Late Glacial PollenDiagram from Taunton, Massachussetts:Bulletin of Torrey Botanical Club, v. 87, n.4, p.258-270.

Debrewer, L., Rowe, G., Reutter, D., Moore,R., Hambrook, J., Baker, N., 2000,Environmental Setting and Effects onWater Quality in the Great and LittleMiami River Basins, Ohio and Indiana inNational Water-Quality AssessmentProgram, Water Resources InvestigationReport 99-4201, U.S. Department of theInterior, p.1-43.

Lewis, C., Anderson, T., 1991, Stable isotope(O and C) and pollen trends in easternLake Erie, evidence for a locally-inducedclimatic reversal of Younger Dryas age inthe Great Lakes basin: Climate Dynamics,v. 6, p. 241-250.

Leverett, F., 1902, Glacial formations anddrainage features of the Erie and Ohiobasins, Washington Government PrintOffices, p.305.

Menking, K., 1997, Climatic signals in claymineralogy and grain-size variations inOwens Lake core OL-92, southeastCalifornia, in Smith, G., and Bischoff, J.,eds., An 800,000-Year PaleoclimaticRecord from Core OL-92, Owens Lake,Southeast California: Boulder Colorado,Geological Society of America, SpecialPaper 317, p. 25-36.

Yu, Z., Wright Jr., H., 2001, Response ofinterior North America to abrupt climateoscillations in the North Atlantic regionduring the last deglaciation: Earth-ScienceReviews v.52 p.333-369.