Wolfe Holocene 2003

The Holocene 131 (2003) pp 29ndash37

Diatom community responses tolate-Holocene climatic variabilityBaf n Island Canada a comparisonof numerical approachesAlexander P Wolfe

(Department of Earth and Atmospheric Sciences University of AlbertaEdmonton AB T6G 2E3 Canada)

Received 13 June 2001 revised manuscript accepted 11 February 2002

Abstract A high-resolution diatom stratigraphy from the late-Holocene sediments of a small Arctic lake onBaf n Island (Nunavut Canada) has been analysed by several numerical methods and considered in relationto independent palaeoclimatic proxies to enable an objective assessment of the response of freshwater diatomcommunities to climatic change Diatom relative frequencies and absolute abundances were subjected to severalmanipulations including application of a weighted-averaging transfer function for summer water temperatureconversion of valve concentrations to cell biovolumes biostratigraphic rate-of-change and taxonomic richnessestimated by rarefaction analysis The major directions of variability within the derived diatom data wereexplored using principal components analysis Diatom-inferred lakewater temperatures and indices of diatompalaeoproductivity declined during late-Holocene cooling and they faithfully record post-lsquoLittle Ice Agersquo warm-ing Between ~2000 cal BP and the lsquoLittle Ice Agersquo which includes the coldest intervals of the Neoglacialdiatom assemblages diversi ed and destabilized re ecting a state of intermediate disturbance Together theseresults examine different methods of interpreting freshwater diatom stratigraphic data with the objective ofgaining palaeoclimatic insights from fossil assemblages

Key words Diatoms palaeolimnology palaeoclimate Neoglacial numerical methods principal componentsanalysis late Holocene Baf n Island Canadian Arctic

Introduction

Diatom-based palaeolimnological records offer considerablepotential for the reconstruction of submillennial-scale Holoceneclimatic variability in part because diatom communities respondrapidly to limnological changes that are directly or indirectlyin uenced by climate The rst efforts to reconstruct palaeocli-matic trends from diatoms invoked indirect relationships involv-ing the climatic control over for example ice cover in high Arcticlakes (Smol 1983) and evaporative solute enrichment in closed-basin systems (Fritz et al 1991) More recently several investi-gators have successfully related diatom assemblages to eitherwater or air temperatures using the training-set approach (Birks1995) and statistical techniques such as weighted averaging (WA)regression and calibration (Pienitz et al 1995 Weckstrom et al1997 Joynt and Wolfe 2001) WA partial least squares (Lotteret al 1997 Rosen et al 2000) and combined or consensusapproaches (Korhola et al 2000) Although most of these studies

Author for correspondence (e-mail awolfeualbertaca)

Oacute Arnold 2003 1011910959683603hl592rp

concede that the direct physiological relationships linking diatomcommunities to temperature remain poorly de ned the statisticalmodels appear to perform acceptably because of the intimatecoupling of climate with several limnological properties of pri-mary importance in shaping the structure of diatom communities(eg nutrient uxes dissolved organic carbon dynamics acidity)Despite the generally perceived success of these approaches theiracceptance has not been universal because of scepticism sur-rounding the ability to isolate a temperature signal from multiplecontrolling factors (Anderson 2000)

Perhaps a more cautious approach to validating the true poten-tial of diatoms as palaeoclimate proxies is the direct comparisonof dated diatom stratigraphic results to independently measured(Korhola et al 2002) or inferred (Anderson et al 1996) climaticgradients Unfortunately in these examples the time-frames con-sidered are relatively short so that anthropogenic in uences maycomplicate the task of attributing diatom assemblage changes topurely climatic forcing The present study continues to exploit thisgeneral approach but over the temporal framework of the past ve millennia and for an Arctic region lacking any direct anthro-

30 The Holocene 13 (2003)

pogenic in uences Small upland Arctic lakes offer additionaladvantages for studies aiming to characterize the climatic controlover diatom assemblages These include reduced trophic com-plexity in the absence of sh and the lack of groundwater in u-ences due to continuous permafrost Although non-climatic factorsassociated with ecosystem ontogeny are known to in uence waterquality and hence diatom assemblages (Engstrom et al 2000) thepresent study addresses a lake having simpli ed soil and veg-etation dynamics in the presence of sparse tundra Furthermorethe longer palaeoenvironmental record suggests that edaphic fac-tors were largely stabilized by the middle Holocene (Wolfe et al2000) in turn maximizing the sensitivity of late-Holocene diatomassemblages to climatic in uences

Study site

Fog Lake (67deg119 N 63deg159 W) is situated at 460 m asl onthe north coast of Cumberland Peninsula Baf n Island NunavutCanada (Figure 1) The lake has a surface area of 0015 km2 anda maximum depth of 95 m The Fog Lake watershed is small(23 km2) occupying an upland that separates two major ordsystems (Padle and Kangert) directly above the highest morainesdated to the last glacial cycle (Steig et al 1998) As a result thelake preserves an unusually long sediment record that includes atleast the latter portion of oxygen isotope stage 5 (Wolfe et al2000) Lakewaters are extremely dilute with electrical conduc-tivity of 10 mS cm21 and a summed concentration of base cations(Na+ + K+ + Mg2++ Ca2+) of 30 mg L21 Summer epilimneticpH is 63 In most years the lake is ice-free from mid-July to mid-September However cold summers may result in the persistenceof a vestigial ice pan throughout the summer

The climate of northern Cumberland Peninsula is severeespecially along the outer coast At Qikiqtarjuaq (formerlyBroughton Island) 40 km northwest of Fog Lake mean annualtemperature is ndash113degC (summer (JJA) mean 12degC winter(DJF) ndash176degC) Annual precipitation at Qikiqtarjuaq averages290 mm a21 of which 85 falls as snow but precipitationincreases dramatically eastward along the coast at Cape Dyer100 km east of Fog Lake mean annual precipitation rises to 660

85 W 80 W 75 W 70 W 65 W

65 N

70 N

BaffinBayBaffin

IslandFoxeBasin

HudsonStrait

400 km

Ungava

Canada

USA

Greenland

1000 km

FogLake

Figure 1 Location map of Fog Lake on the north coast of CumberlandPeninsula Baf n Island Circles indicate the locations of the 61 modernsurface-sediment samples used to develop the summer water-temperatureinference model of Joynt and Wolfe (2001)

mm a21 Local vegetation in the Fog Lake catchment is limitedto a discontinuousheath tundra dominated by Cassiope tetragonaand Empetrum nigrum

Late-Holocene climatic evolution of theBaf n Bay region

The late Holocene in the Baf n Bay region can be characterizedas an interval of progressive summer cooling as inferred bymultiple lines of isotopic palynological and glacial-geologicalevidence (Figure 2) Part of this trend appears to be ultimatelyregulated by decreasing high-latitude summer insolation TermedNeoglacial for associated advances of many local glacial systemslate-Holocene cooling culminated in the lsquoLittle Ice Agersquo (LIA)ad ~1450ndash1850 Since the end of the LIA summer temperatureshave increased rapidly across the Arctic (Overpeck et al 1997)a trend that has probably been ampli ed in recent decades byanthropogenic greenhouse-gas forcing Baf n Island appears toconform closely to this general pattern of climatic evolution(Davis 1985 Hughen et al 2000 Moore et al 2001) Althoughmost of the regionrsquos late-Holocene palaeoclimatic records fromthis region are expressed as relative departures or anomalies theactual amplitude of late-Holocene cooling is probably about 3degC(eg Andrews et al 1981 Short et al 1985 Dahl-Jensenet al 1998)

Methods

Coring and sediment chronologyThe core considered in this study was obtained from the deepestpart of Fog Lake in August 1997 with a modi ed Kajak-Brinkhurst gravity corer (Glew 1989) The 345 cm core wasextruded vertically in the eld in consecutive 05 cm incrementsto eliminate any potential disturbance The corersquos chronology isbased on ve accelerator mass spectrometry (AMS) 14C datesfrom sediment humic acids extracted according to Abbott andStafford (1996) The humic acid 14C ages were corrected byadding 300 years to account for the average lag in the transportof terrestrial dissolved organic carbon to lake sediments (Milleret al 1999) and then calibrated to calendar years before present(Stuiver et al 1998) The resulting age model (Figure 3) is ttedwith a polynomial equation (r2 = 099) that enables the reliableconversion of sediment depth to age and results in an age of 4800cal BP for the base of the core The dating results in addition tothe corersquos uniform visual stratigraphy (olive silty gyttja) implythat the sedimentation rate has varied little over the interval ofdeposition and that it has not been punctuated by low-frequencyhigh-intensity events This is an important consideration withregard to the interpretation of diatom concentration and ux data(Wolfe 1997)

Diatom primary dataDiatom slurries were prepared from each of the 69 samples byoxidation of organic matter in hot 30 H2O2 Slurries were spikedwith a known concentration of marker grains (Eucalyptus pollen)in order to estimate diatom concentrations and then diluted Ali-quots of 200 mL of the diluted slurries were settled ontocoverslips Once air-dried these were mounted on microscopeslides with NaphraxOgrave Diatoms (and markers) were counted fromeach interval in random transects including coverslip edges Dia-tom taxonomy followed primarily Hustedt (1959) Patrick andReimer (1966 1975) Foged (1981) Germain (1981) Krammerand Lange-Bertalot (1986 1988 1991a 1991b) and Camburnand Charles (2000) Diatom sums ranged between 485 and 562

Alexander P Wolfe Diatom community responses to late-Holocene climate ndash numerical approaches 31

Schematic time-distance

Pangnirtung Pass cirque glaciers

advance retreatcold warm

-1 0 +1

Arctic Canada pollen transfer functions

Summer departures(standard deviations)

-30 -275 -25

0

1000

2000

3000

4000

5000

6000

360 380 400

Devon Island Ice Capd18O (permil SMOW)

Summer insolation 65N (W m-2)

Yea

rs B

P

DB C

-1 0 +1

Circum-Arctic summer anomaly (SD units)

2000

1900

1800

1700

1600

1500

Years A

D

2 3 4 5 6 7

S Baffin spring air temperature (C)

A

Upper Soper Lakevarve thicknesses

E

Figure 2 Synthesis of the late-Holocene climatic evolution of the Baf n Bay region with speci c reference to the last 6000 (AndashC) and last 500 (DndashE)years Illustrated records include (A) summer insolation (Berger and Loutre 1991) and the Devon Ice Cap d18O record (Paterson et al 1977) (B) summeranomalies from a synthesis of pollen data from the Baf n Keewatin and Labrador sectors (Andrews et al 1981) (C) schematic time-distance for Neoglacialcirque glacier activity in Pangnirtung Pass (Davis 1985) (D) spring air temperatures on southern Baf n Island (Kimmirut) from varve thicknesses (Hughenet al 2000) and (E) the multiproxy pan-Arctic reconstruction of summer-temperature anomalies for the last 400 years (Overpeck et al 1997)

r2=099

0 2000 4000 6000

0

5

10

15

20

25

30

35

40

age = 071z2 + 11559z + 579

Dep

th (

cm)

Corrected calibrated 14C age BP

Figure 3 Age-depth relationship for the Fog Lake gravity core

valves Tallies for individual taxa were converted to percentagesrelative to the sum of valves counted

Derived diatom dataThe primary diatom data (ie relative frequencies) were manipu-lated in several ways resulting in a second data set of sevenderived variables First a summer water-temperature transferfunction based on WA calibration (Line et al 1994) was appliedto the downcore diatom assemblages This inference model isbased on a training set including 61 lakes and 107 taxa with aroot mean squared error (RMSE) of prediction for summer watertemperature of 194degC (bootstrappedRMSE is 274degC) Completedetails of the training set are presented elsewhere (Joynt andWolfe 2001) Because summer lakewater temperature varies lin-early with locally measured air temperature (Livingstone et al1999) the directions of change in reconstructed summer watertemperature are believed to represent a rst-order response to cli-mate

Second sediment diatom concentrations were estimated fromthe recovery of external markers introduced to the diatom prep-arations (Battarbee and Kneen 1982) These results are tabulatedas the total number of diatom cells per unit dry sediment mass

(frustules gdw21) Using the age model concentrations were alsoconverted to uxes of diatom cells to the sediment surface(frustules cm22 a21)

Because diatom cell sizes vary tremendously even betweencongeneric taxa simple concentrations may produce a misleadingdepiction of total diatom productivity For this reason diatomconcentrationswere converted to biovolumes using taxon-specicmeasurements of whole-cell volumes (Table 1) Although thisapproach is not new to palaeolimnology it has only rarely beenimplemented (Battarbee 1973 Anderson 1994) despite the factthat algal biovolumes have long been a standard and routine unitof measure in phytoplankton ecology For most taxa many speci-mens were individually measured although certain dimensions forrarer taxa were extracted from illustrations in the cited taxonomicliterature Volumes were estimated using simpli ed shapes suchas cylinders (for Aulacoseira spp) rhomboids (eg certain Frus-tulia and Stauroneis taxa) ellipsoids (for fragilarioid and mostnaviculoid taxa) and half ellipsoids (for cymbelloid and eunotioidgenera) Thus frustule volumes could be simply estimated by threemeasurements length width and depth Although more sophisti-cated algorithms for cell volumes exist (Hillebrand et al 1999)these would not re ne the computation of whole-assemblage dia-tom biovolumes given the high degrees of within-taxon varia-bility observed (Table 1) The intraspecic variability of cell vol-umes in natural populations is not an artifact of the measurementsbut rather is due to volumetric reductions following vegetativereproduction For the taxa present in Fog Lake frustule volumescover two orders of magnitude from 30 mm3 (Naviculaschmassmannii) to 9600 mm3 (Eunotia monodon) For taxa withpreviously published biovolumes (Battarbee 1973 Reynolds1984 Anderson 1994) the values presented in Table 1 are ingood agreement The range of calculated biovolumes from theFog Lake ora which include the rst such data for a number ofbenthic forms illustrates how sediments with elevated concen-trations of small frustules may potentially re ect lower diatomproductivity than sediments with lower concentrations of morevoluminous cells Biovolumes of individual taxa from eachstratigraphic interval were summed to produce a total diatombiovolume term which is reported in cm3 gdw21 As with thefrustule concentration data biovolume was also converted to itscorresponding ux (cm3 cm22 a21) using the age model


Table 1 Diatom cell biovolumes used in the study Asterisks indicate infrequent taxa that were estimated from the literature The collective categorylsquosmall Achnanthes spprsquo includes A levanderi A altaica A helvetica var minor A marginulata and A saccula

Taxon Cell volume (mm3) 1 SD (mm3) 1 SD (as ) Specimens measured

Eunotia monodon 9603 2261 24 30Pinnularia viridis 5204 1153 22 42Stauroneis phoenicenteron 5000 ndash ndash Pinnularia borealis 4320 ndash ndash Frustulia rhomboides 4158 372 9 37Pinnularia microstauron 3220 359 11 6Cymbella cuspidata 2928 ndash ndash Eunotia triodon 2857 709 25 14Pinnularia biceps var mesongyla 2699 172 6 24Neidium iridis 2252 1166 52 11Surirella pumila 2094 ndash ndash 1Frustulia rhomboides var crassinervia 1980 ndash ndash Stauroneis anceps fo gracilis 1551 480 31 17Eunotia vanheurkii 1536 169 11 56Aulacoseira lirata and fo biseratia 1415 125 9 2Frustulia rhomboides var saxonica 1260 292 23 72Pinnularia biceps 1008 187 19 44Aulacoseira distans var nivalis 853 233 27 73Neidium bisulcatum 622 29 5 4Pinnularia intermedia 621 ndash ndash Caloneis bacillum 607 46 8 3Navicula cocconeiformis 559 77 14 2Brachysira brebissonii 486 148 30 51Cymbella cesatii 485 48 10 2Eunotia bigibba 475 62 13 7Navicula variostriata 471 ndash ndash Cymbella hebridica 412 30 7 18Aulacoseira distans 407 185 45 24Tabellaria occulosa strain IV 370 60 16 34Peronia bula 351 59 17 23Achnanthes helvetica 306 ndash ndash Brachysira vitrea 302 76 25 9Stauroneis kriegeri 272 5 2 3Eunotia denticulata 253 22 9 9Aulacoseira perglabra 216 27 13 9Navicula perpusilla 209 ndash ndash 1Eunotia rhomboidea 205 41 20 66Nitzschia palea 202 69 34 3Nitzschia perminuta 163 43 27 22Navicula pseudoscutiformis 142 ndash ndash Aulacoseira perglabra var oriniae 140 27 19 16Navicula mediocris 133 11 9 18Cymbella gaeumannii 120 9 8 21Eunotia bilunaris 119 65 55 9Aulacoseira distans var nivaloides 117 7 6 2Navicula soehrensis 113 21 19 2Aulacoseira distans var tenella 113 12 10 3Fragilaria virescens var exigua 111 43 39 81small Achnanthes spp 110 25 22 18Eunotia meisterii 102 20 20 11Eunotia polydentula 97 9 10 2Fragilaria pseudoconstruens 96 ndash ndash Fragilaria construens var venter 55 7 12 14Navicula contenta var biceps 43 ndash ndash Pinnularia balfouriana 43 ndash ndash Achnanthes kriegerii 32 ndash ndash Navicula schmassmannii 30 ndash ndash

The last two classes of derived data aim to evaluate diatomcommunity stability and diversity For the former rate-of-changeanalysis (Jacobson and Grimm 1986) was undertaken This issimply a dissimilarity measure applied between successivesamples and normalized to the amount of time elapsed betweenthe deposition of each as inferred from the core chronologySquared chord distance is the dissimilarity metric used Because

rate-of-change is calculated between contiguous samples thenumber of results is one less than the number of initial samplesIn order to include the rate-of-change results in subsequent com-parisons with the other derived variables one additional diatomassemblage was added to the calculations in order to produce 69results This assemblage was taken from the estimated 4800 calBP sample from the Fog Lake long core (Wolfe et al 2000)


Diatom diversity was quanti ed by estimating the richness offossil assemblages (ie the number of taxa present) using rarefac-tion analysis implemented with the program RAREPOLL (Birksand Line 1992) Rarefaction uses a strategy of random selectionwithout replacement typically selecting a sample of the same sizeas the smallest count from the entire population (in this case 485)In this way realistic estimates of richness are produced withoutany bias associated with the variability of individual samplecount sizes

Finally principal components analysis (PCA) was applied to acorrelation matrix based on the collated derived diatom data(seven parameters) for the 69 samples PCA is a common ordi-nation technique for linear indirect gradient analysis (ter Braak1987) which is especially useful for visualizing large and com-plex data sets in reduced dimensions that are often ecologicallyinterpretable PCA was implemented with the MVSP version 20software package (Kovach 1990)

Results

The Fog Lake diatom stratigraphy (Figure 4) is dominated byAulacoseira distans Fragilaria virescens var exigua (= Fragilar-iaforma exiguiformis) F construens var venter (= Staurosiraconstruens var venter) and Frustulia rhomboides var saxonicaAlthough 62 taxa were recorded in the 69 samples the 14 shownin Figure 4 account for no less than 85 of any one assemblageThe high stratigraphic resolution of the analyses reveals remark-able oristic and ecological variability A major reorganization ofthe diatom ora took place ~2000 cal BP At this time Fragilariawas largely replaced by several Eunotia species (E rhomboideaE vanheurkii E bilunaris) and Peronia bula At the same timeAulacoseira perglabra and A distans largely replaced A distansvar nivalis and Pinnularia biceps more than doubled its represen-

Fragilaria virescensvar exigua

Aulacoseira distansvar nivalis

80

40

0

20

80

40

0

5000 4000 3000 2000 1000 0

Years (cal BP)

4

8

0

2

4

0

10

20

0

1

2

0

4

2

0

6

0

5

8

4

0

Fragilaria construensvar venter

Aulacoseira distans

Aulacoseira perglabra

Frustulia rhomboidesvar saxonica

Achnanthes marginulata 6

4

2

0

10

5

0

10

0

20

10

0

20

10

0

Nitzschia perminuta

Brachysira brebissonii

Eunotia rhomboidea

Peronia fibula

Pinnularia biceps

Eunotia bilunaris

Eunotia vanheurkii

5000 4000 3000 2000 1000 0

Years (cal BP)

rela

tive

freq

uenc

ies

(sca

les

vary

)

Figure 4 Relative frequencies of dominant diatom taxa in the Fog Lake gravity core

tation A second major change occurs in the uppermost foursamples of the core that is within the last 150 years Thesechanges include sudden decreases in frequencies of Eunotia sppthe disappearance of Achnanthes marginulata and Fragilaria con-struens var venter from the stratigraphy and expansions of Nitz-schia perminuta and F virescens var exigua The diatom stra-tigraphy suggests a marked lowering of lakewater pH during theinterval dominated by acidophilous eunotioid diatoms (2000ndash150cal BP) a trend that is expressed regionally and may con rm therelationship between climate and lake acidity suggested byPsenner and Schmidt (1992) This subject is dealt with elsewhere(Wolfe 2002) The present analysis concentrates on re ning tra-ditional approaches to the interpretation of diatom stratigraphicdata in order to extract a potentially meaningful palaeoclimaticsignal

The time series for the seven derived diatom parameters arepresented in Figure 5 Late-Holocene summer water temperaturesinferred by WA calibration have an overall amplitude of about43degC Reconstructed lake temperatures decrease progressivelybetween 4800 and 2000 cal BP from a maximum of 82degC toabout 65degC Thereafter this cooling trend accelerates attainingminimum inferred water temperatures (~40degC) between 1500 and500 cal BP Inferred summer water temperatures at Fog Lakeincrease by at least 15degC in sediments deposited since the end ofthe LIA Although it is possible that water-temperature inferencesoverestimate corresponding air-temperature uctuations(Livingstone et al 1999) the overall pattern of late-Holocenewater-temperature changes from the Fog Lake diatoms is entirelycompatible with the independent regional palaeoclimatic proxies(Figure 2)

Diatom frustule concentrationsand uxes also portray progress-ive decreases over the late Holocene although the rapid changeat ~2000 cal BP is not registered Peak concentrations in theorder of 15 3 107 frustules gdw21 were attained in the 4700ndash


0

1

2

3

4

5Diatom biovolume

(cm3 gdw-1)

0

1

2

3

4

5

0

01

02Rate of change

(squared chord distance 100 yrs-1)

0

20

10

Diatom frustule concentrations (x107 frustules gdw-1)

010002000300040005000

Years (cal BP)

Diatom-inferred summerwater temperature (C)

4

6

8

5

7

3

9

0

10

20 Diatom fluxes

(x104 frustules cm-2 a-1)

Diatom volumetric flux

(cm3 cm-2 a-1)

10

15

20

25

30

35Rarefaction estimated richness (E(S485))

Figure 5 Time series of the derived variables from the diatom assem-blages in the Fog Lake gravity core

4200 cal BP interval Although there is some evidence of increas-ing diatom concentrations in the uppermost (post-LIA) sedimentsthe magnitude of this change is not greater than several othersmall-scale uctuations in the record (Figure 5)

When concentrationsof individual taxa are converted to biovol-umes and summed to estimate total sediment diatom biovolumesvalues in the range of 10ndash45 cm3 gdw21 are obtained Whenconverted to their corresponding volumetric uxes the range ofresults is 10ndash50 cm3 cm22 a21 As with frustule concentrationsand uxes the overall trend is one of gradual decrease throughoutthe late Holocene Minimum values in both cases occurred duringthe LIA

The above trends are contrasted by those relating to diatomcommunity stability (rate-of-change) and diversity (rarefaction-estimated richness) These both remain relatively stable between4800 and 2000 cal BP but thereafter increase dramatically(Figure 5) This clearly signi es that as Neoglacial cooling pro-gressed diatom assemblages simultaneously became less stableand more diverse

Collectively the derived diatom data are well summarized byPCA ordination which expresses two well-de ned and interpret-able gradients (Figure 6) The rst axis which accounts for 671

of variance within the derived data is strongly related to diatom-inferred summer water temperature The four parameters associa-ted with diatom concentrations and biovolumes also load posi-tively on this axis In contrast the second PCA axis (125) ismore directly in uenced by the measures relating to diatom com-munity stability and taxonomic diversity The rst axis of PCAtherefore relates most strongly to climate as expressed by the WAtemperature reconstruction as well as to lsquoquantities of diatomsrsquowithin fossil assemblages The second PCA axis relates moreclosely to the lsquoqualityrsquo of diatom assemblages in terms of their oristic richness and community stability These interpretationsfacilitate the discussion of diatom community responses to cli-matic change

Discussion

Climatic change and diatom productivityThe late-Holocene evolution of diatom communities in Fog Lakeis portrayed by the time-series of the 69 sample scores on the rsttwo PCA axes extracted from the derived data (Figure 6) In thisrepresentation the rst axis bears a close resemblance to the gen-eralized pattern of late-Holocene cooling (Figure 2) with highPCA scores associated with greater relative summer warmth and

-05

0

05

10

Years (cal BP)

010002000300040005000

0

Warming

Neoglacial cooling

Little Ice Age

-04 0 04

richnessrate of changevolumetric fluxsummer water temperaturediatom biovolumefrustule fluxfrustule concentration

0 04 06

richnessrate of change

volumetric flux

summer water temperaturediatom biovolume

frustule fluxfrustule concentration

05

A PCA axis 1 (671)

B PCA axis 2 (125)

Figure 6 Temporal evolution of sample scores on the rst (A) and second(B) PCA axes extracted from the seven derived variables Componentloadings (eigenvectors) for the individual parameters are shown as insets


low scores indicating cooling The rst implication of this simi-larity is that the WA temperature reconstruction from Fog Lakefaithfully tracks palaeoclimate (see also Joynt and Wolfe 2001)The second implication is that diatom productivity whetherinferred by numbers or volumes of diatom frustules and expressedeither per unit mass or as a ux is intimately connected to climateon decadal to millennial timescales Two primary factors are prob-ably responsible for the coupling of diatom productivity to cli-mate First summer temperature regulates the timing of initialmoating of the ice pan and hence the length of the ice-free seasonwhich in turn control both the duration of the growing season andthe availability of littoral habitats for benthic algal growthSecond summer air temperature modulates both terrestrial (soiland plant) productivity and decomposition rates so that attendant uxes of catchment-derived nutrients are greatest during episodesof relative summer warmth and vice versa

Another feature illustrated by upper curve on Figure 6 is theclear diatom response to post-LIA warming The amplitude ofthis signal in the rst PCA axis suggests that the magnitude ofenvironmental change it re ects is perhaps not exceptional in thecontext of the last ve millennia but importantly represents thelargest excursion of at least the last 2000 years This is consistentwith the interpretation of post-LIA warming as a naturallyinitiated but anthropogenically ampli ed climatic phenomenonthat is especially pervasive at high latitudes (Overpeck et al1997)

Further inferences from diatom biovolumesIn addition to providing a meaningful index of diatom palaeopro-ductivity the conversion of diatom concentrations to biovolumesallows additional conclusions to be drawn regarding the compo-sition of late-Holocene sediments in Fog Lake Given that eachextruded 05 cm core increment has a volume of 1916 cm3 andusing average values of 115 g cm23 for sediment wet densityand 80 for water content (Wolfe et al 2000) the biovolumecalculations indicate that diatom frustules comprise by volumebetween 90 (~4500 cal BP Figure 5) and 20 (LIA minimum)of the sediment The diatom biovolume corresponding to 50 ofthe total sediment volume is 23 cm3 gdw21 This implies in turnthat diatoms were the primary constituent of sediments depositedbefore 3000 cal BP (ie 50 by volume) but not thereafter(Figure 5) One possibility is that as Neoglacial cooling pro-ceeded diatoms were progressively replaced by more cold-adapted organisms as the dominant lake biota Climaticallyinduced changes in the representation of different algal groupshave indeed been documented in other Arctic regions (Korholaet al 2002) A second possibility is that diatoms became pro-gressively diluted by increased allochthonous inorganic sedimen-tation which is consistent with records of increased LIA summerstorm activity and erosion elsewhere in the Canadian Arctic(Lamoureux 2000) However there are no lithostratigraphic orgeochronological indications of increased minerogenic sedimen-tation in Fog Lake during the late Holocene (see also Wolfe et al2000) implying that the diatom biovolume trends represent anecological rather than sedimentological signal As summerscooled during the Neoglacial and LIA intervals it appears thatsmall non-glacial Arctic lakes became increasingly isolated fromlocal catchment in uences including nutrients supplied from ter-restrial processes

Late-Holocene diatom community structureAlthough diatom species richness and rate-of-changeare well cor-related (r = 055 p 001) these two variables are uncorrelatedeither to diatom-inferred water temperature or to the indices ofdiatom production discussed above The second axis of PCAextracted from the derived diatom data integrates primarily thein uences of changing species richness and rate-of-change As

can be seen in the time series of sample scores on this axis (Figure6) as well as in the original data (Figure 5) diatom communitieswere relatively stable and characterized by low richness before~2000 cal BP but thereafter destabilized with an attendantincrease of taxonomic richness This signi es that diatom com-munities were fairly resilient during the initial stages of late-Holocene cooling but as cold Neoglacial conditions persistedcommunity structure became disrupted If climatic deteriorationis viewed as a stressor of diatom communities which is a logicalconsequence of the arguments surrounding climatic control overhabitat availability and nutrient uxes then these results conformwell to the intermediate disturbance hypothesis of Connell (1978)In this model which has been successfully adapted and appliedto phytoplankton communities (Reynolds et al 1993 Hambrightand Zohary 2000) maximum species diversity is attained at anintermediate degree of environmental perturbation In a more spe-ci c example Interlandi and Kilham (2001) demonstrated thatdiversity is positively correlated to the number of resources limit-ing phytoplankton growth at a given time These studies aredirectly compatible with the palaeolimnological results from FogLake Simply stated as summers cooled during the Neoglacialhabitats became increasingly fragmented in both space and timewhile nutrient uxes from the catchment to the lake decreasedAccordingly increased competition due to diminished resourceavailability disfavoured dominance by a restricted number ofopportunistic taxa resulting in a diversi cation of the diatom ora

These results are in contrast to the observations of Andersonet al (1996) from northern Sweden where a signi cant reductionin taxonomic richness is registered during the LIA In thisinstance however it is plausible that human activities in thecatchment are compounded with climate deterioration impactingdiatom communities to a degree somewhat greater than that ofintermediate disturbance with the consequence of reducinginstead of augmenting species richness

Unlike the diatom-inferred water temperatures and the indicesof palaeoproductivity the measures of diversity and stability donot show pronounced responses to post-LIA warming This doesnot imply that community reorganizations lag behind environmen-tal change given that marked changes in diatom assemblage com-position are indeed registered (Figure 4) Rather the measures ofdiversity and stability applied here appear relatively insensitiveto these most recent changes because their magnitude remainssomewhat less than the species shifts that occurred earlier in therecord

Conclusion

This study demonstrates that diatom assemblages respond sensi-tively to late-Holocene climatic variability Although theseresponses are shown to be complex they are nonethelessdirectional and in accordance with predictions grounded in limno-logical and ecological theory The transfer-function approachwhich requires statistically explicit considerations of modern dia-tom-environment relationships is not the sole methodology forgleaning useful palaeoclimatic information from diatom strati-graphic data Moreover the congruenceof the WA summer water-temperature reconstruction with indices of diatom palaeopro-ductivity that utilize no a priori ecological characterization ofindividual taxa (with regards to temperature in this case) servesto validate the robustness of training sets designed speci cally forpalaeoclimatic applications From the strict perspective of algalecological physiology several of the points raised by Anderson(2000) such as the occurrence of blooms of certain taxa at timesother than that of the temperature measurements used for model-ling environmental optima cannot be denied and should not be


ignored However the potential problem of multiple environmen-tal variables in uencing WA temperature reconstructions can beaddressed by implementing alternative diatom-based indices thatare sensitive to climate yet completely unrelated to the trainingset in question It is suggested that the generation of such indicesconstitutes a useful strategy to explore limnological responses toclimatic change Furthermore in regions where training sets donot yet exist palaeoclimatic inferences can still be derived fromcreative manipulations of diatom stratigraphic data

Although the differentiation of direct from indirect climaticeffects upon diatom communities can be made conceptually(Smol 1988) the data from Fog Lake indicate that this distinctionmay in actual fact be quite subtle For example the apparentclimatic regulation of late-Holocene diatom productivity borneout by the frustule concentration and biovolume data simul-taneously integrates both the direct physiologicalresponses of dia-toms to changes in growing conditions with respect to tempera-ture light and nutrient concentrations as well as the indirectconsequences of changing lake-ice dynamics and terrestrial pro-cesses In this case because the diatom response is ecologicallyinterpretable and consistent with the independent record of palaeo-climate it becomes a moot point whether lsquodirectrsquo or lsquoindirectrsquoclimatic effects are registered by fossil assemblages or to whatextent these can truly be disentangled from each other

Finally the late-Holocene diatom assemblages from Fog Lakehighlight two intervals of especially rapid environmental changewithin the last ve millennia First species assemblages andinferred water temperatures shifted rapidly between 2500 and2000 cal BP indicating a major episode of accelerated summercooling It is in this interval that near-shore marine molluscassemblages of Arctic af nity migrated southward along the eastcoast of Baf n Island indicating cooling of the waters in westernBaf n Bay presumably in response to an increased delivery ofArctic Ocean water via the Queen Elizabeth Islands (Dyke et al1996) In all likelihood the changes observed at Fog Lake relateto this regional palaeoceanographic forcing given the sitersquoslocation on the outer coast These cold conditions appear to havebeen sustained until the end of the lsquoLittle Ice Agersquo which marks asecond major reorganizationof the Fog Lake diatom assemblagesRapid changes in both the primary (Figure 4) and derived (Figure6) data complement a growing body of evidence for unpre-cedented ecological changes during this interval in a wide varietyof lakes of the Canadian and Fennoscandian Arctic (Douglaset al 1994 Overpeck et al 1997 Sorvari and Korhola 1998Wolfe and Perren 2001 Sorvari et al 2002) These recentchanges can only be partially explained by a purely climaticmechanism since temperatures since the LIA have probably notexceeded maximum Holocene warmth It is therefore hypothes-ized that recent ecological changes are a synergistic response tothe coupled impacts of warming and some as of yet unspeci edatmospheric input of anthropogenic origin

Acknowledgements

This research was initially supported by the National ScienceFoundation (USA) through the PALE and ESH initiatives andsubsequently by the Natural Sciences and Engineering ResearchCouncil of Canada Konrad Hughen and Jay Moore generouslycollected the Fog Lake gravity core Ernest Joynt developed thetraining set and reconstructed water temperatures and GiffordMiller provided a continuous stream of ideas concerning thepalaeolimnology of Baf n Island lakes Both the NunavutResearch Institute (Nunavummi Qaujisaqtulirijikkut)and the ham-let of Qikiqtarjuaq provided logistical support for eld operationson Baf n Island Comments by HE Wright A Korhola and N JAnderson led to substantial improvements This paper is dedicated

to Gina Michl (1973ndash2000) who contributed tirelessly to the dia-tom biovolume measurements

References

Abbott MB and Stafford TW Jr 1996 Radiocarbon geochemistry ofmodern and ancient Arctic lake systems Baf n Island Canada Quatern-ary Research 45 300ndash11Anderson NJ 1994 Comparative planktonic diatom biomass responsesto lake and catchment disturbance Journal of Plankton Research 16133ndash50mdashmdash 2000 Diatoms temperature and climate change European Journalof Phycology 35 307ndash14Anderson NJ Odgaard BV Segerstrom U and Renberg I 1996Climate-lake interactions recorded in varved sediments from a Swedishboreal forest lake Global Change Biology 2 399ndash405Andrews JT Davis PT Mode WN Nichols H and Short SK1981 Relative departures in July temperatures in northern Canada for thepast 6000 yr Nature 289 164ndash67Battarbee RW 1973 Preliminary studies of Lough Neagh sedimentsII diatom analysis from the uppermost sediments In Birks HJB andWest RG editors Quaternary plant ecology New York John Wiley279ndash88Battarbee RW and Kneen MJ 1982 The use of electronically coun-ted microspheres in absolute diatom analysis Limnology and Oceanogra-phy 27 184ndash88Berger A and Loutre MF 1991 Insolation values for the climate ofthe last 10 million years Quaternary Science Reviews 10 291ndash310Birks HJB 1995 Quantitative paleoenvironmental reconstructions InMaddy D and Brew JS editors Statistical modelling of Quaternaryscience data Cambridge Quaternary Research Association TechnicalGuide 5 161ndash254Birks HJB and Line JM 1992 The use of rarefaction analysis forestimating palynological richness from Quaternary pollen-analytical dataThe Holocene 2 1ndash10Camburn KE and Charles DF 2000 Diatoms of low-alkalinity lakesin the northeastern United States Philadelphia Academy of NaturalSciences of Philadelphia Special Publication 18Connell JH 1978 Diversity in tropical rain forests and coral reefsScience 199 1302ndash10Dahl-Jensen D Mosegaard K Gundestrup N Johnsen SJHansen AW Clow GD and Balling N 1998 Past temperaturesdirectly from the Greenland Ice Sheet Science 282 268ndash71Davis PT 1985 Neoglacial moraines on Baf n Island In Andrews JTeditor Quaternary environments eastern Canadian Arctic Baf n Bay andwestern Greenland Boston Allen and Unwin 682ndash718Douglas MSV Smol JP and Blake W Jr 1994 Marked post-18thcentury environmental change in high arctic ecosystems Science 266416ndash19Dyke AS Dale JE and McNeely RN 1996 Marine molluscs asindicators of environmental change in glaciated North America and Green-land during the last 18000 years Geographie Physique et Quaternaire50 125ndash84Engstrom DR Fritz SC Almendinger JE and Juggins S 2000Chemical and biological trends during lake evolution in recently de-glaciated terrain Nature 408 161ndash66Foged N 1981 Diatoms in Alaska Vaduz CramerFritz SC Juggins S Battarbee RW and Engstrom DR 1991Reconstruction of past changes in salinity and climate using a diatom-based transfer function Nature 352 706ndash708Germain H 1981 Flore des diatomees eaux douces et saumatres duMassif Armoricain et des contrees voisines drsquoEurope occidentale ParisSociete Nouvelle des Editions BoubeeGlew JR 1989 A new trigger mechanism for sediment samplers Jour-nal of Paleolimnology 2 241ndash43Hambright KD and Zohary T 2000 Phytoplankton species diversitycontrol through competitive exclusion and physical disturbance Lim-nology and Oceanography 45 110ndash22Hillebrand H Durselen CD Kirschtel D Pollinger U and ZoharyT 1999 Biovolume calculations for pelagic and benthic microalgae Jour-nal of Phycology 35 403ndash24Hughen KA Overpeck JT and Anderson RF 2000 Recent warm-


ing in a 500-year paleotemperature record from varved sediments UpperSoper Lake Baf n Island Canada The Holocene 10 9ndash19Hustedt F 1959 Die Kieselalgen Deutschlands Osterreichs und derSchweiz Teil II Leipzig Academische VerlagsgesellschaftInterlandi SJ and Kilham SS 2001 Limiting resources and the regu-lation of diversity in phytoplankton communities Ecology 82 1270ndash82Jacobson GL and Grimm EC 1986 A numerical analysis of Holo-cene forest and prairie vegetation in central Minnesota Ecology 67958ndash66Joynt EH III and Wolfe AP 2001 Paleoenvironmental inferencemodels from sediment diatom assemblages in Baf n Island lakes(Nunavut Canada) and reconstruction of summer water temperatureCanadian Journal of Fisheries and Aquatic Sciences 58 1222ndash43Korhola A Sorvari S Rautio M Appleby PG Dearing JA HuY Rose N Lami A and Cameron N 2002 A multi-proxy analysisof climate impacts on the recent development of subarctic lake Saanajarviin Finnish Lappland Journal of Paleolimnology 28 59ndash77Korhola A Weckstrom J Holmstrom L and Erasto P 2000 Aquantitative Holocene climatic record from diatoms in northern Fenno-scandia Quaternary Research 54 284ndash94Kovach WL 1990 Multi Variate Statistical Package version 20 usersrsquomanual AberystwythKrammer K and Lange-Bertalot H 1986 Bacillariophyceae 1 TeilNaviculaceae In Ettl H Gerloff J Heynig H and Mollenhauer Deditors Susswasser ora von Mitteleuropa Band 21 Stuttgart GustavFischer Verlagmdashmdash 1988 Bacillariophyceae 2 Teil Bacillariaceae EpithemiaceaeSurirellaceae In Ettl H Gerloff J Heynig H and Mollenhauer Deditors Susswasser ora von Mitteleuropa Band 22 Jena GustavFischer Verlagmdashmdash 1991a Bacillariophyceae 3 Teil Centrales Fragilariaceae Eunotia-ceae In Ettl H Gerloff J Heynig H and Mollenhauer D editorsSusswasser ora von Mitteleuropa Band 23 Jena Gustav Fischer Verlagmdashmdash 1991b Bacillariophyceae 4 Teil Achnanthaceae KritischeErganzungen zu Navicula (Lineolatae) und Gomphonema Gesamtliterat-urverzeichnis Teil 1ndash4 In Ettl H Gartner G Gerloff J Heynig Hand Mollenhauer D editors Susswasser ora von Mitteleuropa Band 24Jena Gustav Fischer VerlagLamoureux S 2000 Five centuries of interannual sediment yield andrainfall-induced erosion in the Canadian High Arctic recorded in lacustrinevarves Water Resources Research 36 309ndash18Line JM ter Braak CJF and Birks HJB 1994 WACALIBversion 33 ndash a computer program to reconstruct environmental variablesfrom fossil assemblages by weighted averaging to derive sample-speci cerrors of prediction Journal of Paleolimnology 10 147ndash52Livingstone DM Lotter AF and Walker IR 1999 The decreasein summer surface water temperature with altitude in swiss alpine lakesa comparison with air temperature lapse rates Arctic Antarctic and AlpineResearch 31 341ndash52Lotter AF Birks HJB Hofmann W and Marchetto A 1997 Mod-ern diatom cladocera chironomid and chrysophyte cyst assemblages asquantitative indicators for the reconstruction of past environmental con-ditions in the Alps I Climate Journal of Paleolimnology 18 395ndash420Miller GH Mode WN Wolfe AP Sauer PE Bennike OForman SL Short SK and Stafford TW Jr 1999 Strati ed inter-glacial lacustrine sediments from Baf n Island Arctic Canada chronologyand paleoenvironmental implications Quaternary Science Reviews 18789ndash810Moore JJ Hughen KA Miller GH and Overpeck JT 2001Little Ice Age recorded in summer temperature reconstruction from varvedsediments of Donard Lake Baf n Island Canada Journal of Paleo-limnology 25 503ndash17Overpeck JT Hughen KA Hardy D Bradley RS Case RDouglas MSV Finney B Gajewski K Jacoby G Jennings AELamoureux S Lasca A MacDonald GM Moore J Retelle M

Wolfe AP and Zielinski G 1997 Arctic environmental change of thelast four centuries Science 278 1251ndash56Paterson WSB Koerner RM Fisher D Johnsen SJ ClausenHB Dansgaard W Bucher P and Oeschger H 1977 An oxygenisotope climatic record from the Devon Ice Cap arctic Canada Nature266 508ndash11Patrick R and Reimer CW 1966 The diatoms of the United Statesexclusive of Alaska and Hawaii Philadelphia Academy of NaturalSciences of Philadelphia monograph 13 volume 1mdashmdash 1975 The diatoms of the United States exclusive of Alaska andHawaii Philadelphia Academy of Natural Sciences of Philadelphia mono-graph 13 volume 2 part 1Pienitz R Smol JP and Birks HJB 1995 Assessment of freshwaterdiatoms as quantitative indicators of past climatic change in the Yukonand Northwest Territories Canada Journal of Paleolimnology 13 21ndash49Psenner R and Schmidt R 1992 Climate driven pH control of remotealpine lakes and effects of acid deposition Nature 356 781ndash83Reynolds CS 1984 The ecology of freshwater phytoplankton NewYork Cambridge University PressReynolds CS Padisak J and Sommer U 1993 Intermediate disturb-ance in the ecology of phytoplankton and the maintenance of speciesdiversity a synthesis Hydrobiologia 249 183ndash88Rosen P Hall R Korsman T and Renberg I 2000 Diatom transferfunctions for quantifying past air temperature pH and total organic carbonconcentrations from lakes in northern Sweden Journal of Paleolimnology24 109ndash23Short SK Mode WM and Davis PT 1985 The Holocene recordfrom Baf n Island modern and fossil pollen studies In Andrews JTeditor Quaternary environments eastern Canadian Arctic Baf n Bay andwestern Greenland Boston Allen and Unwin 608ndash42Smol JP 1983 Paleophycology of a high arctic lake near Cape HerschelEllesmere Island Canadian Journal of Botany 61 2195ndash204mdashmdash 1988 Paleoclimate proxy data from freshwater arctic diatoms Ver-handlungen der Internationalen Vereinigung von Limnologen 23 837ndash44Sorvari S and Korhola A 1998 Recent diatom assemblage changes insubarctic Lake Saanajarvi NW Finnish Lapland and their paleoenviron-mental implications Journal of Paleolimnology 20 205ndash15Sorvari S Korhola A and Thompson R 2002 Lake diatom responsesto recent Arctic warming in Finnish Lapland Global Change Biology 8153ndash64Steig EJ Wolfe AP and Miller GH 1998 Wisconsinan glacialrefugia on eastern Baf n Island coupled evidence from cosmogenicisotopes and lake sediments Geology 26 835ndash38Stuiver M Reimer PJ Bard E Beck JW Burr GS HughenKA Kromer B McCormac FG van der Plicht J and Spurk M1998 INTCAL98 radiocarbon age calibration 24000ndash0 cal BP Radiocar-bon 40 1041ndash83ter Braak CJF 1987 Ordination In Jongman RHG ter Braak CJFand Van Tongeren OFR editors Data analysis in community and land-scape ecology Wageningen Center for Agricultural Publishing and Docu-mentation 91ndash173Weckstrom J Korhola A and Blom T 1997 Diatoms as quantitativeindicators of pH and water temperature in subarctic Fennoscandian lakesHydrobiologia 347 171ndash84Wolfe AP 1997 On diatom concentrations in lake sediments results ofan inter-laboratory comparison and other experiments performed on a uni-form sample Journal of Paleolimnology 18 61ndash66mdashmdash 2002 Climate modulates the acidity of arctic lake on millennialtimescales Geology 30 215ndash18Wolfe AP and Perren BB 2001 Chrysophyte microfossils recordmarked responses to recent environmental changes in high- and mid-arcticlakes Canadian Journal of Botany 79 747ndash52Wolfe AP Frechette B Richard PJH Miller GH and FormanSL 2000 Paleoecological assessment of a 90000-year lacustrinesequence from Fog Lake Baf n Island Arctic Canada QuaternaryScience Reviews 19 1677ndash99


pogenic in uences Small upland Arctic lakes offer additionaladvantages for studies aiming to characterize the climatic controlover diatom assemblages These include reduced trophic com-plexity in the absence of sh and the lack of groundwater in u-ences due to continuous permafrost Although non-climatic factorsassociated with ecosystem ontogeny are known to in uence waterquality and hence diatom assemblages (Engstrom et al 2000) thepresent study addresses a lake having simpli ed soil and veg-etation dynamics in the presence of sparse tundra Furthermorethe longer palaeoenvironmental record suggests that edaphic fac-tors were largely stabilized by the middle Holocene (Wolfe et al2000) in turn maximizing the sensitivity of late-Holocene diatomassemblages to climatic in uences

Study site

Fog Lake (67deg119 N 63deg159 W) is situated at 460 m asl onthe north coast of Cumberland Peninsula Baf n Island NunavutCanada (Figure 1) The lake has a surface area of 0015 km2 anda maximum depth of 95 m The Fog Lake watershed is small(23 km2) occupying an upland that separates two major ordsystems (Padle and Kangert) directly above the highest morainesdated to the last glacial cycle (Steig et al 1998) As a result thelake preserves an unusually long sediment record that includes atleast the latter portion of oxygen isotope stage 5 (Wolfe et al2000) Lakewaters are extremely dilute with electrical conduc-tivity of 10 mS cm21 and a summed concentration of base cations(Na+ + K+ + Mg2++ Ca2+) of 30 mg L21 Summer epilimneticpH is 63 In most years the lake is ice-free from mid-July to mid-September However cold summers may result in the persistenceof a vestigial ice pan throughout the summer

The climate of northern Cumberland Peninsula is severeespecially along the outer coast At Qikiqtarjuaq (formerlyBroughton Island) 40 km northwest of Fog Lake mean annualtemperature is ndash113degC (summer (JJA) mean 12degC winter(DJF) ndash176degC) Annual precipitation at Qikiqtarjuaq averages290 mm a21 of which 85 falls as snow but precipitationincreases dramatically eastward along the coast at Cape Dyer100 km east of Fog Lake mean annual precipitation rises to 660

85 W 80 W 75 W 70 W 65 W

65 N

70 N

BaffinBayBaffin

IslandFoxeBasin

HudsonStrait

400 km

Ungava

Canada

USA

Greenland

1000 km

FogLake

Figure 1 Location map of Fog Lake on the north coast of CumberlandPeninsula Baf n Island Circles indicate the locations of the 61 modernsurface-sediment samples used to develop the summer water-temperatureinference model of Joynt and Wolfe (2001)

mm a21 Local vegetation in the Fog Lake catchment is limitedto a discontinuousheath tundra dominated by Cassiope tetragonaand Empetrum nigrum

Late-Holocene climatic evolution of theBaf n Bay region

The late Holocene in the Baf n Bay region can be characterizedas an interval of progressive summer cooling as inferred bymultiple lines of isotopic palynological and glacial-geologicalevidence (Figure 2) Part of this trend appears to be ultimatelyregulated by decreasing high-latitude summer insolation TermedNeoglacial for associated advances of many local glacial systemslate-Holocene cooling culminated in the lsquoLittle Ice Agersquo (LIA)ad ~1450ndash1850 Since the end of the LIA summer temperatureshave increased rapidly across the Arctic (Overpeck et al 1997)a trend that has probably been ampli ed in recent decades byanthropogenic greenhouse-gas forcing Baf n Island appears toconform closely to this general pattern of climatic evolution(Davis 1985 Hughen et al 2000 Moore et al 2001) Althoughmost of the regionrsquos late-Holocene palaeoclimatic records fromthis region are expressed as relative departures or anomalies theactual amplitude of late-Holocene cooling is probably about 3degC(eg Andrews et al 1981 Short et al 1985 Dahl-Jensenet al 1998)

Methods

Coring and sediment chronologyThe core considered in this study was obtained from the deepestpart of Fog Lake in August 1997 with a modi ed Kajak-Brinkhurst gravity corer (Glew 1989) The 345 cm core wasextruded vertically in the eld in consecutive 05 cm incrementsto eliminate any potential disturbance The corersquos chronology isbased on ve accelerator mass spectrometry (AMS) 14C datesfrom sediment humic acids extracted according to Abbott andStafford (1996) The humic acid 14C ages were corrected byadding 300 years to account for the average lag in the transportof terrestrial dissolved organic carbon to lake sediments (Milleret al 1999) and then calibrated to calendar years before present(Stuiver et al 1998) The resulting age model (Figure 3) is ttedwith a polynomial equation (r2 = 099) that enables the reliableconversion of sediment depth to age and results in an age of 4800cal BP for the base of the core The dating results in addition tothe corersquos uniform visual stratigraphy (olive silty gyttja) implythat the sedimentation rate has varied little over the interval ofdeposition and that it has not been punctuated by low-frequencyhigh-intensity events This is an important consideration withregard to the interpretation of diatom concentration and ux data(Wolfe 1997)

Diatom primary dataDiatom slurries were prepared from each of the 69 samples byoxidation of organic matter in hot 30 H2O2 Slurries were spikedwith a known concentration of marker grains (Eucalyptus pollen)in order to estimate diatom concentrations and then diluted Ali-quots of 200 mL of the diluted slurries were settled ontocoverslips Once air-dried these were mounted on microscopeslides with NaphraxOgrave Diatoms (and markers) were counted fromeach interval in random transects including coverslip edges Dia-tom taxonomy followed primarily Hustedt (1959) Patrick andReimer (1966 1975) Foged (1981) Germain (1981) Krammerand Lange-Bertalot (1986 1988 1991a 1991b) and Camburnand Charles (2000) Diatom sums ranged between 485 and 562





-1 0 +1



-30 -275 -25

0

1000

2000

3000

4000

5000

6000

360 380 400



Yea

rs B

P

DB C

-1 0 +1


2000

1900

1800

1700

1600

1500

Years A

D

2 3 4 5 6 7


A


E


r2=099

0 2000 4000 6000

0

5

10

15

20

25

30

35

40

age = 071z2 + 11559z + 579

Dep

th (

cm)

















Results




80

40

0

20

80

40

0

5000 4000 3000 2000 1000 0

Years (cal BP)

4

8

0

2

4

0

10

20

0

1

2

0

4

2

0

6

0

5

8

4

0


Aulacoseira distans




4

2

0

10

5

0

10

0

20

10

0

20

10

0

Nitzschia perminuta


Eunotia rhomboidea

Peronia fibula

Pinnularia biceps

Eunotia bilunaris

Eunotia vanheurkii

5000 4000 3000 2000 1000 0

Years (cal BP)

rela

tive

freq

uenc

ies

(sca

les

vary

)






0

1

2

3

4

5Diatom biovolume

(cm3 gdw-1)

0

1

2

3

4

5

0

01

02Rate of change


0

20

10


010002000300040005000

Years (cal BP)


4

6

8

5

7

3

9

0

10

20 Diatom fluxes



(cm3 cm-2 a-1)

10

15

20

25

30








Discussion


-05

0

05

10

Years (cal BP)

010002000300040005000

0

Warming

Neoglacial cooling

Little Ice Age

-04 0 04


0 04 06


volumetric flux



05

A PCA axis 1 (671)

B PCA axis 2 (125)










Conclusion






Acknowledgements



References









-1 0 +1



-30 -275 -25

0

1000

2000

3000

4000

5000

6000

360 380 400



Yea

rs B

P

DB C

-1 0 +1


2000

1900

1800

1700

1600

1500

Years A

D

2 3 4 5 6 7


A


E


r2=099

0 2000 4000 6000

0

5

10

15

20

25

30

35

40

age = 071z2 + 11559z + 579

Dep

th (

cm)

















Results




80

40

0

20

80

40

0

5000 4000 3000 2000 1000 0

Years (cal BP)

4

8

0

2

4

0

10

20

0

1

2

0

4

2

0

6

0

5

8

4

0


Aulacoseira distans




4

2

0

10

5

0

10

0

20

10

0

20

10

0

Nitzschia perminuta


Eunotia rhomboidea

Peronia fibula

Pinnularia biceps

Eunotia bilunaris

Eunotia vanheurkii

5000 4000 3000 2000 1000 0

Years (cal BP)

rela

tive

freq

uenc

ies

(sca

les

vary

)






0

1

2

3

4

5Diatom biovolume

(cm3 gdw-1)

0

1

2

3

4

5

0

01

02Rate of change


0

20

10


010002000300040005000

Years (cal BP)


4

6

8

5

7

3

9

0

10

20 Diatom fluxes



(cm3 cm-2 a-1)

10

15

20

25

30








Discussion


-05

0

05

10

Years (cal BP)

010002000300040005000

0

Warming

Neoglacial cooling

Little Ice Age

-04 0 04


0 04 06


volumetric flux



05

A PCA axis 1 (671)

B PCA axis 2 (125)










Conclusion






Acknowledgements



References














Results




80

40

0

20

80

40

0

5000 4000 3000 2000 1000 0

Years (cal BP)

4

8

0

2

4

0

10

20

0

1

2

0

4

2

0

6

0

5

8

4

0


Aulacoseira distans




4

2

0

10

5

0

10

0

20

10

0

20

10

0

Nitzschia perminuta


Eunotia rhomboidea

Peronia fibula

Pinnularia biceps

Eunotia bilunaris

Eunotia vanheurkii

5000 4000 3000 2000 1000 0

Years (cal BP)

rela

tive

freq

uenc

ies

(sca

les

vary

)






0

1

2

3

4

5Diatom biovolume

(cm3 gdw-1)

0

1

2

3

4

5

0

01

02Rate of change


0

20

10


010002000300040005000

Years (cal BP)


4

6

8

5

7

3

9

0

10

20 Diatom fluxes



(cm3 cm-2 a-1)

10

15

20

25

30








Discussion


-05

0

05

10

Years (cal BP)

010002000300040005000

0

Warming

Neoglacial cooling

Little Ice Age

-04 0 04


0 04 06


volumetric flux



05

A PCA axis 1 (671)

B PCA axis 2 (125)










Conclusion






Acknowledgements



References






0

1

2

3

4

5Diatom biovolume

(cm3 gdw-1)

0

1

2

3

4

5

0

01

02Rate of change


0

20

10


010002000300040005000

Years (cal BP)


4

6

8

5

7

3

9

0

10

20 Diatom fluxes



(cm3 cm-2 a-1)

10

15

20

25

30








Discussion


-05

0

05

10

Years (cal BP)

010002000300040005000

0

Warming

Neoglacial cooling

Little Ice Age

-04 0 04


0 04 06


volumetric flux



05

A PCA axis 1 (671)

B PCA axis 2 (125)










Conclusion






Acknowledgements



References













Conclusion






Acknowledgements



References









Acknowledgements



References





Wolfe Holocene 2003

Documents

Transcript of Wolfe Holocene 2003