Stratified interglacial lacustrine sediments from Baffin Island, Arctic Canada: chronology and paleoenvironmental implications

of 22

Please download to get full document.

View again

All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
22 pages
0 downs
Stratified interglacial lacustrine sediments from Baffin Island, Arctic Canada: chronology and paleoenvironmental implications
  * Corresponding author. Fax: 001 303 492 6388; e-mail: gmiller @  Current address: Department of Marine Chemistry and Geochem-istry, WoodsHole OceanographicInstitution,Woods Hole, MA 02543,USAQuaternary Science Reviews 18 (1999) 789 —  810 Stratified interglacial lacustrine sediments fromBaffin Island, Arctic Canada:chronology and paleoenvironmental implications Gifford H. Miller     * , William N. Mode  , Alexander P. Wolfe  , Peter E. Sauer     ,Ole Bennike  , Steven L. Forman  , Susan K. Short  , Thomas W. Stafford  , Jr.   Institute of Arctic and Alpine Research, Uni v ersity of Colorado, Boulder, CO 80309-0450, USA   Department of Geological Sciences, Uni v ersity of Colorado, Boulder, CO 80309-0399, USA   Department of Geology, Uni v ersity of Wisconsin Oshkosh, Oshkosh, WI 54901, USA  Geological Sur  v ey of Denmark and Greenland, Thora v ej 8, Copenhagen, DK-2400, Denmark    Department of Earth and En v ironmental Sciences, Uni v ersity of Illinois, Chicago, IL 60607, USA Abstract Thirteen of 18 piston cores recovered from ‘Robinson Lake’ in the mid-Arctic vegetation zone of Baffin Island, Arctic Canada,penetrated stratified lacustrine sediment beneath a thin over-consolidated diamict (till) and postglacial lacustrine sediment. Thesub-till lacustrine units are up to 120 cm thick, of which the upper several decimeters frequently contain thick, layered mats of aquaticmoss; pollen and diatoms are common throughout both lacustrine units. A series of 23 AMS  C dates defines the chronology of thepostglacial sequence, which records a succession from a pioneer grass- and Oxyria -dominated tundra between 10.4 and 8 ka BP, toa sedge-dominated tundra after 8 ka BP. Limiting  C dates place the sub-till lacustrine sediments more than 40 ka BP; 10luminescence dates centered on 85 ka indicate they were deposited late in oxygen isotope stage (OIS) 5. The dominance of shrub andtree pollen, especially shrub birch and alder, in sub-till lacustrine samples, indicates low-arctic tundra farther north than at any timeduring the Holocene. Pollen concentrationsare comparable to or higher than in the Holocene units. Cooling late in the interglacial isindicated by declining birch and alder pollen percentages in the upper part of the section. Diatom floras in both the sub-till andpostglaciallacustrine sequencesindicate similar developmentof lake-water chemistry,but input of silica and weathering products wasgreater in the older lake cycle, suggesting more vigorous catchment processes. Macrofossils in the sub-till units are characteristic of lakes ice-free in summer. Based on pollen assemblages indicating local and regional vegetation diagnostic of summer temperatureshigher than the Holocene, we interpret the sub-till lacustrine units to be of interglacial character. By analogy with Holocene plantsuccession in central and eastern Canada, all of Keewatin and Labrador/Ungava must have been ice free throughout this interval,suggesting essentially complete deglaciation of the Laurentide Ice Sheet at the time.  1999 Elsevier Science Ltd. All rights reserved. 1. Introduction Although ice-free refugia have been recognized alongthe forelands of eastern Baffin Island since the 1960s(L + ken, 1966; Miller et al ., 1977; Klassen, 1993), stratifiedterrestrial interglacial deposits have not been previouslyreported from the Eastern Canadian Arctic. Last inter-glacial sediments are known from deep-sea cores in theLabrador Sea (de Vernal et al ., 1991) and Baffin Bay(Aksu, 1985). Emerged marine deposits of probable inter-glacialage (OIS 5 sensulato [ s . l .]) have beenrecognizedatseveral sites along eastern Baffin Island and northwestGreenland (Miller et al ., 1992). A variety of lake sedi-ments and buried soils at or beyond the limits of radiocarbon dating have been recovered from Baffin Is-land (Terasmae et al ., 1966; Miller et al ., 1977; Mode,1985; Short et al ., 1985; Morgan et al ., 1993; Wolfe andHa¨rtling, 1996). Although some of these deposits areundeniably of interglacial character, bearing pollen as-semblages diagnostic of summer temperatures above 0277-3791/99/$-see front matter  1999 Elsevier Science Ltd. All rights reserved.PII: S0 2 77 -3 7 9 1 (9 8 ) 0 0 0 75 - 4  present, none can be unequivocally assigned to the lastinterglacial. The lack of a clear stratigraphic successionand chronologic uncertainties limit the interpretation of these sites.Here we report the first demonstrably stratified terres-trial sediments of interglacial character from Baffin Is-land. Eighteen piston cores were recovered from Robin-son Lake; 13 of them penetrated a stiff diamicton that weinterpret to be till, and sampled underlying stratifiedlacustrine sediments up to 120 cm thick, before bottom-ing in a second, less consolidated diamict. We refer to thesub-till lacustrine sediments as interglacial in character,independent of any dating, because they contain pollenassemblages representative of summers consistentlywarmer than the Holocene, and pollen concentrationscomparable, or in excess, of those in overlying Holocenesediment. Abundant lacustrine fossils, including diatomvalves, aquatic bryophytes, and aquatic zooplankton,collectively indicate that the sub-till units were depositedin a permanent lake, with ice-free summers. 2. Study area Robinson Lake (63 ° 23.8  N; 64 ° 15.7  W; unofficial name,after the adjacent Robinson Channel) is a small (17 ha) lakesituated 170 m asl at the divide of an east-west valley thattraverses Brevoort Island, in the Eastern Canadian Arctic(Fig. 1). As a drainage-divide lake there is no primary inletstream, and the 3.0 km  catchment drains into the lake viamany small rivulets and as sheet flow in spring. The lake ispartially dammed by morainal debris that has beenbreached by an outlet stream flowing west into RobinsonChannel. The lake has a simple primary basin that reachesnearly 15 m water depth (Fig. 1c), and a broad regionshallower than 4 m water depth between the primarybasin and the moraine dam. We infer that the lake wasapproximately 7 m shallower prior to emplacement of themoraine dam; at that water level the lake area would havebeen 5 ha, with a maximum water depth of about 9 m.Water samples collected from Robinson Lake in May1993 had a pH of 6.7 and conductivity of 22  Scm \ .The lake is therefore dilute and considered acid-sensitive.Recent field mapping (Miller, unpub. observations,1990), indicates Robinson Lake was situated at the mar-gin of the last glaciation, and was overrun by a thindivergent arm of an outlet glacier flowing SSW throughRobinson Channel; regional deglaciation occurred about10 ka BP (Miller and Kaufman, 1990; Stravers et al .,1992). The bedrock of Brevoort Island, the adjacentchannels, and Hall Peninsula to the west is entirely com-posedof crystallinerocks of the Canadian Shield, provid-ing a mildly acidic terrain without carbonate or othersources of dead carbon.The modern climate of Brevoort Island is maritimearctic; the mean July temperature is # 4.2 ° C, and meanJanuary temperature is ! 22.3 ° C, as recorded at theBAF-3 North Warning Site located 341 m asl, 10 kmsouth of the lake. Mean annual precipitation is 58 cm,43% of which occurs as snow.The vegetation on Brevoort Island is classified as mid-arctic, although it lies close to the low-arctic boundary(Fig.1a). Low-arcticvegetation is characterizedby nearlycomplete plant cover with significant proportions of shrubby plants, predominantly heaths and heath allies( » accinium , Cassiope , ¸ edum , Arctostaphylos , and Empet - rum ), shrub birch ( Betula glandulosa and B . nana ), andwillow ( Salix ). In mid-Arctic regions, plant cover is in-complete, with significantly fewer shrubs and greaterproportions of herbaceous plants, especially sedges (Cy-peraceae) and grasses (Poaceae; Edlund, 1986). Unlikemany low-arctic areas, Alnus (alder) does not occur atpresent on Baffin Island, nor has it been reported inthe Holocene, although its pollen is an important exoticcomponent of pollen spectra (Mode and Jacobs, 1987).Alder macrofossils occur in the Flitaway beds of north-central Baffin Island that were initially assigned to thelast interglacial (Terasmae et al ., 1966), but are nowconsidered to be of Tertiary age (Morgan et al ., 1993).The nearest occurrence of  Alnus today is in northernLabrador, about 500 km south of Brevoort Island. Shrubbirch is restricted to the low arctic and rare isolatedpocketsoflow-arcticvegetationwithinthemid-arcticzonewhere favorable microclimates sustain a richer vegetationassemblage (Jacobs et al ., 1985). Isolated pockets of shrubbirch currently occur 50 km west of Robinson Lake, but itdoes not grow now on Brevoort Island. It is difficult todistinguish reliably the pollen of tree birch from that of shrub birch because of considerable overlap in theirsize-frequency distributions (Ives, 1977), althoughRichard (1981) has shown that small but consistentchanges in the size-frequency of  Betula pollen occur inlake sediment cores from Ungava that may be inter-preted in terms of changing distributions of treeand shrub birch. The closest tree birch currently live innorthern Labrador. 3. Methods 3.1. Coring and physical properties RobinsonLake was coredfrom the lake ice platform inMay 1991 and again in May 1993, utilizing a modifiedNesje percussion-driven piston corer (Nesje, 1992)capable of recovering up to 6 m of sediment in a singledrive in water depths up to 100 m. Because the drivingforce is delivered to the core head at the bottom of thelake, it is possible to penetrate exceptionally stiff sedi-ments without casing. In 1991, four cores of 63 mmdiameter were recovered, two of which penetrated a stiff diamict and sampled underlying lacustrine sediments. In 790 G.H. Miller et al. / Quaternary Science Re v iews 18 (1999) 789 —  810  Fig. 1. (a) Vegetation zones of the Canadian Arctic (Edlund, 1984), and location of Brevoort Island (BI); solid square defines boundaries of Fig. 1b.(b) Location of Robinson Lake, Brevoort Island, with its drainage basin outlined (dashed line); contour intervals are in meters. (c) Outline andbathymetry of Robinson Lake. Shoreline and 2 m isobath from photogrammetry; bathymetry based on GPS-located transects of 37 soundingsthrough the lake ice platform (solid dots). General locations of the primary six cores discussedin the text ae indicated. The interglacial lake shoreline issuggested to fall between the 6 and 8 m isobaths. (d) Cross section of Robinson Lake along profile A/B (Fig. 1c), showing the position of the primarycores and the suggested water level during the Last Interglacial (dashed line). 1993, an additional 14 cores of 75 and 110 mm diameterwererecovered.Elevenof the 1993 cores penetrateda stiff diamict and sampled underlying water-lain sediments(Table 1). The upper diamict is readily distinguishablefrom water-lain units by its high whole-core magneticsusceptibility (MS), which was measured in the field at5 cm intervals using a portable Bartington MS Meter.After measuring MS, some cores were extruded in thefield to calibrate the MS signal with visual description;the remainder were sealed in their srcinal PVC core G.H. Miller et al. / Quaternary Science Re v iews 18 (1999) 789 —  810 791  Table 1Details of sediment cores recovered from Robinson Lake in 1991 and 1993Core ID Water Core Length of Thickness of Units Fate  Analyses  depth diameter recovered sub-till units present(m) (mm) core (m) (m)91-RL1 10.9 63 0.84 0 D, G, H E91-RL2 10.8 63 1.12 0 D, G, H S91-RL3 10.0 63 1.55 0.4 A, B1, C, D, E, F, G L  C91-RL4 9.9 63 1.76 0.3 B1, C, D, E, F, G L P, D,  C93-RL1 14.3 75 2.20 1.1 B, C, E, F, G E93-RL2 14.7 75 1.75 0.4 B, D, E, F, G E93-RL3 14.7 75 1.63 0.3 B1, B2, C, D, F, G L P (stratified B2)93-RL4 14.8 75 2.47 1.2 A, B2, C, E, F, G L D, M93-RL5 14.0 75 2.02 0.3 C, D, E, F, G L TL, M93-RL6 9.6 75 1.20 0.4 C, E, F, G E I93-RL7 11.0 75 1.6 ' 0.4 B1, F, G E93-RL8 9.2 110 0.8  0 E, F, G E93-RL9 10.4 110 1.65  0.6 A, B2, D, E, F L93-RL10 14.4 110 1.80 0.7 A, B2, C, D, E, G E P, D, Saved 35 cmof Unit C93-RL11 14.6 110 1.43 0.5 S93-RL12 7.8 75 0.9 0.6 B2, D, E, G E93-RL13 7.6 75 0.8 S93-RL14 7.8 110 0.8 0 D, E, F, G E  E " extruded and described in the field; not saved, L " opened and described in the lab, S " saved, but not opened).  P " pollen, D " diatoms, M " macrofossils,  C " radiocarbon dates, S " sedimentology, TL " luminescence dating, I " C-isotopes. tubes and opened in our laboratories. Before splitting thecores, X-ray images were made and volume MS wasmeasured at 2 cm intervals. The cores were then split,photographed, and described. One-half of the split corewas sampled (sedimentology, palynology, diatoms, mac-rofossils and radiocarbon dating), and a repository half was sealed; both halves were stored at 4 ° C. Core 93-RL5could be correlated with our primary core (91-RL4)based on their MS stratigraphies, and was preservedunopened for luminescence analyses.Constant-volume sediment samples were analyzed forbulk density and grain-size distribution using a Sedi-graph, and for total organic matter following the Walk-ley —  Black method (Walkley and Black, 1934). Mass MSwas measured at 5 cm intervals in the laboratory forsome cores.To capture undisturbed sediments at the sediment —  water interface we used a 10 cm square Eckman Dredgewith clear polystyrene liner that typically penetrates theupper 10 —  20 cm of sediment and preserves the sediment-water interface. These cores were sampled at 0.5 —  2.0 cmintervals in the field. 93-RLB4, a 10-cm-long box corecollected from 14.7 m water depth, was combined withpiston core 91-RL4 to build a composite Holocene sec-tion for Robinson Lake. 3.2. Palynology Laboratorypreparationof pollen sampleswas by stan-dard methods (Faegri and Iversen, 1989), augmented byfine sieving (Cwynar et al ., 1979). Pollen sums are ' 300grains (excluding aquatics), except where noted. Pollenwas analyzed from the postglacial sediments (UnitsF and G) at 10 cm intervals (ca. 1 sampleka \ ) and at1 cm intervals in the upper part (Unit C) and 2 cm inter-vals in the lower part (Unit B1) of the sub-till lacustrineunits of core 91-RL4. Because the pollen diagram frommassive Unit B1 lacks structure, we analyzed pollen at2 cm intervals throughout a stratified, deeper water facies(Unit B2 from core 93-RL3). To replicate the pollen zonetransition seen in 91-RL4, we analyzed pollen at 2 cmintervals from a similar lithostratigraphic sequence (UnitC over Unit B2) in core 93-RL10. 3.3. Diatoms Volumetric subsamples (1.0 cm  ) were taken from eachinterval and prepared by oxidation in hot 30% H  O  .Diluted aliquots of cleaned slurries were evaporated atroom temperature onto coverslips, and mounted to slideswith Naphrax medium. From each level, between 300and 500 diatom valves were enumerated in transectsincluding coverslip edges. All counting was performed at1000X under oil immersion, using an Olympus Vanoxmicroscope equipped with differential interference con-trast optics. Raw count data were compiled into relativefrequencies to illustrate stratigraphic changes in diatomcommunities. Diatoms were counted from 87 intervals inthe same three cores analyzed palynologically: 91-RL4(46 Holocene samples and 17 from units B1 and C), 792 G.H. Miller et al. / Quaternary Science Re v iews 18 (1999) 789 —  810  93-RL10 (11 samples from units B2 and C), and 93-RL3(13 samples from units B1 and B2). Diatom taxonomyfollowed primarily the floras of Hustedt (1959), Foged(1981), Germain (1981), Patrick and Reimer (1966, 1975),and Krammer and Lange-Bertalot (1986-1991). Nomen-clature and taxonomic authorities follow Hamilton et al .(1994). Several genera ( Fragilaria , Navicula ) have not beensplit according to the revisions of Round et al . (1990),largely to facilitate comparisons with previous studies. 3.4. Macrofossils To test whether some plant taxa that show high pollenpercentage values in the interglacial sediments grew lo-cally, sediment from two cores was analyzed for macro-fossils. The sediment was washed on 0.4 and 0.12 mmsieves. Aquatic and terrestrial macrofossils in the residueleft on the sieves were identified using a dissecting micro-scope. Macrofossils were isolated from eight nearly equalhalf-core segments from a 120-cm-thick sequence of UnitB2 from 93-RL4, and four half-core segments from a 20-cm-thick section of Unit C in 93-RL5. 3.5. Radiocarbon dating  Our objective was to provide a secure time constrainton lacustrine sediments less than 35 ka old, and todetermine whether the sub-till lacustrine units were with-in, or beyond the limits of the method. All radiocarbondating was by accelerator mass spectrometry, due to theneed to date small core increments (1 —  2 cm slices) tocompensate for slow sediment accumulation rates (ca.10 cmka \ ). Macrofossils received a standard acid —  basepretreatment; humic acids were separated by their solu-bility in acid and base extractions following proceduresdescribed by Abbott and Stafford (1996).   C in aquaticmoss from Robinson Lake varies from ! 17 to ! 23  ,with a mean of  ! 21  ( n " 17). One sample of bulk-sediment cellulose from a sub-till lacustrine unit wasdated following the extraction procedures outlined inSauer (1997). 3.6. Luminescence dating  Luminescence dating of water-lain sediments can beproblematic because light-filtering by water and turbid-ityreducetheefficiencyof solar resettingof luminescence.Thermoluminescence (TL) dating, which assumes full so-lar resetting, may yield maximum ages. Infrared stimu-lated luminescence (IRSL; Spooner et al ., 1990) samplestraps that are more sensitive to solar resetting (Godfrey-Smith et al ., 1988), resulting in a more accurate geo-chronometer for waterlain sediments (e.g. Forman et al .,1994; Kaufman et al ., 1996). Red excitation (RSL) mayreflect emissions from even more light-sensitive trapsthan IRSL, particularly for waterlain sediments (Ditlef-sen and Huntley, 1994; Forman et al ., 1994). To evaluatethe potential of luminescence methods applied to lakesediments, we selected a core from the central deep (93-RL5) that contained representative sequences of both theHolocene and interglacial lacustrine sediment (based onMS logs). The core was opened in the luminescencelaboratory where it was described physically, confirmingthe MS-based correlations. The three luminescence tech-niques were applied to sediment from 80 to 85, 90 to 95,and 95 to 100 cm depth in the Holocene sediments to testwhether adequate re-zeroing of the traps was occurringduring transport, and from 190 to 193 and 193 to 196 cmin comparable sub-till lacustrine facies (Unit B2) to datethe interglacial. 3.6.1. TL methods TL emissions were measured on the fine-grained(4 —  11  m) polymineral fraction for two representativesamples each from the Holocene and interglacial sec-tions. Equivalent dose (ED) was determined by the totalbleach method, which results in a near total resetting of TL. All samples were preheated at 124 ° C for 2 days priorto analysis to remove potential instability in the laborat-ory-induced TL signal. After heating, each sample wastested for anomalous fading by storing irradiated sam-ples (100 or 450 Gy) for at least 31 days and compared tothe TL emissions of an aliquot that was analyzed ( 1 hafter preheating. The anomalous fading tests reveal nosignificant instability in the TL emissions of the pre-heated aliquots (Table 2).The rate of TL ingrowth was evaluated by applyingadditive beta doses to the natural signal and fitting thedata with a saturating exponential function. Dose rateestimates were derived from U and Th concentrationsinferred from thick-source alpha counting, and  K con-tent, calculated from the total K concentration (Table 3),and were adjusted for organic content (Divigalpitiya,1982). Holocene and sub-till lacustrine sediments contain80 $ 10 and 60 $ 10% water, respectively. 3.6.2. IRSL method  IRSL stimulation of the fine-grain fraction was derivedfrom infrared emissions (880 $ 80 nm), with an estimatedenergy delivery at the sample position of 17 mWcm \ .IRSL ages were calculated from the same additive betadose, normalization and ED-computational proceduresused in TL analyses, except a shorter and highertemperature preheat (160 ° for 5 h) was employed andmeasurement of IRSL was delayed at least 1 day afterpreheating. Tests for anomalous fading of the laborat-ory-inducedand pre-heatedIRSL signal revealed insigni-ficant reduction in IRSL (Fig. 2). 3.6.3. RSL methods Red excitation (645 $ 10 nm) of the sample was ac-complished by variable narrow bandpass system; the G.H. Miller et al. / Quaternary Science Re v iews 18 (1999) 789 —  810 793
Related Search
Similar documents
View more...
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks

We need your sign to support Project to invent "SMART AND CONTROLLABLE REFLECTIVE BALLOONS" to cover the Sun and Save Our Earth.

More details...

Sign Now!

We are very appreciated for your Prompt Action!