Highresolution Stable Isotope Records from Southwest Sweden the Drainage of the Baltic Ice Lake and Younger Dryas Ice Margin

PALEOCEANOGRAPHY, VOL. 12, NO. 1, PAGES 39-49, FEBRUARY 1997

High-resolution stable isotope records from southwest Sweden: The drainage of the Baltic Ice Lake and Younger Dryas ice margin oscillations PerBodOn, Richard G. Fairbanks,2James D. Wright,3and Lloyd H. Burckle2

Abstract. Benthic foraminifera in two shallow marine sediment cores from southwest Sweden wereanalyzed for oxygenisotopes. Several deglaciation events, previously recognized in terrestrial andlake sedimentsthroughout the Balticregion, are identified and radiocarbon dated. An initial Baltic Ice Lake (BIL) drainageat theMount Billingenthreshold is inferredfrom a distinct2.4%0 •5180oscillation in oneof theinvestigated cores. The estimatedradiocarbon age for thisevent is approximately10,900 14C years. The finaldrainage, which ended BIL history,occurred in two stepsduring the younger part of theradiocarbon plateau at 10,00014C years. Biostratigraphy suggeststhat the final drainagetook placein the climaticallywarm early Preboreal.At approximately11,150 14C years, the deglacial trend toward lighter isotopic composition was interruptedand slightly reversed and is interpretedto representthe onsetof the coldYounger Dryas period.For a few hundredyears prior to the YoungerDryas, melting of the Fennoscandianice sheet appearsto havebeen rapid. Besides the dramaticdrainage events and reduced melting during the YoungerDryas, the recordsdisplay several other salinity variations. These variations may reflect ice marginoscillations during the YoungerDryas, previously identified throughout Scandinavia as ice marginaldeposits. The isotopicrecords suggest approximately 10 climaticallyinduced ice marginrecession/readvances of the southeasternFennoscandian ice sheetduring the YoungerDryas.

Introduction maximum age and is based on terrestrial macrofossilremains found in redepositedsediments in the former BIL area which are Meltwater trappedby retreatingice sheetsand topography thought to representthe final drainage. during deglaciationcreates ice-dammed lakes. When the ice Olausson [1982] presentedthe first stableisotope evidence front drawsback and passesa topographicthreshold, these ice- for short-lived freshwater injections in the shallow marine dammed lakes are drained or catastrophically lowered. In environment of the Swedish west coast during the last Scandinavia,the Baltic Ice Lake (BIL) was createdduring the deglaciation.His •5180 results from benthic foraminifera Bolling and Aller0d interstadials.During the Younger Dryas, revealed two distinct meltwater spikes in one of four cores the BIL was probablyat its maximumsize (Figure 1), with an investigated. Owing to contamination, conventional estimated10,000 km3 of glacialmeltwater [Olausson, 1982]. radiocarbondating on shell fragmentsfound in the cores was A westwarddrainage of the BIL at the northerntip of Mount problematic [Olsson, 1982] and could not be used to establish Billingen (Figure 1) was first hypothesizedover 85 yearsago an age model [ Cato et al., 1982]. Nevertheless,on the basis of [Munthe, 1910], and the geomorphologicalfeatures in this biostratigraphic age control [Knudsen, 1982; Miller, 1982; area have been the subject of debate ever since. More Robertsson, 1982], these meltwater eventswere assignedto conclusiveevidence supportingthe Mount Billingen drainage the Aller0d and Preboreal chronozones [Working Group, hypothesiswas lacking until barrenbedrock and coarsegravel 1982]. Two distinct •5180 oscillationsin benthic foraminifera deposits,indicative of a large flood event with a westward were later also observedin a core from the Skagerrak,south of route, were discovered in an area 5-10 km west of Mount Norway [Erlenkeuser, 1985]. These events were tied by Billingen [StrOmberg,1992]. A radiocarbondate for the final tephrastratigraphyto the Younger Dryas chronozone[Stabell drainageof the BIL indicatesan age of approximately10,300 and Thiede,1985] but were regardedas too old to representthe 14C years [Wohlfarth et al., 1993]. This date is regardedas a final drainageof the BIL [Erlenkeuser, 1985]. It is evident that the final drainage of the BIL passed •Departmentof Geologyand Geochemistry, Stockholm University, throughthe Mount Billingen threshold[StrOmberg, 1992] and Stockholm,Sweden. likely that it left its imprint in the marine sedimentrecord west 2Lamont-DohertyEarth Observatoryof ColumbiaUniversity, of the drainage outlet. The meltwater events observed Palisades,New York. [Olausson,1982; Erlenkeuser,1985], however,can obviously 3SawyerEnvironmental Research Center, University of Maine,Orono. not be correlated, and none seem to agree with the age Copyright1997 by the American Geophysical Union. proposedfor the final BIL drainage[Wohlfarth et al., 1993]. This age discrepancymay be due to differentdating techniques. Papernumber 96PA02879. Moreover, the drainagemay have been a dramaticevent with a 0883-8305/97/96PA-02879512.00 duration of only one or a few years [Johansson, 1926, 1937;

39 40 BODI•NET AL.:THE FINAL DRAINAGE OF THE BALTIC ICE LAKE

30'

70 ø --

60 ø -

Figure 1. Maps showing the areal extent of the Fennoscandianice sheet and Baltic Ice Lake during the Younger Dryas deglaciation standstill.The detailed map showsthe paleogeographyduring Younger Dryas and the position of the ice margin in southwestSweden and its relation to the Mount Billingen threshold. Black area indicates Mount Billingen. Stars indicate the positions of the Moltemyr and Solberga-2 cores. Arrows indicate the drainageroute for the Baltic Ice Lake. (Adapted from Bj6rck and $vensson[ 1994]).

Bj6rck, 1995]. An event of such short duration could easily be extensive examination by the Pleistocene/Holoceneboundary overlooked in a sedimentcore. Thus the final drainagemay not working group. At Solberga, two 27.3 m long cores were have been recordedin previously investigatedcores. collected with a Foil piston corer [Cato, 1982a]; one was used This study is an attempt to unravel some of the ambiguities up by the working group and the other was stored for future concerning the late history of the BIL as revealed in the investigations.The Solberga-2 core was almost complete, and shallow marine environment. To accomplish this goal, we the sampling resolutionused was constrainedonly by the 5-cm reinvestigatedthe core containing the evidence of short-lived storage sections applied throughout the core. We measured meltwater events (Moltemyr) and one additional site oxygen isotopes in benthic foraminifera from 58 samples (Solberga) where these events originally were not observed, from the Moltemyr core and 278 samplesfrom the Solberga-2 using high-resolution sampling and accelerator mass core. Compared with the previous study [Olausson, 1982], spectrometry(AMS) 14Cdating. sampledensity is increasedby nearly a factor of 3 at Moltemyr and 10 at Solberga. Material and Methods Sediment consisted of clay, clayey silt, and silty clays Among 14 sites surveyedon the Swedishwest coast for a [Cato, 1982b], and foraminifer abundancesvary from almost type locality for the Pleistocene/Holocene boundary, barren to more than 700 specimens per gram of sediment Moltemyr and Solbergawere found suitableas stratotypeand [Knudsen, 1982]. Foraminifera were extracted by hypostratotypesections [Working Group, 1982]. Moltemyris disaggregating the samples in hot water and wet-sieving a small bog situatedapproximately 55 m above sea level in a through a 63-•tm mesh. Specimensof the benthic foraminifer hilly area, and the Solbergasite is situatedin a valley a few Elphidiurn excavaturn were picked from the >125-[.tm size meters above sea level [Freddn, 1982]. After ice withdrawal, fraction, and oxygen isotopeswere analyzed using a Finnegan the subsequentisostatic uplift created an archipelago in MAT 251 mass spectrometer with an attached Carousel-48 southwest Sweden (Figure 1). At the beginning of the automatic preparation system. Accelerator mass spectrometry Holocene,the Moltemyr and Solbergasites were locatedin the radiocarbon dating were made at the National Ocean Sciences outer rim of this archipelagoat paleodepthsof approximately AMS Facility at Woods Hole OceanographicInstitution and 30 and 50 m, respectively[Olausson, 1982]. the TandemLaboratory at UppsalaUniversity. All 14Cages are The Moltemyr core is 6.5 m long and was taken with a based on benthic foraminifera and, when possible, modified Russian peat sampler [Cato, 1982a]. The core was monospecific samples were used. Sampling for radiocarbon originally cut into 10-cm sections in the lower part, 5-cm dating was limited to those intervals containing sufficient sectionsin the middle part, and 2.5-cm intervals in the upper foraminifera, and direct dating of the meltwater events part. These samplesare storedin plastic containers,and it is observedin the •5•80records was therefore not possible.In the not possibleto determinethe top and bottom of each interval. Solberga-2core, sampleintervals of up to 30 cm had to be used The sampling resolution used is partly determined by the in order to obtain a sufficient amount of material for dating. storageintervals. Sampling was also limited to those intervals Chronostratigraphic terminology follows Mangerud et al. containing a sufficient amount of sediment left after the [1974]. BODI•N ET AL.: THE FINAL DRAINAGE OF THE BALTIC ICE LAKE 41

Table 1. AcceleratorMass Spectrometer14C Dates

Core Depth,cm 14CAge, years •4C Age* Sourcematerial AccessionNumber

Solberga-2 1125-1140 9,610 + 45 9,170 E. excavatum OS-4532 Solberga-2 1140-1155 8,885 + 90 8,445 E. excavatum Uu-10300 Solberga-2 1330-1355 4,550 + 80 4,110 E. excavatum Uu- 10301 Solberga-2 1530-1555 7,730 + 85 7,290 E. excavatum OS-4527 Solberga-2 1815-1830 10,300+ 110 9,860 E. excavatum OS-4528 Solberga-2 1930-1935 10,800+ 85 10,360 N. labradoricum OS-4526 Solberga-2 2200-2230 10,700+ 75 10,260 E. excavatum OS-4529 Solberga-2 2650-2660 11,600+ 95 11,160 E. excavatum OS-4530 Solberga-2 2730-2735 12,400+ 95 11,960 N. labradoricum OS-4531 Moltemyr 272.5-275 10,650+ 40 10,210 Mixed benthic OS-2350 Moltemyr 277.5-280 10,500+ 40 10,060 Mixed benthic OS-2353 Moltemyr 425-430 10,900+ 55 10,460 Mixed benthic OS-2352 Moltemyr 560-570 12,450+ 55 12,010 N. labradoricum OS-2351 Moltemyr 590-600 12,750+ 45 12,310 N. labradoricum OS-2349

*Correctionis madefor an assumedocean reservoir effect of 440 years.

Radiocarbon ages include the correction of an ocean indicate that the Moltemyr sequenceranged in age from reservoir effect of 440 years, which is based on radiocarbon approximately 12,800 to 10,000 •4C years. As in the ages obtained from recent marine shells from the coast of Solberga-2 core, a small age reversal is recorded at the Norway and southwest Sweden [Mangerud and Gulliksen, radiocarbonplateau at approximately10,000 •4C years.The 1975]. Marine and terrestrial radiocarbon dates for the Vedde averagesedimentation rate at Solberga(Figure 2) during the ash bed indicate a higher ocean reservoir effect during the Older Dryasand Aller0d was approximately 1.5 mrn/•4Cyear Younger Dryas and that the correctioncould be on the order of whichincreased to about6 mrn/14Cyear in theYounger Dryas. 700-800 years [Bard et al., 1994]. It is thus possiblethat some Holocene average sedimentationrates decreasedto about 1 of the estimated •4C ages for the events identified and mm/•4C year. At Moltemyr,the averagesedimentation rate discussed in this paper are a few hundreds years too old. duringthe late Bolling was approximately1 mrn/•4Cyear. The However, as the 400-year correctionhas commonlybeen used almost1500 •4C yearsrepresenting the OlderDryas, Aller0d, in the past, it will make it possibleto compareour resultswith and early Younger Dryas (between560 and 430 cm) have an previous observations of the Younger Dryas in marine estimatedsedimentation rate of lessthan 1 mrn/•4Cyear while environmentsof the North Atlantic region.

Results 4000 6000 8000 10000 12000 14000 1200 I I I I -- 200 Chronology and Estimated Sedimentation Rates 1400 - - 400 Nine monospecificforaminifer sampleswere AMS •4C dated in the Solberga-2 core (Table 1). Using a 440-year 1600 - yr- 600 correctionfor reservoir age [Mangerud and Gulliksen, 1975], 18oo - the sequencecovers the time interval from 11,960 to 4,110 - 800 •4C years.Above 1330 cm, thereis a majorage reversal which 2000 - - 1000 is attributedto reworkedsediment in the upperpart of the core. During isostaticuplift, hills surroundingSolberga reached sea 2200 - - 1200 level long before the Solbergacoring site and were exposedto wave action with subsequentredeposition of surface sediment. 2400 - -1400 Some of the abraded sediment was likely redeposited at Solberga.Samples at 1935 and 2230 cm have ages of 10,360 2600 - - 1600 and 10,260 14C years, respectively,suggesting that these 2800 1800 levels fall into the plateau period of constantages at about 4000 6000 8000 10000 12000 14000 10,000 •4C yearson the radiocarboncalibration curve [Bard et al., 1990; Edwards et al., 1993; Kromer and Becker, 1993; ]4Cyears Goslar et al., 1995]. Rather than indicating sediment Figure 2. Age-depth relations in the Solberga-2 and reworking, this age inversion more likely reflects a high Moltemyr cores. Graphs are based on the younger age sedimentaccumulation rate at Solbergaduring the late Younger alternativeof 10,26014C years at 2230 cm in the Solberga-2 Dryas. coreand 10,060 •4C yearsat 280 cm in theMoltemyr core (see Five samplesin the Moltemyr core were AMS dated (Table text and Table 1). Graphsare plottedat samedepth scale, and 1). Assumed linear sedimentation rates between the dates agesare correctedfor an assumed440-year reservoir age. 42 BODI•N ET AL.: THE FINAL DRAINAGEOFTHE BALTIC ICE LAKE during the late Younger Dryas the estimated average depths.Because the Moltemyr site was situatedabout 20 m sedimentationrate is about4-6 mm/14Cyear. shallowerthan the Solbergasite during the time of sediment deposition,it is likely that the bottomwater compositionat Oxygen Isotopes Moltemyr was more affectedby regionalmeltwater discharge than the deeper Solberga. Although the records display The •180 recordsfrom Solbergaand Moltemyr display the differencesin both patternand amplitude,it is importantto expecteddeglacial trend toward lower values (Tables 2a and 2b note that the two distinct high-amplitudeoscillations found andFigure 3). In Solberga,this changeoccurred in two steps. closely above the 19-m level at Solbergaare also evident in The first was a 1.5%odecrease between approximately 2700 the top of the Moltemyr record at 272.5 cm (9%o)and 267.5 cm (5%ø). and2650 cm, The overlying7 m havea •180 recordmarked by a gradual trend toward higher values. As is evident in the Discussion smoothedrecord (Figure 4), the magnitudeof changeis about 0.5%o. A second decrease of 1.5%o occurs between Aller•d approximately1950 and 1600 cm. A stepwisedeglaciation The deglaciation of southern Sweden occurred in several trend toward generallylighter isotopiccomposition is also steps, as is evident in end moraine lines created in southwest evidentin the less completeMoltemyr core. In the smoothed Swedenduring short glacial standstills or readvancesduring record(Figure 4), the first stepappears as a decreasefrom 3%o the Bolling, Aller0d and YoungerDryas [Fredtin, 1988, and to averagevalues of around-1%o.The seconddistinct drop at 280 cm reflects the dramatic oscillation that occurred in the references therein]. The first fi180 decreasethat could reflect beginningof the final deglacialstep. deglaciationin the Solberga-2core (2700-2650cm) occurred betweenapproximately 11,300 and 11,150 •4C years.This A strikingfeature of the •180 recordsis the appearanceof severaloscillations with amplitudesabove 1%o. These high- abrupt changein fi180 was likely causedby increased amplitudeoscillations, superimposed on the deglacialtrend meltwater discharge when the Fennoscandianice sheet in southwesternSweden retreated rapidly and correspondsto a towardlower fi•80, appearto havebeen created quite rapidly period of increasedsea surfacetemperatures (SST) in the and are in manycases based upon single samples. The most southeastNorwegian Sea [Ko• Karpuz and Jansen, 1992; conspicuousof the oscillationsrecorded at Solbergaare, Lehmanand Keigwin, 1992]. Preceding the distinct drop, there however,represented by two sampleseach. The first of these is a 0.7%ofi•80 increasethat culminatesat about11,600 14C distinctfi•80 spikesoccurs at 2505 cm andshows a maximum years.This increase may reflect reduced glacial melting during amplitudeof 2.4%o.Two otherdistinct oscillations defined by a glacialreadvance that created one of the moraineridges in more than one sampleare alsorecorded at 1890 cm (2.5%o)and southwesternSweden during the Aller0d[e.g., Freddn, 1988]. 1865 cm (3.7%o)in the Solberga-2core. The ageof thisevent closely corresponds to the Norwegian The Moltemyrrecord is differentfrom Solberga-2.This can Seacold SST phaseobserved between 11,800 and 11,500•4C be attributed,in part,to the lowersampling frequency used for years[Koq Karpuz and Jansen, 1992] and at 11,70014C years the Moltemyrcore. The differencein patternand amplitude of [Lehmanand Keigwin, 1992]. In the Moltemyrcore, this the fi•80 oscillationscan also be explainedby differentwater intervalis poorly dated.

Table 2a. TheRecord of •80 in Elphidiumexcavatum Versus Depth From Moltemyr

fi•80, fi180, fi180, 5180, Depth,cm %oPDB Depth,cm %oPDB Depth,cm %oPDB Depth, cm %o PDB

242.5-245 -1.65 317.5-320 -0.61 395-400 -0.29 510-520 0.76 245-247.5 -3.00 320-322.5 0.28 405-410 -3.98 520-530 1.08 262.5-265 -6.14 322.5-325 -0.68 410-415 0.71 530-540 2.84 265-267.5 -6.75 327.5-330 0.08 415-420 1.21 540-550 2.46 267.5-270 -1.42 330-332.5 -0.55 420-425 -1.04 560-570 2.89 270-272.5 -8.83 332.5-335 -3.58 425-430 -2.03 570-580 2.76 272.5-275 -4.60 337.5-340 -0.96 430-435 -1.12 580-590 2.87 275-277.5 -4.18 340-342.5 0.36 435-440 -1.51 590-600 3.17 277.5-280 -2.95 345-347.5 -1.24 445-450 0.00 600-610 2.93 280-282.5 -3.41 347.5-350 -1.90 450-460 0.22 610-620 2.89 282.5-285 -2.43 355-360 -2.02 460-470 -1.68 620-630 2.77 287.5-290 0.17 360-365 0.30 470-480 -1.54 630-640 2.74 302.5-305 -1.94 370-375 -0.95 480-490 -0.76 640-650 2.57 305-307.5 -2.62 375-380 -1.07 490-500 0.95 307.5-310 -0.74 385-390 -1.86 500-510 2.12

PDB, Pee Dee belemnite BOD•N ET AL.: THE FIN• DRAINAGE OF THE BALTIC ICE LAKE 43

Table 2b. The Recordof •5180in Elphidiumexcavatum Versus Depth From Solberga-2

•5180, •5•80, •5•80, •180, Depth,cm %oPDB Depth,cm %oPDB Depth,cm %oPDB Depth, cm %0PDB

1325-1330 0.80 1675-1680 1.58 2025-2030 2.15 2395-2400 1.65 1330-1335 0.59 1680-1685 1.16 2030-2035 2.35 2400-2405 1.91 1335-1340 0.61 1685-1690 1.20 2035-2040 2.23 2405-2410 1.69 1340-1345 0.65 1690-1695 1.16 2040-2045 2.00 2410-2415 1.66 1345-1350 0.26 1695-1700 1.40 2045-2050 2.43 2415-2420 2.13 1350-1355 0.33 1700-1705 1.20 2050-2055 2.51 2420-2425 1.92 1355-1360 0.46 1705-1710 1.44 2055-2060 2.53 2425-2430 1.72 1360-1365 0.49 1710-1715 1.45 2060-2065 2.50 2430-2435 1.63 1365-1370 0.43 1715-1720 1.34 2065-2070 2.24 2435-2440 1.40 1370-1375 0.80 1720-1725 1.69 2070-2075 1.89 2440-2445 1.86 1375-1380 0.82 1725-1730 1.60 2075-2080 1.66 2445-2450 1.99 1380-1385 1.12 1730-1735 1.47 2080-2085 2.42 2450-2455 1.91 1385-1390 1.14 1735-1740 1.54 2085-2090 1.94 2455-2460 2.10 1390-1395 1.12 1740-1745 1.36 2090-2095 1.89 2460-2465 1.62 1395-1400 0.68 1745-1750 1.59 2095-2100 1.05 2465-2470 1.81 1400-1405 0.55 1750-1755 1.71 2105-2110 2.45 2470-2475 1.51 1405-1410 0.49 1755-1760 0.46 2110-2115 2.33 2475-2480 2.12 1410-1415 0.67 1760-1765 1.93 2115-2120 2.32 2480-2485 1.96 1415-1420 0.93 1765-1770 1.68 2120-2125 2.29 2485-2490 1.71 1420-1425 0.53 1770-1775 1.90 2125-2130 2.26 2490-2495 1.92 1425-1430 0.45 1775-1780 1.91 2130-2135 2.00 2495-2500 1.86 1430-1435 0.40 1780-1785 1.93 2135-2140 2.27 2500-2505 -0.54 1435-1440 0.20 1785-1790 1.98 2140-2145 2.16 2505-2510 0.55 1440-1445 0.43 1790-1795 1.98 2145-2150 1.92 2510-2515 1.88 1445-1450 0.73 1795-1800 1.80 2155-2160 2.03 2515-2520 1.07 1450-1455 0.54 1800-1805 1.91 2160-2165 1.76 2520-2525 2.04 1455-1460 0.73 1805-1810 1.63 2165-2170 1.62 2525-2530 2.14 1460-1465 0.97 1810-1815 0.57 2170-2175 1.59 2530-2535 1.66 1465-1470 0.65 1815-1820 2.09 2175-2180 1.54 2535-2540 2.12 1470-1475 0.37 1820-1825 1.78 2180-2185 1.70 2540-2545 1.40 1475-1480 1.40 1825-1830 2.04 2185-2190 1.43 2545-2550 1.63 1480-1485 0.28 1830-1835 1.85 2190-2195 1.65 2550-2555 1.94 1485'1490 -0.11 1835-1840 1.87 2195-2200 1.86 2555-2560 1.99 1490-1495 0.67 1840-1845 2.00 2200-2205 2.23 2560-2565 1.91 1495-1500 0.58 1845-1850 2.07 2205-2210 2.72 2565-2570 1.85 1500-1505 0.53 1850-1855 2.03 2210-2215 2.74 2570-2575 1.79 1505-1510 0.69 1855-1860 0.48 2215-2220 2.71 2575-2580 2.15 1510-1515 0.57 1860-1865 -1.61 2220-2225 2.54 2580-2585 1.64 1515-1520 0.79 1865-1870 1.34 2225-2230 2.48 2585-2590 2.04 1520-1525 1.26 1870-1875 2.23 2230-2235 1.73 2590-2595 1.62 1525-1530 0.76 1875-1880 2.29 2235-2240 1.96 2595-2600 1.93 1530-1535 0.49 1880-1885 2.12 2245-2250 0.93 2600-2605 1.89 1535-1540 1.02 1885-1890 -0.39 2250-2255 1,62 2605-2610 2.08 1540-1545 0.94 1890-1895 0.64 2255-2260 1.82 2610-2615 1.86 1545-1550 0.77 1895-1900 2.09 2260-2265 2.14 2615-2620 1.98 1550-1555 0.78 1900-1905 1.61 2265-2270 2.24 2620-2625 1.91 1555-1560 0.80 1905-1910 2.13 2270-2275 1.44 2625-2630 2.38 1560-1565 1.14 1910-1915 2.14 2275-2280 1.31 2630-2635 2.32 1565-1570 0.84 1915-1920 2.43 2280-2285 2.30 2625-2640 1.81 1570-1575 1.04 1920-1925 2.08 2285-2290 0.66 2640-2645 2.24 1575-1580 0.97 1925-1930 2.18 2290-2295 2.08 2645-2650 1.81 1580-1585 0.85 1930-1935 2.02 2295-2300 1.98 2650-2655 2.70 1585-1590 0.78 1935-1940 1.43 2305-2310 1.94 2655-2660 2.40 1590-1595 0.53 1940-1945 2.19 2310-2315 1.90 2660-2665 2.68 1595-1600 0.73 1945-1950 2.48 2315-2320 1.81 2665-2670 3.06 !600-1605 0.66 1950-1955 1.93 2320-2325 0.22 2670-2675 3.10 1605-1610 0.95 1955-1960 1.93 2325-2330 1.68 2675-2680 3.26 1610-1615 1.19 1960-1965 2.42 2330-2335 1.22 2680-2685 3.39 1,6!5•1620 •1.10 ..... 1965•1970 2.45 233•5,.2340 1,67 2685-2690 3.36 1620-1625 1.19 1970-1975 2.49 2340-2345 1.82 2690-2695 3.46 1625-1630 1.21 1975-1980 2.44 2345-2350 1.59 2695-2700 3.20 1630-1635 0.95 1980-1985 2.57 .2350-23.55 1.47 2700,2705 3.24 1635',1640 1.56 1985-1990 2.52 2355-2360 1.74 2705-2710 3.26 1640-1645 1.49 1990-1995 2.34 2360-2365 1.85 2710-2715 2.75 1645-1650 1.15 1995-2000 2.27 2365-2370 1.77 2715-2720 2.90 1650-1655 1.17 2000-2005 2.25 2370-2375 2.07 2720-2725 3.05 1655-1660 1.43 2005-2010 1.91 2375-2380 2.46 2725-2730 2.84 1660-1665 1.79 2010-2015 2.34 2380-2385 1.65 2730-2735 2.97 1665-1670 1.46 2015-2020 2.27 2385-2390 1.45 1670-1675 1.48 2020-2025 2.13 2390-2395 1.26 44 BODI•NET AL.'THE FINAL DRAINAGE OF THE BALTIC ICE LAKE

•)180(per mil) /)180(per mil)

4 3 2 1 0 -1 -2 4 2 0 -2 -4 -6 -8 -10

1200

4,110 1400 -

_ BILL-2A

_ 7,290 10,210 10,060 1600 -

1800

_ 9,860BILL-2B BILL-2A 10,360 10,460 2000

2200 - 10,260

2400 - 12,010

_ 12,310 BILL- 1

2600 - 11,160 4

_ 11,960 •

2800 ,

SOLBERGA-2 MOLTEMYR

Figure 3. The benthic Elphidium excavatumb180 records in the Moltemyr and Solberga-2 cores. Radiocarbonages are correctedfor an assumed440-year reservoirage.

J180(per mil) •)'80(per mil) 3 2 1 0 -1 -2 4 2 0 -2 -4 -6 -8 -10 1200 , I , I , I , I , I , 200 , I , I , I , I , I , I .

300 1600

• 400 •" 2000 •, 500

2400 600

t movingaverage Five pointmoving average 2800 700

SOLBERGA-2 MOLTEMYR

Figure 4. Samerecords as in Figure 3 but smoothed. BODI•NET AL.:THE FINAL DRAINAGEOF THE BALTICICE LAKE 45

The Younger Dryas and the Drainage of the Baltic the drainageevent and is obtainedfrom radiocarbondated bulk Ice Lake sediments [Svensson, 1989] and AMS dated leaves and insect remains found in redepositedsediments thought to represent Both deep-seaand eustaticsea level recordsshow that the the final drainage [Wohlfarth et al., 1993]. deglaciationoccurred in two majorsteps, separated by reduced Reduced melting of glacial ice sheetsduring the Younger transportof glacialmeltwater to the oceanduring the Younger Dryas and the drainage of the BIL are evident in the isotope Dryas[e.g., Fairbanks,1989; Jansen and Veum,1990; Polyak recordsfrom Moltemyr and Solberga.In the Solberga-2record, et al., 1995]. Paleoclimaticrecords also showthat the Younger 11,150 14C years marks the beginningof the 7-m interval Dryas deglaciationstandstill was accompaniedby a returnto (2650-950 cm) of high-frequency oscillations superimposed nearlyglacial atmospheric temperatures and SST in the North on a trend of fairly constantor slightly increasing•ilsO Atlanticregion [e.g., Dansgaard et al., 1982, 1993; Grooteset (Figures 3 and 4). Considering the vast amount of al., 1993; Ruddimanet al., 1977; Kof Karpuz and Jansen, geomorphological evidence for a standstill and subsequent 1992]. Recent ice core observationsand marine recordsfrom readvance of the Fennoscandianice sheet during the Younger low latitudes[Thompson et al., 1995;Guilderson et al., 1994; Dryas [Andersenet al., 1995a], it is likely that the observed Kennett and Ingram, 1995; Hughen et al., 1996], and •ilsO trend reflectsthe reducedrelease of glacial meltwaterof geomorphologicalevidence from New Zealand[Denton and local origin. Along the coast of Norway, early Younger Dryas Hendy, 1994] indicate that the Younger Dryas cooling was readvancesare radiocarbondated to 11,000-11,300 inc years global in extent. [Andersenet al., 1995b; BergstrOm, 1995]. The beginningof In Scandinavia,the Younger Dryas was characterizedby the slowdown or halt in glacial melting, as indicated in the depositionof extensiveice-marginal end moraines[Andersen Solberga-2 core, also closely correspondsto the onset of the et al., 1995a]. These end moraines occur along the coast of climatic Younger Dryas which is AMS dated to 11,200 •4C Norway, in southernSweden and Finland, and in northwest yearsin SST recordsfrom the NorwegianSea [ Kof Karpuzand Russia[Andersen et al., 1995b;Lundqvist, 1995; Rainio et al., Jansen, 1992; Lehman and Keigwin, 1992]. 1995] and mark the areal extentof the Fennoscandianice sheet In terrestrial records from southern Sweden, a first BIL during the YoungerDryas deglaciationstandstill (Figure 1). drainagepredates the climatically cold Younger Dryas [BjOrck, The Younger Dryas ice-marginal zone in Sweden, which consistsof the Sk0vde and Billingen moraines,is developed 1979]. The freshwater spike recorded as a distinct 5180 both east and west of Mount Billingen [Lundqvist, 1995, and oscillation at 2505 cm in the Solberga-2 core, however, is referencestherein]. These morainesare difficult to radiocarbon apparently of early Younger Dryas age. As the minimum date, and estimatedages for their depositionare based on isotopic ratio of meltwater from the Fennoscandianice sheet correlations to radiocarbon dated moraines in southern was -25%0 [Olausson, 1982], the 2.4%0 change in 5180 Norway, stratigraphy,clay varve chronology,or are inferred indicates that bottom water salinity at Solberga decreased from radiocarbondated marine shell depositsin southwestern approximately3%0. From the estimatedsedimentation rate, the Sweden. Such ages vary from approximately 11,300 to 10, duration of the event appearsto have been rapid and may have 600 InC years for the Sk0vdemoraine and 10,600 to 10,200 occurredon the order of approximatelya decadeor less. Using Inc yearsfor the morenortherly Billingen moraine [Berglund, the radiocarbondate of 10,360 14Cyears at 1935 cm to createa 1979; Bj6rck and Digerfeldt, 1982, 1984; Str6mberg, 1985; linear age depth relation, the estimated age for this low- Freddn, 1988]. Nevertheless, the Sk0vde and Billingen salinity event is 10,995 14C years. Using the younger age morainesprovide clear evidence that duringthe YoungerDryas alternativeof 10,260 14Cyears at 2230 cm givesan estimated the southernmostmargin of the Fennoscandianice sheetwas age of 10,855 14Cyears for the event.The proposedage for a situatedsouth of the topographicthreshold that servedas the first (approximately 10-15 m) lowering of the BIL is 11,200 outlet for the final drainageof the B IL. 14Cyears [Bj6rck, 1979].This radiocarbonage is obtainedon Shoredisplacement records for the Baltic area indicatethat bulk sedimentand may thereforebe a few hundredyears too old onemajor lowering of the BIL occurredlong before its assumed [Wohlfarth et al., 1993, and referencestherein]. As there is no final drainagein the late YoungerDryas. This first loweringof signin the isotopicrecord for a rapid drainageevent in the late the BIL took place at a time close to the Aller0d/Younger AllerOd chron,it may be that the conspicuousfreshwater spike Dryas chronboundary [Donner, 1982, and referencestherein]. with an averageage of 10,925 14C yearsrepresents the first Accordingto Bfiirckand Digerfeldt[1984, 1989], the ice front drainage of the BIL, which is a documentedevent in shore was situated north of Mount Billingen in the late Aller0d displacementrecords in the Baltic region. The first Baltic Ice chron, with a subsequentreadvance during the early Younger Lake Lowering (BILL-1) is not recordedat Moltemyr, probably Dryas.Included in this modelis a first BIL drainage,when the due to the coarser sampling resolutionused in the Moltemyr retreatingice front passedMount Billingenand the lake level core. Since BILLol is recorded in only one core and its droppeddramatically by approximately10-15 m [Bj6rck, estimatedage appearstoo young to representthe assumedlate 1979]. When the ice front readvancedin the early Younger AllerOd drainage, we regard its correlationto the circum-Baltic Dryas and the Sk0vdeand Billingenmoraines were deposited, shore displacementrecords as tentative until its appearance the BIL was again dammed.The secondand final drainageof and age are confirmedin other marine recordsof the Swedish the BIL lowered the lake by approximately 25 m and is west coast. generallybelieved to have occurredin the late YoungerDryas The secondand final drainageof the BIL is clearly evident at 10,300 Inc yearsago [e.g., Bergstenand Nordberg,1992; in both recordsand took place in two steps.In Solberga-2,the Bergsten,1994; Lundqvist, 1995]. This is a maximumage for drainageis recordedas 2.5%0and 3.7%0oscillations occurring 46 BODI•NET AL.: THE FINAL DRAINAGEOF THE BALTICICE LAKE just above 1900 cm, and in the Moltemyr core as 9%0and 5%0 levels (Table 3). Thus, on the basisof biostratigraphy,the oscillations at the top of the record. On the basis of the final drainageappears to shortlypostdate the YoungerDryas. radiocarbondates (Table 1), four differentage estimatescan be The warmingthat marksthe beginningof the Preborealwas made for each event. For the older event (BILL-2A), estimated very rapid and occurredin only a few decades[Alley et al., agesrange between 9,930 and 10,205 •4C years,and between 1993]. The temperaturerise over Greenlandwas of the orderof 9,905 and 10,200 •4C yearsfor the youngerevent (BILL-2B) 8ø-20øC[Grootes et al., 1993; Johnsen et al., 1995' Kerr, (Table 3). Thesedates indicate that the final drainageoccurred 1996] and in the southeastNorwegian Sea SST roseabruptly during the radiocarbonplateau at approximately10,000 •4C by 9øC [Koq Karpuz and Jansen, 1992]. This warming years. apparentlycaused the ice marginat Mount Billingento retreat, The distinct freshwater spike observed by Erlenkeuser and the BIL drained. [1985] in the Skagerrak south of Norway was dated by As seenin the •5180records from Solberga and Moltemyr tephrachronologyand is regardedas too old to representthe andin the Skagerrak[Erlenkeuser, 1985], the final drainage assumedB IL drainageat 10,300 •4C years.The volcanicash took place in two dramaticsteps, the first BILL-2A and the horizonfound, the Vealaleash, was believedto indicatean age final BILL-2B. This may indicatethat a readvance..ofthe ice of 10,600•4C yearsat a level about100 cm belowthe recorded margindammed the BIL afterBILL-2A. As the drainagetook freshwaterspike. New radiocarbondates on terrestrialvegetal place in a period when the climate was warm, anda readvance material from three Norwegiansites, however, now suggest of the ice marginappears unlikely, a second•threshold may that the radiocarbon age for the Vedde ash is closer to have been involved. Such a threshold is •IOCatedi n an area approximately10,300 •4C years [Bard et al., 1994; Birks et approximately30 km east of MOUntBillingen [StrOmberg, al., 1996]. A youngerage for the eventwas alsoindicated by 1992].This threshold, the narrow•Srlen Valley, regulated the radiocarbondated bivalve fragmentsobtained from a 35-cm outflowafter the lake level had been lowered by approximately interval closeto the ash horizon.The age of the fragmentsis 15 m. Anotherpossibility is that icebergsmay havedammed 10,260_+ 280 •4C yearsand is basedon a 400-yearcorrection the narrowoutlet either at thenorthern tip of MountBillingen for reservoirage [ Stabell, 1985]. Thus, the recordedfreshwater or at 6rlenValley, causing drainage to ceasetemporarily or event appearsto be youngerthan previouslyassumed. On the slow down [StrOmberg, 1992, and referencestherein]. basis that the freshwaterevent is recordedas two closely However, a detailedexplanation for the mechanismbehind the spaced/5•80 oscillations,similar to the Solberga-2and two-step drainageis still to be determined. Moltemyr cores,it appearsthat BILL-2A and BILL-2B alsoleft The identificationof BILL-1 is clearevidence of icemargin chemical imprints in the sediment record of the deeper oscillationsduring the Younger Dryas and confirms the Skagerrak. No volcanic ash horizons have been identified in hypothesis of an ice margin position north of Mount the Solberga-2and Moltemyr cores. Billingen before the readvance that created the Skfivde and In a study of lacustrine records from the former BIL area Billingenmoraines [Donner, 1969;Berglund, 1979; BjOrck [Svensson, 1989; see also Wohlfarth et al., 1993], it is and Digerfeldt, 1984, 1989]. Our resultsindicate that the ice suggestedthat the final drainagetook place at, or slightly marginreceded from the northerntip of MountBillingen and before,the YoungerDryas/Preboreal boundary. On the basisof the BIL was drainedat about10,900 14Cyears. Whether the an observed warming, inferred from changes in benthic Skfivdemoraine was deposited before or afterBILL-1 cannotbe foraminifera, the end of the Younger Dryas is identified at determinedsince ages proposed for its depositionrange 1890 cm in the Solbergacore and at 345 cm in the Moltemyr betweenapproximately 11,300 and 10,600.14 C years [ Fredgn, core [Knudsen, 1982]. Using planktonic diatoms, Miller 1988, and referencestherein]. Perhaps the SkfiVdemoraine [1982] observedthis changeat 1935 cm in Solbergaand at marksthe onsetof the YoungerDryas at 11,150•4C years, as 345 cm in Moltemyr. This time lag betweenthe benthicand suggestedby ourisotopic record from Solberga. A readvanee0f planktonic record at the deeper Solberga may be reasonable theice frontis evidentby theBillingen end moraine which has ': becausea climatic warmingwould first affect surfacewaters. In agesvarying between 10,600 and 10,200 14Cyears. This is both cores, BILL-2A and BILL-2B is recorded at or above these alsoconfirmed by the isotopicrecord, since the BIL musthave beendammed again before BILL-2A took place. The Billingen endmoraine appears as a swarmof smallice-marginal ridges Table 3. EstimatedRadiocarbon ages for Baltic Ice Lake within an approximately10-km-wide zone just a few Lowering (BILL) Events kilometerssouth of the drainageoutlet (Figure 5). These featuresmay indicatethat the ice marginin this areaoscillated with several recessions on the order of a few kilometers followedby smallerreadvances which created the ice-marginal Core Depth,cm Event Age, 14Cyears ridges.It is believedthat the final drainageof the BIL started subglacialt.ybefore the ice marginpassed the thresholdandthe main drainagetook place [StrOmberg,1977; BjOrckand Solberga-2 2505 B ILL- 1 10,855-10,995 Solberga-2 1890 BILL-2A 9,930-10,155 Digerfeldt, 1984, 1989]. Consideringthe proximity of the Moltemyr 270 B ILL-2A 10,040-10,205 oscillatingice margin to the Mount Billingen threshold,it is Solberga-2 1865 B ILL-2B 9,905-10,040 possiblethat someice recessionsreached a critical point, Moltemyr 265 B ILL-2B 10,025-10,200 allowingthe BIL to subglaciallytap some of its water.Perhaps the BIL was lowered a few meters or less before a minor ice Age rangeis minimumto maximum. marginadvance closed the subglacialoutlet. BODfiN ET AL.: TI-IEFINAL DRAINAGEOFTHE BALTIC ICE LAKE 47

amounts of meltwater to the sea were complex [e.g., BjOrck, 1995]. Shortly after the drainage, marine waters entered the Baltic basin, and a brackish water phase persistedfor a few hundred years until isostatic rebound narrowed the main connection to the sea and the Ancylus Lake began to rise. Contemporaneouswith land uplift in the north, isostasyhad ceased in the southern Baltic or turned into land subsidence. Meltwater subsequentlyonce again flooded large areas in the south,and a drainageoutlet was developedthrough Denmark in the late Preboreal. Before the outlet was created in the south, the Ancylus Lake was drained through outlets on the Swedish west coast. Owing to land uplift, the inland basin east of Moltemyr and Solberga, the V•nern basin, was isolated from the sea and drained only through two narrow outlets [BjOrck, 0 5 10 km 1995]. These outlets reached the sea north of Moltemyr and I I south of Solberga. Thus the Solberga-2 record monitors the pattern of meltwater outflow from the Ancylus Lake until the southernBaltic outlet was opened approximately 1000 years Figure 5. Detailed map of the ice margin position at the after the final drainageof the B IL. Mount Billingen threshold. Dark patches indicate the ice This part of the record (between approximately 1700 and marginal-deposits of the Billingen end moraine. Arrows 1900 cm; Figure 3) appearsquite different from the Younger indicate a hypothesizedsubglacial drainage of the Baltic Ice Lake. (Adaptedfrom BjOrckand Digerfeldt[1984]). Dryas. With the exceptionof the BIL drainageand two 1-1.5%o spikes at 1815 and 1760 cm, the record is fairly stable and showsa Preborealdecrease in i•180 of approximately0.5%0. The Younger Dryas ice margin oscillationsmay reflect ice This in contrastto the Younger Dryas part of the record,which sheetdynamics caused by differencesin the rate of isostatic shows several i•180 oscillations and an increase of about land rise at different areas along the ice margin or a reorganizationof its mass during the standstill. If the end 0.5%0. The remaining 1%odecrease to mid Holocene values moraines were created due to mechanicalice sheet dynamics, occurred in the succeeding 100 cm (Figure 3) and ends at there would likely be little or no glacial melting. On the other approximately8000 14Cyears, similar to the Skagerrakrecord hand, if they were climatically induced,the recessionswould [Erlenkeuser, 1985]. The more stablepattern of the Preboreal be accompaniedby melting. The meltwater releasedduring record may indicate that the vast amountsof meltwater from these minor ice sheet recessions must have affected the the melting Fennoscandianice sheetgenerally were releasedto salinity of waters proximal to the ice sheet. Apart from the the sea at a fairly constant rate at the Solberga outlet. This major drainage event (BILL-1), the Younger Dryas isotopic pattern can be explained, in part, by the smoothingeffect record in Solberga-2 is characterizedby the appearanceof causedby the lower sedimentaccumulation rate (Figure 2) and numerous 0.5-1%o •180 oscillations and a few oscillations on samplingbias. A minor regulating,and smoothing,effect on the amountof meltwaterreleased to the seamay alsobe linked the order of approximately1.5%o. In the shallowerMoltemyr to the increase in reservoir capacity provided by the record, these salinity variationsare highly amplified. Minor submerginglowlands in the southernBaltic where the Ancylus ice margin oscillationsduring the YoungerDryas, as indicated Lake level rose by 10-25 m [Bj6rck, 1995]. in the Billingen area, are apparently also reflected in the The two •80 oscillations recorded at 1815 and 1760 cm isotopic records of Solberga and Moltemyr. The above speculated"leaking" events of the BIL at Mount Billingen may (estimated ages of about 9800 and 9300 •4C years) may explainsome of the 1-1.5%o•180 oscillationsobserved in the indicate periods of rapid melting of the Fennoscandianice sheet or other dramatic Preboreal ice lake drainage events. interval between BILL-1 and BILL-2A in the Solberga-2 core Such events are known from eastern Finland, where ice- (Figure 3). If our suggestedcorrelation between ice margin oscillations and the observed variations in shallow bottom dammedlakes were drainedshortly after the final drainageof the BIL [Rainio et al., 1995]. Further, in the late Preboreal, the water salinity is correct, the results from Solberga indicate ice-dammedlake Nedre GlfimsjOin Norway was dramatically about 10 climatically inducedrecession/advance events of the drained, and 100 km3 of water may have enteredthe Lake southwesternpart of the Fennoscandianice sheet during the V'•inernbasin in 1 or 2 weeks [Longvaand Thoresen,1991]. As Younger Dryas. As it has been shown that ice margin these spikesare single sampleobservations, they need to be fluctuationsof the Fennoscandianice sheetduring the Younger confirmed in other cores. Dryas differ in age and magnitude in different areas [e.g., Andersen et al., 1995b; Rainio et al., 1995; BergstrOm, Conclusions 1995], it is not possible to link the Solberga and Moltemyr records to variations in total Fennoscandian ice sheet volume. We have tied a late glacial marineoxygen isotope record of the Swedishwest coastto geomorphologicalfeatures that can Early Holocene be seenthroughout the Baltic area, for example,variations in Following the drainage, when the Fennoscandianice sheet the BIL shorelevel, drainagetraces, and ice-marginaldeposits. was melting rapidly, deglaciation and the routing of vast First and foremost,we have capturedthree major drainage 48 BODI•NET AL.' THEFINAL DRAINAGE OF THE BALTICICE LAKE events of the BIL (Table 3) and may have the evidence Atlantic atmosphere-seasurface 14C gradient during the Younger confirming the hypothesisof a first BIL drainageat the Mount Dryas climatic event, Earth Planet. Sci. Lett., 126, 275-287, 1994. Berglurid, B.E., The deglaciationof southernSweden 13,500-10,000 Billingen threshold at a time close to the Aller0d/Younger B.P., Boreas. 8, 89-117, 1979. Dryas chron boundary[BjOrck and Digerfeldt, 1984, 1989]. An Bergsten, H., A high-resolution record of Late glacial and early Holocenemarine sedimentsfrom southwesternSweden; with special estimatedradiocarbon age for this event is about 10,900 emphasis on environmental changes close to the Pleistocene- years. We have also identified the dramatic final drainage of Holocene transition and the influence of fresh water from the Baltic the B IL. Our resultsshow that this drainagewas complex and basin,J. Quat. Sci., 9, 1-12, 1994. Bergsten,H., and K. Nordberg,Late Weichselianmarine stratigraphy of occurredin two steps.The reasonfor this two-stepdrainage is southern Kattegat: evidence for drainage of the Baltic Ice Lake not known but may have been caused by an ice margin between 12,700 and 10,300 yearsB.P., Boreas, 21, 223-252, 1992. readvance,or icebergsmay have dammedone of the two narrow BergstrOm,B., Stratigraphicalevidence of considerableYounger Dryas glacier advance in southeasternNorway, Nor. Geol. Tidsskr., 75, outlets regulating the drainage. On the basis of 127-136, 1995. biostratigraphy[Knudsen, 1982; Miller, 1982], it is suggested Birks, H.H., S. Gulliksen. H. Haflidason,J. Mangerudand G. Possnert, that the final drainage took place at or shortly after a rapid New radiocarbon dates for the Vedde Ash and the Saksunarvatn Ash from westernNorway, Quat. Res., 45, 119-127, 1996. warming in early Preboreal.The estimatedradiocarbon age for Bj6rck, S., Late Weichselianstratigraphy of Blekinge, SE Sweden,and the event falls into the radiocarbonplateau of 10,000 14C water level changesin the Baltic Ice Lake, Univ. of Lurid, Dept. of years [Kromer and Becker, 1993; Goslar et al., 1995]. Quat. Geol., Thesis7, 248 pp., Lurid, Sweden,1979. Bj/Jrck,S., A review of the historyof the Baltic Sea, 13-8 ka B.P., Quat. Numerous/5•80 oscillationsand steps are interpretedto Int., 27, 19-40. 1995. Bj/3rck, S., and G. Digerfeldt, New •4C dates from Hunneberg reflect glacial meltwaterreleased during periodsof melting of supporting the revised deglaciation chronology of the Middle the southeasternportion of the Fennoscandianice sheet.Rapid Swedish end moraine zone, GFF, 103, 395-404, 1982. melting occurred a few hundred years prior to the Younger Bj/3rck,S., and G. Digerfeldt,Climatic changesat Pleistocene/Holocene boundaryin the Middle Swedishend moraine zone, mainly inferred Dryas standstill, which appears to have begun at from stratigraphicindications, in Climatic Changeson a Yearly to approximately 11,150 •4C years, when salinity, although Millenial Basis,edited by N. A. M/3rnerand W. Kar16n, pp. 17-24, D. Reidel, Norwell, Mass., 1984. oscillating,shows no deglacialtendency toward lower values. Bj6rck, S., and G. Digerfeldt, Lake Mullsj/3n - a key site for During the Younger Dryas, approximately 10 salinity understandingthe final stage of the Baltic Ice Lake east of Mt. fluctuations, unrelated to the drainage of the BIL, are Billingen, Boreas,18, 209-219, 1989. Bj/3rck, S., and N-O. Svensson, Geology, National Atlas of Sweden, recognized. These variations are interpreted as climatically editedby C. Fred6n,208 pp., Stockholm,1994. inducedice margin recession/readvanceevents. An oscillating Cato, I., Field investigations, coring and sampling, in The behavior of the Fennoscandianice sheet during the Younger Pleistocene/Holoceneboundary in south-westernSweden, edited by E. Olausson,Sver. Geol. Unders., Ser. C, 794, 27-33, 1982a. Dryas is also indicated by ice-marginal deposits throughout Cato, I., Grain-size distribution of the cores - with emphasison the Scandinavia. sedimentarypatterns around the Pleistocene/Holoceneboundary, in The Pleistocene/HoloceneBoundary in South-WesternSweden, During the Preboreal, when most of the remaining edited by E. Olausson, Sver. Geol. Unders., Ser. C, 794, 54-65, Fennoscandian ice sheet was rapidly melting, there is no 1982b. Cato,I., C. Fred6nand E. Olausson,Summary of the investigation,in The distinctsign in the/5•80 recordof increasedmeltwater outflow. Pleistocene/HoloceneBoundary in South-WesternSweden, edited by With the exception of the two-step drainage of the BIL E. Olausson,Sver. geol. unders.Ser. C, 794, 253-268, 1982. marking the beginning of Preboreal, the record appearsmore Dansgaard,W., H. B. Clausen, N. S. Gundestrup,C. U. Hammer, S. J. Johnsen,P.M. Kristinsdottirand N. Reeh,A new Greenlanddeep ice stable in contrast to the Younger Dryas. This more stable core, Science, 218, 1273-1277, 1982. appearancemay be connectedto the inflow of seawaterinto the Dansgaard,W., et al., Evidencefor generalinstability of past climate Baltic basin, a diminishingdrainage capacity of the outlet due from a 250-kyr ice-corerecord, Nature, 364, 218-220, 1993. Denton,G. H., and C. H. Hendy, YoungerDryas age advanceof Franz to isostaticuplift, or the developmentof a new drainageoutlet Josef Glacier in the southernalps of New Zealand, Science,264, in the southern Baltic. 1434-1437, 1994. Donner, J. J., A profile acrossFennoscandian of Late Weichselianand Flandrianshore-lines, Commentat. Phys. Math., 36, 1-23., 1969. Acknowledgments. We thank Irena Hajdas for commentson an Donner, J. J., Fluctuations in water level of the Baltic Ice Lake, in earlier draft of the manuscript and for discussionof radiocarbon Studieson the Baltic Shorelinesand SedimentsIndicating Relative chronology,and Urve Miller for discussionof the biostratigraphyfor the Sea-Level Changes, edited by T. Aartolahti and M. Eronen, Ann. cores. We thank Ann-Marie Robertsson and Bj/3rn Malmgren for Acad. Sci. Fenn., Ser. A., Math. Phys.III, 13-26, 1982. varioushelp and for permissionto samplethe Moltemyr and Solberga-2 Edwards, R. L., J. W. Beck, G. S. Burr, J. D. Donahue, J. M. A. cores. This study was supported by the Swedish Natural Science Chappell,A. L. Bloom, E. R. M. Druffel andF. W. Talyor, A large Research Council (Per Bod6n). We acknowledge support for AMS drop in atmospheric14C/•2C and reducedmelting in the Younger analyses received from the Climate Center at Lamont-Doherty Earth Dryas, documentedwith 230Thages of corals,Science, 260, 962-968, Observatoryof ColumbiaUniversity. 1993. Erlenkeuser,H., Stableisotopes in benthicforaminifers of Skagerrak core GIK 15530-4: High resolutionrecord of the YoungerDryas and References the Holocene, Nor. Geol. Tidsskr.,65, 49-57, 1985. Fairbanks, R. G., A 17,000-year glacio-eustaticsea level record: Alley, R. B., et al., Abrupt increasein Greenlandaccumulation at the Influenceof glacial meltingrates on the YoungerDryas eventand end of the YoungerDryas event, Nature, 362, 527-529, 1993. deep-oceancirculation, Nature, 342, 637-642, 1989. Andersen,B. G., J. Lundqvist,and M. Saarnisto,The Younger Dryas Fred6n, C., Geologicalnotes of the Moltemyr, Brastadand Solberga margin of the Scandinavianice sheet - An introduction,Quat. Int., areas, in The Pleistocene/HoloceneBoundary in South-Western 28, 145-146, 1995a. Sweden,edited by E. Olausson,Sver. geol. unders.Ser. C, 794, 34- Andersen, B. G., J. Mangerud, R. S0rensen,A. Reite, H. Sveian, M. 35, 1982. Thoresen,and B. Bergstr/3m, Younger Dryas depositsin Norway, Fred6n,C., Marine life and deglaciationchronology of the VfinernBasin Quat. Int., 28, 47-169, 1995b. southwesternSweden, Sver. Geol. Unders., Ser. Ca, 71, 1-80, 1988. Bard, E., B. Hamelin, R. G. Fairbanks and A. Zindler, Calibration of the Goslar, T., et al., High concentrationof atmospheric14C duringthe 14Ctimescale over the past30,000 yearsusing mass spectrometric U- Younger Dryas cold episode,Nature, 377, 414-417, 1995. Th agesfrom Barbadoscorals, Nature, 345, 405-410, 1990. Grootes, P.M., M. Stuiver, J. W. C. White, S. Johnsenand S. Jouzel, Bard, E., M. Arnold, J. Mangerud,M. Pateme, L. Labeyrie, M.-A. J. Comparisonof oxygen isotoperecords from the GISP2 and GRIP Duprat, M61i•res, E. S0nstergaardand J.-C. Duplessy, The North Greenlandice cores,Nature, 366, 552-554, 1993. BODI•N ET AL.: THE FINAL DRAINAGE OFTHE BALTIC ICE LAKE 49

Guilderson, T. P., R. G. Fairbanks and J. L. Rubenstone,Tropical Polyak,L., S. J. Lehman,V. Gataullinand A. J. T. Jull, Two-step temperature variations since 20,000 years ago: Modulating deglaciationof the souteasternBarents Sea, Geology,23, 567-571, interhemisphericclimate change, Science, 263, 663-665, 1994. 1995. Hughen, K. A., J. T. Overpeck,L. C. Petersonand S. Trumbore,Rapid Rainio,H., M. Saarnisto,and I. Ekman,Younger Dryas end moraines in climate changesin the tropical Atlantic region during the last Finland and NW Russia,Quat. lnt., 28, 179-192, 1995. deglaciation,Nature, 380, 51-54, 1996 Robertsson,A.-M., Pollen,in The Pleistocene/HoloceneBoundary in Jansen,E., and T. Veum, Evidence for two-step deglacitionand its South-WesternSweden, edited by E. Olausson,Sver. Geol. Unders., impact on North Atlantic deep-watercirculation, Nature, 343, 612- Ser. C, 794, 224-242, 1982. 616, 1990. Ruddiman,W.F., C. D. Sancettaand A. Mcintyre,Glacial/interglacial Johansson,S., Baltiskaissj6ns tappning, GFF, 48, 186-263, 1926. responserate of subpolarNorth Atlanticwaters to climaticchange: Johansson, S., Senglaciala och interglaciala avlagringar vid The record in oceanic sediments,Philos. Trans. R. Soc. London,Ser. iindmorfinstrfiketi V•isterg6tland, GFF, 59, 379-454, 1937. B, 280, 119-142, 1977. Johnsen, S.J., D. Dahl-Jensen, W. Dansgaard and N. Gundestrup, Stabell,B., Shell materialin core 15530-4:Its radiocarbonage, Nor. Greenland palaeotemperaturesderived from GRIP bore hole Geol. Tidsskr.,65, 35, 1985. temperatureand ice coreisotope profiles, Tellus, Ser. B, 47, 624-629, Stabell,B., and J. Thiede,Absolute chronology: Summary core GIK 1995. 15530-4, Nor. Geol. Tidsskr.,65, 25, 1985. Kennett,J.P., and L. Ingram, A 20,000-yearrecord of oceancirculation Str6mberg,B., Deglaciationenvid Billingen och Baltiska issj6ns and climate changefrom the SantaBarbara basin, Nature,377,510- tappning,GFF,99, 92-95, 1977. 514, 1995. Str6mberg,B., Revisionof the late glacial Swedishvarve chronology, Kerr, R.A., Ice bubbles confirm big chill, Science, 272, 1584-1585, Boreas, 14, 101-105, 1985. 1996. Str6mberg, B., The final stage of the Baltic Ice Lake, Sver. Geol. Knudsen,K. L., Foraminifers,in The Pleistocene/HoloceneBoundary in Unders. Ser. Ca, 81,347-353, 1992. South-WesternSweden, edited by E. Olausson,Sver. Geol. Unders., Svensson, N.-O., Late Weichselian and early Holocene shore Ser. C, 794, 148-177, 1982. displacementin the centralBaltic, basedon stratigraphicaland Koq Karpuz, N., and E. Jansen,A high-resolutiondiatom recordof the morphologicalrecords from easternSmfiland and Gotland, Sweden, last deglaciation from the SE Norwegian Sea: Documentationof Univ. of Lund, Dept. of Quat. Geol., Thesis25, 195 pp., Lund, rapid climaticchanges, Paleoceanography, 7, 499-520, 1992. Sweden, 1989. Kromer, B., and B. Becker,German oak pine 14Ccalibration, 7,000 BC Thompson,L. G., E. Mosley-Thompson,M. E. Davis, P. N. Lin, K. A. to 9400 BC, Radiocarbon, 35, 125-135, 1993. Henderson, J. Cole-Dai, J. F. Bolzan and K. B. Liu, Late Glacial Lehman, S. J., and L. D. Keigwin, Suddenchanges in North Atlantic stageand Holocenetropical ice core recordsfrom Huascarfin,Peru, circulationduring the last deglaciation,Nature, 356, 757-762, 1992. Science, 269, 46-50, 1995. Longva, O., and M. K. Thoresen, Iceberg scours, iceberg gravity Wohlfarth,B., S. Bj6rck, G. Possnert,G. Lemdhal,L. Brunnberg,J. cratersand currenterosion marks from a giganticPreboreal flood in Ising, S. Olssonand N.-O. Svensson,AMS datingSwedish varved southeasternNorway, Boreas,20, 47-62, 1991. clays of the last glacial/interglacial transition and the Lundqvist, J., The Younger Dryas ice-marginalzone in Sweden,Quat. pontential/difficultiesof calibrating Late Weichselian "absolute" lnt., 28, 171-176, 1995. chronologies,Boreas, 22, 113-128, 1993. Mangerud, J., and S. Gulliksen, Apparentradiocarbon ages of recent Working Group, Proposal by the working-group, in The marine shellsfrom Norway, Spitsbergen,and Arctic Canada,Quat. Pleistocene/HoloceneBoundary in South-WesternSweden, edited by Res., 5, 263-273, 1975. E. Olausson,Sver.Geol. Unders.,Ser. C, 794, 269, 1982. Mangerud, J., S. T. Andersen, B. E. Berglund and J. J. Donner, Quaternarystratigraphy of Norden,a proposalfor terminologyand classification,Boreas, 3, 109-128, 1974. P. Bod6n, Departmentof Geology and Geochemistry,Stockholm Miller, U., Diatoms,in The Pleistocene/HoloceneBoundary in South- University, S-106 91 Stockholm, Sweden. (e-mail: WesternSweden, edited by E. Olausson,Sver. Geol. Unders.,Ser. C, per.boden@ geokem.su. se) 794, 187-210, 1982. L. H. Burckle and R. G. Fairbanks, Lamont-Doherty Earth Munthe, H., Studiesin the Late Quaternaryhistory of southernSweden, Observatory,Columbia University, Palisades,NY 10964. (e-mail: GFF, 32, 1197-1293, 1910. burckle•lamont.1deo.columbia. edu; fairbank•lamont. 1deo. columbia. edu) Olausson,E., Stableisotopes, in The Pleistocene/HoloceneBoundary in J. D. Wright,Sawyer Environmental Research Center, University of South-WesternSweden, edited by E. Olausson,Sver. Geol. Unders., Ser. C, 794, 82-92, 1982. Maine,Orono, ME 04469.(e-mail: [email protected]) Olsson, I. U., Radiocarbon dating, in The Pleistocene/Holocene Boundary in South-WesternSweden, edited by E. Olausson,Sver. (ReceivedApril 24, 1996;revised August 29, 1996; Geol. Unders., Ser. C. 794, 243-252, 1982. acceptedSeptember 10, 1996)