JOURNAL OF QUATERNARY SCTENCE Q006) 21(1) 2s4SCopyright @ 2006 John Wiley & Sons, Lrd.Published online in wiley Interscience (www.interscience.wiley.com). Dol: '1o.1002/jqs.94o

Relative sea-level changes s¡nce 1s 000 cal. yr Bpin the Nanortalik area, southern Greenlandc. f. SPARRENBOM,I* O. BENN|KE,2 S. gjöncrt and K. LAMBECK3' centre for ceoBiosphere science, Quaternary ceology, Lund University, sweden' Ceological Survey of Denmark and Creenland, Copenhagen, Denmark' Research School of Earth Sciences, Australian National University, Canberra, Australia

Sparrenbom, C' J., Bennike, O., Björck, S. and Lambeck, K. 2006. Relative sea-level changes since 15000cal.yrBP in the Nanortalik area, sourhern Creenland.J. Quaternary Sci., Yol. 21 pp. 2948. ISSN 0262-81 29.Received 12 January 2005; Revised 14 April 2OO5; Accepted 29 April 2005

ABSTRACT: We present new results for relative sea-level change for southern Greenland for theinterval from 9000cal.yrBP to the present. Together with earlier work from the same region thisyields a nearly complete record from the time of deglaciation to the present. lsolation and/ãr trans-gression sequences in one lake and five tidal basins have been identified using lithostratigraphic ana-lyses, sedimentary characteristics, magnetic susceptibility, saturated induced *rn"nent rãgnetisation(SIRM), organic and carbonate content, and macrofossil analyses. AMS radiocarbon datinf of macro-fossils and bulk sediment samples provides the timescale. Relative sea level fell rapidly añd reachedpresent-day level at -9300cal.yrBP and continued fall ing unti l at least 9000cal.yrBP. Between8000 and 6000 cal. yr BP sea level reached ¡ts lowest level of around - 10 m below highest astronom-ical tide. At around 5000 cal. yr BP, sea level had reached above 7.8 m below highest astronomicaltide and slowly continued to rise, not reaching present-day sea level until toãay. The isostaticrebound caused rapid isolation of the basins that are seen as distinct isolation contacts in the sedi-ments. In contrast, the late Holocene transgressions are less well defined and occurred over longertime intervals. The late Holocene sea-level rise may be a consequence of isostatic reloadingtyadvancing glaciers and/or an effect of the delayed response to isostatic rebound of the Laurentideice sheet. One consequence of this transgression is that settlements of Palaeo-Eskimo cultures maybe missing in southern Creenland. Copyright @ 2006 John Wiley & Sons, Ltd.KEYWORDS: sea level; isolation basin; Lateglacial; Holocene; Creenland.

Journal of Quaternãry Science

Introduction

Observations of relative sea-level changes through time pro-vide constraints on volumes of past ice sheets as well as theirretreat history, and data from iifferent localities from aroundthe world have been used to constrain the ice volumes of indi-vidual ice sheets as well as to establish the total change in icevolume for the Last Clacial Maxímum and later times (Lambeckand Chappell,2OOl). Creenland is one such ice sheet that hasbeen investigated in this way (Fleming and Lambeck, 2OO4)and one that we address here. Holocene relative sea-levelchanges around Creenland have earlier been observed andreported in a number of investigations from different sites,for example in NE Creenland (Funder, '1978; Björck andPersson, 1981; Björck et al., '1994; Bennike and Weidick,2001), NW Creenland (Kelly er al., 1999) and W Creenland(Rasch, 2000; Rasch andJensen, 1 992; Rasch et al., 1997; Long

*Correspondence to: Charlotte Sparrenbom, Centre for CeoBiosphere Science,Quaternary Ceology, Lund UniversiÇ, Sölvegatan 12, 223 62 Lund, Sweden.E-mail: charlotte.sparrenbom@geol-lu.se

et al., 1999; Long and Robefts, 2002; Bennike, 2OO2; Longet al.,2OO3). But only limited data are available from southernCreenland (Fredskild, 1973; Funder, 1979; Bennike ef a/.,2OO2; Weidick et a1.,2004) and more data are needed to beable to understand the regional postglacial sea-level changesand associated isostatic rebound in this sector.

In this paper we present results from the first of two investi-gations on relative sea-level changes in the southern Creenlandsector. lt is an extension of the shoreline displacement curvepublished by Bennike et al. (2002) from the Nanortalik areain which they show that the local sea level fell during the Late-glacial period and reached present level at - 1 0 000 years agoand remained below this level until the present. In the currentstudy we have addressed the sea-level change for the morerecent period by coring five basins situated below presentspring-tide sea level and one basin above this level.

Other evidence for Late Holocene relative sea-level changesin southern Creenland comes from marine geological andarchaeological studies: (i) a subaerially weathered 'dry crust,found at ca.9.3-1 0m below present sea level (b.s.l.) in acore taken in Narsaq harbour dated to >2800cal.yrBp(Bennike et al., 2OO2; Weidick et al., 2OO4); (ii) an undateddrowned beach at 3-4m b.s.l. situated close to the Norse

JOURNAL OF QUATERNARY SCIENCE

settlement at Brattahlid (Kuijpers et al., '1999); and (iii) Norseand Neo-Eskimo ruins situated close to the present spring-tidewater level (Mathiassen, 1936; Bak, 1969). Furthermore, fromtidal measurements betlveen 1BB3 and 1885, and again from1932to 1934, Cabel-JØrgensen and Egedal (1940) reported thatthe land in southern Creenland was sinking at a rate of3.9 + 0.38 mm per year during the 48 years between the twomeasurements. Evidence for the presence of Palaeo-Eskimo set-tlements is extremely rare in southern Greenland, possiblybecause their areas of habitation have been inundated by therising sea through Holocene time, and new information onregional sea-level change, especially between 4000ca|. yrBPand 250cal.yrBP, may shed light upon where investigationsneed to be focused to find evidence of Inuit ancestors possiblyreaching southern Creen land.

Field area

The Nanortalik area is situated in south Greenland (Fig. 1)ataround 60" N. The landscape is dominated by fjords reachingdepths over 600m and alpine mountains reaching heights of1500m near the coast and up to 2000m further inland, withthe ice-free land area being around 100 km wide. The offshore

. ä q

shelf is relatively narro4 mostly less than 100km. Claciallyabraded, rounded and flattened mountains are common inthe outer archipelago, especially at the mouth of the fjords.The bedrock in the region mainly consists of granites, gneissosegranites, diorites, migmatised meta-sediments and meta-volca-nics (Escher and Watt, 1976).

During winter, the snow cover in the Nanortalik area is deepand constant apart from windswept ridges. 'Storisen', a widezone of drift ice following the East Creenland Current, reachesthe area in late winter (ca. March) and melts off during the sum-mer, normally disappearing from the area around late June.When the cold East Creenland Current meets the warmerlrminger Current it causes cool, cloudy and foggy conditionsduring the summer months with mean temperatures around5-6"C at the outer coast.

Dwarf-shrub heaths with mosses and lichens dominate thevegetation. Empetrum nigrum, Betula glandulosa and Salixglauca constitute major elements in the heaths.

Methods

Fieldwork was carried out during three weeks in August 2001 .A large Zodiac inflatable boat was used to transport coring

FÁ o å

e g ,u h u . o€

% "ü t i

iJ [Þ í.t= ,"Wfå,,rîq#"f,

e 'r/"r)¡

dü :

Figure 1 (A Map of Creenland where the rectangle marks the sector where the investigated sites are located. (B) Map of the southern Creenlandsector with the rectangle defines the field area. (C) Map showing the location of investigated sites (N1zt-N24 investigated in 1999 (Bennike etal.,2002) and N25-N30 investigated in 2001)

Copyright @ 2006 John Wiley & Sons, Ltd. J. Quaternary Sci., Vol. 21(11 2948 (2006l

SEA-LEVEL CHANCES IN SOUTHERN GREENLAND 3 l

equ¡pment and people from a base in Nanortalik to the differ-ent cor¡ng sites. Coring posit¡ons were determined with CpS or,in the case of sites N29 and N30, taken from the sea chart. Thebathymetry of the submarine basins was investigated with anecho sounder to find suitable coring sites. A small speciallydesigned Zodiac, with a funnel hole in the centre, was usedas a coring platform. During coring, the small Zodiac wasusually roped tightly to three points on the shore. Using aRussian corer, multiple overlapping core sediment sequenceswere collected from the different basins. The cores weredescribed and then wrapped for shipment back to the labora-tory in Lund, where the cores are stored at 4.C.

As the aim of the investigation is to establish when sea levelwas situated at a certain level, it is important to take tides intoconsideration. The marine influence in a basin wil l start andstop when the saline oceanic water ceases to cross the sill ofthe basin. The highest astronomical tide is therefore of impor-tance, as that controls when marine waters can enter the basin.Storm events may also be of importance but harder to establishin detail, and will not affect the different sites evenly because ofdifferences in location and exposure. The highest astronomicaltide has been calculated by O. Andersen (KMS) using the globaltidal model AG (Andersen Crenoble, 'l 995.1) for each site andit varies between sites from 184 cm to 1BB cm above mean sealevel. The mean tidal amplitude for the area is ca. 0.25 m and1.5 m at neap and spring tides, respectively (Farvandsvæsenet,2000,2001).

The alt¡tude of the sill of the lake was measured w¡th a d¡gitalaltimeter and a clinometer, while the threshold levels of thetidewater basins were measured using an echo sounder andcoring rods. The accuracy of the sill measurements dependson the method used (Table 1). By noting the time for every mea-surement, tidal corrections were possible. The total uncertaintvat each site depends on: the accuracy ofthe measurement; th;interpreted threshold development; and the uncertainty in thetidal correction. The total uncertainty is then the square root ofthe sum of the variances of the three contributinq factors, i.e.

d ro ta r : f 4 +4+4 ) r / ,where { is the measurement var¡ance , fi, is the variance of thethreshold development and { is the variance of the tidal cor-rection. The total uncertainty of the relative sea-level determi-nation for each site is given in Table 1.

Detailed l ithological descriptions were carried out in thelaboratory before subsampling. Measurements of magnetic sus-ceptibil¡ty and saturation isothermal remanent magnetisation(SIRM), organic and carbonate contents were carried out onclosely spaced samples from all cores. The magnetic suscept-ibility (x) is a measure of the extent to which a sediment canbe magnetised and is expressed as (pm3 kg-l), and dependson the mineral composition of the sediment. High susceptib¡l-ity values may indicate a higher content of magnetic particles,i.e. a sediment containing more minerogenic particles, and lowvalues may indicate the absence of magnetic particles, i.e.

Table 1 Assessment of total uncertainty for each site investigated in 200i

often a more organic sediment. However, this is a simplifica-tion since the magnetic susceptibil¡ty may also be a measureof fhe amount of magnetotactic bacter¡a. The measurementon naturally wet samples is particularly useful to facilitate corecorrelations, but in combination with SIRM it is also a supple-mentary method to recognise isolation events, i.e. changes indepositional environment (Björck, 1979). The susceptibilitymeasurements were made with a Ceofozika Brno KLy-2 sus-ceptib¡l lty bridge on naturally wet samples.

A Redcliff pulse-magnetiser (model 700 BSM) has been usedto induce a field of 1 T to achieve saturation isothermal rema-nent magnetisation, and a Molspin Minispin spinner magnet-ometer was used to measure the remanent magnetisation.SIRM was measured on naturally wet sediment as well as onsediments that were dried at 55'C for 12-24hr. The authigeni-cally formed magnetic mineral gregite (Fe:S¿) forms in anoxicenvironments in fresh, marine or brackish water where sul-phate suppfy is l imited (Walden et al., 1999; Sandgren andSnowball, 2001 ). lt can be expected in isolation sequenceswhen the environment changes from marine aerobic condi-tions to brackish anaerobic conditions.

Analyses of plant and animal macrofossils were used tocheck the ¡nterpretations of environmental changes, particu-larly isolations/transgressions, implied by the other analyses.Samples for macrofossil analyses were wet-sieved using a0.2-mm sieve.

AMS radiocarbon dating was carried out on selected terres-trial or freshwater macrofossils and bulk sediment sampleswhere suitable macrofossils were not found. The radiocarbondating was carried out at the Radiocarbon Dating Laboratoryin Lund and the Poznan Radiocarbon Laboratory. The radiocar-bon dates were calibrated using the software Oxcal v. 3.5(Bronk Ramsey, 2OO1) and v. 3.9 (minor changes in 2003),based on the INTCALSS dataset of Stuiver et al. (i998), andusing the terrestrial calibration curve. The radiocarbon ana-lyses provide ages for the time of isolation and transgressionof each basin.

Results

The six investigated sites are described below. We continue touse site name codes from Bennike et al. (2OO2), where N standsfor Nanortalik. Sediment depths in the descriptions refer todepth below water sur{ace. Altitude of the thresholds arerelated to metres above ( + ) or below ( - ) the highest astro-nomical t ide (m a.h.a.t. and m b.h.a.t., respectively).

Tidal basin N25 (N 60.1896",W 45.0760")

The site is a marine tidal basin with a threshold consisting of acoarse sediment lag (pebbles and rocks) at 6.3 m b.h.a.t. The

Site Measuredelevation (m a.m.s. l .)

Measurementmethod

Measurementuncertainty (m)

Assessment of Tidal Totalthreshold uncertainty (m) uncertainty (m) uncertainty (m)

N25N26N 2 7N2BN29N30

-4.55

-0.45-5.95-1 .45

1 . 2 5

0 .1o.20.10 .10.10.1

+ 0.3+0.' l+ 0.0+1/-2+1/-2+0.3

RodsClínometer

RodsRodsRodsRods

u-50.50 .50.50 .50.5

+0 .6+0.5+ 0.5

+1 .1/-2.1+1 .11-2.1

+0.6

Copyright @ 2006 John W¡ley & Sons, Lrd. J- Quaternary Sci., Vol. 2-l(11 2948 (2006)

JOURNAL OF QUATERNARY SCIENCE

basin is located 10 km northeast of Nanortalik in the Tasermiutfjord and is the most inland site investigated to yield isolation ortransgression results. lt is oval in shape, ca. 1500m long and400m wide, and a series of lakes drain into it via a waterfall.The water depth at the coring site varies with the tides, between9.5 m and 11 m, but in some parts of the basin, depths of up to22m were recorded. At the isolation contact at 1200 cm, thesediments change from greenish grey sandy clayey gyttja silt,into laminated greyish brown sandy clayey algal-rich silt gyttja(Fig. 2). Marine algae, foraminifera and Mytilus edulls also dis-appear above this level and a freshwater fauna with Cladoceraand Chironomidae start to dominate (Fig. 3). The organic car-

N25

bon and carbonate content increase at around 1200cm and asmall peak is visible in all the magnetic parameters (Fig. 2). Thetransgression event is harder to establish from the lithologicalanalyses alone as there are several possibilities indicated bythe sedimentary changes and changes in the magnetic para-meters, as well as by the organic carbon and carbonate contentat 1181 cm, 1163.5cm, 1O7Bcm, 1061 cm, 1Ù44.5cm and1027 cm. But from the macrofossil analyses the transgress¡oncan be established at 1061 cm (Fig. 3). The transgression is amore gradual transition than the isolation event, withfreshwater taxa such as Isoëtes lacustris, Rhabdocoela andCladocera slowly decreasing in numbers while marine algae,

0 1 0SIHMI¿,

go c0 4t

CARBO|\IATEÛ0NTENTI "i.T

* -0 , 5 1 ' 1 . 5 S e ç 0 9 , 5

MINËROGËNIOFË$nLrE {%}

8 4 æ g â g ê f f f iôÈ4r

1000

f¡t

(t{l)îtF{D

t,

:O-{l¡Ê

'fo$q

:l'!0Ð11, t rt120:t1êtl t4û1150I'lÉ0T17 'o11È0

tT{lo1 J 1 01 Ieû11At1 t40I150,'1I6,0;

11701 180,1ftöæ00f91ûfi241e3û124û1å50

lsolalion- 11s

--:1e0o

121ût22ô1290124ft12Só

0 " 1 2 û . 1 6 0 ^ S 0 , 2 4 0 2 4 € 8 1 0 0 4 8 1 2MAGNETIû $lBM(mAm 2 ks-1 ¡ 9åF,99lt- *SUSÊEPTIBILITY (Ð - DryS¡RM trLrlrrrtrrrr ['?cJ(X1$6 m3 Xc'l ) æ Wát Sin¡¡

LegendI silty_and sendysilfy ...gFga clay or o|ayey s|n

sandy clayey algae-rich silt gyttia,laminated wilh grey/black FeS-bands

E clayeysilty sard gynia

ffîl clayey Êiltl¿ gyttja mnd

[:f] sandy c{ayey algae"rich eilt gyttja or: clayey silly algae-rich sand gyttjal-Jj sandy silty algae-rich clay gyttja,

laminated wiih grey/black FeS-bands

ffi sandy clayey gyttþ silr

Figure 2 Core log for site N25, the magnetic susceptibility, SlRM/¡ (scale bar on top), SIRM wet and dry, carbon content (massTo of dry weight),carbonate contents (massTo of dry weight, scale bar on top) and minerogenic residue (massTo of dry weight). Ages given are calibrated years BP (cal. yrBP) or interpreted interpolated ages between several dates, all in cal. yr BP

i\

ïransgre*sion at

4 4? - È

r - <7- aF'|-li

II:"

[ ,\ \f t\It '

F .

\4 *

i i

Copyright @ 2006 John Wiley & Sons, Lrd. J. Quaternary Sci., Vol. 21(-ll 2948 (2006)

ô!

=.æoN

o

7f

ãoPc

f

r

25 -a- :;"^1"""'" 7 /Lacustríne-':7

.r*iffi ';î$.ç".1f '5*91,.;|ffit::-^

_--=.='='Þ .-L þþþ þþþ rrana - ¡ - =¡

1 060

1070

1 080

1 090

_. : 1100U)3 1 1 1 0oõ.o 1120E_c 1 130o_E i14o

1 1 5 0

1 1 6 0

1170

1 180

' 1 1 9 0

1 200

-cÐo=Þ

.9o-

N@L

þO

c

counts per 30 ml sample (Note the different scales for the different taxa)

c. 6360 cal yr BP mP-mfi-ôTzoñ7oC--lTñzômmz-z0

c. 18700 cal yr BP

c. 9500 cal BPlsolationc. 10000 cat Bp

34 JOURNAL OF QUATERNARY SCIENCE

Foraminifera and Mytilus edulis increase in numbers at thesame time. The small number of poorly preserved marinemacrofossils appearing in the interval between 1200cm and1063 cm can possibly be explained by wave-induced erosionand re-sedimentation of older sediments. The presence ofreworked material is supported by the old radiocarbon age of15690+90 14C yr BP for a bulk sed¡ment sample f rom11.635 cm (Poz-7175, Table 2). Alternatively, the occurrenceof marine macrofossils in small numbers during periods clearlydominated by freshwater species may be a result of stormevents that transported marine waters and organisms over therelatively narrow and low threshold which is located towardsthe south-southwest and is exposed to southwesterly storms.The age of the isolation of the basin is determined to around9500cal.yrBP based on a bulk sediment sample (9470-9630 cal. yr BP) just above the isolation contact, a macrofossilsample of Nitella sp. (9020-9490cal.yrBP) at and abovethe isolation contact and a sample of the moss DrepanocladuscÍ. badius (9750-l}49ocal.yrBP) just below the contact(Table 2). The time of the transgression is dated to around25OO cal.yr BP based on a sample of the following taxa; Bryop-sida indet., Nitellasp., Empetrum sp., /soëtes lacustris, Daph-nia pulex, Batrachium confervoides and chironomids (2340-274O and 2150-2250cal.vrBP) at 1O57-106'l cm ffable 2).

Table 2 AMS radiocarbon and calibrated dates for sites N25-N30

Lake N26 (N 60.1 278" , W 45.0621")

The lake is located ca. 9 km east of Nanortalik close to themountain Jakobinerhuen at the mouth of Tasermiut fjord. Thethreshold consists of rock at an altitude of ca.3.2m a.h.a.t.With an irregular shape, the basin measures ca. 150 x 500 m.There is no obvious inlet to the lake, so inflow of material islimited to that brought by surface runoff. The water depth atthe coring site is 5.5 m, but depths of up to 10 m were recordedin the basin. Sedimentary changes from brownish grey siltysand to dark brown and black FeS-laminated sandy clayeyalgal-rich silt gyttja overlain by clayey silt gyttja can beobserved around the isolation (Fig. a). Marine algae, Hydroideaindet. and Mytilus edulrs disappear at the isolation and a fresh-water fauna with the cladocerans Daphnia sp. and Ceriodaph-nia sp. and chironomids appear and increase in abundance(Fig. 5). Organic carbon and carbonate content increase atthe same level (Fig. 4). SIRM and susceptibil i ty also increaseduring the transition from the marine to the freshwater environ-ment at 920 to 915 cm. The age of the isolation is established toca. 10 300 cal. vr BP on the basis of three bulk sediment sam-ples (one at 916-91Bcm with the10650cal .yrBP, one at 9 '14-916cm10600cal .yrBP and one at 910-914cm

age of 9900-dated ro 9900-

with an age of

Site Laboratory Depthno. (cm)

Age (taC yr BP) Calibrated age2o (95.40/")

Material Comments

N25 Poz-7174N25. LUA-5846

N25 Poz-732'lN25 Poz-7175N25 Poz-7320

N25 LuA-5848N25 Poz-74'15N25 LUA-5847N26 LUA-5225N26 LUA-5226N26 LUA-5227N27 Poz-7172N27 LUA-5349N27 Poz-7173

N27 LuA-5348N2B LuA-5351N2B Poz-7178N28 LuA-5350N29 LUA-5849

N29 Poz-7'176

N29 LuA-531 6

N29 LuA-531 5N30 LuA-531 3

10.23510.57-10.61

10.765r 1 . 6 3 51 1 . 9 5

1 1 .98-12.0012.O2

"t2.o1-"t2.o39.10-9. ' t49."t4-9. ' t69.1 6-9.1 B

6.445-6.4556.68-6.6856.94s-6.95

7.58-7.5913.275-13.28513.39- i 3 .40

13.419-13.42711.77-11.855

11.79-11.80

11.86

1't .971 . 0 9

i 2 8 0 + 3 02450+85

5580 + 401 5690 + 908560 + 50

8270X.1008720L508990 + 90B B 4 5 + 1 1 59 1 1 0 + 1 1 09 1 3 0 f 1 0 51 5 5 0 + 3 0705 + 85

2 8 1 0 + 3 0

8 4 1 0 + 1 1 54375+958250 É 508 3 1 0 + 9 52395 + B5

2290+30

7300 + 95

8085 + 95945 + 90

1725+.901 0 2 0 + 8 08350 + 809325 È 1 30

1140-12902340-2740 (92.8%)

6290-64501 81 50-1 9450

9680-9640 (2.6%)9630-9470 92.8'/.)

902c-94809550-991 0

9750-1 04009550-1 02509900-1 06009900-1 06501 350-1 530520-790

2840-3000 (92.0./.)2790-2830 (3.40/.)

9000-96004650-53509030-94309030-9500

2300-2750 (93.1%)2"tso-2250 (2.3%)

2300-2360 (s9.8%)2',tBj-2250 (33.6%)2150-2170 (2.'t%)

7940-8340

8600-93001030-1060 (1 .9%)670-1000 (e3.s%)

1 41 0-1 830730-1 090

91 30-95301 0 2 0 0 - 1 1 1 0 0

Transgression

N25N25

lsolat ion

lsolationIsolationlsolationIsolationlsolationlsolation

Above transgressionTransgression

Below transgression

lsolationTransgression

lsolationlsolation

Transgression

Transgression

Transgression(below hiatus)

lsolationTransgression

TransgressionTransgression

lsolation

N30 LuA-5352 1.09-1.095N30 LuA-5312 1 .13N30 LuA-5311 1 .39N30 LuA-5314 1 .9

CyttjaBryopsida indet., Nite//a

Empetrum, Isoëtes, Daphiia,B atrachi um and Chi ronomidae

CYttJaCYttjuCyttja

Nitel laCyttja silt

Drepanocladus cÍ. badiusCyttjaGyttjaCynjaCYttjaGYttjaCYttju

GYttjaCYttjaGynjuGYnja

Bryopsida indet., Nite//a,Empetrum, lsoëtes, Daphnia,

Juniperus, Chironomidae adu ltCYttju

Silty gyttja

Silty gyttjaJ uni perus com mu n i s, Empetrum

GYttjaJ u n i perus com m u n i s, EmpetrumEmpetrum and Hippuris vulgaris

Cyttjasilt

Copyright @ 2006 John Wiley & Sons, Ltd. J- Quaternary Sci., Vol. 21(11 2948 (2006l

SEA-LEVEL CHANCES IN SOUTHERN CREENLAND 35

N26ñ t

SIRM/¿4 6 8 1 û 1 2

TAEBO|\IATE - -coNTENr{"/o}t l 2 3 4

I

> t60@

Ñ.ø(/l

Ë * tEË="eorâ.@Ê

92t

l\I

l€ôlatiún atô. t0300 râl

0.04 0.0É Õ-r2 0'16 0.2 0_24 tMAGNETICsuscEPTlBtL|TY (g{xl0'6 ¡¡3 ftg:l)

Legend,silty gravelly sand -silty sandy gravel

o5 1.r 1.5 2.O 2,5 t 10 2t 30 tÐ 70 I,O 90 ìt0SlHlul {mAmz tçg:l } CAFBO¡¡ MTNEHOGENICe - ¡ Dry $lffM COHTENT {%) HËS¡DUE {o/4

Wei SIRM

F=r-laminated silty clayeyalgae gyttiäalgae gyttjâ crntaining variousproportions of silt ãnd clay

depths of up to 15 m were recorded. Today the environment inthe basin changes twice a day because of the tides, from beingpart of the sea to becoming an isolated water body.

Bef ow the isolation found at ca.757-759 cm, the sedimentsconsist of greyish gyttja with gravel and sand, changing intolaminated brown gyttja containing black iron sulphide-richlaminae (Fig. 6). Above the isolation contacr, laminated g¡tjacontinues another ca. 5 cm, with colours that varv from reddishto more greyish brown. Marine brown algae, Mytitus edutisand Tricladida indet. disappear above the isolation and afreshwater fauna with Cladocera especially Daphnia sp.,Chironomidae and Plumatella sp. starts to dominate (Fig. 7).The organic carbon and carbonate content increase below

sandy,Çlalt€y.silt ii:sandy clayey algae-rich gyttja,laminated with FeS-bandssilty sand

Figure 4 Core log for site N26, the magnetic susceptibility, SIRM/x (scale bar on top). S|RM wet and dry, carbon contents (mass% of dry weight).carbonate contents (masso/" of dry weight, scale bar on top) and minerogenic residue (massTo of dry weight). Ages given are cal. yr Bp or interpretedinterpolated ages between several dates

Iffi[ilffi

9550-1 0250cal.yrBP, Table 2) at the transition between thetwo sedimentation environments.

T ida l bas in N27 (N 60.1321", W 45.0513' )

The site is a tidal basin located ca. 9 km east of Nanortalik closeto site N26 and the mountain Jakobinerhuen at the mouth of theTasermiut fjord. Bare bedrock constitutes the threshold at analtitude of 2.3 m b.h.a.t. A series of lakes drain into this oval-shaped basin. The basin measures ca. 1250 x 160 m. The waterdepth at the coring site is ca. 6 m, but in some parts of the basin,Copyright @ 2006 John Wiley & Sons, Ltd. J, Quaternary Sci., Vol. 2-l(1) 2948 (2006)

JOURNAL OF QUATERNARY SCIENCE

L-L- L--_-l- t--_l_J

lsolation atc. 10300 cal. yr BP

0 5 0 i 0 2 0 0 1 0 0 0 5 0 0 r 0 0 2 4 0 s 0 1 0 0 0 s 0 i 0 0 0 2 5 s 0 0 2 5 o 2 0 2 0 0 0 0 5 0 1 0 0 0 2 5 0 5 0 0 0 5 0 0 0 5 0 1 0 0 0 1 0 0 0

Counts per 30 ml sample (Note the different scales for the different taxa)

Figure 5 Simplified macrofossil diagram showing the results from the macrofossil analyses of site N26. Ages given are cal. yr BP or interpreted inter-polated ages between several dates, a l f in cal . yr BP. Notethedi f ferentscalesfor thedi f ferent taxa

9r3

; s14o:I e1563E e16oo

_o 9f7E()

r 9 1 8

O 919

920

the isolation contact, probably as a consequence of the thresh-old reaching above the low tide level and the basin becoming'semi-isolated', i.e. isolated during low tides twice a day result-ing in brackish water conditions. Both the organic and the car-bonate content are rather high during the isolation phasecomoared to the marine environment before the isolation.The magnetic susceptibil i ty curve shows a small rise at the iso-lation and decreases above the contact at ca. 750cm (Fig. 6).The magnetic susceptibility varies as a consequence of the con-centrat¡on of magnetic particles, which is mostly controlled bythe variation of the minerogenic content. Figure 6 shows thatsusceptibility values increase and decrease when the minero-genic residue increases and decreases, respectively.

The sedimentary transition from a freshwater environmentinto a marine environment at the transgression (level 669cm)is reflected by a change from dark brown algae gyttja into abrown, green, and grey laminated algae gyttja with markedblack iron sulphide lamination (Fig. 6). Allthe magnetic para-meters measured show a marked decrease at this level. Carbonand carbonate content show large fluctuations, probably as aresult of the constantly changing environment due to saltwaterinflow twice a day. Foraminifera, Hydroidea and some brownalgae enter the basin while the freshwater taxa Cladocera, Chir-onomidae and /soëtes lacustris slowly decrease in numbers(Fig.7). The age of the isolation of the basin is established toaround ca. 9300 cal. yr BP (Table 2) based on a bulk sedimentsamole at the isolation contact at 758-759cm (9000-9600cal.yrBP). The time of the transgression is determinedto around 2'l 00 cal. vr BP (at level 668 cm) on the basis of aninterpolation between two bulk sediment samples, one fromabove the transgression at 645 cm (1 350-1 530 cal. yr BP) andthe other taken below the transgression at 694.5-695cm(2840-3000 and 2790-2830cal.yrBP) (Table 2). The uncer-tainty of this transgression date is hard to estimate, especiallyconsidering any marine reservoir effect, but it certainly

Copyright @ 2006 John Wiley & Sons, Ltd.

amounts to at least a few hundred years. The error bars for thisinterpolated age have been set to -750 and + 900, which arethe minimum and maximum departures of the measured agesof the bulk samples from the interpolated mean. A third bulksediment sample from level 668-668.5 cm gave an age of52O-790cal.yrBP. This date is l ikely to be too young, as thesample above this level is older and the sample below wasdated to ca. 2900ca|. yrBP.

Mar ine embayment N2B (N 63.0316' ,w 44.9083' ) '

This marine embayment is located ca. 23 km southeast ofNanortalik at Narlusoq island and measures ca. 250 x 125 m.Bedrock or boulders form the threshold at 7.Bm b.h.a.t. andthe maximum depth measured in the basin is 2O m. The waterdepth at the coring sites varied between ca. 8.3 m and 1 2 m as aconsequence oftidal changes during the day and shift ing posi-tions of the coring platform but the overlapping cores could besuccessfully correlated visually as well as with the magneticscans. The isolation contact is olaced at 1340.5-1 341 .5 cm atthe transition from dark brown to black FeS-laminated silt gynjawlth a high content of minerogenic material to lighter brownsandy clay algal rich silt gyttja (Fig. B). Marine brown algae,Foraminifera, Tricladida and Hydroidea disappear above thislevel and a freshwater fauna with Cladocera, Chironomidaeand Plumatella sp. starts to dominate (Fig. 9).

The organic carbon and carbonate contents first increase justabove 1345cm and continue to increase unti l level 1332cmwhere they both suddenly decrease. The magnetic parameters(Fig. 8) show high peaks at the isolation contact at 1 341 cm andthen fluctuate highly until 1326cm where they reach the samehigh values as before the isolation. The mid- to late Holocene

J. Quaternary Sci., Vol. 2-l(1\ 2948 (2006)

SEA-LEVEL CHANCES IN SOUTHERN CREENLAND

slnM/0¿)r0 ?0 ô0 40

CARBONATEcoflrENr {%)

0 9 4N27 - ;

õ 6ÊûÞg<c 70ûrl)ft)F 7P0t]-Ê T4ÙEil zroËõ: 7EoÉ

It

1f

\t

II

,l

I1ï

J

8@

0.05 Ð.15 a.Æ t_ MAGNETIç

SUSCFP-rlBlflTY (¡¿){ x10€m3¡ga ¡

Legend

4 8 0SINM WETAND DFY'{mAm2¡g-I1

; - * D r v

ü

ffiffi

Figure 6 Core log for site N27, the magnetic susceptibility, SIRM/¡ (scale bar on top), SIRM wet and dry, carbon contents (massyo of dry weight),carbonate contents (masso/o of dry weight, scale bar on top) and minerogenic residue (massTo of dry weight). Ages given are cal. yr BP or interpretedinterpolated ages between several dates, all in cal. yr BP

transgression in south Greenland, from a freshwater environ-ment into full marine conditions, has usuallv been found tobe a gradual process, but in the cores from this site the litholo-gic change is very sharp. The freshwater sequence is only

Copyright @ 2006 john Wiley & Sons, Ltd.

8 '16 24 70 gÉ 90CARFON-** MINIEROGENIGÛONTENT (%l RES|ÐUË (%)

$andy silty clayeydetritus gyttjãGffia wittr vârylng mnt6ntsof sand, silt and:elaySandy silty õlay

ca. 6 cm thick, and from the radiocarbon analyses we concludethat the sequence is incomplete and that erosion has removedsignificant parts of it. The first incursion of marine water into thebasin is seen at 1334cm on the basis of macrofossil analvses

100

Algae Syttjã with variousctRtÉnts of sand, silt and clayFeS larninated gyftja withvår¡ous contents of sand,silt änd cläyEnricliment horizonof sand and grêvelin gyttja

l. Quaternary Sci., Vol. 21(11 2948 (2006)

JOURNAL OF QUATERNARY SCIENCE

\ , -"uI ***

l *= \t \|

*"*--

# voô'è''

'=

E *i a . "%%

P""&\ " %_%

ä I I ' p F p(srxe ueìolq eql eloN) 'l 's ^ olêq ulc q¡deq

(Fig. 9). Foraminifera, Pectinaria sp. and some marine brownalgae enter the basin and freshwater species such as Cladocera,Chironomidae and Plumatella so. start a slow decline.

The isolation of the basin is dated to ca. 9250cal.yrBP(Table 2) based on two bulk sediment samples (one above theisolation at 1339-1 340cm dated to 9030-9430cal.yrBP andone below the isolation ar 1341 .9-1342.2cm dated to 903G-9500 cal. yr BP). The time of the transgression is determined tohave occurred before 5000cal.yrBP based on a marine bulksed i ment sample taken at 1 327 .5-1 328.5 cm (Table 2). Th is datepresents a minimum age of the transgression event as it is takenabove the hiatus in marine sediments. The transgression musttherefore have started at this time or earlier.

T ida l bas in N29 (N 60.0300' , W 44.9303' )

The threshold of this ca. 300 x 250m marine embayment ismade up of boulders located 3.3 m b.h.a.t. lt is located ca.23 km southeast of Nanortalik by Narlusoq island and the max-imum depth measured in the basin is 22m. As a consequenceof tidal changes during the day and horizontal movements ofthe coring platform between each coring, water depth variedbetween ca. 1 1.6 and 12.8 m. The isolation contact is olacedaI 1197.5 cm, at the transition from light greyish brown sandyclayey silty gyttja to brown, black and grey laminated sandyclayey silty gyttja w¡th FeS laminae (Fig. 10). At this level, mar-ine algae, foraminifera, Tricladida and Hydroidea disappearand freshwater organisms w¡th Cladocera, Chironomidae andNlte//a sp. start to dominate (FiB. 1 1). The organic carbon con-tent first increases just below the isolation contact at ca.1203cm and continues a general increase unti l ca. 1194cmwhere it suddenly decreases. The carbonate curve has roughlythe same shape as the carbon curve, except between 1200-1207 cm where there is a high peak in the carbonate curve cor-responding to a high concentration of shell fragments (Fig. 10).The magnetic parameters (Fig. 10) decrease from around theisolation with a high peak just above the onset of marine con-ditions at 1 185.5 cm. The lithology at the assumed lacustrinèmarine transition shows a sharp erosive contact and the brown,black and grey laminated sandy clayey silty gyttja containingFeS-(bands) laminations changes abruptly into a brownish greysandy silt gyttja. Two toc-dated bulk samples confirm the ero-sive contact; the sample below the contact at 1 186 cm has anage of ca. 8140cal.yrBP and the sample above was dated toca.2300cal.yrBP (Table 2). The macrofossil analyses show afairly abrupt change from freshwater taxa to marine taxa/ com-pared to most of the other sites; from an environment domi-nated by Cladocera, Chironomidae and Nife//a sp. to anenvironment dominated by marine algae, foraminifera andPectinaria sp. (Fig. 11).

The age of the isolation of the basin is established as ca.9000 cal. yr BP (Table 2) on the basis of a bulk sediment sampleat the isofation contact at 1197 cm (8600-9300 cal. yr BP). Thetransgression is determined to have occurred sometimebetween B140cal .yrBP and 2300cal .yrBP on the basis ofthe bulk sediment sample taken from both above the hiatusat 1 179-1 1 B0 cm and below the hiatus at 1 1 86 cm (Table 2).

T ida l bas in N30 (N 60.0408",W 44.9033' )

This tidal basin is also located ca. 23 km southeast of Nanorta-lik north east of the island of Narlusoq and measures ca.500 x 200 m. The threshold, situated at an altitude of ca.

o. 9õo

q)

õZo-

Jõc:õôoo

Gq)

o

ooìo

3

0)b¡õ

70)dõo

.goo

o.;o-cô

JõoõC)'õo

om

ñNzo.=

a).>(d

d

õEo

Eg

q)

o

þo

}-

ñboõõ=

b x: P

H C )

c -ì = ¿x €O ø

:a^ ;

6

(\Jxo ( ú

co Q

- - õ(1)

ìó c)- - ( Ú.J)

= c )(l)

o õ

<)o q )

^ zì á v

(l)

È^ =R o

F Ao U ,

c

- O

ll

l

f-C\Iz.

7.s,

Þ\

þçooo

| "w '-u

I o-u*

E a . ZE %..'-õ

''oo-

(l) "è¿.c ' v

î

oYeo¿¿oø

| - - ç

\ " %4".

Copyright @ 2006 John Wiley & Sons, Ltd. J. Quaternary Sci., Vol. 21(1) 2948 (2006)

SEA-LEVEL CHANGES IN SOUTHERN CREENLAND

N28slFlM, - -(rnAm2kg'1)

o o.8 1.ۂARBONATEcoNrENT{%}

C I 4 Ð

130ü131013m1330134CI13501360f370:f380139014ffi

1470148t1490r5ùtI51t

I

I

130û131CI13201330134013501360r37013t01390140014101420143014Æ14501460I4701480149t15001510

1410tr4æ1,4ffi144û14501460

or before

c. 9250 cal. yr BP 6=tIII, \It

. ¿ - \r'-r*:TryT*-T!-"*"7

0.08 0.12 0.16 0.Êo 4 I 12 ,t60 ffi 4050 7Ð g0MAGNETIC r SIRM/i- CARBONSUSCEPTIBILITY (¡) CONTËNT(%I ffEglDUË (%){xlt-6m3tcg-1¡

Legend

f Gffia sand

ffiffiWNll¿í'4

ffi

Grey elayey såndy siltSilty gravelly andclayey silty sand

Sandy clayey, gyttja sífrtlayey to sand¡¡ silt gyttja

Sandy and silty algae gytt¡a

Fe$-laminated sandy clayeyalgaerich silt gyttja

Figure B Core log for site N2B, the magnetic susceptibility. dry SIRM/¡ (scale bar on top), dry SIRM, carbon contents (mass% of dry weight), carbo-nate contents (mass7" of dry weíght, scale bar on top) and minerogenic residue (massTo of dry weight). Ages given are cal. yr BP or interpieted inter-polated ages between several dates

0.6 m b.h.a.t., is made up of boulders and the maximum depthin the rather shallow basin was measured to around 2.6 m. Thewater depth at the coring site was measured to ca. 0.8 m and asthe coring was done during low tide, no variations in depth atthe coring site were registered. The isolation contact is placedat 139.5 cm at the transition from greenish brown sandy clayeysilty gyttja to brown sandy silry algal rich gynja (Fig. 't2). Belowthe isolation contact Tricladida indet. dominates the environ-ment, but disappear above this level and a freshwater flora

Copyright @ 2006 John Wiley & Sons, Ltd.

and fauna w¡th Cladocera, Chironomidae and lsoëtes lacustristakes over (Fig. 13).

Both the organic carbon and carbonate content increaseat the isolation contact level. The magnetic susceptibil i ty(Fig. 12) decreases at the isolation contact while the SIRMmeasurements show litt le change. The transgression, with afirst marine incursion seen from macrofossil analyses at112cm, was gradual and the sediments change from darkbrown clayey silty sandy gyttja into grey clayey silry sandy

f . Quaternary Sci., Vol. 21 (1) 2948 (2006)

a)

=.t@N

o

Ía

Io

F

-

N28,-

lvlarineorganisms

."0*' .al. ^Ql. "rt"* øQ, ô, ,-uP

-""'C or*'*t* ^,r"t"*' --a".

..a- "."a- "*.&N -oøQ' ,oøQ' **o* *uuuq' .on*' ^."ttt* ."a- .-"t"- ,-t-

-'ØÈõEo

ã 1ss5ê

,ooð' ^øQ,

ÞoaÞa.

-N

NetN

o

7 ,-""racustrineoro;isms7d ^oo -.---"ttt"C .".ð -"'"-

.."C .'o'-'oN .r--.

À

oCz-ôT

c--tmz

ñzC)m

1 2 0Í-rl-rTl [-rTTrn0 3200 0 200 400

SEA-LEVEL CHANCES IN SOUTHERN CREENLAND

StRM/¡1.6 2.4 s-2

CÁf;IBONATETONTENT{"L}

0 2 4 6

TI

B

I

\ttI

Iq

-CI

TJ

GLí)Ê

2300 ÊaÍ BF

! C -\-rt

\ IIIII

0.08 0.1 0.12 0.1 0.2 0,3 0.4 0-,MAGNETIC SIHM

suscEPTtEtLtTY WET & DRY{xlûsrn3¡g-1¡ (m4mÊ ks-l1

DHY SIRMWÊT SIRM

Legend

5 10 15 ffi ?570 8û s0 100CAFIBTN M]NEROGENICCONTENTi"Io) RËSIDUE (o/o)

sandy clayey silt gyttja,larninated with Fes-bands

V....77V elayey sândy silt gyttia7:-À*l and sandy clayey sitt gyftiaffi sandy silt gyttja

gyttja with visible shell fragments. The macrofossil analysesshow that Pectinaria sp. and Macoma balthica enter the basinat 1 12 cm while at the same time Cladocera, Chironomideaand /soëfes lacustris decrease in numbers (Fig. 13). Thefreshwater flora and fauna almost totally disappear at ca.102- ' lO4cm.

The age of the isolation of the basin is established to ca.9300cal .yrBP (9130-9530cal .yrBp) (Table 2) based on amacrofossil sample containing Empetrum nigrum and Hippurisvulgaris from 1 39 cm depth. The time of transgression is deter-mined to ca. 900 cal. yr BP from a macrofossil sample, contain-Copyright @ 2006 John Wiley & Sons, Ltd.

| _j rclayey sílt gyttjå

ing luniperus communis and Empetrum nigrum, taken at1 13 cm, just below the first marine incursion (Table 2).

Discussion

Basin isolation

In the Nanortalik area, early Holocene isolation sequencesshow abrupt lithological changes from coarser minerogenicsediments deposited in a (glacioJ marine environment forming

N29I120

1130

't140

'r150

'11€t

1170

i180

1190

12t0

12I0

r?Jãg

I230

724Ð

Ë.'-H ctayey siltyalgae-rich ffiaFigure 10 Core log for site N29, the magnetic susceptibility, SIRM/x (scale bar on top), SIRM wet and dry, carbon contents (mass% of dry weight),carbonate contents (masso/" of dry weight, scale bar on top) and minerogenic residue (massTo of dry weighi). Ages given are cal. yr Bp or interpretedinterpolated ages between several dates

¡ 1120¡

11S0

r140

1150

1160

'1170

1160

1190

1200

1Ê1t

1220

1230

1240

/ 8l4t cal. yrBP

J. Quaternary Sci., Vol. 21(11 2948 (2006)

Õ!

f .@

oñôîf

fõ'ee

l-Ø-

ôÞc.o-N

N@tñÕOo

Figure 11 Simplified macrofossil diagram showingthe results from the macrofossil analyses of site N29. Ages given are cal. yr BP or interpreted interpolated ages betweendifferent taxa

-)Cz-ôT

c-ñz

Øamze)m

Counts per 30 ml sample(Note the different scalesfor the ditferent taxa)

several dates. Note the different scales for the

SEA-LEVEL CHANCES IN SOUTHERN CREENLAND 15

NBüItu1t01 1 012t13t14t15016017û18t1gû200210220ä3t?4CIÊ50ã602702802903t03 f t32CI330349350360370

3t * *SIRM/7¿t 4 8l ¡ t , t , t . , t , i

CéRBONATEcoNlËNTt%)r , r ¡ [ , 1r t 6û

9tr00

Transgressicn at '

93ût cal. yr

--r-T-

o,ffi 0.16 û 0.8-"_ ßTIAGNET|C $tRF/i€USTEPTIEILITY WET & DRY(x tg-6 m3 tg-1 ) (mAmZ ¡ag-1¡- - w

Legend

;1g. :;1 gyqa with varytng cantentsÌ: ú sand, siltandday[]l ntu*ly sandy ffia silt

laminated sediment, often with black iron sulphide-rich lami-nat¡ons deposited in a brackish water environment, and oftenshifting abruptly into a freshwater deposited brownish algaegyttja. The macrofossil analyses show that pre-isolation sedi-ments are dominated by marine brown algae such as Sphace-lariasp., Rhizoclonium sp. and Desmarestia sp., Foraminifera,the blue mussel Mytilus edulis, hydrozoans of the order Hydro-Copyright @ 2006 John Wiley & Sons, Ltd.

1 1 012Ar3ü

J

\+

*14015t't60170r8019020021û

;vi"ciEg

1ì-{¡)o

\t ll jtr't lllr

t'f it{

f {\ ' l_I -.=*

Ìlìì,tir Í4¡i,|ll

23û24û250

]

f . 6 0 ,16 2478 8û g0 lffiCARBON "-* MINERTGENIC -*

coNTENrri%) RESfDUE t%)

idea and flatworms of the order Tricladida. Even tiny bones ofthe small fish Casferosteus aculeatushave been found in one ofthe basins below the isolation contact. In connection with iso-lation and transgression sequences, the oribatid mite Amerono-thrus lineatus appears in the transitional sediments, possiblyfavoured by the changes between marine and lacustrine condi-tions. This oribatid lives at or near the sea shore (Hammer,

I*I,ft

Þ' l' t14,

'Ft ; -i { _( d

\ (

{€l' {,¡;ì

H1\ l ,( Lì ì

:,Þ

ffi cytr¡"

algae-rich gffi a ivitr variousccrrtenfs of sard, silt and clay

gyttja sílt with variouswntents Õf sänd and claysiftysandyffiaæntain¡ng gravel

laminated ffij a with imrylngnts of sard, sift ard dãy

Figure12 CorelogforsiteN3o,themagneticsusceptibi l i ty,SlRM/¡(scalebarontop),SlRMwetanddry,carbonconrents(massyoofdryweight),carbonate contents (massTo of dry weight, scale bar on top) and minerogenic residue (massTo of dry weight). Ages given are cal. yr Bp or interprãtedinterpolated ages between several dates

J. Quaternary Sci., Vol. 21(1t 2948 (2006l

aloE

=.oN

o

=:Iôe"

-Ø-

N30

,"nd ^-,.'"**9095

1 0 01 0 5

1 1 01 1 51201251 3 0

1 3 5140145

1 5 0

1 6 01 6 51701751 8 01 8 51 9 0

1 9 5

Marine organisms

=

o"**t* *u"".- --.-'"-

J

Ø=o

E

i1-o

ÐõÐ! .

N

NaI

ñ6

øQ,..-""'1$

r----r----t--T I----r----T f---r---l---r- [------r------l r----r------l t-------r------r-

Lacustrine organisms

'"t"'.øp

d.t" .-*"".-d,*".-'.""*"..oC

z-oT

C--lmz

am7(^)m

r--------r------- r------f----

SEA.LEVEL CHANCES IN SOUTHERN CREENLAND 45

1944) and has been reported from isolation sequences beforeby, for example, Bennike (1992, lggï) and Bennike ef a/.(2002). The presence of marine organisms often ends abruptlyat the isolation contact, to be replaced quickly by freshwatertaxa. Most of the freshwater taxa probably arrive at the isolatedlake by surface runoff from small rivers, draining lakes furtherinland, but some species may also be transported by birds. Themost commonly found freshwater organisms are cladoceranslike Daphnia sp. and Alona sp., Chironomids, lsoëtes lacustris,Nitella sp., Plumatella repens, Rhabdocoela flatworms, thewater-crowfoot Batrachium confervoides, larval cases fromTrichoptera, some Hippuris vulgaris and Lepidurus arcticus.Occasionaf bones from the fish Gasterosteus aculeatus havealso been found above the isolation contact from site N26.

Basin ingression

The transition from freshwater sediments into marine sedimentsis often not as clearly visible as the isolation contacts. This is aconsequence of the gradient of the sea-level change being highduring the isolation and lower during the transgression. lt isusually obvious in the sediments that a marine ingressionoccLrrs, but in most cases macrofossil analyses were necessaryto establish exactly where to place it. The changes in the sedi-ments are more gradual than during the isolation and it is there-fore difficult to define a 'transgression contact,. physically, thechange is often seen as a higher amount of mineral particles,which often but not always shows up as higher magneticsusceptibil i ty and higher SIRM values; in fact, biogenic magne-tite can give higher signals than minerogenic sediments. Insome cases/ as in sites N2B and N29, the in- and out-flowingtidal currents have eroded parts of the sedimentary column dur-ing the marine ingression. The macrofossil analyses haveshown that the environmental change is gradual with severalmarine species showing up in small numbers at the same timeas freshwater species gradually decline in numbers of indivi-duals. lt is not unusual to find both marine and freshwater spe-cies in fairly large numbers in the first 10-1 5cm of thesediment column above the registration of the first marineingression. The marine taxa that first migrate into the basinsare often Foraminifera, brown algae such as Sphacelaria sp.,Rhizoclonium sp. and Desmarestia sp., bivalves such as Myti-I u s edu I i s or Macom a balth i ca, hydrozoans of the order Hydro-idea, the ice-cream-cone worm Pectinariaand flatworms of theorder Tricladida. Remains of the marine fish Mallotus villosushave been found in the transgression sequence of N25 and infully marine sediments in N2B.

A Lateglacial and Holocene sea-levelcurve flrom the Nanortalik area

The isofation and transgression ages shown in Fig. 14 area summary of the seven sites from Bennike et at. (2002)(Table 3) and the six sites described in this paper. The sea levelsshown are the inferred values for the individual sites and noadjustment has been made for possible differential isostatic sig-nals between sites that do not lie on the same isobase for thereason that the orientation and gradient of the isobases cannotbe determined with accuracy from the available data. The sitesN16, N18, N19, N21, N22, N24, N2B, N29 and N30 areat very comparable distances from the present ice marginand from the shelf edge and differential isostatic signalsfor these sites are l ikely to be small. The oldest isolation basinCopyright @ 2006 John Wiley & Sons, Ltd.

(site N14) also lies further offshore and its elevation would beincreased if projected onto a sea-level curve for the main clus-ter of sites. Therefore Fig. 14 should not be interpreted as asea-level curve for any specific location but as a general repre-sentation of change over the area spanned by the data points.

Alf ages are calibrated to yr BP using OxCal (v. 3.9) with 2oerror bars, with the exception of the transgression age of basinN27 for which the age is the result of an interpolation betweendates above and below the transgression as discussed above.The altitude error bars are based on the total uncertainty pre-sented in Table 1 . No basins suitable for coring with thresholdsbelow -2.8 m b.h.a.t. have been found in the area and we can-not establish firmly the maximum lowest Holocene sea levelattained in the area. The sediments from sites N2B and N29show erosive contacts (i.e. hiatuses) and this implies a dynamicenvironment with strong bottom currents eroding the soft gyttjasediments and could indicate, especially for site N2B wherethe threshold is wide, that the lowest sea level in the areawas close to this threshold altitude. The fact that marine macro-fossils are found in the sediment column from N25 during afreshwater-dominated period, could also imply that sea levelwas close to the threshold and that marine sediments werewashed in during storm events. Thus the lowest sea level inthe Nanortalik region could not have been much lower thanca. 10m below highest astronomical t ide and occurredbetween 8000 and 6000cal.yrBP (Fig. 1 ). The observationspresented in Fig. 14 show a slow regression between ,|4000and 12000cal.yrBP and an accelerated regression between12000 and 9000cal.yrBP. This implies a fast recession ofthe ice sheet during at least parts of the Lateglacial and theearly Holocene. The slower transgression starting at sometimebefore 5000cal.yrBP could reflect either a readvance ofthe Creenland ice sheet and/or a delayed collapse of theLaurentide peripheral bulge. lt has been suggested that theice sheet margin reached its present position at around 9500BP, and from 9000 BP to 3000 BP the ice margin was behindthis position (Weidick et a1.,2004). This suggestion correlaresvery well with the sea-level observations presented here, as sealevel fell below the present at around 9300cal.yrBp and thenrose from around the mid-Holocene to reach the present levelfairly recently. Neoglacial growth andlor a delayed isostaticresponse to the changes in the Laurentide and/or the Creenlandice sheets can explain such behaviour in relative sea level. ltcould also be a combination of both processes.

The data in the diagram in Fig. 14 are uncorrected for differ-ential isostatic rebound and should be seen as raw data and notas a relative sea-level curve for any specific locality. We arecurrently working on relative sea-level changes in a secondarea of southern Creenland (Qaqortoq-Narsarsuaq). When thisanalysis is completed we will be in a better position to establishthe gradients and rates of crustal rebound and to draw conclu-sions about ice thickness and ice margin locations for the Late-glacial period. This work is important as observations ofrelative sea-level changes in southern Greenland are few andprovide few constraints for glacial-isostatic adjustment models.Poor agreement is seen between results from glacial-isostaticadjustment models reconstructing sea-level and glacial historywith those few observations available at present (Bennike etal.,2OO2; T arasov and Peltier, 2002; F lem i ng and Lambeck, 2OO4).Fleming and Lambeck (2004) present a curve for rhe Kap Farvelarea that shows a predicted relative sea-level rise for the entireLateglacial and Holocene, and conclude that their modelunderestimates the amounts of ice melted from the region. Thisconclusion agrees well with our observations. Tarasov andPeltier (2002) produce a curve for the Julianehåb area showinga transgression from the LCM until around 1 0 000 BP, when thepresented relative sea-level curve reaches almost 60m a.s.l.

.1. Quaternary Sci., Vol. 21(11 2948 (2006)

46 JOURNAL OF QUATERNARY SCIENCE

40Èc;Ëd 3 0çL

oo'= zo(Ú.9Eoc9 1 0Ø(ÚØo-o ) u=oo-o(Ú -10oEf.==

-20

20004000800010000 600016000 14000 12000

Material CommentsSite Laboratory no. Depth (cm) Age laC yr BP Calibrated age2o (95.4'/") OxCal v. 3.9

N26

Age (calibrated years BP)Figure 14 Relative sea-level changes in the Nanortalik area du ring the Lateglacial and Holocene as indicated by isolation and transgression ages of13 different basins. Altitude errors are listed in Table 1 and age error bars of 2o are shown as given in Table 2. Note that some age spans are dividedinto two or three periods. Note also that the transgression for site N29 is dated by two samples on each side of a hiatus and shows a large span for thisreason. The transgression age for site N28 is a minimum date, taken from a sample above a hiatus in marine sediments

Tabfe 3 Selected AMS radiocarbon ages from the basins studied by Bennike et al. (2OO2)

N 1 4

N l 8N24N l 9N 1 6

N22

N 2 1

Ua-'14844

Ua-'154'14Ua-1 5902Ua-'154'17Ua- I5412

Ua-1 5419

Ua-1 541 B

765-771

367.5-370262

793-796589-591.5

1125-1128

700-704.5

11 665 + .125

1 0 2 0 0 + 1 1 01 0 0 1 5 + 1 2 09810 +1759240+.95

8905 + 90

8930+80

14 850-1 5 050 p.3 %)13 150-' ,14 i 50 (93.1'/")

11 250-126501 1 1 5 0 - 1 2 1 5 01 0 5 5 0 - 1 2 0 5 0i 0 6 1 0 - 1 0 6 8 01 0 2 1 0 - 1 0 6 0 0

9700-10250

9750-10240

B r y u m s p .

Daphnia pulexBu lk

Wa rnsto rfi a exan n u I ataBatrachi u m confervoides,

Empetrum nigrum

B. confervoides, Hi ppu ris vu lgaris,E. nigrum, Bryum sp., W. exannulata

B. confervoides, E. nigrum

lsolat ion

lsolat ionAbove isolationAbove isolation

lsolat ion

Above isolation

Isolation

From 10 000 years BP and onwards the curve shows a regres-s¡on that continues unti I the present. Tarasov and Peltier (2002)compare their modelled relative sea-level curve for southGreenland to a single observation and because this one pointdoes not match their model prediction, they suspect the obser-vation to be erroneous. Their model prediction does notshow much resemblance to the observat¡ons presented in this

Copyright @ 2006 John Wiley & Sons, Ltd.

work either, implying that their model needs to be modified. InBennike et al. (2002) the analysis is more extens¡ve for the areabecause it deals only with southern Creenland. Even thoughthe comparison between the observations and the predictionsare unsatisfactory, they conclude that the model underesti-mates the reduction in ice thickness and that the timing ofthe recession needs to be considered in more detail. The

J. Quaternary Sci., VoL2'l(1\ 2948 (2006)

SEA-LEVEL CHANCES IN SOUTHERN CREENLAND

analyses imply rapid melting before the earliest isolations andthis conclusion agrees well with our observations. Observa-tional data from only one area wil l not give a unique solutionand it is therefore important that more data are collected beforenew extensive modelling analyses can be made.

Conclusions

Holocene relative sea-level changes in southern Greenlandhave previously been observed mostly on land and in lakesand they indicate that levels were below present sea-level forthe past 10000yr. In this paper we present evidence that thelocal sea-level reached -10m below highest astronomicaltide (slightly lower than 7.Bm b.h.a.r.) in the interval from8000 to 6000 cal. yr BP. Combined with earlier results we candraw the following conclusions about sea-level changes in theNanortalik area of southern Creenland:

1. The relative sea-level fall in the earlv Holocene was raoidfrom the time the area became ice-free, whereas the relativesea-level rise in the mid- to late Holocene was slower andmore gradual.

2. The fall in sea level between 12000 and 10200cal.yrBpshown by Bennike et al. (2002) continued until at least9000cal.yrBP, by which time sea level reached below-7.8m b.h.a.t. and continued to fall some time after that.The relative uplift of the area during that time was ca.12mmyr- ' .

3. Between 8000 and 6000 cal. yr BP, sea level in the Nanonalikarea reached its lowest level at around 10 m b.h.a.t.

4. During the mid-Holocene, at or before ca. 5000ca|. yrBp,sea level again reached above -7.8 m b.h.a.t. and continuedto slowly rise until the present day. The mean submergenceof the area during this time was 1.5-2 mm yr-l.

Our results explain why Norse ruins are found so close to thepresent shore (Bak, 1 969); because since the Norse people builttheir houses around 1 000 years ago, relative sea level has risenca. 0.5-l m. Also, evidence for the presence of palaeo-Eskimocultures is extremely rare in southern Greenland. The lack ofPalaeo-Eskimo ruin sites is not necessarily evidence of theirabsence. lf these early Palaeo-Eskimos reached Creenland, set-tlements are likely to be found below present sea level becauselocal sea level has risen 5-6 m since 4000 cal. vr Bp. The oues-tion remains Did Palaeo-Eskimo cultures reach this oart ofCreenland? lf so, has the evidence been preserved, or havetidal currents and wave action destroyed them?

Acknowledgemenfs We thank the Carlsberg Foundation and theCrafoord Foundatíon for all funding that made this study possible,and the Swedish Polar Research Council for their logistic support.Kurt Lambeck thanks the Swedish Research Council for the financialsupport as Tage Erlander Professor. We also acknowledge VasileErsek and Niklas Sparrenbom for all their work and help in thelaboratory, Dr Kevin Fleming for fruitful discussions on geodynamics,Palle Bo Nielsen at Farvandsvæsenet for making tidal charts easilyaccessible and discussions on tidal issues, Ole Andersen at Kort &Matrikelstyrelsen for calculating the astronomical tides, Dr CöranSkog and Dr hab. Tomasz Coslar for help and discussion about radio-carbon ages, Professor Mats Sonesson for help in identifying mossremains from the cores, Dr Poul Møller pedersen for help withidentification of algae, Dr Lena Barnekow and professor CunnarDigerfeldt for support in identifying macrofossils and all others thatin any way have contributed to making this investigation successful.Last but not least we would like to express our gratitude to Morten

Copyr¡ght @ 2006 John Wiley & Sons, Ltd.

Rasch and the anonymous reviewer who helped improve themanuscript.

References

Bak O. 1969. Herjolfsnæs. Grønland 1969: 349-352.Bennike O.1992. Palaeoecology and paleoclimatology of a late Holo-

cene peat deposit from Brændevinsskær, central West Creenland.Arctic and Alpine Research 24: 249-252.

Bennike O. 1995. Palaeoecology of two lake basins from Disko, WestCreenf and. Journal of Quaternary Science 10: i 49-1 55.

Bennike O.2002. Late Quaternary history of Washington Land. NorthCreenland. Boreas 31 : 260-272.

Bennike O, Weidick A. 2001. Late Quaternary history around Nio-ghalvfjerdsfjorden and Jøkelbugten, North-East Creenland. Boreas3O:205-227.

Bennike O, Björck S. Lambeck K.2OO2. Estimates of South CreenlandIate-glacial ice limits from a new relative sea level curve. Earfh andPlanetary Science Letters 197: 171-186.

Björck S. 1979. Late Weichselian stratigraphy of Blekinge, SE Sweden,and water level changes in the Baltic lce Lake. LUNDeUAThesis 7,248 pp.

Björck S, Persson T. 1981 . Late Weichselian and Flandrian biostratigra-phy and chronology from Hochstetter Forland, Northeast Creenland.Meddelelser om Grønland, Geoscience S: 19.

Björck S, Bennike O, Ingólfsson Ó, Barnekow L, penney DN. 1994.Lake Boksehandsken's earliest postglacial sediments and theirpalaeoenvironmental implications, Jameson Land, East Creenland.Boreas 23: 459472.

Bronk Ramsey C. 2001. Development of the radiocarbon programOxCal. Rad i ocarbon 43: 355-363.

Escher A, Watt WS. 1976. Summary of the geology of Creenland. InGeology of Greenland, Escher A, Watr WS (eds). The CeologicalSurvey of Creenland: Copenhagen; 1 G-l 6.

Farvandsvæsenet. 2000. Tidevandstabeller 2O0l for GrønlandskeF arva nde. Farvandsvæsenet: Copenhagen.

Farvandsvæsenet. 2001. Tidevandstabeller 2002 for GrønlandskeF arvande. Farvandsvæsenet: Copenhagen.

Fleming K, Lambeck K.2OO4. Constraints on the Greenland lce Sheetsince the Last Clacial Maximum from sea-level observat¡ons and gla-ciaf -rebound models. Quaternary Sc¡ence Reviews 28: 1053-'1077.

Fredskild B. 1973. Studies in the vegetational history of Creenland,palaeobotanical investigations of some Holocene lake and bogdeposits. Meddelelser om Grønl and 1 9B(4): 245.

Funder S. 1978. Holocene stratigraphy and vegetation history in theScoresby Sund area, East Creenland. Bulletin Crønlands CeologiskeUndersøgelse no. 1 29; 66 pp.

Funder S. 1979. The Quaternary geology of the Narssaq area, SouthCreenland. Rapport Crønlands Geologiske Undersøgelse no. 86;24 pp.

CabelJørgensen CCA, Egedal J. 1940. Tídal observations made atNanortalik and Julianehåb in 1932-1934. Meddelelser om Grønland1O7(Ð :47 .

Hammer M. 1944. Studies on the Orbatids and Collemboles ofCreenland. Meddelelser om Grønland 141(3): 21O.

Kelly M, Funder S, Houmark-Nielsen M, Knudsen KL. Kronborg C.Landvik l, Sorby L. i999. Quaternary glacial and marine environ-mental history of northwest Creenland: a review and reappraisal.Quaternary Science Reviews 18: 373-392.

Kuijpers A, Abrahamsen N, Hoffman C, Hühnerbach V, Konradi p,Kuzendorf H, Mikkelsen N, Thiede j, Weinrebe W, shipboard scien-tific party of RV Poseidon, surveyors of the Royal Danish Administra-tion for Navigation and Hydrography 1999. Climate change and theViking-age environment of the Eastern Settlement, South Creenland.Ceology of Greenland Survey Bulletin 183: 61-67.

Lambeck K, Chappell 1.2001. Sea level changethroughthe lastglacialcycle. Science 292: 679-486.

Long AJ, Roberts DH. 2OO2. A revised chonology for the ,Fjord Stade,moraine in Disko Bugt, west Creenland. lournal of euaternaryScìence 17: 561-579.

,. Quaternary Sci., Vol. 21(1) 2948 (2006l

4B JOURNAL OF QUATERNARY SCIENCE

Long AJ, Roberts DH, Rasch M. 2003. New observations on relative sealevel anddeglacial historyof Greenlandfrom Innaarsuit, Disko Bugt.Quaternary Research 6O: 162-17'l .

Long AJ, Roberts DH, Wright MR. 1999. lsolation basin stratigraphyand Holocene relative sea-level change on Arveprinsen Ejland,Disko Bugt. West Creenland. lournal of Quaternary Science 14:323-345.

Mathiassen T. 1936. The eskimo archaeology of Julianehaab District.Meddelelser om Crønland 118(): 141 .

Rasch M. 2000. Holocene relative sea level changes in DiskoBugt, West Creenland. Journal of Coastal Research 16:306-31 5.

Rasch M, Jensen )F.1997. Ancient Eskimo dwell ing sites and Holocenerelative sea level changes in southern Disko Bugt, central WestCreenland. Polar Research 16: 1 01-1 1 5.

Rasch M, Holm Jakobsen B, Nielsen N. 1997. Ceomorphology andsedimentary record of three cuspate forelands as indicators of late

Holocene relative sea-level changes, Disko, West Greenland.Danish Journal of Ceography 97 : 3346.

Sandgren P. Snowball l. 2001 . The Late Weichselian sea level history ofthe Kullen Peninsula in northwest Skåne, southern Sweden. Eoreas3 0 : 1 1 5 - 1 3 0 .

Stuiver M, Reimer P, Bard E, Beck WJ, Burr CS, Hughen KA, Kromer B,McCormac C, van der Plicht J, Spurk M. 1998. INTCAL 98 radiocar-bon age calibration, 24,000-Ocal. BP. Radiocarbon 40: 1041-i 083.

Tarasov L, Peltier WR. 2002. Greenland glacial history and local geo-dynamic consequences. Ceophysícal Journal International 15O:198-229.

Walden J, Oldfield F, Smith JP (eds). 1999. Environmental Magnetism:A Practical Cuide. Technical Cuide, No. 6. Quaternary ResearchAssociation: London; 243 pp.

Weidick A, Kelly M, Bennike O.2004. Late Quaternary developmentof the southern sector of the Creenland lce Sheet, with particularreference to the Qassimiut lobe. Boreas 33:284-299.

Copyright @ 2006 John Wiley & Sons, Ltd. l. Quaternary Sci., Vol. 21(1) 2948 (2006\