The evolution of Ace Lake, Antarctica, determined from sedimentary diatom assemblages

ELSEVIER

Palaeogeography, Palaeoclimatology,Palaeoecology124 (1996) 73-86

The evolution of Ace Lake, Antarctica, determined from sedimentary diatom assemblages Serena P. Fulford-Smith a, Elisabeth L. Sikes b,c Institute of Antarctic and Southern Ocean Studies, University of Tasmania, GPO Box 252C, Hobart, Australia b Antarctic Co-operative Research Centre, GPO Box 252C, Hobart, Australia c Australian Geological Survey Organisation, GPO Box 378, Canberra, Australia

Received 6 May 1994; revised and accepted I November 1995

Abstract

The sediments in Ace Lake in the Vestfold Hills, East Antarctica, contain a continuous sequence of diatom frustules that record the lake's history since its formation during the retreat of continental ice more than 9200 years ago. Frustules from six indicator groups; Pinnularia microstauron, Nitzschia cylindrus, Nitzschia curta, Eucampia antarctica, a freshwater Stauroneis species and several centric species in a sediment core were used to determine the sequence of the lake's evolution. The history of the lake falls into five stages. Ace Lake began as a marine inlet influenced by dynamic mixing of ocean and meltwater inputs. As the ice sheet retreated, isostatic uplift isolated the lake allowing it to be flushed by meltwater input from the retreating ice sheet. Over the course of 800 years (~9200-8400 14C yr B.P.) the lake became meromictic supporting a freshwater diatom assemblage. Approximately 6700 years ago, coinciding with Antarctic sea level maxima, diatom assemblages indicate that seawater flooded over the sill into Ace Lake disturbing the freshwater meromixis. The sediments in this period were laminated and contained elemental sulphur suggesting that the marine input was limited in extent and energy. Approximately 5500 years ago this marine input ceased and the lake again became a meromictic basin which stabilised over 1700 years to become the lake that it is today and has been with little change for about the past 4000 years.

1. Introduction

The lakes of the ice-free oasis, Vestfold Hills, East Antarctica, provide in their sediments a continuous record of the physical changes that have occurred since at least 9200 years ago. Because Ace Lake's sediments are subject to little bioturbation and reworking they are an accurate record of the phytoplankton assemblages at the time of deposition. Ace Lake is situated on Long Peninsula in the Vestfold Hills (Fig. 1). It was chosen for this palaeoclimatic study because it has been the site of numerous studies since 1978 (Burton and 0031-0182/96/$15.00 © 1996 Elsevier Science B.V. All rights reserved PII S0031-0182(96)01657-0

Barker, 1978; H a n d and Burton, 1981; Volkman et al., 1986, 1988; Burch, 1988; Van den H o f f e t al., 1989; Mancuso et al., 1990; Bird et al., 1991). Although it has been suggested that portions of the Vestfold Hills (including the area of Ace Lake) remained ice free during the last glaciation (Pickard, 1986), prior study indicates that lake formation in the Vestfold Hills began when the Antarctic ice cap retreated over 8000 years ago (Fitzsimons and Domack, 1993). Abundant fossils of extant mosses and lichens and the ventifacts and striae of the rocks of the Vestfold Hills suggest that the climate and wind direction of the region

74

S.P. Fu[fbrd-Smith, E.L. Sike,s/Palaeogeography. Palaeo{'limatology, Palaeoecoh)gy 124 ( 1996J 73 86

£,

o

5 I~

km

'

f,

' d '.., fa / ,9~"-'0 '

i

Ace Lake

PRYDZ BAY

'¢~

: PENINSULA

68° 30'

BROAD PENINSULA Lake 2 ~

, g DAVIS

Scale Lake - l['LakeC

~

a

k

e

~ ~

Lake 3 Lake

Lake

FJORD

MULE

~°

~,o- Plain

7 8 ° 00' E

Fig. 1. The Vestfold Hills, East Antarctica, showing the position of Ace Lake on Long Peninsula and other topographical feature,~. Unofficial lake names are numbered (Lake / = Lake C. Lake 2 - Medusa Lake, Lake 3 = Cat Lake).

have been constant since the last ice retreat in the early Holocene (Pickard et al., 1983; Pickard et al., 1986). Climate records of the region have been collected since 1957 from Davis Station until present, excluding the period November, 1964 February, 1969, when the station was closed. Palaeoecological and geochemical studies have been carried out in numerous lakes in the region,

including Ace Lake, as well as Ellis Fjord and Taynaya Bay (Burton and Barker, 1978; Pickard et al., 1986; Volkman et al., 1986; Bronge, 1989; Mancuso et al., 1990; Bird et al., 1991). These studies included sulphur isotope measurements, ~4C and fossil dating, lipid examinations, and sediment core analyses. The diatom composition in sediment records is

S.P. Fulford-Smith, E.L. Sikes/Palaeogeography, Palaeoclimatology, Palaeoecology 124 (1996) 73 86

a widely applied palaeoecological tool used to determine the evolution of lakes, fjords, rivers, and marine basins as well as the climate change of a region (Bronge, 1989; Leventer and Dunbar, 1988; Stockwell et al., 1991; Wasell and Hakansson, 1992; Pichon et al., 1992). Diatom analysis of sediment cores has been used extensively in Prydz Bay (Fryxell, 1989) and other marine areas of Antarctica (Leventer and Dunbar, 1988); the species that occur at certain temperatures and geographical areas are well detailed. Wasell and Hakansson (1992) studied the diatoms in a sediment core from Skua Lake on Horseshoe Island, off the Antarctic Peninsula. In this work they were able to relate the changes in the diatom assemblage of the lake to brackish, marine, and freshwater conditions. Burton and Barker (1978), first examined the plankton assemblage of Ace Lake. They found that the lake's water column contained a dominance of diatoms, mostly Fragilaria sp. and Navicula sp. Studies by Burch (1988) and Volkman et al. (1988) indicated four main species in the phytoplankton assemblage; a pymnesiophyte, a dinoflageUate, Cryptomonas sp. and the antarctic prasinophyte, Pyramimonas gelidicola. Volkman et al. (1988) found no evidence of diatom-derived compounds in their analysis of organic compounds but did report a few diatom cells in the preserved lake samples. In their lipid analysis of the surface sediments of the lake, Mancuso et al. (1990), found evidence of a variety of diatom species not noted in the plankton assemblage. These conflicting reports on the presence of some diatoms in the lake could be due to their association with the benthic mats that cluster at the edge of the lake. This association would make the diatom distribution in the water column patchy over time (Burch, 1988), but provide a source for frustules in the sediments. Burton and Barker (1978) noted the presence of two zooplankton; Paralabidocera antarctica and Acartia sp. in the lake. However, the reported presence of Acartia sp. in Ace Lake was not confirmed by Bayly and Burton (1987), who only found P. antarctica. With only one species of zooplankton confirmed to exist in the lake, Paralabidocera antarctica, grazing is probably not

75

a major restriction on biomass (Volkman et al., 1986). Examinations of the sedimentary sulphur chemistry of Ace Lake by Burton and Barker (1978) suggested that there were three different stages in the lake's development. The first phase included several cycles of mixing and meromixis, with most of the seawater being replaced by fresh runoff and 76% of the sulphur disappearing from the system through biological reduction in the monimolimnion. This was followed by a phase of complete mixing which involved the replenishment of the sulphur concentrations. The final and present stage is meromixis where the local freshwater is layered on older water left after the general mixing ceased. Burton and Barker (1978) suggested that the holomictic stage was caused by a period of warmer weather which increased the surface salinity and decreased the stabilising ice cover. The three stages of lake development suggested by Burton and Barker were reaffirmed by Bird et al. (1991), based on the analysis of sulphur, 613C, salt and water contents of the same sediment core used by the present study (Fig. 2). Bird et al. (1991) noted that whilst the sediments in their unit 2 of the core had marine characteristics, those of unit 1 and 3 did not. Therefore, they concluded unit 2 was marine but the origin of unit 3 was difficult to interpret, based on the sedimentological evidence. They hypothesised that Ace Lake contained no marine sediments, and that all sediments in the core were lacustrine in origin. They also noted algal mats in the sediments which are similar in description to those found in Highway Lake and in the Antarctic Dry Valley lakes (Wharton et al., 1983). These results correlate well with the work of Volkman et al. (1986) on the same core, whose analysis of sedimentary hydrocarbons suggests the lake was initially oxic, then went through a low oxygen phase, before the lake developed its permanent anoxic basin in the last 1000 years. Ace Lake's suitability for palaeoclimatic research rests on these factors: it is free from external contaminations, it is a closed meromictic system with a limited biota and has been monitored extensively for the last 24 years. Past studies conducted on Ace Lake did not include a downcore analysis of diatom frustules. This analysis is poten-

S.P. Fu[iord-Smith, E.L. 5"ikes,.'Pa/aeogcograplo'. Pahwoclimamlogy. Palaeoecology 124 (1996) 73 86

76 100 "

80"

•~-

Water (wt %)

60"

~'~ 40" C3 20-

Salt (wt %) 10

Sulphur (wt % ) l , . - -

", m

0

Organic carbon (wt %

m

-10

& d

, *

*

!

*t

*!

-20

13C Organic (%0 PDB) -30 0

2

6

8

10

Age (ky BP) Fig. 2. Physical parameters of the Ace Lake sedimem core (after Bird et al., 1991 ). 14C dates are indicated by asterisks at 5310yr(20 35cm),6110yr(35 75cm).6740yr(135 140cm). and 8380 yr (165 170cm). The marine influenced period of the lake (6500 5500 yr B,P.) is illustrated by the change in almost all physical parameters measured in this core: water content increases 50%. salt content increases by three li)ld. elemental sulphur is relatively and consistently low and organic (5L3C is enriched. The overlap of the Bird et al. ( 1991 ) units I -3 and our units a e are designated by dashed lines.

tially valuable for determining the past environmental conditions of Long Peninsula and will allow comparison with studies of other lakes. The purpose of this study was to determine the evolution of Ace Lake since the retreat of the continental ice by examining the changes in the diatom assemblages preserved in the lake's sediments. By looking at these results in conjunction with prior geochemical studies we can refine our understanding of the lake and the history of the region.

2. Methods

The core used in this study is the same core that was used by Volkman et al. (1986) and Bird et al. (1991). It was collected in 1978 through the ice

cover of the lake using a Zullig corer. The core comes from the centre of the lake at a water depth of 23 m, in the anoxic zone. After retrieval, the core was cut into 5 cm blocks and then stored at 18 C . It had been partially melted and refrozen during previous sampling. Sediment samples were desalinated to prevent salt crystal formation during the preparation of the slides. A small amount of wet sediment was rinsed three times by suspension in 10 ml of distilled water followed by centrifugation at 2000 rpm t\~r 3 minutes and removal of the supernatant. Once rinsed, the sample was resuspended and allowed to settle for 3 seconds, after which approximately l ml was placed on a 22mm x 40ram coverslip and heated at 5 0 C for an hour or until the water in the sample had evaporated. It was then mounted with 4 drops of Naphrax fixative and covered with a slide. Completed slides were then left on a warm (25c'C) hotplate to facilitate the spreading and setting of the Naphrax for approximately I hour. Three replicate slides were made for each sediment sample. Three hundred diatoms frustules were counted from each of the replicate slides, differentiating them into six parameter groups. If the slides did not contain three hundred diatoms the average number of diatoms on the three replicate slides was taken as representative of the assemblage at that time. The indicator species chosen to be representative of different conditions were: Pinnularia microxtauron (freshwater benthic), Stauroneis sp. A t fresher benthic ), Nitzschia cylindrus (pioneer benthic), Nitzschia curia (pioneer benthic), Eucampia antarctica (marine); Thalassiosira, Porosira, Coscinodiscus and Asteromphalus, were categorised as centrics (marine). Other diatoms were noted as such and used to determine the total cells counted. The species were chosen by reference to Wasell and Hakansson (1992) as representative of those species that were freshwater, those that preferred marine environments and those that were halotolerant (pioneers). The salinity preferences of the parameter groups were used to determine if the changes in the preserved diatom assemblage in the lake sediment, represented changes in the salinity of the lake. The two Nitzschia species were chosen as the "pioneers" because they are frequently

S.P. Fulford-Smith, E.L, Sikes/Palaeogeography, Palaeoclimatology, Palaeoecology 124 (1996) 73 86

found in sea ice communities (Leventer and Dunbar, 1988; Fryxell, 1989; Stockwell et al., 1991) and are therefore exposed to salinities of 5-60 ppt (Johnsen and Nost Hegseth, 1991). Wasell and Hakansson (1992) documented the freshwater species Pinnularia microstauron in the fresh waters of Skua Lake, a shallow lake on Horseshoe Island off the Antarctic peninsula. Stauroneis species are associated with shallow brackish waters and were associated with fresher lake conditions in Skua Lake (Wasell and Hakansson, 1992). Eucampia antarctica is a planktonic diatom which like centric species, is associated with the open ocean and sea ice edges (Leventer and Dunbar, 1988; Fryxell, 1989). The sediments of five lakes (Medusa, Scale, Cat, Collerson, and Lake C) were used for comparison.

77

These contained large amounts of fine clastic material. The sediment samples from these lakes were filtered through 0.8 ~tm filter paper with distilled water before enumeration, which may have caused some loss of particles. Such errors could not have occurred in the Ace Lake data because the samples were not filtered.

3. Results

3.1. Diatom assemblage Five main assemblage changes were observed in the diatom frustules throughout the Ace Lake core (Fig. 3).

O

¢-

W o

•

% Centrics (Open Marine)

[]

% E. antarctica(Open Marine)

[]

% N. cyclindrus(Pioneer)

•

% N. curta (Pioneer)

•

% Staumneissp. (Freshwater)

•

% P. microstauron(Freshwater)

°L,

°

|Ill

0

0

O

Depth (cms) Fig. 3. Diatom stratigraphy of Ace Lake sediment core. High abundance of E. antarctica and centric species indicate a marine environment from 25 to 135 cm. Dominance of freshwater species Stauroneis sp. and P. microstauron from 135 to 170 cm suggest lacustrine conditions. A mixed assemblage of open marine species co-existing with freshwater diatoms from 170 to 180 cm, indicates mixing of the two environments occurred in the early stages of the lake's history.

78

S.P. Fu!/m'd-Smith, E. L 5"ike,~,Pa/aeoeeo,erat>/O', Palaeoc/imato/ogy,

Unit a: 0 22 cm

Benthic diatoms, potentially from mats at the edge of the lake, dominate the assemblage in the top segment of the core, with freshwater dwelling Stauroneis being the main species ( 14'7,, total cells). This assemblage also contained the two pioneer benthic species. N. curta ( 1.8% total cells) and N. cylindrus (1.0% total cells) and numerous other benthic species, but no open marine species were present ( Fig. 3 ).

Palaeoecology 124 ~1996) 73 86

4. Discussion

The six species selected as reference species illustrate the change in salinity from freshwater to marine in the lake's history. The species were used singly or were grouped as a particular parameter type: pioneer, freshwater dwelling, or open marine species. The concentrations of these species and parameters in the core samples varied dramatically indicating at least one marine and one freshwater phase in the lake's history (Fig. 3).

Unit b. 22 35.5 cm

Freshwater species were present, but the main species present are the pioneer benthics (26'I~i, of total cells) and open marine dwelling forms (6.7'!, total cells). Unit c. 35.5 135 cm

Open marine (4.8'7,, total cells) and pioneer benthic species (40% total cells) were found throughout this section of the core. It was the only section of the core that contained Eucampia antarctica (0.5% total cells). Large numbers of the fi'eshwater benthic species, Staurom'is sp., were found only in the deepest sample of this unit t3(t",,, total cells). Unit d." 135 170 cm

This unit was dominated by the freshwater dwelling species, P. microstauro~l (86% total cells), however, the pioneer species were present m deeper samples of the unit (6.3% total cells). Unit e." 170 185 cm

This unit was mixed and had three distinct regions. From 170 175cm there was a concentration of the pioneer species (26% total cells), with small numbers of open marine species present (1.3% total cells). From 175 180cm there was a mix of N. cvlindrus, the pioneer benthic species (9% total cells), the centric open marine species (8% total cells) and Stauroneis sp., the freshwater species (2.5% total cells). From 180 185cm all species were present in similar proportions, freshwater dwellers (7"/,, total cells), pioneer benthic species (12% total cells) and open marine species (6% total cells), except for the open marine species E. antarctica.

4. l. Stratigraphy

In order to determine absolute ages for the sequence of these changes, 14C dates were measured for the Ace Lake core. Initial ~4C dates were reported by Bird et al. (1991). They dated two sediment levels in the core, one at their unit 1 ~ transition (our unit a b transition: 20 35 cm) and the other in the middle of their unit 2 (our unit c: 35 75cm). The dates were 5310+_90 14C yr B.P. (ANU-6414) and 6110_+180 14C yr B.P. (ANU-6419), respectively. Additional dating for this study was provided by M. Bird (Research School of Earth Science, Australian National University in 1992), which produced ages of 6740 ~4C yr B.P. at 135 140cm and 8380 14C yr B.P. (ANU-8166) at 165-170cm. Note that large amounts of sediment were needed to obtain reliable dates in the initial study, making the ages from Bird et al. (1991) the average of a large sediment time slice. Carbon dates in Antarctic waters are subject to a false age of 1000-1500 years. This is a result of low' concentrations of ~4C in the Southern Ocean waters caused by the upwelling of old bottom water and the input of 14C-depleted glacial melt water (Omoto, 1983). This *'reservoir" effect for the terrestrial environment is zero and for marine areas in the vicinity of Davis Station is 1000-1300 years, while linear regression of ~4C dates in Highway Lake indicate a lacustrine reservoir effect of approximately 1000 years (Bird et al., 1991 ). Based on this, the total reservoir effect for sediment deposited under marine conditions in the lakes of the Vestfold Hills was estimated by Bird et al. (1991) to be 2200 years.

S.P. Fulford-Smith, E.L Sikes/Palaeogeography, Palaeoclimatology, Palaeoecology 124 (1996) 73-86

Because previous interpretations of the history of Ace Lake have suggested there are no marine sediments in the lake and that conditions have been lacustrine since its original formation (Bird et al., 1991 ), we applied two different corrections to the 14C dates. To assess maximum and minimum sedimentation rates in Ace Lake, a reservoir effect of 1000 years was used to adjust dates in that portion of the core which was determined to be lacustrine from the presence of freshwater dwelling diatoms. An effect of 2200 years (after Bird et al., 1991) was used as a total reservoir effect for sediments which appeared to be deposited under marine conditions. Hence, the dates were adjusted to obtain a minimum time estimate for the duration of a marine phase in order to rule out the possibility of a "splashover" rather than a true marine incursion at around 6500-5500 14C yr B.P. This adjustment indicated sedimentation rates of approximately 1.4 mm/yr for deposition in the marine environment (55 cm-137 cm) and a sedimentation rate of 0.18 mm/yr for deposition in lacustrine conditions (137cm-167cm), with a total minimum sedimentation rate of 110 cm in 1200 years of marine incursion. This gives a minimum time for the sedimentation of marine diatoms that is a longer period than could be expected for a "splashover" event. These sedimentation rates compare favourably with those of Highway Lake and Organic Lake (Bird et al., 1991). With these corrections, however, the corrected date at 35-75 cm is younger (3900 14C yr B.P.) than the corrected date at 20-35 cm above it (4200 a4C yr B.P.). This indicates the reservoir age corrections determined for the Davis Station area today do not apply strictly in the past. Nonetheless, it is certain that a reservoir effect must occur, but, because it is not well known, we plot all data using uncorrected ages (Fig. 4). The five sediment units determined from the diatom stratigraphy demonstrate the occurrence of five distinct phases in the evolution of the lake (Fig. 4). This refines the three phases suggested by Burton and Barker (1978) and Bird et al. (1991). Unit a and b are equivalent to the Bird et al. (1991) unit 1. Similarly, our unit c corresponds to their unit 2 and our units d and e to their unit 3.

79

4.2. Evolution of Ace Lake With ages determined for the sediment levels in the core the diatom stratigraphy can be used to interpret the evolution of the lake in relation to the changes in the regional environment: Unit e (185-170 cm; 9200~8400 14C yr B.P.) This unit is the most complicated and is interpreted as comprising the sediment of the lake when it was first forming. This segment of the core shows an upcore increase in all of the physiochemical parameters except 613C (Bird et al., 1991) (Fig. 2). The diatom assemblage is mixed throughout, with the presence of both fresh and open marine diatoms (Fig. 4). A strong melt water influence at that time is indicated by the presence of Pinnularia microstauron, a species which is found primarily in freshwater lakes and melt water streams of the Vestfold Hills. The simultaneous presence of the open marine centric species suggests a strong marine influence. The sediments in this portion of the core are not laminated, indicating that circulation was not restricted. Low diatom concentrations in this section suggest that they did not actually grow in this location but may have been transported there by water movement. Therefore, an unrestricted marine input and a meltwater input could result in a contemporaneous mix of both freshwater and marine species in the sediment column. The relative increase in the pioneer species in the top segments of this unit suggest a decrease in marine influence and a shift towards lacustrine conditions through time (Fig. 4). This is supported by the slight increase in sulphur content and organic carbon (Fig. 2). As the developing lake became more isolated the mixing influence of the marine and freshwater inputs decreased and the lake became increasingly stratified, developing an anoxic basin. This is registered by an increase in the sediment sulphur levels as well as evidence of a dominant assemblage of pioneer species. The marine species of phytoplankton found in many of the inland lakes in the Vestfold Hills are often dispersed by the sea spray as demonstrated elsewhere in Antarctica (Schmidt et al., 1990). To determine if the marine input was via direct circula-

0.0

L

0.4

1~. . . . . . . . . .

i

0.2

:_

0.6 ,

I

2 ,

(

4 ,

I

6 ,

% Centrics

I

8 ,

10 0

20

40

60

80

1060

% P. microstauron

L

10 I

20 =

30

% N. cyclindrus

40 0

2

4

6

% N. curta

8

i

,

L

10

2/

10 0

i

I

unit b

20

% Stauroneis sp.

Fig. 4. Changes in the parameter species against time. t4C dates (uncorrected) are indicated by asterisks at 5310 yr (20 35cm), 6110 yr (35 75 cm), 6740 yr (135 140 cm), and 8380 yr (165 170 cmL The dominance of P. micro.rtaurorl at 8400 6700 yr B.P. (unit d) indicates that Ace Lake was freshwater at that time. The rapid change from a freshwater stage to the marine stage (6700-5500 yr B.P.; unit c) is evident in the sudden decline in the freshwater species which are replaced by the centrics and pioneer species. An increase in the pioneer marine species N. crlindrus and the freshwater Stauroneis species with loss of the open marine species in the top of the core (4400 yr B.P., unit a) reflects the freshening of the lake alter it was again closed off from the oceanic inputs of the marine phase. Bird et al. ( 1991 ) units I-3 are designated by dashed lines and our units a e are designated by solid lines.

8

4

0

% E. antarctica

30

,2

2.

~2

.m.

-.x

9~

X P. Fulford-Smith, E.L. Sikes/Palaeogeography, Palaeoclimatology, Palaeoecology 124 (1996) 73-86

tion rather than spray, five other lakes from the Vestfold Hills (Medusa, Scale, Cat, Collerson, and Lake C) were chosen as a comparison for Ace Lake. These lakes are inland and have never been flooded by a marine incursion (M. Bird, pers. comm., 1992). The basal sediments of these lakes were examined in order to determine the marine diatom composition and abundance in purely freshwater lakes for comparison with Ace Lake. The sediments of all five lakes contained only fragmented diatom frustules in low numbers (Table 1). Centrics were the most dominant diatom group sampled. The results suggest that the input of marine diatoms by sea spray produces mostly diatom fragments in low densities. We interpret the large abundance of whole marine diatom frustules in unit e in Ace Lake to indicate a direct marine influence. An alternate interpretation is that the lake may have been fresh when it was first formed and the marine and pioneer species present in the sediments may have been introduced to the lake after formation by animal vectors, such as penguins. Studies carried out on freshwater Skua Lake, by Wasell and Hakansson (1992), indicated that the input of marine diatom species, both pioneer and centric, by animal vectors was associated with an increase in organic carbon in the lake. Rookery Lake on the end of Long Peninsula also has a large organic input from the rookeries situated around it (Burton, 1981). The organic content of the Ace Lake core is relatively low throughout the core and we consider it unlikely that there was any

Table 1 Fragments (greater than half frustules) of parameter species found in filtered samples of sediment cores from 5 lakes of the Vestfold Hills. Values represent number of fragments per slide that were greater than half the original frustule size, except for Ace Lake where whole frustules were sampled as well as fragments Species

Scale Collerson Lake C Medusa Cat

Freshwater 2 Pioneer 3 Open marine 6

Total cells

11

0.3 6 8

5 1 3

14.3

9

1 3 3 7 11

3 1.3 6

Ace

18.3 5.3

10.3 26.6

81

association with animal vectors (Fig. 2). Ace Lake is also further from coastal regions with large bird populations than either of the above lakes (Fig. 1) and is considered less likely to have been visited regularly by seals or penguins. The presence of whole marine diatom specimens in the unlaminated basal sediments of Ace Lake suggests that the lake began as a tidally affected inlet on the coast, a mechanism proposed by Adamson and Pickard (1986). As the lake formed, it appears to have been influenced by two inputs, tidal marine and meltwater inflow. Meltwater would have delivered freshwater species to the sediment. Concurrently the tidal influence would have imported marine species to the inlet, which caused the co-occurrence of both open marine and freshwater species throughout the sequence. Whether the basin of Ace Lake was gouged by ice advance of the Last Glacial Maximum or whether this area of the Vestfold Hills remained ice free at that time (Pickard, 1986) cannot be resolved by this study. It is believed that the core used in this study represents the entire sequence of sediments from the lake (Bird et al., 1991). If the area remained ice free it is likely that lower precipitation and colder temperatures caused proto-Ace Lake to dry out at the last glacial maximum at which time any previous sediments may have been removed. However, the absolute age of the Ace Lake basin can only be addressed by further coring and study.

Unit d (170-135 cm." 8400,,~6719 14Cyr B.P.) Throughout this period the sediments increased in organic content and the presence of high sulphur levels indicate the formation of an anoxic basin similar to such basins in Ellis Fjord (Fig. 2) (Bird et al., 1991). Acyclic isoprenoids (phytane, 2,6,10,15,19-penta-methyleicosane and pristane) are indicators for methanogens and the presence of moderate levels of these compounds in this unit indicate slightly anoxic conditions in the bottom waters at that time (Volkman et al., 1986). The dominant species in the sediment was Pinnularia microstauron, with varying numbers of a Stauroneis species. Only a small number of pioneer species were present in the lower sediments and their

82

S.P. Fullord-Srnith, E. L, Sikes/Palaeogeography. Palaeoclimutology, Palaeoecology 124 (1996~ 73 86

number diminished up-core ( Fig. 5). This suggests that the lake became fresher during this period. As Ace Lake is only 2 m below its sill at present, it is probable that the initial marine influence was not long lasting and only minor isostatic uplift was needed to isolate the lake. We suggest that as the melt water input into the lake increased with ice cap retreat, the lake was flushed and the salinity dropped, causing the elimination of marine species from the system. This flushing is supported by the change in the sulphur isotopes observed by Burton and Barker (1978). Watts Lake is believed to have undergone similar flushing about 2000 years ago (Pickard et al., 1986), causing a similar change in assemblage from brackish pioneers to freshwater species as the lake formed a permanent freshwater lens (Bronge, 1989). However, Bronge (1989) determined that Watts Lake became fresh over a period of only 300 years, due to the stratification and upper mixing of the lake, with only a moderate assumed melt-water input (2 m3/s for 2 summer months). His figure is much shorter than the 2000 years for the flushing of Watts Lake suggested by

100 80

% Total Freshwater

I

% Total Pi. . . . .

II

Pickard et al. (1986). Even taking into account possible reservoir corrections discussed above, Ace Lake became purely freshwater over a minimum of 1700-2500 years. This compares favourably to the time Bronge (1989) proposed for Watts Lake, which is situated next to a large drainage system (Druzhby). Ace Lake is not situated near a similar system and therefore freshwater input may have been less than that experienced by Watts Lake and consequently, the flushing time longer.

Unit c ( 1 3 5 - 3 5 cm. 6 7 1 9 ~ 5 5 0 0 14C yr B.P. )

The diatom assemblage changed dramatically through this section with P. microstauron disappearing in favour of both the pioneer species (N. curta and N. cylindrus) and marine species (centrics and E. antarctica) ( Fig. 4). Increased species diversity is indicated by the large number of "other species" and the lower number of pioneer species: Nitzschia slo. (Fig. 3, Fig. 4). This was the only section of the core where E u c a m p i a antarctica was observed (Fig. 3). This is interpreted as the only " o p e n " marine section of the core, as E u c a m p i a

~

°JoO0onMa0nor/

....

='., ~I

II

~1 'i

~,)

II

(J I , ~

"O

~ '

,

i JJ

60"

=

~)

=

I

"1

*

+,

*

40"

i I

I ,,

...... 0

, _ ~ . ~-,, : _- . . . . . . . . . . 1

2

3

4

Age

,.. 5

., 6

_ ; ._':..: . . . . . 7

8

, .... 9

10

(kyBP)

Fig. 5. The diatom parameters in the Ace Lake sediment core show the various stages in the evolution of the lake. Unit e has both open marine and freshwater species reflecting the mixed conditions. L.)zit d has over 80% freshwater species strongly indicating a freshwater environment. Unit c has 20 40% open marine species indicating a marine influence. Assemblagesin unit b indicate the lake was stabilising after the marine phase. Unit a with about 20% freshwater species and a low percent of pioneer species represents the present stage of stable meromixis. ~4C dates are indicated by asterisks at 5310 yr (20 35 cm), 6110 yr (35 75 cm), 6740 yr (135 140cm), and 8380 yr (165 170cm). Bird et al. (1991) units 1 3 are designated by dashed lines and our units a e are designated by solid lines.

S.P. FulJbrd-Smith, E.L. Sikes/Palaeogeography, Palaeoclimatology, Palaeoecology 124 (1996) 73-86

antarctica is a open marine species, not associated with sea ice or fast ice (Stockwell et al., 1991). The distinct shift in 613C, salt content and the constant sulphur and organic carbon trends of unit c (Fig. 2), suggest a marine environment. However, the 613C values of the Ace Lake core are much lower than those recorded in Ellis Fjord, Organic Lake or Taynaya Bay. Very low concentrations of the methanogen marker compounds indicate that the bottom waters at this time were oxic, but only slightly; they were perhaps disoxic (Volkman et al., 1986). In addition, this sediment horizon in Ace Lake is banded, suggesting that the marine influence was restricted and low energy (similar to Ellis Fjord) with the lower t~13C values indicating that the system was not as productive (Bird et al., 1991). It is likely that the marine source was tidal with water entering only at high tides. The presence of intermixed diatom and sandy lamina in the sediments has been shown to be suggestive of a tidal input (Bronge, 1989). However, grain size analysis and X-ray examination of the Ace Lake core were not carried out before the core was initially thawed, so the presence or absence of intermixed lamina could not be determined. The 1~13Cvalues in the core rise slightly at ~ 6400 ~4C yr B.P. (95 cm; Fig. 2), which may indicate the transition from a marine environment to a restricted marine inlet. This change is also evident in the diatom stratigraphy. There is a distinct increase in the concentration of the pioneer species and a decrease in the open marine species in unit c (Fig. 5). This change is similar to the results of Wasell and Hakansson (1992), who found that Skua Lake had three distinct zones of diatoms: marine, brackish and fresh. The brackish zone was signified by an increase in Nitzschia species, in particular N. curta and N. cylindrus, which show a preference for shallower regions. There was however, no increase in the sulphur and organic carbon content in Ace Lake nor in the 613C values at this level and so, evidence of a prolonged marine connection is lacking (Fig. 2). These physical parameters increased at 5500 14C yr B.P. (35-40 cm; Fig. 2) and are taken to indicate the separation of the lake from its marine source. There is also a definite break shown in the diatom stratigraphy, as E. antarctica was not sampled

83

after 5500 14C yr (Fig. 4). The emergence of the lake could have been due to continuing isostatic uplift. The diatom and physical changes in this segment of the core are interpreted as a tidal marine input to the lake which took place over a minimum of 1200 years. A full marine incursion would have disturbed the meromixis of the lake and made it oxic. However, the sulphur levels in the core are equivalent to those of Ellis Fjord (Bird et al., 1991) which has anoxic bottom waters. Therefore it is suggested that there were deep anoxic pockets, similar to those seen in Ellis Fjord today or, that the bottom of the lake had very low oxygen as indicated by the low abundance of methanogen marker compounds (Volkman et al., 1986). That the sediments remain laminated indicates that the incursion was restricted and low energy, perhaps entering from a protected fjord, until the lake was last isolated about 5500 years ago. Unit b (35-20 cm." 5500~3800 14C yr B.P.) Over the 1700 years that these sediments were laid down, the organic carbon and sulphur in the core increased and the water and salt content dramatically decreased (Bird et al., 1991 ). A dramatic increase in the abundance of the acyclic isoprenoids by more than 20 fold occurred in this section of the core (Volkman et al., 1986). This indicates methanogenesis became an important process in the bottom waters of the lake. These data suggest that the lake became more stabilised, fresh and bottom waters became strongly anoxic (Fig. 2). The diatom species in the core reflect this change with an increase in the freshwater species and a decrease in the pioneer species upcore (Fig. 5). The formation and deepening of a fresher water lens at the surface of the lake and the stratification of the water column would promote these conditions. A similar change from marine to freshwater flora was also observed in Skua Lake when it became isolated 3080-3260 I4C y BP (Wasell and Hakansson, 1992). As the salinity of the surface waters of Ace Lake today is about 14%0 (Burch, 1988), it is suggested that the melt water input into the lake has decreased with time, with flushing decreasing then ceasing completely. This decrease in input may have allowed the lake

84

N P. Fu([brd-Smith. L~L. A'ikes/Pahwogeograp/ly, Pa[aeoc/inlatoh~Kv. Palaeoecology [24 (1996) 73 86

level to decrease through evaporation to the present level of 2 m below sill height. If this is correct it suggests the precipitation of the region may have decreased during the past 8000 years. Alternately, the drainage pattern of the region may have changed, channelling less meltwater to the lake's drainage basin and reducing the freshwater flowing into Ace Lake.

Unit a ( 2 0 - 0 cm: 3 8 0 0 ~ 5 0 0 ~4(, vr B.P.

Ace Lake appears to have stabilised at the present conditions about 3800 years ago. The physical and organic chemical parameters remain at the same levels as in the transitional unit below (unit b), but the assemblages are distinct. The overall diatom stratigraphy suggests that the lake reached an equilibrium during this period. In this unit there is an increase in the freshwater species over the levels in unit b, the marine species disappear, and only low numbers of pioneer species are present ( Fig. 5). Burch (1988) suggested that these pioneer species could be associated with the benthic mats of the lake, which conforms with the observation of Bird et al. (1991), that laminated algal material, possibly from prostrate algal mats, occurred in the sediment at this depth. The large concentration of the fresh-brackish species, Stauroneis sp. (Fig. 4), indicates its establishment as part of the lake's plankton assemblage rather than a meltwater immigrant like P. microstauron. This suggests the formation of a fresher water lens at the surface of the lake indicating increased input of meltwater into the lake or decreased evaporation. At 3800 14C yr B.P., P. microstauron was 8% of the total frustule number, with this percentage decreasing to 1% 400 years ago, suggesting that a decrease in meltwater input into the lake is more likely. However, there was an increase in the water level of the lake of 1 m in the period of 1978 1982 (Pickard et al., 1983) and of 24 cm from 1985 to 1989 (J. van den Hoff, pers. comm., 1992). These recent water level changes indicate there have been at least two changes in the short term climatic history of the region, as the lake could not have increased 1 m every four years for the past 3800 years,

5. Conclusions A study of diatom assemblages, combined with the results of physical parameters of Bird et al. (1991) indicate that Ace Lake was first isolated from the continental ice more than 9000 years ago. This date is earlier than the 7000-8000 years suggested by Adamson and Pickard, (1986). Our results suggest Ace Lake was subject to two main influences during its formation: a melt water input from retreating continental ice and a marine influence ~¥om the open ocean. We propose that the relative importance of these competing influences caused five distinct stages in the evolution of the lake which we have designated units a-e. The dual marine and lacustrine influence is evident in the diatom stratigraphy and unlaminated sediment of the bottom-most unit; unit e. With isostatic uplift the lake is believed to have become isolated, decreasing the mixing capability of the marine influence via the mechanism proposed by Adamson and Pickard (1986). As the marine influence decreased, freshwater input flushed the lake over the course of approximately 1700 years (~8400 6700 14C yr B.P.) and the lake became stable and meromictic (unit d). Sea level maxima in this area is estimated to have occurred 5000 6000 years ago (M. Bird, pers. comm., 1992), and we suggest that seawater flooded over the 2 m sill into Ace Lake and disturbed the freshwater meromixis. The lake was again a marine basin for approximately 1200 years, from 6700-5500 ~4C yr B.P. (unit c). The sediments laid down in this period were laminated and sulphur was present indicating the marine input was restricted. 8t3C data indicates that production in the lake at this stage was at its peak, but less than that found in Ellis Fjord and other lakes of the region at that time (eg. Organic Lake: Bird et al., 1991 ). Isostatic uplift caused re-isolation of the lake at approximately 5500 years ago, which was followed by about 1700 years of slow flushing to a stable meromictic basin, with the lake developing a fresher surface layer through melt water input ( unit b). The diatom and geochemical data indicate that the lake has been stable and meromictic for approximately the past 4000 years (unit a). The decrease in the P. microstauron data seen

S.P. Fulford-Smith, E.L. Sikes/Palaeogeography, Palaeoclimatology, Palaeoecology 124 (1996) 73-86

in unit a over approximately 3800 years could be indicative of a decreasing meltwater input because the species is introduced into the lake by meltwater streams. The levels of this species in the diatom assemblages in other lakes is potentially useful as a regional meltwater indicator. The local influences of marine incursions and freshwater flushing masked any record of other climate changes that were occurring at the time of the formation of Ace Lake. The process of the evolution of the lake does, however, give an indication of the broad scale climatic and geological changes that were affecting Long Peninsula and the Vestfold Hills in general. In particular, the presence of a marine incursion at 6700-5500 14C yr B.P. in Ace Lake indicates that eustatic sea level rise overtook the isostatic rebound in this area for that time period.

Acknowledgments We thank Dr. Michael Bird for all the sediment samples and dates used in this study and also Dr. Andrew McMinn, who helped with the initial selection of the diatom species and gave advice on the structuring of the initial sediment experiments. Harry Burton was responsible for initiating the research into Pyramimonas gelidicola scales in Ace Lake, from which this study evolved. Thanks also to Naomi Parker and Lisette Robertson for drafting assistance. We would also like to thank Dr John Pickard for his review of the manuscript. This work was funded by a Co-operative Research Centre student grant in the Natural Variability subprogramme of the Antarctic CRC.

References Adamson, D.A. and Pickard, J., 1986. Cainozoic history of the Vestfold Hills. In: J. Pickard (Editor), Antarctic Oasis. Academic Press, Sydney, pp. 63-97. Bayly, I.A.E. and Burton, H.R., 1987. Vertical distribution of Paralabidocera antarctica (Copepoda: Calanoida) in Ace Lake, Antarctica, in Summer. Aust. J. Mar. Fresw. Res., 38: 537-543. Bird, M.I., Chivas, A.R., Radnell, C.J. and Burton, H.R., 1991. Sedimentological and stable-isotope evolution of

85

lakes in the Vestfold Hills, Antarctica. Palaeogeogr. Palaeoclimatol. Palaeoecol., 84: 109-130. Bronge, C., 1989. Holocene climate fluctuations recorded from lake sediments in Nicholson Lake, Vestfold Hills, Antarctica. Univ. Stockholm Dep. Phys. Geogr. Res. Rep., 74, 21 pp. Burch, M.D., 1988. Annual cycle of phytoplankton in Ace Lake, an ice covered, saline meromictic lake. Hydrobiologia, 165: 59-76. Burton, H.R., 1981. Chemistry, physics and evolution of antarctic saline lakes: a review. Hydrobiologia, 82: 339-362. Burton, H.R. and Barker, R.J., 1978. Sulfur chemistry and microbiological fractionation of sulfur isotopes in a saline antarctic lake. Geomicrobiol. J., 1: 329-340. Fitzsimons, S.J. and Domack, E.W., 1993. Evidence for Early Holocene deglaciation of the Vestfold Hills, east Antarctica. Polar Rec., 29(170): 237-240. Fryxell, G.A., 1989, Marine phytoplankton at the Weddell Sea ice edge: seasonal changes at the specific level. Polar Biol., 10: 1-18. Hand, R.M. and Burton, H.R., 1981. Microbial ecology of an antarctic saline meromictic lake. Hydrobiologia, 82: 363-374. Johnsen, G. and Nost Hegseth, E., 1991. Photoadaption of sea-ice microalgae in the Barents Sea. Polar Biol., 11: 179-184. Leventer, A. and Dunbar, R.B., 1988. Recent diatom record of McMurdo Sound, Antarctica: Implications for history of sea ice extent. Paleoceanography, 3: 254-274. Mancuso, C.A., Franzmann, P.D., Burton, H.R. and Nichols, P.D., 1990. Microbial community structure and biomass estimates of a methanogenic antarctic lake ecosystem as determined by phospholipid analysis. Microb. Ecol., 19: 73-95. Omoto, K., 1983. The problem and significance of radiocarbon geochronology in Antarctica. In: R.L. Oliver et al. (Editors), Antarctic Earth Science. Aust. Acad. Sci., Canberra, pp. 450-452. Pichon, J.J., Labeyrie, L.D., Bareille, G., Labracherie, M., Duprat, J. and Jouzel, J., 1992. Surface water temperature changes in the high latitudes of the southern hemisphere over the last glacial-interglacial cycle. Paleoceanography, 7: 289-318. Pickard, J., 1983. Surface lowering of ice-cored moraine by wandering lakes. J. Glaciol., 29:338 342. Pickard, J., 1986. The Vestfold Hills: A window on Antarctica. In: J. Pickard (Editor), Antarctic Oasis. Academic Press, Sydney, pp. 333 351. Pickard, J., Selkirk, P.M. and Selkirk, D.R., 1983. Holocene climates of the Vestfold Hills, Antarctica, and Macquarie Island. In: SASQUA Int. Symp. 29 August-2 September, Swaziland. Pickard, J., Adamson, D.A. and Heath, C.W., 1986. The evolution of Watts Lake, Vestfold Hills, East Antarctica, from marine inlet to freshwater lake. Palaeogeogr. Palaeoclimatol. Palaeoecol., 53: 271-288. Schmidt, R., Mausbacher, R. and Muller, J., 1990. Holocene diatom flora and stratigraphy from sediment cores of two

86

S.P. I~)~!/brd-&nith, E.L. Sikes/Palaeogeography, Pal~woclimatology, Palaeoecology 124 (1996) 73 -86

antarctic lakes (King George lsland i. ,I. Palaeolimnol.. 3:55 74. Stockwell, D.A., Kang, S.-H. and Fryxell, G., 1991. Comparisons of diatom biocoenoses with Holocene sediment assemblages in Prydz Bay, Antarctica. Proc. ODP Sci. Results, 119: 667. Van den Hoff, J., Burton, H,R. and Vesk, M., 1989. An encystment stage, bearing a new scale type, of the antarctic prasinophyte t~vramimonas gelidicola and its palaeolimnological and taxonomic significance. J. Phycol., 25: 446-454. Volkman, J.K., Allen, D.I.. Stevenson, P.L. and Burtom H.R,, 1986. Bacterial and algal hydrocarbons in sediments from a saline antarctic lake, Ace Lake. Organic Geochem.. 10: 671 681.

Volkman, J,K., Burton, H.R., Everitt, D.A. and Allen, D.I., 1988. Pigment and lipid compositions of algal and bacterial communities in Ace Lake. Vestfold Hills, Antarctica. Hydrobiologia, 165:41 57. Wasell, A. and Hakansson, H., 1992. Diatom stratigraphy in a lake on Horseshoe Island, Antarctica: A marine--brackish freshwater transition with comments on the systematics and ecology of the most common diatoms. Diatom Res., 7:157 194. Wharton Jr., R.A., Parker, B.C. and Simmons Jr., G.M., 1983. Distribution, species composition, and morphology of algal mats in Antarctic Dry Valley Lakes. Phycologia, 22:355 365.