Halogen geochemistry of the McMurdo dry valleys lakes, Antarctica: Clues to the origin of solutes and lake evolution

Geochimica et Cosmochimica Acta, Vol. 69, No. 2, pp. 305–323, 2005 Copyright © 2005 Elsevier Ltd Printed in the USA. All rights reserved 0016-7037/05 $30.00 ⫹ .00

doi:10.1016/j.gca.2004.06.040

Halogen geochemistry of the McMurdo dry valleys lakes, Antarctica: Clues to the origin of solutes and lake evolution W. BERRY LYONS,1,* KATHLEEN A. WELCH,1 GLEN SNYDER,2 JOHN OLESIK3 ELIZABETH Y. GRAHAM,4 GILES M. MARION5 and ROBERT J. POREDA6 1

Byrd Polar Research Center, The Ohio State University, Columbus, OH 43210-1002, USA 2 Department of Geological Sciences, Rice University, Houston, TX USA 3 Microscopic and Chemical Analysis Research Center, The Ohio State University, Columbus, OH 43210, USA 4 Department of Geological Sciences, University of Alabama, Tuscaloosa, AL 35487, USA 5 Desert Research Institute, Reno, NV 89512, USA 6 Department of Earth and Environmental Sciences, University of Rochester, Rochester, NY 14627, USA (Received October 30, 2003; accepted in revised form June 30, 2004)

Abstract—We have determined the halogen and boron concentrations in the ice-covered lakes of Taylor Valley, Antarctica, to better establish the sources of salts and evolutionary histories of these unusual water bodies. In addition, we report on a series of 129I measurements that were compared with previous 36Cl data that also help constrain the source of solutes and histories of the lakes. The new data, when put into context of previous work on these systems over the past forty years, allow us to make the following conclusions. The primary source of solutes to Lake Hoare, the youngest of the lakes, is the dissolution of marine aerosols and aeolian salts and the chemical weathering of dust on the glaciers. The geochemistry of Lake Fryxell, the brackish water lake, is primarily dominated by the diffusion from a halite-saturated brine at the sediment-water interface and the recent infilling of the lake by glacier meltwater. These waters have chemical weathering and marine aerosols components. Lake Bonney has two distinct lobes whose hypersaline hypolimnia have different chemistries. Both of the lobes are remnants of ancient marine waters that have been modified by the input of weathering products. This lake has also been modified by periods of cryogenic concentration when solutes have been lost via mineral precipitation. Thus the geochemistry of Lake Bonney owes its unusual geochemistry, in part, to variations in the climate in the Taylor Valley over at least the past 300kyr. The 129I data from the Taylor Valley are similar to those from fracture fluids in crystalline rocks from the Northern Hemisphere. Copyright © 2005 Elsevier Ltd 1964; Hendy et al., 1977; Matsubaya et al., 1979; Lyons et al., 1998b, Poreda et al., 2004). Stream flow from the glaciers currently occurs for 4 –10 weeks during the austral summer. During periods of cooler summers, stream flow diminishes. If cool conditions last for long periods, eventually the ice covers on the lakes are lost through sublimation, the lake volumes are substantially decreased, and the salinity increases. When the climate ameliorates and austral summers are milder, the stream flow increases and the ice-covers are re-established as the lake volumes increase. These drawdown-refill cycles have apparently occurred many times over at least the past ⬃300,000 yrs (Hall and Denton, 2000; Hendy, 2000) and may have taken place over even a longer time period (Barrett and Hambrey, 1992). Lake Bonney and Lake Fryxell in Taylor Valley (Fig. 1) and Lake Vanda in Wright Valley represent the consequences of at least one such drawdown event, in that the salinities in their hypolimnia are greatly elevated with respect to the surface waters. Therefore, it is now clear that the strongly stratified density and chemical profiles observed in some of the lakes are produced by climatic variations. The amelioration of the climate over the past ⬃1 kyr has led to a substantial refilling of the majority of these lakes with freshwater (Wilson, 1964; Lyons et al., 1998a, 1998b). Although our overall understanding of the role of climate on hydrologic processes in the dry valleys is at a much better stage of development than in the 1960s, the questions of solute sources and a more detailed understanding of the chemical evolution of these unusual lakes remain unanswered. In syn-

1. INTRODUCTION

The geochemistry of the ice-covered lakes of the McMurdo Dry Valleys region of Antarctica (henceforth referred to as “the dry valleys”) has been investigated for over forty years. These lakes are unusual in that over a relatively small geographic area, the geochemistry of the lakes varies dramatically from entirely freshwater (i.e., Lake Hoare) to the hypersaline hypolimnia of Lake Bonney. In addition to the differences in salinity, the lakes have very different major ionic ratios as well (Lyons et al., 1998a). The earliest investigations of these systems noted the diversity in their chemistries, and speculated on the origins of these differences (Angino et al., 1962; Angino and Armitage, 1963). Recent work has utilized more sophisticated analytical techniques and approaches than the pioneering work of Angino and his co-workers, yet the ultimate explanation of the sources of solutes to the lakes and geochemical evolution of the saline waters is still enigmatic (Wilson, 1979; Green et al., 1988; Lyons and Mayewski, 1993; Lyons et al. 1998a, 1998c). What is agreed upon is that climatic variations affecting the hydrologic cycle within the valleys have greatly influenced the amount of freshwater flow entering into these closed-basin systems. These variations in freshwater input through time have greatly affected the dynamics of these lakes, thereby changing their size and salinity through time (Wilson, * Author to whom correspondence should be addressed (lyons.142@ osu.edu). 305

306

W. B. Lyons et al.

Fig. 1. Location map.

thesizing all the previous work cited above, four major sources of solutes to these lakes have been proposed: 1) hydrothermal; 2) remnant seawater; 3) glacier melt and precipitation; 4) a combined glacier melt-precipitation and chemical weathering source. Of these four possibilities, even after 40 yr. of investigation, none of these can be eliminated conclusively (Lyons and Mayewski, 1993), but our most recent work suggests the first one is highly unlikely (Poreda et al., 2004). In this paper, we present both chemical and isotopic measurements from the lakes of Taylor Valley, including halogens and boron, and combine these measurements with previous data to provide the most conclusive evidence to date on the geochemical origins of these lakes. In addition, this is the first investigation to utilize iodine-129 to determine the sources of halides into the lakes of the dry valleys. Until now, 129I investigations of brines in the polar regions have been limited to fracture fluids emplaced in the Canadian Shield and northern Europe during previous ice ages (Fabryka-Martin et al., 1989; Bottomley et al., 2002; Starinsky and Katz, 2003). This study complements work of the past decade, where the environmental isotope 36Cl was employed to determine salinity sources to the lakes in the dry valleys (Carlson et al., 1990; Lyons et al., 1998c) and also provides direct evidence of the active processes leading to the observed 129I/I and 36Cl/Cl ratios in polar

regions. This multi-faceted approach provides compelling evidence that the age of the lakes, their location within the landscape, and their overall “sensitivity” to climate change have all influenced their isotopic composition and geochemical evolution. 2. STUDY AREA AND LAKE DESCIPTION

The McMurdo Dry Valleys (77°40’S, 163°E) are the largest ice-free region in Antarctica. Even though the environment in the dry valleys is considered to be a polar desert, with a mean annual temperature of ⫺18°C and a precipitation rate of ⬍10 cm yr⫺1 water equivalent, currently there is sufficient glacier melt to produce stream flow into the lakes. In Taylor Valley there are three major lakes, Lake Bonney, Lake Fryxell and Lake Hoare (Fig. 1). Lake Bonney has two lobes separated by a 13 m deep sill. Since 1993, Taylor Valley has been the location of the McMurdo Dry Valleys (MCM) Long-Term Ecological Research (LTER) site supported by the National Science Foundation, where meteorological, glaciological, hydrological, biogeochemical and ecological data are collected on a regular basis. This collection of long-term data has allowed researchers to better establish the ecological response to climatic forcing, especially in the lake systems (Doran et al.,

Halogens in Antarctic Dry Valley Lakes

2002). The establishment of the LTER has also allowed more comprehensive investigation of the geochemistry of these lakes. This recent geochemical work includes a series of isotopic measurements and trace element determinations (Lyons and Welch, 1997;; Lyons et al., 1998c, 1999, 2003; Neumann et al., 1998; Poreda et al., 2004) that have all supported the earlier research demonstrating significant geochemical differences between the lakes. The interpretations of the available geochemical data include the following: Lake Hoare is the “youngest” of the lakes having begun to fill at the onset of a warming event beginning at ⬃ 1ka. The chemistry of Lake Hoare indicates no evidence of a previous drawdown either by an enriched ␦D or ␦18O signal or higher salinity at depth (Lyons et al., 1998a, 1998b). In addition, the surface 36Cl isotopic ratios are similar to the deeper waters (Lyons et al., 1998c). Lake Fryxell has a brackish water hypolimnion. Its current salinity profile is controlled by diffusion and it has been generated from the increased inflow of fresh water beginning ⬃1 ka when the lake was essentially a hypersaline playa (Lyons et al., 1999b). The two lobes of Lake Bonney have the most complex histories. The west lobe did not undergo the drastic drawdown from 3ka to ⬃1ka that the other lakes in both Taylor and Wright Valleys experienced (Hendy et al., 1977; Matsubaya et al., 1979; Poreda et al., 2004). It has maintained its ice cover for some undetermined period of time, but it is probably at least a remnant of Glacial Lake Washburn, the large lake that filled the entire Taylor Valley during the Last Glacial Maximum (LGM) into the early Holocene (Hall and Denton, 2000). Poreda et al. (in press) have suggested, based on helium isotopic data, that the west lobe of the lake could be much older than this. The east lobe did lose its ice cover as the ␦D and ␦18O of the hypersaline hypolimnion are extremely enriched, demonstrating a strong evaporation signal (Matsubaya et al., 1979). Before ⬃2.7 ka the east and the west lobes of the lake were not connected (Poreda et al., 2004 ). At that time, water began to flow from the west lobe into the east lobe as the water level rose. The permanent ice-cover on the east lobe did not return until ⬃300 years ago (Poreda et al., 2004). Lake Vanda, in Wright Valley, shows a similar general history, with a high stand at ⬃2.7 ka (Smith and Friedman, 1993) and a hypersaline, low stand sans ice-cover at ⬃1 ka (Wilson, 1964). At the terminus of the Taylor Glacier, west of Lake Bonney, a saline discharge termed Blood Falls exists. The relationship of Blood Falls to Lake Bonney has been speculated (Black et al. 1965). Its current discharge into Lake Bonney is thought to be very important to the chemical mass balance of the lake (Lyons et al., 1998c). 3. MOST RECENT GLACIAL HISTORY OF TAYLOR VALLEY

307

2000; Hall et al., 2000; Hendy, 2000). The WAIS probably advanced during earlier glacial periods as well. Bonney Drift, which represents the advance of the Taylor Glacier, contains dated lacustrine carbonates, interpreted to have been deposited in shallow, proglacial lakes that are dated from 70 to 130, 160 –240, 270 –330 and older than 400 ka (Hendy, 2000). The ␦18O of these carbonates indicates lake water ␦18O values from ⫺34.4 to ⫺42.9‰; or essentially water derived primarily from Taylor Glacier (Higgins et al., 2000b). This Bonney Drift is overridden by glaciolacustrine facies of the Ross Sea Drift in the current location of the Canada Glacier (Fig. 1) (Higgins et al., 2000b; Hall and Denton, 2000). Exposure ages of morainal materials in Arena Valley to the southeast of Taylor Valley generally corroborate this work, as they document Taylor Glacier advances at 208 ⫾ 67 and 335 ⫾ 187 ka (Brook et al., 1993). In addition, another moraine yields an age of 1.2 ⫾ 0.2 Ma, but this is considered a minimum age due to potential diffusional loss of helium (Brook et al., 1993). A somewhat older age would be in accord with moraines located higher on the slopes of Taylor Valley that date to ⬎1.5 Ma (Higgins et al., 2000b). Presently, the Taylor Glacier is thought to be at its most extensively advanced position during the Holocene. Ice from the Ross Sea blocked Taylor Valley from 9.4 to ⬎24 ka, with a maximum extent between 15 and 17.5 ka (Hall and Denton, 2000). Grounded ice retreated from Explorer’s Cove (the seaward edge of Taylor Valley) by ⬃7.4 ka. The blockage of Taylor Valley from the east produced Glacial Lake Washburn, which reached an elevation of almost 300m within the Fryxell basin, with datable delta sediments (i.e., melt stream-lake contacts) as high as 240m in the Fryxell basin, and 314m in the Bonney basin (Hall et al., 2000). Hendy (2000) provides very convincing arguments that there have been lakes in Taylor Valley for at least 300 kyr and perhaps much longer, with the interglacial lakes being smaller than the large, entire basin filling, proglacial lakes of the glacial periods. Clearly, the advance and retreat of both the EAIS and the WAIS have been instrumental in the rise and fall of lacustrine systems in Taylor Valley. There is evidence of fjord–like conditions in Wright Valley to the north (a higher elevation) before the Pleistocene, with the most recent deposits dating at 3.9 Ma (Prentice et al., 1993). In addition, Taylor Valley may have contained seawater as cores drilled at the mouth of the valley contain marine deposits of late Miocene and/or early Pliocene age (Porter and Beget, 1981). Higgins et al. (2000a) have proposed that the original source of salt in the Bonney basin is evaporated seawater of Miocene age, introduced when Taylor Valley was also a fjord. 4. METHODS

The glacial geological reconstructions of Taylor Valley for the Pleistocene and Holocene provide strong evidence that the East Antarctic Ice Sheet (EAIS) (represented by the Taylor Glacier) as well as the alpine glaciers in the valley, advanced during interglacial times and retreated during glacial ones (Hendy, 2000; Higgins et al., 2000a, 2000b). It is also clear that the West Antarctic Ice Sheet (WAIS) (represented by Ross Sea Ice) advanced during the LGM and blocked the entrance to Taylor Valley (Hall and Denton,

In November 2002, a series of lake water profiles were obtained from Lake Bonney, Fryxell and Hoare. Samples were collected using standard oceanographic/limnetic techniques in 5L Niskin bottles. Samples were filtered using 0.4 ␮m Nucleopore filters within a few hours of collection, into precleaned polyethylene bottles, using precleaned plastic filtering apparatus (Welch et al., 1996). Because of the differences in salinity and ionic composition among the lakes, different filtering equipment was used for each lake. The samples were stored chilled and dark and transported back to McMurdo Station for analysis. The halogens Cl⫺, Br⫺ and F⫺ were analyzed

308

W. B. Lyons et al.

using ion chromatography (IC). The details of the analytical and cleaning methods are outlined in Welch et al. (1996). The precision of the Cl⫺, Br⫺ and F⫺ measurements is ⫾ 1, 5 and 2%, respectively. The accuracy of the measurements based on the analysis of check standards is ⫾ 1, 1 and 10%, respectively. With the exception of the east lobe of Lake Bonney, iodine measurements were made at the Microscopic and Chemical Analysis Research Center (MARC) at the Ohio State University in February 2003 on the same samples analyzed at McMurdo Station by IC. The samples had been stored at room temperature in the dark until that time. Unlike other halogens, iodine exists in several oxidation states, as iodide and iodate, as well as organic iodine, in natural settings. Total I was measured using a Thermo Finnigan Element 2 ICP-MS on diluted samples with 1% tetramethyl ammonium hydroxide (TMAH). This was added to stabilize the sample and minimize I⫺ vaporization, to reduce washout time and to decrease I sorption onto plastic (Takaku et al., 1995). A standard solution was run repeatedly during the analysis to allow correction for instrumental drift. A number of samples were spiked with a known concentration of I⫺. Most had spike recoveries between 88 and 108%, whereas one from Lake Bonney yielded only 66%. We have not evaluated these measurements for yield, but have assumed a precision of 12% or less. The east lobe of Lake Bonney was sampled for iodine analysis in January 2000 and also analyzed using ICP-MS techniques, but at the University of Rochester. Iodine for 129I analysis was extracted and purified at the University of Rochester Cosmogenic Isotope Laboratory (Fehn et al., 1992) and the extracted samples were sent to Purdue Rare Isotope Measurement Laboratory (PRIME) for carrierfree 129I determinations through accelerator mass spectrometry (AMS) (Sharma et al. 2000). Chloride measurements were made on the same samples using the IC method discussed above. In addition to the halogen data, we also present dissolved B concentrations for samples collected from the lakes in December 1997. These samples were analyzed in July 1999 at the Geochemistry Laboratory, the University of Alabama, using a Perkin Elmer Inductively Coupled Plasma Spectrometer. The precision of the B measurements at 4.6 ␮M was ⬍1%. Chloride measurements were also made on these samples using the same methods discussed above. The Cl⫺ measurements are similar to those observed in November 2002, but there are annual variations in the surface waters (above the chemocline) of the lakes (Welch et al., 2000). The B/Cl ratio presented is for samples that were collected in 1997.

tions vary from a low of ⬃1.5 ␮M in Lake Hoare to ⬃3.5 mM in the east lobe of Lake Bonney. The solute/Cl ratios also vary with depth from lake to lake and reflect either different sources of the solutes to these lakes or different processes that have affected solute concentrations in the lakes, or both (Figs. 3a-d). The Br/Cl profiles are the most uniform for the lakes showing consistency with depth. However, there are slight increases in this ratio below the chemoclines in Lake Fryxell and the east lobe of Lake Bonney (Fig. 3). The ratio of F/Cl in Lake Hoare is essentially constant, whereas the other profiles demonstrate a decrease with depth (Fig. 3). The I/Cl profiles are the most complex with Lake Hoare and the east lobe of Lake Bonney decreasing with depth. The profiles from Lake Fryxell and the west lobe of Lake Bonney go through middepth minima, and then increase below the chemocline to constant values that are essentially equal to the surface values (Fig. 3). Lakes Hoare and Fryxell show a slight increase in the B/Cl ratio with depth, whereas both lobes of Lake Bonney show slight decreases in the B/Cl ratio with depth. In addition to the ion data, a number of lake samples were collected in January 2000 and six were subsequently analyzed for 129I. 129I has a long half-life (15.7 Myr) and has been used to differentiate between sources of I⫺ to aquatic systems as well as to better understand I⫺ cycling (Moran et al., 1998; Snyder and Fehn, 2004). Because of the potential importance of Blood Falls in understanding both the current geochemical mass balance and the past history of the lake, the MCM-LTER team has collected samples emanating from this saline discharge for over a decade. These samples have also been analyzed for their major solute composition to better establish the flux of salt into the west lobe of Lake Bonney.

5. RESULTS

6. DISCUSSION

The Cl⫺, Br⫺, F⫺, I⫺, and B profiles from Lake Bonney, Fryxell and Hoare are shown in Figures 2a, b, c, d and e and the data are tabulated in Appendix A. Although the Cl⫺ profiles from all these lakes (i.e., Lyons et al. 1998a) and the Br profile from Lake Bonney (Lyons et al., 1998c) have been previously reported, the Br data from the other lakes and the F⫺, I⫺ and B data are new. The Br/Cl, F/Cl, I/Cl, and B/Cl ratios vs. depth for each lake are shown in Figures 3a, b, c, and d. The ratios of each of the halogens and B relative to Cl⫺ in seawater are also shown for reference. Normalizing the solute data to Cl⫺ allows us to look at the variations in chemistry in more detail because of the large differences in salinity between lakes and the large salinity gradients within the lakes. The Cl⫺ profiles demonstrate the extreme range of salinities in these lakes with Lake Hoare being fresh water and the hypolimnia of both lobes of Lake Bonney being hypersaline. In general, the halogens and B increase with depth in all the lakes, but like Cl⫺, their concentrations vary dramatically from ⬃2 ␮M in Lake Hoare to ⬃20 mM in the east lobe of Lake Bonney for Br⫺; from ⬃10 ␮M in the surfaces of all the lakes to ⬃1.2 mM in the east lobe of Lake Bonney for F⫺; and from ⬃20 nM in Lake Hoare to 8 ␮M in the west lobe of Lake Bonney for I⫺. The B concentra-

In the following discussion we assess the potential sources of solutes to the Taylor Valley lakes and then assess the importance of the recent geologic history of the Taylor Valley on the chemical evolution of these lakes. 6.1. Halogen and Boron Concentrations in Streams of Taylor Valley The Br⫺ concentrations in the streams of Taylor Valley are mostly below our detection limit of ⬃2 ⫻ 10⫺4 mM. Where we can quantify Br⫺, the average molar ratio of Br/Cl is ⬃0.001 for the Taylor Valley streams. This value is lower than the seawater ratio suggesting that there is not a large source of Br due to chemical weathering in the streams (Appendix B). The F⫺ concentrations from stream samples collected during the 2003– 04 field season range from below the detection limit of ⬃1 ␮M to 8.8 ␮M with an average F/Cl molar ratio of 0.026 (Appendix C). The F/Cl ratio in seawater is much lower than this (Fig. 3), strongly suggesting an additional source of F⫺ other than marine aerosol. The B/Cl ratios in the Lake Fryxell and Lake Hoare streams are 0.0026 (n ⫽ 7) and 0.0030 (n ⫽ 3), respectively, whereas the Bonney basin streams have lower ratios (0.0023, n ⫽ 9)

Halogens in Antarctic Dry Valley Lakes

309

Fig. 2. Concentration vs. depth profiles of a) Cl, b) Br, c) F, d) I, and e) B in the Taylor Valley lakes.

(Appendix D). However, there are no B data from Blood Falls. These ratios are all higher than seawater (Fig. 3) and the surface waters of the lakes generally reflect the stream B/Cl ratios. This also indicates a potential weathering source of B to the lakes. Unfortunately, we have not measured I⫺ in any stream samples from Taylor Valley.

6.2. Halogen and Boron/Chloride Ratios Relative to Seawater in the Lakes The Lake Bonney profiles are slightly enriched in Br relative to Cl, suggesting some precipitation of halite especially in the east lobe of Lake Bonney (Fig. 3) (Lyons et al., 1999c). The

310

W. B. Lyons et al.

Fig. 3. Depth profiles of a) Br/Cl, b) F/Cl, c) I/Cl, and d) B/Cl ratios in the Taylor Valley lakes.

Br/Cl ratio in Lake Hoare is slightly lower than the seawater ratio. The hypolimnia of Lake Bonney have F/Cl very similar to that of seawater, whereas all the other waters are enriched in F relative to Cl compared to seawater. The surface waters of Lake Bonney have F/Cl ratios ⬃10x greater than seawater, whereas Lake Hoare has an enrichment of ⬃100x relative to seawater (Fig. 3b). In Lake Fryxell, the F/Cl is ⬃20x greater than the seawater ratio. The bottom waters of Lake Bonney are within approximately a factor of 4 of the I/Cl ratio of seawater although the east lobe appears to be depleted in I and the west lobe is enriched. Lake Fryxell and Lake Hoare show I enrichments of ⬃10 to 20x relative to the I/Cl ratio in seawater (Fig. 3c). The hypolimnetic waters of Lake Bonney are very similar to the seawater B/Cl ratio, whereas all the other lakes investigated have B/Cl ratios that are slightly (2 to 4x) enriched compared to seawater (Fig. 3d). 6.3. Sources of Halogens and Boron and the Implication of Seawater Ratios 6.3.1. Lake Bonney The similarity of the F/Cl, B/Cl, Br/Cl and I/Cl ratios of the hypolimnetic waters in Lake Bonney to seawater ratios suggest

that the source of solutes was originally seawater (Fig. 3). The seawater source of halogens to the hypolimnia of Lake Bonney also explains the seawater-like 36Cl values of these waters (Lyons et al., 1998c). Closer inspection of the profiles demonstrates that in the west lobe of Lake Bonney, the B, Br⫺ and I⫺ to Cl⫺ ratios are slightly enriched relative to seawater, whereas in the east lobe, Br⫺ is even further enriched but B and I⫺ are depleted. These minor deviations from the seawater ratios likely reflect past and present processes occurring within the lake. We have previously interpreted the enrichments of the Br/Cl ratios relative to seawater to indicate previous removal of some Cl⫺ as halite or hydrohalite. The sediments in the east lobe of Lake Bonney have NaCl salts associated with them and the previous interpretation of the 36Cl data also suggest that some Cl⫺ removal would have previously occurred thereby increasing the Br/Cl ratio in the residual brine (Lyons et al., 1999). The profiles suggest that B and I⫺ may have been lost as the brine cryoconcentrated beyond halite saturation. Interestingly, F⫺ is also slightly enriched relative to the seawater F/Cl ratio. The lack of F⫺ removal in the most saline waters probably results from the relatively low Ca2⫹ activities in these waters, which has been used to explain the quasi-conservative

Halogens in Antarctic Dry Valley Lakes

behavior of F⫺ in other hypersaline systems (Jones et al., 1977; Levy et al., 1999). As mentioned above, the west lobe of Lake Bonney may be very old, whereas the east lobe has clearly undergone a recent drawdown-refill event. It is also significant that the east lobe of Lake Bonney has undergone a concurrent loss of iodine and boron, relative to chloride. The west lobe of Lake Bonney did not undergo this event, as its ␦D and ␦18O signatures are similar to that of the Taylor Glacier ice showing no evaporative signal (Matsubaya et al., 1979). The slight enrichment of I/Cl at depth in the west lobe may reflect input via decomposition of organic matter below the chemocline as is observed in the ocean (Chester, 2003). The west lobe of Lake Boney and Lake Fryxell are the two most biologically productive lakes in Taylor Valley (Priscu et al., 1999) and both show an increase of I⫺ below the chemocline and their deep chlorophyll maxima. These enrichments must be related to biogeochemical cycling of I⫺ within these lakes. 6.3.2. Lake Fryxell and Lake Hoare The Br/Cl of Lake Fryxell, as noted above, is very similar to seawater. However, all the other solute to Cl ratios investigated are enriched relative to seawater. Lake Fryxell was essentially a playa with a shallow (⬍ 1m) halite-supersaturated brine at the terminus of the last drawdown event (⬃1 ka) (Lyons et al., 1998a). As the lake has refilled since this time, the ␦D and ␦18O of the water indicate modern meteoric water as its source, strongly suggesting that although there was salt within the lake (as evidenced by its brackish water TDS and solute profiles), there was little water. Diffusion of solutes such as Cl⫺ as well as constituents such as DOC occurs from the surficial sediments (Aiken et al., 1996, Lyons et al., 1998b). The enrichments of F⫺, I⫺ and B relative to Cl⫺ in Lake Fryxell are likely to be related to the input of weathering products from the lake basin as the stream flow into the lake increased. Lake Fryxell has the most streams associated with it of any Taylor Valley lakes (Fig. 1) and these streams contain the highest concentration of weathering produced solutes of all the streams in the dry valleys (Nezat et al., 2001). Over the past few years it has been clearly demonstrated that chemical weathering within the hyporheic zones of the streams in the Taylor Valley is a major process contributing solutes to the lakes (Lyons and Welch, 1997; Nezat et al., 2001; Maurice et al., 2002; Gooseff et al., 2002; Lyons et al., 2003). Therefore the most logical explanation for the sources of solutes in Lake Fryxell is the following: as the WAIS moved into Taylor Valley at the height of the LGM it pushed seawater into the valley. Freezing of the lake was accompanied by an increase in Br:Cl, as Glacial Lake Washburn shrunk to its last playa stage (⬃1 ka). This increase was subsequently countered by dilution from streams, which returned Br:Cl ratios to near marine values in Lake Fryxell (Starinsky and Katz, 2003). The dilution process also resulted in a dramatic increase in Na:Cl ratios because of an increased Na input from chemical weathering of the stream channel sediments (Nezat et al., 2001). As stated above, recent work suggests that Lake Hoare is only ⬃1 kyr old (Lyons et al., 1998b). The enrichments of the halogens, F⫺ and I⫺, and B to Cl⫺ relative to seawater suggest input from other sources besides the melting of primary pre-

311

cipitation. We hypothesize that the majority of the “excess” F(i.e., above the F/Cl seawater ratio) entering the lakes is derived either from the weathering of minerals within the streams or from the dissolution of dust upon the glacier surfaces as they melt in the austral summer (Lyons et al., 2003). There is no evidence of F⫺ enrichment above the seawater ratio during marine aerosol formation (Duce and Hoffman, 1976). There is also little evidence for B enrichment in marine precipitation (Fogg and Duce, 1985), but recent work suggests that gas phase fractionation, dust and volcanic input could lead to B/Cl ratios greater than marine values in aerosols (Park and Schlesinger, 2002). Boron minerals have been observed in ice-free regions of interior Antarctica (Fitzpatrick et al., 1990), although none has been specifically identified in Taylor Valley (Keys and Williams, 1981). However, the global boron flux into the ocean is dominated by chemical weathering input (Lemarchand et al., 2000) and specific boron-rich mineral phases are not needed to explain the B enrichment observed in the lakes. As noted above, the B/Cl ratio in the streams of Taylor Valley is also greater than the seawater ratio, indicating a rock-weathering source of B. It is possible that F⫺ and B rich mineral dust is blown onto glacier surfaces, especially during the very strong katabatic wind events that occur frequently in the austral winter. During the short melt season in the austral summer, solutes are transferred into the streams and then transported into the lakes. This process is an important one in controlling Ca2⫹, HCO3⫺ and SO42⫺ concentrations (Lyons et al., 2003). Recent work has also demonstrated that temperate rivers have I/Cl that can be much greater than the seawater value (Moran et al., 2002). The lowest observed ratios are associated with the highest Cl⫺ concentrations (Moran et al., 2002). The source of the “excess” I⫺ in these systems is primarily thought to be from soil or sedimentary rock– derived organic matter. Because there is limited hydrologic exchange between the soils and the lakes in Taylor Valley and there is little sedimentary rock associated with the stream sediments, these sources of I⫺ to the lakes are minimal to non-existent. The I⫺ content of igneous and metamorphic rocks, the primary component of the Taylor Valley floor, is small and hence not an available source (Muramatsu and Wedepohl, 1998). The most likely source of “excess” I⫺ to the Taylor Valley system is from the atmosphere. Aerosol enrichments of 100 to 1000 are not unusual and I/Cl ratios from 750 to 28,000 times enrichment above the seawater ratio have been measured in Antarctica (Duce and Hoffmann, 1976). The I/Cl ratios in Lake Hoare are ⬃10x that of seawater suggesting that aerosol enrichment rather than chemical weathering can explain them. 6.4. Sources of Solutes to Lake Bonney 6.4.1. Seawater as the major source As mentioned above, the hypolimnetic waters in Lake Bonney have B, Br, and F/Cl ratios fairly similar to seawater. Boswell et al. (1967) were the first to propose that Lake Bonney deep waters were of marine origin. A plot of Na/Cl vs. Br/Cl of Lake Bonney deep waters also suggests that these waters evolved from seawater (Fig. 4). There is geologic evidence that a deep fjord existed in Taylor Valley between 6 and 7 Ma (Elston and Bressler, 1981). At 4.6 –5.1 Ma the central and

312

W. B. Lyons et al.

6.4.2. Blood Falls and its role in Lake Bonney geochemistry Blood Falls was first described by Black et al. (1965) as “a striking reddish-yellow ice cone built from a saline discharge at the terminus of the Taylor Valley.” The color was due to abundant iron oxides within the frozen salty ice (Black et al., 1965). These authors recognized that the chemistry of the discharge changed with time and pondered as to whether the discharge was adding salts to Lake Bonney, or whether the saline deposit was instead derived from the lake. Black and Bowser (1967) were unable to constrain the sources of salts but again pointed out that the saline discharge changes with time, being NaCl rich initially, and then “aging” to a more Na2SO4

Fig. 4. Freezing and evaporation trajectories for brines, showing modification of seawater (after Herut et al., 1990). Taylor Valley lake samples are comparable with brines from the Fennoscandian shield (Starinsky and Katz, 2003) and the Canadian shield (Bottomley et al., 1999).

western portion of the valley experienced partial emergence, but there is strong evidence that the sediments in the eastern portion of Taylor Valley were still dominated by marine deposition until 1.8 to 1.7 Ma (Elston and Bressler, 1981; Ishman and Rieck 1992). During the development of Glacial Lake Washburn during the LGM, any marine waters present in the sediments of the central portion of the valley were flushed (Stuvier et al., 1981), but it appears that the western end of the valley may have maintained its marine salt accumulation. Previously published Cl isotope data (Lyons et al., 1998c, 1999) and the data presented here strongly support the notion that the solutes in the hypolimnia of both lobes of Lake Bonney were initially derived from a marine source. The geologic history of the region suggests that these salts may have become isolated from their oceanic source between 1.7 to 5.1 Ma. The connection between the production of cryogenic brines derived from seawater and Pleistocene glaciation has been argued for many locations in the polar regions and in the northern temperate regions of North America (Bein and Arad, 1992; Starinsky and Katz, 2003). Based on the conceptual models put forth by these authors, an advancing East Antarctic Ice Sheet during a previous interglacial time could have depressed the crust sufficiently so that the seawater could infiltrate into a marginal trough between the ice edge and the ocean. As mentioned above, the EAIS has advanced toward the Ross Sea many times during interglacials of the Pleistocene (Hendy, 2000). In addition, the sedimentary record obtained via the Dry Valley Drilling Project (DVDP) suggests that the Taylor Glacier advanced into the Lake Hoare region through the late Neogene (Ishman and Rieck, 1992). Unfortunately the age of the solutes in Lake Bonney is unknown (see below), so we are unable at this time to constrain the exact time of marine isolation in western Taylor Valley.

Fig. 5. a: Blood Falls major ion data showing the variations in concentrations and ratios over time.

Halogens in Antarctic Dry Valley Lakes

313

Fig. 5. b: Blood Falls major ion data plotted vs. Cl and Ca. The solid line represents the seawater ratio except in the case of SO4 vs. Ca where it represents the 1:1 line.

dominated chemistry. Since the onset of the MCM-LTER we have measured the geochemistry of the major stream draining Blood Falls. The Cl⫺ concentrations and the Na/Cl, Ca/Cl and SO4/Cl ratios plotted through time since December, 1993, are shown in Figures 5 a and b. Our results agree strongly with what Black and his co-workers observed in the 1960s. The higher Cl⫺ concentrations (highest TDS) are associated with the lowest Na/Cl, Ca/Cl and SO4/Cl ratios, with a Na/Cl ratio essentially that of seawater (Fig. 5a). The lower TDS samples are associated with the highest ion/Cl ratios, suggesting salts such as NaSO4●10H2O, CaSO4●2H2O and perhaps, CaCO3 are dissolving. This so-called “aging” effect was the same as observed by Black et al. (1965) and Black and Bowser (1967). There are a couple of lines of evidence to suggest that the source of Blood Falls dissolved load is marine. The Na/Cl and Mg/Cl ratios at essentially all Cl⫺ concentrations plot on the

seawater line (Fig. 5b). At Cl⫺ values less than 250 mM, the K⫹ values also plot on the seawater line (Fig. 5b). Secondly, the 36Cl values from Blood Falls are essentially seawater values (i.e., “dead”) (Lyons et al., 1998c). Because the stable isotopic geochemistry of Blood Falls is very similar to that of the hypolimnion of the west lobe of Lake Bonney (http://huey. colorado.edu), we assume that they are related. Currently Blood Falls is supplying solutes to Lake Bonney, but like Black et al. (1965), we hypothesize that Blood Falls, and the probable saline deposits under Taylor Glacier, are actually a frozen portion of a larger saline lake, which we will call, Greater Lake Priscu. The age and origin of this lake is unknown, but as stated above, it could be seawater of Miocene age (Higgins et al., 2000a). The Ca/Cl and SO4/Cl ratios of Blood Falls do not fall on the seawater line (Fig. 5b). All samples from 1993 to 2003 col-

314

W. B. Lyons et al.

lected by us are enriched in Ca2⫹ relative to seawater. The majority of samples less than 500 mM Cl⫺ are enriched in SO42⫹, whereas those above 500 mM Cl⫺ are depleted in SO42⫺, relative to seawater (Fig. 5b). Interestingly, most of the Ca2⫹ and SO42⫺ data fall on or very close to the 1:1 line although SO42⫺ is systematically enriched at Ca concentrations below ⬃20 mM and is depleted at higher Ca2⫹ concentrations (Fig. 5). We initially hypothesized that this “excess” SO42⫺ (that in excess of Ca2⫹) might originate from oxidation of sulfide or dimethylsulfide as Blood Falls is exposed to the atmosphere (Mikucki et al., 2004). However recently ␦34S measurements have been made on three Blood Falls samples, having Ca/SO4 ratios of 0.39, 0.67, and 1.07. The measurements yielded sulfate-␦34S signatures to be 20.06, 20.28 and 19.87‰, respectively, strongly suggesting that the SO42⫺ originated directly from seawater and is not an oxidation product of a more reduced sulfur species (Chivas et al., 1991). The 87Sr/86Sr ratio of seawater has changed through geologic time and this change is well documented (Faure, 1986). Because there is little to no fractionation of Sr between the salt phase and the water from which it precipitates, the 87Sr/86Sr has been used as a dating tool and a hydrologic tracer (Faure, 1986). We have measured the 87Sr/86Sr ratio in the waters of Lake Bonney and Blood Falls (Lyons et al., 2002). The Blood Falls ratio is 0.71146. Seawater has never reached a 87Sr/86Sr ratio as high as this during Phanerozoic time (Burke et al., 1982). This indicates that the Blood Falls brine has been modified by the introduction of 87Sr via chemical weathering, either before or during its evolution. So although the majority of solutes in Greater Lake Priscu are derived from seawater, there clearly has been some modification of this primary source. The current hypolimnetic waters of Lake Bonney range between 0.7121 and 0.7124 demonstrating even larger enrichments in 87Sr from a more radiogenic weathered source (Lyons et al., 2002). The very high concentration of Fe3⫹ currently observed in Blood Falls also indicates a significant modification of the original seawater. It may suggest that the brine was suboxic or anoxic (possibly a fjord) and accumulated Fe2⫹. It is currently hypothesized that as the brine is exposed to oxygen the reduced iron is converted to Fe3⫹OOH. Where this actually takes place (under the glacier or as the frozen brine is initially exposed to the atmosphere) is not documented. Very high concentrations of dissolved Fe and Mn are observed in the hypolimnia of Lake Bonney (Ward et al., 2003) again providing supporting evidence that the geochemistry of the saline discharge and Lake Bonney are closely related and both are derived from modified seawater. 6.4.3. Geochemical modeling and the source of salts to Lake Bonney The chemical thermodynamic model FREZCHEM (Marion and Grant, 1994; Marion and Farren, 1999; Marion, 2001) was used to evaluate the evolution of the surface waters of both Lakes Fryxell and Hoare if they were to be cryoconcentrated to brines. This was undertaken to assess if Lake Bonney deep waters could be produced from the current fresh water in lakes in Taylor Valley. The model was run until the eutectic temperature for freezing of the waters was obtained (Table 1). The

Table 1. Modeling (FREZCHEM) results following a freezing pathway for surface waters from Lakes Fryxell and Lake Hoare and measured ion concentrations in Lake Bonney hypolimnia. Lake Fryxell surface water at 250.15K at eutectic Final conc., moles/kg (water) Na K Ca

4.86 0.76 0.00005

Mg Cl

0.047 5.53

SO4 HCO3 ⫹ CO3

0.017 0.147

Phases present Ice NaCl · 2H2O KCl Na2SO4 · 10H2O CaCO3 NaHCO3 3MgCO3 · Mg(OH)2 · 3H2O

Lake Hoare surface water at 237.15K at eutectic Final conc., moles/kg (water) Na K Ca

1.14 0.37 0.0055

Mg Cl SO4 HCO3 ⫹ CO3

2.48 6.26 0.098 0.018

Phases present Ice NaCl · 2H2O KCl MgCl2 · 12H2O CaCO3 Na2SO4 · 10H2O 3MgCO3 · Mg(OH)2 · 3H2O

Measured ion concentrations in Lake Bonney hypolimnia, moles/L

Na K Ca Mg Cl SO4 HCO3 ⫹ CO3

Lake Bonney, east lobe

Lake Bonney, west lobe

2.16 0.078 0.043 1.29 4.69 0.037 0.0044

1.71 0.036 0.057 0.369 2.47 0.050 0.078

freezing of these waters lead to the development of Na-Cl brines that have Na/Cl and SO4/Cl within the range of Lake Bonney bottom waters. Freezing of the Lake Fryxell surface water would result in a brine having Mg/Cl ratios greatly depleted with respect to Lake Bonney bottom waters (Table 1). The brines evolved from both Lake Hoare and Lake Fryxell also have K/Cl that is enriched and Ca/Cl that is depleted relative to Lake Bonney (Table 1). During the cryocentration of the solutes, hydrohalite (NaCl●2H2O), sylvite (KCl), MgCl2●12H2O, calcite (CaCO3), mirabilite (Na2SO4●10H2O), and hydromagnesite (3MgCO3●Mg(OH)2●3H2O) would precipitate from both lake waters. In addition nahcolite (NaHCO3) would precipitate from Lake Fryxell water. This modeling suggests, but does not prove, that the development of the hypersaline brines in Lake Bonney did not occur as dilute Taylor Valley water was freeze-dried. Because the freezing of seawater can undergo two pathways (stable “Gitterman” vs. metastable “Ringer-Nelson-Thompson”), gypsum is not always produced as seawater freezes

Halogens in Antarctic Dry Valley Lakes

315

Table 2. Representative end-member compositions. Source

Cl (mM)

I (␮M)

Modern meteoric water Pre-anthropogenic meteoric Formation waters Seawater

0.081 0.081 52305 5508

0.052 0.052 1656 0.58

Cl/Cl (10⫺15)

129

400 ⫾ 2003 400 ⫾ 2003 0.5 ⫾ 0.57 0

45 1.5 0.15 1.5

36

I/I (10⫺12) ⫾ ⫾ ⫾ ⫾

222 0.24 0.057 0.24

1 Fabryka-Martin et al. (1989), 2Antarctic snowmelt (Snyder and Fehn, 2004), 3Antarctic glacier melt (Lyons et al., 1998c), 4Moran et al. (1998), 5East Bonney bottom water (this publication), 6Highest concentration observed in Canadian fracture fluids (Bottomley et al., 2002), 7typical ratios for Upper Cretaceous to Paleocene formation waters (Snyder et al., 2003; Fehn et al., 2000; Fehn et al., 1992), 8Broecker and Peng (1982).

(Marion and Farren, 1999). The Gitterman pathway would yield the precipitation of mirabilite at ⬃ ⫺6°C (i.e., at lower Cl⫺ concentrations) and gypsum precipitation at ⫺22°C (i.e., at higher Cl⫺ concentrations) (Marion and Farren, 1999). The increased Ca2⫹ and SO42⫺ concentrations relative to Cl⫺ indicate that these excess ions must have originated from a brine that initially reached ⫺22°C, but had an excess source of these ions above the initial seawater value (i.e., open-system). This excess was eventually lost to salt precipitation while Na⫹, Mg2⫹ and K⫹ maintained their seawater ratios relative to Cl⫺. Finally, these ions were precipitated as hydrohalite and MgCl2●12H2O, which are along the Gitterman pathway at ⫺23°C and ⫺36°C, respectively. The loss of K⫹ is perplexing and indicates that a K-rich salt may have precipitated as the brine evolved. The modeling suggests that this K-rich salt would probably be sylvite (Table 1). The extent of the saline deposition that is manifested as a surface feature through Blood Falls is unknown. However, it is probable that there is a portion underneath the Taylor Glacier that is enriched in K⫹ relative to Cl⫺. Interestingly enough, the hypolimnia of Lake Bonney are slightly depleted in K⫹ relative to seawater Cl⫺, but are slightly enriched in Mg2⫹. The west lobe of Lake Bonney has a similar Ca/Cl ratio as seawater. Marion et al. (1999) have argued that to precipitate gypsum (i.e., the Gitterman pathway), there must be mirabilite dissolution to provide the SO2⫺ 4 . A sufficient mass of mirabilite must be present within the system as a sulfate source, and its removal would greatly inhibit gypsum production (Marion et al., 1999). Therefore, the brine must have been in contact with the mirabilite to produce the gypsum present. The conservative behavior of Mg2⫹ is expected as it should only start to be removed at the eutectic at ⫺36°C. There is general agreement between the data from Blood Falls and the Gitterman pathway of seawater freezing as detailed in Marion et al. (1999). 6.5. Implications of 129I/I and Source of Saline Waters

36

Cl/Cl Ratios on the

Typical values of 129I and 36Cl for the major reservoirs are summarized in Table 2. Natural production of 129I in the upper atmosphere via cosmic ray interaction with xenon has led to a fairly uniform marine isotopic ratio of 129I/I ⫽ 1.5 ⫻ 10⫺12 (Moran et al., 1998). This ratio can also be considered to be representative of preanthropogenic values in surface waters because marine aerosols dominate the deposition of iodine on land. Iodine in marine organic matter has a similar initial ratio such that the present 129I/I ratio may be used to derive a

minimum iodine age in brines of marine or estuarine origin (Moran et al., 1995; Snyder et al., 2003; Muramatsu et al., 2001). Iodine in the atmosphere is predominantly derived from marine sources (Yoshida and Muramatsu, 1995) where it is either methylated by macroalgae (Manley et al., 1992; Carpenter et al., 1999) or through marine microbial activity (Amachi et al., 2001). Anthropogenic 129I is presently encountered in shallow marine waters (Edmonds et al., 2001) and 129I/I ratios are generally two orders of magnitude above preanthropogenic values. These elevated marine ratios are due to runoff from rivers in the northern hemisphere which have been heavily influenced by nuclear reprocessing (Buragilio et al., 2001; Moran et al., 2002) and are subsequently returned to the atmosphere through the marine production of methyl iodide. Although the precise mechanisms of atmospheric transport and deposition are unknown, similar ratios of 129I/I in both the Arctic and Antarctica reflect anthropogenic input that has been diluted by a marine iodine source (Snyder and Fehn, 2004). Where brines are hosted in rocks with high uranium and thorium, in situ production of 129I through spontaneous fission may also be appreciable ( Fabryka Martin et al., 1985, 1989; Andrews et al., 1989; Bottomley et al., 2002). Anthropogenic input from nuclear weapons testing and fuel reprocessing has also released significant quantities of 129I into the environment over the past 40 yr (Moran et al., 2002). This third source of 129I is most pronounced in North American and Europe, but also in portions of the southern hemisphere (Fehn and Snyder, 2000). Anthropogenic input has elevated 129I/I ratios both in the Arctic and Antarctica. In McMurdo Station, Antarctica, 129I/I ratios in snowmelt are presently over an order of magnitude above preanthropogenic values, at 45 ⫾ 22 ⫻ 10⫺12, and similarly, Mount Erebus snowmelt is at 86 ⫾ 17 ⫻ 10⫺12 (Snyder and Fehn, 2004). Despite the modest elevation in 129I/I, these ratios are roughly an order of magnitude lower than those reported for samples collected from glacially fed streams in the Arctic (Snyder and Fehn, 2004). The 129I/I ratios in the Taylor Valley lakes range from 45.7 ⫻ ⫺12 10 to 1.7 ⫻ 10⫺12 and with the exception of the ⬎35m depth of the east lobe of Lake Bonney, the highest values exist above the chemoclines in the lakes and the lowest values are in the hypolimnia (Appendix A). Figure 6 illustrates how dilution of a brine source by meteoric water can influence both iodine concentrations and 129I/I ratios. Although the source of the brine is likely to have been derived from a complex series of events as suggested above, it is likely to bear some resemblance to iodine-rich formation waters derived from late Cretaceous to Paleocene which are

316

W. B. Lyons et al.

Fig. 6. Plot of 129I/I ratios vs. iodine concentrations showing mixing trajectories between old formation water brines with anthropogenic and preanthropogenic snowmelt. Canadian shield brines from Bottomley et al. (2002), Stripa brines from Fabryka-Martin et al. (1989) and Milk River brines from Fabryka-Martin et al. (1991). End-members for mixing trajectories as in Table 2.

widely distributed globally (Moran et al., 1995; Fehn et al., 2000; Muramatsu et al., 2001; Snyder et al., 2003). These typically have 129 I/I ratios of 0.1 ⫻ 10⫺12 to 0.2 ⫻ 10⫺12. Assuming an anthropogenic end-member similar to the McMurdo Station snow melt, most samples from the Taylor Valley lakes show mixing between the formation waters and anthropogenic waters less than 50 years old. Formation waters from the Milk River Aquifer, Alberta, Canada (Fabryka-Martin et al., 1991), are also presented in the plot, indicating dilution by a predominantly preanthropogenic source. Interestingly, the fracture fluids from the Canadian shield (Bottomley et al., 2002) and the granites of Stripa, Sweden (FabrykaMartin et al., 1989, 1991) bear some similarity to the Taylor Valley lake waters. This general observation is highly significant because the waters in Taylor Valley are essentially open systems containing the brine component, which is a cumulative result of ongoing surficial processes. On the other hand, the fracture fluids from the Canadian Shield and Stripa have been assumed to be connate in origin with infiltration of seawater as far back as the Devonian (Bottomley et al., 1999; Bottomley et al., 2002). The basic working assumption in fracture fluid studies has been that the marine cosmogenic 129I component has completely decayed away, leaving only stable marine 127I and an 129I component derived solely from in situ fissiogenic production (Andrews et al., 1989; Fabryka-Martin et al., 1989; Bottomley et al., 2002; Starinsky and Katz, 2003). This assumption raises several questions, as illustrated in Figure 6. Firstly, samples from the east lobe of Lake Bonney are from shallow waters, which are essentially identical in 129 I/I composition to the Stripa Granite waters. Secondly, the trend observed in the Stripa Granite samples parallels the mixing trend observed in the Taylor Valley lakes. Addition of anthropogenic waters with a higher 129I/I ratio would explain the shift in the dilute brines, but not the concentrated brines. Likewise, since

Fig. 7. Ratios of 36Cl/Cl vs. ratios of 129I/I, showing mixing between preanthropogenic and anthropogenic sources with formation waters. Symbols as in Figure 6. Milk River data from Fabryka-Martin et al. (1991), Stripa granite data from Fabryka-Martin et al. (1989), and 36Cl data for the Antarctic lake waters from Carlson et al. (1990) and Lyons et al. (1998c). End-members for mixing lines as in Table 2.

fissiogenic production of 129I in granites is independent of the concentration of 127I in the fluids, the iodine-poor brines should show a proportionately larger increase in 129I/I ratios than the iodine-rich brines especially if the 129I is derived from uranium and thorium in the host rocks. Also, because fissiogenic 129I is produced within the rock matrix, the resultant amount of 129I that reaches the fluids is proportional to the release efficiency from the mineral matrix, and inversely proportional to the effective porosity of the granites (Andrews et al., 1989; Fabryka-Martin et al., 1989). In cases where the stable iodine concentration is high, it is questionable whether enough 129I will be released to significantly alter 129 I/I ratios (Snyder et al., 2003). Considering the overall chemistry of both the Canadian Brines and these Antarctic lake waters, it is more likely that the iodine concentrations have increased due to freezing and evaporation, rather than the 129I/I ratios increasing because of fissiogenic production. As with 129I, the radioisotope 36Cl is also derived from cosmogenic interactions, from in situ production, and from the nuclear bomb tests of the 1960s. The 36Cl system differs in a number of respects. Firstly, 36Cl has a much shorter half-life (0.3 Myr). In addition, the predominance of stable 35Cl in the oceans renders marine 36Cl/Cl below the limits of AMS detection (Sharma et al., 2000). The predominance of chloride in sea spray as well as latitudinal variations in the cosmogenic production of 36Cl have led to regional variations in meteoric 36Cl/Cl (Davis et al., 2000) and the dilution of the 36Cl bomb signature by marine aerosols has led to present 36Cl/Cl ratios which are indistinguishable from preanthropogenic meteoric values (Cornett et al., 1997). We have plotted the 129I/I data from January 2002 vs. the previously published 36Cl/Cl data from the lakes (Carlson et al., 1990; Lyons et al., 1998c) in Figure 7. Both the Stripa Granite and the Milk River aquifer data are shown to illustrate the difference

Halogens in Antarctic Dry Valley Lakes

between fracture-fluid and hydrocarbon systems, respectively. With the possible exception of the east lobe of Lake Bonney samples, the Antarctic lake waters appear to be a mix of ancient brine and either post or prebomb meteoric waters. This is a similar conclusion as that derived from the chloride isotopes alone (Lyons et al., 1998c, 1999). The east lobe of Lake Bonney appears to have higher 129I:I ratios than anticipated using this simple mixing approach. The lack of precision for the deeper east lobe sample does not allow us to better delineate a closer relationship between the deep west lobe sample and the current saline discharge at Blood Falls (Lyons et al., 1998c), however, the decrease in I/Cl ratios with depth (Fig. 3c) would at least suggest that dissolution of halite both low in 36Cl and with negligible iodine in deeper waters is at least partially responsible for the profile. Again, the surface waters of Lake Bonney closely resemble the values from Precambrian crystalline rocks in Sweden. Although all of the 36Cl in the formation water end-member is most likely derived from in situ production, Figure 7 clearly suggests mixing with a predominantly anthropogenic source. The Milk River aquifer data, on the other hand, show mixing between older formation waters and preanthropogenic groundwater. A portion of the solutes in Lake Bonney is derived from Blood Falls (Spigel and Priscu, 1998). The 36 Cl:Cl value for Blood Falls is 9.5 ⫻ 10⫺15 (Lyons et al., 1998c). It is likely that subglacial seawater-rock interactions have not only altered the 87Sr/86Sr ratio of Blood Falls, but also the 129I/I and 36 Cl/Cl ratios as observed in the Canadian and Swedish brines. 7. CONCLUSIONS

he halogen and boron geochemistry presented here provides important information regarding the origin and the evolution of the three major lakes in Taylor Valley. The conclusions of this work are as follows. 1. All the data collected over the past decade suggest that Lake Hoare is the youngest of the Taylor Valley lakes having begun to fill ⬃1ka. The advance of the Canada Glacier keeps Lake Hoare from flowing into the Lake Fryxell basin and it is assumed that the advance of the glacier established the lake (Lyons et al., 2000). The solutes in the lake have been derived from glacier melt that flows directly into the lake, as well as input from proglacial streams. This includes input of the primary precipitation and the dissolution of aeolian derived salts from the glacier surfaces (Lyons et al., 1998a, 2002). The halogen and B to Cl ratios in the lake indicate significant sources of F⫺, I⫺, and B from either the primary precipitation or the dissolution of salts and dust. 2. The bottom waters of Lake Fryxell are derived from the diffusion of a halite-saturated brine from the pore waters of a playa that began to fill with glacier meltwaters ⬃1 ka. This brine is probably not derived from ancient seawater because of its substantially higher 36Cl:Cl ratio. It is probable that it is brine derived from freezing of contemporary freshwaters within the valley as evidenced by the modeling exercises above. Much of the refilling of the lake to its current level took place from the 1950s to the early 1990s (Hood et al., 1998). The Br/Cl ratios confirm that seawater and/or marine aerosol has been a major contributor of solutes to the lake. The higher than seawater F/Cl ratios indicate a major contribution of weathering derived solutes to the lake. Because of the large number of and length of streams in the Fryxell

317

basin, this lake has the highest current input of silicate weathering products (Nezat et al., 2001). The differences in surface water composition between Lakes Fryxell and Hoare reflect differences in their landscape position as well as their age and previous histories. The solute input to Lake Hoare is dominated by the melting of very old ice and the dissolution of aeolian derived dust on the glaciers (Lyons et al., 2003). The composition Lake Fryxell, as stated above, is related to the input of chemical weathering products derived from the stream channels and the diffusional input of solutes from an older brine. The lower than seawater Br/Cl ratios in Lake Hoare along with the 36Cl data also support this notion. 3. The hypolimnia of Lake Bonney are derived from ancient seawater as originally proposed by Boswell et al. (1967) and more recently argued by Higgins et al. (2000a). Our data strongly support this hypothesis, although it is clear that the ancient seawater has been modified by the input of some weathering products (i.e., radiogenic Sr) and, in the case of the east lobe, by freeze drying and subsequent loss of solutes via precipitation of minerals. The most intriguing question related to this hypothesis is, when was this seawater first emplaced? There are many lines of geological evidence that suggests the inflow of seawater to the western end of the Taylor Valley and could be a remnant of Miocene time. In reality there are no current geochemical data to confirm the “age” of these salts, however. 4. Although the 129I and 36Cl data do not confirm an age of the saline source, they clearly indicate mixing trends between this source and snowmelt containing an anthropogenic signature. Because these open systems bear some resemblance to fracture fluids in the northern hemisphere, modeled ages based entirely on in situ production are possibly overestimated and may require revision. The distinction between static fracture fluids and those actively exchanging with surface waters is not trivial. Ongoing investigations will provide further insight as to whether or not the Taylor Valley lakes provide an active analog for the processes that led to the formation of saline fluids in cratons of the northern hemisphere. In conclusion, the original work of Angino and his co-workers has to a large extent been confirmed, in that the origin of the solutes of these Taylor Valley lakes is very complex. They include ancient seawater, direct glacier-melt and a contribution from the chemical weathering of the stream channel, as well as the dissolution of aeolian transported dust and salts. In addition, in the cases of Lake Bonney and Lake Fryxell, the effect of climate change on lake level and ice cover has also had a major impact on the chemistry of these lakes. The loss of ice-cover and the freezedrying of the lakes have led to the modification of the ion/Cl ratios through loss of salts (Eugster and Jones, 1979). One of the remaining mysteries involving these lakes that still remain is the age of Lake Bonney. Acknowledgments—This work was supported by NSF grants: OPP9211773 and OPP-9813061. We are greatly appreciative to our colleagues John C. Priscu, Diane M. McKnight and Peter T. Doran for their insights into the development and chemical evolution of these lakes. Numerous individuals have helped in sample and data collection over the past ten years and we thanks them all, but especially Rob

318

W. B. Lyons et al.

Edwards and Craig Wolf. We thank Brandy Anglen and Lisa Pratt from Indiana University for the ␦34S measurements of Blood Falls sulfate. We also thank Udo Fehn for use of the Cosmogenic Isotope Laboratory, and David Elmore of PRIME Laboratory for the 129I determinations. We thank Brenda Hall for guidance on the ages of climactic events in Taylor Valley. We would like to thank Joris Gieskes and two anonymous reviewers whose comments and suggestions helped us to greatly improve the earlier version of this manuscript. Associate editor: K. K. Falkner REFERENCES Aiken G., McKnight D., Harnish R. and Wershaw R. (1996) Geochemistry of aquatic humic substances in the Lake Fryxell Basin, Antarctica. Biogeochem. 34, 157–188. Amachi S., Kamagata Y., Kanagawa T. and Muramatsu Y. (2001) Bacteria mediate methylation of iodine in marine and terrestrial environments. App. Environ. Microbiol. 67, 2718 –2722. Andrews J. N., Davis S. N., Fabryka-Martin J., Fontes J-Ch., Lehmann B. E., Loosli H. H., Michelot J-L., Moser H., Smith B. and Wolf M. (1989) The in situ production of radioisotopes in rock matrices with particular reference to the Stripa granite. Geochim. Cosmochim. Acta 53, 1803–1815. Angino E. E. and Armitage K. B. (1963) A geochemical study of Lakes Bonney and Vanda, Victoria Land, Antarctica. J. Geol. 71, 89 –95. Angino E. E., Armitage K. B. and Tash J. C. (1962) Chemical stratification in Lake Fryxell, Victoria Land, Antarctica. Science 138, 34 –36. Barrett P. J. and Hambrey M. J. (1992) Plio-Pleistocene sedimentation in Ferrar Fiord, Antarctica. Sedimentol. 39, 109 –123. Bein A. and Arad A. (1992) Formation of saline groundwaters in Baltic region through freezing of seawater during glacial periods. J. Hydrol. 140, 75– 87. Black R. F. and Bowser C. J(1967) Salts and associated phenomena of the termini of the Hobbs and Taylor Glaciers, Victoria Land, Antarctica. In Extracts of “Commission of Snow and Ice” General Assembly of Bern, Sept.-Oct. 1967, 226 –238. Black R. F., Jackson M. L. and Berg T. E. (1965) Saline discharge from Taylor Glacier, Victoria Land, Antarctica. J. Geol. 74, 175–181. Boswell C. R., Brooks R. R. and Wilson A. T. (1967) Some trace elements in lakes of the McMurdo Oasis, Antarctica. Geochim. Cosmochim. Acta 31, 731–736. Bottomley D. J., Katz A., Chan L. H., Starinsky A., Douglas M., Clark I. D., Raven K. D. (1999) The origin and evolution of Canadian Shield brines: evaporation or freezing of seawater? New Lithium isotope geochemical evidence from the Slave craton. Chem. Geol. 155, 295–320. Bottomley D. J., Renaud R., Kotzer T. and Clark I. (2002) Iodine-129 constraints on residence times of deep marine brines in the Canadian Shield. Geol. 30, 587–590. Broecker W. S. and Peng T.-H (1982) Tracers in the Sea, Eldigio Press, Columbia Univ., New York, 690pp. Brook E. J., Kurz M. D., Ackert R. P. Jr., Denton G. H., Brown E. T., Raisbeck G. M. and Yiou F. (1993) Chronology of Taylor Glacier advances in Arena Valley, Antarctica, using in situ cosmogenic 3He and 10Be. Quat. Res. 39, 11–23. Buraglio N., Aldahan A., Possnert G. and Vintersved I. (2001) 129I from the nuclear reprocessing facilities traced in precipitation and runoff in northern Europe. Environ. Sci. Technol. 35, 1579 –1586. Burke W. H., Denison R. E., Heatherington E. A., Koepnick R. B., Nelson N. F. and Otto J. B. (1982) Variation of seawater 87Sr/86Sr throughout Phanerozoic time. Geol. 10, 516 –519. Carlson C. A., Phillips F. M., Elmore D. and Bentley H. W. (1990) Chlorine-36 tracing of salinity sources in the Dry Valleys of Victoria Land, Antarctica. Geochim Cosmochim. Acta 54, 311–318. Carpenter L. J., Sturges W. T., Penkett S. A., Liss P. S., Alicke B., Hebestreit K. and Platt U. (1999) Short-lived alkyl iodides and bromides at Mace Head, Ireland: Links to biogenic sources and halogen oxide production. J. Geophys. Res. 104, 1679 –1689. Chester R., (2003) Marine Geochemistry, 2nd ed. Blackwell Publ. Malden, MA 505pp.

Chivas A. R., Andrew A. S., Lyons W. B., Bird M. I. and Donnelly T. H. (1991) Isotopic constraints on the origin of salts in Australian playas. I. SulfurPaleo3 84, 309 –332. Cornett R. J., Andrews H. R., Chant L. A., Davies W. G., Greiner B. F., Imahori Y., Koslowsky V. T., Kotzer T., Milton J. C. D. and Milton G. M. (1997) Is 36Cl from weapons’ test fallout still cycling in the atmosphere? Nucl. Inst. Met. Phys. Res. B. 123, 378 –381. Davis S. N., Fabryka-Martin J., Wolfsberg L., Moysey S., Shaver R., Alexander E. C. and Krothe N. (2000) Chlorine-36 in ground water containing low chloride concentrations. Ground Water 38, 912– 921. Doran P. T., Priscu J. C., Lyons W. B., Walsh J. E., Fountain A. G., McKnight D. M., Moorhead D. L., Virginia R. A., Wall D. H., Clow G. D., Fritsen C. H., McKay C. P. and Parsons A. N. (2002) Antarctic climate cooling and terrestrial ecosystem response. Nature 415, 517–520. Duce R. A. and Hoffman E. J. (1976) Chemical fractionation at the air/sea interface. In Annu. Rev. Earth Planet. Sci. (eds. F.A. Donath, F G. Stehli and G.W. Wetherill) vol. 4 pp. 187–228. Annu. Rev. Inc.. Edmonds H. N., Zhou Z. Q., Raisbeck G. M., Yiou F., Kilius L. and Edmond J. M. (2001) Distribution and behavior of anthropogenic 129 I in water masses ventilating in the North Atlantic Ocean. J. Geophys. Res. 106, 6881– 6894. Elston D. P. and Bressler S. L(1981) Magnetic Stratigraphy of DVDP Drill Cores and Late Cenozoic History of Taylor Valley, Transantarctic Mountains, Antarctica. In Dry Valley Drill Project, Antarctic Research Series 33, 413– 426. Eugster H. P. and Jones B. F. (1979) Behavior of major solutes during closed-basin lake evolution. Am. J. Sci. 279, 609 – 631. Fabryka-Martin J., Bentley H., Elmore D. and Airey P. L. (1985) Natural iodine-129 as an environmental tracer. Geochim. Cosmochim. Acta 49, 337–347. Fabryka-Martin J. T., Davis S. N., Elmore D. and Kubik P. W. (1989) In situ production and migration of 129I in the Stripa granite, Sweden. Geochim. Cosmochim. Acta 53, 1817–1823. Fabryka-Martin J., Whittemore D. O., Davis S. N., Kubik P. W. and Sharma P. (1991) Geochemistry of halogens in the Milk River aquifer, Alberta, Canada. App. Geochem. 6, 447– 464. Faure G (1986) Principles of Isotope Geology, 2nd ed, Wiley, New York, 589pp. Fehn U., Holdren G. R., Elmore D., Brunelle T., Teng R. and Kubik P. W. (1986) Determination of natural and anthropogenic 129I in marine sediments. Geophys. Res. Lett. 13, 137–139. Fehn U., Peters E. K., Tullai-Fitzpatrick S., Kubik P. W., Sharma P., Teng R. T. D., Gove H. E. and Elmore, D., (1992) 129I and 36Cl concentrations in waters of eastern Clear Lake area, California: residence times and source ages of hydrothermal fluids. Geochim. Cosmochim. Acta 56, 2069 –2079. Fehn U. and Snyder G. (2000) 129I in the Southern Hemisphere: Global redistribution of an anthropogenic isotope. Nucl. Inst. Met. Phys. Res. B. 172, 366 –371. Fehn U., Snyder G. T. and Egeberg P. K., (2000) Dating of pore fluids with 129I: Relevance for the origin of marine gas hydrates. Science 289, 2332–2335. Fitzpatrick J. J., Muhs D. R. and Jull A. J. T(1990) Saline minerals in the Lewis Cliff ice tongue, Buckley Island quadrangle, Antarctica. In Contributions to Antarctic Research I, Antarctic Research Series 50, 57– 69. Fogg T. R. and Duce R. A (1985) Boron in the troposphere: Distribution and fluxes. J. Geophys. Res. 90 (D2), 3781–3796. Green W. J., Angle M. P. and Chave K. E. (1988) The geochemistry of Antarctic streams and their role in the evolution of four lakes of the McMurdo Dry Valleys. Geochim. Cosmochim. Acta 52, 1265– 1274. Gooseff M. N., McKnight D. M., Lyons W. B. and Blum A. E(2002) Weathering reactions and hyporheic exchange controls on stream water chemistry in a glacial meltwater stream in the McMurdo Dry Valleys. Water Resour. Res. 38(12) 1279, doi:10.1029/ 2001WR000834. Hall B. L. and Denton G. H (2000) Radiocarbon chronology of the Ross Sea Drift, eastern Taylor Valley, Antarctica: Evidence for

Halogens in Antarctic Dry Valley Lakes grounded ice sheet in the Ross Sea at the Last Glacial Maximum. Geogr. Ann. A 82 (2–3), 305–336. Hall B. L., Denton G. H. and Hendy C. H (2000) Evidence from Taylor Valley for the grounded ice sheet in the Ross Sea, Antarctica. Geogr. Ann. A 82 (2–3), 275–303. Hendy C. H (2000) Late Quaternary lakes in the McMurdo Sound region of Antarctica. Geogr. Ann. A 82 (2–3), 411– 432. Hendy C. H., Wilson A. T., Popplewell K. B., and House D. A (1977) Dating of geochemical events in Lake Bonney, Antarctica and their relation to glacial and climatic changes. New Zealand J. Geol. Geophys. 20 (6), 1103–1122. Herut B., Starinsky A., Katz A. and Bein A. (1990) The role of seawater freezing in the formation of subsurface brines. Geochim. Cosmochim. Acta 54, 13–21. Higgins S. M., Denton G. H. and Hendy C. H (2000a) Glacial geomorphology of Bonney Drift, Taylor Valley, Antarctica. Geogr. Ann. A 82 (2–3), 365–389. Higgins S. M., Hendy C. H. and Denton G. H (2000b) Geochronology of Bonney Drift, Taylor Valley, Antarctica: Evidence for interglacial expansions of Taylor Glacier. Geogr. Ann. A 82 (2–3), 391– 409. Hood E. M., Howes B. L. and Jenkins W. J. (1998) Dissolved gas dynamics in perennially ice-covered Lake Fryxell, Antarctica. Limnol. Oceanogr. 42 (2), 265–272. Ishman S. E. and Rieck H. J(1992) A late Neogene Antarctic glacioeustatic record, Victoria Land basin margin, Antarctica. In The Antarctic Paleoenvironment: A Perspective on Global Change, Antarctic Research Series, 56, pp. 327–347, American Geophysical Union. Jones B. F., Eugster H. P. and Rettig S. L., (1977) Hydrochemistry of Lake Magadi basin, Kenya. Geochim. Cosmochim. Acta 41, 53–72. Keys J. R. and Williams K. (1981) Origin of crystalline, cold desert salts in the McMurdo region, Antarctica. Geochim. Cosmochim. Acta 45, 2299 –2309. Lemarchand D. J., Gaillardet J., Lewin E. and Allègre C. J. (2000) The influence of rivers on marine boron isotopes and implications for reconstructing past ocean pH. Nature 408, 951–954. Levy D. B., Schramke J. A., Esposito K. J., Erickson T. A. and Moore J. C. (1999) The shallow ground water chemistry of arsenic, fluorine and major elements: Eastern Owens Lake, California. Appl. Geochem. 14, 53– 65. Lyons W. B. and Mayewski P. A(1993) The geochemical evolution of terrestrial waters in the Antarctic: The role of rock-water interactions. In Physical and Biogeochemical Processes in Antarctic Lakes, Antarctic Research Series, 59 135–143. Lyons W. B. and Welch K. A. (1997) Lithium in waters of a polar desert. Geochim. Cosmochim. Acta 61, 4309 – 4319. Lyons W. B., Welch K. A., Neumann K., Toxey J. K., McArthur R., Williams C., McKnight D. M. and Moorhead D. (1998a) Geochemical linkages among glaciers, streams and lakes within the Taylor Valley, Antarctica. In Ecosystem Dynamics in a Polar Desert: The McMurdo Dry Valleys, Antarctica (ed. J. C. Priscu) pp. 77–92. Antarctic Research Series 72, American Geophysical Union, Washington, D.C. Lyons W. B., Tyler S. W., Wharton R. A., McKnight D. M. and Vaughn B. H. (1998b) A Late Holocene desiccation of Lake Hoare and Lake Fryxell, McMurdo Dry Valleys, Antarctica. Antarct. Sci. 10, 247–256. Lyons W. B., Welch K. A. and Sharma P. (1998c) Chlorine-36 in the waters of the McMurdo Dry Valley lakes, southern Victoria Land Antarctica: Revisited. Geochim. Cosmochim. Acta 62, 185–191. Lyons W. B., Frape S. K. and Welch K. A. (1999) History of McMurdo Dry Valley lakes, Antarctica, from stable chlorine isotope data. Geology 27, 527–530. Lyons W. B., Fountain A., Doran P., Priscu J. C., Neumann K. and Welch K. A. (2000) Importance of landscape position and legacy: The evolution of the lakes in Taylor Valley, Antarctica. Freshwater Biol. 43, 355–376. Lyons W. B., Nezat C. A., Benson L. V., Bullen T. D., Graham E. Y., Kidd J., Welch K. A. and Thomas J. M. (2002) Strontium isotopic signatures of the streams and lakes of Taylor Valley, Southern Victoria Land, Antarctica: Chemical weathering in a polar climate. Aquat. Geochem. 8, 75–95.

319

Lyons W. B., Welch K. A., Fountain A. G., Dana G. L., Vaughn B. H. and McKnight D. M. (2003) Surface glaciochemistry of Taylor Valley, southern Victoria Land, Antarctica and its relationship to stream chemistry. Hydrol. Proc. 17, 115–130. Manley S. L., Goodwin K. and North W. J. (1992) Laboratory production of bromoform, methylene bromide and methyl iodide by macroalgae and distribution in nearshore southern California waters. Limnol. Oceanogr. 37, 1652–1659. Marion G. M. (2001) Carbonate mineral solubility at low temperatures in the Na-K-Mg-Ca-H-Cl-SO4-OH-HCO3-CO3-CO2-H2O system. Geochim. Cosmochim. Acta 65 (14), 1883–1896. Marion G. M. and Farren R. E. (1999) Mineral solubilities in the Na-K-Mg-Ca-Cl-SO4-H2O system: A re-evaluation of the sulfate chemistry in the Spencer-Møller-Weare model. Geochim. Cosmochim. Acta 63, 1305–1318. Marion G. M., Farren R. E. and Komrowski A. J. (1999) Alternative pathways for seawater freezing. Cold Reg. Sci. Tech. 29, 259 –266. Marion G. M. and Grant S. A(1994) FREZCHEM: A chemical-thermodynamic model for aqueous solutions at subzero temperatures. CRREL Spec. Rep. 94-18. USA Cold Regions Research and Engineering Laboratory, Hanover, NH. Matsubaya O., Sakai H., Torii T., Burton H. and Kerry K. (1979) Antarctic saline lakes-stable isotopic ratios, chemical compositions and evolution. Geochim. Cosmochim. Acta 43, 7–25. Maurice P. A., McKnight D. M., Leff L., Fulghum J. E. and Gooseff M. (2002) Direct observations of aluminosilicate weathering in the hyporheic zone of an Antarctic dry valley stream. Geochim. Cosmochim. Acta 66, 1335–1347. Mikucki J. A., Foreman C. M., Sattler B., Lyons W. B. and Priscu J. C. (2004) Geomicrobiology of Blood Falls: An iron-rich saline discharge at the terminus of the Taylor Glacier, Antarctica. Aquat. Geochem. 10, 199 –220. Moran J. E., Fehn U. and Hanor J. S. (1995) Determination of source ages and migration patterns of brines from the U.S. Gulf Coast basin using 129I. Geochim. Cosmochim. Acta 59, 5055–5069. Moran J. E., Oktay S. D. and Santschi P. H(2002) Sources of iodine and iodine 129 in rivers. Water Resour. Res. 38(8), 10.1029/ 2001WR000622. Moran J. E., Fehn U. and Teng R. T. D. (1998) 129I/I ratios in Recent marine sediments: Evidence for a fossil organic source component. Chem. Geol. 152, 193–203. Muramatsu Y., Fehn U. and Yoshida S. (2001) Recycling of iodine in fore-arc areas: evidence from the iodine brines in Chiba, Japan. Earth Planet. Sci. Lett. 192, 583–593. Muramatsu Y. and Wedepohl K. H. (1998) The distribution of iodine in the earth’s crust. Chem. Geol. 147, 201–216. Nezat C. A., Lyons W. B. and Welch K. A. (2001) Chem. weathering in streams of a polar desert (Taylor Valley Antarctica). Geol. Soc. Am. Bull. 113, 1401–1407. Neumann K., Lyons W. B. and DesMarais D. J. (1998) Inorganic carbon-isotopic distribution and budget in the Lake Hoare and Lake Fryxell basins, Taylor Valley, Antarctica. Ann. Glaciol. 27, 685– 690. Park H. and Schlesinger W. H(2002) Global biogeochemical cycle of boron. Global Biogeochem. Cyc. 16(4), 1072, doi: 10.1029/ 2001GB001766. Poreda R. J., Hunt A. G., Lyons W. B. and Welch K. A. (2004) The helium isotopic chemistry of Lake Bonney, Taylor Valley, Antarctica. Aquat Geochem. 10, 353–371. Porter S. C. and Beget J. E(1981) Provenance and depositional environments of late Cenozoic sediments in permafrost cores from lower Taylor Valley, Antarctica. In Dry Valley Drilling Project, Antarctic Research Series 33, 351–363. Prentice M. L., Bockheim J. G., Wilson S. C., Burckle L. H., Hodell D. A., Schlüchter C. and Kellogg D. E(1993) Late Neogene Antarctic glacial history: Evidence from central Wright Valley. In The Antarctic Paleoenvironment: A Perspective on Global Change, Antarctic Research Series 60, 207–250. Priscu J. C., Wolf C. F., Takacs C. D., Fritsen C. H., Laybourn-Parry J., Roberts E. C., Sattler B. and Lyons W. B. (1999) Carbon transformations in a perennially ice-covered Antarctic lake. BioSci. 49, 997–1008.

320

W. B. Lyons et al.

Sharma P., Bourgeois M., Elmore D., Granger D., Lipschutz M. E., Ma X., Miller T., Mueller K., Rickey G., Simms P. and Vogt S. (2000) PRIME lab AMS performance, upgrades and research applications. Nucl. Inst. Met. Phys. Res. B. 172, 112–123. Smith G. I. and Friedman I(1993) Lithology and paleoclimatic implications of lacustrine deposits around Lake Vanda and Don Juan Pond, Antarctica. In Physical and Biogeochemical Processes in Antarctic Lakes (eds. W. J. Green and E. I. Friedmann) Antarctic Research Series 59, 83–94. Snyder G. T., Riese W. C., Franks S., Fehn U., Pelzmann W. L., Gorody A. W. and Moran J. E. (2003) Origin and history of waters associated with coal-bed methane: 129I, 36Cl and stable isotope results from the Fruitland Formation, CO and NM. Geochim. Cosmochim. Acta 67, 4529 – 4544. Snyder G. and Fehn U. (2004) Global distribution of 129I in rivers and lakes: Implications for iodine cycling in surface waters. Nucl. Inst. Met. Phys. Res. B. 223–224, 579 –586. Spigel R. H. and Priscu J. C. (1998) Physical limnology of the McMurdo Dry Valley Lakes. In Ecosystem Dynamics in a Polar Desert: The McMurdo Dry Valleys, Antarctica (ed. J. C. Priscu), pp. 153–187. Antarctic Research Series 72, American Geophysical Union, Washington, D.C. Stuiver M., Yang I. C., Denton G. H. and Kellogg T. B(1981) Oxygen isotope ratios of Antarctic permafrost and glacier ice. In Dry Valley Drilling Project, Antarctic Research Series 33, 131–139.

Takaku Y., Shimamura T., Masuda K. and Igarashi Y. (1995) Iodine determination in natural and tap water using inductively coupled plasma mass spectrometry. Anal. Sci. 11, 823– 827. Ward B. B., Granger J., Maldonado M. T. and Wells M. L. (2003) What limits bacterial production in the suboxic region of permanently ice-covered Lake Bonney, Antarctica? Aquat. Microb. Eco. 31, 33– 47. Welch K. A., Lyons W. B., Graham E., Neumann K., Thomas J. M. and Mikesell D. (1996) Determination of major element chemistry in terrestrial waters from Antarctica by ion chromatography. J. Chromatogr. A 739, 257–263. Welch K. A., Neumann K., McKnight D. M., Fountain A. and Lyons W. B. (2000) Chemistry and lake dynamics of the Taylor Valley lakes, Antarctica: The importance of long-term monitoring. In Antarctic Ecosystems: Models for Wider Ecological Understanding (eds. W. Davison, C. Howard-Williams and P. Broady), pp. 282–287. New Zealand Natural Sciences, Christchurch. Wilson A. T. (1964) Evidence from chemical diffusion of a climate change in the McMurdo dry valleys 1,200 years ago. Nature 201, 176 –177. Wilson A. T. (1979) Geochemical problems of the Antarctic dry areas. Nature 280, 205–208. Yoshida S. and Muramatsu Y. (1995) Determination of organic, inorganic and particulate iodine in the coastal atmosphere of Japan. J. Radioanal. Nucl. Chem. 195, 295–302.

APPENDIX A. CL, F, BR, I AND B CONCENTRATIONS, AND

129

I:Iⴛ10ⴚ12 FOR THE TAYLOR VALLEY LAKES

Samples were collected and analyzed for different constituents at different times as described in the Methods section, although Cl⫺ was analyzed every time a lake chemistry sample was obtained. In the table below, we report the Cl⫺, F⫺, and Br⫺ data for the samples that were collected in November 2002 (with exceptions noted below). The I analyses for Lake Fryxell, Lake Hoare and the west lobe of Lake Bonney were also performed on the samples collected November 2002. The samples for I analyses for the east lobe of Lake Bonney were collected in January 2000. The samples for B analysis were collected in December 1997. The Cl⫺ and Br⫺ data reported for the 18m sample from Lake Fryxell, the 12m and 22m samples from the west lobe of Lake Bonney and the 30m sample from the east lobe of Lake Bonney are from samples that were collected in January 2000. The I datum for the 18m sample from Lake Fryxell is also from January 2000. Location L. L. L. L. L. L. L. L. L. L. L. L. L. L. L. L. L. L. L. L. L. L. L. L. L. L. L. L. L. L. L. L. L.

Fryxell Fryxell Fryxell Fryxell Fryxell Fryxell Fryxell Fryxell Fryxell Fryxell Hoare Hoare Hoare Hoare Hoare Hoare Hoare Hoare Hoare Hoare Hoare Bonney, Bonney, Bonney, Bonney, Bonney, Bonney, Bonney, Bonney, Bonney, Bonney, Bonney, Bonney,

east east east east east east east east east east east east

lobe lobe lobe lobe lobe lobe lobe lobe lobe lobe lobe lobe

Depth (m)

Cl (mM)

F (mM)

Br (mM)

I (␮M)

B (mM)

5 6 7 8 9 10 11 12 15 18 4.5 5 6 8 10 12 14 16 18 20 22 5 6 8 10 12 13 15 18 20 22 25 30

2.74 13.0 15.6 17.7 22.5 32.7 46.9 53.3 80.8 106 1.24 1.22 1.63 3.05 3.99 4.17 4.64 5.22 5.55 5.51 5.51 9.8 14.9 29.6 45.9 128 289 431 1063 1957 2995 3924 4784

0.018 0.063 0.070 0.080 0.098 0.139 0.196 0.215 0.290 n.a. 0.019 0.020 0.027 0.050 0.065 0.069 0.077 0.087 0.092 0.091 0.093 0.013 0.017 0.026 0.027 0.040 0.061 0.074 0.221 0.537 0.675 0.951 n.a.

0.016 0.018 0.020 0.024 0.030 0.048 0.067 0.094 0.122 0.133 0.0023 0.0025 0.0025 0.0025 0.0027 0.0032 0.0049 0.0054 0.0054 0.0053 0.0053 0.028 0.032 0.035 0.050 0.12 0.45 0.76 1.18 7.43 10.0 14.6 21.5

0.037 0.084 0.087 0.086 0.144 0.345 0.836 0.976 1.42 1.83 0.0225 0.0224 0.0259 0.0327 0.0394 0.0347 0.0276 0.0273 0.0307 0.0275 0.0275 n.a. n.a. n.a. 0.168 0.317 n.a. 0.463 0.792 1.10 1.23 1.05 1.08

0.031 0.032 0.036 0.046 0.074 0.110 0.153 0.157 0.227 0.276 0.0020 0.0030 0.0070 0.0120 0.0130 0.0170 0.0200 0.0220 0.0230 0.0230 0.0230 0.014 0.019 0.040 0.045 0.137 0.254 0.347 0.844 1.67 2.30 2.93 3.64

129

I:1 ⫻ 10⫺12

1.7 ⫾ 1.3

45.7 ⫾ 23.3 8.3 ⫾ 6.4

Halogens in Antarctic Dry Valley Lakes APPENDIX A. Cl, F, Br, I AND B CONCENTRATIONS, AND Location L. L. L. L. L. L. L. L. L. L. L. L. L. L. L. L. L. L.

Bonney, Bonney, Bonney, Bonney, Bonney, Bonney, Bonney, Bonney, Bonney, Bonney, Bonney, Bonney, Bonney, Bonney, Bonney, Bonney, Bonney, Bonney,

east lobe east lobe west lobe west lobe west lobe west lobe west lobe west lobe west lobe west lobe west lobe west lobe west lobe west lobe west lobe west lobe west lobe west lobe

129

321

I:Iⴛ10ⴚ12 FOR THE TAYLOR VALLEY LAKES (Continued)

Depth (m)

Cl (mM)

F (mM)

Br (mM)

I (␮M)

B (mM)

35 37 4.5 5 6 8 10 12 13 14 15 17 20 22 25 30 35 38

4696 4937 11.3 9.02 22.8 31.0 38.2 168 272 902 1283 1488 1693 1933 2110 2325 2345 2486

1.18 1.20 0.010 0.010 0.014 0.026 0.026 n.a. 0.052 0.128 0.165 0.189 0.189 n.a. 0.263 0.214 0.239 0.226

20.2 20.3 0.02 0.029 0.033 0.018 0.079 0.45 0.62 2.44 2.92 3.11 3.46 4.32 4.50 5.01 5.29 5.09

1.00 1.01 0.062 0.061 0.090 0.132 0.137 0.242 0.155 0.802 1.71 2.45 3.53 4.35 4.83 7.16 7.52 7.88

3.46 3.37 0.037 0.018 0.023 0.032 0.051 0.136 0.372 1.18 1.26 1.39 1.59 1.64 1.86 2.00 2.07 2.12

129

I:1 ⫻ 10⫺12 33 ⫾ 8

1.7 ⫾ 1.3

2.4 ⫾ 2

APPENDIX B. BR AND CL CONCENTRATIONS IN TAYLOR VALLEY STREAMS Br concentrations in Taylor Valley streams are typically below our detection limit of 0.0002 mM. The data compiled below represent the stream samples collected during 5 field seasons, 1997–98 though 2001– 02, where Br was detected and quantified. The Cl concentrations for those samples and the molar ratios of Br/Cl are also reported. Date

Cl (mM)

Br (mM)

Br/Cl (molar)

Lat (°S)

Long (°E)

Lake Bonney basin Blood Falls Blood Falls Blood Falls Blood Falls Blood Falls Blood Falls Blood Falls Blood Falls Blood Falls Blood Falls Blood Falls Blood Falls Bohner Lawson Lawson Lawson Lyons Lyons Lyons Lyons Lyons Priscu Priscu Santa Fe Santa Fe Santa Fe Santa Fe Santa Fe Santa Fe Santa Fe Santa Fe Santa Fe Santa Fe Sharp Sharp

11/23/97 1/2/98 1/11/98 1/9/99 1/18/00 11/30/00 12/7/00 12/7/01 12/22/01 1/4/02 1/12/02 1/24/02 12/22/00 11/29/00 11/30/00 12/7/00 12/23/99 1/6/00 1/9/00 12/7/00 1/4/02 12/9/00 1/6/01 11/30/98 12/1/98 12/9/98 12/30/98 1/9/00 11/22/00 11/29/00 11/30/00 12/29/00 1/6/01 11/23/97 12/7/00

1045 1279 942 663 1424 14.8 20.2 78.7 907 1254 677 1450 0.647 0.286 0.230 0.299 1.074 4.466 1.094 0.490 1.350 0.687 0.386 1.309 1.760 0.527 0.860 1.209 5.022 0.929 1.090 0.625 0.954 1.19 1.784

1.64612 1.57278 1.12811 0.86371 1.812 0.02637 0.02345 0.07969 1.068 1.22738 0.66901 1.71044 0.00022 0.00023 0.00025 0.00051 0.00175 0.00679 0.00224 0.00089 0.00199 0.00032 0.00021 0.00197 0.00265 0.00083 0.00122 0.00183 0.00868 0.00183 0.00213 0.00089 0.00131 0.00015 0.00032

0.00157 0.00123 0.00120 0.00130 0.00127 0.00178 0.00116 0.00101 0.00118 0.00098 0.00099 0.00118 0.00035 0.00082 0.00109 0.00170 0.00163 0.00152 0.00205 0.00182 0.00147 0.00046 0.00055 0.00150 0.00150 0.00158 0.00142 0.00152 0.00173 0.00197 0.00195 0.00143 0.00137 0.00013 0.00018

77.72 77.72 77.72 77.72 77.72 77.72 77.72 77.72 77.72 77.72 77.72 77.72 77.69 77.72 77.72 77.72 77.72 77.72 77.72 77.72 77.72 77.70 77.70 77.72 77.72 77.72 77.72 77.72 77.72 77.72 77.72 77.72 77.72 77.72 77.72

162.26 162.26 162.26 162.26 162.26 162.26 162.26 162.26 162.26 162.26 162.26 162.26 162.56 162.26 162.26 162.26 162.27 162.27 162.27 162.27 162.27 162.53 162.53 162.26 162.26 162.26 162.26 162.26 162.26 162.26 162.26 162.26 162.26 162.24 162.24

Lake Fryxell basin Aiken Aiken Aiken

12/19/00 12/26/00 1/20/02

0.425 0.442 0.998

0.00058 0.00047 0.00054

0.00136 0.00107 0.00055

77.60 77.60 77.60

163.26 163.26 163.26

322

W. B. Lyons et al. APPENDIX B. Br AND Cl CONCENTRATIONS IN TAYLOR VALLEY STREAMS (Continued)

Lake Fryxell Basin Aiken Aiken Aiken Aiken Aiken Aiken Aiken Aiken Commonwealth Commonwealth Commonwealth Delta Delta Delta Harnish Harnish Harnish Huey Lost Seal Lost Seal Lost Seal Lost Seal Lost Seal Lost Seal Lost Seal Lost Seal McKnight McKnight McKnight McKnight McKnight Von Guerard Von Guerard Von Guerard

Date

Cl (mM)

Br (mM)

Br/Cl (molar)

Lat (°S)

Long (°E)

11/26/97 12/28/97 1/7/98 1/16/98 1/23/98 12/10/99 12/19/00 12/26/00 12/20/00 12/30/00 12/23/97 12/23/00 12/27/00 12/24/98 1/21/02 1/9/98 1/15/98 1/6/98 1/20/02 12/10/00 12/19/00 12/27/00 12/24/97 12/28/97 1/7/98 1/16/98 12/19/00 1/20/02 11/26/97 12/28/97 1/23/98 12/23/00 12/26/00 12/26/97

0.715 0.515 0.531 0.763 0.782 1.207 0.264 0.359 0.279 0.090 0.145 1.203 2.086 0.661 1.175 0.880 0.899 0.582 1.462 0.410 0.236 0.212 0.547 0.285 0.737 0.486 0.353 0.334 0.569 0.245 0.297 0.218 0.288 0.661

0.00058 0.00035 0.00024 0.00030 0.00027 0.00134 0.00020 0.00035 0.00031 0.00016 0.00015 0.00043 0.00060 0.00032 0.00034 0.00040 0.00045 0.00033 0.00050 0.00047 0.00036 0.00027 0.00061 0.00029 0.00049 0.00036 0.00034 0.00026 0.00053 0.00022 0.00021 0.00025 0.00032 0.00056

0.00081 0.00068 0.00045 0.00040 0.00034 0.00111 0.00074 0.00096 0.00111 0.00172 0.00105 0.00036 0.00029 0.00048 0.00029 0.00045 0.00050 0.00056 0.00034 0.00114 0.00151 0.00130 0.00112 0.00100 0.00066 0.00075 0.00096 0.00078 0.00093 0.00091 0.00071 0.00117 0.00112 0.00084

77.60 77.60 77.60 77.60 77.60 77.60 77.60 77.60 77.56 77.56 77.56 77.62 77.62 77.62 77.62 77.62 77.62 77.60 77.59 77.59 77.59 77.59 77.59 77.59 77.59 77.59 77.59 77.59 77.59 77.59 77.59 77.60 77.60 77.60

163.26 163.26 163.26 163.26 163.26 163.26 163.26 163.26 163.38 163.38 163.38 163.10 163.10 163.10 163.26 163.26 163.26 163.12 163.24 163.24 163.24 163.24 163.24 163.24 163.24 163.24 163.26 163.26 163.26 163.26 163.26 163.25 163.25 163.25

APPENDIX C. F AND CL CONCENTRATIONS FOR TAYLOR VALLEY STREAM SAMPLES COLLECTED DURING THE 2003– 04 FIELD SEASON Stream

Date

F⫺ (mM)

Cl⫺ (mM)

F/Cl (molar)

Lat. (°S)

Long. (°E)

Lake Fryxell Basin Aiken Aiken Bowles Canada Canada Canada Canada Commonwealth Commonwealth Delta Delta Delta Green Green Green Green Green Huey Lost Seal Lost Seal Lost Seal Lost Seal McKnight McKnight McKnight VonGuerard

1/6/04 12/10/03 12/29/03 12/12/03 12/18/03 12/29/03 1/21/04 1/26/04 12/16/03 1/21/04 12/12/03 12/29/03 1/21/04 1/21/04 12/12/03 12/18/03 12/29/03 12/12/03 1/6/04 12/10/03 12/19/03 12/30/03 1/6/04 12/10/03 12/30/03 12/27/03

0.0088 0.0056 0.0034 ND ND ND 0.0023 0.0018 0.0031 0.0050 0.0053 0.0050 0.0050 0.0028 ND 0.0014 0.0014 0.0052 0.0034 0.0016 0.0039 0.0042 0.0085 0.0048 0.0086 0.0050

0.934 0.258 0.102 0.054 0.044 0.056 0.063 0.129 0.223 0.584 0.368 0.695 0.358 0.043 0.034 0.038 0.033 0.180 0.209 0.113 0.094 0.120 0.307 0.193 0.276 0.224

0.009 0.022 0.033

77.60 77.60 77.62 77.61 77.61 77.61 77.61 77.56 77.56 77.62 77.62 77.62 77.62 77.62 77.62 77.62 77.62 77.60 77.59 77.59 77.59 77.59 77.59 77.59 77.59 77.60

163.27 163.27 163.05 163.05 163.05 163.05 163.05 163.38 163.38 163.10 163.10 163.10 163.05 163.05 163.05 163.05 163.05 163.12 163.24 163.24 163.24 163.24 163.26 163.26 163.26 163.25

0.036 0.014 0.014 0.009 0.014 0.007 0.014 0.065 ND 0.036 0.042 0.029 0.016 0.015 0.042 0.035 0.028 0.025 0.031 0.023

Halogens in Antarctic Dry Valley Lakes

323

APPENDIX C. F AND Cl CONCENTRATIONS FOR TAYLOR VALLEY STREAM SAMPLES COLLECTED DURING THE 2003–04 FIELD SEASON (Continued) Stream

Date

F⫺ (mM)

Cl⫺ (mM)

F/Cl (molar)

Lat. (°S)

Long. (°E)

VonGuerard Lake Hoare Basin Andersen House House McKay McKay Wharton Wharton

1/21/04

0.0063

0.237

0.027

77.60

163.25

1/26/04 12/6/03 12/13/03 12/6/03 12/13/03 12/6/03 12/13/03

0.0057 ND ND ND ND ND ND

0.500 0.108 0.044 0.078 0.022 0.058 0.033

0.011 ND ND ND ND ND ND

77.62 77.64 77.64 77.64 77.64 77.64 77.64

162.90 162.74 162.74 162.74 162.74 162.74 162.74

Lake Bonney Basin Priscu Priscu Priscu Lawson Lawson Lawson Lawson Lyons Lyons Lyons Lyons Blood Falls Blood Falls Blood Falls Santa Fe Santa Fe Santa Fe Santa Fe Sharp

1/23/04 12/17/03 12/29/03 1/5/04 1/23/04 12/16/03 12/26/03 1/5/04 1/23/04 12/16/03 12/26/03 1/5/04 1/23/04 12/26/03 1/5/04 1/23/04 12/16/03 12/26/03 1/23/04

0.0071 0.0078 0.0067 0.0034 0.0036 0.0039 0.0039 ND ND ND ND ND ND ND ND ND ND ND 0.0088

0.514 0.810 0.432 0.083 0.066 0.126 0.070 0.104 0.191 0.082 0.356 1.712 15.782 10.074 0.222 0.337 0.329 0.350 0.770

0.014 0.010 0.016 0.040 0.055 0.031 0.056 ND ND ND ND ND ND ND

77.70 77.70 77.70 77.72 77.72 77.72 77.72 77.72 77.72 77.72 77.72 77.72 77.72 77.72 77.72 77.72 77.72 77.72 77.72

162.53 162.53 162.53 162.26 162.26 162.26 162.26 162.27 162.27 162.27 162.27 162.26 162.26 162.26 162.26 162.26 162.26 162.26 162.26

0.011

APPENDIX D. SELECTED STREAM SAMPLES COLLECTED DURING THE 1997–1998 AUSTRAL SUMMER SEASON WERE ANALYZED FOR B AND CLⴚ CONCENTRATION Stream

Date

B (␮M)

Cl (␮M)

Lake Fryxell Basin Canada Canada Lost Seal Lost Seal McKnight Aiken Aiken Delta Green Green

12/26/97 1/18/98 12/28/97 1/7/98 12/28/97 11/26/97 12/28/97 12/24/97 11/26/97 12/31/97

⬍0.04 ⬍0.04 1.47 1.05 0.94 2.43 1.70 0.58 0.08 ⬍0.04

63 45 285 737 245 715 515 661 151 61

Lake Hoare Basin Andersen Andersen Andersen House

12/25/97 1/5/98 1/12/98 1/5/98

0.17 0.24 1.19 ⬍0.04

115 151 179 66

Lake Bonney Basin Lawson Lawson Santa Fe Santa Fe Bartlette Bartlette Priscu Priscu Priscu

11/23/97 1/2/98 11/22/97 1/2/98 11/23/97 1/17/98 11/23/97 12/30/97 1/17/98

0.87 0.09 2.03 0.21 0.36 0.57 0.54 0.36 0.83

158 114 553 186 213 425 313 211 244

B/Cl (molar)

Lat. (°S)

Long. (°E)

77.61 77.61 77.59 77.59 77.59 77.60 77.60 77.62 77.62 77.62

163.05 163.05 163.24 163.24 163.26 163.27 163.27 163.10 163.05 163.05

0.0015 0.0016 0.0066

77.62 77.62 77.62 77.64

162.90 162.90 162.90 162.74

0.0055 0.0008 0.0037 0.0011 0.0017 0.0013 0.0017 0.0017 0.0034

77.72 77.72 77.72 77.72 77.72 77.72 77.70 77.70 77.70

162.26 162.26 162.26 162.26 162.41 162.41 162.53 162.53 162.53

0.0051 0.0014 0.0038 0.0034 0.0033 0.0009 0.0005