Sulfur Cycling in a Permanently Ice-Covered Amictic Antarctic Lake

Home , Lake Fryxell

to additional dry valley sites and systems. Analysis of soils Freckman, D.W. 1982. Parameters of the nematode contribution to returned to the United States for physical and chemical prop- ecosystems. In D.W. Freckman, (Ed.), Nematodes in soil ecosystems. erties will allow us to identify the soil factors that best explain Austin, Texas: University of Texas Press. nematode abundance, distribution, and community structure Freckman, D.W., and R.A. Virginia. 1989. Plant-feeding nematodes in the McMurdo Dry Valleys. in deep-rooting desert ecosystems. Ecology, 70(6), 1,665-1,678. This work was supported by National Science Foundation Maslen, N.R. 1981. The Signy Island terrestrial reference sites: XII. grants DPP 88-18049 and DPP 89-14655. Population ecology of nematodes with additions to the fauna. British Antarctic Survey Bulletin, 53, 57-75. References Timm, R.W. 1971. Antarctic soil and freshwater nematodes from the McMurdo Sound region. Proceedings of the Helminthological Society of Block, W. 1984. Terrestrial microbiology, invertebrates and ecosys- Washington, 38(1), 42-52. tems. In R.M. Laws, (Ed.), Antarctic ecology. New York: Academic Vincent, W.F. 1988. Microbial ecosystems of Antarctica. Cambridge: Cam- Press. bridge University Press.

Sulfur cycling lake area), we are in the process of constructing a sulfur balance for the lake. The goals of this effort are to determine: in a permanently ice-covered • the long-term fate of sulfur entering the lake, particularly if amictic antarctic lake, a sulfate sink exists within the lake as is suggested from estimates of stream input into the closed basin (Green et al. Lake FryxeD 1989); • the importance of organic-matter remineralization by microbial sulfate reduction in the carbon and nitrogen cycles; BRIAN L. HOWES • the rate of internal recycling of sulfur; and • to begin to address the long-term redox stability of Lake Biology Department Fryxell. Woods Hole Oceanographic institution Woods Hole, Massachusetts 02543 During 1988 and 1989, measurements of chloride concentration in the water column supported the concept of amixis and a diffusion dominated transport (Lawrence and Hendy RICHARD L. SMITH 1985; Aiken et al. in press). While measured chloride profiles were nearly identical to earlier measurements by Toni et al. Water Resources Division (1975) and Green et al. (1989), our measurements of sulfate U.S. Geological Survey concentrations (by both turbidimetnic and ion chromatographic Arvada, Colorado 80002 methods) showed no significant year-to-year differences but were lower at depth than reported by previous studies. The dissolved sulfide and sulfate profiles indicated sulfate-reduc- The ice-free valleys of southern Victoria Land contain a va- ing activity within the sediments. Using the slope of these riety of perennially ice-covered closed-basin lakes. We are profiles and applying Ficks first law of diffusion (Li and Gre- studying one of these dry-valley lakes, Lake Fryxell (77°37S gory 1974) yields a rate of sulfate consumption below 18 meters 163°07E) in the lower Taylor Valley. The lake is approximately of 0.64 micromole per square centimeter per year and an up- 5.5 kilometers long and 2 kilometers wide, with a surface area ward flux of sulfide of 0.48 and 0.65 micromole per square of 7.06 square kilometers (Lawrence and Hendy 1985, 1989), centimeter per year from the bottom and near the oxycline, a maximum basin depth of 18.9 meters and a center ice thick- respectively. These rates indicate a system in relative balance ness of 5 meters. The lack of wind-driven mixing and the saline and a turnover of the water-column sulfate pool (9.5-18.7 me- bottom waters resulting from the upward diffusion of brines ters) of approximately 1,750 years, emphasizing the long time or redissolved salts from evaporative concentration of lake water scales upon which the biogeochemical cycling with Lake Fryxell (Lawrence and Hendy 1989), coupled with concentration of must be gauged. glacial meltwater inflow (Green et al. 1989), has resulted in an For comparison, we made direct measurements of sulfate amictic water column. Upon this setting of amixis and a closed- reduction over 36-72 hours in surficial sediment (0-20 centi- basin is a redox stratified water column. In the euphotic zone meters) using a sulfur-35/sulfate injection technique (Jorgensen (5.0-9.5 meters), oxygen concentrations in excess of air equi- 1978) and recovery of reduced label following treatment with libration exist due to exclusion of gases in the formation of hot chromous chloride under anaerobic acid conditions (Howes, new ice at the base of the ice sheet (Wharton et al. 1986) and Dacey, and King 1984). There was significant sulfate-reducing oxygen production resulting from carbon fixation (Vincent 1981). activity at the sediment water interface, which is expected at In contrast, the hypolimnion is anoxic with concentrations of the benthic boundary layer, especially under these strongly hydrogen sulfide approaching 1.25 millimole (figure 1) as a stratified conditions. The maximal activity, however, was found result of the settling of autochthonous organic matter into sa- in the 0-2-centimeter section with a rapid decrease with depth line bottom waters during decay. and undetectable activity below 8 centimeters. Almost 60 per- Due to the apparent importance of sulfur transformations cent of the activity was found in the 0-2-centimeter section to the geochemical cycles in Lake Fryxell (the anoxic basin (figure 2). The absence of significant sulfate-reducing activity covers approximately 2.2 square kilometers or one-third of the below 4 centimeters is related to the depletion of sulfate and

230 ANTARCTIC JOURNAL OXYGEN/SULFIDE (mM) SULFATE (mM) 0.5 1.0 1.5 0.5 M. am 1.0 1.5 2.0

5 5

(I) U) Q) a) -4-J -J Q) ,10 ,10

I I H H 0 0 LU a aLU

15 15

20 20 0 20 40 60 80 100 120 CHLORIDE (mM) Figure 1. Water-column concentrations of dissolved oxygen and sulfide (December 1989) and dissolved sulfates and chloride (December 1988; open symbols) in the mid-lake basin. (mM denotes millimole.)

is consistent with the high concentrations of methane found cromoles per square centimeter per year of inorganic sulfur in these sediments (data not shown). The lower rate measured over the past approximately 8,000 years. This burial of inor- from time-course incubations (0.38±0.05 micromole per square ganic reduced sulfur is close to the 0.63 micromole per square centimeter per year) versus sulfate diffusion may be due to a centimeter per year of stream input into Lake Fryxell estimated lower present day annual rate or to seasonal changes in sed- by Green et al. (1989) (figure 3). It appears then that the "lost" iment sulfate reduction related to seasonal changes in the or- sulfate in Lake Fryxell can be accounted for by microbial re- ganic-matter supply from phytoplankton production. duction and burial into sediments. The high rate of burial Although the sediment efflux of sulfide was of the same relative to the amount reduced would require anoxic condi- magnitude as our estimate of sulfide production, measurement tions over this period which is supported by sediment organic of sedimentary sulfur pools indicate a significant accumulation carbon to reduced sulfur ratios of 1.3 (below the depth of active of inorganic sulfur in AVS or acid volatile (hydrogen sulfide, sulfate reduction) much smaller than the 2.8 (carbon/sulfur) iron sulfide), CRS or chromium reducible (pyrite, elemental found in oxic marine basins and is similar to other anoxic basins sulfur) and "organic" (non-sulfate, non-CRS) sulfur with depth. such as the Black Sea (Berner and Raiswell 1983). The sulfur pools in the surface 4 centimeters are still increasing The present sulfur balance of the anoxic basin of Lake due to the continuing input from sulfate reduction at these Fryxell suggests a system in near equilibrium with sulfate depths (figure 2). Inorganic reduced sulfur pools are signifi- inputs (figure 3). The rate of apparent sulfur oxidation near cantly higher than found in freshwater sediments and are sim- the oxycline suggests, however, that we must also determine ilar to marine sediments consistent with the saline nature of the potential for the removal by oxidation of sulfide via an- the lake waters. Using the 10,410 date of the aragonite lens oxygenic photosynthesis and/or the precipitation of metal sul- from Lawrence and Hendy (1985) (found at 25 centimeters at fides below the interface. In addition, the estimated rates of this site), we calculate an average accumulation of 0.43 mi- sulfate reduction relative to sulfur burial and upward diffusion

1990 REVIEW 231

SULFATE REDUCTION RATE (umol/cm3/yr) SULFUR (umol/cc)

0 0 50 0.02 004 006 0.08 0.1 0.12 100 150 200

INTERFACE

0-2

2-4

4-6

6-8 Z 10 8-10

10-12

12-14 15

14-16

16-18

18-20 I ACID VOLATILE Is CHROMIUM REDUCIBLE ES "ORGANIC" Figure 2. Vertical distribution of microbial sulfate reduction (n=3 cores) and fractionation of sulfur pools in sediments (n=7 cores) from 18.5 meters in the mid-lake anoxic basin of Lake Fryxell, December 1988. (umol/cm3/yr denotes micromole per cubic centimeter per year. umol/cc denotes micromole per cubic centimeter.)

Figure 3. Present status of sulfur balance for the anoxic basin of Lake Fryxell. Sulfate (SO4 ) and sulfur (S =) diffusion, sulfate re- L1 duction, sulfur sedimentation, and sulfur pools (for burial calcu- lations) are from the present study; stream input of sulfate calculated from Green et at. 1989; and sediment accretion rate calculated from Lawrence and Hendy 1989. Question marks represent areas where no data is yet available. All units are micromoles per cubic centi- 02 Sed meter per year. L),L 11 II 0.01 of sulfide indicate the need to quantify water-column sulfate Stream AEROBIC ZONE Input reduction and the areal extent of sediment sulfate-reducing activity within the lake. We thank M. Brooks, R. Harnish, and R. VanEtten for as- INTERFACE sistance in the field; D. Goehringer, S. Brown-Leger, J . Weber, and C. Stewart for laboratory analysis; and C. Taylor for help- ANAEROBIC ZONE ful discussions of the data. This research was supported by 0.65 the U.S. Geological Survey and National Science Foundation Water Column grant DPP 88-18782 and is Woods Hole Oceanographic Insti- so— Sulfate tution contribution number 7442. 4 Reduction S ? A7 References F1 V Aiken, C., D. McKnight, R. Wershaw, and L. Miller. In press. Evi-

0.64 0.48 dence for the diffusion of aquatic fulvic acid from the sediments of Lake Fryxell, Antarctica. Proceedings of the American Chemical Society. Berner, R.A., and R. Raiswell. 1983. Burial of organic carbon and pyrite sulfur in sediments over Phanerozoic time: A new theory. Geochimica Sediment et Cosmochimjca Acta, 47, 855-862. Sulfate Reduction Green, Wj., T.J. Gardner, T.G. Ferdelman, M.P. Angle, L.C. Varner, and P. Nixon. 1989. Geochemical processes in the Lake Fryxell Basin (Victoria Land, Antarctica). Hydrobiologia, 172, 129-148. SEDIM Howes, B.L., J.W.H. Dacey, and G.M. King. 1984. Carbon flow through oxygen and sulfate reduction pathways in salt marsh sediments. Lunnology and Oceanography, 29, 1,037-1,051.

232 ANTARCTIC JOURNAL Jorgensen, B. 1978. A comparison of methods for the quantification Toni, T., N. Yamagata, S. Nakaya, S. Murata, T. Hashimoto, 0. Mat- of bacterial sulfate reduction in coastal marine sediments. 1. Mea- subaya, and H. Sakai. 1975. Geochemical aspects of the McMurdo surement with radiotracer techniques. Geonicrohiology Journal, 1, 11- saline lakes with special emphasis on the distribution of nutrient 27. matters. In T. Toni (Ed.), Memoirs of National Institute of Polar Research: Lawrence, M.J.F., and C. H. Hendy. 1985. Water column and sediment Geochemical and Geophysical Studies of Dry Valleys, Victoria Land in characteristics of Lake Fryxell, Taylor Valley, Antarctica. New Zealand Antarctica. (Special Issue No. 4.) Journal of Geology and Geophysics, 28, 543-552. Vincent, W.F. 1981. Production strategies in Antarctic inland waters: Lawrence, M.J.F., and C.H. Hendy. 1989. Carbonate deposition and Phytoplankton eco-physiology in a permanently ice-covered lake. Ross Sea ice advance, Fryxell basin, Taylor Valley, Antarctica. New Ecology, 62, 1,215-1,224. Zealand Journal of Geology and Geophysics, 32, 267-277. Wharton, R.A., C.P. McKay, G.M. Simmons, and B.C. Parker. 1986. Li, Y.-H., and S. Gregory. 1974. Diffusion of ions in seawater and in Oxygen budget of a perennially ice-covered Antarctic lake. Limnology deep-sea sediments. Geochi,nica et Cosinochimica Acta, 38, 703-714. and Oceanography, 31, 437-443.

Bacterial biomass involves enumerating and characterizing the total microbial populations within the lake. Little is known about microor- and heterotrophic activity ganisms in such systems, especially the planktonic bacteria, in the water column even though it appears that much of the carbon, nitrogen, and sulfur cycling may occur in the water column (Canfield and of an amictic antarctic lake Green 1985; Vincent, Downs, and Vincent 1981). Lake Fryxell is one of the most productive lakes in the McMurdo Dry Valleys (Vincent 1981). It contains a relatively RICHARD L. SMITH uniform salinity gradient and a corresponding increase in dissolved organic matter with depth. The dissolved organic mat- Water Resources Division ter reaches a maximum concentration of 25 milligrams of carbon U.S. Geological Survey per liter at 18 meters (maximum depth, 18.5 meters) (McKnight Denver, Colorado 80225 et al. 1988). The water column is composed of an upper aerobic zone that contains oxygen concentrations well in excess of BRIAN L. HOWES atmospheric equilibrium and an anoxic zone that contains reduced compounds such as ammonium (figure 1), hydrogen sulfide, and methane (data not shown). The oxycline, the re- Biology Department gion of transition between these two zones, has a very steep Woods Hole Oceanographic Institute Woods Hole, Massachusetts 02543 oxygen gradient; dissolved oxygen values decrease from 0.9 to 0 millimoles per liter in only 1.5 meters (from 8.0 to 9.5 meters). Very high oxygen concentrations within the zone of The lakes in the McMurdo Dry Valleys have several unique light penetration have been found in other antarctic lakes and physical characteristics that make them unusual environments result because the ice cover restricts exchange of gases with for aquatic microorganisms. The thick, permanent ice cover on the atmosphere (Wharton et al. 1986). Immediately beneath these lakes maintains a constant temperature within each lake the oxic-anoxic interface in Lake Fryxell there is a turbidity by insulating the water column. The ice also markedly reduces maximum (9.5 meters, figure 1). At least part of this high light penetration and prevents exchange of gases and nutrients turbidity was due to populations of bacteria. Bacterial abun- between the water column and the atmosphere (Vincent 1981; dance increased dramatically from 9.0 to 10.5 meters, with a Wharton et al. 1986). Many of the dry valley lakes contain large peak value at 10.5 meters that was nearly eightfold greater salinity gradients that extend vertically throughout the entire than that in the aerobic portions of the water column (figure water column, and these gradients effectively prevent mixing 2). Morphologically, the bacteria were much larger within this from occurring (Toni et al. 1975). Lakes of this type are rare depth interval than elsewhere in the depth profile, which would examples of true amixis and, as such, are systems in which also help account for the turbidity maximum. Interestingly, solute movement is predominantly a diffusion-controlled pro- bacterial abundance was 2-3 times higher in the deeper, anoxic cess. The lakes are situated in closed-basin drainages noted zones (12-18 meters) than in the aerobic zone, a phenomenon for their barren moonscape appearance and the nearly com- that has also been reported for a saline, meromictic, temperate plete absence of any plants and animals. Hydrologic recharge lake (Zehr et al. 1987) and may be the result of grazing by to the lakes occurs only during a 6-8-week period from glacial protozoan populations in the aerobic zone. meltwater and does not contain significant concentrations of In general, the microbial populations within Lake Fryxell dissolved organic matter or nutrients (Howard-Williams, Priscu, were physiologically stressed. The adenylate energy charge and Vincent 1989; Green, Angle, and Chave 1988). In essence, values for planktonic microorganisms throughout most of the these lakes may represent the natural environment that most depth profile were 0.5-0.6 (figure 2), indicative of a starving closely approximates a closed aquatic ecosystem; systems that or extremely stressed metabolism. The energy charge of a cell are controlled entirely by internal processes. estimates the energy potential of the cell in a manner analogous We are studying the biogeochemical processes affecting the to a battery potential; the theoretical range is 0-1, but the actual carbon, nitrogen, and sulfur cycles within one of these dry limits for pure cultures are 0.4-0.5 for dying or senescing cells valley lakes, Lake Fryxell (77°37S 163°8E). A part of the study and 0.75-0.8 for exponential growth (Atlas and Bartha 1981).

1990 REVIEW 233