Info

Climate and Hydrologic Variations and Implications for Lake and Stream Ecological Response in the McMurdo Dry Valleys, Antarctica

Kathleen A. Welch W. Berry Lyons Diane M. McKnight Peter T. Doran Andrew G. Fountain Diana Wall Chris Jaros Thomas Nylen Clive Howard-Williams

Introduction

Because polar regions may amplify what would be considered small to moderate climate changes at lower latitudes, Weller (1998) proposed that the monitoring of high latitude regions should yield early evidence of global climate change. In addition to the climate changes themselves, the connections between the polar regions and the lower latitudes have recently become of great interest to meteorologists and paleoclimatologists alike. In the southern polar regions, the direct monitoring of important climatic variables has taken place only for the last few decades, largely because of their remoteness. This of course limits the extent to which polar records can be related to low latitude records, even at multiyear to decadal timescales. Climatologists and ecologists are faced with the problem that, even though these high latitude regions may provide important clues to global climatic change, the lengths of available records are relatively short.

The McMurdo Dry Valleys Long-Term Ecological Research (MCM LTER) program was established in 1993. This program built on the monitoring begun in the late 1960s by researchers from New Zealand, who collected records of climate, lake level, and stream discharge in the Wright Valley, Antarctica. Griffith Taylor's field party obtained the first data related to lake level in 1903 as part of Scott's Discovery expedition. Analysis of the more recent data from the New Zealand Antarctic and MCM LTER programs when compared to the 1903 datum indicates that the first half of the twentieth century was a period of steadily increasing streamflows, followed in the last half of the century by streamflows that have resulted in more slowly increasing or stable lake levels (Bomblies et al. 2001). Thus, meteorological and hydrological records generated by the MCM LTER research team, when coupled with past data and the ecological information currently being obtained, provide the first detailed attempt to understand the connection between ecosystem structure and function and climatic change in this region of Antarctica. In addition, the program helps to fill an important gap in the overall understanding of climatic variability in Antarctica.

Even though most climatic records from the Antarctic continent are relatively short and/or lack associated biological monitoring to be useful in an ecological sense, a number of investigators have demonstrated interannual variations in Antarctic climate signals and responses. For example, Cullather et al. (1996) have shown that precipitation in west Antarctica covaries with ENSO, but the sign of the correlation changed in 1990. White and Peterson (1996) have speculated that a number of circum-Antarctic climatic parameters (i.e., sea level pressure, sea surface temperature, sea ice extent, and meridional wind stress), termed the Antarctic Circumpolar Wave (ACW), show interannual variability that may be related to ENSO. Sea ice extent in the Ross Sea region of Antarctica has also been shown to vary with the higher latitude ENSO signal (Ledley and Huang 1997). Finally, a detailed ice core record from West Antarctica, extending back in time approximately 1100 years, also shows dominant periodicities in chemical concentrations that are coincident with the Southern Oscillation Index (SOI) (Kreutz et al. 2000).

The primary emphasis of this chapter is on which types of ENSO or other mul-tiyear climate variations might be observed at the MCM site and how these variations impact the ecosystem of the dry valleys. These considerations are aided by a basic physical understanding of climate and hydrology linkages. The key climatic parameters influencing ecosystem structure and function in the McMurdo Dry Valleys are the ones that affect the physical state of water. Small interannual variations in summer temperatures, the number of days above freezing, and solar radiation can have a large impact because the availability of liquid water is such an important driver for the ecosystem (Fountain et al. 1999). Absorption of solar radiation by the surfaces and faces of the glaciers generates meltwater that either soaks into the alluvium of the streambeds or is carried by streams to the lakes in the valley floors. Meltwater generation depends on a fine balance of radiation, temperature, and the albedo of the glacier surface, which can be increased by snowfall or decreased by the input of eolian dust. Water is lost to the atmosphere from the stream and lake systems by sublimation from streambeds and ablation of the ice covers on the lakes. Because the transport and chemistry of water are the primary factors controlling habitat characteristics of the streams and lakes (Kennedy 1993), we focus our discussion on the aquatic components of the MCM ecosystem.

0 12 3 4 5 KILOMETERS

Figure 10.1 Map of Taylor Valley, Antarctica.

0 12 3 4 5 KILOMETERS

Figure 10.1 Map of Taylor Valley, Antarctica.

Site Description

The McMurdo Dry Valleys are the largest single ice-free expanse in Antarctica (~ 4800 km2). The valleys are a mosaic of glaciers, ephemeral streams, perennially ice-covered lakes, soils, and bedrock (Moorhead et al. 1999). They are among the driest and coldest deserts on the planet, with annual precipitation of <10 cm yr-1 and mean annual temperatures between -14.8 and -30.0°C on the valley floor at different locations (Doran et al. 2002a). Taylor Valley has been the focal point of MCM LTER research (figure 10.1). Taylor Valley consists of three major, closed-basin ice-covered lakes and about 25 streams (McKnight et al. 1999). For 4-10 weeks of the austral summer, the streams are fed by glacial meltwater from the surrounding glaciers (figure 10.1). Because the distribution of liquid water greatly influences the function and biodiversity of the MCM ecosystem, investigation of the role of climate variability has been a major emphasis of the MCM LTER (Fountain et al. 1999). To accomplish this, the delineation of the hydrologic budget of each subbasin within Taylor Valley (Bonney, Hoare, and Fryxell), has been undertaken, with glacier mass balance, stream discharge, and lake level variations being closely monitored. These, along with other data from MCM LTER, can be found at our web site, http://huey.colorado.edu. Descriptions of the methods used to collect meteorological data are in Doran et al. (1995). The methods used to collect stream dis-

Figure 10.2 Mean monthly temperatures from meteorological stations in the Lake Bonney, Hoare, and Fryxell basins.

charge data are described in Von Guerard et al. (1995) and McKnight et al. (1994). Detailed discussions of the Taylor Valley physiochemical and biogeochemical systems have recently been published in Priscu (1998) and volume 49, no. 12 of BioScience and are not repeated here.

Results and Discussion Meteorological Data

Mean monthly air temperatures for the three major lake basins in the Taylor Valley are shown in figure 10.2. Mean monthly air temperatures are at or below freezing even in the summer months. Temperatures are similar for all basins in the summer, with Lake Bonney being approximately 1°C warmer than Lake Hoare and Lake Fryxell. The occurrence of warm, dry katabatic winds flowing from the polar plateau and relatively cold easterly winds from the sea may explain the spatial differences in temperature (Clow et al. 1988). The Lake Bonney basin is the farthest inland and is most influenced by these winds from the polar plateau, whereas the influence is less pronounced nearer the coast. Mean daily air temperatures in the Lake Hoare basin (figure 10.3) have reached 2-4°C in the summer months during this period. The number of days above freezing also varies from year to year, as shown for the Lake Hoare basin. Even during the summer, there may be only a few days with temperatures above freezing (figure 10.3).

One major constraint on annual productivity of the aquatic ecosystem is the total darkness for approximately 4 months of the year in winter, from May through Au-

Lake Hoare

Lake Hoare

-50 ri 1 1111111111111

1993 1994 1995 1996 1997 1998

Figure 10.3 Mean daily air temperatures in the Lake Hoare basin.

gust. During the other seasons, there is interannual variability in the light regime. Solar radiation, as short wave radiation (SW), is shown for the Lake Hoare basin (figure 10.4). This parameter is important to monitor because it can influence the production of glacial melt. Water vapor and cloud cover are the primary variables responsible for the interannual variability in solar radiation (Dana et al. 1998). Because of the total darkness in winter, most of the variability in the ecosystem response will occur in the summer when there is sufficient solar radiation to drive photosynthesis and melt snow and ice. These seemingly small interannual variations in summer temperatures, the degree-days above freezing, and solar radiation can have a large impact on the availability of liquid water (Doran et al. 2002b; Fountain et al. 1999). The generation of liquid water on a glacier surface is determined by very small changes in surface temperature.

Stream Discharge

As part of the hydrologic monitoring component of the LTER, the major inflow streams in each of the three large lake basins in Taylor Valley are gauged. Stream discharge data are shown for selected streams in the Lake Bonney, Lake Hoare, and Lake Fryxell basins (figure 10.5). The inflow streams in the Lake Fryxell basin were monitored beginning in the austral summers of 1990-1991 and 1991-1992,

Lake Hoare Mean Daily Radiation

Lake Hoare Mean Daily Radiation

Figure 10.4 Mean daily short wave radiation for the Lake Hoare basin.

1993 1994 1995 1996 1997 1998 1999

Figure 10.4 Mean daily short wave radiation for the Lake Hoare basin.

before the start of the MCM LTER in 1993. The relatively high streamflows in the Lake Fryxell basin at that time have not recurred through the 2000-2001 flow season. The stream-gauging network was expanded to include the other lake basins in the Taylor Valley in 1993, and in 1994 the LTER project assumed responsibility for the Onyx River stations in the Wright Valley that had been operated by the New Zealand Antarctic Programme since 1968. These data are also available on the MCM LTER web site.

Since 1993, the patterns of interannual variations in streamflow have differed among the lake basins, and in some cases, within basins. The different response is in part related to differences in glacier position within the valley and to stream length (Fountain et al. 1999). In addition, studies of the radiation balance of the glaciers indicate that during low flow years more of the meltwater comes from the glacier faces rather than from the subhorizontal surfaces (Fountain et al. 1998). This is supported by more pronounced diurnal variation in streamflows during low flow periods, when peak flows are associated with the time of day that solar radiation directly impacts the face of the glaciers (Conovitz et al. 1998). Storage of water in the alluvium underneath and adjacent to the stream (referred to as the hyporheic zone) is a greater control on annual discharge in low flow years than in high flow years. Storage of water in the hyporheic zone can increase with stream length. During low flows years the shorter streams in the Lake Fryxell basin account for a greater pro-

Figure 10.5 Total stream discharge data for each summer for selected streams in the Lake Bonney, Lake Hoare, and Lake Fryxell basins.

Figure 10.5 Total stream discharge data for each summer for selected streams in the Lake Bonney, Lake Hoare, and Lake Fryxell basins.

portion of total inflow to the lake than they do in the high flow years (House et al. 1995; Conovitz et al. 1998).

Table 10.1 compares variations in annual discharge for Andersen Creek in the Lake Hoare basin to available meteorological data. As shown in figure 10.1, Andersen Creek is immediately adjacent to the west side of the Canada Glacier and flows through a short (200-m) delta before entering Lake Hoare. Algal mats are sparse in this short deltaic reach, and hyporheic zone interactions do not influence flow as much as they would in the longer streams in the Lake Fryxell and Lake

Bonney basins (Conovitz et al. 1998). Annual stream discharge is positively correlated to average summer temperature, degree-days above freezing, and solar radiation (both photosynthetically active radiation, [PAR] and SW) (table 10.1). Even maximum temperatures hover around the freezing point, and hence, small variations in solar radiation can also greatly influence glacier melt and subsequent streamflows (Lewis et al. 1998; Fountain et al. 1999). During the two coldest summers (1994-1995 and 1997-1998), the Lake Hoare average summer (December and January) temperatures were only -3°C. Solar radiation was also low, and streamflows were the lowest observed during this period. The highest streamflow occurred during 1998-1999. This was not the warmest summer, but the solar radiation was relatively high (table 10.1).

Variations in annual discharge for Canada and Crescent streams in the Lake Fryxell basin are also compared in table 10.1. Canada Stream drains the east side of the Canada Glacier, and the lower 1.5 km of the stream flows in a channel that is 0.2-0.5 km east of the glacier face. Crescent Stream is 5.4 km long and flows from the Crescent Glacier on the south side of the lake. Canada Stream has very abundant algal mats and mosses as does Crescent Stream in its upper reaches near the glacier source (Alger et al. 1997; McKnight, pers. comm.). During this period, Crescent Stream had much lower annual discharges than either Andersen Creek or Canada Stream, reflecting its greater length and greater storage of meltwater in the hyporheic zone. The annual discharge of Canada Stream is positively correlated with annual discharge in Andersen Creek to a greater extent than to Crescent Stream. The annual discharge in Canada Stream is also positively correlated with solar radiation and air temperature.

The correlations with PAR and air temperature are weaker for annual discharge in Crescent Stream, but the correlation with SW radiation is stronger. The highest annual discharge in Crescent Stream occurred in 1993-1994 when SW radiation was greatest, and the lowest annual discharges occurred during the two coldest summers. The weaker relationships with meteorological data in Crescent Stream are attributable to a greater influence of hyporheic zone processes in longer streams. The meltwater stored in the hyporheic zone supports growth of mosses and cyanobacterial mats along the edges of the streams, but does not directly contribute to the lake ecosystem. Storage in the hyporheic zone may also influence total discharge for a particular stream. Relatively short streams such as Andersen Creek should be influenced less by hyporheic zone storage, and therefore, should respond rapidly to local environmental variables, making them more important streams to monitor to assess hydrologic responses to climate variability.

Streamflow Variability and ENSO

Given that there are previously documented ENSO climate effects in Antarctica, we postulate that the interannual variability of meteorological parameters, and hence streamflow in the dry valleys, might be influenced by ENSO. As previously described, temperature, barometric pressure, solar radiation, and albedo on glacier surfaces are thought to affect glacier melt and hence streamflow in the dry valleys. How might the ENSO influence these variables in the region of the dry valleys?

Table 10.1 Meteorological data from the Lake Hoare basin and total stream discharge from Andersen Creek, Canada Stream and Crescent Streama

SW

Barometric

Degree-

Air

Andersen

Canada

Crescent

Flow

PAR

radiation

Pressure

days above

Temperature

Discharge

Discharge

Discharge

year

|imol s-1 m-2

W m-2

mbar

0°C

°C

103 m3

103 m3

103 m3

93/94

595

322

8.8

-2.0

54.0

130.0

39.0

94/95

534

273

976

9.0

-3.1

15.0

32.0

4.8

95/96

562

267

978

32.3

-1.5

41.5

109.8

8.6

96/97

573

271

982

17.4

-1.9

44.8

50.9

7.9

97/98

529

254

983

28.2

-3.3

28.0

74.2

5.4

98/99

600

288

978

24.7

-1.6

88.6

125.9

10.2

99/00

609

308

971

3.8

-2.6

SW

Barometric

Degree days

Air

PAR

radiation

Pressure

above 0°C

Temperature

Andersen

Canada

Crescent

PAR

1

SW RAD

0.81

1

Pressure

-0.57

-0.85

1

Degree-days

-0.38

-0.70

0.63

Temperature

0.60

0.27

-0.03

0.29

1

Andersen

0.89

0.48

-0.12

0.20

0.72

1

Canada

0.73

0.59

-0.19

0.28

0.63

0.75

1

Crescent

0.60

0.92

-0.18

-0.49

0.31

0.31

0.61

1

a Meteorological data are averages for December and January with the exception of degree-days above 0°C which is summed for the entire summer. A correlation matrix is included in the lower part of the table.

a Meteorological data are averages for December and January with the exception of degree-days above 0°C which is summed for the entire summer. A correlation matrix is included in the lower part of the table.

30 i—i—i—i—i-1—j—i—i—i—i-1—i—i—i—i—i—i—i i g—i—i-1—i—j-1—i—i-r

30 i—i—i—i—i-1—j—i—i—i—i-1—i—i—i—i—i—i—i i g—i—i-1—i—j-1—i—i-r

_50 * 1 1 ' 1 1 ' 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 * 1 1 1969/70 1974/75 1979/80 1984/85 1989/90 1994/95 1999/2000

Figure 10.6 Southern Oscillation Index (SOI) for January for the period 1969-2000. Strongly negative SOI indicates the El Niño (warm) phase, whereas strongly positive SOI indicates the La Niña (cold) phase of the ENSO.

To determine the possible manifestations of ENSO in the dry valleys, we need a better understanding of the climate mechanisms related to ENSO that might influence the regional climatic and hydrologic records. As mentioned, previous links between the SOI and Antarctic climate have been observed, for example, higher sea surface temperatures and reduced sea-ice concentrations in the Ross Sea region during El Niño years (Ledley and Huang 1997). Variations in sea-ice extent may influence moisture availability, as well as the extent of cloud cover and precipitation in the dry valleys. Lower sea-ice extent and warmer sea surface temperatures during El Niño may lead to cloudier and cooler conditions in the region as well as increased precipitation. The presence of fresh snow on the glacier surfaces can increase the surface albedo and effectively limit meltwater production from the glaciers (Fountain et al. 1998).

In addition to variation in sea-ice extent and sea surface temperatures, atmospheric circulation, which influences extreme warm and cold winter temperatures in the Antarctic Peninsula region, is thought to be related to ENSO (Marshall and King 1998). It is not clear how atmospheric circulation in the Ross Sea region would be influenced by ENSO. However, variations in circulation patterns could influence such things as storm tracks and the episodes of katabatic winds in the dry valleys, which in turn could influence temperature, cloud cover, precipitation, and solar radiation.

Figure 10.6 shows the Southern Oscillation Index (SOI) for January for the pe riod from 1969 to 2000. The SOI is calculated from normalized Tahiti and Darwin sea-level pressure anomalies. The SOI is negative during the El Niño (warm) phase and positive during the La Niña (cold) phase of the ENSO.

The high streamflows observed in the Lake Fryxell basin in 1990-1991 and 1991-1992 have not recurred through 2000-2001 (figure 10.5). The austral summer of 1991-1992 was an El Niño year, and streamflows in the Lake Fryxell basin were high. However, during the strong El Niño of 1997-1998, streamflows in the Lake Fryxell and Lake Hoare basins were relatively low (figure 10.5). The relatively high streamflows in Andersen Creek in 1995-1996, 1996-1997, and 19981999 coincided with a positive SOI (cold, La Niña) phase, whereas 1994-1995 and 1997-1998 were cooler summers with lower streamflows, and the January SOI was in a negative (warm, El Niño) phase. However, 1993-1994 was a year with near-neutral SOI, but with high streamflows in Andersen Creek, Canada Stream, and Crescent Stream. These records from the Taylor Valley span a short time period; therefore, it is difficult to quantify the relationship between streamflows and ENSO. In addition, the Lake Bonney, Lake Hoare, and Lake Fryxell basins, as well as other locations in the dry valleys, may exhibit a different climate response because of their relative position in the dry valleys and their slightly different microclimates (Fountain et al. 1999).

To further examine the possible connection between ENSO and climate in the dry valleys, longer time period records from this region of Antarctica are needed. The longest record of stream discharge in Antarctica is for the Onyx River, located approximately 20 km to the north of Taylor Valley in Wright Valley. The Onyx River is the longest river in Antarctica, flowing approximately 40 km from the Lower Wright Glacier inland to Lake Vanda (figure 10.7). In addition, mean summer (December and January) air temperature (figure 10.8) and mean summer (December and January) barometric pressure (figure 10.9) records are available for McMurdo Station, Antarctica, 100 km east of the dry valleys.

The annual discharge of the Onyx River varies by an order of magnitude during the period 1970-1999 (figure 10.7). The correlation matrix for the Onyx River discharge, McMurdo Station temperature and pressure, and the SOI shows that the discharge in the Onyx River is positively correlated to temperature and pressure and negatively correlated to the January SOI (table 10.2). This analysis suggests that during this time period higher discharges in the Onyx River may be associated with the El Niño phase of ENSO. None of the correlations is significant at the 95% confidence level. However, during summers when temperatures in the region are high, streamflow also tends to be high, although during some warm summers, for example, in 1974, streamflow is low. The McMurdo barometric pressure was also low in 1974, which could indicate lower incoming solar radiation because of increased cloud cover and storms in the region. As mentioned, the correlations between the SOI and McMurdo temperature and pressure are not statistically significant. However, when the SOI is low, both temperature and pressure tend to be higher.

Four of the five summers with the highest streamflows on the Onyx River (1971, 1985, 1987, and 1991) occurred when the SOI was near neutral. The SOI was low

Figure 10.7 Total stream discharge for each summer from the Onyx River for the period 1969-1999.

(El Niño) in the high flow summer of 1992. During the years of lowest flow, the SOI is neutral or high (La Niña). For example, the SOI was high in the low flow summers of 1974 and 1976.

This analysis is somewhat problematic because the discharge measured on the Onyx River near Lake Vanda is also influenced by other nonclimatic factors such as potential lag between the generation of meltwater and streamflow, as well as variability of storage in the hyporheic zone. However, there is an additional record of discharge on the Onyx River near the Wright Lower Glacier, the major source of meltwater for the Onyx River. The annual discharge records between the two sites are strongly correlated (R = 0.83, n = 23) during this period. Comparing data from McMurdo Station and the Wright Valley can also be somewhat problematic because of regional climate differences. Although McMurdo Station and the dry valleys are far enough apart that they can experience different weather, it is thought that the general trends in climate of the sort that might be influenced by ENSO should be similar throughout this region. The limited (more recent) meteorological data from Lake Hoare and Lake Vanda that can be compared to the records from McMurdo Station show that, although the mean summer temperatures are different, the trends are correlated (significant at 95% confidence level). The various records from the Ross Sea region, including the ones presented here from Taylor Valley, Wright Valley, and McMurdo Station can be used to examine the influence of ENSO on the regional climate of the southern Ross Sea.

1969/70 1974/75 1979/80 1984/85 1989/90 1994/95 1999/2000

Figure 10.8 Mean summer temperature recorded at McMurdo Station, Antarctica. The record represents the average of December and January mean monthly temperatures for each summer.

During El Niño years, perhaps conditions in the Ross Sea region are such that barometric pressures are high, allowing for relatively high solar radiation, warmer air and sea surface temperatures, and high streamflows. Alternatively, lower sea-ice concentrations or sea-ice extent that might be expected during El Niño (Ledley and Huang 1997) could lead to conditions that would not favor high incoming solar radiation as a result of increased moisture availability and cloud cover. At longer timescales, high lake stands during the last glacial maximum are attributed to cool and dry conditions, which are linked to fewer clouds, less snowfall, higher radiation, and more melt (Hall et al. 2001). Clearly, additional work is needed to elucidate possible linkages among the SOI, sea-ice, solar radiation, and other climate parameters.

Climate Variability and the Hydrologic Cycle Within the McMurdo Dry Valleys

It is well documented that the closed basin lakes within the Taylor and Wright valleys have waxed and waned since the last glacial maximum. The timing and evidence for these large-scale volume changes in the lakes will not be discussed here, as they have been reviewed recently (Doran et al. 1994; Hall et al. 2001; Lyons et al. 1998b). These fluctuations in the sizes of the lakes were brought about by subtle century- to millennial-scale climatic perturbations (Fountain et al., chapter 16 this volume; Hall et al.

1969/70 1974/75

1979/80

1984/85

1989/90 1994/95 1999/2000

Figure 10.9 Mean summer pressure recorded at McMurdo Station, Antarctica. The record represents the average of the December and January mean monthly temperature.

1969/70 1974/75

1979/80

1984/85

1989/90 1994/95 1999/2000

Figure 10.9 Mean summer pressure recorded at McMurdo Station, Antarctica. The record represents the average of the December and January mean monthly temperature.

2001). Similar century-scale climate variations have been observed in the Antarctic coastal marine system (Leventer et al. 1996; Smith et al. 1999).

Next, we will examine the impact short-term climatic variations might have on the McMurdo Dry Valleys ecosystem. Again, our discussion will be based on the fact that the McMurdo Dry Valleys ecosystem is sensitive to very small variations in climate because the change between solid and liquid water is delicately poised, and thus small changes in temperature and radiant energy are amplified by large nonlinear changes in the hydrologic budgets that can cascade through the system.

Table 10.2 Correlation matrix for the Onyx River discharge, McMurdo mean summer (Dec-Jan) temperature and pressure, and January SOI for the period 1970 to 1996 (N=28)

Onyx River McMurdo mean McMurdo mean discharge summer temperature summer pressure January SOI

Discharge 1

Temperature 0.21 1

Pressure 0.27 0.13 1

Impact on Stream Ecosystems: Cascade I

Glacial melt and subsequent streamflow may respond differently to several consecutive warm and sunny days than they would to many days that hover near freezing. The result of differing flow rates may lead to the same volume of water generated over an entire summer season, however, the response of the algal community in some stream habitats can be significantly different for a short-duration pulse of water versus a more continuous low flow. The volume of meltwater per unit time from the glacier has a direct impact on the stream ecosystems. In addition, glacial routing differences, rapid versus slow melting rates on the glacier, and variations in hyporheic exchange with discharge affect water quality, including nutrient distribution, hence they influence biogeochemical processes within the streams.

The algal mats in the streams are composed of cyanobacteria and persist from summer to summer in a freeze-dried state (Vincent and Howard-Williams 1986). The stream reaches with high abundance of algal mats typically have moderate gradients and a streambed armored by a stone pavement. The stream reaches with less or sparse algal mats are typically steep gradient or sandy deltaic reaches that do not have an armored streambed. These latter reaches are turbid and sediment laden during high flow events, which can scour the streambeds, limiting the accumulation of algae to form a mat (Vincent and Howard-Williams 1986; Howard-Williams et al. 1986; Alger et al. 1997; McKnight et al. 1998). Our stream algal surveys in 1994 (2 years after the high flows in 1990-1991 and 1991-1992) and in 1998 (after 6 years of lower flow) indicate that stream reaches with steep gradients or deltaic reaches with shallow gradients but unstable substrates actually accumulate pho-totrophic biomass as well as algal diversity under low flow conditions (McKnight et al. 1998). We attribute these increases to recovery from the effects of scouring in 1990-1992 during the current period of low and stable streamflows. Stream morphology changed as well, with a widening of stream channel and redistribution of surface flow paths. The primary phototrophs that expand their coverage of the streambed during low flow conditions are orange algal mats and mosses in some locations. However, the stream reaches that have a stone pavement and support abundant algal mats have stable mat communities and do not change much with in-terannual variation in streamflow.

In addition, although decreased flow may reduce the total amount of nutrients passing through the stream ecosystem, it allows for more efficient uptake of nutrients because of slower flow rates and greater contact with algal mats within the stream environment. Our modeling of nitrogen uptake in the streams of the Lake Fryxell basin indicates that uptake rates are extremely sensitive to flow regime in the streams (Moorhead and Priscu 1998), with higher uptake at lower discharges. In addition, there is the influence of diel changes in flow on nutrient concentrations. Concentrations increase at night and early morning as flows decrease. These changing concentrations must influence nutrient uptake through straightforward Michaelis Menten kinetics. All these data suggest that stream algal communities may be more viable and more efficient at removing nutrients at lower flow rates when the hy-drologic connectivity (e.g., from glacier to lake) and sediment transport are minimized. This finding implies that low streamflow years lead not only to increased

50 100 150 200 250

0 —'—>—'—i—i—"—i—i—i—i—i—i—i—i—i—i—i—i—i—

50 100 150 200 250

0 —'—>—'—i—i—"—i—i—i—i—i—i—i—i—i—i—i—i—i—

Figure 10.10 Cl versus depth at the beginning of the austral summer at Lake Hoare from 1993 to 1999.

biomass within the streams, but also to decreased nutrient and suspended sediment load concentrations to the lakes.

Impact on Lake Ecosystems: Cascade II

To evaluate the role of hydrologic and, hence, climatic variability on the lake environments, we will focus our attention on Lake Hoare, the least saline of the Taylor Valley lakes. Figure 10.10 shows a time series of Cl versus depth during early summer (i.e., prior to the initiation of streamflow) from the same location in Lake Hoare. Over the period 1993-1999, the surface water Cl (and hence total dissolved solids, TDS) has increased as a result of decreased inputs of glacial meltwater during this time period (Welch et al. 2000). As described previously, water is lost from these lakes through ablation of the perennial ice covers. During the winter, new ice is accreted onto the bottoms on the ice covers, and salts are exsolved from the ice matrix. Unless the water loss via sublimation is balanced by an input of freshwater, this input of salt will lead to an increase in salinity in the upper water column and a decrease in water column stability over time. This is clearly what we have observed in Lake Hoare from 1993 to 1999 during these relatively low flow years (figure 10.10).

Lyons et al. (1998a) noted that geochemistry data from Lake Fryxell in the Mc-Murdo Dry Valleys indicated an overall change in the ionic composition of the lakes when data collected in the mid-1990s are compared to older, but reliable, data obtained in the early 1960s. Our recent data demonstrated an increase in Ca:Cl, HCO3:Cl, and SO4:Cl ratios during this time period (30-35 years). This change in ratios is attributed to an increase of weathering products into the lake because of increased streamflows from the 1970s to 1991 (Lyons et al. 1998a). No matter what the cause, it is clear that variations in the hydrologic cycle within the McMurdo Dry Valleys do impart measurable geochemical changes in the streams and lakes.

In addition to changes in major ion chemistry, low streamflows also lead to a decrease in nutrient fluxes in the lakes, which are accentuated by increased algal mat uptake of fixed nitrogen species within the streams, as detailed previously. In the Taylor Valley lakes, inorganic nutrients such as nitrate, ammonium, and phosphate are low or below detection in the upper part of the water column prior to stream-flow. This has an important consequence because there is little nutrient available to sustain primary production when the lakes first receive light in the spring. Another important consequence of lower streamflows is a decrease in CO2 input into these lakes. Unlike most lakes in other parts of the world that are supersaturated with CO2 in their surface waters (Cole et al. 1994), the Taylor Valley lakes can be extremely undersaturated in CO2 (with respect to the atmosphere), because of their ice covers and the lack of terrestrial organic carbon input (Neumann et al. 2001). Decreased streamflow enhances this undersaturation and may lead to CO2 limitation (Neumann et al. 2001).

Phytoplankton enumeration from 1989 to 1999 in Lake Fryxell has allowed us to evaluate the influence of streamflow and climate on the phytoplankton populations (McKnight et al. 2001). During low flow years, Chroomonas lacustris is the dominant algal species, whereas in high flow years, Cryptomonas sp. is dominant. In addition, filamentous cyanobacteria have a maximum abundance at depth during low flow years. Because phytoplankton populations persist during the winter in the upper water column, although at a lower concentration than in the summer (McKnight et al. 2000), the species distribution of the phytoplankton in early summer may be influenced by conditions during the previous summer as well as by the nutrient and light regime at that time. Although we are just beginning to digest and analyze these data, clearly the variability in the hydrologic cycle greatly influences the biological communities of these lakes (Doran et al. 2002b).

Impacts on Soil Ecosystems: Cascade III

We have primarily focused our discussion on the aquatic systems within the McMurdo Dry Valleys as they may be affected by ENSO-type climatic fluctuations, but soil environments may also be influenced by short-term climate variations. Tre-onis et al. (2000) have recently observed that soil nematodes are in a state of an-

hydrobiosis in soils with moisture contents less than 2%. Wetting soils to 12% leads to the animals coming out of anhydrobiosis within 6 hours (Treonis et al. 2000). Recent snow patch experiments by our group initiated to simulate increased snow input to the landscape have shown that soils covered by increased snowpack can have moisture contents as high as 12%, compared to nearby control soils with 0.2 to 0.7% moisture (Gooseff et al. 2000). Therefore, variations in the amount of precipitation within the McMurdo Dry Valleys brought about by both short- and long-term climate variability could have an important impact on soil moisture and hence changes in population dynamics of the dominant animals in the dry valleys. Ongoing research will continue to evaluate this concept.

Summary of Potential and Observed Cascades through the Ecosystem

The key climatic parameters influencing ecosystem structure and function in the McMurdo Dry Valleys are the ones that affect the conversion of solid to liquid water. These include, but are not limited to, temperature, solar radiation, and precipitation changes. Variations in meltwater production on the glaciers could be due to changes in solar radiation as a function of cloud cover and/or storminess, surface albedo of the glaciers, and temperature. Variations in melt will then impact stream discharge, which, in turn, can impact the distribution, species composition, and nutrient uptake rate of the algal mats within the stream channels. Fluxes of water, as well as dissolved and particulate matter, into the lakes will also be affected. This, in turn, will lead to chemical and biological changes within the lakes themselves. Thus the short-term climate fluctuations that have been observed can have multiyear impacts on the aquatic ecosystems of the dry valleys (Doran et al. 2002b).

Although it is still not clear just how ENSO and other types of climatic phenomena influence the climate of the dry valleys, it is clear that even very small changes in climate can play an important role in the ecosystem dynamics of the McMurdo Dry Valleys. This is evident from the information presented within this chapter and that of Doran et al. (2002b), as well as from our long-term records (e.g., see Fountain et al. 2003, chapter 16). Lyons et al. (2000) termed the McMurdo Dry Valleys an "unstable" system because the recovery from past climatic disturbance is quite long. Its sensitivity to small-scale (by temperate latitude measures) climate variability is an extraordinary feature of the McMurdo Dry Valley ecosystem. Although the records are not long enough to extract an ENSO signal from the clima-tological and hydrological data from Taylor Valley, our primary LTER site, there is little doubt that variations in local sea-ice extent, which has been related to ENSO, can have a major impact on climate and water budgets in the McMurdo Dry Valleys as well (e.g., Welch et al. 2000; Hall et al. 2001). Model simulations under a global warming scenario indicate an increasingly positive Antarctic Oscillation (AAO), or in other words, lower barometric pressures over the polar regions (Fyfe et al. 1999). Therefore, our understanding of the impact of storminess, increased snowfall, and albedo change may become more important in attempting to predict the impact of climate change on the McMurdo Dry Valley system.

Conclusion

A number of conclusions can be drawn from this research. Data collected by the MCM LTER scientists since 1993 clearly demonstrate that small variations in climate, especially in the austral summer, can have a significant multiyear impact on the ecosystem. Although we have observed interannual climate variability in the dry valleys region of Antarctica, it is unclear whether it is related to ENSO, the AAO, the ACW, or some other climatic forcing. Our record from Taylor Valley is simply too short to discern the long-term trends. The longer term climate records from the McMurdo region and Onyx River discharge may vary with the SOI; however, the correlations are not statistically significant. Nonetheless, the available records indicate that high streamflows in the Onyx River in Wright Valley occur during neutral to low SOI and that low streamflows occur during neutral to high SOI, suggesting the influence of ENSO. However, stream discharge in Andersen Creek in Taylor Valley seems to be positively correlated to SOI, with higher flows occurring during the positive (La Niña) phase of the SOI. This difference between discharge and SOI between the Onyx River and Andersen Creek probably represents differences in the physical characteristics of these systems. Much larger storage capacity in the hyporheic zone of the Onyx River mediates the flow. If this is true, the monitoring of the shorter, higher gradient streams like Andersen Creek may lead to better records of climate teleconnections. Clearly, additional work is needed to elucidate possible linkages among the SOI, sea ice, solar radiation, and other climatic parameters in this region of Antarctica, and longer climate records are needed.

The variability in climate manifested through changes in the hydrologic cycle within the McMurdo Dry Valleys cascade through the aquatic portion of the ecosystem. Variations in temperature, snowfall, and solar radiation influence melt-water production on the glaciers. The resulting streamflow influences stream mi-crobial populations and morphology. The closed basin lakes that are fed by the streams can respond in a variety of ways, including changes in phytoplankton populations, which have been observed to be different in high and low streamflow years, possibly driven by changes in lake water chemistry. We speculate that climate variability can also cascade through the soil environment as well. Only when longer climatic records from the McMurdo Dry Valleys and from other parts of Antarctica are available will scientists be able to evaluate the overall ecological impact of ENSO climate variations in the southern polar regions.

Acknowledgments This work was supported by NSF grants OPP-9211773 and OPP-9813061. We thank David Greenland for allowing us to participate in the LTER ASM workshop that was the original focus of this work. We thank all our MCM LTER colleagues who helped with collection and analysis of these data. We thank the New Zealand National Institute of Water and Atmospheric Research Ltd. and particularly Pete Mason and Kathy Walter for hydrological data analysis on the Onyx River. SOI data were obtained from The National Centers for Environmental Prediction (NCEP) Climate Prediction Center web site. Meteorological data for McMurdo Station were compiled from the Antarctic Journal of the United States. We appreciate the discussions regarding short-term climatic variations in

Antarctica with Matt Lazara, University of Wisconsin, Paul Mayewski, University of Maine, and Chris Shuman, University of Maryland. We also thank Carmen Nezat for reviewing the original manuscript. We especially thank two anonymous reviewers, whose thoughtful suggestions greatly improved the manuscript.

References

Alger, A. S., D. M. McKnight, S. A. Spaulding, C. M. Tate, G. H. Shupe, K. A. Welch, R. Edwards, E. D. Andrews, and H. R. House. 1997. Ecological processes in a cold desert ecosystem: The abundance and species distribution of algal mats in glacial meltwater streams in Taylor Valley, Antarctica. University of Colorado, Institute of Arctic and Alpine Research, Occasional Paper, 51. 108pp.

Bomblies, A., D. M. McKnight, and E. D. Andrews. 2001. Retrospective simulation of lake level rise in Lake Bonney based on recent 21-year record: Indication of recent climate change in the McMurdo Dry Valleys, Antarctica. Journal of Paleolimnology 25(4): 477-492.

Cole, J. J., N. F. Caraco, G. W. Kling, and T. K. Kratz. 1994. Carbon dioxide supersaturation in the surface waters of lakes. Science 265:1568-1570.

Clow, G. D., C. P. McKay, G. M. Simmons, Jr., and R. A. Wharton. 1988. Climatological observations and predicted sublimation rates at Lake Hoare, Antarctica. Journal of Climatology 1:715-728.

Conovitz, P. A., D. M. McKnight, L. H. MacDonald, A. G. Fountain, and H. H. House. 1998. Hydrologic processes influencing streamflow variation in Fryxell basin, Antarctica. Pages 93-108 in J. C. Priscu, editor, Ecosystem Dynamics in a Polar Desert: The McMurdo Dry Valleys, Antarctica. American Geophysical Union Antarctic Research Series 72, Washington, D.C.

Cullather, R. I., D. H. Bromwich, and M. L. Van Woert. 1996. Interannual variations in Antarctic precipitation related to El Nino-Southern Oscillation. Journal of Geophysical Research D 101:19109-19118.

Dana, G. L., R. A. Wharton, and R. Dubayah. 1998. Solar radiation in the McMurdo Dry Valleys, Antarctica. Pages 39-64 in J. C. Priscu, editor, Ecosystem Dynamics in a Polar Desert: The McMurdo Dry Valleys, Antarctica. American Geophysical Union Antarctic Research Series 72, Washington, D.C.

Doran, P. T., G. Dana, J. T. Hastings, and R. A. Wharton. 1995. The McMurdo LTER Automatic Weather Network (LAWN). Antarctic Journal of the United States 30(5):276-280.

Doran, P. T., C. P. McKay, G. D. Clow, G. L. Dana, A. G. Fountain, T. Nylen, and W. B. Lyons. 2002a. Valley floor climate observations from the McMurdo Dry Valleys, Antarctica, 1986-2000. Journal of Geophysical Research, Atmospheres, 10.1029/ 2001JD002045.

Doran, P. T., J. C. Priscu, W. B. Lyons, J. E. Walsh, A. G. Fountain, D. M. McKnight, D. L. Moorhead, R. A. Virginia, D. H. Wall, G. D. Clow, C. H. Fritsen, C. P. McKay, and A. N. Parsons. 2002b. Antarctic climate cooling and terrestrial ecosystem response. Nature 415:517-520.

Doran, P. T., R. A. Wharton, Jr., and W. B. Lyons. 1994. Paleolimnology of the McMurdo Dry Valleys, Antarctica. Journal of Paleolimnology 10:85-114.

Fountain, A. G., G. L. Dana, K. J. Lewis, B. H. Vaughn, and D. McKnight. 1998. Glaciers of the McMurdo Dry Valleys, Southern Victoria Land, Antarctica. Pages 65-75 in J. C. Priscu, editor, Ecosystem Dynamics in a Polar Desert: The McMurdo Dry Valleys,

Antarctica. American Geophysical Union Antarctic Research Series 72, Washington, D.C.

Fountain, A. G., W. B. Lyons, M. B. Burkins, G. L. Dana, P. T. Doran, K. J. Lewis, D.M. McKnight, D. Moorhead, A. N. Parsons, J. C. Priscu, D. H. Wall, R. A. Wharton, Jr., and R. A. Virginia. 1999. Physical controls on the Taylor Valley Ecosystem Antarctica. BioScience 49(12):961-971.

Fyfe, J. C., G. J. Boer, and G. M. Flato. 1999. The Arctic and Antarctic oscillations and their projected changes under global warming. Geophysical Research Letters 26:1601-1604.

Gooseff, M. N., J. E. Barrett, and P. Doran. 2000. Seasonal snow pack coverage and its influence on soils of a polar desert. ESA meeting, February 2001, Santa Fe, NM. Abstracts with Program.

Hall, B. L., G. H. Denton, and B. Overturf. 2001. Glacial Lake Wright, a high-level Antarctic lake during the LGM and early Holocene. Antarctic Science 13:53-60.

House, H. R., D. M. McKnight, and P. von Guerard. 1995. The influence of stream channel characteristics on streamflow and annual water budgets for lakes in the Taylor Valley. Antarctic Journal of the United States 30(5):284-287.

Howard-Williams, C., C. L. Vincent, P. A. Broady, and W. F. Vincent. 1986. Antarctic stream ecosystems: Variability in environmental properties and algae community structure. Internationale Revue der gesamten Hydrobiologie 71:511-544.

Kennedy, A. D. 1993. Water as a limiting factor in the Antarctic terrestrial environment: A biogeographical synthesis. Arctic and Alpine Research 25:308-315.

Kreutz, K. J., P. A. Mayewski, I. I. Pittalwala, L. D. Meeker, M. S. Twickler, and S. I. Whitlow. 2000. Sea-level pressure variability in the Amundsen Sea region recorded in a West Antarctic glaciochemical record. Journal of Geophysical Research 105(D3):4047-4059.

Ledley, T. S., and Z. Huang. 1997. A possible ENSO signal in the Ross Sea. Geophysical Research Letters 24:3253-3256.

Leventer, A., E. W. Domack, S. E. Ishman, S. Brachfeld, C. E. McClennen, and P. Manley. 1996. Productivity cycles of 200-300 years in the Antarctic Peninsula region: Understanding linkages among the sun, atmosphere, oceans, sea ice, and biota. Geological Society of America Bulletin 108:1626-1644.

Lewis, K. J., A. G. Fountain, and G. L. Dana. 1998. Surface energy balance and meltwater production for a Dry Valley glacier, Taylor Valley, Antarctica. Annals of Glaciology 27:603-609.

Lyons, W. B., A. Fountain, P. T. Doran, J. C. Priscu, K. Neumann, and K. A. Welch. 2000. Importance of landscape position and legacy: The evolution of the lakes in Taylor Valley, Antarctica. Freshwater Biology 43:355-367.

Lyons, W. B., S. W. Tyler, R. A. Wharton, Jr. D. M. McKnight, and B. H. Vaughn, 1998a, A late Holocene desiccation of Lake Hoare and Lake Fryxell, McMurdo Dry Valleys, Antarctica. Antarctic Science 10:247-256.

Lyons, W. B., K. A. Welch, K. Neumann, J. K. Toxey, R. McArthur, C. Williams, D. M. McKnight, and D. Moorhead. 1998b. Geochemical linkages among glaciers, streams and lakes within the Taylor Valley, Antarctica. Pages 77-92 in J. C. Priscu, editor, Ecosystem Dynamics in a Polar Desert: The McMurdo Dry Valleys, Antarctica. American Geophysical Union Antarctic Research Series 72, Washington, D.C.

Marshall, G. J., and J. C. King. 1998. Southern Hemisphere circulation associated with extreme Antarctic Peninsula winter temperatures. Geophysical Research Letters 25:24372440.

McKnight, D., H. House, and P. Von Guerard. 1994. McMurdo LTER: Streamflow measurements in Taylor Valley. Antarctic Journal of the United States 29(5):230-232.

McKnight, D. M., A. Alger, C. M. Tate, G. Shupe, and S. Spaulding. 1998. Longitudinal patterns in algal abundance and species distribution in meltwater streams in Taylor Valley, Southern Victoria Land, Antarctica. Pages 109-127 in J. C. Priscu, editor, Ecosystem Dynamics in a Polar Desert: The McMurdo Dry Valleys, Antarctica. American Geophysical Union Antarctic Research Series 72, Washington, D.C.

McKnight, D. M., B. L. Howes, C. D. Taylor, and D. D. Goehringer. 2000. Phytoplankton dynamics in a stably stratified Antarctic lake during winter darkness. Journal of Phy-cology 36:852-861.

McKnight, D. M., D. K. Niyogi, A. S. Alger, A. Bomblies, P. A. Conovitz, and C. M. Tate. 1999. Dry Valley streams in Antarctica: ecosystems waiting for water. BioScience 49(12):985-995.

McKnight, D. M., E. VonMaytre, J. Priscu, and W. B. Lyons. 2001. Response of an Antarctic lake ecosystem to climate variation: Linkages between phytoplankton species dynamics and streamflow. ASLO meeting, Albuquerque, NM, February 12-16. Abstracts with the Program.

Moorhead, D. L., P. Doran, A. G. Fountain, W. B. Lyons, D. M. McKnight, J. C. Priscu, R. A. Virginia, and D. H. Wall. 1999. Ecological legacies: Production, persistence and influence on soil and lake ecosystems of the Antarctic dry valleys. BioScience 49(12):1009-1019.

Moorhead, D. L., and J. C. Priscu. 1998. Modeling nitrogen transformations in Dry Valley streams, Antarctica. Pages 141-151 in J. C. Priscu, editor, Ecosystem Dynamics in a Polar Desert: The McMurdo Dry Valleys, Antarctica. American Geophysical Union Antarctic Research Series 72, Washington, D.C.

Neumann, K., W. B. Lyons, J. C. Priscu, and R. J. Donahoe. 2001. CO2 concentrations in perennially ice-covered lakes of Taylor Valley, Antarctica, Biogeochemistry 56:27-50.

Priscu, J. C. 1998. Ecosystem Dynamics in a Polar Desert: The McMurdo Dry Valleys, Antarctica. American Geophysical Union, Antarctic Research Series 72, Washington,

Smith, R. C., D. Ainley, K. Baker, E. Domack, S. Emslie, B. Fraser, J. Kennett, A. Leventer,

E. Mosley-Thompson, S. Stammerjohn, and M. Vernet. 1999. Marine ecosystem sensitivity to climate change. BioScience 49(12):393-404.

Treonis, A. M., D. H. Wall, and R. A. Virginia. 2000. The use of anhydrobiosis by soil nematodes in the Antarctic dry valleys. Functional Ecology 14:460-467.

Vincent W. F., and C. Howard-Williams. 1986. Antarctic stream ecosystems: Physiological ecology of a blue green algal epilithon. Freshwater Biology 16:219-233.

Von Guerard, P., D. M. McKnight, R. A. Harnish, J. W. Gartner, and E. D. Andrews. 1995. Streamflow, water-temperature, and specific-conductance data for selected streams draining into Lake Fryxell, Lower Taylor Valley, Victoria Land, Antarctica, 1990-92. U.S. Geological Survey Open-File Report 94-545.

Welch, K. A., K. Neumann, D. M. McKnight, A. Fountain, and W. B. Lyons. 2000. Chemistry and lake dynamics of the Taylor Valley lakes, Antarctica: The importance of long-term monitoring. Pages 282-287 in W. Davison, C. Howard-Williams, and P. Broady, editors, Antarctic Ecosystems: Models for Wider Ecological Understanding. New Zealand Natural Sciences, Christchurch.

Weller, G. 1998. Regional impacts of climate change in the Arctic and Antarctic, Annals of Glaciology 27:543-552.

White, W. B., and R. G. Peterson. 1996. An Antarctic circumpolar wave in surface pressure, wind, temperature and sea-ice extent. Nature 380(25):699-702.

Was this article helpful?

0 0

Post a comment