The McMurdo Dry Valleys (MCM), the largest ice-free region on the Antarctic continent (~4800 km2), is a mosaic of perennially ice-covered lakes, ephemeral streams, soils and glaciers (Fig. 2.1). The MCM is one of the driest and coldest ecosystems on earth (Fig. 2.2; Table 2.1) and harbors a distinctive biota capable of surviving in the diverse habitats of the MCM (Priscu, 1999). The most complex life forms in the MCM are small invertebrates. Other eukaryotic phyla included protozoans, fungi and microalgae found in glacier surfaces, in soils and in streams and lakes.

Research in MCM-I (1993-1999) demonstrated that physical constraints control the structure and function of this polar desert (Fountain et al., 1999). We discovered that subtle changes in temperature, precipitation, and albedo have a major influence on the hydrologic cycle, biogeochemistry and productivity within the valleys. For example, a decadal scale cooling significantly reduced stream flow, lowered lake levels, and increased lake ice thickness thus reducing lake primary productivity, and decreasing soil biomass (Doran et al., 2002b). The unusually warm austral summer of 2001-2002 greatly increased glacier melt, which rapidly restored lake levels to those that existed at the beginning of MCM-I. The increased meltwater enhanced productivity and related biogeochemical processes across the landscape units (Foreman et al., in review). These events underscore the sensitivity of the MCM ecosystem to small variations in climate and the importance of the transition between solid and liquid water phases. Consequently, small changes in temperature and solar energy are amplified by large, non-linear changes in hydrologic budgets that cascade through the ecosystem (Welch et al., 2003).

The central hypothesis for MCM-II (1999-2005) was that "legacies" of past climatic events are a major factor regulating contemporary ecosystem processes (Moorhead et al., 1999; Lyons et al., 2000; Burkins et al., 2000). The term "legacy" refers to the carry-over or ecosystem "memory" of past events (Vogt et al., 1997). For example, during MCM-II we showed that the presence of Lake Washburn, which inundated Taylor Valley (TV) between 24,000-6,000 yr BP (Denton et al., 1989), followed by a subsequent cold, dry period ending about 1000 yrs BP (Lyons et al., 1998a), created old pools of organic carbon and nutrients in the soils and lakes that explain many of the ecosystem processes we now observe (Fig. 2.3). Legacy effects at other LTER sites are typically viewed as human-induced disturbance such as forest cutting (HBR, HFR, CWT, AND), conversion to agriculture (KBS), or shorter-term climate events including desertification (JRN, SEV, SGS) and catastrophic disasters, like hurricanes (LUQ).

Another result of MCM-II demonstrated that habitat suitability, overprinted by wind and water transport, largely dictates biodiversity and community structure. Many of the phototrophs in the soils and streams may represent the legacy of benthic mat communities deposited by paleolakes, whereas other species reflect more contemporary origins (Brambilla et al., 2001; Nadeau and Castenholz, 2001). Similar phylotypes exist within stream mats, lake ice, and glacier cryoconite holes (Gordon et al., 2000; Christner et al., 2003) implying that wind dispersion may be key in regulating biodiversity among landscape units in the MCM. Conversely, we also observe extensive patchiness throughout the MCM. For example, 40% of the soils in TV (Freckman and Virginia, 1998; Virginia and Wall, 1999) and the cryoconite holes on the glacier surfaces lack nematode communities (Porazinska et al., in press). Clearly dispersion, particularly by wind, is important. But, the ability of organisms to establish a population also can be strongly dependent on physical and chemical habitat factors (i.e. water, salinity and carbon). These results raise important questions concerning the relation between biodiversity and ecosystem function. A major goal for MCM-III will be to understand both how the environment controls the diversity of organisms and how diversity itself controls the functioning of the MCM ecosystem. This is one of the hottest topics in modern ecological research (Pfisterer and Schmid, 2002; Chase and Liebold, 2002), and the MCM lend themselves to answering these questions in a unique way.

Limnological Background

Research in MCM-I and II showed that the Taylor Valley (TV) lakes have highly contrasting geochemistry and evolutionary histories (Fig. 2.7) (Lyons et al., 1998a,b; 2000). We contend that contemporary phytoplankton and bacterioplankton production in the deep water of the lakes is controlled by the upward diffusion of relict or legacy nutrients, whereas near surface assemblages are fueled by nutrient loading by extant stream flow (Priscu, 1995; Takacs et al., 2001). P:R ratios in the water columns of certain lakes are < 1 implying that ancient carbon is supporting much of the present day biological activity, particularly in the deeper waters (Priscu et al., 1999). This is in contrast to more temperate region lakes where P:R < 1 results from contemporary input of terrestrial carbon (Pace et al., 2004). Despite the importance of resource legacy, relatively low stream nutrient loads from 1992 and 2002 presumably have influenced phytoplankton biodiversity by reducing the cryptophyte populations (Tursich, 2003; Fig. 2.8). Consequently, both contemporary and ancient resources balance the activity and biodiversity of microbial populations within the lakes.

Given the microbial dominance in the MCM, we have begun to utilize molecular methods to address questions related to prokaryotic diversity and ecosystem function. Preliminary results showed that bacterial density, activity and diversity change significantly at the oxic/anoxic interface in the water column of Lake Fryxell. Molecular data also show differences in vertical structure of Lake Bonney and that the majority of phylotypes identified in the ice cover differ from those in the water column implying that aeolian deposition does not directly regulate the biodiversity of the water. The planktonic food web is well developed and consists of photosynthetic nanoflagellates, cyanobacteria, bacterioplankton, heterotrophic nanoflagellates, ciliates and rotifers (Priscu et al., 1999; Roberts et al., in press).

Research on benthic processes in TV lakes is logistically difficult because of the requirements for SCUBA diving beneath the perennial ice covers. In Lake Hoare, benthic production rivals that of the water column (Hawes and Schwarz, 2001); however, the contribution of the benthos to whole lake productivity depends largely on the lake morphology. In lakes with a low surface area to volume ratio (e.g. Lake Bonney), the importance of benthic primary productivity is less than in lakes with a higher ratio (e.g. Lake Fryxell).

Hypotheses

CENTRAL HYPOTHESIS: Biodiversity and ecosystem structure and function in the MCM are dictated by the interactions of climatic legacies with contemporary biotic and physical processes.

1. Hydrology

The availability of liquid water is the overriding factor constraining biological activity in the MCM. Past variations in water balance produced the legacies we now observe. The relationships between biodiversity and ecosystem function in the MCM requires understanding hydrological responses to climate change, and we will continue to study climate controls on the hydrology. Our new focus in MCM-III will be on transport of organic and inorganic material by water and wind, and its influence on biological diversity and stoichiometric (C:N:P) relationships.

Hypothesis 1a. The variability of glacier melt depends on the interaction of climate and landscape position, and sediment on the glacier surfaces.
Hypothesis 1b. Meltwater in interconnected subsurface passages on the glaciers creates cryoconite habitats, and biological processes in these habitats influence aquatic geochemistry throughout the MCM.
Hypothesis 1c. At the stream/soil interface, thawing of the active layer in summer drives the expansion of the hyporheic zone, increasing storage of water, solute flux from weathering reactions and microbial cycling of nutrients in the streams.
Hypothesis 1d. Spatial variation in the hydrologic characteristics of the hyporheic zone creates a mosaic of biogeochemically heterogeneous subzones.

2. Biodiversity

Biodiversity in many temperate and tropical systems is largely governed by biological processes, including growth rate, competition, predation, and human disturbance (Luck et al., 2004). The MCM environment constrains biological processes in all landscape units. Given these constraints, and the absence of large organisms capable of active immigration, we contend that fluvial and aeolian transport may be the primary dispersive agents for organisms in the MCM. Fluvial transport carries organisms from the cryoconite holes on the glacier (cryoconites), through streams to the lakes, and may explain the recent establishment of cyanobacterial populations in Lake Fryxell (Spaulding et al., 1994). The constant and often strong winds in the MCM (Doran et al., 2002a; Nylen et al., in press) are another powerful dispersive agent for organisms. We propose that the biodiversity of the MCM is sustained by physical processes and modified by growth and survival rates in habitats of varying quality. Within landscape units, biodiversity is affected by spatial heterogeneity in the physiochemical characteristics of the habitat. This presence or absence of habitat suitability is primarily driven in the MCM by past legacy effects. Therefore, the distribution of organisms within the landscape units of the MCM reflects both ecological legacies and contemporary physical processes.

Hypothesis 2a. Stream and lake microbial mats provide seed populations to the contemporary MCM landscape.
Hypothesis 2b. Diversity of stream microbial mats is controlled by flood frequency, streambed stability, and small-scale patchiness related to hyporheic exchange.
Hypothesis 2c. The diversity of soil invertebrate communities is determined by soil legacies and hydrology.
Hypothesis 2d. Phytoplankton diversity is controlled by a combination of "old" nutrient input via diffusion from below the chemocline and "new" nutrient input from streams.

3. Ecosystem Structure and Function

Our ecosystem-based approach to the MCM will expand what is known about biodiversity in an extremely cold and arid environment, and will allow us to define functional relationships that allow life to survive in extreme environments.

Hypothesis 3a. Biodiversity in all landscape units is linked to ecosystem functioning (diversity begets function). Higher diversity is associated with greater C, N and P cycling.
Hypothesis 3b. Biogeochemical activity, nutrient deficiency, and biodiversity relationships within the MCM landscape are reflected in the elemental stoichiometry.
Hypothesis 3c. The low biodiversity and slow growth rates of its organisms make MCM ecosystems highly susceptible to local and global human disturbance.
Hypothesis 3d. Geomorphological features interacting with resource legacies of landscape units dictate the distribution of species and ecosystem function.

Research Plan Overview

Hydrology. Hypothesis 1a addresses the role of climatic variation on the distribution of liquid water and will be addressed by the meteorological, glacial mass balance, lake level and streamflow monitoring networks. Hypothesis 1b, on cryoconites, will be studied using field measurements of cryoconite holes that include the monitoring of their physical dimensions, and biogeochemical characteristics and physical modeling. Hypotheses 1c-d will examine hyporheic zone development and heterogeneity. Hyporheic processes will be examined through meteorological and streamflow monitoring, geochemical measurements (including isotopes) and supporting tracer experiments.

Biodiversity. Hypothesis 2a focuses on the export of algal mat material from the streams and lakes. Their landscape influence will be investigated by comparison of biotic material collected in aeolian traps and cryoconite holes to mat material from the streams and littoral zones of the lakes. Hypothesis 2b is related to the diversity of stream microbial mats and will be addressed through on-going hydrological/ecological measurements and expanded to include morphological and molecular-based characterizations of the mats. Hypothesis 2c concerns soil biodiversity and will be addressed through continued long-term experiments and stable isotopic investigations. Hypothesis 2d focuses on the vertical distribution of phytoplankton and bacterioplankton species and will be addressed through bioassay experiments and stoichiometric measurements in concert with 16S rDNA sequencing.

Ecosystem structure and function. In our conceptual model (Fig. 2.4), biodiversity lies at the base of ecosystem structure and function, which is in-turn related to biogeochemistry, stoichiometry and nutrient availability. Hypotheses 3a-b address directly ecosystem biodiversity and function. Our measurements of C:N:P conducted in the continuing and expanded monitoring programs for the streams, soils and lakes will allow us to examine the diversity and nutrient relationships in the landscape units. Hypothesis 3c addresses the critical issue relating human disturbance to biodiversity, and ecosystem change to human well-being and will be addressed through the monitoring programs. By addressing these issues, we can cast our results within the context of other LTER projects, most of which include human disturbance as a major research theme. Hypothesis 3d will address the role of geomorphic features in determining habitat quality and associated diversity-functional relationships. This hypothesis will be supported through the monitoring programs and the new nutrient enrichment experiments that will be conducted in habitats in all landscape units. All hypotheses in portfolio 3 will increase our understanding of the evolution of this diversity-functional relationship through time (based on the concept of resource legacy) within the MCM.

Limnological Long-Term Monitoring

The lakes lie at the end of the hydrologic continuum and represent a repository of past conditions within the MCM. We will maintain our routine suite of limnological measurements and will add measurements to address the hypotheses of MCM-III. To define the phylogenetic and functional aspects of biodiversity of the lake biota, we will add measurements of molecular diversity, phytoplankton pigment diversity, sestonic stoichiometry (C:N:P) and microbial exoenzyme production. To further resolve the resource legacy in the lake water column, the racemization state (D- or L- isomer) of aspartic acid will be measured to define the biological age of the water (Grutters et al., 2002). Data collection and sample processing will be closely aligned with the recently funded 5-year dry valley lake microbial observatory (J. Priscu, PI; http://mcm-dvlakesmo.montana.edu) that is searching for novel prokaryotes and novel physiologies.

We will continue to work with the NSF to extend our field season into late fall/early winter to examine water column microbial processes during the transition from complete sunlight to total darkness to complement previous results for the polar sunrise (Priscu et al., 1999). In addition, we still desire to field a winter-over team when logistics and resources become available, as documented in our last Five Year Plan submitted to NSF-OPP in May 2002. The site review strongly urged a winter component to our science. We will begin addressing this issue via our proposed in-situ sensor development program.

Expansion of long-term lake monitoring:

Under ice PAR.. Photosynthetically active radiation (PAR) has been shown to be a primary driver for phytoplankton photosynthesis (Neale and Priscu, 1998) and will be used to model depth-integrated primary production within the lakes (Priscu et al., 1999).

Elemental stoichiometry. Particulate carbon (PC), NH4+, NO2-, NO3-, particulate N (PN) and soluble reactive phosphorus (SRP) were measured routinely during MCM-I and MCM-II. We will add dissolved organic nitrogen (DON), particulate P (PP), and dissolved organic P (DOP) to this suite of measurements to allow detailed C:N:P ratios to be determined for both the particulate and dissolved fractions at selected depths within the water column.

Phytoplankton pigment diversity. In addition to our routine microscopic phytoplankton identification and enumeration, which is tedious and subjective, we will begin using a commercially available submersible spectrofluorometer. This instrument will allow us to differentiate, in real time, the diversity of the main photosynthetic groups within the lakes.

Molecular phylogeny. Molecular phylogeny will be used to characterize the microbial diversity annually within each lake at depths representing biogeochemically distinct layers within the water columns. Similarity indices and cluster analysis (CA) will be used to compare the level of similarity of sequence data over the geochemical gradients that exist within each lake and between lakes. These statistical methods will also be applied to the geochemical and exoenzyme parameters we measure and the resulting dendrogram(s) will be compared with the phylogram derived from the DNA data. These statistical methods will allow us to assess the features within each lake (and between lakes) most closely related to the phylogenetic data.

Bacterial biomass and productivity. Our past work has focused on measurements of phytoplankton biomass and productivity within the lakes. It recently became clear that the bacterioplankton are a major component of lake biomass and their phylogenetic and functional diversity may provide a link between past and present biogeochemical function and structure (e.g., Lisle and Priscu, in press; Lee et al., 2004). Consequently, bacterial biomass and bacterial productivity will be determined throughout the water column.

Exoenzyme-based in situ analyses. Most heterotrophic bacteria readily assimilate simple organic monomers, such as monosaccharides and amino acids. However, such compounds usually account for only a small fraction of the available DOC pool; macromolecules compose the bulk of the DOC. Hence, the rate-limiting step in DOC utilization is not monomer uptake, but monomer generation from macromolecules, a process mediated by extracellular enzymes (exoenzymes). Arrieta and Herndl (2002) have recently suggested that an increase in substrate availability induces changes in aquatic bacterial species composition, favoring the growth of beta-glucosidase producers. This, in turn, acts as a feedback loop, increasing the hydrolysis rate at the community level. We propose a series of exoenzyme studies which will provide us with two important clues to the microbial ecology: (1) a snapshot of heterotrophic processes in situ, and (2) the data necessary to determine if the response is the result of regulation of enzyme expression or is due to temporal trends in species diversity.

Mass loss to the sediments. The lakes represent the "sink" in the long cascade of material flow in the MCM. Hence, the determination of particulate fluxes to the lake sediments in concert with their stoichiometry will provide important information on biogeochemical processes within the ecosystem. We will deploy sediment traps just off the bottom of Lake Bonney to generate a time series of particulate C, N and P flux. If this deployment is successful, the traps will be deployed in Lakes Fryxell and Hoare for a similar period.

Amino acid racemization. All amino acids except glycine may occur as either L- or Dstereoisomer. L-amino acids are most abundant in nature because they are constituents of all living material. The formation of D-amino acids is primarily restricted to the aging of amino acids in organic matter (the L-form can racemize to the D-form). Preliminary data (Fig. 2.16) from Lake Bonney indicate that the deep saline waters are biologically old, supporting results from isotope dating (Lyons et al., 1998b). Detailed profiles will be made within each lake to determine the biological age of the water column and relate it to contemporary biogeochemicalprocesses.

New experiments integrating stoichiometry and biodiversity across the MCM landscape

Nutrient bioassays. A series of nutrient bioassay experiments will be conducted from selected depths within each lake, within cryoconite holes from glaciers in both the Fryxell and Bonney basins, and within a selected stream from the Fryxell and Bonney basin 3 times during MCM-III. The lake water and cryoconite experiments will be based on time-course incorporation of 14CO2 at saturating irradiances under a variety of nutrient amendments (see Priscu, 1995). Stream experiments will be done in conjunction with stream-scale nutrient tracer experiments in streams with both abundant and sparse algal mats, carried out for 4-6 days, which represents the relevant time-scale for response given the dynamic patterns of streamflow. Because growth of algal mats on various artificial substrates has been unsuccessful (Chatfield et al., in review; Vincent and Howard-Williams, 1986) and unrepresentative of the regrowth of mats on natural substrates, we will use existing mats in our nutrient bioassay for the streams. We will identify mats in comparable subhabitats above and below the site of nutrient injection and collect 4-5 small cores of mat on a daily basis for chlorophyll-a, ash-free dry mass, C, N and P content, and bacterial cell counts (see Alger et al., 1997 and Table 2.4 for methodology) to monitor the microbial response to nutrient enrichment. Enrichments will be the same as those used in the lake and cryoconite hole bioassays. Dissolved nutrients will be monitored during the tracer experiment at a site near the subhabitat during the experiment to determine relationships between nutrient amendment and stoichiometry.

New measurements using in situ sensors

We currently maintain a network of sensors that measure physical parameters (e.g. meteorology, underwater PAR, stream flow, lake level, etc.) at a high frequency (several times an hour). Yet, as the site review pointed out, there is a logistical delay in retrieving portions of the previous season's data until the following year. In addition, our basic biogeochemical analysis is restricted to, on average, about 3 to 5 samples per year, which are only taken during summer months due to logistics. This low sampling frequency greatly limits our understanding of the coupling of hydrological (measured several times an hour) and ecosystem processes in the system. Furthermore, winter-over lake data have not been collected due to the difficulty in maintaining a human presence year-round. To overcome these limitations we will augment our existing physical network and develop a biogeochemical sensor network in the streams and lakes that will provide, at a minimum, high resolution measurements of temperature, nitrate, turbidity, chlorophyll-a and dissolved oxygen. Other sensors will be added as the network develops. Osmotic pumps can provide continuous samples for later analysis of other parameters in the lab. Ballast-controlled profilers could provide regular (e.g. weekly) top to bottom lake measurements. Funds are insufficient in the current budget to support a full network, but initial instrumentation could be deployed through regular equipment procurement in Antarctica, with supplemental funds being supplied through outside sources (e.g. the NSF Sensors and Sensor Network Program). As the site review pointed out, data telemetry is also considered important in reducing the need for site visitation and speeding up data delivery time. Telemetry of the new and existing sensor data will be sought in collaboration with NSF's logistics providers and was put forward as a MCM priority in our last Five Year Plant to NSF-OPP. Clearly this telemetry effort will also provide interesting outreach opportunities.