Bacterial dissolved organic carbon demand in McMurdo Dry Valley lakes, Antarctica

The proximal substrate source of planktonic bacteria is dissolved organic carbon (DOC), and the combined sources of DOC (bulk, phytoplankton production, and advected) set an upper limit on how much C is available for bacterial respiration (BR). We compared measurements of bacterial production (BP) and estimates of BR to measurements of what we assumed to be the major DOC inputs for three permanently ice-covered lakes in Antarctica: Lakes Fryxell, Hoare, and Bonney. These measured inputs, which included phytoplankton extracellular release (ECR), stream input, and upward diffusion of DOC across the chemocline, sediments, and benthic microbial mats, were three to eight times smaller than planktonic BR, suggesting that a major source of bacterial C was unaccounted for. Despite overestimating DOC and doubling bacterial growth efficiency (BGE), BR in the lakes was 1.25– 2 times greater than our estimate of DOC supply. We hypothesize that a major source of organic C for planktonic bacteria in these lakes comes from drawdown of bulk DOC and/or decomposition of particulate material. Dissolved organic carbon (DOC) represents the proximal substrate of heterotrophic planktonic bacteria production, and the sum of DOC inputs (bulk, phytoplankton production, and advected) restricts how much C is available for bacterial respiration (BR). The classical view holds that phytoplankton production sustains planktonic communities; however, depending on the temporal and spatial scales used to analyze the ecosystem, respiration often exceeds production in many oligotrophic waters (Coveney and Wetzel 1995; del Giorgio et al. 1997; Carignan et al. 2000). Determining the balance of production and respiration is fundamental to understanding C flow in aquatic ecosystems. The Taylor Valley Lakes (Lakes Fryxell, Hoare, and Bonney) of the McMurdo Dry Valleys, Antarctica, provide a unique system in which to study bacterioplankton DOC demand because phytoplankton production is limited to approximately 5 months each year, whereas bacterial activity occurs throughout the year (Takacs and Priscu 1998; Priscu et al. 1999). Additionally, the lakes are virtually closed systems owing to the permanent 4to 5-m ice cover, lack of outflow, and relatively low inflow. Although 20% of the DOC pool in Lakes Fryxell and Hoare is comprised of lowmolecular-weight fulvic acid that is microbial in origin (McKnight et al. 1991, 1993), the remainder of the bulk DOC pool of these lakes is relatively high-molecular-weight DOC that is presumably recalcitrant. Phytoplankton extracellular release (ECR) and upward diffusion from the nutrient-rich hypolimnia (Howes et al. 1992; Priscu 1995) are believed to be the two dominant mechanisms that supply new DOC to the trophogenic zone of the lakes. To test this assumption, we constructed partial C budgets for Lakes Fryxell, Hoare, and Bonney to compare measurements of DOC supplied from primary production, stream input, and diffusion to bacterial production (BP). DOC demand and supply relationships were further explored by including estimates of BR to determine the role of unaccounted-for sources of DOC in the regulation of bacterial productivity during the Austral summer. The results of our budgets are discussed with respect to the annual light : dark cycle in Taylor Valley Lakes. Site description—Our study concentrated on Lakes Fryxell, Hoare, and the east and west lobes of Lake Bonney, which lie in the Taylor Valley ( 77 37 S, 163 00 E). While these lakes have varying degrees of chemical stratification, generally all contain nutrient-rich deep water covered by a relatively nutrient-poor trophogenic zone (Priscu 1995; Spigel and Priscu 1998). The trophogenic zone is defined as the layer where oxygenic photosynthesis is measurable: to 18 m in Lake Hoare and the east lobe of Lake Bonney, to 17 m in the west lobe of Lake Bonney, and to 9 m in Lake Fryxell. The permanent ice cover of these lakes prevents wind-driven mixing, which, coupled with low advective stream input, allows vertical chemical and biological gradients to develop and persist (vertical mixing is at the molecular level throughout the water column; Spigel and Priscu 1998). DOC in these lakes increases with depth, and in the highly stratified Lakes Fryxell and Bonney, the gradients are extreme (Fig. 1). DOC concentrations range from 0.16 mg L 1 in the surface waters of the lakes to 34 mg L 1 in the bottom waters. Temporal changes in trophogenic zone and water-column DOC concentrations (volume weighted and depth integrated) were most variable in Lake Bonney’s east lobe and Lake Fryxell during the Austral summers of 1993–1997, whereas DOC concentrations in the west lobe of Lake Bonney and Lake Hoare were less variable (Fig. 2). Although DOC varied over the season, a clear temporal trend was not apparent. Phytoplankton and bacterioplankton production (Fig. 3) are greatest just below the ice cover (5 m) in the spring, but production peaks at the chemocline become more pronounced than the 5-m peaks as the summer progresses (Priscu 1995; Lizotte et al. 1996; Takacs and Priscu 1998). Detailed descriptions of the lakes and streams may be found in Green and Friedmann (1993) and Priscu (1998). Partial DOC budgets—Budgets based on the presumed major DOC inputs (phytoplankton ECR, stream input, and diffusion) were calculated for each sampling season (October–January, except for the 1995–1996 season, which was September–January) in the trophogenic zone and water column of each of the lakes. Lake Bonney has two basins (east and west lobes), connected by a narrow ( 20 m wide), shallow (12 m) sill (Spigel and Priscu 1998) that prevents mixing of waters between the two lobes below 12 m. Therefore, separate budgets were constructed for the east and west lobes of Lake Bonney because of the lack of deep-water interchange between the two lobes. Samples were collected at least once a month during the Austral summer over the deepest portion of the lakes at 2–3-m intervals throughout the