Comparing carbon dynamics of pelagic food webs in two Antarctic lakes: A practical application of ecological network analysis to an ecological network

A model was developed to explore patterns of carbon flow through the pelagic food webs of Lake Fryxell, Taylor Valley and Ace Lake, Vestfold Hills, Antarctica. These studies evaluated structure, quantified patterns of carbon flow and tested top-down versus bottom-up controls on these systems. The steady-state model calculated carbon flows between trophic groups needed to maintain steady state biomass, based on estimates of basal metabolism, assimilation and production efficiencies, and predation losses. Independent measures of total community respiration, in Lake Fryxell, or NPP, in Ace Lake, validated model behavior. A measure of trophic efficiency (λ) was defined as the ratio of carbon uptake by the top predator to total carbon uptake by all other trophic groups. Lambda from both lakes was always <5% of total community carbon flow and seldom correlated to other flows of carbon. Within both lakes, major fluctuations in total community biomass and carbon flows followed patterns of seasonal and inter-annual variations of a number of abiotic parameters suggesting bottom-up control on these systems.;The simple pelagic food webs of these lakes provided a common set of data that was used to evaluate within- and between system functional characteristics with Ecological Network Analysis (ENA). I compared ENA metrics generated by analysis of these two lakes, to more commonly used ecological indices such as production: biomass (P/B), Shannon diversity index (H') and trophic efficiency (λ). Ascendency (A/C) had strong relationships in both Antarctic lakes with the Shannon indices used, including flow diversity (Si). Additionally, differences in Ascendency between years, months, presence/absence of the predator, and lakes was lower during periods of high biomass and productivity. The potential value of ENA to ecosystem analysis was illustrated by the results of PCA where the first three axes explained up to ca. 93% of variability within the scaled metrics, and almost 80% for the smaller subset of predator only data; ENA metrics loaded on all three axes. From these combined analysis using both ENA and the more common metrics, a clearer picture of system function as it relates to structure was developed, suggesting that ENA metrics provided additional unique information about these communities.