Biological implications of empirical models of winter oxygen depletion

We compiled data for 23 North American temperate zone lakes to assess three alternative winter O-2 depletion models for estimating O-2 dynamics from freezing to thawing. Dissolved O-2 concentrations were constant or declined slightly for an average of 40 days after freezing and then declined rapidly. Once O-2 concentrations reached 1-3 mg .L-1, consumption slowed. No model that we fit captured O-2 dynamics shortly after freezing. The best fit was a one-pool exponential decay model after one to four initial data were removed. Photosynthesis and freeze-out estimates suggest that O-2 inputs are more important in shallow than in deep lakes. Oxygen decay rates (k) correlated with morphometry in shallow lakes and chlorophyll a, Secchi depth, and the sediment surface area to volume ratio in deep lakes. We hypothesize that the failure of chlorophyll a to correlate with k in shallow lakes is because macrophytes are the primary source of decaying organic matter but have not been included in assessments of winter O-2 depletion. Thus, some processes in deep lakes cannot simply be scaled to smaller scales in shallow lakes.