Patterns and mechanisms of year-to-year variability in winter oxygen depletion rates in ice-covered lakes
Ten years of winter oxygen data were aoalyzed from the North Temperate Lakes, Long Term Ecological Research site io Vilas County, Wisconsin, to characterize the year-to-year variability of oxygen depletion rates aod to determine the factors which iofluence this variability. Meao oxygen depletion rates among the lakes raoged from 0.023 to 0.084 g ·m-3 ·d-1 Depletion rates varied by as muchas 50\% from year to year. Lakes with similar morphometry exhibited temporal coherence in depletion rates over the ten-year period, suggesting that different mechaoisms were influencing the variability in deep aod shallow lakes. The sediment area to lake volume ratio best explained the variability in meao depletion rates among lakes, although summer chl a was also a factor. While differences in morphometry aod productivity explained the variability in oxygen depletion rates among lakes, these variables did not account for the variability observed among years. Cover variables, such as snow depth aod ice thickness, were correlated to yearly depletion rates for several lakes. There was strong relationship between light extinction coefficients of the ice cover aod depletion rates for several lakes indicating that oxygen production was causing the year-to-year variability in depletion rates. The importance of oxygen production suggests that year-to-year variability in depletion rates is caused by variability in net water column oxygen consumption. Lake-to-lake variability, then, is caused by differences in sediment oxygen consumption related to the morphoQ.letric and average productivity characteristics of the lakes. These observations have implications for assessing the impact of climate change, as changes in lake productivity, morphometry, and winter cover characteristics could affect depletion rates. Variability from year to year is also important for predicting the risk of winterkill over a period of years. Several multiple regression models are presented which combine lake-specific and year-specific variables to make predictions about winter oxygen depletion rates.