Convective mixing and high littoral production established systematic errors in the diel oxygen curves of a shallow, eutrophic lake

The diel (24-h) oxygen (O-2) curves approach has become a popular method for analyzing gross primary production (GPP) and ecosystem respiration (ER) rates in aquatic systems. Despite the simplicity of this approach, there remain aspects of the calculation and interpretation of diel O-2 curves which may skew results, with potentially large implications for estimates of metabolic rates. One common problem in lakes is the occurrence of unexpected changes in O-2 concentrations (for instance, increasing overnight O-2 concentrations). Such changes have typically been ascribed to the random mixing of pockets of O-2. It has thus been suggested that negative GPP or positive ER values should be included in calculations, on the assumption that under-and overestimates should occur with equal frequency, and thus cancel each other out. Our data from a shallow, eutrophic lake provided a high share of negative GPP values. We argue that these may have been the result of elevated littoral productivity coupled with convective currents produced by consistent differences in the heating or cooling of littoral and offshore waters. Such phenomena might be common in small, sheltered lakes where the role of mixing by wind is diminished. We conclude that a failure to account for consistent metabolic gradients and periodic convective mixing may lead to a chronic underestimation of metabolic rates in lakes when using the diel O-2 curves method.