Factors explaining the yearly changes in minimum bottom dissolved oxygen concentrations in Lake Biwa, a warm monomictic lake

Abstract

Vertical profiles of dissolved oxygen (DO) and water temperature (WT) measured bi-monthly for 36 years (1980–2015) near the deepest part of a warm monomictic lake were analyzed with special reference to yearly minimum DO at bottom (DOmin). DOmin changed yearly (3.0\u2009±\u20091.2\u2009mg\u2009l−1) and significant differences in DOmin were not observed between Period I (1980–1993; cooler and worse in water quality) and Period II (1994–2015; warmer and better in water quality). This unclear trend in DOmin was probably due to the offsetting influences between warming induced by global warming and oligotrophication attempted by local governments etc. for the study period. DOmin was positively correlated with disturbance time (timing of last cold water intrusion observed from Mar to Aug), which could be related to the start of DO depletion at bottom. Thus, the linear model using this parameter could predict yearly DOmin fairly well for the entire study period (r2\u2009=\u20090.60). In addition, DOmin and time of disturbance were correlated negatively with water density at bottom in Jan and positively with water density equilibrated to air temperature (AT) in Mar. Higher lake water density after full depth mixing advances the disturbance time. In contrast, lower AT in Mar and/or higher density of influent water after Mar delays the time likely due to the larger amount of snowfall in the watershed. Further, DOmin was positively correlated with maximum wind velocity in Sep which probably induced the recovery of DO. Multiple-regression models to predict DOmin using these meteorological and water quality parameters were developed (r2\u2009≥\u20090.38, worse performances than the model using disturbance time) to forecast future trends of DOmin through global warming and/or climate change. Significant influences of water or sediment oxygen demands on DOmin were not detected. We also discuss the applicability of the proposed models.