No abstract

Light environment variability was investigated in the two Japanese lakes Biwa and Kasumigaura, which offer a broad range of optical conditions in the water bodies due to their diverse morphometries and limnological characteristics. To elucidate their light environments, Secchi depths (SDs) were related to long-term monitored datasets of concentrations of optically active substances (OASs) using two approaches based on statistical and mechanistic models. A good estimate for the nonphytoplanktonic suspended solids (NPSS) concentration gained using a monthly factor d (which represents the phytoplanktonic portion in total suspended solids) from a longterm analysis was utilized to develop robust models. Using the mechanistic model, the contribution of each OAS to the SD can be understood and investigated in more detail than possible with a statistical approach, but the statistical model yields better results in terms of SD prediction. Based on the results of an analysis of the contribution of each OAS to the SD, it was clear that NPSS was the component that exerted the most influence on the light environments in the two lakes; in this respect, this study agrees with other studies that show the importance of suspended inorganic particles as the main contributor to the SD in inland waters. Using ANOVA, we analyzed how specific inherent optical properties may have changed spatially and temporally, and the results indicated that the temporal (monthly) effect was primarily responsible for the loss of accuracy in the models. In addition, the ANOVA analysis suggested that grouping the data improved the predictive performances of the statistical models. Finally, we concluded that combining the two models yields the most reliable results in terms of SD prediction and determining the contribution of each OAS to the SD at present.