Abstract
In this research, we evaluated errors of meteorological data generated by a dynamical downscaling using a regional weather forecast model of WRF from JRA25 data along with several bias-correction technique applications. Then we investigated uncertainties of the water quality prediction in a reservoir due to those errors using a 3-D hydrodynamic and ecological model of ELCOM-CAEDYM. Comparisons of several bias-correction methods revealed that the quantile mapping method is most suitable. But a statistical downscaling was better for wind direction and wind speed correction might need very fine grid scale simulation to express wind field in a reservoir. In-reservoir water quality simulations revealed that errors of air temperature and solar radiation, which can be bias-corrected, were two major factors that affect the seasonal change of water temperature profile and the phytoplankton biomass in a reservoir.
