Masaomi KIMURA

Our research sub-team is trying to develop innovative machine learning models that can reproduce rainfall-inundation phenomena in low-lying agricultural areas and saltwater intrusion in agricultural canal networks with high accuracy and low computational burden acceptable to practical applications. For the input training datasets, we attempt to use the artificially generated data, which is simulated by using the process-based hydrodynamic models, in addition to the observed data. The developed models will be expected to contribute in supporting the managers of water-utilizing facilities to make operating decisions by providing the short-term forecasted data.