レーダ雨量を考慮したニューラルネットワークによる流出計算

Abstract

Neural networks, are used to calculate runoff from weather radar data and ground rain gauge data. Compared to usual runoff models, it is easier to use radar data in neural network runoff calculation. Basically you can use the radar data directly, or without transforming them into rainfall, as the input of the neural network. A situation with the difficulty of ground measurement is supposed. To cover the area lacking ground rain gauge, radar data are used. In case that the distribution of ground rain gauges is non-uniform, filling up the vacancy with radar data is very efficient to get good runoff estimation.