Precipitation Nowcasting from Geostationary Satellite: Neural Approaches Trained By Polar Orbiting and Ground-Based Data

This work explores possible improvements of the Neural Combined Algorithm for Storm Tracking (NeuCAST) proposed in Marzano et al. [2007]. In its single channel version, developed for the Visible-Infrared Imager (VIRI) onboard Meteosat-7, this technique has been successfully applied to the rainfall field nowcast from thermal infrared (TIR) and microwave (MW) passive-sensor imagery aboard, respectively, Geostationary-Earth-Orbit and Low-Earth-Orbit satellites. The multi-channel NeuCAST methodology is here introduced. It extends the single-channel NeuCAST technique to infrared (IR) multi-channel data available from Meteosat Second Generation (MSG) and MW data from ground based meteorological Radar.