Neural-network retrieval of integrated precipitable water vapor over land from satellite microwave radiometer

A method based on neural networks is proposed to retrieve precipitable water vapor (IPWV) over land from brightness temperatures measured by the Advanced Microwave Scanning Radiometer - Earth Observing System (AMSR-E). Water vapor values provided by European Centre for Medium-Range Weather Forecasts (ECMWF) were used to train the network. The performance of the network was demonstrated by using an independent dataset of AMSR-E observations and the corresponding IPWV values from ECMWF. This work has been developed as a part of the Mitigation of Electromagnetic Transmission errors induced by Atmospheric Water Vapor Effects (METAWAVE) project. Therefore, our study was optimized over two areas, centered on the METAWAVE test sites in Como and Rome, Italy. Results were compared with the IPWV measurements obtained from in situ instruments, a ground-based radiometer and a GPS receiver located in Rome, and a local network of GPS receivers in Como.