GloWS-Net: A Deep Learning Framework for Retrieving Global Sea Surface Wind Speed Using Spaceborne GNSS-R Data

Spaceborne Global Navigation Satellite System Reflectometry (GNSS-R) is a new remote sensing technology that uses GNSS signals reflected from the Earth’s surface to estimate geophysical parameters. Because of its unique advantages such as high temporal and spatial resolutions, low observation cost, wide coverage all-weather operation, it has been widely used in land ocean fields. Ocean wind monitoring main objective recently launched Cyclone (CYGNSS). In previous studies, speed was usually retrieved using features extracted delay-Doppler maps (DDMs) empirical model functions (GMFs). However, challenge employ GMF method if multiple sea state parameters input. Therefore, this article, we propose an improved deep learning network framework retrieve global spaceborne GNSS-R data, named GloWS-Net. GloWS-Net considers fusion auxiliary information including swell significant wave height (SWH), rainfall direction build end-to-end retrieval model. order verify improvement proposed model, ERA5 Cross-Calibrated Multi-Platform (CCMP) data were reference for extensive testing evaluate performance models (i.e., GMF, fully connected (FCN) convolutional neural (CNN)). The results show that, when winds ground truth, root mean square error (RMSE) 23.98% better than MVE method. Although FCN have similar RMSE (1.92 m/s), absolute percentage (MAPE) former by 16.56%; CCMP 2.16 m/s, which 20.27% Compared with MAPE 17.75%. Meanwhile, outperforms FCN, traditional CNN, modified CNN (MCNN) CyGNSSnet especially at speeds.