Spaceborne GNSS-R Sea Surface Rainfall Intensity Retrieval Considering the Effects of Wind and Wave

Sea surface rainfall intensity is an important sea state parameter and has an important impact on marine navigation safety and global climate. At present, the methods for measuring sea surface rainfall intensity include rain gauge, weather radar and remote sensing. However, increasing the spatial and temporal resolution and decreasing the cost are still a challenge. GNSS reflectometry (GNSS-R) is an emerging remote sensing technology, which may effectively handle the resolution and cost issues. In this paper, by making use of spaceborne GNSS-R data, three models based on random forest are developed for retrieving sea surface rainfall intensity. The normalized bistatic radar cross section (NBRCS), the leading edge slope (LES), and the signal to noise ratio (SNR) from the CYGNSS are used as the key variables for the rainfall intensity retrieval. The results show that the model only considering wave effects has the best accuracy, with the coefficient of determination (fi² ) of 0.79 and the root mean square error (RMSE) of 1.05 mm/hr. Compared with the model without considering wind and wave effects, the fi² improves by 12.9% and the RMSE improves by 16%. The results also show that the effect of wind on rainfall intensity retrieval should not be considered, if the effect of wave is already considered.