Daily precipitation datasets across the Chinese Mainland

Some scholars have developed multi-source data-fusion approaches to estimate regional precipitation. In this study, we addressed the potential ranges of accuracy improvement for satellite and reanalysis rainfall products using various machine learning fusion approaches over the Chinese mainland. Multivariate linear regression (MLR), feedforward neural network, random forest (RF), and long–short–term memory (LSTM) networks were used. The results indicate that all four fusion methods reduce errors in the original precipitation products.