A Downscaling–Merging Scheme for Improving Daily Spatial Precipitation Estimates Based on Random Forest and Cokriging

High-spatial-resolution precipitation data are of great significance in many applications, such as ecology, hydrology, and meteorology. Acquiring high-precision high-resolution a large area is still challenge. In this study, downscaling–merging scheme based on random forest cokriging presented to solve problem. First, the enhanced decision tree model, which from machine learning algorithms, used reduce spatial resolution satellite daily 0.01°. The downscaled satellite-based then merged with gauge observations using method. applied downscale Global Precipitation Measurement Mission (GPM) product over upstream part Hanjiang Basin. experimental results indicate that (1) downscaling model can correctly spatially GPM data, retains accuracy original greatly improves their details; (2) be seasonal scale; (3) merging method data. This study provides an efficient for generating high-quality area.