Evaluation of SMAP-Enhanced Products Using Upscaled Soil Moisture Data Based on Random Forest Regression: A Case Study of the Qinghai–Tibet Plateau, China

The evaluation of satellite soil moisture is a big challenge owing to the large spatial mismatch between pixel-based products and point-based in situ measurements. Upscaling measurements obtain “true value” content at grid/footprint scale can make up for difference improve validation. Many existing upscaling methods have strict requirements regarding distribution quantity sensors. However, reality, soil-moisture-monitoring networks are commonly sparse with low sensor density, which increases difficulty obtaining accurate upscaled data limits validation products. For this reason, paper proposes scheme upscale using five machine learning along Landsat 8 datasets DEM validate accuracy SMAP-enhanced passive product network on Qinghai–Tibet Plateau. proposed realizes pixel (30 m × 30 m) then coarse grid (9 km 9 km) by multi-source remote sensing as bridge conversion. long-time SMAP SM since April 2015 Plateau were validated based data. results show that (1) random forest regression performs best, reflect region-average conditions be used evaluating data; (2) meets its scientific measurement requirements; (3) generally underestimates study area.