Downscaling of SMAP Soil Moisture Using Land Surface Temperature and Vegetation Data

Abstract

The reference component of the algorithm was developed at the NLDAS grid size (12.5 km) to downscale the SMAP Level 3 radiometer‐based 9‐km soil moisture to 1 km. The downscaled results were validated using data acquired in Soil Moisture Active Passive Validation Experiment 2015 (SMAPVEX15) that included in situ soil moisture and Passive Active L‐band System (PALS) airborne instrument observations. The resulting downscaled SMAP estimates better characterize soil moisture spatial and temporal variability and have better overall validation metrics than the original SMAP soil moisture estimates. Additionally, the overall accuracy of the downscaled SMAP soil moisture is comparable to the PALS high spatial resolution soil moisture retrievals. The method demonstrated in this study downscales satellite soil moisture to produce a 1‐km product that is not site specific and could be applied to other regions of the world using the publicly available NLDAS/Global Land Data Assimilation System data.