Abstract:
Soil moisture is a key parameter in hydrological and environmental processes such as global water, carbon, and climate cycling. Measuring the soil moisture content of agricultural soil is paramount to determining vegetation health and potential food availability as well as monitoring climate events such as droughts and flooding. The SMAP satellite (Soil Moisture Active Passive) estimates soil moisture through spaceborne retrieval on a global scale. These retrievals should be validated with in situ ground measurements to ensure adequate performance of the satellite. To investigate and assess SMAP performance for agricultural application, the satellite retrievals were validated with ground-based soil moisture measurements at five agricultural sites in Argentina, Canada, Spain, and the United States (US) for the period of April 1, 2015 to October 31, 2020. The evaluation included the dual channel algorithm (DCA) and the V-polarized single channel algorithm (SCA-V) based soil moisture from the Level 3 (L3) product and the model-assimilation-based Level 4 (L4) soil moisture. The L4 product was assessed at the satellite overpass times (as opposed to the 3-hr modeling interval) to make the comparisons between the products equitable. Various statistical measures such as unbiased root mean square error value (ubRMSE), root mean square error value (RMSE), mean difference (MD), Pearson correlation coefficient (R) and anomaly R were used to determine the accuracy of the retrieved soil moisture products. We found that the SCA-V soil moistures were within mission requirement of 0.04 m^3/m^3 ubRMSE at one site, the DCA soil moistures were within the requirement at three sites, and the L4 soil moistures were within the requirement at two sites. To further investigate the root causes of the challenge in meeting the requirements, we assessed the seasonal performance by separating spring (MAM), summer (JJA), fall (SON), and winter (DJF). The results show that the performance of the data products is indeed seasonally dependent. During the growing season, the increasing vegetation and surface conditions affect the performance metrics as well as the seasonal changes in the soil moisture. We found that the changes in seasonal metrics were different for each site depending on the particular conditions of each site.