Soil Moisture Retrieval from Soil Moisture Active Passive (SMAP) Satellite: A Validation and Error Analysis Using Ground-Based Observation Over Bangladesh

Abstract

The Soil Moisture Active Passive (SMAP) mission, the latest satellite dedicated to soil moisture monitoring, provides global soil moisture maps through simultaneous L-band radar and radiometer measurements. This study aimed to support the ongoing authentication of SMAP soil humidity products by inter-comparing five products: L4_SM, L3_SM_P_E, L3_SM_P, L2_SM_P_E, and L2_SM_P. Ground-based soil moisture data from nine locations in Bangladesh—Cumilla, Ishwardi, Mymensingh, Dinajpur, Rajshahi, Rangpur, Joydebpur, Khulna, and Sreemongal weather stations were used for validation. The analysis also investigated potential sources of error in SMAP products. Results revealed that L4_SM had the lowest error values among the five products, with ubRMSE (5.3001 m³/m³), RMSE (5.3077 m³/m³), and Bias (-0.2841 m³/m³). However, these values fell short of the SMAP mission's accuracy goal of 0.04 m³/m³, primarily due to the use of point location data for validation. The correlation factor between in-situ observations and L4_SM soil moisture was probabilistically significant (r = 0.63, p < 0.01). Among the evaluated products, L4_SM demonstrated a relatively better agreement with ground observations and could contribute to enhancing the SMAP algorithm's accuracy. Further research, particularly involving core validation sites, is recommended to address the identified error sources and improve the precision of SMAP soil moisture retrievals.