Validation of Satellite-Derived Soil Moisture Products Using Ground Observations in Southern Europe

Soil Moisture (SM), acknowledged by the Global Climate Observing System (GCOS) and the European Space Agency’s Climate Change Initiative (ESA CCI) as an Essential Climate Variable (ECV), is a fundamental hydrological parameter that plays a pivotal role in bridging Earth's surface and atmospheric interactions. Understanding SM status and dynamics is critical for various meteorological, hydrological, and climatological applications. Furthermore, it provides insights into the water, energy, and carbon cycles while aiding in the forecasting of extreme climatic events, such as droughts and floods. In consequence, accurate global monitoring of SM with suitable temporal and spatial resolutions is imperative.

This study focuses on the validation of multiple satellite-derived near-surface SM products against field measurements to evaluate their accuracy and reliability. The research was conducted over the northeastern Spain and southern France, covering a 7-year span from January 2015 to December 2021. Ground truth data were obtained from the International Soil Moisture Network (ISMN) database, which included observations from 30 stations across four networks (COSMOS, FR-Aqui, IPE, and SMOSMANIA). The analysis assessed four microwave-based sensors, encompassing both active and passive systems: ASCAT (Advanced Scatterometer), SMOS (Soil Moisture and Ocean Salinity), SMAP (Soil Moisture Active Passive), and CCI.

Following data acquisition and processing for both satellite images and ground observations, a comprehensive validation was performed using statistical metrics, scatter plots, and linear regression analysis of the respective time series. Results highlighted that the SMAP mission delivered the most reliable outcomes, achieving a near-unity slope, an intercept close to zero, a correlation coefficient of R = 0.72, and a Root Mean Square Error of RMSE = 0.07 m³/m³. The CCI product followed, while ASCAT and SMOS showed larger uncertainties and weaker correlations, respectively. In addition, an analysis of the in situ depth effect using SMAP indicated that measurements at 0–6 cm (integrated) and 5 cm (point-specific) depths yielded optimal results. Nevertheless, despite remarkable advances in SM monitoring, this work underscores the need for further research to align satellite-derived data more closely with field-level precision.

Acknowledgements: This study was carried out in the framework of the PID2020-118797RBI00 (Tool4Extreme) project, funded by MCIN/AEI/10.13039/501100011033, and also the PROMETEO/2021/016 project, funded by Conselleria d’Educació, Universitats i Ocupació de la Generalitat Valenciana.