Satellite-based soil moisture product performance assessment among the EU Ecoregions

Several remote sensing (RS) microwave-based SM products are available in recent years and offer an extraordinary opportunity to quantify land surface soil moisture (SSM). These products provide soil moisture (SM) estimates with different levels of accuracy which are influenced by  climate, vegetation, and soil features. For this reason, several studies aimed at assessing satellite SM data performance also for comparison with in situ measurements (e.g., the International Soil Moisture Network – ISMN),  have already tried to investigate the relationship among SM and the type of coverage or the climatic conditions.

In any case, no one of these studies (i) have considered together climate, vegetation, and soil features when characterising accuracy of SM derived from RS, or (ii) analysed separately the uncertainty due to interaction with the water-soil cycle variables and the uncertainty due to the wetting condition of the upper layer; indeed, the topsoil wetness variability affects the penetration depth of microwave radiation bringing additional errors when comparing information collected at different depths, from surface to the root zone.

In this context, the present study aims (i) to assess the accuracy of SSM measurement through the implementation of an intercomparison between satellite and the terrestrial International Soil Moisture Network data among the European ecoregions which are considered the largest homogeneous area in terms of climate, vegetation and potentially investigable soil cover. Furthermore, considering that soil characteristics add further uncertainty due to the soil saturation condition when the upper soil layer is excessively dry or excessively wet, the study explores (ii) the local dynamics of soil moisture described by the probability density function of SM. 

Five satellite SM products have been studied, considering those derived from the National Aeronautics and Space Administration (NASA) mission (SMAP), as well as those generated by the European Space Agency (ESA) mission (SMOS, ASCAT, ESA CCI, SENTINEL -1) while the ISMN data were considered as a ground truth.

The results show the best or worst performance of the above cited satellite retrievals in different climate, vegetation, and soil features by looking at their variability at ecoregion scale. Moreover, the approach of multimodality using the ASCAT product, which is provided in % of saturation, validated by the test of the excess mass of Ameijeiras-Alonso, following the removal of phenological seasonality, has proven to be an excellent tool for characterising errors in dry areas, confirming that worse performance occurs in areas with a dry phase observed.