Inter-comparison of Soil Moisture Satellite products on European Ecoregions

Soil moisture (SM) content is a crucial parameter for an extensive range of fields (e.g., hydrology cycle, smart agriculture, environmental risk management, climate system) as it regulates the water balance, land surface energy, and the carbon cycle. However, the non-homogeneous horizontal and vertical distribution of water content in the soil complicates SM evaluation. The integration of in-situ measurements with those remotely acquired or produced by models may help in overcoming such a problem.

Focusing on satellite data, it is worth noting that the growing availability of sensors (active or passive) working in the microwave spectral region has increased the capability to have SM information on a regional scale with a level of accuracy depending on the selected data, the characteristics of the study area as well as the metric considered for their evaluation.  

This study aims to compare the accuracy of several freely available microwave-based SM satellite products with in-situ measurements distributed, after quality control and harmonization, by the International Soil Moisture Network (ISMN) for several stations located in the European (EU) Ecoregions for rivers and lakes (WFD 2000/60/CE) in the time frame 2015-2020.

The satellite products investigated are based on the acquisition by: i) the National Aeronautics and Space Administration (NASA) Soil Moisture Active Passive (SMAP) mission, ii) the European Space Agency (ESA) Soil Moisture and Ocean Salinity (SMOS) mission, iii) the Advanced Scatterometer (ASCAT) aboard of MetOp satellites and iv) the radar onboard ESA’s SENTINEL-1 platforms. In particular we have used processed the following SM products: the SMAPL4 V5 (3-hourly and 9 km of spatial resolution) is based on the assimilation of SMAP (operated in L-band ) observations into a customized version of the NASA Goddard Earth Observing System Version 5 (GEOS-5) land data assimilation system (LDAS); the SMOS-IC V2.0 is the second version of a physically-based algorithm applied to SMOS retrievals operating in L-Band; the H115 and H116 SM products from the ASCAT backscatter observations provided on a fixed Earth grid (12.5 km sampling) in time series format. Finally, the SSM1km -CGLS V retrieved by Sentinel-1 radar images have been also considered (available only for the European continent every 1.5-4 days at spatial resolution of 1km).

Satellite SM retrievals performances are evaluated against ground-based measurements in terms of Bias, Root Mean Square Error (RMSE), unbiased RMSE, and Pearson correlation (considering both original observations and anomalies). On average, SMAP and SMOS-IC highlight the best performance.

The proposed inter-comparison offers both guidelines for choosing among available satellite products and insights on SM retrieval products and versions. As the EU Ecoregions outline is based on a large scale, they enclose areas affected by several climate change impacts (such as drought, changes in relative sea level, salinity, etc.). Thus, the outcomes can be used to develop novel satellite-based integrated methods for modelling the hydrologic response to climate change.