A synergetic use of observations from modis, SEVIRI MSG, ASAR and AMSR-E to infer a daily soil moisture index

The objective of this study is to infer a soil moisture index from an approach mainly based on the concept of apparent thermal inertia (ATI). To reduce the effect of spurious variability and cloud presence, soil moisture temporal trend derived from passive microwave based product, namely the NASA AMSR-E-soil moisture product, are used as a tool to filter the data. The AMSR-E data due to their coarse resolution can be considered as natural "low pass filter" thus reducing the effect of noise. Furthermore, the approach considers the soil moisture estimates derived from SAR sensors and use them to spatially calibrate the information coming from the optical data. The algorithm has been validated over two different test areas in Italy and France where ground truth measurements were available. Four main clusters of ATI have been identified and classified into 4 different levels of wetness. In densely vegetated areas, only three classes of soil moisture were distinguishable. The comparison with ground measurements indicates an accuracy of around 88% on the Italian test sites and of 73% on the French test sites, the last mainly characterized by densely vegetated fields.