Data Assimilation to Extract Soil Moisture Information From SMAP Observations

Abstract: Statistical techniques permit the retrieval of soil moisture estimates in a model climatology while retaining the spatial and temporal signatures of the satellite observations. As a consequence, they can be used to reduce the need for localized bias correction techniques typically implemented in data assimilation (DA) systems that tend to remove some of the independent information provided by satellite observations. Here, we use a statistical neural network (NN) algorithm to retrieve SMAP (So…