On the Disaggregation of Passive Microwave Soil Moisture Data Using A Priori Knowledge of Temporally Persistent Soil Moisture Fields

Water and energy fluxes at the interface between the land surface and atmosphere are affected by the surface water content of the soil, which is highly variable in space and time. The sensitivity of active and passive microwave remote sensing data to surface soil moisture content has been investigated in numerous studies. Recent satellite mission concepts as, for example, the soil moisture and ocean salinity (SMOS) mission, are dedicated to provide global soil moisture information with a temporal frequency of a few days to capture the high temporal dynamics of surface soil moisture. SMOS soil moisture products are expected to have geometric resolutions on the order of 40 km. Mesoscale flood forecasting or water balance models typically operate at much higher spatial resolutions on the order of 1 km. It seems therefore essential to develop appropriate disaggregation schemes to benefit from the high temporal frequency of the SMOS data for hydrological applications as well as, for example, local numerical weather prediction models that are operated at a resolution of a few kilometers. This paper investigates the potential of using prior information on spatially persistent soil moisture fields to disaggregate SMOS scale soil moisture products. The approach is based on a ten-year soil moisture climatology for a mesoscale hydrological catchment, situated in southern Germany, which was generated using a state-of-the-art land-surface process model. The performance of the disaggregation algorithm is verified by comparison of disaggregated soil moisture fields with another ten-year period. To investigate the potential of the suggested disaggregation method for SMOS soil moisture products, a ten-year synthetic brightness temperature data set is generated at the 1-km scale. soil moisture is then retrieved from the aggregated brightness temperature data at the SMOS type scale of 40 km and then disaggregated using the suggested approach. The results are compared against reference soil moisture at the 1-km scale. Uncertainties in the retrieval of the SMOS soil moisture products are explicitly considered, and the uncertainties of the disaggregated fields are quantified. The developed method shows a generally good performance for large parts of the test site, where soil moisture can be disaggregated with an accuracy that is better than the 4 vol.% benchmark of the SMOS mission. As the suggested method shows high sensitivity to biased soil moisture retrievals, uncertainties of the SMOS soil moisture products will directly reflect on the absolute accuracy of the disaggregated soil moisture fields, resulting in a much worse performance under noisy conditions. Nevertheless, the resulting soil moisture distributions show that it is feasible to derive relative soil moisture distributions in these cases.