Simultaneous estimation of both soil moisture and model parameters using particle filtering method through the assimilation of microwave signal

Soil moisture is a very important variable in land surface processes. Both field moisture measurements and estimates from modeling have their limitations when being used to estimate soil moisture on a large spatial scale. Remote sensing is becoming a practical method to estimate soil moisture globally; however, the quality of current soil surface moisture products needs to be improved in order to meet practical requirements. Data assimilation (DA) is a promising approach to merge model dynamics and remote sensing observations, thus having the potential to estimate soil moisture more accurately. In this study, a data assimilation algorithm, which couples the particle filter and the kernel smoothing technique, is presented to estimate soil moisture and soil parameters from microwave signals. A simple hydrological model with a daily time step is utilized to reduce the computational burden in the process of data assimilation. An observation operator based on the ratio of two microwave brightness temperatures at different frequencies is designed to link surface soil moisture with remote sensing measurements, and a sensitivity analysis of this operator is also conducted. Additionally, a variant of particle filtering method is developed for the joint estimation of soil moisture and soil parameters such as texture and porosity. This assimilation scheme is validated against field moisture measurements at the CEOP/Mongolia experiment site and is found to estimate near‐surface soil moisture very well. The retrieved soil texture still contains large uncertainties as the retrieved values cannot converge to fixed points or narrow ranges when using different initial soil texture values, but the retrieved soil porosity has relatively small uncertainties.