ABSTRACT

L-band passive microwave remote sensing is currently the optimum tool for the large-scale soil moisture mapping. However, due to the coarse spatial resolution of passive microwave remote sensing, the existence and dynamic changes of water bodies in the field of view would affect the performance of soil moisture retrieval. The Qinghai-Tibet Plateau (QTP) is a densely populated area of lakes in China, and the correction of water contamination effect is much more challenging. In this letter, we proposed a simulation-based approach which utilizes a radiative transfer model to simulate the brightness temperature (TB) of water bodies to correct the Soil Moisture Active Passive (SMAP) data over the QTP, taking into account of wind speed, surface temperature and salinity of lakes. The simulated data and water area fraction obtained from a dynamic map of water bodies were applied to remove the effect of water contamination. Results showed that the relatively low SMAP TBs surrounding both small and large lakes could be eliminated and do not have correlation with the water area fraction after using the proposed approach. Validation results in the Naqu network demonstrated that the simulation-based approach can effectively remove the water contamination from SMAP data with improved soil moisture retrievals.