Active-passive microwave remote sensing data combination for retrieval of soil moisture

Abstract

Microwave remote sensing has become an important way of soil moisture retrieval because of its better penetration into vegetation and soil layers, and its higher sensitivity to the content of soil moisture. Retrieval by active or passive method has a series of mature algorithms, and the combined active-passive methods are the recent research highlights. But the existing combined algorithms are not mature enough, with their weaker applicability and lower accuracy. This study aims to solve the problem of low accuracy, and proposes a new algorithm: based on the imitation of water-cloud model and IEM model, with a BP neural network, obtaining its own sensitive surface parameters from active and passive microwave remote sensing data, and developing a new combined active-passive retrieval model for soil moisture. Finally with testing and verifying in SMEX02 experiments, the model presented the results of higher accuracy.