Joint utilization of closure phase and closure amplitude for soil moisture change using interferometric synthetic aperture radar

Abstract

The sensitivity of microwave data in soil moisture is attributed to radar wave penetration depth and signal attenuation. However, current soil moisture models rarely consider the simultaneous effects of amplitude and phase induced by soil moisture. This study proposes an innovative InSAR Bias Soil Moisture Model (IBSMM) that jointly exploits closure phase and closure amplitude. Compared with traditional models, IBSMM considers the dual physical change process of microwave signals in soil moisture change. The IBSMM includes a three-step framework to estimate soil moisture. First, conventional repeat-pass InSAR datasets are generated. Second, the bias in closure characteristics is estimated using Regularized Maximum Likelihood Estimation (RMLE) and a dynamic nested sampling strategy. Third, a forward model for soil moisture change is constructed based on the backscattering field. The simulation results indicate that the dynamic nested sampling strategy has a deviation of only 0.042 from the logarithm evidence value. Moreover, the insensitivity and saturation thresholds in the soil moisture model are quantified. Subsequently, the results of two practical case experiments in different land cover types confirm the effectiveness of IBSMM. In Castrejón de Trabancos, Spain, from January 12, 2020 to February 5, 2020, the model had an overall average correlation coefficient (R value) of 0.57 and a root mean square error (RMSE) of 3.39%. Similarly, in Guyuan County, China, from October 5, 2018 to October 29, the model recorded an average R value of 0.46 and an RMSE of 3.1% in grassland. The proposed IBSMM effectively enhances soil moisture estimation accuracy and explains the physical process of soil moisture change.