Modeling, Prediction, and Retrieval of Surface Soil Moisture from InSAR Closure Phase

We present a discretized, multi-layer soil moisture model that links soil moisture variability to single-look SAR measurements. Our model reveals distinct closure phase signatures arising from variation of soil moisture, radar frequencies, and soil textures. Specifically, our model predicts that positive asymmetric soil moisture anomalies produce positive closure phase step-changes, and negative asymmetric anomalies yield negative closure phase step-changes, consistent with observed data. In addition, our analysis shows that low-frequency radar (e.g., L-band) exhibits heightened sensitivity to the vertical distribution of soil moisture. We introduce a scalable algorithm for retrieving the InSAR Soil Moisture Index. We demonstrate the retrieval algorithm in two diverse environments: the Mojave desert and the Central Valley in California. Good agreements between the derived InSAR Soil Moisture Index, in situ soil moisture measurements, and SMAP/Sentinel-1 soil moisture measurements highlight the potential for large-scale soil moisture monitoring using InSAR closure phase.