Soil Moisture Retrieval with High Spatial-temporal Resolution by Fusion of CYGNSS and SAR Data

As the probability and intensity of global drought events continue to increase, soil moisture is an important basis for drought monitoring and drought assessment, and the need to accurately obtain soil moisture distribution information with high spatial-temporal resolution is becoming extremely important. CYGNSS data based on spaceborne GNSS-R has the advantage of high temporal resolution, while SAR data can provide information on surface features with high spatial resolution, and the combination of the two provides favourable conditions for obtaining soil moisture with high spatial-temporal resolution. This paper proposes a soil moisture retrieval method with high spatial-temporal resolution by the fusion of spaceborne GNSS-R (CYGNSS) and SAR (Sentinel-1) data. This method constructs a function relationship between surface reflectivity of spaceborne GNSS-R and backscattering coefficient of SAR, with the aim of preparing for fusion of CYGNSS and Sentinel-1. By fusing sentinel-1 data, a two-layer machine learning framework based on CYGNSS data is constructed to retrieve the soil moisture with high spatial-temporal resolution, and the retrieval results are compared with the measured data and soil moisture products of SMAP. The results indicate that, the surface reflectivity of spaceborne GNSS-R shows an approximate linear relationship with the backscattering coefficient of SAR. The constructed first-layer framework is able to supplement CYGNSS surface reflectivity data, and verifies the feasibility of converting backscattering coefficients of SAR to the CYGNSS surface reflectivity. The soil moisture retrieval by the two-layer framework method in this paper is comparable to the soil moisture product of SMAP in terms of retrieval accuracy (average ubRMSE = 0.070cm ³ /cm ³ , average R = 0.65) at the same spatial resolution (3 km), and the temporal resolution is improved by 3.9 times on average, which confirms the feasibility of soil moisture retrieval by CYGNSS at 3 km spatial resolution.