A new method for retrieving daily land surface soil moisture using CYGNSS reflectometry data and coherency ratio

 Soil Moisture (SM) plays a crucial role in the hydrological characteristics of the Earth’s land surface, energy exchange, and climate change. The spaceborne GNSS-R (Global Navigation Satellite System Reflectometry) technique has developed as an effective method to retrieve land surface soil moisture. This study proposes a new semi-empirical method to estimate the land surface soil moisture from the spaceborne GNSS-R data. First, the Fresnel reflectivity is linearly modeled with the GNSS-R-derived reflectivity and coherency ratio and the environment variables (i.e. vegetation water content and surface roughness). Then the Fresnel reflectivity is used to estimate the soil moisture by the dielectric constant model. We apply our method to the Cyclone GNSS (CYGNSS) reflectometry data globally collected from 2021 to 2023. The CYGNSS-derived reflectivity and coherency ratio and the Soil Moisture Active Passive (SMAP) data in 2021 are used to construct the linear model. Then the global daily soil moisture data with a spatial resolution of 36 km from 2022 to 2023 are retrieved with the model. Our soil moisture retrievals and the official CYGNSS soil moisture product (SM_CYGNSS) are compared to the SMAP SM. The results show that our soil moisture retrievals perform well (ubRMSE=0.043 cm³/cm³; R=0.62) and are superior to that of the SM_CYGNSS (ubRMSE=0.059 cm³/cm³; R=0.34), with ubRMSE decreasing by 27.1%. The proposed method improves the soil moisture estimation and will benefit the physical interpretation of hydrological issues with GNSS-R.