First quasi-global soil moisture retrieval using Fengyun-3 GNSS-R constellation observations

Abstract

Global Navigation Satellite System-Reflectometry (GNSS-R) has considerable potential for large-scale soil moisture (SM) monitoring. With the Fengyun-3 (FY-3) E, F, and G satellites currently in orbit, the FY-3 satellite series has formed the GNSS-R constellation. A comprehensive analysis and validation of the SM retrieval capability of the FY-3 GNSS-R constellation observations are essential. This study is the first to use FY-3 GNSS-R constellation observations to evaluate their performance in quasi-global daily SM retrieval. Specifically, this study proposed an effective SM retrieval method for obtaining an FY-3 GNSS-R SM with minimal ancillary data. Compared with the Soil Moisture Active Passive (SMAP) reference SM, the FY-3 SM exhibited a reasonable global spatial pattern as SMAP, with a root mean square error (RMSE) of 0.039 ${\mathrm{cm}}^3$/${\mathrm{cm}}^3$ in low vegetation areas. Validation results from over 200 independent in situ stations showed that the unbiased RMSE and correlation for FY-3 SM were 0.039 ${\mathrm{cm}}^3$/${\mathrm{cm}}^3$ and 0.60, respectively. Triple collocation (TC) analysis showed that the standard deviation and correlation for the FY-3 SM were 0.017 ${\mathrm{cm}}^3$/${\mathrm{cm}}^3$ and 0.62, respectively. Global and local validations indicate that the SM derived from the FY-3 GNSS-R constellation observations has well-defined accuracy and effectively captures spatiotemporal variations. Compared to the contemporaneous Cyclone GNSS official SM, the accuracy of the FY-3 SM retrieved using the proposed method improved by 17.1 %. Consequently, the SM from the FY-3 GNSS-R constellation observations can be an invaluable complement to the global SM dataset. Furthermore, this method effectively reduced systematic bias and random errors in SM retrievals (unbiased RMSE (ubRMSE) from 0.041to 0.034 ${\mathrm{cm}}^3$/${\mathrm{cm}}^3$and TC standard deviation from 0.034to 0.017 ${\mathrm{cm}}^3$/${\mathrm{cm}}^3$), which may provide a valuable reference for generating SM products from subsequent FY-3 GNSS-R constellations.