A Novel Sliding Window Algorithm-Based Approach for Global Bias Correction in CYGNSS Soil Moisture Retrievals

The spaceborne Global Navigation Satellite System Reflectometry (GNSS-R) has been shown its capacity for global soil moisture (SM) retrieval. However, the traditional SM retrieval method with introduction of redundant auxiliary parameters may lead to increased dependence on auxiliary data sources. Conversely, relying solely on spaceborne GNSS-R reflectivity for global SM retrieval limits accuracy and overlooks the influence of complex surface conditions. To address these issues, a bias correction model for SM product of Cyclone GNSS (CYGNSS), a spaceborne GNSS-R mission, is proposed. In this model, the bias of CYGNSS SM product is refer to soil moisture active passive (SMAP) SM. To capture a small range of the bias changes, a sliding window algorithm is introduced to divide the global region into regular windows of equally sized. The bias corrected results are verified with SM derived from SMAP and the International SM Network (ISMN). The results show that, after bias correction, the root mean square error (RMSE) was 0.059 cm3/cm3, representing a 21.33% improvement, and the correlation coefficient R increased from 0.801 to 0.876, improving 8.56%. The validation with ISMN stations showed that the average RMSE decreased from 0.088 to 0.070 cm3/cm3 with a reduction of 20.45%, while the correlation coefficient R improved by 29.37% as well as the unbiased RMSE (ubRMSE) decreased by 11.76%. The result demonstrates that the proposed model further strengthens the capability of spaceborne GNSS-R in SM monitoring.