Soil Moisture Retrieval Using Single-Frequency Dual-Polarization GNSS-R Data From Airborne GLORI Experiment

In Global Navigation Satellite System Reflectometry (GNSS-R) research, polarization was historically overlooked. However, it has garnered increasing attention in recent years. This paper focuses on analyzing the first publicly available dual-polarization, single-frequency dataset from the airborne GLORI (GLObal navigation satellite system Reflectometry Instrument), collected during the summer of 2021. Based on theoretical analysis, multiple soil moisture retrieval algorithms were implemented. Initially, only the surface reflectivity of LR (LHCP-receiving/RHCP-transmitting) and RR (RHCP-receiving/RHCP-transmitting) polarizations was investigated. As additional surface parameters, such as surface roughness and vegetation, were integrated into the algorithm, the retrieval accuracy, measured by RMSE, improved significantly from approximately 0.07 cm3/cm3 to 0.03 cm3/cm3. The retrieval accuracy of RR polarization is slightly better than that of LR polarization. Nevertheless, when both dual polarizations were considered, the retrieval accuracy was comparable to that of using only one polarization. When surface roughness, Leaf Area Index (LAI), and incidence angle are taken into account, the soil moisture retrieval accuracy, indicated by RMSE, reaches 0.0344 cm3/cm3. This clearly demonstrates the great potential of dual polarization in soil moisture estimation. GLORI data is the first publicly available dual - polarization GNSS - R data that encompasses not only coherent and but also non - coherent scattering scattering information. This paper further discusses the non - coherent scattering properties of LR and RR polarizations. In the context of coherent scattering, it is found that the scattering properties at LR polarization are stronger than those at RR polarization. Conversely, for non - coherent scattering, the scattering properties at LR polarization are weaker than those at RR polarization for corresponding land surface types. By employing both cohrent and noncoherent scattering properties with the consideratons of LR and RR polarizations, the soil moisture retrieval accuracy indicated by RMSE(Root Mean Square Error) and correlation coefficients(R) will be 0.0316 and 0.9043, respectively. The analysis of dual-polarization data incorporating both coherent and non-coherent scattering mechanisms holds significant potential for advancing soil moisture retrieval accuracy through future data mining efforts, while also informing the design of next-generation polarimetric GNSS-R payloads such as ESA's HydroGNSS and China's GNSS-ReSAR.