First-Time Soil Moisture Retrieval Using Single-Frequency Dual-Polarization GNSS-R with Airborne GLORI Data

Краткое изложение: Polarization was once overlooked in GNSS-R studies. However, in recent years, it has garnered increasing interest. This paper explores dual polarization for single-frequency data from the airborne GLORI experiment. Building on theoretical analysis, we have applied several retrieval algorithms for soil moisture estimation. Initially, only the surface reflectivity of LR and RR polarizations was examined. As additional surface parameters, such as surface roughness and vegetation, were incorporated into the algorithm, the retrieval accuracy, as indicated by RMSE, improved from around 0.07 to 0.03. The accuracy retrieved by RR polarization appears slightly better than that of LR polarization. However, when both dual polarizations were considered, the retrieval accuracy matched that of using only one polarization. When surface roughness, LAI, and incidence angle are taken into account, the retrieval accuracy, as indicated by RMSE, is 0.0344. This suggests that dual polarization holds great potential for soil moisture estimation. GLORI data is the first publicly available dual polarization GNSS-R data that includes both coherent and noncoherent scattering. This paper also discusses the noncoherent scattering properties of LR and RR polarizations. In the realm of coherent scattering, it is observed that the scattering properties at LR polarization exceed those at RR polarization. However, this trend reverses for noncoherent scattering, wherein the scattering properties at LR polarization are diminished compared to those at RR polarization for corresponding land surface types. The analysis of dual polarizations data will benefit future data mining for more accurate soil moisture retrieval and the design of future polarization GNSS-R payloads. While the retrieval accuracy with the consideration of noncoherent scattering properties indicates that not only coherent scattering but also noncoherent scattering should be included in future GNSS-R data sets, they are comparable for future soil moisture retrieval algorithms.