Quantitative analysis of small-scale roughness effects in spaceborne GNSS reflectometry informed by L-band microwave radiometry

Spaceborne Global Navigation Satellite System Reflectometry (GNSS-R) has shown great potential for terrestrial monitoring, but the compensation of small-scale surface roughness remains an important challenge. In particular, the root-mean-square surface height σ h , its incidence angle dependence represented by n in c o s n θ , and their relationships with soil moisture (SM) are still insufficiently characterized for spaceborne GNSS-R. In this study, CYGNSS surface reflectivity (SR) observations are combined with model-simulated smooth bare-soil reflectivity derived from SMOS SM and vegetation attenuation represented by SMOS-IC L-band Vegetation Optical Depth (L-VOD). Global σ h and n are estimated on a 25 km grid by fitting the model predictions to CYGNSS observations. The retrieved σ h is compared to multiple topographic variables, SMAP and SMOS-IC roughness parameters, and Saharan aerodynamic roughness measurements, while the dependence of n on vegetation and land cover is also investigated. Results show that the retrieved σ h is more strongly associated with terrain descriptors such as uniformity, Shannon index, entropy, Northness, and Eastness than with conventional elevation- and slope-based variables. Both σ h and n also exhibit relationships with SM and vegetation, indicating potential coupling between effective roughness characterization and vegetation-related uncertainties. After compensation for vegetation and roughness effects, CYGNSS SR shows good consistency with simulated bare soil SR. Additional evaluation using in situ SM observations from the International Soil Moisture Network further supports the effectiveness of the retrieved parameters. Overall, these findings offer valuable insights into the use of physics-based empirical or semi-empirical models in spaceborne GNSS-R for soil moisture retrievals and other terrestrial applications.