Estimation of significant wave height using GNSS interferometric reflectometry

Abstract

Reflected GNSS signals contain valuable information about their reflecting surface. While reflections are typically discarded in geodetic applications, GNSS-R makes use of them, particularly in environments with large reflective areas such as water. In coastal regions, surface waves are a dominant environmental feature. A commonly used parameter to describe wave conditions is the significant wave height (SWH). Its estimation using satellite altimetry can be challenging in coastal zones due to the mixture of land and water within the sensor footprint. In such cases, in-situ measurements are often required to bridge these observational limitations. This study investigates the estimation of SWH from ground-based Global Navigation Satellite System Interferometric Reflectometry (GNSS-IR) using standard geodetic antennas. Reflected GNSS signals contain interference signatures visible in the signal-to-noise ratio (SNR), from which surface roughness can be inferred. Following an modeling approach, a roughness parameter s of the reflecting surface was estimated for individual reflection arcs across multiple GNSS frequencies. In a second step, s was related to in-situ SWH from nearby moorings using a Total Least Squares regression. Two coastal stations along the German North Sea coast and one station on the U.S. West Coast were analyzed. Results show that only 10% of the available time series is sufficient to derive stable parameters, enabling robust SWH predictions over the remaining 90% of the observation period. The resulting RMSE values range from approximately 18 cm to 34 cm across the analyzed frequencies. A transferability analysis demonstrates that parameter sets derived at one site can be applied to others without recalibration. For two of the stations, such transfer is successful for 6-7 of the 9 available frequencies, highlighting the potential for generalized parameter models. An azimuth dependent analysis showed that signals arriving from the direction of the incoming waves exhibited smaller RMSE values than those from other directions, with the opposite direction resulting in the largest errors. Time-series analyses indicate that GNSS-IR-derived SWH generally follows the temporal evolution of the in-situ observations. Moreover, periods without mooring measurements can be reconstructed using GNSS-IR.