Bridging Satellite Observations: A Comprehensive Intercomparison of Wave Spectra

Satellite remote sensing has fundamentally advanced the observation of directional ocean wave spectra, progressing from empirical and model-assisted retrievals toward physically based, direct spectral estimation methodologies, now supported by simulation capabilities enabled by higher computing performances. Current missions provide complementary spectral coverage across a broad range of ocean wave wavelengths. Sentinel-1/SAR (Copernicus missions) retrieves intermediate- to long-wavelength swell up to 800 m, primarily from Wave Mode (WV) acquisitions in the open ocean. More recently, Terrain Observation by Progressive Scans (TOPS) acquisitions, particularly over coastal regions, have enabled advanced mapping of the SAR-observed cross-spectra, through advanced signal-processing algorithms. CFOSAT/SWIM (Franco-China cooperation) provides multi-incidence directional measurements resolving intermediate wave components up to about 800m, while SWOT (Franco-American cooperation) extends observations toward longer swell wavelengths, reaching approximately 1200 m for some extreme cases. Together, these missions enable a multi-scale, multi-geometry characterization of ocean wave spectra. Forthcoming missions, currently under design and development, are expected to provide sensitivities to longer ocean wave and result in different imaging sensitivities for directional ocean wave spectra (ROSE-L) and offering multi-static measurements to enhance angular sampling and directional retrievals (Harmony). 

Despite these advances, no single sensor provides complete coverage of the ocean's directional wave spectra and wave products are often expressed across different spectral domains (frequency vs. wavenumber), physical variables (wave height vs. slope), and coordinate systems (Cartesian vs. polar). Limitations arise from radar band, incidence geometry, capability to resolve wave direction ambiguity and imaging mechanism: long-period swells remain largely beyond the reach of SWIM, which also suffers from contamination such as a long-track speckle noise; SAR systems are limited by azimuth cut-off effects and rely on quasi-linear inversion of the measured cross-spectra ; the directional wave spectra are ambiguous with respect to satellite azimuth. Additional constraints include contamination from non-geophysical signals (rain cells, atmospheric front, low winds, etc.) leading to non-geophysical wave products. 

To address these limitations, integrating multiple missions and performing systematic inter-comparisons and “normalization” is essential. By aligning overlapping spectral ranges, filling observational gaps, and accounting for mission-specific artefacts, this study aims at reconstructing a continuous and geophysical consistent directional wave spectrum, including the separation and partition-based evaluation of individual swell systems and the local wind-sea component. Complementary strategies—including advanced AI- or data-driven algorithms applied across the full spectral field—support robust quality control and the accurate computation of total and wind-sea significant wave heights. 

Integrating multi-mission observations with full-spectrum retrieval enables the most comprehensive reconstruction of directional ocean wave spectra from space. Independently validated against numerical wave models and in situ measurements, this framework provides a quantitative and robust characterization of wave spectra across multiple spatial and temporal scales. It establishes a scalable methodology for ensuring observational consistency and lays out the groundwork for next-generation remote sensing missions and advanced algorithmic developments in global ocean wave dynamics.