Correction and validation of Sentinel-1 IW radial velocity products using drifter and HF radar across the entire ocean environment

Since Sentinel-1 synthetic aperture radar (SAR) was launched in 2014, Interferometric Wide swath (IW) mode Level-2 radial velocity (RVL) products have been widely used to map fine-scale ocean surface current (OSC) in coastal zones. However, RVL product applications are restricted by non-geophysical and Wind-wave Induced Artifact Surface Velocity (WASV) errors. Previous studies have focused on improving the current retrieval accuracy in coastal zones, while neglecting open ocean regions and insufficient uncertainty analysis. To address these issues, a non-geophysical correction scheme suitable for both coastal and open sea is proposed by considering land coverage within SAR scenes. Corrected RVL products are validated using 1282 drifters and 78,054 HF radar points collected from the U.S. East Coast, West Coast, and Hawaiian Islands, showing overall accuracy improvements exceeding 60 %. To investigate the impact of WASV correction under different sea states (e.g. pure wind wave, wind wave dominant mix sea, swell dominant mix sea, and pure swell), a total of 127,534 matching points collected from January 2018 to May 2019 are used to assess the performance of four correction schemes. These include CDOP, KaDOP with wind and swell inputs, KaDOP with wind and wind-sea inputs, and CDOP-Y_n. A comprehensive comparison with HF radar current reveals that CDOP performs poorly in pure wind wave sea (RMSE up to 0.34 m/s), while incorporating sea state parameters enhances the retrieval accuracy. KaDOP and CDOP-Y_n yield comparable performance, while KaDOP performs better in pure wind or wind wave dominant mix sea, achieving RMSE of 0.21 m/s and a correlation coefficient (r) of 0.62. The correlation between SAR-derived and in-situ currents also varies with incidence angle, satellite track, and polarization. Overall, these results provide reliable OSC data for mesoscale and sub-mesoscale ocean dynamics research.