Retrieval of ocean surface current using single-antenna synthetic aperture radar (SAR) relies on the removal of the wind and wave signal (wind-wave bias) from the observed SAR Doppler shift. This is often performed using deterministic atmospheric and wave models input into a geophysical model function, though atmospheric models are chaotic, and their predictability is flow-dependent. 251 Sentinel-1 SAR scenes, obtained in 2023 over Skagerrak and Kattegat, are analyzed to assess the uncertainty in the radial velocity estimation caused by uncertainties in the wind speed and direction provided by MEPS, an operational regional atmospheric ensemble system. Using ensemble surface wind direction and speed from MEPS, we investigate the conditions under which the radial velocity retrieval uncertainties exceed community-guided thresholds. The maximum wind speed and direction values that meet the specified thresholds are also estimated using two approaches. Our findings show that retrievals have a higher degree of uncertainty when the wind speed uncertainty exceeds 20% of the mean field, with larger uncertainties associated with low-wind speed conditions. A strong dependency between satellite antenna-look and maximum uncertainty values is reported. Employing ensemble models on radial velocity uncertainty quantification has promising potential and may be extended to operational global products.