Evaluation of Nearshore QuikSCAT 4.1 and ERA-5 Wind Stress and Wind Stress Curl Fields over Eastern Boundary Currents

Abstract

Fields of coastal wind stress and wind stress curl in the 10–100 km next to the land control the processes of upwelling and downwelling of nutrients and water properties that are vital to highly productive coastal marine ecosystems. Here we ask the question: Do the present surface wind stress products from a satellite-borne scatterometer (QuikSCAT) and an atmospheric reanalysis model (ERA-5) systematically overestimate the magnitude of wind speed and stress in the 10–50 km next to the coast? We compare QuikSCAT wind speed retrievals to the relatively unused wind speed retrievals from satellite altimeters, which are able to approach closer to the coast than scatterometers without land reflections, due to their smaller radar footprints. Altimeter data on tracks approaching and crossing the coast indicate that the increases in coastal QuikSCAT wind speed values and ERA-5 coastal wind stress values are unrealistic. For analyses of wind speed and stress requiring high accuracy, especially those involving wind stress curl, we suggest considering individual Level 2B scatterometer wind retrievals as suspect at distances of 10 km and less from the coast, along with use of the Poor Coastal Processing flag. We found that similar increases in wind stress values next to the coast in gridded ERA-5 fields are not due to errors in the model physics or wind speeds. They are created during the interpolation of wind stress from the original model grid to a regular rectangular grid. We recommend that researchers who are analyzing wind stress and wind stress curl should calculate wind stress themselves from the gridded ERA-5 vector wind speed fields, rather than using the interpolated model wind stress or curl fields.