On differences in satellite wind products and their effects in estimating coastal upwelling processes in the south-east Pacific

Alongshore wind stress and curl obtained from ERS and QuikSCAT satellite wind products are routinely used to investigate coastal upwelling dynamics in the south-east Pacific. North of 17 degrees S and south of 26 degrees S there is a good agreement in the seasonal cycle of both products. Data from Peruvian coastal wind stations are used to validate satellite coastal winds. Between 17 degrees S and 26 degrees S seasonal cycles from coastal stations are in phase with QuikSCAT but in opposition with ERS. This discrepancy is confirmed during the overlapping period of the satellite products. Therefore, data from ERS do not seem to capture the strong cross-shore wind gradient in that area, while QuikSCAT renders at least part of it. These differences are dramatically enhanced when computing the wind stress curl along the coast: estimates of Ekman pumping obtained from QuikSCAT data in a coastal band are 2 - 3 times higher than those achieved with ERS data.