Net sea–air CO₂ flux uncertainties in the Bay of Biscay based on the choice of wind speed products and gas transfer parameterizations

Abstract. The estimation of sea–air CO₂ fluxes is largely dependent on wind speed through the gas transfer velocity parameterization. In this paper, we quantify uncertainties in the estimation of the CO₂ uptake in the Bay of Biscay resulting from the use of different sources of wind speed such as three different global reanalysis meteorological models (NCEP/NCAR 1, NCEP/DOE 2 and ERA-Interim), one high-resolution regional forecast model (HIRLAM-AEMet), winds derived under the Cross-Calibrated Multi-Platform (CCMP) project, and QuikSCAT winds in combination with some of the most widely used gas transfer velocity parameterizations. Results show that net CO₂ flux estimations during an entire seasonal cycle (September 2002–September 2003) may vary by a factor of ~ 3 depending on the selected wind speed product and the gas exchange parameterization, with the highest impact due to the last one. The comparison of satellite- and model-derived winds with observations at buoys advises against the systematic overestimation of NCEP-2 and the underestimation of NCEP-1. In the coastal region, the presence of land and the time resolution are the main constraints of QuikSCAT, which turns CCMP and ERA-Interim in the preferred options.