Surface wind stress from global atmospheric analyses

Evaluations and comparisons of global analyses from the European Centre for Medium Range Weather Forecasts (ECMWF) and the US National Meteorological Center (NMC) have allowed an appreciation to be gained of the characteristics, problems, and errors in these data sets. For surface wind, the ECMWF's 1000-mb analyses are regarded as most suitable for computing surface wind stresses over the global oceans because ship data were applied directly at 1000 mb prior to September 1986. The global analyses use a four-dimensional data assimilation system and have the advantage of high spatial and temporal resolution. New climatologies of surface wind stress, the curl of the wind stress, and the Sverdrup transport stream functions have been computed on the basis of the 1980-1986 analyses. The wind stress was computed using a wind speed and surface-atmospheric-stability-dependent drag coefficient from W. G. Large and S. Pond, and the effective mean drag coefficient required to convert a pseudostress into a stress has been evaluated as a function of location and time of year. Comparisons with the S. Hellerman and M. Rosenstein (1983) wind stress climatology are presented.