In this study, the predictive capability of air-sea coupling is examined within the Navy's mesoscale predictive system COAMPS, forced at the lower boundary by an analyzed sea surface temperature (SST) updated at a 12-h frequency. We find that the COAMPS model contains robust air-sea coupling features congruent with satellite observations over the California current. Statistics from a 9-km horizontal resolution reanalysis over U.S. West coastal and offshore waters encompass 4 summer time periods between 2002-2005. Summertime means, monthly/daily/hourly standard deviations of SST, wind stress and their gradient or derivative fields are computed to ascertain variability and coupling between the ocean and overlying atmosphere. Evaluation of correlation maps and grid point time series of overlapping monthly averages suggests regions of strongest coupling reside upwind of persistent, localized ‘expansion fan' features in the marine atmospheric boundary layer wind field. These locations have strong wind forcing that drive coastal upwelling and transport SST filaments offshore, tightening SST fronts that then feedback to the atmosphere through the wind stress curl and divergence. In comparison with observed coupling coefficients documented by Chelton et al. (2007), COAMPS yields a weaker atmospheric response to SST gradients by 25%, about half of which is related to COAMPS SST which has 4 times the resolution of the satellite SST retrievals.