Errors of Atlantic air-sea fluxes derived from ship observations

Abstract

Using individual ship reports of the Comprehensive Ocean-Atmosphere Data Set (COADS) monthly 1° × 1° averages of the air-sea flux fields in the Atlantic are computed to investigate the variance on a seasonal-to-interannual timescale. As an exemplary parameter the latent heat flux is chosen. The total temporal variance at each grid point is split up into four components: the error of the longtime mean, the seeming extramonthly variability, the mean error variance of monthly means, and the intramonthly variance. The spatial distributions of these components are discussed. In most regions the extramonthly variability is dominated by the error. One grid point in the North Atlantic is investigated in more detail. Even in this region of highest data density within the central Atlantic, it turns out that more than half of the temporal variance is caused by the errors of the monthly means.