Another approach to the problem of evaluating the water cycle over the ocean is to examine the output of numerical weather prediction models (NWP). There is optimism that the surface fluxes of the hydrologic cycle might be more accurately modeled than the heat flux, because the constraint of water conservation and lack of complex radiation terms makes for a potentially simpler problem. Recent improvements in precipitation algorithms have made the models more accurate, at least over land where comparisons can be made [ Environment Canada, 1993]. Preliminary comparisons of the monthly average E-P for two years from the European Center for Medium-range Weather Forecasts (ECMWF) model and the Schmitt, et al. [1989] 30 year climatology for the Atlantic are promising in the position and amplitude of major features. Both ECMWF and National Meteorological Center are planning ``reanalysis'' programs, in which the newer, more accurate models are run with data for the past one or two decades. If run with sufficient resolution, the time series of surface water flux over the ocean could be of great utility for climate studies.