The Impact of Analysis Error on Medium-Range Weather Forecasts

Abstract

All meteorological analyzed fields contain errors, the magnitude of which ultimately determines the point at which a given forecast will fail. Here, the authors explore the extent to which analysis difference fields capture certain aspects of the actual but unknowable flow-dependent analysis error. The analysis difference fields considered here are obtained by subtracting the NCEP and ECMWF reanalysis 500-hPa height fields. It is shown that the magnitude of this 500-hPa analysis difference averaged over the North Pacific Ocean has a statistically significant impact on forecast skill over the continental United States well into the medium range (5 days). Further, it is shown that the impact of this analysis difference on forecast skill is similar to that of ensemble spread well into the medium range, a measure of forecast uncertainty currently used in the operational setting. Finally, the analysis difference and ensemble spread are shown to be independent; hence, the impact of these two quantities upon forecast skill is additive.