Intercomparison of Global Research and Operational Forecasts

Abstract

Current limitations of atmospheric predictive skill are investigated through comparison of correlation and error statistics of operational and research global models for two winter seasons. In 1993, bias-corrected models produced anomaly correlations of 0.6 after 6.5?7 days, with relatively little forecast skill beyond that point. In 2003, the forecast skill of a more developed, higher-resolution operational model has been extended 36 h, while the skill of the unchanged, low-resolution research model has been extended 6 h. This implies more predictable patterns in 2003 or model and initial state improvements made since 1993. The relative importance of improved model resolution/physics and improved initial state to the lengthening of forecast skill is diagnosed through the evaluation of rms evolution of analyzed and forecast differences of 500-mb height and meridional wind. Results indicate that forecast sensitivity to initial data is less important than is the sensitivity to the model used. However, the sensitivity to model used (rms of model forecast differences) is smaller than the rms forecast error of either model, indicating model forecasts are more similar to each other than to reality. In 1993, anomaly correlations of model forecasts to each other reach 0.6 by roughly 8 days; that is, the models predict each other's behavior 1.5 days longer than they predict that of the real atmosphere. Correlations of model errors to each other quantify this similarity, with correlations exceeding the asymptotic value of 0.5 through the 14-day forecasts. Investigations of initial state error evolution by wavenumber show long waves (0?15) account for 50% more of the total uncertainty growth in 14-day research model integrations than do short waves (16?42). Results indicate current predictive skill may be impacted by model sophistication, but error pattern similarities suggest a common deficiency of models, perhaps in the initial state uncertainty.