Investigation of high and low predictability periods in medium range forecasts.

Abstract

Medium-range five-day forecasts from the U.S. Navy Operational Global Atmospheric Prediction System (NOGAPS) are investigated to study high and low predictability periods from two winter seasons. Northern hemisphere 500 mb height fields are scored using the anomaly correlation coefficient- An objective method is used to choose high and low scoring periods which are analyzed using height tendencies and wavenumber structure. Results show that it is possible to objectively determine why some high and low periods occurred. Flow characteristics leading to high scoring five-day forecasts include: long wave amplitude decay, transition from meridional to zonal flow, and more meridionally extensive flow patterns. This study revealed that persistence is not a good indicator of model performance, and no appreciable skill difference exists between good and poor five-day forecasts at the 48 hour point- However, no single measure of the flow patterns is found to be a unique indicator of high or low scoring forecasts.