Ensemble classification as a method for improvement of the long-term weather forecast quality
V.Yu. Tsepelev
Russian State Hydrometeorological University, St. Petersburg, Russia
Keywords: ensemble weather forecast, monthly weather forecast, ensemble classification, development scenario, macro-sinoptic process, post-processing, forecast quality assessment, fields of sea level pressure and temperature anomalies
Abstract
A method is suggested for classifying weather forecast ensemble members and identifying the set of the ensemble members which most likely reflects the future state of the atmosphere. The first forecast month of the ensemble is used for the comparison between every selected class and observation data in order to identify the most realistic scenario of the development of atmospheric processes. The best class is used for prediction of the sea level pressure and surface temperature anomaly fields for the next month. The method suggested allows improving the quality of forecasts for north-west of the Russian Federation and the Arctic.
|
