Using Synoptic Classification to Evaluate an Operational Weather Forecasting System

Verifying forecast quality over extended periods, i.e. monthly, seasonally or annually, can provide misleading results since performance is averaged over a wide range of meteorological conditions. The danger is that differences in forecast quality can be masked or results can be biased when the data are not homogeneous, even in terms of flow regimes. A relevant weather classification scheme must therefore be developed and applied. The aim of the study is to identify weather situation-dependent weaknesses and strengths of the COSMOGR model that is used operationally at the Hellenic National Meteorological Service (HNMS). A tailor-made weather classification was constructed focusing mainly on the origin of the air masses that are responsible for the presence of certain weather regimes in the region. The period between December 2009 and June 2011 was categorized and the model outputs were subsequently verified, with a particular emphasis on the predicted precipitation, cloudiness and temperature.