Evaluation of Uncertainty in the Area Related Quantitative Precipitation Forecast of Heavy Convective Precipitation

Abstract. Extreme convective events have been studied with numerical weather prediction model LM COSMO (horizontal resolution of 2.8km) and radar data. An ensemble of 13 forecasts, which followed from modyfying initial fields, was created for two of analyzed events. Uncertainty in area-related quantitative precipitation forecast (QPF) was evaluated by fraction skill score (FSS) to quantify an ensemble spread and skill. The spread assesses a difference between a control forecast and ensemble member forecasts and the skill evaluates a difference between a precipitation forecast and radar derived precipitation. The uncertainty analysis shows an effect of forecast lead time and space scale as well as a sensitivity to the area structure of precipitation field. In this paper we summarize our first results dealing with the scale influence on the ensemble spread and skill.