Extremeness of meteorological variables as an indicator of extreme precipitation events

Extreme large-scale precipitation events in central Europe are determined by the synoptic pattern, which is typically characterized by cyclones moving from the south into this region. Herein, we demonstrate that this synoptic pattern is accompanied by large-scale anomalies of dynamic and thermodynamic variables that occur in specific regions before or during the precipitation event. The extremeness of individual variables was evaluated by ERA-40 data relative to the corresponding climatology by the obtained distribution function values. Precipitation events leading to 14 significant summer floods in the period between 1958 and 2002 were used as reference events. The temporal correspondence of the anomalies among reference events was employed by selecting 26 synoptic predictors of extreme precipitation in the Czech Republic. A predictor evaluates the extremeness of a variable in a limited area on a specific day relatively to the start of the precipitation event (day D). While the predictors cover a five-day period from D − 2 to D + 2, most are relevant to the days with precipitation (from D to D + 2).The predictors cannot be considered individually because only opportune coincidences make the synoptic situations dangerous. In the present study, a combined EM index was used by evaluating the predictors as a whole, without considering their relationships. Days with minimum EM values correspond to significant flood events in the Czech Republic. Our results confirm that the synoptic pattern can be properly characterized by the extremeness of selected dynamic and thermodynamic variables, and that the anomalies are distinctly correlated with the extremity of large-scale precipitation and flood events in the Czech Republic. This approach is applicable to other regions where precipitation is significantly related to large-scale synoptic patterns, and could be employed also by operational detection of upcoming extreme precipitation events.