Predictability of flood events in view of current meteorology and hydrology in the conditions of the Czech Republic

In central Europe, floods are natural disasters causing the greatest economic losses. One way to reduce partly the flood-related damage, especially the loss of lives, is a functional objective forecasting and warning system that incorporates both meteorological and hydrological models. Numerical weather prediction models operate with horizontal spatial resolution of several dozens of kilometres up to several kilometres, nevertheless, the common error in the localisation of the heavy rainfall characteristic maxima is mostly several times as large as the grid size. The distributive hydrological models for the middle sized basins (hundreds to thousands of km2) operate with the resolution of hundreds of meters. Therefore, the (in) accuracy of the meteorological forecast can heavily influence the following hydrological forecast. In general, we can say that the shorter is the duration of the given phenomenon and the smaller area it hits, the more difficult is its prediction. The time and spatial distribution of the predicted precipitation is still one of the most difficult tasks of meteorology. Hydrological forecasts are created under the conditions of great uncertainty. This paper deals with the possibilities of the current hydrology and meteorology with regard to the predictability of the flood events. The Czech Hydrometeorological Institute is responsible by law for the forecasting flood service in the Czech Republic. For the precipitation and temperature forecasts, the outputs of the numerical model of atmosphere ALADIN are used. Moreover, the meteorological community has available operational outputs of many weather prediction models, being run in several meteorological centres around the world. For the hydrological forecast, the HYDROG and AQUALOG models are utilised. The paper shows examples of the hydrological flood forecasts from the years 2002-2006 in the Dyje catchment, attention being paid to floods caused by heavy rainfalls in the summer season. The results show that it is necessary to take into account the predictability of the particular phenomenon, which can be used in the decision making process during an emergency.