The outputs from the global and regional climate models are fundamental data sources for the climate prediction and research. However, the data suffer substantially from the systematic errors, caused primarily by the rough spatial resolution of the climate models. A statistical correction of the climate model data is necessary prior to their usage in hydrological and other impact studies.
A broad spectrum of the mathematical statistics techniques is used to solve the problems mentioned above. The analysis of the probability distributions (parametric approach, the kernel density estimates etc.) is the essential topic with regard to the mutual transformations of investigated distributions. Also the time series analysis, the correlation and profile analyses are widely used. The principal components, canonical correlations and other techniques are used to study multidimensional data.