Reliability of probabilistic analysis of flood events can only be assured by a sufficient deeply engaged analysis of flood waves. In-depth analysis includes an analysis of peaks, volumes and durations of flood waves.
The theoretical part of the thesis comprises analysis of hydrograph and baseflow. Methods of baseflow extraction are presented, with an emphasis on automated procedures whose aim is to establish a simple and repeatable models for the extraction of baseflow. The program R is also presented, which includes packages for different baseflow separation methods.
The practical part of the thesis includes the comparison of different baseflow separation methods for the selected flood wave of the gauging station Radovljica I on Sava River in 2004. We compare the BaseflowSeparation method and BFI method with different filters and number of passages. The method that gave us the best result was chosen for the analysis of flood waves of selected gauging stations in Slovenia. We analysed peak, volume, duration and slopes of the flood waves.
The lingage between indivudual pairs of variables was investigated for 50 gauging stations in Slovenia. The results show that maximum correlation is between the peak and the volume of flood wave, medium correlation between the volume and the duration of flood wave and the minimum
correlation is between the peak and the duration of the flood wave.