Floods are a natural and a very dynamic phenomena. In Slovenia they may occur due to intense rainfall, snowmelt or a combination of both. Quality and reliable flood frequency analyses are necessary for effective flood protection, as floods caused extensive damage and can also threaten human lives.
In the thesis flood frequency analyses were carried out with annual maximum series method which is most often used in practice. Different theoretical distributions are presented and the method of Lmoments for parameters estimation. Some statistical tests used to determine the most appropriate distribution are described. A statistical Mann-Kendall test for determing trends is also presented.
In the practical part of the thesis flood frequency analyses for different 30-year periods of measurements for gauging station Slovenia are made. Six most commonly used distribution functions were used and parameters were estimated with the method of L-moments. Different tests were used to select the best fitting distribution for each measuring period and station. We compared the design flow of the selected return periods between the periods 1961-1990 and 1981-2010 and found that the selected measurement period has a significant impact on the results of the flood
frequency analyses or in other words that climate variability has a significant impact on the design flow of the majority of gauging stations in Slovenia.