Sekundarni povzetek: |
For the statistical analysis of high flows, the Log-Pearson III method with moments is usually used in Slovenia. This method gives unrealistically high values for higher return periods. The main reason for this study is to compare the usage of L-moments (Hosking and Wallis, 1997), as proposed in the Flood Estimation Handbook (FEH), with the existing statistical methods in use in Slovenia for single-site analysis. For this reason, I compared the Pearson III distribution with L-moments with the Pearson III distribution and the Log- Pearson III distribution using moments. As the data set (a series of annual maximum discharges Qmax) I used available data from Slovenian hydrological network of 300 stations. Even though the specialised software WINFAP-FEH is available on the market, I found it inconvenient to use for the whole Slovenian network (there is no direct support for Slovenian data). That is why I developed a MS Excel computer program running in the Visual Basic environment, supporting different two- and three-parameter statistical frequency distributions using the L-moments, and directly using the available files created in the Environmental Agency of the Republic of Slovenia. The output files of this newly created software can be easily used in Windows Office environment. The results obtained using the L-moments were compared for three return periods (50, 500, and 5000 years) with the other two applied methods using moments in terms of average values, absolute differences, catchment area of the measuring station, and the size of the data set. After all these criteria the new method with the L-moments proved stable, and the results ranged in-between the results yielded by the other two methods using moments. |