Changes in characteristics of maximum precipitation intensity in the period 1961-2017 in Zagreb

Cilj rada je određivanje projektne oborine na temelju proračuna vjerojatnosti maksimalnih intenziteta oborina trajanja 10 min do 24 h primjenom empirijskih i teorijskih distribucija vjerojatnosti te korištenjem povijesnih podataka o izmjerenim oborinama preuzetih s meteorološke stanice Maksimir za razdoblje 1961. – 2017. Podaci o maksimalnim količinama oborina su sortirani na način da se omogući korištenje empirijske distribucije vjerojatnosti po Čegodajevu. Korištene su teorijske distribucije vjerojatnosti pogodne za analizu ekstremnih vrijednosti, Gumbelova i Log – Pearson III raspodjela. Provedbom Kolmogorov-Smirnov testa zaključeno je da Log – Pearson III najbolje opisuje zadani skup podataka te su izrađene ITP krivulje. Nepoznati koeficijenti funkcije ITP krivulje određeni su metodom najmanjih kvadrata. ITP krivulje su prikazane analitički i grafički te su s njih očitane projektne oborine kao cilj ovog rada. Postupak je ponovljen za 2 podrazdoblja, 1961. - 1990. i 1991. – 2017., te je provedena usporedba s rezultatima za cijeli skup podataka.The goal of this research is determination of design storm based on the calculation of the probability of maximum precipitation intensity for durations from 10 min to 24 h using empirical and theoretical probability distributions. Historical data of precipitation is taken from meteorological station Maksimir for the period 1961 - 2017. Maximum precipitation data were sorted for the use of the empirical probability distribution according to Chegodayev. Theoretical probability distributions suitable for the analysis of extreme values were used, Gumbel and Log - Pearson III distributions. Kolmogorov-Smirnov test was conducted, and it was concluded that the Log - Pearson III best describes the data and the IDF curves were constructed. The unknown coefficients of the IDF curve function were determined by the least square method. IDF curves were presented analytically and graphically, and the design storms were determined as the goal of this paper. The procedure was repeated for 2 subperiods, 1961-1990 and 1991-2017, and a comparison with the results for the whole data set was performed