Best fitting distributions for the standard duration annual maximum precipitations in the aegean region

Abstract

Knowing the properties like amount, duration, intensity, spatial and temporal variation etc. of precipitation which is the primary input of water resources is required for planning, design, construction and operation studies of various sectors like water resources, agriculture, urbanization, drainage, flood control and transportation. For executing the mentioned practices, reliable and realistic estimations based on existing observations should be made. The first step of making a reliable estimation is to test the reliability of existing observations. In this study, Kolmogorov-Smirnov, Anderson-Darling and Chi-Square goodness of distribution fit tests were applied for determining to which distribution the measured standard duration maximum precipitation values (in the years 1929-2005) fit in the meteorological stations operated by the Turkish State Meteorological Service (DMI) which are located in the city and town centers of Aegean Region. While all the observations fit to GEV distribution according to Anderson-Darling test, it was seen that short, mid-term and long duration precipitation observations generally fit to GEV, Gamma and Log-normal distribution according to Kolmogorov-Smirnov and Chi-square tests. To determine the parameters of the chosen probability distribution, maximum likelihood (LN2, LN3, EXP2, Gamma3), probability-weighted distribution (LP3, Gamma2), L-moments (GEV) and least squares (Weibull2) methods were used according to different distributions.