Selection of the Best Probability Distributions of Rainfall Events and Development of the Intensity Duration Frequency Curves in Al-Khoser River Basin, Northern Iraq

It is crucial to intensively study the anticipation of rainfall events and select the most appropriate probability distribution when conducting hydrogeological studies. This study uses frequency analysis to investigate and extract the probability of future events in Al-Khoser River Basin based on the maximum annual rainfall depth. Eleven distinct probability distribution functions, including exponential, generalized extreme value, Gumbel, Weibull, normal, log-normal, log-normal (3 parameters), gamma, generalized gamma, inverse gamma, and Pearson Type III, were fitted by using Hyfran-Plus software to simulate the observed probability of rainfall depth. The selection of the best probability distribution was determined through the use of Akaike information criteria and Bayesian information criteria tests. Among the tested distributions, the Gumbel probability distribution is applied by using the maximum likelihood method, demonstrated the best fit for frequency analysis in the study area, followed by log-normal and gamma distributions. The Soil Conservation Service method type IΙ was employed to generate synthetic storms lasting 24 hours, considering cumulative and incremental rainfall depth distribution. These synthetic storms were designed for the arid and semi-arid regions, with periods of return of T = 5, 10, 25, 50, and 100 years. In order to enhance the storm management approaches in this region, rainfall intensity duration frequency curves were constructed using simulated storms.