Identification of the most suitable probability distribution models for monthly and annual rainfall series in Guzelyurt Region, Northern Cyprus

Precise information of rainfall probability distribution is truly significant for many hydrological studies such as extreme flood analysis, drought investigations, reservoir volume studies, and time-series modeling. The objective of the present study is to identify the suitable probability functions for estimating the rainfall distribution in Guzelyurt Region, Northern Cyprus. Moreover, the aim of the study is to evaluate whether proposed distribution models have more satisfying performance than the commonly used models in previous scientific studies. Based on 33-y rainfall data collected from Meteorology Department located in Lefkosa, this study presents a statistical analysis of the average monthly, maximum monthly, and annual rainfall characteristics in the selected region with the assistance of 37 distribution models. Goodness-of-fit tests including KolmogorovSmirnov (K-S) test, Anderson-Darling (A-D) test, and Chi-squared (C-s) test are used to select the best-fit probability distribution model. Easyfit software was used in this study. The results indicate that Beta, Log-Pearson 3, and exponential (2P) distributions have the lowest value of K-S, A- D, and C-s, respectively, which are considered as the best distribution functions to study the average rainfall characteristics. Furthermore, among the best score results, Burr, Wakeby, and Nakagami distributions are giving the best-fit of actual data of total rainfall based on Goodness-of-fit tests. Additionally, the best fit probability distribution model for maximum monthly rainfall at the selected region was Nakagami, Wakeby, and Cauchy distributions based on Goodness-of-fit tests. The choices of the selected distribution model may be used for forecasting hydrologic events and defining the policies regarding water resource management and a source of data for flood hazard mitigation. This study provides a significant contribution to the current understanding of predicting hydrological events for various purposes.

Automated summary

This study aims to identify suitable probability functions for estimating rainfall distribution in the Guzelyurt Region, Northern Cyprus, and evaluate their performance in comparison to commonly used models. Based on statistical analysis and goodness-of-fit tests, the study found that Beta, Log-Pearson 3, and exponential (2P) distributions are the best for studying average rainfall characteristics, while Burr, Wakeby, and Nakagami distributions are the best fit for actual total rainfall data.