Probability Modeling of Rainfall for Some Regions in Nineveh Governorate

Abstract

With the rise in extreme weather events like floods and droughts and their effects on agriculture, water dam building, and operation, it is critical to study the chance of monthly precipitation. However, the distribution of precipitation in the regions of Nineveh Governorate is still unknown. Thus, the aim of this work is to evaluate the probability distribution of precipitation frequency in the main regions of Nineveh Governorate (Mosul, Bashiqah, Tal Afar,and Rabe’ea). A probabilistic analysis was conducted using a 33-year historical precipitation dataset (1990-2022) obtained from the Meteorological Center of the Iraqi Ministry of Agriculture and the global weather data website Meteoblue. To achieve high accuracy in the probabilistic modeling of the data, the data for the June, July, August, and September months were not included in the analysis because rainfall is almost non-existent in these months. Seven advanced probability distributions_ Generalized Extreme Value, Generalized Gamma, Generalized Logistic, Pearson Type VI, and Pert were compared with the monthly data for total precipitation over the specified regions. EasyFit statistical software was used to estimate the probability functions of the mentioned distributions. And by finding the sum of the scores for three tests of goodness of fit, it was determined that the Generalized Extreme Value distribution is the best probability distribution to represent the monthly precipitation data over the city of Mosul with a fitting percentage of (76%), the Pert distribution is best fitted for the data of Bashiqah with a fitting percentage of (67%), the Gumbel Max distribution is fitted for the data of Tal Afar regions with a fitting percentage of (95%), and the Generalized Logistic distribution is the best for the data of Rabe’ea region with a fitting percentage of (95%). These results were confirmed by the root mean square error criterion, which was at its lowest value when using the mentioned distributions in accordance with the chosen research regions. The scientific results clearly demonstrate that the analytical process developed and evaluated in this study can be appropriately applied to determine the most appropriate probability distribution for weather parameters.