Author(s): Sania B. Mahtab
Linked Author(s):
Keywords: Generalized Extreme Value; Generalized Normal; Generalized Logistic; Generalized Pareto; Pearson Type III; Wakeby
Abstract: In recent years, it has been recognized that society has become more vulnerable to extreme storms. Many studies have hence been carried out to investigate the variation of these extreme events. Of particular interest for urban water infrastructure design is how to estimate accurately the extreme rainfall amount and its probability of occurrence using frequency analysis methods. Consequently, several different probability distributions have been recommended for these rainfall frequency analyses. However, there is no general agreement as to which distribution should be used. This paper proposed therefore an assessment procedure for evaluating systematically the performance of some commonly used probability distributions in order to identify the most suitable models that could provide the most accurate extreme rainfall estimate for urban infrastructure design. More specifically, the proposed procedure was based on the evaluation of the descriptive ability of six popular probability distributions in the estimation of extreme rainfalls: the Generalized Extreme Value (GEV), Generalized Normal (GNO), Generalized Logistic (GLO), Generalized Pareto (GPA), Pearson Type III (PE3), and Wakeby (WAK) distributions. To test the feasibility of the proposed method, an illustrative application was carried out using annual maximum rainfall data for various durations (3h, 6h, 9h, 12h, 18h, and 24h) from a network of eight stations located in Bangladesh. These stations represent the diverse climatic conditions over the country. Based on numerical and graphical goodness-of-fit criteria, results of this application have indicated that the GEV, GNO and PE3 distributions are the best models that can be used for describing accurately the distributions of annual maximum rainfalls in Bangladesh.
Year: 2024