Statistical modeling of extreme rainfall in peninsular Malaysia

Liew, Woon Shean (2020) Statistical modeling of extreme rainfall in peninsular Malaysia. Master dissertation/thesis, UTAR.

Preview

PDF Download (1250Kb) | Preview

Abstract

Flash floods are known as one of the common natural disasters that costs over billions of Ringgit Malaysia throughout the history, and monsoon season that is known as the rainy phase of a seasonally changing pattern is the main period of occurrence of flash floods. Academically, an extreme rainfall model is effective in modeling, so as to predict and prevent the occurrence of flash floods. A reliable extreme rainfall model would help in reduce the cost and rate of mortality in the occurrence of f lash floods. This study is to compare four probability distributions, which include Exponential distribution, Generalized Extreme Value (GEV) distribution, Gamma distribution, and Weibull distribution, with the data for rainfall from 10 stations in Peninsular Malaysia for the period of northeast monsoon from November to February. The time span of the data is from 1975 to 2008. The comparison is based on the performance of descriptive and predictive analytics of models. Rainfall data is cleansed by applying peak-over-threshold approach to obtain data that are more suitable to be use in modelling extreme rainfall. Determination of the most effective model is relying on both numerical result through Kolmogorov-Smirnov, Anderson-Darling, and Chi-Squared Test, and graphical result through using the quantile-quantile plot. Result shows that GEV is the most preferred extreme rainfall model to the rainfall cases in Peninsular Malaysia.

Actions (login required)