Modeling Monthly Precipitation in Central Northeast of Thailand

Abstract

The purpose of this research is to model monthly extreme precipitation in central northeast of Thailand; Khon Kean, Kalasin, Roi-Et and Mahasarakham Provinces, by using the Generalized Extreme Value distribution. The maxima monthly rainfall, which are obtained from the Meteorological Department of Thailand, during January 1987 to September 2013; from six stations in the central northeast of Thailand were studied. There are four models via Generalized Extreme Value distribution with stationary (Form 1), the Generalized Extreme Value distribution in which the location parameter changes depending on linear trend (Form 2), the Generalized Extreme Value distribution in which the location parameter changes depending on quadratic trend (Form 3) and the Generalized Extreme Value distribution in which the location parameter changes depending on linear trend and the scale parameter changes depending on exponential (Form 4). The return levels for various return periods are estimated and the likelihood-Ratio test is evaluated as a criterion to find the best form. The study found that the Fréchet distribution is best fitted for all location when the process is stationary. Since the Kamalasai station of Kalasin province has the highest return level among various return periods,
so it should be the first to be considered station in preventing or reducing the severity of floods. Furthermore, if the process is non-stationary there is only the Roi-et Station at Muang district of Roi-et province is form 1, and the rest are form 4.