Probability Distributions to Describe Rainfall Pattern in North-Western and Eastern Coasts of Sri Lanka

Abstract

The changes in the global rainfall have raised concerns over the present and future use of coastal areas in many Asian countries including Sri Lanka. The coasts of the North-western and Eastern provinces of Sri Lanka harbor varied important dynamic industries such as aquaculture, agriculture, and tourism industry. However, no in-depth analysis of rainfall of coastal lines of Sri Lanka has been performed in the past to understand any changes in the pattern in general and to mitigate the effects of those changes on these sectors. In this study, an attempt has been made to determine the best-fit distribution from the monthly rainfall data spanning 30 years of observations from 1987 to 2016 at 19 rain gauge stations as an initial step to identify rainfall patterns and future shifts. To fit the observed data, 45 different distributions were considered. The efficacy of the fits for these distributions was evaluated using three empirical nonparametric goodness-of-fit tests, namely Kolmogorov–Smirnov, Anderson–Darling, and Chi-square test. Results showed that the. Johnson SB was the best-fit distributions in stations of North-western province and Gen. Pareto, Generalized Extreme Value distribution, and Frechet (3P) were the best-fitted distributions for Eastern province. The results of this study would be useful to the water resource engineers, policymakers, and planners for the agricultural development and conservation of natural resources of North-western and Eastern coastal lines.