Page 423 - Contributed Paper Session (CPS) - Volume 6

CPS2008 Syafrina A.H et al.
of Saudi Arabia, Abdullah and Al-Mazroui (1998) analysed annual rainfall and
found out that the main season for rainfall was spring, followed by summer.
Gamma distribution was found to be the best model to describe the rainfall
analysis in Saudi Arabia. In Ibadan metropolis, Oseni and Ayoola (2012)
compared the statistical distributions for rainfall data by using two types of
test from goodness of fit model, which were Chi Square (CS) and Kolmogorov-
Smirnov (K-S) tests. Based from the study, Exponential distribution was the
best fitting distribution followed by Poisson, Normal and Gamma distributions.
In Malaysia, Suhaila and Abdul (2007) observed that Mixed Exponential was
the best model for estimating the daily rainfall amount in Peninsular Malaysia
with four different types of wet day. The selection was performed by testing
the median of absolute difference (MAD) between the empirical and
hypothesized distributions, the traditional Empirical Distribution Function
(EDF) Statistics which were K-S statistic, Anderson Darling (AD) statistic and
Cramer-von Mises (CvM) statistics.
In 2016, Poisson-Gamma distribution was found to be the most
appropriate model for the two components which are the amount (rainfall
totals on wet days) and occurrence (dry/wet days) of rainfall simultaneously
(Rossita et al., 2016). The study was based on the four selected stations in
Peninsular Malaysia namely Bayan Lepas, Jelebu, Kuala Terengganu and
Mersing. The most common parametric distribution models that usually been
used by the researchers were Gamma, Weibull, Exponential, Gumbel and
Kappa. Based from the previous studies, Gamma distribution is widely used to
evaluate the rainfall data. One of the reasons is Gamma distribution is well
known and easy to understand with the parameters are used to analyse the
characteristics of the rainfall regimes (Husak et al., 2006). Gamma distribution
with two parameters gave considerable weight for the wet-day amounts
(Thom, 1951). For another distribution, Weibull family of distributions has
various sensible characteristics, which are invertibility, integrability and closed
form expressions for all moments (Duan et al., 1998). Having characteristic of
heavy tail distribution, Weibull was widely used in analysing the rainfall data
(Syafrina et al., 2017).

2. Methodology
The daily rainfall data at Penang International Airport (Station ID: 486010)
from the year 1990 to 2017 was provided by the Global Summary of the Day
(GSOD) and National Atmospheric and Oceanic Administration (NOAA). The
latitude and longitude of this study are 5.2961°N and 100.2752 °E respectively.
Penang International Airport was hit by flood caused by the rainstorm on July
2016 and about 14 flights have been put on hold due to heavy rain and strong
wind (Predeep, 2016). All these problems had affected to the economy, people

412 | I S I W S C 2 0 1 9