Stationary and non-stationary generalized extreme value modelling of extreme rainfall with small sample size.

Autor: Siow, Chen Sian, Ghazali, Khadizah, Yusoff, Rubena, Gabda, Darmesah
Předmět:
Zdroj: AIP Conference Proceedings; 2023, Vol. 2746 Issue 1, p1-9, 9p
Abstrakt: This paper aims to fit the Generalized Extreme Value (GEV) distribution to the annual maximum rainfall at several sites of the West Coast district in Sabah. We compare the stationary and non-stationary GEV models to fit the annual maximum rainfall with a linear function of temperature as a covariate in the location parameter. In addition, the available observations at each site were limited; therefore, the generalized maximum likelihood estimation (GMLE) was applied to estimate the parameters of GEV distribution. Since the GEV was fitted independently at each site, therefore we applied the sandwich estimator to adjust the standard error. It was found that the additional covariate has no impact on the GEV parameter. Hence, the return level of each station has been obtained by using the stationary GEV model. [ABSTRACT FROM AUTHOR]
Databáze: Complementary Index