Spatial extreme modeling using student t copula approach in Ngawi Regency

Abstract

Extreme rainfall is an unpredictable phenomenon which causes suffering effect such as flooding. Located on the Equator area, Indonesia results in a high intensity of extreme rainfall. Initial information regarding the patterns, characteristics, and rainfall prediction is needed in order to minimize the negative effect of such phenomenon. A method that can be used to predict extreme rainfall is the spatial extreme value using the copula approach. The copula approach used in this study is student-T copula. The Generalized Extreme Value (GEV) distribution used for the student-T copula with parameter estimation is Pseudo-Maximum Likelihood Estimation (PMLE). The proposed method was applied to model the extreme rainfall atNgawi Regency. An extreme spatial dependency on location is shown by extreme coefficient graphic. The best model that is obtained is based on Akaike Information Criteriation's (AIC) lowest value. The best model then continues to be used to predict the rainfall intensity return level. The prediction result of the rainfall intensity return level value shows that the maximum value of rainfall intensity increases from year to year in each station.