Determination of minimum inter-event time for storm characterisation in Johor, Malaysia

Abstract

Rainfall events are important in the analysis and design of hydraulic structures. To obtain valid statistical results, choosing an accurate minimum inter-event time (MIT) value is very important. However, an independent event is difficult to identify. The determination of MIT was done using two approaches; the first used an autocorrelation analysis, and the second examined the relationship between the estimated MIT and the observed average annual number of rainfall events. An autocorrelation analysis was performed by determining the lag time when an autocorrelation coefficient is sufficiently close to zero, and the examination approach was done by observing the time when further increase in MIT does not result in significant changes in the number of events observed. The analysis used hourly rainfall data from 15 stations in the state of Johor, Malaysia from the years 2007 to 2011. Both methods produced an MIT value of 8h. The characteristics of each storm, that is, depth, duration and intensity, were determined and fitted into various distribution models. Rainfall depth, duration and intensity are best fitted by the lognormal distribution using Anderson Darling, whereas gamma is the best distribution for rainfall duration and intensity, and Weibull distribution is the best for rainfall depth based on the Akaike Information Criterion.