Bivariate Flood Frequency Analysis of Sungai LUI Catchment in Malaysia Using Archimedean Copulas
In order to formulate mitigating measures against flood damages, flood frequency analyses are required. There are
several methods of flood frequency analyses applied throughout the world. This study will focus on bivariate flood frequency
analyses. Archimedean copulas have been adopted, using Gumbel, Frank and Clayton copulas. Sungai Lui catchment in
Malaysia was adopted as the study area. Akaike Information Criterion (AIC) was used to assess the copulas. L-moment was
used to assess the distributions. Graphical methods were used to find relationships between Flood Peak, Flood Volume and
Flood Duration. Results show that Generalised Extreme Value (GEV) was the best fitted distribution for Flood Peak, and
Weibul were the best fitted distribution for Flood Volume and Flood Duration. Gumbel was found to be the best copulas with
minimum AIC values. Relationships were established between Flood Peak-Flood Volume, Flood Peak-Flood Duration and
Flood Volume-Flood Duration.
Index Terms- Akaike Information Criterion, Clayton Copula, Flood frequency analysis, Frank Copula, Gumbel Copula, Joint
Cumulative Distribution, Marginal Distribution Parameter.