Predicting Future Flood Frequency Under Climate Change Using Copula Function

Abstract

AbstractThe joint behaviour of flood variables under climate change is of high importance for the economics of projects and risk reduction. This study investigates the implications of climate change using Gumbel–Hougaard copula function for future bivariate of flood peak and volume variables, in Azarshahr chay watershed. Canadian Earth system model (CanESM2) under three Representative Concentration Pathways (RCPs) along with statistical downscaling method (SDSM) and soil and water assessment tool (SWAT) were adopted to assess both baseline (1976–2005) and future (2030–2059) periods. Bivariate analysis of Copula improved the accuracy of model with an average NSE of 0.97 for all scenarios. Joint return period for severe floods has declined in the future, especially in RCP8.5. For a constant discharge and volume, joint return periods at the base period, RCP2.6, RCP4.5 and RCP8.5 were 24, 10, 13 and 9 years, respectively. Multivariate analysis may also provide useful information for flood risk assessment.