Bayesian Markov Chain Monte Carlo-Based Copulas: Factoring the Role of Large-Scale Climate Indices in Monthly Flood Prediction

Floods are caused by heavy rainfall associated with variation of large-scale climate index, El Nino–Southern Oscillation (ENSO). The chapter applies an advanced statistical copula approach to model lag relationships between monthly Southern Oscillation Index (SOI), an ENSO indicator, and monthly Flood Index (FI) that can be used for flood prediction. Copula parameters were numerically derived from under a hybrid-evolution Markov chain Monte Carlo (MCMC) approach within a Bayesian framework. The empirical findings showed that monthly SOI data from Aug to Dec have a significant correlation with monthly FI that can be predicted at least four months ahead using SOI information. These advanced flood prediction models, presented in this chapter, are indeed imperative tools for civil protection and important to early warning and risk reduction systems.