A Bayesian peaks-over-threshold analysis of floods in the Itajaí-açu River under stationarity and nonstationarity

摘要: In this paper we revisit the case study of Silva et al. (Stoch Env Res Risk A. doi:10.1007/s00477-015-1072-y, 2015), Itajai-acu River at Apiuna (Southern Brazil), with an augmented data set and Bayesian inferential techniques. Nonstationary Poisson-GP models are used to joint influence El Nino-Southern Oscillation (ENSO) upstream flood control structures on regime site. The Nino3.4 DJF index a dimensionless reservoir as covariates. Prior belief about GP shape parameter is elicited by fitting GEV distribution AMS samples from 138 sites in Southern Brazil 40 or more years deriving estimates that parameter. Following data-driven exploratory analysis, Markov chain Monte Carlo (MCMC) procedure sample posterior parameters. Model evaluation selection Bayes factors two information criteria. Results show evidence that, while dams play significant, though small, role reducing hazard, climate covariate stronger, increase ENSO amplitude last decades has led occurrence higher annual maximum floods. MCMC derive predictive quantiles design life levels. Uncertainty analyses based parameters presented.