Evaluation of the Influence of Climate Variability on Flood Risk in Moderately Impaired Watersheds

Annual maximum flood risk is commonly assessed under the assumption of stationarity, i.e., flood risk is unaffected by trends, periodicity, climatic variables, and anthropogenic activities. In light of recent findings, however, this assumption should be reconsidered. Analyses of the nonstationary character of flood series induced by natural climatic variation, climate change, and/or anthropogenic activities are essential to develop more appropriate statistical models for flood risk assessment. This study aimed to investigate nonstationarity in annual maximum flood series for moderately impaired sites in the Eastern United States. Linear trends in flood series were identified for almost 30% of the investigated sites. A larger number of sites exhibited a change point implying a shift in the mean annual maximum flood magnitude. A high degree of correlation was observed between flood magnitude and associated precipitation series; however, no significant influence of temperature on flood magnitude was detected. Investigation of teleconnections between flood series and climatic indices revealed significant relationships with the Atlantic Multidecadal Oscillation (AMO) and the Pacific Decadal Oscillation (PDO) for sites located in Indiana, Ohio, Kentucky, and the New England region; few sites exhibited significant correlation with the Multivariate ENSO Index (MEI). Efforts were also made to distinguish the effects of anthropogenic activities on flood response from the influence of natural climatic variation.