Climate informed flood frequency analysis and prediction in Montana using hierarchical Bayesian modeling

Abstract

It is widely acknowledged that climate variability modifies the frequency spectrum of extreme hydrologic events. Traditional hydrological frequency analysis methods do not account for year to year shifts in flood risk distributions that arise due to changes in exogenous factors that affect the causal structure of flood risk. We use Hierarchical Bayesian Analysis to evaluate several factors that influence the frequency of extreme floods for a basin in Montana. Sea surface temperatures, predicted GCM precipitation, climate indices and snow pack depth are considered as potential predictors of flood risk. The parameters of the flood risk prediction model are estimated using a Markov Chain Monte Carlo algorithm. The predictors are compared in terms of the resulting posterior distributions of the parameters that are used to estimate flood frequency distributions. The analysis shows an approach for exploiting the link between climate scale indicators and annual maximum flood, providing impetus for developing seasonal forecasting of flood risk applications and dynamic flood risk management strategies