Effects of climate model interdependency on the uncertainty quantification of extreme rainfall projections

Changes in rainfall extremes under climate change conditions are subject to numerous uncertainties. One of the most important uncertainties arises from the inherent uncertainty in climate models. In recent years, many efforts have been made in creating large multi-model ensembles of both Regional Climate Models (RCMs) and General Circulation Models (GCMs). These multi-model ensembles provide the information needed to estimate probabilistic climate change projections. Several probabilistic methods have been suggested. One common assumption in most of these methods is that the climate models are independent. The effects of this assumption on the uncertainty quantification of extreme rainfall projections are addressed in this study. First, the interdependency of the 95% quantile of wet days in the ENSEMBLES RCMs is estimated. For this statistic and the region studied, the RCMs cannot be assumed independent. Then, a Bayesian approach that accounts for the interdependency of the climate models is developed in order to quantify the uncertainty. The results of the Bayesian approach show that the uncertainty is narrower when the models are considered independent. These results highlight the importance of accounting for the climate model interdependency when estimating the uncertainty of climate change projections.