Extreme Precipitation and Flood Estimation Using Large Ensemble Climate Model Outputs: A Review

Extreme precipitation, flood estimation, and associated risk assessment encountered a paradigm shift in the big-data era with the emergence of large ensemble climate model outputs (LECMOs). By increasing the sample size, LECMOs strongly support a more robust estimation of extreme floods than those based solely on observational data or limited climate model outputs. Although several case studies on the development and applications of LECMOs in climate change were presented in the late 2010s, international reviews of such practices are lacking; consequently, up-to-date advancements are hindered from being incorporated into global flood risk management studies. This review provides an overview of recent cutting-edge developments in LECMO-based extreme flood estimation, including flood frequency analysis and probable maximum precipitation, as well as emerging applications in river/coastal flood risk assessment, compound flooding, and LECMO bias correction. A cross-sectional review of application studies of LECMOs from several countries/organizations revealed that they enable nonparametric extreme flood estimation and provide insights into the uncertainty of parametric functions, necessary sample sizes, and unprecedented events. Finally, this review raises several future perspectives and challenges with respect to handling big data, selecting subsamples, computational costs of impact assessment models, product selection, nonstationarity, and the combination of mathematical techniques in LECMOs for further research. Our cross-sectional review is expected to clarify methods for estimating extreme floods using LECMOs worldwide.