Probabilistic projections of hydrologic extremes under a warming climate

Abstract

Hydrologic extremes (i.e., droughts and floods) have become frequent and intense over many regions of the world during the past decades. Such extreme events have led to significant socioeconomic losses and serious environmental problems in the past. The intensification of the water cycle under a warming climate is likely to increase the frequency and intensity of hydrologic extremes in the future. Understanding hydrologic extremes and their impact thus plays a crucial role in climate change adaptation and disaster risk reduction. Although tremendous efforts have been made in the detection and projection of hydrologic extremes, most previous studies have their limitations: 1) the neglect of the relative performance of climate models and the potential uncertainty. 2) the inappropriateness of the stationary assumption in the precipitation intensity–duration–frequency (IDF) curves. 3) the heavy computational burden and structural variability of physically based hydrologic models. 4) the complex interaction of multiple mechanisms causing extreme storm surges (ESS) that is not explicitly uncovered. 5) emerging drought-downpour weather whiplash that remains poorly understood in terms of trends, drivers, and exposures. To address the abovementioned issues, novel approaches and high-resolution climate simulations are developed to provide comprehensive assessments of hydrologic extremes on different spatial scales. Specifically, 1) A Bayesian multi-model ensemble projection of multidimensional drought risk is developed for China. 2) A vine copula-based projection of future IDF curves is developed for China based on multi-model ensemble climate simulations. 3) A vine copula-based polynomial chaos framework is developed to improve multi-model projections of hydroclimatic regimes at a convection-permitting scale over a river basin in South China. 4) A robust data-driven framework is developed to improve the characterization of ESS in Hong Kong. 5) A comprehensive assessment of emerging drought-downpour weather whiplash is conducted on a global scale. The main findings are as follows. 1) The likelihood of extreme droughts is projected to increase in China under a warming climate. 2) China's 196 cities are projected to experience an increase in extreme precipitation of up to 30% in intensity and nearly two times the frequency of historical events under RCP8.5. 3) The vine copula-based polynomial chaos framework achieves an acceleration of more than 4,000 times faster and higher reliability than physically based hydrologic models. 4) The proposed data-driven framework demonstrates higher skill than previous statistical models in terms of characterizing ESS. 5) The frequency of drought-downpour weather whiplash shows significant increases over the poor region but shows insignificant change over the non-poor region during the period of 1980‒2010. The findings of this dissertation suggest that explicitly characterizing the interaction of hydroclimate variables can improve the reliability of risk assessments of hydrologic extremes. The relative performance of climate models should be considered to generate reliable future projections of hydrologic extremes; otherwise, misleading projections might be generated. The design guidelines of Chinese urban infrastructure should be updated to consider possible changes in the IDF curves induced by the changing climate. Growing poverty exposure to weather whiplash demands greater support for climate adaptation to reduce poverty and inequalities than previously understood.