Rainfall and temperature, simulated using Global Climate Models (GCMs), serve as key inputs for hydrological models in studying catchment responses to climate scenarios. GCM simulations of rainfall and temperature, however, are uncertain due to model structure, scenarios and initial conditions, which results in biased outcomes if used for hydrological models without due consideration of the uncertainties. In this study, we develop a novel uncertainty metric, square root error variance (SREV), to quantify uncertainties involved in GCM rainfall and temperature simulations as well as illustrate its application for water resources assessment. The uncertainty metric involves converting multiple GCM simulations into their percentile, estimating uncertainties at each quantile and translating these uncertainties into time-series. We apply the method to estimate uncertainties in rainfall and temperature simulations using multiple GCM, scenarios and ensemble runs. The utility of the uncertainty estimate for water resources assessment is illustrated through two case studies: (1) future drought analysis across the world; and (2) water availability study at the Warragamba catchment, Sydney, Australia. In the first case, future drought is estimated using Standard Precipitation Index (SPI) with simulation-extrapolation (an algorithm that reduces parameter bias when input errors are known) being used to reduce biases in SPI parameter. In the second case, an additive error model is proposed to generate rainfall and temperature realizations that are used to simulate streamflow. Future storage requirement of the reservoir is then evaluated with its associated uncertainty using behavior analysis. The results suggest that GCM uncertainty arises mainly from model structural errors, for both rainfall and temperature. Consideration of these uncertainties in drought analysis is vital, as drought values with and without considering the uncertainties are significantly different. It is also found that the existing storage capacity of the Warragamba reservoir suffices the future requirements, although large uncertainty exists in the storage estimates.
Woldemeskel, Fitsum Markos; Sharma, Ashish; Sivakumar, Bellie; and Mehrotra, Rajeshwar, "Quantifying GCM Simulation Uncertainty And Incorporating Into Water Resources Assessment" (2014). CUNY Academic Works.