Equidistance Quantile Matching Method for Updating IDFCurves under Climate Change

Abstract

Recent increase in intensity and frequency of catastrophic hydrologic events is shown to be a major threat to the global economy. These major events have been strongly linked to climate change and are expected to become worse in the future. The intensity-duration-frequency (IDF) curves quantify the extreme precipitations which are commonly used in planning and design of hydraulic structures. Since it is expected that the trends in extreme precipitation events will alter in the future, this will impact the IDF curves and they will have to be updated. In this study we present a methodology based on equidistance quantile matching (EQM) for updating the IDF curves under climate change. The two main steps in the proposed methodology are: (i) spatial downscaling of the maximum daily precipitation values from the global climate model/s (GCM) data to each of the sub-daily maximums observed at a station under consideration; (ii) explicit description of the changes in the GCM data between the baseline period and the future period (temporal downscaling). The main advantage of the proposed method compared to the existing methods which only use the spatial downscaling/disaggregation methods at baseline period, is that the proposed methodology additionally incorporates the changes in the distributional characteristics of the GCM model between the baseline period and the projection period. In addition, the method is simple to adopt and computationally efficient. To demonstrate the utility of the proposed methodology we use: (i) Canadian GCM model CanESM2 and (ii) its Representative Concentration Pathways (RCPs) for greenhouse gas concentration trajectories adopted by the IPCC for its Fifth Assessment Report (AR5) for the description of future conditions. The sub-daily annual maximum intensities are obtained for four stations in Canada located at London (Ontario), Hamilton (Ontario), Calgary (Alberta) and Vancouver (British Columbia). It is observed that the proposed approach is skillful in capturing and replicating the historical intensities and frequencies. The results indicated that for all RCP simulations considered in this study there is increase in precipitation intensities for all return periods. The relative increase in precipitation extremes is consistent with the RCP scenarios, i.e., the intensity of RCP-26 is lower than the RCP-45 which in-turn is lower than RCP-85. The quantile-based modeling without the temporal downscaling consistently underestimates the precipitation intensity when compared to the proposed method. The proposed method offers a valuable contribution to water resources planning and management of future extreme conditions.