Development of Water Resources Drought Early Warning System

Signs of impending drought are often vague and result from hydrologic uncertainty. Because of this, determining the appropriate time to enforce water supply restrictions is difficult. This study proposes a drought early warning index (DEWI) that can help water resource managers to anticipate droughts so that preparations can be made to mitigate the impact of water shortages. This study employs the expected-deficit-rate of normal water supply conditions as the drought early warning index. An annual-use-reservoir-based water supply system in southern Taiwan was selected as the case study. The water supply simulation was based on reservoir storage at the evaluation time and the reservoir inflow series to cope with the actual water supply process until the end of the hydrologic year. A variety of deficits could be realized during different hydrologic years of records and assumptions of initial reservoir storage. These deficits are illustrated using the Average Shortage Rate (ASR) and the value of the ASR, namely the DEWI. The ASR is divided into 5 levels according to 5 deficit-tolerance combinations of each kind of annual demand. A linear regression model and a Neuro-Fuzzy Computing Technique model were employed to estimate the DEWI using selected factors deduced from supply-demand traits and available information, including: rainfall, reservoir inflow and storage data. The chosen methods mentioned above are used to explain a significant index is useful for both model development and decision making. Tests in the Tsengwen-Wushantou reservoir system showed this DEWI to perform very well in adopting the proper mitigation policy at the end of the wet season.