No abstract

This study aims to construct two hedging policies based on current storage and two auxiliary factors (past storage trend and future standardized streamflow index (SSI) and compare the effects on reservoir performance in terms of shortage characteristics (maximum single-period shortage ratio, total shortage ratio, average water shortage per shortage period, and risk) during droughts. The proposed approach is applied to the Nanhua Reservoir located in southern Taiwan. The results reveal that Model S (demandbased rule curves associated with fuzzified storage) efficiently improves shortage characteristics during droughts and outperforms Model C (current operation). Further improvements are obtained by incorporating past storage trends (Model STs: Model S with different periods of past storage trends) and future SSIs (Model SHs: Model S with different time-scale SSIs) into Model S. Exactly known SSIs in Model SHs derive optimistic hedging policies that have fewer less-than-1 rationing coefficients and significantly reduce shortage duration and total deficits. In contrast, Model STs lack future inflow information and lead to conservative hedging policies, which have early hedging and effectively decrease the maximum single-period water shortage. The effects of past storage trends and future SSIs on shortage characteristics decrease with longer periods since models with short-term information effectively capture the inherent variations and derive more effective hedging policies. According to the overall objective, Model SHs generally outperform Model STs, models with short-term information outperform the long-term models, and all the proposed optimization models outperform the current operation.