Joint probability analysis of storage reservoir system characteristics

A comprehensive evaluation of the frequency distribution analysis of reservoir storage capacity plays an important role in reducing the uncertainty in the design and operation of reservoir systems. In most studies, a simplified reservoir system (such as over-year system and using annual data, without incorporating evaporation losses and performance measures, i.e. reliability, vulnerability, and constant monthly demand) was investigated. On the other hand, the required reservoir capacity variable is not independent of other characteristics of the reservoir system such as performance measures, critical period, and evaporation losses, and has a high correlation with these variables. Therefore, in this study, the joint probabilities of different characteristics of the reservoir system were comprehensively developed using copulas. Hence, an extensive Monte Carlo reservoir simulation procedure incorporating monthly streamflow data, performance measures, evaporation losses, and demand was applied to extract the population of dependent variables of an actual reservoir. The results showed that a) the required storage capacity in the reservoir system is highly dependent on other reservoir characteristics such as performance measures, critical period, and evaporation losses; b) the joint probability of the reservoir system characteristics are well modeled by a Gaussian copula and c) the joint probability of the reservoir characteristics has a systematic behavior for given reliability and vulnerability values.

Automated summary

A comprehensive evaluation of the frequency distribution analysis of reservoir storage capacity significantly contributes to reducing uncertainty in the design and operation of reservoir systems. This study incorporated copulas to develop joint probabilities of different reservoir system characteristics, such as performance measures and evaporation losses, which are highly correlated with the required reservoir capacity. Results from the extensive Monte Carlo reservoir simulation demonstrated the dependence of the required storage capacity on other reservoir characteristics, and the systematic behavior of joint probabilities for given reliability and vulnerability values.