Decision Tree-Based Conditional Operation Rules for Optimal Conjunctive Use of Surface and Groundwater

Conjunctive use (CU) of surface water (SW) and groundwater (GW) implies the optimal operation of water resources to reduce the negative effects compared to when each of these resources is managed separately. Optimal operation of CU systems is not possible without providing operation rules that are valid for future uncertain conditions. The current study presents decision tree-based conditional operation rules for a distributed CU system in which the SW bodies and GW aquifer(s) are treated as physically interconnected and operationally joint parallel storage facilities. To derive the operation rules, the operation optimization model of the CU system is implemented with the objective function of minimizing water shortage for different river flow time series. Using five efficiency criteria, the comparison of conditional and single linear regression operation rules is conducted. The results show that the conditional operation rules reduce relative absolute error by a minimum of 39% and a maximum of 71% compared to single linear regression. The maximum water shortage is also projected to be close to 16.61 MCM over ten years.