An Uncertainty-Based Random Boundary Interval Multi-Stage Stochastic Programming for Water Resources Planning

Given the increasing population growth and rapid global economic development, the conflict between reduced available water resources and increased water demand in many countries has become a challenging issue; confirming the necessity of allocating optimal water resources to equate water conservation and environmental sustainability. Uncertainties and climate change together deteriorate the complexity of optimal allocation related to water resources used for the agricultural sector. The present study sought to design an uncertainty-based random boundary interval multi-stage stochastic programming (URBIMSP) model by integrating the Kataoka's criterion method and chance-constrained programming within a general optimization framework to control the uncertain information which exists in both constraints and goal functions. The developed model was applied to a case study of water resources allocation in the Marand basin, Iran. Then, a conjunctive water allocation system was arranged to identify the introduced model in two planning periods which were 2023 to 2025 and 2026 to 2028. The results indicated that different water allocation schemes and water shortages with varying net system benefits would be achieved under different stream flow levels and different risk scenarios. The result was significant for carrying out a comprehensive analysis of the tradeoffs between total net benefits and related risks. Finally, the findings demonstrated that improving irrigation efficiency and upgrading irrigation systems are the key strategies for farmers to achieve economic benefits while also help to mitigate environmental impacts.