Decision Tools for Integrated Water Management: Examining Distributed Capacity For Reuse

Abstract

In water-short areas, particularly those in high growth regions, water reclamation and reuse can comprise a significant portion of the regional water budget. The integration of wastewater reclamation and reuse systems within regional water and sanitation facilities plans remains a complex matter of engineering importance. Scenario planning can be a useful tool for integrating water and wastewater system development in the presence of future uncertainties regarding growth, water demand characteristics, legal/administrative constraints, water resources availability and so forth. A high growth portion of the Tucson, AZ, water management area was selected as a study site. Stress on Southwestern water resources, including those of southern Arizona, is widely acknowledged (USBOR 2018). To illustrate the utility of scenario planning for water resources/sanitation management in water-stressed areas, a multi-step process was undertaken to (i) define the problem to be solved with scenario planning, (ii) identify the most relevant planning uncertainties, and (iii) establish rational bounds for those uncertainties. The process led to the development of several scenarios, or alternative futures for the Tucson planning area that attempt to define the probable limits of factors affecting water supply and wastewater treatment in the future. Each scenario, defined by a set of water demands, a quantity of water available through the Central Arizona Project (CAP), and a decision about whether certain demands outside the city limits would be served as part of the system, was imagined to be equally probable. The regional water/sanitation system was reduced to an array of nodes (demand points) and links (potential connecting pipes) on a grid consisting of square mile units. A genetic algorithm was then used to find a near-optimal infrastructure development plan specific to each scenario. Inputs to the model included topographic data, a projected pattern of development, trajectory of water demand, limits to the future availability of Colorado River water to the region, and potential sites of wastewater reclamation. Outputs were near-optimal pipe and pump locations and sizes and reclamation facility locations and capacities. By comparing the solutions for several individual scenarios, the range of possible values (e.g. commercial pipe sizes) for each decision variable was reduced to the range encountered among the various scenario-dependent solutions. A family of compromise solutions was then obtained using the NSGA II multi-objective heuristic algorithm. These outcomes were selected to provide both low total present worth costs, expressed as the sum of overpayment and supplementary cost, and low variation of costs among scenarios. The sensitivity of the compromise solution with respect to total demand, available supply, cost of additional water supplies, and discount rate was determined. Results were most sensitive to the cost of additional water supplies.