Spatial optimization and interpretation of citywide sustainable promotion of building-based decentralized water technologies

Decentralized water technologies (DWTs) are emerging as alternative solutions to reduce freshwater withdrawal from the centralized infrastructure and improve water security. The benefits of DWTs effectively arise when they are widely implemented, requiring an appropriate sustainability assessment of economic and environmental impacts on urban water systems. However, existing decision-supporting tools remain inadequate regarding 1) evaluation of the trade-off between environmental benefits and costs and 2) interpretation of variations in the optimal selection of locations as the financial support increases. To fill the gaps, this study presents an integrated analytic framework combining multi-objective optimization and interpretation to facilitate citywide sustainable promotion of building-based DWTs. We selected single-family households in Boston (U.S.) as target users and applied our framework to study the promotion of rainwater harvesting (RWH) and greywater recycling (GWR) technologies. We used genetic algorithm to generate a set of optimal solutions by evaluating the trade-off between energy savings and investment costs. Our spatial optimization suggests RWH is more applicable than GWR for single-family house communities in Boston. In the interpretation phase, census blocks were grouped into two clusters with distinct selections of the prioritized locations for investing in RWH and GWR given different investment levels. One cluster located in the southern part of the city demonstrates a higher priority for installing RWH. We created a random forest classifier to analyze the critical features of the cluster to interpret why it is prioritized. Overall, our framework can be an additional supporting tool to direct the sustainable citywide promotion of DWTs.