Advancing sustainable urban development through integrated water resources management and land use planning

Abstract

Urbanization, population growth, and climate change are placing significant strain on urban water systems. As cities expand, the conversion of natural landscapes to impervious surfaces increases flood risks and degrades water quality through contaminated runoff. Simultaneously, growing populations drive rising demand for already limited water resources. Despite the interconnected nature of these challenges, current planning practices often treat land use and water management as independent concerns. This fragmented approach limits the capacity of communities to develop holistic solutions that address both immediate water needs and long-term urban sustainability. Therefore, this dissertation addresses the key challenges in urban water sustainability through five research objectives designed to bridge the gap between water resource management and urban planning. The initial objective of this research aimed to validate the CLASIC (Community-enabled Life-cycle Analysis of Stormwater Infrastructure Costs) web-based urban hydrology model against the complex SWMM model. CLASIC is a cloud-based web tool that enables municipalities to evaluate and compare the lifecycle costs, hydrologic performance, water quality impacts, and co-benefits of different stormwater control measures, including green infrastructure, traditional gray infrastructure, and hybrid approaches. It was found that CLASIC accurately represents urban hydrological processes and effectively quantifies stream discharge for temporal scales greater than the catchment's time of concentration. Critically, this tool demonstrated a significant advantage of computational efficiency, especially as the drainage area increases, making it highly suitable for municipal-scale analyses. This validated tool was then used to quantify nutrient loads from Municipal Separate Storm Sewer Systems (MS4s) across Colorado, establishing a 2015 baseline and projecting future impacts under various Colorado Water Plan scenarios in 2050. Results revealed that the South Platte and Arkansas basins are the primary sources of statewide nutrient pollution, and while continued population growth threatens to worsen this issue, the implementation of advanced technologies and stricter regulations could lead to substantial load reductions. However, meeting stringent water quality targets, such as those in Regulation #31, was found to be a significant challenge for current technologies. Building on these findings, the research explored the trade-offs of vegetation planting strategies for stormwater management practices in Philadelphia. Through multi-objective optimization, the study demonstrated that for cities like Philadelphia, a combination of trees and sand filters provides a superior benefit-to-cost ratio, delivering higher triple bottom line benefits at a lower life-cycle cost than systems with surface vegetation like rain gardens. The optimal strategy involves prioritizing tree planting, followed by the incorporation of diverse surface vegetation based on available budgets and water quality goals. This highlights that simply implementing green infrastructure is not enough; the design and selection of vegetation type are crucial, as a lack of appropriate design and maintenance can lead to lower effectiveness in water quality enhancement and higher water demand and maintenance costs. The fourth objective was built upon these findings by developing a regional meta-model to estimate monthly water consumption at the zoning scale, only based on the characteristics of zones where no historical consumption is available. This model enables planners to evaluate the effects of different development patterns on water demand without sharing sensitive meter-level data. Utilizing the same dataset from the fourth objective, the model achieved strong predictive performance (R2 > 0.95) for both domestic and non-domestic water use. This study also confirmed again that per capita consumption decreases with increasing housing density due to reduced outdoor irrigation, while water use per acre rises with mixed-use intensification in denser areas. The fifth objective focused on developing a spatiotemporal machine learning (ML) framework to forecast monthly water consumption at the zoning scale. This work addressed the need for accurate, data-driven demand projections that integrate demographic, land use, and climate variables for cities. Using data compiled from eight Colorado utilities, the ML framework was trained to predict water use intensity metrics, including gallons per capita per day (GPCD) and million gallons per acre (MGA), for residential, CII, irrigation, and other user sectors. The model achieved high predictive accuracy (R2 = 0.90–0.98) and identified the history of water use, temperature, developed area, and household occupancy as key predictors. Scenario analysis demonstrated that infill development could reduce water consumption largely due to decreased outdoor irrigation, providing a valuable tool for evaluating future water demand under alternative growth and climate scenarios. Together, the forecasting model and the regional meta-model formed the analytical foundation for the POLARIS tool, an integrated platform linking land use and water resource planning to support sustainable decision-making for community development. In conclusion, this research provides holistic analysis and models for addressing urban water sustainability. It not only develops and validates new robust and computationally efficient models but also delivers critical insights into how strategic green infrastructure design and integrated land use planning can mitigate the impacts of urbanization and climate change. By providing decision-makers with the means to understand and optimize the complex trade-offs between hydrologic performance, economic costs, and community benefits, this dissertation lays a foundation for building more resilient, efficient, and sustainable urban environments for the future.