Diverse paradigms of residential development inform water use and drought-related conservation behavior

Abstract

Abstract Widespread urbanization has led to diverse patterns of residential development, which are linked to different resource consumption patterns, including water demand. Classifying neighborhoods based on urban form and sociodemographic features can provide an avenue for understanding community water use behaviors associated with housing alternatives and different residential populations. In this study, we leveraged built environment data from the online real estate aggregator Zillow to develop neighborhood typologies and community clusters via a sequence of unsupervised learning methods. Five distinct clusters, spatially segregated despite no geospatial inputs, were associated with unique single-family residential water use and conservation patterns and trends. The two highest-income clusters had divergent behavior, especially during and after a historic drought, thus unraveling conventional income–water use and income–water conservation relationships. These clustering results highlight evolving water use regimes as traditional patterns of development are replaced with compact, water-efficient urban form. Defining communities based on built environment and sociodemographic characteristics, instead of sociodemographic features alone, led to 3% to 30% improvements in cluster water use and conservation cohesion. These analyses demonstrate the importance of smart development across rapidly urbanizing areas in water-scarce regions across the globe.