One Model Does Not Fit All: Bottom-Up Indicators of Residential Water Use Provide Limited Explanation of Urban Water Fluxes

Abstract

Demand forecasting for water resources is an essential planning element for cities to provide secure and sustainable water supplies. A common way of determining water demands is to identify indicators at the household level across a city or region, utilizing socioeconomic and behavioral characteristics of homes to identify patterns and trends associated with metered water demand. This bottom-up approach to water demand forecasting is in contrast to top-down approaches characteristic of urban metabolism studies. In urban metabolism studies, the city is viewed as an entity with material demands and wastes, including the withdrawal, abstraction, and discharge of water. Household studies have identified several important demographic, infrastructural, and socioeconomic indicators that provide insight into water demand. However, do these or similar indicators scale from the metered household to the monitored city? In this study, we approach water resource demand from a top-down perspective, considering the city as a singular entity. We utilize city-specific, readily available indicators, scaled from bottom-up studies, for social, environmental, economic, and infrastructure conditions to model the variability of water demand across the country. Acknowledging that there are seasonal dependencies of water use, we investigate water fluxes at an urban scale using a previously assembled database. We find that there are minimal similarities between indicators from the household level to the city level, suggesting the limits of top-down urban metabolism studies.