| description abstract | Waste at the household level is an unwarranted strain on water supply systems, particularly during droughts. Conservation measures enhance resilience. Some advocate for price as an efficient instrument for encouraging conservation, whereas others contend that residential water demand is price inelastic, and mandates and education campaigns can be more effective. Consumers too are concerned that future droughts may curtail water availability and are in favor of voluntary water conservation campaigns. Here we show how education alone, without any pricing or legal interventions, can help utilities achieve substantial savings. We focus on the impacts that a set of educational interventions had on about 15,000 single family residences (SFRs) in College Station, Texas for the period 2008 to 2018. The study period includes one of the worst one-year drought on record. We first compare consumption patterns before, during, and after the drought to establish a baseline consumption, and then analyze the impacts of educational interventions that were implemented as a response to the drought. We compute the landscape irrigation budget for each SFR to determine if they are wasting water. We found that winter consumption—representative of indoor use—did not change drastically, whereas summer consumption was significantly influenced by weather. We show, by mining water use data, that consumption patterns and trends were influenced by both the drought and the educational interventions. Notably, the interventions contributed to a reduction in supply, despite an increase in the number of SFRs. Furthermore, through clustering, we identified and profiled the largest water wasters that should be prioritized to enhance conservation. We therefore suggest that a policy approach involving both mandatory and voluntary educational conservation programs may prove to be most efficacious in reducing outdoor water use and waste. Water utilities collect a wealth of data which can be mined to extract valuable information on consumption patterns—such as the impact of drought restrictions, the effect of educational interventions that promote conservation, or identify the highest water users including the water wasters—through a process known as knowledge discovery in databases. In this paper, using consumption data from about 15,000 single-family residences in College Station, Texas, over an 11-year period, we demonstrate how we can use a combination of statistical and machine learning tools to extract spatial and temporal patterns and trends in winter and summer consumption as well as outdoor water use. The study period includes one of the most severe one-year drought that this region experienced. We found that winter consumption—representative of indoor use—remained relatively stable, whereas summer consumption was significantly influenced by weather. We also found that targeted voluntary water conservation campaigns, without any pricing or legal interventions, can help utilities achieve significant savings. The data mining approach developed in this paper can be easily replicated by utility managers and other stakeholders. | |
