Estimating Actual Water Savings: A Different Approach

摘要

In 2007, Dallas launched a toilet replacement and rebate program, called New Throne forrnYour Home. The program offers a financial incentive to home and multi-family buildingrnowners to replace old toilets which use 3.5 gallons per flush (gpf) or more with modern, lowflowrntoilets that use less than 1.6 gpf. Since its inception, over 21,000 toilets have beenrnreplaced. In an effort to determine the true water savings and cost-benefits derived from thernNew Throne program, statistical models were used to evaluate the program’s effectiveness.rnTraditional water savings estimates derived from engineering end-use studies heavily reliantrnon the technical characteristics of devices or programs have often proven to be inaccurate. Asrna result, Dallas used multiple linear regression models (MLR) to approximate actual waterrnsavings, while accounting for the factors of weather, seasonality and customer characteristics.rnThe New Throne program was used to demonstrate the application.rnOver 2,200 New Throne program participants were selected for this study. Customerrncharacteristics, seasonality, and weather effects were taken into account regarding both indoorrnand outdoor water use. An average of 1.66 high water use toilets were replaced per family.rnOur results indicate that each household reduced their monthly water usage by 982.5 gallons.rnThis amount is 46.9% greater than engineering end-use study estimates of 669 gallons. In thisrncase, the use of MLR models in lieu of engineering end-use studies signal the likelihood thatrngreater water savings were actually achieved.