Skip to main content
Article
Forecasting urban water demand: A meta-regression analysis
Journal of Environmental Management (2016)
  • Maamar Sebri
Abstract
Water managers and planners require accurate water demand forecasts over the short-, medium- and long-term for many purposes. These range from assessing water supply needs over spatial and temporal patterns to optimizing future investments and planning future allocations across competing sectors. This study surveys the empirical literature on the urban water demand forecasting using the meta-analytical approach. Specifically, using more than 600 estimates, a meta-regression analysis is conducted to identify explanations of cross-studies variation in accuracy of urban water demand forecasting. Our study finds that accuracy depends significantly on study characteristics, including demand periodicity, modeling method, forecasting horizon, model specification and sample size. The meta-regression results remain robust to different estimators employed as well as to a series of sensitivity checks performed. The importance of these findings lies in the conclusions and implications drawn out for regulators and policymakers and for academics alike.
Keywords
  • Meta-analysis; Forecasting; Urban water demand; Accuracy; Mean absolute percentage error (MAPE); Box-Cox transformation
Publication Date
Fall September 30, 2016
DOI
http://dx.doi.org/10.1016/j.jenvman.2016.09.032
Citation Information
Maamar Sebri. "Forecasting urban water demand: A meta-regression analysis" Journal of Environmental Management (2016)
Available at: http://works.bepress.com/maamar_sebri/19/