U.S. flag

An official website of the United States government

Dot gov

Official websites use .gov
A .gov website belongs to an official government organization in the United States.


Secure .gov websites use HTTPS
A lock ( ) or https:// means you’ve safely connected to the .gov website. Share sensitive information only on official, secure websites.


Main content area

Prediction of urban residential end-use water demands by integrating known and unknown water demand drivers at multiple scales I: Model development

K. Rathnayaka, H. Malano, M. Arora, B. George, S. Maheepala, B. Nawarathna
Resources, conservation, and recycling 2017 v.117 pp. 85-92
consumer demand, irrigation, models, people, planning, prediction, residential areas, stochastic processes, urban areas, water shortages, water utilization
Detailed prediction of water demand by their end-uses at multiple scales is essential to support planning of Integrated Urban Water Management, an increasingly applied approach to deal with the problem of water scarcity. This paper presents an urban residential water demand modeling framework that can predict end-use water demand at multiple scales, especially at small scales with a robust explanatory capacity. This is achieved by integrating the complex water demand dynamics of urban residential water use and their underlying variables into a single model. The model described in this study can predict shower, toilet, tap, dishwasher, clothes washer, irrigation, evaporative cooler, bath, and other uses which account for the entire household water use. The model aims to predict water demand at multiple spatial (household/cluster/suburb) and temporal scales (hourly, daily, weekly and seasonal) by considering behavioral differences triggered by factors such as seasonality and presence of people at home. The model incorporates an improved representation of spatial variability by considering behavioral differences between customer groups, and improves the capability to deal with areas with different demographic and housing characteristics. This research confirms the capacity of stochastic modeling methods to represent unexplained behavior of water consumers.