Statistical characterisation and estimation of non-domestic water demand

Abstract

Estimation of annual average water demand figures is critical for the design and evaluation of water distribution systems. This study evaluated the metered water consumption of more than 67,000 non-domestic consumers in six categories from cities and towns in South Africa. It was found that lognormal distributions provide good descriptions of the annual average daily demand (AADD) distribution in each category. The land use categories Business Commercial, Industrial, Agricultural holdings and Sports & Parks displayed similar median AADDs of between 1.5 and 1.7 kl/property/day. Educational properties used substantially more water (4.7 kl/property/day), while Government & Institutional properties used substantially less water (0.7 kl/property/day). A step-wise regression analyses showed that property size has the greatest impact on water demand for most categories. Finally, a novel statistically based method is proposed for estimating the average AADD of a given number of properties based on an acceptable risk of non-exceedance.

Keywords:

• Water demand
• non-domestic
• design demand

Acknowledgement

The authors would like to acknowledge the water demand data supplied by GLS Consulting.