Classifying households for water demand forecasting using physical property characteristics

Changes in population distribution across Europe are driving the construction of substantial numbers of new houses, creating a need to forecast water demand for new housing developments. The most certain information available on new households during planning are the physical characteristics of the properties themselves. This paper sets out to establish how to classify properties in terms of their physical characteristics for the purpose of forecasting water demand. Analysis of household water demand under a univariate classification of property type showed significant differences for properties of different size (number of bedrooms), architectural type (e.g. flats vs. terraced) and garden presence but not for age or for garden aspect. Analysis of household water demand under a multivariate classification of property type showed fewer significant differences between property types. The results of the study were compared to studies and found to fit qualitatively. However, quantitative differences were noted indicating geographical and sampling variation which requires further investigation. In addition, further research is required to determine the relative certainty of forecasts derived from physical vs. socio-economic or demographic characteristics. (C) 2008 Elsevier Ltd. All rights reserved.