ESTIMATING WATER DEMAND DETERMINANTS AND FORECASTING WATER DEMAND FOR NZIOA CLUSTER SERVICES AREA

Abstract

The accuracy of water demand projections depends on the availability of reliable population and water use data as well as an understanding of the distribution of different types of users within the community. The underlying problem for this study is that water demand in Kenya is based on the fact that operational demand of drinking water is based on experience and appropriated practices, rather than local empirical evidence. There is limited number of analytical studies on water demand and supply reliability. In the face of limited knowledge, per capita use statistics adapted from developed countries are applied to estimate water consumption in Kenya, and most probably will fail to depict the water use patterns. At the same time, there is the unknown component of suppressed consumption induced scarcity and water quality problems. Almost certainly, will release these constraints, will modify and disrupt the water demand and design baseline. Finally it is crucial to establish time varying water consumption patterns and the critical demand values. Correct prediction of these factors determines the extent to which a network can satisfy critical demand and maintain economic efficiency. The objective of this study was to model water demand mathematically to determine the significance of water determinants for systems design and operations management. To achieve this objective, a survey of water usage in towns in Bungoma and Trans Nzoia counties was done. The survey was done in Bungoma town and its environs, Webuye town and its environs, kitale town and its environs of Nzoia Water Services clustered company (NZOWASCO). Out of the sample size of 23,000 population a sample size of 517 consumers was chosen across the five categories of consumers namely; domestic, low income, commercial, industrial and institutional. The presentation has focused on development of water determinant for domestic and low income consumers. Out of 517 population, the Primary data was collected in the field by use of structured questionnaire, whereas secondary data was collected from secondary sources of NZOWASCO. The primary data was used to develop the regression model. The secondary data was used for sensitivity tests and validation of the regression model. The results of the model were presented in various forms and compared with actual data for a period of 2005 and 2014. The model was able to determine historical water demand and water forecasts for domestic and low income consumers. The model generated values did not vary significantly with actual data for historical water demand and forecasting.