Document Type


Publication Date

January 1979


Multiple regression and frequency analysis of average month, peak month, peak day, and instantaneous water use by various water supply systems in Utah and Colorado are used to develop water demand functions. The research objective was to predict water demand as a function of a small number of independent variables for which data were easily obtainable and thereby provide an attractive method for use by consulting engineers inf uture planning studies. The independent variables in landscaped area and accounts for use of supplementary ditch or pressure irrigation systems. The demand functions were developed with data from a sample of 14 systems varying in size from very small low density rural systems to Salt Lake City's water system. The correlation coefficients (R^2) vary from 0.80 to 0.95. The demand fucntions were validated by comparing calculated to measured water use for more than 40 other Utah systems. Instantaneous demands are determined for any desired recurrence interval as a function of number of connections. The demand functions are presented both at best fit (expected value) levels for average month, peak month, and peak day and at recommended design levels for the same time durations. The design levels were calculated by adding to expected values an increment which was based upon standard deviation of the samples. Instanataneous demand peaks which can be expecte donce in about 30 years in Utah systems are under 2 gpm per connection for lines serving more than 50 families, 3 gpm for lines serving 10 families, and 5 gpm for lines serving 4 connections.