Laserfiche WebLink
USE AND EFFECTIVENESS OF MUNICIPAL WATER RESTRICTIONS DURING DROUGHT IN COLORADO <br />account for temporal persistence in the time series of <br />accuracy in predicting water use (as shown in the <br />municipal water use. Similar approaches have been <br />right -hand column of Table 3); r- squared values <br />used in earlier drought studies, including those by <br />ranged from 0.62 to 0.77. Undoubtedly, more sophisti- <br />Anderson et al. (1980), Lee and Warren (1981), Maid - <br />cated advanced regression techniques .could lead to <br />ment et al. (1985), Maidment and Miaou (1986), Shaw <br />even greater accuracy. However, this level of accuracy <br />and Maidment (1988), and Shaw et al. (1992). <br />is more than sufficient for our purpose of describing <br />The regression model has the form <br />drought response in this case study. The regression <br />Water Use Use <br />equations were applied to data from the summer of <br />Y = 00 + 01x1 + 02x2 + ... + Rkxk + E <br />2002 to estimate expected use (i.e., what per capita <br />Use Use <br />use would have been absent restrictions and given cli- <br />where y is the response variable (i.e., per capita water <br />mate conditions) during periods of watering restric- <br />use), Ro is the regression constant, R1 is the slope coef- <br />tions. The difference between expected (calculated) <br />ficient for the first explanatory variable (xl), 02 is the <br />water use and actual (measured) water use provides <br />slope coefficient for the second explanatory variable <br />an estimate of the water savings that can be attribut- <br />(x2), Rk is the slope coefficient for the kth explanatory <br />ed to the drought inspired water restrictions. This is <br />variable (xk), and a is the remaining unexplained <br />shown graphically in Figure 1 for the city of Westmin- <br />noise in the data (the error). This model uses three <br />ster. <br />explanatory variables: daily data on maximum tem- <br />Lafayette 46 49 50 — — — 53 55 <br />perature (xl), daily data on precipitation (x2), and <br />0.69 <br />one -day lag variable of water use (x3). The coefficients <br />Presentation of Results <br />in the regression equations were estimated using data <br />per capita use (deliveries) in 2002 with that level of use anticipated in 2002 had water restrictions not been in effect and given the adverse <br />from the year 2000 and tested on the year 2001. Both <br />Tables 3 and 4 summarize the calculated effective - <br />summers of 2000 and 2001 had no watering restric- <br />ness of water restrictions, both voluntary and manda- <br />tions. This cross - validation exercise demonstrated <br />tory, over the study period. Percent savings in Table 3 <br />that these very simple equations had considerable <br />are based on the methods described above [i.e., "net <br />TABLE 3. Water Savings During Water Restrictions (May through August, 2002). <br />Basis of Percent Savings, Calculation1 <br />Entire Study Period <br />Voluntary Restrictions Period <br />Mandatory Restrictions Period <br />Per <br />Expected <br />Per <br />Expected <br />Per <br />Expected <br />Municipal Net Capita <br />Use Per <br />Net Capita <br />Use Per <br />Net Capita <br />Use Per Model <br />Water Use Use <br />Capita <br />Use Use <br />Capita <br />Use Use <br />Capita Skill <br />Provider ( %) ( %) <br />( %) <br />( %) ( %) <br />( %) <br />( %) { %) <br />( %) (r2) <br />Providers Limiting Lawn Watering to Once Every Three Days (2 -1/3 times/week) <br />Thornton -8 1 9 -7 2 10 — — <br />— <br />0.71 <br />Aurora 9 12 16 — — — 13 15 <br />18 <br />0.72 <br />Denver Water 7 10 13 2 5 7 14 16 <br />21 <br />0.67 <br />Westminster 4 7 14 3 6 11 17 19 <br />27 <br />0.70 <br />Average 3 7 13 0 4 9 14 17 <br />22 <br />-- <br />Cities Limiting Lawn Watering to Twice a Week <br />Fort Collins 9 13 18 3 7 12 17 20 <br />24 <br />0.63 <br />Boulder 24 24 27 -2 -2 4 29 28 <br />31 <br />0.62 <br />Louisville 39 39 41 — — — 43 43 <br />45 <br />0.77 <br />Average 24 25 29 0 2 8 30 31 <br />33 <br />-- <br />Cities Limiting Lawn Watering to Once a Week <br />Lafayette 46 49 50 — — — 53 55 <br />56 <br />0.69 <br />1 "Net use" compares daily system wide water deliveries in 2002 to the 2000 to 2001 average for the same dates. "Per capita use" standardizes <br />the net use calculation by accounting for population growth over the 2000 to 2002 period. "Expected use per capita" is a comparison <br />of actual <br />per capita use (deliveries) in 2002 with that level of use anticipated in 2002 had water restrictions not been in effect and given the adverse <br />climatic conditions associated with drought. In all cases, negative numbers indicate an increase in water use. <br />2 Averages are calculated from nonrounded values. <br />.JOURNAL OF THE AMERICAN WATER RESOURCES ASSOCIATION 81 <br />JAWRA <br />