Laserfiche WebLink
<br />INHERENT INACCURACIES OF FORECASTS <br /> <br />Summary of Comments <br /> <br />Numerous comments were received concerning the fact that pro- <br /> <br />jecting or forecasting future conditions is not an exact science and <br /> <br />inaccuracies must be expected and planned for. It was stressed in the <br /> <br />comments that there is a need for quantification of inaccuracies, risk <br /> <br />factors, and margins of error in all phases of the analysis, including <br /> <br />data gathering, population projection, and model development. Several <br /> <br /> <br />people felt that draft Technical Appendix 2 presented modeling and <br /> <br />forecasting as an exact science and that the report presents a very <br /> <br />deceptive implication of validity and reliability with respect to the <br /> <br />final demand forecasts. <br /> <br />Response <br />The COE recognizes that the variables and the degree of variation <br />reflected in the demand forecasts are those that were evident during <br />the 1974 through 1982 period. It is also recognized that the influence <br />these variables have on water demand could change over time, and that <br />variables that do not appear in this analysis might influence water <br />demand in the future. However, it is believed that the variables <br />contained in the model have been affecting water demand in the Denver <br />metropolitan area for an extended period of time; therefore, there is <br />no reason at the present time to believe they will cease to influence <br />water demand during the forecasting period. <br /> <br />The inherent limitations in the process of forecasting water <br /> <br />demand will be explained in greater detail in the final Task 2 report. <br /> <br />In addition, the demand forecasts will be accompanied by sensitivity <br /> <br />analyses to enable risk assessments to be made. The final Task 2 <br /> <br />technical appendix will contain additional comments relative to error <br /> <br />margins, where applicable. However, a substantial portion of this work <br /> <br />5 <br />