WSP02699 - Laserfiche WebLink

Home Browse Search

to significant physical and climatic characteristics are defined. The standard error of estimate for each equation Is also given. For example, the relation for the mean annual flow Is Qa a 0.106 AO.947 (p - 14.0) 1.61 Ev.0.771 and the standard error of estimate Is 28 percent. The standard errors achieved In the regression are compared with the accuracy goals for each flow characteristic In table 6. Except for the low-flow characteristics, the accuracy goals are met for the Important characteristics. It is important at this point, however, to consider the range of basin characteristics represented by the gaging stations 'used In the regression analysis. Table 6.--Comparison of regression equation to the accuracy goals for minor streams with natural flow. Streamflow characteristics Standard error (percent) 10-year record Regression equation Mean annual flow Standard deviation of annual flow Mean monthly flow (average) Standard deviation of monthly flow 50-year peak flow 50-year 7-day high flow 2-year 7-day low flow 20-year 7-day low flow 30 25 50 25 58 62 53 73 28 30 38 44 57 41 306 410 Applicability of regression equations Because regression equations usually do not represent the actual physical relationships, they should not be applied to conditions outside the range of data from which they were developed. Size of drainage area Is the most obvious limitation; areas used ranged In size from 110 to 9,110 sq mi. In a multiple regression, not only the range of each Independent vari- able but also the range of combinations of two or more Independent variables should be considered. Examples of combinations of two variables are shown on figures 4, 5, and 6 (combinations of more than two variables are not shown because,of the difficulty of graphical representation). Both the 28