FLOOD06431 - Laserfiche WebLink

Home Browse Search

peak discharges; s is an index of variability equal to the mean standard deviation of the logarithms (base 10) of the observed annual-peak discharges at the gaging stations used in each respective regression-model data set; and SEp is the average standard error of prediction, in log units (base 10), estimated usmg the Press statistic. Several of the primary drainage-basin characteristics used in the regression equations listed in table 2 are map-scale dependent. Use of maps of scales other than the scales used to develop the equations may produce results that do not conform to the range of estimation accuracies listed for the equations in table 2. The scale of map to use for manual measurements of each primary drainage-basin characteristic is outlined in Appendix A and Appendix B. An additional constraint in the application and reliability of the channel-geometry characteristic equations is thE' requirement to obtain onsite measurements of bankfull or active-channel width, and possibly bankfull depth. Training and experience are required to properly identify the bankfull and active- channel features in order to make these measurements. The variability in making these measurements can be large, even among experienced individuals. As reported by Wahl (1976), based on a test condueted in northern Wyoming, the standard error in estimated discharge due to variation in width measure- ments alone was about 30 percent (0.13 log unit). Variation in bankfull-depth measure- ments probably would increase this standard error in estimated discharge. Wahl (1976) also noted an average bias with respect to the mean channel width of about 14 percent (0.06 log unit). A truer total standard error, in log units, for a channel-geometry discharge estimate is calculated by Wahl (1984, p. 6;3) as the square root of the sums of the squares of the errors of the regression equation and of the variation and average bias in width measurements. Using the standard error of estimate for the Region I, 100-year flood bankfull equation (table 4) and assuming the standard errors for measuring channel width reported by Wahl (1976), the true standard error = [(0.192)2 + (0.13)2 + (0.06)2] 0.5, ~ 0.240. This yields an average standard error of 59.6 percent compared to 46.4 percent for the regression equation alone. Wahl (1984, p. 64) notes that the variability of the measurements collected in the Wyoming test probably is greater than normally would be encountered in applying channel-geometry measurements in a particular hydrologic area. Sites in the Wyoming test were chosen for their diversity, and they ranged from ephemeral streams in a nearly desert environment to perennial streams in a high mountain environment. Despite the limitations associated with the channel-geometry method, the equations presented in this report are considered to be useful as a corroborative flood-estimation method with respect to the drainage-basin method. The channel-geometry equations are applicable to all unregulated, stabilized stream channels in the State, whereas the drainage- basin equations are applicable only to stream sites with drainage areas less than 1,060 mi2. Although the error of measurement may be larger for channel-geometry characteristics than for drainage-basin characteristics, the variability of channel-geometry measurements made in Iowa are assumed to be not as great as reported by Wahl (1984) for the Wyoming test. An additional advantage in utilizing the channel-geometry method is that design-flood discharge estimates obtained from each flood-estimation method can be used to calculate a weighted average as described in the following section. Weighting Design-Flood Discharge Estimates Design-flood discharges determined using both the drainage-basin and channel-geometry flood-estimation methods are presumed to be independent from each other. Each flood- estimation method thus can be used to verify results from the other; when design-flood discharge estimates are independent, the independent estimates can be used to obtain a weighted average (lACWD, 1982, p. 8-1). 34 ESTIMATING DESIGN-FLOOD DISCHARGES FOR STREAMS IN IOWA