Laserfiche WebLink
<br />0.035 for natural sandy channels. 10 associate all of the uncertainty on the roughll\:SS coefficient <br />would suggest a "true" coefficient of 0.055. Without knowledge of the coefficient used in the <br />indirect measurements, it is difficult to assess sources of the discrepancy. <br /> <br />The paleoflood investigation at Owl Creek tributary is likely to be picking up the large flood that <br />fell within the systematic record. It is possible the flow was slightly supercritical and the <br />reported peak of 3,400 cfs will be assumed accurate. From the analysis and PSI data, it may be <br />stated that this flood was probably the largest in last 100 to 1,000 years. <br /> <br />4.2 Flood frequency results <br /> <br />The following three sections discuss the flood frequency characteristics for each of the sites. <br />Each site provided different difficulties and insights into the flood frequency analysis. Such <br />uniqueness warrants adetailed discussion of each analysis. <br /> <br />4.2.1 Igo Creek tributary <br /> <br />Igo Creek tributary is 42 percent censored with five years of peaks falling below the minimum <br />recordable discharge of 5.0 cfs. One hundred scenarios were run weighting the station skew <br />against the regional skew of -0.1 (IACWD, 1982). This value of the regional skew was used at <br />all sites. Figure 10 illustrates the peak discharge data against the frequency curve. The Weib\lll <br />plotting position was used and is based on the .average plotting position for each flow point. A <br />Komolgorov-Smimov goodness of fit test was performed on each scenario. The log-Pearson III <br />distribution could not be rejected in any of the scenarios. <br /> <br />Storm data were available for this site. The storm used for calibration was that of July 6, 1978. <br />This storm measured 1.26 inches with duration of 55 minutes. The temporal distribution is \ <br />similar to the design storm pattern used in this study. The storm corresponds roughly to a 2- to \ <br />5-year precipitation event and produced a peak discharge near the 5-year flood. The basin was <br />calibrated and reproduced the peak within -4.5 percent and -17 percent on volume in HEC-HMS. l/-: <br />The SCS CN model was calibrated to +0.6 percent of the peak and +21 percent on total volume. . <br /> <br />The model parameters suggest a well drained, predominantly sandy soil with very low effective <br />impervious area. Parameters appear consistent with published values (Smith, 1999; Rawls, et aI, <br />1993). The curve number used, 75, suggests a fair to poor SCS Class B soil in semi arid regions <br />(Rawls, et ai, 1993). A full set of model parameters may be found in Appendices D and E. <br /> <br />Figure II illustrates the model simulations compared to the Bulletin 17B frequency curve. The <br />HEC-HMS simulation is good for return periods beyond 10 years. There is a little noise between <br />the 25- and 1 DO-year events, but the runoff simulations remain within the ensemble limits and <br />fairly close to the mean frequency curve. The 5-year event simulation is questionable and just <br />outside the upper limit of the ensemble bounds. The storm used for calibration was similar in <br />depth to the 5-year precipitation depth, but the design distribution resulted is a slightly greater <br />maximum storm intensity. The higher design intensity results in a smaller maximum infiltration <br />rate early in the storm. <br /> <br />14 <br /> <br />~/ <br />