Laserfiche WebLink
<br />OJ13~8 <br /> <br />19 <br /> <br />Similar flood peak relationships could be developed for other study <br />regions provided the data listed in Table II-l is available. The pro- <br />cedures for developing these relationships are essentially identical to <br />those discussed earlier. However, instead of applying selected rainfall <br />events, the entire extended rainfall record is applied to the calibrated <br />model to generate a continuous runoff hydrograph. The flood frequency <br />analysis is then performed on some series (partial duration or annual) <br />of flood peaks extracted from this hydrograph. <br />An alternative continuous hydrologic simulation approach has been <br />reported by Lumb and James (63). Instead of developing general flood- <br />frequency-urbanization regression relationships, they develop four <br />series of runoff responses and. store them in computer files. Each file <br />consists of the runoff response (as a function of area) from four <br />typical "sub-units" within a watershed (impervious, high infiltration, <br />medium infiltration, and low infiltration soils) for various recurrence <br />intervals. The total runoff hydrograph from a particular watershed and <br />recurrence interval can then be computed by dividing the watershed into <br />subareas whose soil characteristics match one of the defined "subunits", <br />and routing the appropriate runoff response (adjusted for area) through <br />the entire watershed. <br />The modeling of watershed hydrologic response with either <br />continuous or non-continuous physically-based rainfall-runoff models is <br />indeed a tremendous advancement in the field of hydrology. Modeling <br />should not, however, be viewed as the hydrologist's panacea. Data <br />generated from simulation models can only be as accurate as the input <br />data available. That is, the generation of extended runoff hydrograph <br />