Laserfiche WebLink
<br />therefore, the data required to develop unit hydro graphs in the above manner seldom <br />exists. However, if individual bursts of rain in the storm result in well-defined peaks, <br />it is sometimes possible to separate the hydrographs produced for the various bursts <br />by estimating the recession of runoff from each burst. These hydrographs may then <br />be used as runoff from independent storms for the development of unit hydrographs <br />in the manner described above. <br /> <br />The result of the above derivation, or where unit hydrographs from <br />different storms at the same location are derived, is a series of unit hydrographs with <br />differences due to observation errors and other factors as described previously. For <br />the convenience of dealing with only one unit hydrograph, and to minimize any errors <br />due to separation of the hydrographs in the above procedure, the unit hydrographs are <br />normally averaged. All unit hydrographs to be averaged must have the same unit <br />duration, which may require converting to a common unit duration. The averaging is <br />usually done graphically to prevent reducing the peak incorrectly which is the likely <br />result when an arithmetic average is applied to concurrent ordinates. The <br />instantaneous peak flows should be averaged, regardless of differences in lag time and <br />plotted at the average lag time. The average unit hydrograph can then be sketched <br />to conform to the shape of the graphs, passing through the computed average peak, <br />and having a volume of 1 unit of runoff. <br /> <br />By inspection of the heaviest rainfall and snowmelt excesses that caused <br />the peak flows of the flood hydrograph, unit hydrograph ordinates around the peak are <br />first estimated. The remainder of unit hydrograph is then estimated so that the <br />volume is equal to 1 unit of runoff. In estimating the overall shape of the unit <br />hydrograph, consideration is given to the rate of rise and recession of the flood <br />hydrograph and the estimated lag time from the heaviest excess period to the peak of <br />the flood hydrograph. This first-trial unit hydrograph is then applied to the excess <br />estimates, and the resulting computed hydrograph is checked against the observed <br />flood hydrograph. The unit hydrograph is then adjusted as needed, and the process <br />is repeated until the computed hydrograph approximates the observed hydrograph to <br />the desired accuracy. <br /> <br />The computation of a hydrograph requires not only developing a unit <br />hydrograph but estimating rainfall and snowmelt losses as well, and ordinarily the <br />analysis involves both procedures at the same time. This is normally required since <br />it is not always obvious whether the unit hydrograph or the loss rates should be <br />adjusted to obtain a better fit between the computed and observed hydrographs. <br /> <br />The reconstitution of historical flood hydrographs involves the estimation <br />of base flow, loss rates, and unit hydrographs; and it becomes clear that many <br />combinations of these three items can reconstitute the same flood hydrograph equally <br />as well. Consequently, a considerable amount of engineering judgment will be <br />required to establish reasonable estimates for each of these items. Some subjectivity <br /> <br />7-27 <br />