Laserfiche WebLink
<br />Calculated 100-year-storm intensity-duration curves based on the Gumbel, <br />log-Gumbel, and log-Pearson Type III distributions are also plotted on <br />figure 5. For comparison purposes, the observed intensity-duration curve <br />for the storm of August I, 1985, is plotted on the same figure. The same <br />data are presented in tabular form in table 3. Intensities for selected <br />durations for storms of other recurrence intervals are presented in table 4. <br /> <br />Results <br /> <br />The intens ity-durat ion curves (f ig. 5) based only on data collected <br />through 1984 approximately overlie each other, and, as a group, lie below <br />the curves based on data collected through 1986 and on the method of Steel <br />and McGhee (1979). The curves representing the log-Gumbel distribution lie <br />above the other curves; however, the log-Gumbel distribution probably is <br />overpredicting the intensity of precipitation because the skewness of the <br />actual data (table 2) is different than the skewness of 1.14 assumed in the <br />Gumbel distribution. Similarly, the arithmetic Gumbel distribution may be <br />slightly underpredicting the intensity of precipitation. The log-Pearson <br />Type III distribution may better fit this data because only the log-Pearson <br />distribution incorporates the value of skewness of the observed data in <br />calculating the fitted curve. Although the calculated value of skewness may <br />be affected by sampling error, this effect should be minimized in the case <br />of the Cheyenne data due to the long period of record. <br /> <br />The curve for the storm of August I, 1985, lies above all the curves <br />plotted on the figure for the 5-minute and I-hour durations except those <br />representing the log-Gumbel distributions, and the. curve for the storm of <br />August 1 lies above all the other curves for the 2-hour duration, indicating <br />that this storm was unusually intense for these durations. <br /> <br />HYETOGRAPH SHAPE AND PATTERN INDEX <br /> <br />In order for runoff to occur, precipitation rate must exceed <br />infiltration rate. Only the amount of precipitation in excess of the <br />infiltration rate (infiltration rate is zero for fully saturated ground) is <br />available to run off. If a period of intense precipitation occurs when the <br />ground is dry (as at the beginning of a storm), some of the precipitation <br />will infiltrate into the soil, and only the remainder will run off. <br />However, when the ground is saturated, virtually all of the precipitation <br />will be available to run off. If precipitation occurs when the ground is <br />dry, both a smaller volume of runoff and a smaller peak flow will result <br />than if precipitation of the same intensity and duration occurs when the <br />ground is saturated. <br /> <br />Calculation of pattern index (Hasfurther and Tyrrell, 1984) is one <br />technique for examining storm hyetographs to determine whether the period of <br />maximum precipitation intensity occurred early or late during the storm. <br />Dimensionless hyetographs (cumulative precipitation versus time) are drawn <br />for each storm, and the area under the curve is calculated either by <br />graphical methods or numerical integration. This area, greater than zero <br />but less than one, is the pattern index. A pattern index less than 0.5 <br />indicates that the major part of the precipitation occurred after the <br /> <br />38 <br />