FLOOD06061 - Laserfiche WebLink

Home Browse Search

41 A family of depth-duration curves that would cover the range required in the Southwest was then developed. First, a base depth-duration curve was es- tablished using all recorder data for least-orographic stations in the August 1951 and September 1970 storms. These storms are the closest to the proto- type PMP storm for most of the Southwest. Averages of 6/24-, 12/24-, 18/24-, 48/24-, and 72/24-hr ratios are shown by the large dots in figure 2.20. The 72-hr dot is based solely on August 1951 data. A smooth line was drawn through these dots. Next, we expanded this base depth-duration curve to a family of curves constrained by the limits: a. Contant rainfall rate. A straight line from 0 to 100% at 24 hours to 300% at 72 hours. b. All rain in the first instant, or 100% at all durations. These two constraints are represented by the straight lines in figure 2.20. There is great flexibility in how to draw additional curves between these two lines. We selected 6/24-hr ratios at increments of 30, 40,..., 90% and drew smooth curves between 0 and 24 hours that were consistent with the curvature of the basic relation and somewhat symmetrical about a perpendicular bisector to the curves. The 6 additional curves were then extended to 48 and 72 hours as smooth (not necessarily straight) lines. Further adjustments were made to the in- crements between curves beyond 24 hours in order to maintain a gradual in- crease (smooth gradient) in the increment between successive curves as the 72/24-hr ratios increased. The control for this gradient was the range in individual recorder durational curves for the stations used in the August 1951 and September 1970 storms. Although the family of curves in figure 2.20 suggests a broad range of 72/24-hr ratios, a much smaller range is ap- plicable to the Southwest as discussed under seasonal and regional variation& The PMP study for the Northwestern States, HMR No. 43, used a similar gen- eralized set of depth-duration relations for convergence PMP. While not de- veloped in the same manner as in the present study, the results are similar. Adopted smooth relations from the two studies are compared in figure 2.21. 2.4.3. Seasonal variation It is to be expected that the 6/24-hr ratio should have a seasonal varia- tion, i.e., because of greater convective activity ratios should be higher in summer than in winter. In figure 2.19, a check was made of two points (labelled A and B) that ap- pear to be extremes relative to the seasonal distribution of points indicated in this figure. Hourly precipitation records and synoptic weather analyses indicate that point A is the result of 3 days of isolated afternoon thunder- showers. Thus, it is not representative of a general-storm summer rainfall. Point B results from one-day rainfall associated with a rapidly moving and dissipating low-latitude cold front with light post-frontal showers on the