|
<br />Doesken (Colorado Climate Center, personal commun., 1995) indicated that occasionally when
<br />precipitation is measured in a standard 8-inch diameter National Weather Service (NWS) raingage, the
<br />amount of rainfall was inadvertently measured with a nonnal ruler (l-inch depth = l-inch of rainfall) rather
<br />than with a special ruler (l-inch depth = O.l-inch of rainfall) for use with the 8-inch precipitation gage that is
<br />used to increase measurement accuracy. However, if a nonnal ruler is used, a 1.D-inch rainfall would
<br />incorrectly be measured and reported as 10 inches (or a 10:1 error). For example, Jarrett (1988)
<br />conducted an onsite paleoflood investigation and reviewed hydrometeorologic data for the reported
<br />rainstonns of 11.4 inches on July 31 and 9.6 inches at August 18, 1961, in Mayflower Gulch near Frisco,
<br />Colorado (Curry, 1966). The elevation of the Mayflower raingage was 11,600 feet Rainstonns of this
<br />magnitude (11.4 and 9.6 inches in 24 hours) are unprecedented at higher elevations of the Rockies (even
<br />one such rainstonn). Nearby rainfall gages had an order-of-magnitude less rainfall on the same days,
<br />there was no substantial rainfall runoff at downstream streamflow-gaging stations, and there was no
<br />paleoflOod evidence in Mayflower Gulch or in any other nearby streams (Jarrett, 1988). Thus, he
<br />concluded that the large rainstonns in 1961 were erroneously measured, most likely due to the order of
<br />magnitude (10:1) error.
<br />
<br />The 5.25 inches of rain reported at Cimarron may be another example of a 10: 1 recording error. I n this
<br />case, a reported 5.25-inch rain actually would have been 0.52 inches. A value of 0.52 inches at Cimarron
<br />on June 3rd would be more consistent with nearby rainfall observations (figure 2). By using the simplified
<br />rainfall-runoff relation for Colorado (figure 10), and for a rainfall of 0.5 inches the peak discharge would
<br />range from about 150 to 250 cubic feet per second. Because the Cimarron rainstonn did not cover much
<br />area, a runoff value lower than estimated using figure 10 would have occurred. Therefore, this review also
<br />supports that the June 3rd rainfall probably was less than one inch.
<br />
<br />Over the years, the accuracy and validity of dozens of other large recorded rainstonnslfloods and
<br />anecdotal reports of large rainfall amounts throughout the Rocky Mountain region have been reviewed for
<br />their accuracy (eg., Jarrett, 1987b, 1988, 1990; Costa and Jarrett, 1981). A review of available rainfall
<br />and streamflow data and lack of onsite paleoflood evidence in streams subject to these rainstonns has
<br />identified numerous questionable or erroneous extreme rainstonn records, which resulted from a variety of
<br />reasons. Historical-flood estimates made in cobble and boulder-bed streams have been shown be
<br />overestimated by an average of 50 to 60 percent (Jarrett, 1986, 1987a, 1987b, 1994). The main reason
<br />for the overestimation is that selected Manning's n-values typically ranged from 0.025 to 0.035 for streams
<br />with boulders up to 3 feet in diameter (Jarrett, 1994). With today's understanding of flow-resistance
<br />coefficients in cobble and boulder-bed streams (Bames, 1967; Umerinos, 1970; Aldridge and Garrett,
<br />1973; Jarrett, 1984, 1985, 1986, 1994; Trieste and Jarrett, 1987; Hicks and Mason, 1993), flow-resistance
<br />coefficients for these types of streams are about 1.5 to 2 times larger than the historical estimated values.
<br />The methods and flow-resistance coefficients used to estimate the discharge of historic floods in Colorado,
<br />which used the best understanding available at the time, were used for mountain rivers throughout United
<br />States (Murphy and others, 1904; Follansbee and Jones, 1922; Follansbee and HodQes, 1925;
<br />Follansbee and Sawyer, 1948). Logically, those flood estimates also are highly questionable and
<br />probably overestimated (Jarrett, 1986, 1987b, 1994).
<br />
<br />Because of questionable flood-discharge data and their effects on design flood hydrology, floodplain
<br />management, and related environmental studies, Jarrett (1994) and Jarrett and Petsch, (1985) proposed a
<br />critical need exists for interdisciplinary documentation of extreme-flood processes for future flooding. To
<br />best implement advances in improving the understanding of flood processes, there is a need to focus
<br />more efforts and resources on completely documenting all processes for individual floods rather than
<br />studying individual components of many separate floods. These processes indude meteorologic,
<br />hydrologic, hydraulic, sediment transport, fluvial-geomorphologic, and water-chemistry processes, and
<br />their interrelations. These research needs are based on an integrated, river-system-process approach
<br />that utilizes various disciplines, better communication among water scientists, engineers, planners, and
<br />managers, and greater emphasis on societal needs to mitigate flood losses. Such a research program will
<br />benefit the public through improved engineering designs, flood-plain management, and environmental
<br />investigation that require good data and methodologies to describe flooding.
<br />
<br />Another example of questionable extreme-rainfall amount is the 1935 rainstorm that occurred on the
<br />Palmer Divide, which had centers at the headwaters of Cherry Creek basin and another center near Hale,
<br />Colorado. The Palmer Divide, which fonns the boundary between the Arkansas and South Platte River
<br />basins, is suspected to produce orographically increased precipitation. A rainfall amount of 24 inches in
<br />
<br />9
<br />
|