Laserfiche WebLink
<br />One method to reduce prediction uncertainty when extending the systematic record is through <br />incorporating historic data into the frequency analysis. Historic data consists of flood data that <br />pre- and/or post-dates the systematic record. Evidence of historic flood levels is preserved in the <br />sediment deposits and damage left behind by the flood as well as documented water levels by <br />residents and victims of flooding. The study of this data is called paleohydrology and can give <br />reliable estimates of the magnitude of past flood levels. <br /> <br />Another method of improving the frequency statistics is by using a calibrated rainfall-runoff <br />model. This method is especially useful and important as many drainage and flood control <br />structures are designed by running a design storm through a calibrated model to obtain an <br />estimate of the associated runoff. The flexibility exists in modeling to simulate both pre- and <br />post-development conditions and how these changes affect flood frequency. It is generally <br />assumed that a precipitation event of a given frequency will produce a runoff event with the <br />same frequency (Urbonas, 1979). With this assumption, it is possible to simulate several runoff <br />events with storms of various recurrence intervals to construct a flood frequency curve. This <br />method will be discussed in more detail in subsequent sections. <br /> <br />This paper investigates the flood hydrology of three sites in the northeastern plains region of <br />Colorado. Each of these sites provide different challenges in attempting to construct the flood <br />frequency relationship. This paper examines the application of standard flood frequenl:Y <br />methods and illustrates some methods used to accommodate these data difficulties. Floods <br />generated by rainfall are the only ones considered in this study. Two rainfall-runoff models are <br />used to develop the flood frequency relationship and compared against the frequency curve <br />derived from the systematic and paleoflood data. This illustrates the validity of using rainfall- <br />runoff modeling to estimate the flood frequency relationship. The resulting calibrated models <br />may be used to extrapolate the flood frequency curve for use in design. <br /> <br />2. Site Description <br /> <br />The study area resides in the northern portion of Weld County, Colorado. The USGS operated a <br />network of 5-minute rainfall-runoff gages in this area between 1969 and 1979 to support flood <br />frequency analysis studies (Cochran, et aI, 1983). Three sites within this network were chosen <br />for this study and are summarized in Table I. 'Their location within Colorado is shown in <br />Figure I. ~) <br /> <br />The three sites drain the southern slopes of the Chalk Bluffs egion and lie within the Pawnee <br />National Grassland. The bluffs are confined to the extreme northern portion of the study area <br />and consist of mesa formations with maximum elevations near 6,200 feet. The remainder of the <br />area is dominated by rolling prairie. The prairie portions are covered with short, native prairie <br />grasses and yucca. Tree outcrops are sparse and confined to areas where water is readily <br />accessible. The grassland is a conglomerate of federal and privately owned land. Much of the <br />region is open to cattle grazing. <br /> <br />~ <br /> <br />2 <br />