|
<br />
<br />2958
<br />
<br />JARRElT AND TOMLINSON: REGIONAL INTERDISCIPLINARY PALEOFLOOD METHOD
<br />
<br />~
<br />:".
<br />
<br />The NRC stresses that this should be the basis for evaluaiion
<br />of the risk and consequences of any decision.
<br />Efforts have been made to assign a frequency to the PMF,
<br />but they are subjective [Interagency Committee on Water Data,
<br />1986; NRC, 1988J. There is disagreement about the range of
<br />frequency for the PMF and whether any frequency should be
<br />assigned to a deterministic PMF estimate. Extensions of flood-
<br />frequency relations to rare floods (e.g., 10,000-year recurrence
<br />interval) are tenuous for short streamft.ow records but are
<br />enhanced when paleoflood data are included in flood-
<br />frequency anaiysis [Kochel and Baker, 1982; NRC, 1988; Jarrett,
<br />1987; Jarrett and Costa, 1988; Levish et al., 1994; Ostenaa and
<br />Levish, 1995J. Jarrett and Cosla [1988] proposed using regional
<br />flood-frequency estimates, which incorporated paleoflood in-
<br />fonnation, to estimate recurrence intervals for rare floods in.
<br />cluding PMF values. These extensions provide an approach to
<br />place PMF estimates in perspective with regional gaged and
<br />paleoflood estimates of flood potential [Jarrett and Costa,
<br />1988J.
<br />Paleoflood hydrology is the study of past or ancient floods
<br />[Baker, 1987]. Kochel and Baker [1982], Gregory [1983J, Bakeret
<br />al. (1988], Costa [1987c], Sledinger and Baker [1987J, Stedinger
<br />and Cohn [1986J, Hupp [1988J, and Jarrett [1987, 1990b, 1991]
<br />provide summaries of paleoflood hydrology. Although most
<br />studies involve prehistoric floods, the methodology is applica-
<br />ble to historic or modern floods [Baker et aI., 1988; Jarrett,
<br />1990bJ. Paleoflood studies provide important information that
<br />can be used in risk assessments and in the assessment of cli-
<br />matic change on flooding and droughts [Jarrett, 1991]. Paleo-
<br />flood data are particularly useful in providing upper limits of
<br />the largest floods that have occurred in a river basin in long
<br />time spans [Jarrett, 1990b; Enzel el aI., 1993].
<br />A regional interdisciplinary paieoflood study was conducted
<br />in northwestern Colorado (Figure 1) to help assess the flood
<br />hydrology for Elkhead Reservoir in Elkhead Creek basin near
<br />Craig. The objective of the paleoflood study was to estimate
<br />prior maximum flooding during the Holocene from evidence
<br />preserved in the floodplain. The interdisciplinary components
<br />included documenting maximum paleofloods and regional
<br />analyses of contemporary (-155 years in Colorado) extreme
<br />rainfall and flood data in and near Elkhead basin. The major
<br />drainages within the regional study area are the Yampa River
<br />and White River basins. Hydroclimatic conditions are rela-
<br />tively bomogeneous in northwestern Colorado [Miller el al.,
<br />1973; Kircher et aI., 1985J. A primary focus of U.S. Geoiogical
<br />Survey (USGS) interdisciplinary research is to develop cost-
<br />effective paleoflood techniques that can be used to comple-
<br />ment meteorologic, hydrologic, and engineering methods to
<br />improve estimation of the magnitude, frequency, and risk of
<br />floods.
<br />The paleoflood study was conducted for the Colorado River
<br />Water ConseIVation District to complement a site-specific
<br />PMP by Tomlinson and Solak [1997] and a PMF study by Ayres
<br />Associates, Inc. in Fort Collins, Colorado, for Elkhead Reser-
<br />voir. Elkhead Reservoir was being recertified for hydrologic
<br />safety by the Colorado State Engineer. PMP estimates are
<br />considered of lesser reliability along the Continental Divide,
<br />which includes the upper Yampa River basin [Hansen el al.,
<br />1977J. Therefore a site-specific PMP study was conducted to
<br />address issues raised by the NRC [1988, 1994] pertaining to the
<br />hydrometeorology for the basin and the surrounding geo-
<br />graphically and climatologically similar region. Inherent in a
<br />site-specific PMP study are analyses of extreme storms that
<br />
<br />have occurred in the region since the generalized hydrometeo~
<br />rology report was published. Site-specific hydrometeorologic
<br />studies are being conducted because dam safety officials rec-
<br />ognize the difficuit problems inherent in PMP estimates in the
<br />Rocky. Mountains. Utilization of an interdiscipiinary regional
<br />paleoflood study provides additional supporting information
<br />for understanding the magnitude of the iargest contemporary
<br />floods and paleofloods with estimates of the PMF potential for
<br />a particular basin.
<br />
<br />~~
<br />
<br />2. Background
<br />
<br />Substantial uncertainty and controversy exists in estimating
<br />flood magnitudes and frequencies, particularly those or ex-
<br />treme floods in the Rocky Mountains. This resuits from a
<br />misunderstanding of the complex hydrometeorological pro-
<br />cesses involved and a lack of data on extreme rainstorms and
<br />flooding. Jarrett (1993] made a systematic evaluation of flood-
<br />frequency estimates for 25 long-term streamflow-gaging sta-
<br />tions throughout Coiorado where published Federal Emer-
<br />gency Management Agency (FEMA) floodpiain reports, which
<br />were derived by various methods, also were available. On av-
<br />erage, the published FEMA 100-year flood is 35% greater than
<br />the gaged lOO-year flood obtained by fitling a log-Pearson type
<br />III distribution to streamflow data for the 25 stations. Similar
<br />differences were identified for the 10-, 50-, and 500-year floods
<br />and demonstrate the need to improve flood-frequency estimates
<br />used for floodplain management and other uses in Colorado.
<br />Interdisciplinary flood research in Colorado [Jarrett and
<br />Costa, 1983, 1988; Jarrett, 1987, 1990a, b; Grimm, 1993; Pitlick,
<br />1994; Waythomas and Jarrett, 1994; Pruess, 1996; Capesius,
<br />1996; Pruess er al.. 1998; McKee and Doesken, 1997] and in the
<br />Rocky Mountains [Henz, 1991; Jarrett, 1993; Jensen, 1995;
<br />Buckley, 1995; Eastwood, 1995; Brien, 1996; Parrett, 1997, 1998J
<br />provides new insight into the hydrometeorology of extreme
<br />flooding in the Rocky Mountains. In an analysis of USGS
<br />streamflow-gaging station and paleoflood data in the Rocky
<br />Mountain region, which included stations in New Mexico, Col-
<br />orado, Wyoming, Idaho, and Montana, envelope curves of
<br />maximum unit discharges (maximum peak flow divided by
<br />drainage area) and gage elevation identified elevation limits
<br />for large unit discharges [Jarrett, 1993]. In Colorado the eleva-
<br />tion limit is about 2300 m for an envelope curve value of 1 m3
<br />S-1 km-2 for basins less than about 10 km2; as basin size
<br />increases, unit discharge decreases. Such low-magnitude unit
<br />discharges result in only minor flooding in Colorado; flows
<br />seldom exceed the top of the main-channei banks, and when
<br />they do, flow depths usually are insuffipient to modify the
<br />floodplain. These peak discharges are caused prim8rily by
<br />snowmelt runoff, relatively small amounts of rainfall (as com-
<br />pared to lower-elevation rainfall amounts). or a combination
<br />of rainfall on snowmelt. Below about 2300 m in eastern Col-
<br />orado, unit discharges as large as 38 m3 S-1 km-2 have oc-
<br />curred. Above about 2400 m in Coiorado, maximum observed
<br />6-hour rainfall is about 100 mm; in eastern Colorado at lower
<br />elevations, maximum observed 6-hour rainfall is about 610
<br />mm, which also is the maximum 24-hour value [Hansen et al.,
<br />1978].
<br />The interdisciplinary research cited above has provided de-
<br />finitive information that substantial flooding in Colorado has
<br />not been observed above an elevation of about 2300 m [Jarrett,
<br />1987, 1990b, 1993; Jarrett arid Costa, 1983, 1988; Grimm et aI.,
<br />1995; Pruess, 1996]. The limit varies somewhat because of
<br />
|