|
<br />rock surfaces [White, 1971J, take about 50 years to become
<br />established. Environmental factors known to affect the growth
<br />of R. geographicum include rock type, shading, temperature,
<br />moisture, and stability of the substrate, and thus they need to
<br />be fac.fored into age assignments. Growth increases with
<br />coarser texture, moisture, temperature. Abrasion during sedi-
<br />ment transport in higher-energy streams common to floods in
<br />mountain regions removes most lichen thus essentially reset-
<br />ting the time "clock" to zero for lichen growth. Benedict [1967,
<br />1968J indicated that growth curves are fairly constant between
<br />3125 m and 4047 m (corresponding to a mean air temperature
<br />change of about 6.60C) for his sites along the crest of the
<br />Continental Divide in Colorado. Benedict's [1967, 1968J sites
<br />are located about 80 Ian southeast of this study area. Lichen
<br />growth curves have a maximum age utility of about 3000 years
<br />in Colorado [Benedict, 1967, 1968J, probably a shorter time at
<br />lower elevations in Colorado where climate is more conducive
<br />to faster growth rates. Maximum thalli diameter and percent
<br />lichen on similar rock types for 25-50 clasts on the flood
<br />deposit and other rock surfaces were made at each site.
<br />
<br />
<br />2966
<br />
<br />JARRETT AND TOMLINSON: REGIONAL INTERDISCIPLINARY PALEOFLOOD METIJOD
<br />
<br />their relative age. Older deposits have extensive surface pit-
<br />ting; rougher surfaces, and increasing grain relief because of
<br />differential weathering of minerals. As less resistant minerals
<br />decompose, quartz and feldspar grains tend to stand out in
<br />relief. Pitting is common if present on more th.:m 75% of the
<br />clast surfaces and is rare (or incipient) if present on less than
<br />10% of the clast surfaces. Flood-deposited clasts are compared
<br />with end-member clasts from the streambed (fr'~sh) and clasts
<br />much higher On the land surface (extremely weathered).
<br />4.2.3. Boulder burial. Many recent flood deposits in
<br />Colorado (e.g., Figure 3) and other mountain rivers consist of
<br />little or modest amounts of matrix-supported cobble and boul-
<br />dery deposits, particularly around surface clasts [e.g., Matthai,
<br />1969; McCain et ai., 1979; Costa, 1983; Waythomas and Jarrett,
<br />1994; Jarrett el ai., 1996J. Boulder burial refers to the percent-
<br />age of the total boulder surface exposed above ground. For
<br />flood deposits the amount of cobble and houlder burial by
<br />addition of newer sediment at the site (colluvium, eolian, and
<br />slope wash) also can be used to estimate the relative age of a
<br />deposit [Waythomas and Jarrett, 1994J. Thus tirr.e or age since
<br />flood depositinn is inferred from depth ofburia!. The older the
<br />deposit is, the greater is the percentage of these clasts that are
<br />covered by postflood deposition.
<br />4.2.4. Surface morphology. Formation of stream terraces
<br />involves changes in the behavior of a fluvial system [Bull, 1990].
<br />Terraces may form because of a variety of intenal or external
<br />changes, climate, tectonics, base level, slope, complex re.
<br />sponsc, and thresholds [Patlan and Schumm, 19i5, 1981; Wom-
<br />ack and Schumm, 1977; Bull, 1990). Remnants of the former
<br />streambed are preserved as terrace treads. Terrace features
<br />such as riser angle become muted with time by faunal, water,
<br />and wind action, and rates of change depend On factors such as
<br />cohesion of material, vegetation, and flow stress [Lewin, 1978;
<br />Birkeland et al., 1979]. Younger terraces tend ta be more an-
<br />gular and have steep slopes; with increasing age, terrace scarp
<br />slopes become flatter unless maintained by cut-bank erosion.
<br />Along rivers in glaciated basins in the Yamp 1 River basin,
<br />early Holocene to late Pleistocene terraces, which are 1-2 In
<br />above the present floodplain, are covered with eolian (loess)
<br />deposits [Madole, 1991aJ. These surfaces are relatively easily
<br />erodible if flooded thus providing unique sites for paleoflood
<br />investigations. Over time, local hillslope runoff produces trans-
<br />verse gullies or channels on terraces and colluvial surfaces;
<br />greater development generally requires longer time. Also, if
<br />alluvial (or colluvial) surfaces are inundated by flood waters,
<br />microchannels and surface deposits are prodt: ced longitudi.
<br />nally, which are somewhat similar to crevasse channels and
<br />splays [Lewin, 1978J. With time, geomorphic expression of
<br />these features is muted. Lack of such flood features or noni-
<br />nundation surfaces [Jarrett and Costa, 1988; Levish el al., 1994;
<br />Ostenaa attd Levish, 1995J provides an upper bound of flood
<br />height on a surface of known age.
<br />4.2.5. Lichenometry. A common RD technique used for
<br />dating glacial deposits is lichenometry [Benedict, 1966, 1967,
<br />1968; While, 1971; Beschel, 1973J. Its use has Jrimarily been
<br />high-elevation or arctic climates, and the transf'~r value is lim-
<br />ited, particularly for growth curves, Rhizocarpon geographicum,
<br />the most commonly used lichen for dating, grows throughout
<br />most of Colorado. Most lichenometric studies use a combina-
<br />tion of maximum thallus diameter and percentage cover on
<br />clast surfaces to determine the ages of late Holocene deposits
<br />[Benedicl. 1967, 1968; Beschel, 1973; Birkeland et al., 1979J.
<br />Lichens, which grow on all but freshly exposed or deposited
<br />
<br />~
<br />W
<br />"
<br />"
<br />;.
<br />
<br />!
<br />
<br />4.3. Regional Analyses of Maximum Rainfall
<br />and Flood Data
<br />
<br />A iack of flood evidence, particularly of extremely rare
<br />floods, in one basin such as Elkhead Creek basin could result
<br />from pure chance. Thus it is essential to ascertain the flood
<br />history for other basins in the region [NRC, 1988]. Regional
<br />analysis extends hydrometeorologic records and provides a
<br />tool to estimate discharge at ungaged sites [JalTett and Costa,
<br />1988; NRC, 1988; Hosking and Wallis, 1998J. In addition, re-
<br />gional analyses provide improved estimates of precipitation
<br />and streamflow characteristics for gaged sites by decreasing
<br />time-sampling errors for relatively independent samples.
<br />Predicting the upper limits to the magnitudes of floods in a
<br />specific tegion has been a long-standing challenge in flood
<br />hydrology. Envelope curves encompassing maximum rainfall
<br />[Linsley et al., 1982; Jarrell, 1987, 1990bJ and floods in a homo-
<br />geneous hydrometeorologic region have long been used in
<br />flood hydrology [Crippen and Bue, 1977; Costa, 1987a; Jarrett,
<br />1987, 1990b; Enzel et al., 1993). Utilization of envelope curves
<br />for a hydro meteorologic region can be evaluated by examining
<br />maximum floods in nearby basins. A premise for envelope
<br />curves is that not all basins in the region are expected to have
<br />had the maximwn flood, but no basin has yet had a flood that
<br />exceeds the envelope curve for the specific region. The primary
<br />limiting factors for extreme floods are amount, intensity, du-
<br />ration, and spatial distribution of rainfall, which includes oro-
<br />graphic enhancement effects and basin slope [Cosra, 1987b;
<br />Pitlick, 1994]. Incorporating paleoflood data for various basins
<br />in the region provides an opportunity to add a new level of
<br />confidence to envelope curves [NRC, 1988; Enzel et al., 1993J.
<br />
<br />4.4. Rainfall Data
<br />
<br />Extreme rainfall data for the last 100 years were compiled
<br />from 181 official precipitation gages and numerous supple-
<br />mental rainfall-bucket surveys in western Colorado [McKee
<br />and Doesken, 1997J. Four, long-term precipitation stations in
<br />the study area, at Steamboat Springs, Hayden, Lay, and
<br />Meeker (Figure 1), have been operated from 61 to 94 years.
<br />Rainfall-hucket survey data, primarily collected by the Na-
<br />tional Weather Service, Army Corps of Engineers, and Bureau
<br />of Reclamation, were compiled for Colorado [Jarrett, 1987,
<br />199Gb] and updated through 1997 for this study. Although very
<br />
<br />-~
<br />
|