<br />2964
<br />
<br />JARRETT AND TOMLINSON: REGIONAL INTERDISCIPLINARY PALEOFLOOD METHOD
<br />
<br />. ,
<br />in steep mountain basins [Boner and Stemtitz, 1967; Snipes et'
<br />af., 1974; Schwarz el af., 1975; McCain el al., 1979; Cosla, 1983;
<br />Jarrell and Cosla, 1986; Jarrett, 1987, 199Ob, 1991; Grimm el al.,
<br />1995; Waylhomas and Jarrett, 1994] is unequivocal (e.g., Figures
<br />3-5). Paleoflood evidence is relatively easy to recognize and
<br />long lasting because of the quantity, morphology, and structure
<br />and size of sediments deposited by floods. Lack cf movement
<br />of all sizes of clasts on a streambed with an unlimited size of
<br />clasts is a direct measure of flow competence [Costa, 1983;
<br />Komar, 1987; Bull, 1990; Wilcock, 1992] and an indirect mea-
<br />sure of flood discharge. In many channels in Colorado, there is
<br />an "unlimited" size of clasts present in channels because of
<br />past glacial outwash or rockfall that are available for transport.
<br />Point, longitudinal, and transverse bars are built up of layers of
<br />silt, sand, gravel, cobble, and boulders from bed load to about
<br />the height of maximum flood water [Jarrett el af., 1996J. For a
<br />given reach of channel the relation of maximum clast size in a
<br />flood deposit to the available clast size in the upstream channel
<br />was noted. Clasts show little reworking on the stream bed
<br />(random pattern due to winnowing of finer ma1:erials) when
<br />floods are small; however, large floods tend to produce well-
<br />developed, fluvial-depositional features. In adjition, when
<br />flows exceed the main channel by even small dep:hs (<0.5 m),
<br />coarser-grained bed material is transported onto the floodplain
<br />and deposited [Jarrett el ai., 1996J such as shown in Figure 3;
<br />thus lack of coarse material on floodplain surfaces indicated
<br />minimal inundation.
<br />There are three main sources of uncertainty in paleoflood
<br />reconstructions [Jarrett and Maide, 1987]: selection of flow-
<br />resistance coefficients, channel changes, and representative-
<br />ness of PSIs of flood height Selecting flow-resistance coeffi-
<br />cients, whether Manning's n value as in this s':udy or other
<br />resistance forms (e.g.. Darcy's or Chezy's), can be problematic
<br />for flood studies, particularly for paleoflood e~ timates when
<br />riparian vegetation may have been different and unquantifi-
<br />able, Jarrett [1986] and TrieSIe alld Jarrett [198~'] indicate for
<br />most higher-gradient streams and many lower~gradient rivers,
<br />flow is about critical or slightly subcritical even in channels
<br />having substantial channel and floodplain vegetation prior to a
<br />flood, Large floods tend to remoVe most vegetation prior to the
<br />peak stage [Matthai, 1969; McCain el ai., 1979; Phillips and
<br />Ingersoll, 1998], and because flow is nearly critical, some of the
<br />uncertainty in estimating n values due to unquantifiable past
<br />vegetation is removed.
<br />To minimize effects due to channel changes, the second
<br />sourCe of uncertainty, cross sections were loca:ed in bedrock
<br />reaches or in relatively stable alluvial reaches, !'>Jthough most
<br />of the Yampa River is alluvial, there are good constraints on
<br />relative stability or slight rates of incision during the Holocene
<br />[Madole, 1991a]. For alluvial tributary streams that may have
<br />undergone cyclical aggradation and degradation or if these
<br />cycles were difficult to ascertain, then very restrictive (conser-
<br />vative) ages were assigned to the paleoflood estimate, usually
<br />less than 100 years.
<br />The third source of uncertainty in paleoflood reconstruc-
<br />tions is maximum flood height inferred from PSIs. In most
<br />paleoflood studies, PSIs have been assumed to be slightly lower
<br />than maximum flood height [Kochel and Baker, 1982; NRC,
<br />1988; Baker et ai., 1988], New research on the ,levation of the
<br />top of flood-deposited sediment (new PSIs) and high-water
<br />marks (HWMs) of recent flooding in 90 streams primarily in
<br />the western United States that has been dore [JalTett et ai.,
<br />1996J challenges this assumption. HWMs are the evidence of
<br />
<br />the highest stage reached by a flood and primarily consist of
<br />fine woody debris, leaves, grass, needles, other floatable mate-
<br />rials and mud lines. These streams have drainage areas that
<br />range from about 0.004 lan2 to more than 6000 lan'. Peak
<br />discharge ranged from about 0.03 to 2500 m3 s", and the
<br />majority .were iarger than 100-year floods. Stream gradient at
<br />these sites ranges from about 0.0007 to 0.35 m m ". Tbe size of
<br />flood~deposited sediments ranged from silt to boulders more
<br />than 4 m in diameter; only those deposits considered to have
<br />long-term preservation potential were documented. Analysis
<br />of the differences in PSIs and HWMs indicates that the eleva-
<br />tions of tbe top of flood-deposited sediments (PSIs) generally
<br />are within ::!:O.2 m of flood HWM elevations, Therefore use of
<br />the top of flood-deposited sediments as PSIs for streams in this
<br />study provides a reliable estimate of the maximum paleoflood
<br />depth that is used to reconstruct tbe discharge of paleofloods.
<br />Good HWMs for the 1995 near-record peak flows during
<br />fieldwork were established, which were not that much smaller
<br />than maximum paleofloods. Paleoflood techniques were used
<br />to estimate peak discharge for the 1995 flood where stream-
<br />flow-gaging station estimates were available but without prior
<br />discharge knowledge, Although this comparative approach
<br />does not consider all the uncertainties, the comparison does
<br />provide an estimate of the uncertainty of the critical depth and
<br />slope-conveyance methods for estimating peak discharge in the
<br />study.
<br />
<br />4.2. Paleoflood Chronology
<br />
<br />In this study, a variety of relative-dating (RD) techniques
<br />were used to estimate the approximate length of time (age)
<br />corresponding to the highest flood deposits emplaced during
<br />the Holocene for subsequent use in flood-frequency analysis,
<br />RD techniques based on landform modification, rock-
<br />weathering features, and soil development have long been used
<br />to differentiate and map Quaternary deposits in the western
<br />United States with emphasis on dating glacial deposits [Birke-
<br />iand el ai., 1979; Colman and Pierce, 1983; Birkeland, 1990].
<br />Rock-weathering, soil, and geomorphic parameters all change
<br />with time [Coiman and Pierce, 1983; Birkeland el ai., 1979J.
<br />Although mucb dating has involved fluvial deposits, ages pri-
<br />marily are determined with absolute-dating methods such as
<br />14C, thermoluminesce, and dendrochronology [Kochel and
<br />Baker, 1982; Hupp, 1988; Coiman and Pierce, 1991J. Some
<br />investigators have used a variety of RD methods, which utilize
<br />the collective strengths of each method [Burke and Birkeland,
<br />1979; Harden, 1986, 1990; Waythomas and Jarrett, 1994; Mills
<br />and Allison, 1995J. The chances of arriving at a valid approxi-
<br />mate age is greatly enhanced if a variety of RD methods are
<br />measured [Birkeland et ai., 1979]. '
<br />RD methods appiied to surficial deposits are based on post-
<br />depositional modifications that vary with age [Birkeland el aI.,
<br />1979J. Field evidence of age is usually derived from soil prop-
<br />erties, rock-weathering characteristics, changes in landform
<br />morphology, and lichenometry. Their usefulness for character-
<br />izing surficial deposits is related to the degree to which they
<br />can be quantified and to their rate of change [Birkeland et ai.,
<br />1979]. Episodic flooding produces deposits of different ages
<br />that can be separated by long periods; thus the deposits have
<br />distinctive properties that are amenable to measurement by
<br />RD methods. It is assumed that when flood~deposited sedi.
<br />ments undergo transport and reworking during high~energy
<br />flood transport, the weathering "clock" essentially is reset to
<br />zero [Waythomas and Jarrett, 1994). Certainly, a limitation is
<br />
<br />
<br />."
<br />
<br />l
<br />
<br />,
<br />~
<br />~.
<br />.
<br />
<br />I
<br />,
<br />
<br />;;
<br />'.
<br />,
<br />
<br />l
<br />:1
<br />i.
<br />i,
<br />t
<br />~
<br />,
<br />~
<br />"
<br />
|