|
<br />however, can layover and even remove vegetation,
<br />thereby decreasing channel roughness and increasing
<br />channel conveyance (Burkham, 1976; Phillips and
<br />Hjalmarson, 1994),
<br />The f1ow,induced changes in vegetation
<br />conditions are commonly assumed to occur before
<br />peak flow, and postflow vegetation conditions are
<br />assumed to reflect the roughness conditions when the
<br />high,water marks were formed. However, for certain
<br />conditions, this may be an erroneous assumption,
<br />Water,level indicators in vegetated channels surveyed
<br />following flows may not actually represent the water,
<br />surface during the time of peak discharge. The force
<br />and power of flow may be large enough to substantially
<br />affect (Jay over or remove) vegetation just before peak
<br />flow and thus cause the water surface to drop
<br />dramatically because of the decrease in flow resistance,
<br />This phenomenon may result in a lower water,surface
<br />elevation during peak discharge than that suggested by
<br />the high, water indicators, This phenomenon may have
<br />occurred for the flow of February 9, 1993, at
<br />Hassayampa River near Arlington. The high,water
<br />indicators may actually have been deposited before and
<br />not during peak discharge, This verification
<br />measurement, therefore, was rated as estimated,
<br />Vegetation documented during and shortly after flow at
<br />the other study sites appeared to have little or no
<br />change compared to preflow conditions.
<br />
<br />Uncertainties of Discharge Measurements
<br />
<br />The accuracy of the verified 11 values directly
<br />depends on the precision of the measured discharge. As
<br />mentioned previously, discharge was determined either
<br />by a current,meter measurement or from a well-defined
<br />stage,discharge relation. Current-meter measurements
<br />made by the USGS are rated as excellent, good, fair, or
<br />poor depending on factors that include the number of
<br />subsections in the measurement, stability of the
<br />channel, and accuracy of the equipment (Rantz and
<br />others, 1982). These ratings correspond to possible
<br />errors of less than 2, 5, 8, or greater than 8 percent of
<br />the actual discharge, respectively. Errors for discharges
<br />detennined from a well-defined and stable stage,
<br />discharge relation are assumed to be less than
<br />10 percent. Discharge accuracy is considered when
<br />determining the overall accuracy rating for each of the
<br />n-verification measurements.
<br />
<br />Flow Depth and Magnitude
<br />
<br />In the absence of bank vegetation and other
<br />obstructions, the roughness for low flows in a uniform
<br />grave],bed stream generally decreases with increasing
<br />depth of flow, As flow approaches bank-full stage,
<br />however, roughness may asymptotically approach a
<br />constant value, as shown by several previous
<br />investigations (Limerinos, 1970; Bray, 1979; Sargent,
<br />1979; Griffiths, 198]; Jarrett, 1985; Blodgett, 1986;
<br />Hicks and Mason, 1991; Coon, 1995),
<br />The basic roughness coefficient for these streams
<br />should not vary greatly with depth of flow if the
<br />relative roughness is greater than about 5 (Benson and
<br />Dalrymple, 1967). Many of the verification
<br />measurements presented in this report, however, have
<br />values of relative roughness that are less than or close
<br />to 5, and the variation in Manning's roughness
<br />coefficient with depth is apparent.
<br />For many previously published l1-verification
<br />manuals and for this report, verified 11 values are
<br />obtained from flow discharge data that may not result
<br />in a reliable value for studies requiring estimates of
<br />roughness coefficients for design purposes, Design
<br />discharge typically is determined on the basis of the
<br />estimated flood having a particular recurrence interval
<br />(lOO-year flood, for example), Roughness-coefficient
<br />verification studies, however, generally are limited to
<br />flows that may not exceed even the 5'year flood (Wahl,
<br />] 994), This limitation is caused by the relatively short
<br />duration of l1,value studies and the difficult logistics
<br />involved in making l1,verification measurements
<br />during f1oodflows. Roughness coefficients can be
<br />extrapolated to the design discharges using relations
<br />between hydraulic components, such as R and 11;
<br />however, large uncertainties may be associated with
<br />these extrapolations (Wahl, 1994). Water,resource
<br />managers and engineers need to be aware of these
<br />limitations when using this and other l1,verification
<br />manuals.
<br />For sand,dominated streams, the amount of
<br />variance in Manning's 11 with depth is much more
<br />difficult to describe and quantify compared to gravel-
<br />bed streams, Roughness in sand,dominated streams
<br />depends not only on grain size but also on flow regime
<br />and type of bedform manifested. As indicated by the
<br />verification measurements made at the Hassayampa
<br />River near Morristown, the relation of Manning's 11 to
<br />flow depth (hydraulic radius) can actually be the
<br />inverse of the relation for gravel-bed streams (fig. 9C,
<br />see p, 37),
<br />
<br />8 Verification of Roughness Coefficients for Selected Natural and Constructed Stream Channels in Arizona
<br />
|