9633

• UCREFRP Catalog Number: 9633
• Author: Schmidt, J. C. and J. B. Box
• Title: Application of a Dynamic Model to Assess Controls on Age-0 Colorado Pikeminnow Distribution in the Middle Green River, Colorado and Utah
• USFW Year: 2004
• USFW - Doc Type: Annals of the Association of American Geographers
• Copyright Material: NO

462

Schmidt and Box

surface of emergent, bank-attached compound bars and
in the lee of migrating dunes along bar edges (Rakowski
1997; Day, Christopherson, and Crosby 1999). In debris
fan-affected reaches, backwaters occur in eddies where
circulation has slowed or stopped when bars become
emergent during flow recession (Rubin, Schmidt, and
Moore 1990; Goeking, Schmidt, and Webb 2003).
There is substantial year-to-year variation in the size
of backwaters, the discharge at which they exist, and the
geomorphic features that create them. Rakowski (1997)
studied the geomorphology of backwaters in reach H
(Figure 1) in 1993 and 1994 and found that the total
area of nursery habitat changed by as much as 25 percent
from one year to the next because the topography of
alluvial bars changes in response to floods and winter

Figure 3. Photograph showing the Green River in part of Reach H
in the Uinta Basin. This reach has restricted meanders, and back-
waters occur in embayments adjacent to emergent bank.attached
compound bars. Channel form and distribution of backwater hab.
itats in the downstream part of this photograph were modeled by
Andrews and Nelson (1989) and Rakowski (1997). Flow is from
bottom to top.

base flows. Goeking, Schmidt, and Webb (2003) found
similar year-to-year differences in the debris fan-affected
Grand Canyon.
Pucherelli, Clark, and Williams (1990) found that the
relationship between discharge and the area of back-
waters was unimodal in the Green River, based on
measurements in one year at five different discharges in
five reaches, but other studies show that the relation-
ship sometimes has secondary modes (Rakowski 1997;
Goeking, Schmidt, and Webb 2003). The differences in
the shape of this relationship are due to the different
response of bars to large floods, which create high ele-.
vation bars with relatively simple topography, and small
floods, which create complex patterns of scour and fill on
existing alluvial bars and tend to eliminate backwaters in
eddies.

Analysis of Sampling Data

Methods

We analyzed existing data about larval pikeminnow
drift and transport into backwaters. Since the ratio of the
number of larvae that enter the study area in early
summer to the number and distribution of age-O fish in
backwaters in the fall is potentially affected by dam re-
leases in the intervening period, we sought to compare
the two. Large errors are associated with estimates of the
number of larvae that enter the study area because
sampling at the mouth of the Yampa River is confounded
by many factors. Sampling took place for two hours each
day at dawn during the study period (Bestgen, Muth,
and Trammell 1998). We estimated the total number of
larvae entering the study area by multiplying the dawn
sampling data, reported as an hourly rate, by 24. The
hourly drift rates themselves were based on the uncer-
tain assumption that larval drift was well mixed across
the channel, because Bestgen, Muth, and Trammell
(1998) set three nets for about two hours in 30-40 cm
deep flow and estimated transport abundance by divid-
ing the number of larvae sampled by the proportion of
the total discharge that flowed through the sample nets.
Sampling also may have been inefficient, because larvae
were only 8 to 9 mm in length at the time of sampling.
Estimates of fall backwater populations have even
larger variance, because fall populations are small, patchy,
and sampled inefficiently. Of the 472 seine hauls in
backwaters during the six years of concern to this study,
approximately 58 percent contained no fish. About 150
river km were sampled each year, and 33 percent of the
backwaters in the 8-km reaches in the six years of
sampling contained no fish. The backwater sampling