9665

• UCREFRP Catalog Number: 9665
• Author: Schmidt, J. C. and J. B. Box.
• Title: Controls on the Longitudinal Distribution of Age-0 Colorado Pikeminnow in the Middle Green River, Colorado and Utah - Draft Report.
• USFW Year: 2001.
• USFW - Doc Type: Logan, UT.
• Copyright Material: NO

11/14/01 draft report, Schmidt and Box

observations indicate that the time of first arrival at Ouray is consistent wi th model predictions. The
model predicts that larvae enter backwaters in every reach, because every reach has shorelines of
sufficient complexity to produce backwaters. In the early stages of drift, water exchange between
the main flow and backwaters is predicted to be high and the swimming ability of larvae is low, so
larvae are swept back into the main current and continue downstream. The period of transport into
backwaters in the downstream part of the study area is of longer duration because there is low,
persist drift caused by leakage from backwaters further upstream.
The proportion of the drifting population that is predicted to be transported into backwaters
depends on the value of cr and the rate of fish growth, as shown in (3). We evaluated the effect of
values of cr on model predictions. Larger values of cr lead to higher predicted rates of transport
into backwaters. As cr increases from 0.5 to 0.8, the model predicts that the peak population in
backwaters increases by 54% in reach 345335 and by 47% in reach 335325 (Table 5). Higher
values of cr lead to predictions of larger backwater populations throughout the study area, because
the larger initial transport into backwaters in the upstream reaches provides a higher predicted late
season drift due to leakage. Differences in model parameters never lead to predictions of transport
into backwaters of more than about 10% of the drifting population, however. Thus, the predicted
date of arrival of the first larval fish and the date when the majority of the total drifting population
pass a reach do not change, regardless of value of cr. Model predictions are consistent with field
observations. Differences in model parameters do not result in significant differences in predicted
populations in late summer during the period of field sampling, because sampling takes place more
than two months after the peak concentrations of larval fish have passed downstream.
Discharge at the time of drift plays an important role in determining the proportion of the .
simulated population transported into backwaters, because discharge determines flow speed and
shoreline complexity (Fig. 9). Larval fish entering the Green River earlier in the runoff season of
1992 drift at higher discharges, are transported downstream at faster rates, and encounter a more
complex shoreline that leads to higher predicted peak populations of larval fish during the first few
days of drift (Fig. 10).

Model Simulations
Model predictions of backwater populations in September do not predict mean CPE (Fig.
11), although the general longitudinal pattern of larval fish distribution is consistent with some of
the field data The coefficient of determination (RZ) is 0.14 for 7 reaches in the 6 years of this
study (n=42), and most of the variation in mean CPE is not predicted by the regression curve
between predicted population and mean epE.
Discrepancies between field data and model results may not be solely related to
inadequacies of the model, and these discrepancies may highlight problems in the present sampling
program. For example, the model predicts that the total number of larval fish was greater in 1990

13