9702

• UCREFRP Catalog Number: 9702
• Author: Petersen, J.H., and C.P. Paukert.
• Title: Development of a bioenergetics model for humpback chub and evaluation of water temperature changes in the Grand Canyon, Colorado River.
• USFW Year: 2005.
• USFW - Doc Type: Transactions of the American Fisheries Society
• Copyright Material: NO

f
964
PETERSEN AND PAUKERT
tui chub (Table 1). Activity had bounds of 0.5-
2.0, and p was allowed to range from 0.0 to 1.0
(Table 2).
A laboratory growth experiment conducted at
24°C by Gorman and VanHoosen (2000) was used
to establish the test criteria for Monte Carlo fil-
tering. Gorman and VanHoosen (2000) conducted
growth experiments at three temperatures (12, 18,
and 24°C) with four native fishes of the Colorado
River, including humpback chub. Their experi-
ments were conducted from January through Sep-
tember, starting with fish that were approximately
4 g. We used results from experiments at the high-
est temperature, 24°C, assuming this would give
the best estimates of optimum growth for these
warmwater fish. Fish were fed flaked krill, com-
mercial pelletized feed (Silvercup #2), and thawed
frozen brine shrimp three times daily (morning,
midday, and early evening). Fish were offered 12%
of their body weight daily, and tanks were si-
phoned daily to remove excess food; however, the
amount of food that was uneaten was not reported.
In Gorman and VanHoosen's experiment, hump-
back chub started at 3.9 g and grew to 21.7 g over
a 238-d period (Figure IA).
A bioenergetics model was constructed to grow
simulated fish, and Monte Carlo methods were
used to sample parameters within reasonable
bounds (Table 2). Simulated fish started at 3.9 g
and were grown for 238 d at 24.2°C, the measured
temperature (Gorman and VanHoosen 2000). To
test the adequacy of a randomly drawn parameter
set, the size of the modeled fish was compared to
the observed size halfway through the growth pe-
riod (120 d; 12.3 g) and at the end of the growth
period (238 d; 21.7 g). To be acceptable and in-
cluded for further analysis, a parameter set had to
produce growth of modeled fish that was within
one standard error of the mean observed size at
the halfway point and the end of the experiment.
Preliminary parameter filtering indicated that test-
ing only the final size of fish was insufficient be-
cause certain parameter sets met the end criteria,
but size trajectories were largely outside the gen-
eral pattern of growth (Figure 113).
Monte Carlo sampling of temperature-related
parameters for consumption and respiration re-
quired special attention. In the formulations that
we used for temperature functions, optimum and
maximum temperature parameters were required
and were thus sampled in the Monte Carlo pro-
gram. The optimum temperature was assumed to
be no greater than the temperature of the labora-
tory growth experiment used to fit parameters
30
25
20
15
10
5
Cl 30
25
20
15
10
5
0
Day of year
FIGURE 1.-Fit of a bioenergetics model to humpback
chub growth. In panel (A) average parameter values
from Monte Carlo filtering were used to simulate the
growth of a subadult humpback chub (solid line), which
was compared with the observed laboratory growth at
24°C (means [circles] ± SEs). In panel (B) examples of
growth trajectories are shown, each with a different set
of randomly chosen parameter values. In this example,
two of the six trajectories met the growth criteria (ver-
tical heavy bars), which were based on laboratory size
on days 128 and 246 (see panel A). Laboratory data are
from Gorman and VanHoosen (2000).
(24°C), and the optimum temperature was less than
the maximum temperature, that is,
temperature of experiment
<_ optimum temperature
< maximum temperature.
This rule was applied to both the consumption
(CTO, CTM) and the respiration (RTO, RTM) pa-
rameters. Although the optimum temperature
might have been less than the temperature used in
the growth experiment (24°C), Gorman and
VanHoosen (2000) as well as other information
suggest that optimum temperature for humpback
chub is likely quite high and is above 24°C because
they evolved in environments where temperatures
were often above this value (Stevens et al. 1997).
Without inclusi
sible to get non
Carlo sampling
maximum teml
rameter combi
temperature pai
test criteria.
Many paran
growth rates th
other paramete
test criteria at
time but not at
combinations 1
duced growth c
in the 24°C lab
1B). Using the
acceptable sets
laboratory gro,
model paramet
tion.
Final param
and sensitivity.
to Carlo mode]
ing that fish in
either their ma
maximum rate,
fered a 12% ra
man and Vanl
below their in
perature was n
competition, c
For corroborat
values, we co
parameters to f
two assumpti(
ability was in
feeding at thei
1.0), and (2) b
in growth expl
below maxims
to vary during
1.0. For this s
sampled as a I
the Monte Ca:
ters.
To corrobol
select final pai
growth rates 1
chub from va
ran the model
and fish size;
p values nea
Growth under
reasonable p
0 50 100 150 200 250