9724

• UCREFRP Catalog Number: 9724
• Author: Coggins, L.G., W.E. Pine, C.J. Walters, D.R. VanHaverbeke, D. Ward and H.C. Johnstone.
• Title: Abundance trends and status of the Little Colorado River population of humpback chub.
• USFW Year: 2006.
• USFW - Doc Type: North American Journal of Fisheries Management
• Copyright Material: YES

240

COGGINS ET AL.

to represent a range of possibilities for explaining the
observed changes in catch rate in the LCR as well as
the ontogeny of downstream migration. This is
unsatisfactory as a long-term solution to the problem
of disentangling movement and sampling intensity
effects, but it allowed us to determine whether
recognition of seasonal movement results in different
abundance trend estimates than those obtained from the
spatially aggregated methods. The results of these
scenario tests were essentially the same as those of the
annual ASMR and Jolly-Seber models shown in Fig-
ure 5. We were not able to fit multistate models in
MARKJor reasons similar to those described above for
ASMR: low capture probabilities between states and
the associated failure to estimate a transition probabil-
ity.

Discussion
Population Trend and Abundance Assessments

All of our evidence implies that the adult humpback
chub population in the LCR has experienced a large
decline in abundance since 1989. Low sampling effort
in the late 19908 (1996-1999), heterogeneity in capture
probability related to age, and a large number of age-
classes make complicated models with large numbers
of parameters necessary to realistically estimate
population parameters. However, population trend
and annual population size estimates from all catch
indices and methods indicate declines in population
size of 30--60% since the early 1990s.
A comparison of the recent closed-population
(Lincoln-Petersen) estimates with other closed-popu-
lation estimates from the early 1990s indicates large
declines in population size. The only published
population estimates for humpback chub in the LCR
are closed-population estimates of fish exceeding 150
nun TL by Douglas and Marsh (1996). Their estimates
for the spring of 1992 ranged from about 4,300 to
5,500 (Figure 3). Our closed-population estimates from
spring 2000-2003 ranged from about 2,000 to 3,400
fish (TL > 150 nun; Figure 3). A comparison of
closed-population estimates for. fish occupying the
inflow reach suggests a decline in adult abundance of
about 60% between the early 1990s and 2001 (Valdez
and Rye11995; Trammell and Valdez 2003). Although
the assumptions of closed-population models (particu-
larly the lack of animals moving into and out of the
sampling area) may not have been met, Lin-
coln- Petersen estimates do allow some relaxation of
this closure assumption (pollock et al. 1990). For
example, if unmarked animals immigrate into the
sampling area, then the Lincoln-Petersen population
estimate is not biased for the second sample. If both
marked and unmarked animals emigrate between the

samples, then the Lincoln-Petersen estimate is un-
biased by emigration for the first sample. IT only
untagged animals emigrate between samples, then
population size estimates would be negatively biased.
However, it is unlikely that tagging effects only
occurred in the early 2000s, and any bias caused by
tagging probably occurred across all closed-population
estimators in the 1990s and 2000s. As such, these
estimates still serve as a relative indicator of an
apparently large decline in the humpback chub
population.
Using open-population models (Jolly-Seber and
ASMR), .the estimated LCR humpback chub popula-
tion of adult fish (i.e., TL >200 mm, age 4 and older)
is currently between 2,400 and 4,400 individuals
(Figure 5). No other open-population size estimates
have been published for this endangered species. The
TSM open-population trend estimates provide an
alternative to the age-dependent Jolly-Seber and
ASMR model annual population estimates. Population
growth estimates using TSM methods are robust to
heterogeneity in capture probability under the assump-
tion that the heterogeneity of the population does not
change over time (e.g., capture rates for smaller fish are
always lower than those for larger fish; Williams et al.
2002). Estimates of population growth also do not
depend on geographic closure of the sample area as
long as the measured population change of the sampled
area matches that of the population as a whole
(Schwarz 2001; Hines and Nichols 2002). All bi-
ologically reasonable TSM models, including con-
strained models that allow survival and capture
probability to remain time dependent and population
change to be time independent (Franklin 2001),
indicate annual declines in population size of up to
14%. IT the trends in annual estimated population size
from open-, closed-, or TSM population models are
examined, all approaches agree that the abundance of
the LCR humpback chub population has declined
significantly since at least the early 199Os.

Model Performance and Potential Assessment Errors

Abundance trends from the Jolly-age and ASMR
models are similar, though the absolute abundance
estimates differ between the two approaches, particu-
larly in the early years of the study. These differences
arise because the ASMR model predicts the number of
fish available for capture by using the existing age
structure of the population at the beginning of the study
while the Jolly-age approach only uses information
gained from recapture of tagged fish. Thus, as the
tagged-fish population increases in the middle to late
years of the study, the estimates from the two
approaches converge (Figure 4).