See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/281547921

Assessing Juvenile Native Fish Demographic
Responses to a Steady Flow Experiment in a
Large Regulated River
Article in River Research and Applications · March 2015
DOI: 10.1002/rra.2893

CITATIONS

READS

2

40

8 authors, including:
William E. Pine

Charles B Yackulic

University of Florida

United States Geological Survey

75 PUBLICATIONS 1,259 CITATIONS

54 PUBLICATIONS 833 CITATIONS

SEE PROFILE

SEE PROFILE

Michael Yard

Brandon Gerig

United States Geological Survey

University of Notre Dame

14 PUBLICATIONS 196 CITATIONS

6 PUBLICATIONS 21 CITATIONS

SEE PROFILE

All in-text references underlined in blue are linked to publications on ResearchGate,
letting you access and read them immediately.

SEE PROFILE

Available from: William E. Pine
Retrieved on: 21 October 2016

RIVER RESEARCH AND APPLICATIONS

River Res. Applic. (2015)
Published online in Wiley Online Library
(wileyonlinelibrary.com) DOI: 10.1002/rra.2893

ASSESSING JUVENILE NATIVE FISH DEMOGRAPHIC RESPONSES TO A STEADY
FLOW EXPERIMENT IN A LARGE REGULATED RIVER
C. FINCHa, W. E. PINE IIIa*, C. B. YACKULICb, M. J. DODRILLa,b, M. YARDb, B. S. GERIGa,
L. G. COGGINS JR.c AND J. KORMANd
a

Department of Wildlife Ecology and Conservation, University of Florida, Gainesville, Florida, USA
US Geological Survey, Grand Canyon Monitoring and Research Center, Flagstaff, Arizona, USA
c
US Fish and Wildlife Service, Yukon Delta National Wildlife Refuge, PO Box 346, Bethel, Alaska, USA
d
Ecometric Research, Vancouver, British Columbia, Canada
b

ABSTRACT
The Colorado River below Glen Canyon Dam, Arizona, is part of an adaptive management programme which optimizes dam operations to
improve various resources in the downstream ecosystem within Grand Canyon. Understanding how populations of federally endangered
humpback chub Gila cypha respond to these dam operations is a high priority. Here, we test hypotheses concerning temporal variation in
juvenile humpback chub apparent survival rates and abundance by comparing estimates between hydropeaking and steady discharge regimes
over a 3-year period (July 2009–July 2012). The most supported model ignored ﬂow type (steady vs hydropeaking) and estimated a declining
trend in daily apparent survival rate across years (99.90%, 99.79% and 99.67% for 2009, 2010 and 2011, respectively). Corresponding abundance of juvenile humpback chub increased temporally; open population model estimates ranged from 615 to 2802 individuals/km, and
closed model estimates ranged from 94 to 1515 individuals/km. These changes in apparent survival and abundance may reﬂect broader
trends, or simply represent inter-annual variation. Important ﬁndings include (i) juvenile humpback chub are currently surviving and
recruiting in the mainstem Colorado River with increasing abundance; (ii) apparent survival does not beneﬁt from steady fall discharges from
Glen Canyon Dam; and (iii) direct assessment of demographic parameters for juvenile endangered ﬁsh are possible and can rapidly inform
management actions in regulated rivers. Copyright © 2015 John Wiley & Sons, Ltd.
key words: peaking ﬂows; hydropower; endangered species; Colorado River; regulated river
Received 27 June 2014; Revised 21 October 2014; Accepted 10 February 2015

INTRODUCTION
In over half the world’s large rivers, dams modify ﬂows by
storing water in upstream reservoirs and regulating downstream discharges to facilitate shipping, power production
and ﬂood control, or to provide water for industry, municipalities and irrigation (WCD, 2000; Poff et al., 2006; Richter
and Thomas, 2007). Although the magnitude of impacts are
as varied as the dams themselves (ranging from small ‘runof-the-river’ type dams to large structures impounding more
than the mean annual ﬂow of a river), the attendant physical
and biological changes incurred due to river regulation are
similar across climates, latitudes and elevations (Blinn and
Poff, 2005; Nilsson et al., 2005; Poff et al., 2006). Changes
to the physical environment include sequestration of woody
debris and sediment in upstream reservoirs (Stanford and
Ward, 1991; Kearsley et al., 1994; Schmidt et al., 1998),
modiﬁcation of temperature regimes (Stanford and Ward,
*Correspondence to: W. E. Pine III, Department of Wildlife Ecology and
Conservation, University of Florida, 110 Newins-Ziegler Hall, Gainesville,
Florida, 32611, USA..
E-mail: billpine@uﬂ.edu

Copyright © 2015 John Wiley & Sons, Ltd.

1991; Clarkson and Childs, 2000) and alteration of the
timing, magnitude and frequency of ﬂows (Poff et al.,
1997; Lovich and Melis, 2007; Burgess et al., 2013; Dutterer
et al., 2013). Ecological impacts range from altered trophic
dynamics (Stevens et al., 1997; Kennedy and Gloss, 2005)
to development of novel ﬁsh communities, often dominated
by non-native species (Mueller and Marsh, 2002; Coggins
et al., 2011).
Where dams are used for hydropower production, diel variations in downstream river discharge represents a speciﬁc
form of habitat modiﬁcation that affects the physical conditions within rivers (i.e. depth and velocity) creating
unnatural and altered habitat conditions (Fette et al., 2007).
Juvenile ﬁsh that predominately use nearshore or shallow
areas are thought to be most affected by ﬂuctuations in discharge (Gaudin, 2001) that force them away from preferred
habitats (Bunt et al., 1999) leading to slower growth rates
(Korman and Campana, 2009) and potentially exposing them
to higher predation rates by altering foraging arenas (Walters
and Juanes, 1993; Ahrens et al., 2012). In shallow littoral habitats (especially vegetated cobble bars or backwaters), warm
water microhabitats can develop under steady ﬂow conditions

C. G. FINCH ET AL.

(Korman and Campana, 2009; Ralston, 2011), which may be
important to juvenile life stages of warm water native ﬁsh
(Trammell et al., 2002; USDOI, 2008; Ralston, 2011).
One study of the Tallapoosa River below Harris Dam in
the southeastern USA documented reduced amounts of shallow water habitat and decreased young-of-year ﬁsh abundance because of water level ﬂuctuations, which resulted in
changes in the ﬁsh community composition (Freeman
et al., 2001). Similarly, Glen Canyon Dam in northern
Arizona (Figure 1) is operated to match electrical power
demand in the southwestern USA, with diel ﬂuctuations in
dam discharge (hydropeaking; Figure 2). These ﬂuctuations
can shift the availability of different nearshore habitat types
in the downstream Colorado River, especially where shoreline angles are low (Korman and Campana, 2009). Habitat
alterations in Grand Canyon as a result of the federally operated Glen Canyon Dam may have signiﬁcant repercussions
on the imperilled native ﬁsh community downstream
(Minckley, 1991; Converse et al., 1998).

Despite the effects of diel ﬂuctuations on habitat, there is
wide variation in community (Murchie et al., 2008) and individual species responses (Steele and Smokorowski, 2000;
Young et al., 2011) to ﬂow stabilization. Ecological complexity and uncertainty often necessitate an experimental
approach towards optimizing management policies to reach
conservation targets (e.g. waterfowl harvest, Williams
et al., 1996; Everglades restoration, Gunderson and Light,
2006). Below Glen Canyon Dam, experimental management
approaches are challenged by the constraints of minimizing
physical and ecological impacts while still meeting the water and power needs of tens of millions of people (Reisner,
1986; Anderson and Woosley, 2005; Richter and Thomas,
2007). The dam was built on the Colorado River (Figure 1)
as a water supply and hydropower project before the
drafting of the National Environmental Policy Act, the
Endangered Species Act, or the Clean Water Act, yet must
now operate within the conﬁnes of this legislation. To facilitate environmental compliance, the 1995 Environmental

Figure 1. Map of study area near conﬂuence of Colorado and Little Colorado Rivers in Grand Canyon, northern Arizona, USA. Black box
denotes the area containing the Little Colorado River aggregation of humpback chub Gila cypha, and black dots represent the river kilometre
(rkm) measured from Lee’s Ferry, Arizona
Copyright © 2015 John Wiley & Sons, Ltd.

River Res. Applic. (2015)
DOI: 10.1002/rra

DAM OPERATION EFFECTS ON FISH

Figure 2. Daily discharge in the Colorado River at Lee’s Ferry over the 4-year period of this study (2009–2012). Sampling intervals during
2009–2011 occurred in July, August, September and October, and sampling in 2012 occurred in late-April and July and are, as indicated
by the blue rectangles. Data from USGS gage 0938000 available at waterdata.usgs.gov (Accessed February 2014). This ﬁgure is available
in colour online at wileyonlinelibrary.com/journal/rra

Impact Statement on Glen Canyon Dam in Grand Canyon
National Park identiﬁed a number of important uncertainties
for the Colorado River related to water releases from the
dam (United States Department of Interior, 1996), which
led to the creation of the Glen Canyon Dam Adaptive
Management Programme (GCDAMP; USBR, 2013).
Provisions of the GCDAMP intrinsically include experimental testing of operational alternatives to discover which
discharge regimes will simultaneously satisfy the myriad
social and legal constraints of operating Glen Canyon Dam.
As an example, GCDAMP has implemented engineered
ﬂoods across a range of magnitudes, durations and upramp/
downramp rates to identify policies that will maximize retention of seasonally available sediment and construct sand bars
that are important for camping, archaeological preservation
and possibly ﬁsh habitat (Melis et al., 2011). In practice,
the range and scope of testable alternatives is limited by
infrastructure and stakeholder interests. These experimental
constraints are typiﬁed by the aforementioned engineered
ﬂoods, which are limited to a fraction of the predam ﬂood
discharge because of the diameter of the discharge tubes.
However, considerable uncertainty still remains even within
Copyright © 2015 John Wiley & Sons, Ltd.

the available range of management options, especially when
considering non-stationary resources that are difﬁcult to
measure like migratory birds, rare plants or ﬁshes.
One goal of the GCDAMP is population recovery of
federally endangered humpback chub Gila cypha, a largebodied, morphologically distinct cyprinid endemic to the
Colorado River basin whose largest population is in Grand
Canyon. The reasons humpback chub populations in Grand
Canyon are imperilled are unresolved but likely include
negative interactions with non-native ﬁsh (Coggins et al.,
2011; Yard et al., 2011); loss of essential habitats due to ﬂow,
temperature and sediment input modiﬁcations (Converse
et al., 1998; Clarkson and Childs, 2000; Stone and Gorman,
2006); and non-native parasites (Minckley, 1991; Choudhury
et al., 2004).
Most humpback chub currently occupy the Colorado
River near the conﬂuence with the Little Colorado River,
the largest tributary within Grand Canyon (Figure 1).
Adults participate in a spawning migration from the
mainstem Colorado River to the Little Colorado River, after
which the majority return to the mainstem Colorado River
(Kaeding and Zimmerman, 1983; Valdez and Ryel, 1997;
River Res. Applic. (2015)
DOI: 10.1002/rra

C. G. FINCH ET AL.

Gorman and Stone, 1999). Unknown but potentially significant proportions of each year’s cohort emigrate to the
mainstem Colorado River as larvae or small juveniles,
where their survival is thought to be low because of cold
water temperatures and negative interactions with nonnative species (Valdez and Ryel, 1997; Clarkson and
Childs, 2000; Yard et al., 2011). These mainstem-resident
juveniles are the targeted benefactors of experimental discharges from Glen Canyon Dam. Alternatively, juvenile
humpback chub may remain in the Little Colorado River
for several years (Gorman and Stone, 1999; Clarkson and
Childs, 2000; Limburg et al., 2013), where warmer water
temperatures and lower predation risk from non-native species (Kaeding and Zimmerman, 1983; Marsh and Douglas,
1997) improve vital rates like growth and survival, thus reducing time spent at smaller, more vulnerable size classes
and ultimately increasing the adult spawning population
(Gorman and Stone, 1999).
Field experiments implemented by GCDAMP intended to
beneﬁt juvenile humpback chub occupying the Colorado
River include non-native predator removal (Coggins et al.,
2011; Yard et al., 2011), ﬂuctuating ﬂow mitigation
(Ralston, 2011) and habitat restoration (Melis et al., 2011;
USDOI, 2008). A critical objective of these experiments
is to increase recruitment to the adult population by enhancing vital rates of juvenile humpback chub. However,
interpreting effects of these ecosystem experiments on juvenile humpback chub is challenging because of the difﬁculty
of directly monitoring population dynamics of juveniles.
Trends in juvenile humpback chub have traditionally been
made through inferences based on catch-rate indices as well
as from reconstructions of historical recruitment patterns
from recaptures of tagged adult humpback chub as part of
the long-term adult humpback chub assessment programme
(age-structured mark-recapture or ASMR; Coggins et al.,
2006a, 2006b). Because of seasonal contrasts in growth rates
between the Colorado and Little Colorado Rivers (Finch
et al., 2013), as well as the propensity of humpback chub
to unpredictably occupy either or both locations (Limburg
et al., 2013; Yackulic et al., 2014), a 100-mm humpback
chub (size at ﬁrst tagging) could be anywhere from 1 to
5 years old (Limburg et al., 2013). This uncertainty reduces
the annual variability of reconstructed recruitment trends
and challenges the ASMR model’s ability to determine
which years (and corresponding dam operations/habitat conditions) caused observed ﬂuctuations in juvenile population
dynamics (Coggins and Walters, 2009). Directly measuring
juvenile humpback chub abundance and survival would
reduce uncertainty and shorten the time lag between experiment results and informing future policies.
We conducted a mark-recapture study of juvenile humpback chub in Grand Canyon to address sampling and modelling uncertainties, as well as to substantiate the prevailing
Copyright © 2015 John Wiley & Sons, Ltd.

hypothesis that ﬂuctuating ﬂows in Grand Canyon are detrimental to juvenile humpback chub (Ralston, 2011). The
GCDAMP designed an experiment to address these questions by implementing steady ﬂows (i.e. constant water
release from Glen Canyon Dam) during September and
October 2008–2012 (Figure 2). Our monitoring could then
assess trends in apparent survival and abundance of juvenile
humpback chub, with the experimental goal being improving recruitment and population growth rate of adult humpback chub.
Although a steady ﬂow policy to effect these changes
may be intuitive by design, the beneﬁts to native ﬁsh are
both highly uncertain (non-native species can beneﬁt from
steadier ﬂows; Korman et al., 2009, 2012) and expensive
to implement because of lost power revenue (Palmer et al.,
2004; Ralston, 2011). Additionally, detecting the potentially
small effects of this policy is difﬁcult because of several
experimental design issues including conducting the experiment during late summer and fall (when river warming is
less likely to occur), short duration of the experiment
(approximately 2 months each year) and limited replication
(3 years of evaluation). Nonetheless, experimental manipulations of entire river ecosystems are rare and represent
opportunities for substantial learning (Finch et al., 2013;
Gerig et al., 2014). In this paper, we develop a capture–
recapture framework to directly estimate abundance and
apparent survival of juvenile humpback chub and assess
trends in juvenile humpback chub apparent survival and
abundance in three of the ﬁve experimental steady ﬂow
years (2009–2011). We present ﬁeld techniques and analytical approaches used for this work and then discuss how
our results ﬁt within the broader context of understanding
juvenile humpback chub population dynamics in Grand
Canyon. If juvenile humpback chub apparent survival rates
or abundance improve as a direct result of experimental
steady ﬂows, then steady ﬂows could be considered as part
of a long-term management plan for Glen Canyon Dam
operations. From a global perspective, understanding the
ﬁsh-ﬂow relationship is increasingly important in order to
design dam operations that mitigate associated negative
ecological changes. The results presented here highlight
the value of empirically assessing ﬁsh responses to the
elimination of ﬂuctuating ﬂows, typiﬁed by our experience
with humpback chub in Grand Canyon.

METHODS
Current summertime operations at Glen Canyon dam attempt to match power demand according to the stipulations
of a management policy known as ‘modiﬁed low ﬂuctuating
ﬂows’, which nearly doubles discharges every day from
around 255 to around 480 m3/s (USBR, 1996; Figure 2).
River Res. Applic. (2015)
DOI: 10.1002/rra

DAM OPERATION EFFECTS ON FISH

This policy is a milder version of the ‘No Action’ alternative
in operation from the 1960s until the 1990s, where daily
ﬂuctuations ranged as much as 28–892 m3/s (USBR,
1996). For the experimental ﬂows we evaluated, modiﬁed
low ﬂuctuating ﬂows were replaced with steady discharges
of 295 and 230 m3/s during September and October of
2009 and 2010, respectively (Figure 2). Flows for the entire
summer 2011 period were steady (September and October
discharges in 2011 were 453 m3/s) as a result of equalization
ﬂows triggered by high snowmelt in spring 2011, which
provided important contrast to separate steady ﬂows and
seasonality (Figure 2; see discussion). Our study location
was ~125 km below Glen Canyon Dam, so water arrived
about 24 h after being released, with slight attenuation of
temperature and discharge extremes according to local
weather and channel roughness. Local peak discharge
occurred during the day, with minimum discharge around
midnight. The Colorado River channel in our study reach
is heavily inﬂuenced by debris fans, rapids and large eddy
pools. Shorelines are mostly steep talus, cliffs and vegetated
cut banks, with very few sandbars or low-angle areas.
We based inferences on 14 sampling events (river trips)
over 2009–2012 in a section of the Colorado River
102–104.5 river km below Lee’s Ferry (126.1–128.6 km below Glen Canyon Dam) and about 2 km downstream of the
conﬂuence with the Little Colorado River. We sampled for
12 consecutive days in July, August, September and October
of 2009–2011, and nine consecutive days in April and July
2012. Across trips, we sampled the ﬁsh community using
two gear types: (i) unbaited mini hoopnets (50 cm diameter,
100 cm long, single 10-cm throat, made of 6-mm nylon
mesh, ﬁshed over 24-h intervals and checked daily); and
(ii) boat electroﬁshing at night during low ﬂows (pulsed direct current at 15–20 amps, 200–300 volts, 7–10 s per metre
of shoreline, repeated 24–72 h apart for three passes per trip).
Initial sampling in July 2009 included 48 hoopnets in the
upper 1.5 km of the sampling reach, which expanded to
60 hoopnets in August–October, all of which were ﬁshed
12 consecutive days for about 24 h each. In 2010 and 2011,
we further expanded hoopnet efforts to include 12 passes
of 80 hoopnets over the upper 2 km of the sampling reach.
In April and July 2012, hoopnet sampling was further
increased to 90 hoops over the entire 2.5 km, but with only
seven or eight, 24-h passes. All available shoreline were
electroﬁshed during the 12 sampling trips in 2009–2012
following guidelines in Korman and Campana (2009)
targeting small ﬁsh with two additional sampling passes in
2012 using standard Grand Canyon ﬁsh sampling electroﬁshing protocols.
Following capture, we measured, tagged and then released
humpback chub to within 25 m of the captured location. We
gave humpback chub from 40 to 99 mm total length (TL)
batch marks that were identiﬁable to speciﬁc gears and
Copyright © 2015 John Wiley & Sons, Ltd.

sampling trips using visible implant elastomer (VIE, Northwest Marine Technologies, Shaw Island, Washington, United
States). Any time a humpback chub was recaptured on a subsequent trip or with a different gear while still <100 mm TL,
we applied a new VIE batch mark. Recapturing within a trip
with the same gear type did not receive an additional mark.
Any humpback chub ≥100 mm TL (including VIE recaptures)
received a uniquely coded 134.2 kHz passive integrated transponder (PIT) tag (12-mm long, Biomark). Although humpback chub less than 40 mm are present in the river, these
individuals were not equally vulnerable to our gear and could
not be included in our mark-recapture study without introducing substantial sampling bias. Additionally, the technology is
not yet available for mark-recapture of such small ﬁsh without
increasing risk of handling mortality, which is especially
problematic for federally endangered species.
Because our ﬁeld programme was speciﬁcally designed in
an experimental framework to document any changes in
apparent survival or abundance that might occur relative to
the steady ﬂow (vs regular or annual monitoring without an
experimental objective), sampling trips were clustered around
1 September of each year, when the ﬂow experiment began.
Upon completion of the ﬁrst 3 years of ﬁeld work, the sampling calendar then transitioned into a regular monitoring programme with larger gaps between trips to increase annual
resolution (hence the clustering of trips in the ﬁrst 3 years vs
2012; Figure 2). We also had little precedent for our ability
to directly monitor juvenile humpback chub via markrecapture. With incremental successes and failures, we
adapted our efforts (i.e. increasing hoopnet effort) to expand
the pool of marked animals and increase statistical inference.
Using simulation to evaluate model accuracy
In capture–recapture experiments, it is possible to confound
apparent survival rates and capture probability estimates
due to the inability to distinguish between ﬁsh that died
(low survival) and ﬁsh that lived but were not observed
(high survival and low capture probability). We expected
juvenile humpback chub capture probabilities to be low
based on pilot samples and other capture–recapture experiments in large rivers (Coggins et al., 2006a; Korman et al.,
2009; Lauretta et al., 2013). These low capture probabilities
combined with short discrete sampling intervals (designed
around the ﬂow experiment) could lead to erroneous
survival estimates. To assess this, we generated data sets
(programme GENCAPH1; http://www.mbr-pwrc.usgs.gov/
software/gencaph1.shtml) with known populations, survival
rates, capture probabilities and sampling intervals similar to
our expected values based on a pilot study and experimental design theory. We then assessed the accuracy and precision of survival estimates from a Cormack–Jolly–Seber
(CJS) model using these input data (Appendix).
River Res. Applic. (2015)
DOI: 10.1002/rra

C. G. FINCH ET AL.

Apparent survival rate and abundance estimation using an
open model
In most survival studies, unique marks allow the fates of individuals to be monitored through time. We were interested
in survival patterns of humpback chub too small to be
marked individually with PIT tags, necessitating the use of
batch marks (Pine et al., 2013). Batch marks present a peculiar problem in developing capture histories because individuals are not discernible. However, compilation of capture
histories is possible because all recaptured individuals (other
than recaptures using the same gear within a trip) received a
new, separate mark and could then be subtracted from the
number of individuals previously marked but not recaptured.
This framework is different than batch marks that only distinguish marked versus unmarked ﬁsh, but it still does not
contain information that allows for unique identiﬁcation of
each individual ﬁsh. By not having individual marks, we
are unable to analyse these data using models with behavioural effects (Otis et al., 1978) or robust design approaches
(Pollock et al., 1990). In our study, once ﬁsh recruit into subadult sizes (≥100-mm TL), they receive unique PIT tags.
These larger PIT tagged ﬁsh were then excluded from our
analyses of juvenile humpback chub because they are presumed to be more robust to hydropower ﬂuctuations and
are not the current primary management target for optimizing
Glen Canyon Dam operations.
We developed a set of a priori models to test hypotheses
about effects of experimental steady ﬂows on humpback
chub apparent survival and capture probability over the
3-year study period. These are open, CJS (Pollock et al.,
1990) models and are based on current understanding of
humpback chub biology and life history, coupled with
discharge information. We chose to present results as daily
apparent survival rates, which need to be exponentiated by
365 to allow them to be comparable with annual estimates
such as from the ASMR model (Coggins et al., 2006b).
In our suite of models, we parameterized both apparent
survival and capture probability according to six different
ecological assumptions (36 total models): (i) constant,
implying that the parameter does not change based on ﬂow
type or any other factors; (ii) ﬂow, where the parameter is
inﬂuenced by the ﬂow type (steady or ﬂuctuating) that
occurred during intervals between sampling trips; (iii)
ﬂow*year, which considers the ﬂow treatment within a year
as independent from other years, because the magnitudes of
steady ﬂows were different in 2009, 2010 and 2011; (iv)
year, to determine if the interaction of ﬂow*year performs
better than considering year alone; (v) time, a global model
allowing parameter estimates to differ independently on
each occasion; and (vi) mod_ﬂow. This last model is an extension of the ﬂow model because not all survival intervals
(at-large periods between trips; Figure 2) are composed of
Copyright © 2015 John Wiley & Sons, Ltd.

a single-ﬂow treatment (Figure 2, between blue boxes); thus,
the mod_ﬂow model estimates a third parameter for intervals that are both steady and ﬂuctuating. Model ﬁt was
assessed using Akaike Information Criterion for ﬁnite
sample size (AICc) and support criteria from Burnham and
Anderson (2002).
We also used these marked ﬁsh to estimate juvenile
humpback chub abundance in an open population framework using the Horvitz–Thompson estimator, which is
based on the capture probabilities from the most supported
CJS apparent survival model and the corresponding catch
information (CJS models explained in detail earlier). This
abundance calculation is simply catch divided by capture
probability (Williams et al., 2002), and variances for abundance estimates were estimated using the method described
in McDonald et al. (2005). This method of estimating abundance differs from the sequential Bayes’ closed estimate
method described in the succeeding text because estimates
of capture probability are derived from data spanning multiple trips. This difference in abundance estimation methods is
important because the population used in open models will
be larger than that considered in the sequential Bayes’
closed capture approaches described in the succeeding text,
where all ﬁsh are <100 mm TL.
Abundance of juvenile humpback chub—closed model
^ ) of juvenile humpback chub
We estimated abundance ( N
(40–99 mm TL) for each gear using sequential Bayesian
closed population models described by Gazey and Staley
(1986) based on a running tally of the marked population
size that accumulates individuals with each pass. In the
Gazey and Staley (1986) framework, the distribution of the
population size is unknown and assumed to be represented
by the non-informative discrete uniform distribution (i.e.
prior distribution), where the ﬁrst prior is uninformative.
^ i ≥ the number of animals marked in the
We assume that N
time interval over which closure is assumed (each river trip).
The number of animals marked each trip, captured each trip
and recaptured are then combined with the prior distribution
to create the posterior distribution (i.e. the probability of
each possible Ni given the data) using Bayes’ theorem
(Gazey and Staley, 1986). In this framework, the posterior
distribution becomes more concentrated (i.e. smaller distribution) about the true population size as the number of
sequential samples increases because the cumulative
number of animals captured at each time increases (example in Figure 3). We calculated the mode of the posterior
distribution as the Bayesian Point Estimate of the sampling
distribution to determine the population estimate (Gazey
and Staley, 1986). The 2.5% and 97.5% values (quantiles,
when normally distributed) are reported as approximate
95% credible intervals. Our basic assumptions for this
River Res. Applic. (2015)
DOI: 10.1002/rra

DAM OPERATION EFFECTS ON FISH

Simulation results to inform analyses of ﬁeld data
Simulations demonstrated that apparent survival and capture
probability can be estimated using certain models, even with
a small population (~500 individuals) and low capture probabilities (5%). We found that the most precise estimates
of apparent survival came from models where survival was
a shared parameter across annual, seasonal or ﬂow treatments. Models that assumed independent time effects for
apparent survival did not converge (Appendix 1).
Assessing the effects of the ﬂow experiment on juvenile
humpback chub apparent survival

Figure 3. Likelihood estimates of juvenile humpback chub
40–99 mm TL abundance from hoopnetting data collected for July
2010. Thick black line represents the Bayesian point estimate of
abundance after 12 nights of hoopnet sampling, while the thin
horizontal line represents the (diffuse) likelihood estimate of abundance after 1 night of sampling. Other thin lines represent likelihood estimates after sequential nights of hoopnetting showing
that a minimum of seven nights of hoopnetting were required
before a reasonable (i.e. informed) estimate of abundance was
reached. With increasing samples (nights of hoopnetting), the likelihood estimate becomes better deﬁned and resulting maximum
likelihood estimation is plotted as the thin black line

approach are the same as for other closed population
models including (i) marks are not lost or overlooked
within a trip; (ii) there is no immigration to or emigration
from the study reach within a trip; (iii) mortality within a
trip is assumed to be zero; and (iv) recapture probability
of marked individuals is assumed to be equivalent to
capture probability of unmarked individuals (Pine et al.,
2003).

RESULTS

The most-supported model in our AIC framework parameterized apparent survival by year, ignoring the effect of
ﬂow treatment. Ignoring this model (which did not explicitly test ﬂow treatments—the purpose of our experiment),
highlighted an important result that the steady ﬂow periods
resulted in a signiﬁcant negative effect on apparent survival
of juvenile humpback chub (Table 2). The 95% conﬁdence
intervals for apparent survival according to both the Phi
(ﬂow) and the Phi(mod_ﬂow) models did not overlap with
the estimate for apparent survival during ﬂuctuating ﬂows.
Models parameterizing apparent survival by time (Phi(t))
and by ﬂow treatment nested within each year
(Phi(ﬂow*year)) did not converge and were excluded from
consideration. However, the Phi(year) model that we
constructed to substantiate the potential results of the
Phi(ﬂow*year) was the most supported model overall
(Table I). This suggests that inter-annual variability
(especially considering the high 2011 discharges) likely
swamps the effect of ﬂow treatment on juvenile humpback
chub apparent survival (Table II), but overall our results suggest that juvenile humpback chub apparent survival declined
over our study period. Regardless of how apparent survival
was parameterized, capture probability varied independently
by trip in all but the least supported models.

Table I. The four best apparent survival models for juvenile
humpback chub in Grand Canyon from 2009–2012 ranked by
lowest AICc

Field results

Model name

Over the 3-year duration of this project, we captured and
VIE marked 5575 juvenile humpback chub. Average catch
per trip based on calendar year was 221, 331, 588 and 794
juvenile humpback chub for 2009, 2010, 2011 and 2012,
respectively. Twenty-seven percent (1479) of juvenile
humpback chub were recaptured at least once. Of
individuals marked in 2009 and 2010, 11% were not
recaptured for the ﬁrst time until they had spent at least
1 year at large.

Phi(year) p(t)
Phi(mod_ﬂow) p(t)
Phi(ﬂow) p(t)
Phi(.) p(t)
Other models

Copyright © 2015 John Wiley & Sons, Ltd.

AICc

ΔAICc

No. Par.

10 175.4
10 188.7
10 190.7
10 201.2
>10 225

0
13.3
15.3
25.7
>50

16
16
15
14
—

AICc, Akaike Information Criterion for ﬁnite sample size.
Model parameter estimation is symbolized as follows: constant (.), timespeciﬁc (t), annual (year), ﬂow-speciﬁc (ﬂow) and modiﬁed ﬂow-speciﬁc
(mod_ﬂow). Models that did not converge or had ΔAICc values >50 are
not displayed.
River Res. Applic. (2015)
DOI: 10.1002/rra

C. G. FINCH ET AL.

Table II. Daily apparent survival estimates for juvenile humpback chub in Grand Canyon from 2009–2012
Model Parameterization
2009
2010
2011
Fluctuating
Both
Steady

Phi(year)

Phi(mod_ﬂow)

Phi(ﬂow)

Phi(.)

99.90% (99.86–99.93)
99.79% (99.74–99.83)
99.67% (99.57–99.75)
—
—
—

—
—
—
99.92% (99.86–99.95)
99.84% (99.76–99.90)
99.52% (99.27–99.72)

—
—
—
99.90% (99.85–99.93)
—
99.44% (99.23–99.64)

99.83% (99.81–99.85)
99.83% (99.81–99.85)
99.83% (99.81–99.85)
99.83% (99.81–99.85)
99.83% (99.81–99.85)
99.83% (99.81–99.85)

Results are from the top four models ranked by ΔAICc (Table I). All four models parameterized capture probability (p) by time. Values in parentheses under
each estimate are 95% conﬁdence intervals.

Open model abundance estimate
Open population model capture probability estimates ranged
from 6% to 19% per trip with 95% conﬁdence intervals of
4–22%. Abundance estimates ranged from 615 to 2,802
humpback chub per km of study reach with 95% conﬁdence

intervals from 437 to 3551 individuals (Figure 4(A)) and
varied independent of ﬂow treatment. The capture probabilities used in this open-framework abundance estimation are
for all cumulative passes within a trip (both electroﬁshing
and hoopnets combined). In other words, they reﬂect the
probability of capturing an individual at least once on a trip
(no distinction between multiple recaptures with the same
gear) as opposed to capture probabilities from the closed
abundance estimates, which apply to a single pass of a
single gear. These estimates also include individuals over
99 mm TL until they are ﬁrst PIT tagged. This is a larger size
range (and proportion of the juvenile humpback chub population) than considered in the closed model, which contributes to the difference in abundance trends between closed
and open models because the number of animals captured
in each event differs between the two model structures
(Figures 4(B) and 4(C)).
Closed model abundance of humpback chub 40–99 mm TL
(sequential Bayes’ framework)

Figure 4. Abundance of visible implant elastomer tagged juvenile

humpback chub per km with 95% conﬁdence by sample month
(month indicated by number above upper 95% conﬁdence interval)
2009–2012 in the nearshore ecology sampling reach in the
Colorado River using three different models. Tagged individuals
were included from when they were ﬁrst VIE-tagged (40-99 mm TL)
until they were ﬁrst PIT-tagged (100+ mm TL). Estimates in
Panel A are from a Cormack-Jolly-Seber model based on data
from both hoopnet and electroﬁshing gear types. Panel B abundance and 95% credible intervals are estimated using a simple
sequential Bayesian closed capture model and electroﬁshing
data only. Upper credible interval estimates for July 2012
(Panel B) extend off graph to a value of 9542. Panel C
abundance and 95% credible intervals are estimated from
hoopnet data only, using the same sequential Bayesian closed
capture model as Panel B
Copyright © 2015 John Wiley & Sons, Ltd.

Using closed population models for juvenile humpback
chub 40–99 mm TL, the Bayesian point estimate of our
capture probability ranged from 0.05% to 14% per pass with
either gear. Corresponding closed population estimates per
km of study reach using electroﬁshing recaptures ranged
from 116 to 1515 with 95% credible intervals of 45–3817
(Figure 4(B)). Closed-population abundance estimates based
on hoopnets per km ranged from 94 to 1124 humpback chub
with 95% credible intervals of 26–2695 (Figure 4(C)). In
general, closed estimates of abundance by gear type were
lowest during 2009–2010 and increased beginning in
September 2011–July 2012. There was also no apparent relationship between closed model abundance estimates and
ﬂow experiments (Figures 4(B) and 4(C)).

DISCUSSION
Juvenile humpback chub abundance demonstrated no appreciable relationship with ﬂow treatment, while apparent
River Res. Applic. (2015)
DOI: 10.1002/rra

DAM OPERATION EFFECTS ON FISH

survival declined during steady ﬂows. However, the effect
of ﬂow type (steady vs ﬂuctuating) was dwarfed by annual variation in apparent survival, as the Phi(year) p(t)
model was better supported by over 13 ΔAICc points than
either model considering ﬂow (Table I). Importantly, the
structure of the ﬂow model is similar to the annual model
(it is weighted towards 2011) because three of the ﬁve
steady ﬂows intervals that we considered were from the
last year (July 2011–July 2012); thus, it is not completely
clear if apparent survival of juvenile humpback chub declines with time, by ﬂow type or possibly by ﬂow magnitude (Figure 2). What is clear is that steady ﬂow
experiments, including one lasting over 9 months, did not
improve apparent survival of juvenile humpback chub.
Whether the declining trends in apparent survival according to the Phi(year) p(t) model represent long-term shifts
or natural temporal ﬂuctuations in demographic rates is
not yet known.
This result differs from the expectation that steady ﬂows
would improve demographic rates by stabilizing available
habitats and potentially reducing energetic costs and predation risks associated with juvenile ﬁsh movement. Several
limitations of the experimental design may have contributed
to this counterintuitive outcome. Because the experimental
manipulations were constrained within the conﬁnes of current dam operation, large contrasts and replication were difﬁcult. In addition, the regulatory need to release excess
water downstream to Lake Mead in summer 2011 (Figure 2)
because of wet conditions in the upper basin usurped our
third steady ﬂow treatment. Serendipitously, because these
unexpected water releases were steady, 2011 allowed us to
estimate apparent survival during summer steady ﬂows in
addition to fall and escape the perfect confounding of ﬂow
treatment with seasonality. Yet apparent survival still declined in 2011, which had the warmest (Finch et al., 2013)
and steadiest (Figure 2) conditions we observed. It is possible that the higher discharges or longer experimental duration from 2011 could alter our capture efﬁciency, but
humpback chub catches in 2011 were higher than the previous 2 years combined. This leads us to conclude that apparent survival, not capture probability, was primarily affected
by equalization ﬂows, which was corroborated by our simulation models.
One of the postulated beneﬁts of stabilizing nearshore
habitats under steady ﬂow operations was increase of primary and secondary production, primarily from backwater
habitats (Grand et al., 2006; USDOI, 2008), but also from
cobble or larger substrates (Blinn et al., 1995). This
increased production is thought to beneﬁt ﬁsh because of
the potential for food limitation, evidenced by high
consumption rates of ﬁsh within this section of river (Cross
et al., 2013). However, the fall timing of the steady ﬂow
experiments usually corresponded with monsoonal ﬂooding
Copyright © 2015 John Wiley & Sons, Ltd.

in the Little Colorado River (LCR), which increased turbidity and potentially decreased autochthonous production
within the mainstem. Finch et al. (2013) attributed at least
part of the lack of response in humpback chub growth to
the steady ﬂow experiments occurring during fall, when
primary productivity is generally declining throughout the
system. The authors also documented a similar counterintuitive response where growth rates of juvenile humpback
chub were actually higher during ﬂuctuating ﬂows compared with steady ﬂow periods (Finch et al., 2013). While
the reasons for this difference in growth between the two
ﬂow regimes are unknown, potential factors include increased availability of invertebrate drift as a prey resource
during ﬂuctuating ﬂows compared with steady ﬂows.
Previous investigations have demonstrated that in some
river systems, juvenile or larval ﬁsh that often occupy
shallow water habitats are most affected by elimination of
ﬂuctuating ﬂows (Freeman et al., 2001; Gaudin, 2001).
Fluctuating ﬂows affect these small life stages because of
factors such as stranding (Halleraker et al., 2003), downstream displacement (Young et al., 2011) and dewatering
effects on nest survival (McMichael et al., 2005; Grabowski
and Isely, 2007). However, the pre-dam Colorado River
experienced the lowest baseﬂows of the year in fall, at a time
when monsoon ﬂoods from the Little Colorado River and
other tributaries could more than double the discharge
almost instantaneously. It is likely that the same behaviour
that allowed young humpback chub to historically persist
in this dynamic environment likewise allows them to avoid
stranding and dewatering effects due to dam operations,
leading to the apparent lack of improvement in apparent
survival (this paper) and growth (Finch et al., 2013) from
the ﬂow experiment as designed. Additionally, the lack of
a ﬂow-related improvement in apparent survival and abundance could be associated with the persistence of suitable
shoreline habitat under both of the ﬂow regimes observed.
Korman et al. (2004) examined the amount of suitable
shoreline habitat under different dam operations and concluded that although ﬂuctuations decreased the amount of
persistent shoreline habitat, similar amounts of each habitat
type were still available over a large range of ﬂow conditions. The relative rarity of low-angle shorelines in our study
reach likely contributed to this homogeneous habitat
availability across the range of ﬂows that we observed. This
ﬂow insensitivity was corroborated by Alvarez and
Schmeekle (2012), who demonstrated that our study reach
(just below the Little Colorado River conﬂuence) was
among the most insensitive to discharge ﬂuctuations of the
entire Grand Canyon corridor.
Backwaters are one habitat type present in Grand Canyon
that are sensitive to temperature and discharge ﬂuctuations
because they are shallow and generally associated with
ephemeral sand bars (vs bedrock or cobble bars); the
River Res. Applic. (2015)
DOI: 10.1002/rra

C. G. FINCH ET AL.

suspected relationship between backwaters and juvenile
humpback chub prioritizes backwater construction and
retention as part of GCDAMP (USDOI, 2008; Melis et al.,
2011). However, Dodrill et al. (2014) found that while backwaters often contain higher densities of juvenile humpback
chub, the total abundance of humpback chub using backwaters in our study area is low because backwaters are locally
short-lived and rare. A change that only occurs within a
small proportion of the total population would be difﬁcult
to reasonably detect. Additionally, the low steady ﬂow
experiments occurred when air temperatures, solar insolation and warming rates of nearshore environments are lower
(September and October of each year; Yard et al., 2005;
Korman et al., 2006), which was corroborated when Ross
and Grams (2013) did not ﬁnd warmer water nearshore
compared with offshore during fall steady ﬂows. Even if
air temperatures were conducive to nearshore warming
(June and July), the turbulent, well-mixed hydraulics of
our study reach could have prevented the localized warming
of microhabitats that is observed in calmer river reaches like
Lee’s Ferry (Korman et al., 2006).
Ecological factors that could lead to changes in survival
over this time include potential changes in prey resources,
density dependent effects or predation. In March 2008, a
high ﬂow experiment was conducted from Glen Canyon
Dam that triggered several important ecosystem responses.
Survival rates of juvenile rainbow trout in the Lees Ferry
reach below Glen Canyon Dam (approximately 100-km upstream of our study site) increased substantially following
the ﬂood in 2008, and their survival rates remained higher
than normal even a year later in 2009 before declining to
the preﬂood levels in 2010 (Korman et al., 2011; Melis
et al., 2011). This improvement in rainbow trout survival
led to a subsequent increase in abundance of older rainbow
trout in 2009–2011 and may have resulted in higher levels
of predation or competition with native ﬁsh including humpback chub. Second, shifts in aquatic invertebrates were also
observed in response to the 2008 ﬂood event, with large declines in New Zealand mudsnail Potamopyrgus antipodarum
and amphipod Gammarus lacustris production and large
increases in drift invertebrate families Chironomidae and
Simuliidae. Cross et al. (2013) also found high interaction
strengths between ﬁsh and prey because of low secondary
production of prey items combined with very high consumption of available items by ﬁsh. The vast majority of humpback chub spawning occurs in the Little Colorado River
about 100-km downstream of the Lees Ferry reach, so
changes in spawning substrate at Lees Ferry are unlikely to
beneﬁt humpback chub spawning. However, it is possible
that juvenile humpback chub in the mainstem Colorado
River did beneﬁt from increases in the invertebrate community in 2008 and 2009 before these benthic and drift species
declined in 2010.
Copyright © 2015 John Wiley & Sons, Ltd.

Abundance was not related to apparent survival, because
the highest abundance of juvenile humpback chub that we
observed corresponded with the lowest apparent survival,
from July 2011 to July 2012. This may simply be because
of higher juvenile humpback chub production in the Little
Colorado River that year (Van Haverbeke et al., 2013),
accompanied by lower survival or higher emigration of
these juveniles in the mainstem Colorado River. Low apparent survival in 2011 may also come from density dependent
competition that existed in the Colorado River, as prey consumption rates have previously been documented to equal or
exceed invertebrate production in this reach (Cross et al.,
2013), while at the same time invertebrate drift was declining from the levels observed following the 2008 experimental ﬂood (Cross et al., 2011).
It is also possible that individual juvenile humpback
chub survived but eluded capture, causing a downward bias
in parameter estimates of apparent survival. Because apparent survival includes emigration (apparent survival = 1-true
survival-emigration), humpback chub in our study reach
may have emigrated downstream or upstream where we
were not permitted to sample. However, our observed
movement patterns for juvenile humpback chub between
our study reaches within a year suggest restricted movement patterns and small home ranges (Pine et al., 2013;
Gerig et al., 2014).
Another potential mechanism for the low apparent
survival in 2011 is direct or indirect effects of predation by
rainbow trout Oncorhynchus mykiss, brown trout Salmo
trutta or even older humpback chub. Although humpback
chub are not highly piscivorous (Kaeding and Zimmerman,
1983), they do consume ﬁsh, and the increasing abundance
of adults (Van Haverbeke et al., 2013) may have depressed juvenile survival through predation or competition;
if true, we see no clear management necessity or solution.
A more likely scenario was described by Yard et al.
(2011), who documented signiﬁcant piscivory near the
conﬂuence of the Colorado and Little Colorado Rivers
from both brown and rainbow trouts. This mortality does
not include any negative effects on humpback chub from
the risk of predation such as changes in habitat use
(He and Kitchell, 1990; He et al., 1993). Following the
experimental removal of non-native ﬁsh in this study reach
(Coggins et al., 2011), juvenile native ﬁsh abundance
increased; however, the mechanisms are still unclear
(Coggins et al., 2011). In our study, catch rates of rainbow
trout increased in each year of our ﬁeld work, increasing
overall by a factor of four. Electroﬁshing capture probabilities of rainbow trout in this reach are relatively consistent
across turbidity levels (J. Korman, unpublished data) so it
is likely that our catch rate trends are reﬂective of actual
trout abundance in our study reach. The interaction
between trout and humpback chub abundance and survival
River Res. Applic. (2015)
DOI: 10.1002/rra

DAM OPERATION EFFECTS ON FISH

in the mainstem Colorado River and the timing of humpback chub migration to the mainstem Colorado River
(Yackulic et al., 2014) are important areas for continued
research and experimental management.
ACKNOWLEDGEMENTS

This paper was developed as part of the ‘Nearshore
Ecology Project’ funded by US Bureau of Reclamation to
the US Geological Survey Grand Canyon Monitoring and
Research Center and the University of Florida. We would
like to thank our many cooperators including Navajo
Nation Department of Fish and Wildlife, United States
Fish and Wildlife Service, Arizona Game and Fish Department and United States National Park Service for permitting
and providing technical and ﬁeld assistance. We are very appreciative of our many boatmen and volunteers who assisted
with this project, and we thank Humphrey Summit Support
and GCMRC logistics for their many hours of hard work to
make this project possible. We thank J. Hines for the assistance with the simulations and C. Walters and M. Allen for
reviewing the earlier drafts of this work. We thank the
University of Florida and the Florida Cooperative Wildlife
Research Unit for the administrative and technical support.
The use of trade, ﬁrm or corporation names in this publication is for informational use only and does not constitute an
ofﬁcial endorsement or approval by the US government.

REFERENCES
Ahrens RNM, Walters CJ, Christensen V. 2012. Foraging arena theory.
Fish and Fisheries 13: 41–59.
Alvarez LV, Schmeekle MW. 2012. Erosion of river sandbars by diurnal
stage ﬂuctuations in the Colorado River in the Marble and Grand
Canyons: full-scale laboratory experiments. River Research and Applications 29: 839–854.
Anderson MT, Woosley LH Jr. 2005. Water availability for the western
United States—key scientiﬁc challenges. U.S. Geological Survey Circular 1261, Denver, Colorado.
Blinn DW, Poff NL. 2005. Colorado river basin. In Rivers of North America,
Benke AC, Cushing CE (eds). Elsevier Academic Press: Burlington,
Massachusetts; 483–526.
Blinn DW, Shannon JP, Stevens LE, Carder JP. 1995. Consequences of ﬂuctuating discharge for lotic communities. Journal of the North American
Benthological Society 14: 233–248.
Bunt CM, Cooke SJ, Katopodis C, McKinley RS. 1999. Movement and
summer habitat of brown trout (Salmo trutta) below a pulsed discharge
hydroelectric generating station. Regulated Rivers: Research and
Management 15: 395–403.
Burgess OT, Pine WE III, Walsh SJ. 2013. Importance of ﬂoodplain connectivity in the Apalachicola River, Florida. River Research and Applications 29: 718–733. DOI: 10.1002/rra.2567
Burnham KP, Anderson DR. 2002. Model Selection and Multimodel Inference. Springer: New York, New York.
Choudhury A, Hoffnagle TL, Cole RA. 2004. Parasites of native and
nonnative ﬁshes in the Little Colorado River, Grand Canyon, Arizona.
Journal of Parasitology 90: 1042–1052.

Copyright © 2015 John Wiley & Sons, Ltd.

Clarkson RW, Childs MR. 2000. Temperature effects of hypolimnialrelease dams on early life stages of Colorado River Basin big-river ﬁshes.
Copeia 2000: 402–412.
Coggins LG Jr, Walters CJ. 2009. Abundance trends and status of the Little
Colorado River population of humpback chub: an update considering data
from 1989–2008. US Geological Survey Open-File Report 2009–1075.
Coggins LG Jr, Pine WE III, Walters CJ, Van Haverbeke DR, Ward D,
Johnstone HC. 2006a. Abundance trends and status of the Little
Colorado River population of humpback chub. North American Journal
of Fisheries Management 26: 233–245.
Coggins LG Jr, Pine WE III, Walters CJ, Martell SJD. 2006b. Agestructured mark-recapture analysis: a virtual-population-analysis-based
model for analyzing age-structured capture–recapture data. North
American Journal of Fisheries Management 26: 201–205.
Coggins LG, Yard MD, Pine WE III. 2011. Nonnative ﬁsh control in the
Colorado River in Grand Canyon, Arizona: an effective program or
serendipitous timing? Transactions of the American Fisheries Society
140: 456–470.
Converse YK, Hawkins CP, Valdez RA. 1998. Habitat relationships of
subadult humpback chub in the Colorado River through Grand Canyon:
spatial variability and implications of ﬂow regulation. Regulated Rivers:
Research and Management 14: 267–284.
Cross WF, Baxter CV, Donner KC, Rosi-Marshall EJ, Kennedy TA,
Hall RO, Wellard-Kelly HA, Rogers RS. 2011. Ecosystem ecology
meets adaptive management: food web response to a controlled ﬂood
on the Colorado River, Glen Canyon. Ecological Applications 21:
2016–2033.
Cross WF, Baxter CV, Rosi-Marshall EJ, Hall RO Jr, Kennedy TA, Donner
KC, Wellard-Kelly HA, Seegert SEZ, Behn KE, Yard MD. 2013.
Foodweb dynamics in a large river discontinuum. Ecological Monographs 83: 311–337.
Dodrill MJ, Yackulic CB, Gerig B, Pine WE III, Korman J, Finch C. 2014. Do
management actions to restore rare habitat beneﬁt native ﬁsh conservation?
Distribution of Juvenile native ﬁsh among shoreline habitats of the
Colorado River. River Research and Applications. DOI: 10.1002/rra.2842
Dutterer AC, Mesing C, Cailteux R, Allen MS, Pine WE III, Strickland PA.
2013. Fish recruitment is inﬂuenced by river ﬂows and ﬂoodplain inundation at Apalachicola River, Florida. River Research and Applications
29: 1110–1118 DOI: 10.1002/rra/2604
Fette M, Weber C, Peter A, Wehrli B. 2007. Hydropower production and
river rehabilitation: a case study on an alpine river. Environmental
Modeling and Assessment 12: 257–267.
Finch C, Pine WE III, Limburg KE. 2013. Do hydropeaking ﬂows alter
juvenile ﬁsh growth rates? A test with juvenile humpback chub in the
Colorado River. River Research and Applications. DOI: 10.1001/rra.2725
Freeman MC, Bowen ZH, Bovee KD, Irwin ER. 2001. Flow and habitat
effects on juvenile ﬁsh abundance in natural and altered ﬂow regimes.
Ecological Applications 11: 179–190.
Gaudin P. 2001. Habitat shifts in juvenile riverine ﬁshes. Archiv für
Hydrobiologie - Supplement 135: 2–4.
Gazey WJ, Staley MJ. 1986. Population estimations from mark-recapture
experiments using a sequential Bayes algorithm. Ecology 67: 941–951.
Gerig B, Dodrill MJ, Pine WE III. 2014. Habitat selection and movement of
adult humpback chub in the Colorado River in Grand Canyon, Arizona,
during an experimental steady ﬂow release. North American Journal of
Fisheries Management 34: 39–48.
Gorman OT, Stone DM. 1999. Ecology of spawning humpback chub, Gila
cypha, in the Little Colorado River near Grand Canyon, Arizona.
Environmental Biology of Fishes 55: 115–133.
Grabowski TB, Isely JJ. 2007. Effects of ﬂow ﬂuctuations on the spawning
habitat of a riverine ﬁsh. Southeastern Naturalist 6: 471–478.
Grand T, Railsback CSF, Hayse JW, LaGory KE. 2006. A physical habitat
model for predicting the effects of ﬂow ﬂuctuations in nursery habitats of

River Res. Applic. (2015)
DOI: 10.1002/rra

C. G. FINCH ET AL.

the endangered Colorado pikeminnow Ptychocheilus lucius. River
Research and Applications 22: 1125–1142.
Gunderson L, Light SS. 2006. Adaptive management and adaptive governance in the everglades ecosystem. Policy Sciences 39: 323–334.
Halleraker JH, Saltveit SJ, Harby A, Arnekleiv JV, Fjeldstad HP, Kohler B.
2003. Factors inﬂuencing stranding of wild juvenile brown trout (Salmo
trutta) during rapid and frequent ﬂow decreases in an artiﬁcial stream.
River Research and Applications 19: 589–603.
He X, Kitchell JF. 1990. Direct and indirect effects of predation on a ﬁsh
community – a whole-lake experiment. Transactions of the American
Fisheries Society 119: 825–835.
He X, Wright RA, Kitchell JF. 1993. Fish behavioral and community
responses to manipulation. In The Trophic Cascade in Lakes, Carpenter
SR, Kitchell JF (eds). Cambridge University Press: Cambridge; 69–84.
Kaeding LR, Zimmerman MA. 1983. Life history and ecology of the humpback chub in the Little Colorado and Colorado rivers of the Grand
Canyon. Transactions of the American Fisheries Society 112: 577–594.
Kearsley LH, Schmidt JC, Warren KD. 1994. Effects of Glen Canyon dam on
Colorado River sand deposits used as campsites in Grand Canyon National
Park, USA. Regulated Rivers: Research and Management 9: 137–149.
Kennedy TA, Gloss SP. 2005. Aquatic ecology: the role of organic matter
and invertebrates. In The State of the Colorado River Ecosystem in Grand
Canyon, Gloss SP, Lovich JE, Melis TS (eds). United States Geological
Survey Circular 1282, Geological Survey: Reston, Virginia; 87–101.
Korman J, Campana SE. 2009. Effects of hydropeaking on nearshore
habitat use and growth of age-0 rainbow trout in a large regulated river.
Transactions of the American Fisheries Society 138: 76–87.
Korman J, Wiele SM, Torizzo M. 2004. Modelling effects of discharge on
habitat quality and dispersal of juvenile humpback chub (Gila cypha) in
the Colorado River, Grand Canyon. Regulated Rivers: Research and
Management 20: 379–400.
Korman J, Kaplinski M, Buszowski J. 2006. Effects of air and mainstem
water temperatures, hydraulic isolation and ﬂuctuating ﬂows from Glen
Canyon Dam on water temperatures in shoreline environments of the
Colorado River in Grand Canyon. Final Report to the Grand Canyon
Monitoring and Research Center, Flagstaff, Arizona.
Korman J, Yard M, Walters C, Coggins LG. 2009. Effects of ﬁsh size,
habitat, ﬂow, and density on capture probabilities of age-0 rainbow trout
estimated from electroﬁshing at discrete sites in a large river. Transactions of the American Fisheries Society 138: 58–75.
Korman J, Kaplinski M, Melis TS. 2011. Effects of ﬂuctuating ﬂows and a
controlled ﬂood on incubation success and early survival rates and
growth of age-0 rainbow trout in a large regulated river. Transactions
of the American Fisheries Society 140: 487–505.
Korman J, Martell SJD, Walters CJ, Makinster AS, Coggins LE, Yard MD,
Persons WR. 2012. Estimating recruitment dynamics and movement of
rainbow trout (Oncorhynchus mykiss) in the Colorado River in Grand
Canyon using an integrated assessment model. Canadian Journal of
Fisheries and Aquatic Sciences 69: 1827–1849.
Lauretta MV, Camp EV, Pine WE III, Frazer TK. 2013. Catchability model
selection for estimation the composition of ﬁshes and invertebrates
within dynamic aquatic ecosystems. Canadian Journal of Fisheries and
Aquatic Sciences 70: 381–392.
Limburg KE, Hayden TA, Pine WE III, Yard MD, Kozdon R, Valley JW.
2013. Of travertine and time: otolith chemistry and microstructure detect
provenance and demography of endangered humpback chub in Grand
Canyon (USA). Public Library of Science 8: e84235.
Lovich J, Melis TS. 2007. The state of the Colorado River ecosystem in
Grand Canyon: lessons from 10 years of adaptive ecosystem management. International Journal of River Basin Management 5: 207–221.
Marsh PC, Douglas ME. 1997. Predation by introduced ﬁshes on endangered
humpback chub and other native species in the Little Colorado River,
Arizona. Transactions of the American Fisheries Society 126: 343–346.

Copyright © 2015 John Wiley & Sons, Ltd.

McDonald TL, Amstrup SC, Regehr EV, Manly BF. 2005. Examples. In
Handbook of Capture–recapture Analysis, Amstrup SC, McDonald TL,
Manly BF (eds). Princeton University Press: Princeton; 196–262.
McMichael GA, Rakowski CL, James BB, Lukas JA. 2005. Estimated fall
Chinook salmon survival to emergence in dewatered redds in a shallow
side channel of the Columbia River. North American Journal of Fisheries
Management 25: 876–884.
Melis TS, Grams PE, Kennedy TA, Ralston BA, Robinson CT, Schmidt JC,
Schmit LM, Valdez RA, Wright SA. 2011. Three experimental high-ﬂow
releases from Glen Canyon Dam, Arizona-effects on the downstream
Colorado River ecosystem. United States Geological Survey fact sheet
2011–2012.
Minckley WL. 1991. Native ﬁshes of the Grand Canyon region: an obituary?
In Colorado River Ecology and Dam Management, Marzolf GR (ed.).
National Academy Press: Washington, DC; 124–177.
Mueller GA, Marsh PC. 2002. Lost, a desert river and its native ﬁshes: a
historical perspective of the lower Colorado River. United States Geological
Survey Information and Technology Report USGS/BRD/ITR 2002–0010,
Fort Collins, Colorado.
Murchie KJ, Hair KPE, Pullen CE, Redpath TD, Stephens HR, Cooke SJ.
2008. Fish response to modiﬁed ﬂow regimes in regulated rivers:
research methods, effects and opportunities. River Research and Applications 24: 197–217.
Nilsson C, Reidy CA, Dynesius M, Revenga C. 2005. Fragmentation and
ﬂow regulation of the world’s large river systems. Science 308: 405–408.
Otis DL, Burnham KP, White GC, Anderson DR. 1978. Statistical inference from capture data on closed animal populations. Wildlife Monographs 62: 3–135.
Palmer SC, Burbidge C, Patno H. 2004. The ﬁnancial impacts of the low
summer steady ﬂow experiment at Glen Canyon Dam. Western Area
Power Administration, June 2004. Available: http://www.gcmrc.gov/
library/reports/other/glencanyondam/palmer2004b.pdf (January 2013).
Pine WE III, Pollock KH, Hightower JE, Kwak TJ, Rice JA. 2003. A
review of tagging methods for estimating ﬁsh population size and
components of mortality. Fisheries 28: 10–23.
Pine WE III, Hightower JE, Coggins LG, Lauretta MV, Pollock KH. 2013.
Design and analysis of tagging studies. In Fisheries Techniques, 3rd edn,
Zale AV, Parrish DL, Sutton TM (eds). American Fisheries Society
Education Section Publication: Bethesda, Maryland; 521–572.
Poff NL, Allan JD, Bain MB, Karr JR, Prestegaard KL, Richter BD, Sparks
RE, Stromberg JC. 1997. The natural ﬂow regime. Bioscience 47: 769–784.
Poff NL, Bledsoe BP, Cuhaciyan CO. 2006. Hydrologic variation with land
use across the contiguous United States: geomorphic and ecological
consequences for stream ecosystems. Geomorphology 79: 264–285.
Pollock KH, Nichols JD, Brownie C, Hines JE. 1990. Statistical inference
for capture–recapture experiments. Wildlife Monographs 107: 3–97.
Ralston BE. 2011. Summary report of responses of key resources to the
2000 low summer steady ﬂow experiment, along the Colorado River
downstream from Glen Canyon Dam, Arizona. US Geological Survey
Open-File Report 2011–1220. Available: http://pubs.usgs.gov/of/2011/
1220/of2011-1220.pdf (January 2013)
Reisner M. 1986. Cadillac Desert: The American West and its Disappearing
Water. Viking: New York.
Richter BD, Thomas GA. 2007. Restoring environmental ﬂows by modifying
dam operations. Ecology and Society 12: 12.
Ross R, Grams PE. 2013. Nearshore thermal gradients of the Colorado
River near the Little Colorado conﬂuence, Grand Canyon National Park,
Arizona, 2010. USGS Open-File Report: 2013–1013.
Schmidt JC, Webb RH, Valdez RA, Marzolf GR, Stevens LE. 1998.
Science and values in river restoration in the Grand Canyon. Bioscience
48: 735–747.
Stanford JA, Ward JV. 1991. Limnology of Lake Powell and the chemistry
of the Colorado River. In National Resource Council (U.S.) Committee to

River Res. Applic. (2015)
DOI: 10.1002/rra

DAM OPERATION EFFECTS ON FISH

Review the Glen Canyon Environmental Studies. Colorado River
Ecology and Dam Management: Proceedings of a Symposium, May
24–25 1990, Santa Fe, NM. National Academy Press: Washington,
DC; 75–102.
Steele RJ, Smokorowski KE. 2000. Review of literature related to the
downstream ecological effects of hydroelectric power generation. Fisheries and Oceans Canada.
Stevens LE, Shannon JP, Blinn DW. 1997. Colorado River benthic ecology
in Grand Canyon, Arizona, USA: dam, tributary and geomorphical inﬂuences. Regulated Rivers: Research and Management 13: 129–149.
Stone DM, Gorman OT. 2006. Ontogenesis of endangered humpback chub
(Gila cypha) in the Little Colorado River, Arizona. The American
Midland Naturalist 155: 123–135.
Trammell M, Valdez RA, Carothers SW, Ryel R. 2002. Effects of a low
steady summer ﬂow experiment on native ﬁshes of the Colorado River
in Grand Canyon—ﬁnal report: Flagstaff, Arizona. SWCA, Inc. Environmental Consultants, submitted to U.S. Geological Survey, Grand Canyon
Monitoring and Research Center, cooperative agreement no. 99–FC–40–
2260, 77 p.
US Endangered Species Act of 1973, as amended, Pub. L. No. 93–205, 87
Stat. 884. Dec. 28, 1973. Available: http://www.fws.gov/endangered/esalibrary/pdf/ESAall.pdf (January 2013).
USBR (United States Bureau of Reclamation). 1996. Record of decision, operation of Glen Canyon Dam, ﬁnal environmental impact statement, Appendix
G. Available: https://www.usbr.gov/uc/rm/amp/pdfs/sp_appndxG_ROD.
pdf (October 2014).
USBR (United States Bureau of Reclamation). 2013. Glen Canyon Dam
adaptive management program. Available: http://www.usbr.gov/uc/rm/
amp/ (December 2013).
USDOI (United States Department of the Interior). 2008. Final biological
opinion for the operation of Glen Canyon Dam. Phoenix, Arizona,
U.S. Fish and Wildlife Service, AESO/SE 22410-1993-F-167R1, 88
p. Available: http://www.usbr.gov/uc/envdocs/bo/FinalGCDBO2-2608.pdf (January 2013).
Valdez RA, Ryel RJ. 1997. Life history and ecology of the humpback chub
in the Colorado River in Grand Canyon, Arizona. In Proceedings of the
Third Biennial Conference on the Colorado Plateau, National Park
Service, Flagstaff, AZ; 3–31.

Copyright © 2015 John Wiley & Sons, Ltd.

Van Haverbeke DR, Stone DM, Coggins LG, Pillow MJ. 2013. Long-term
monitoring of an endangered desert ﬁsh and factors inﬂuencing population dynamics. Journal of Fish and Wildlife Management 4: 163–177.
Walters CJ, Juanes F. 1993. Recruitment limitation as a consequence of
natural selection for the use of restricted feeding habitats and predation
risk taking by juvenile ﬁshes. Canadian Journal of Fisheries and Aquatic
Sciences 50: 2058–2070.
Williams BK, Johnson SA, Wilkins K. 1996. Uncertainty and the adaptive
management of waterfowl harvests. The Journal of Wildlife Management
60: 223–232.
Williams BK, Nichols JD, Conroy MJ. 2002. Analysis and Management of
Animal Populations. Academic Press: San Diego, California.
World Commission on Dams (WCD). 2000. Dams and development:
a new framework for decision making. The Report of the World
Commission on dams. Available: http://www.internationalrivers.org/
ﬁles/world_commission_on_dams_ﬁnal_report.pdf (September 2011).
Yackulic CB, Yard MD, Korman J, Van Haverbeke DR. 2014. A quantitative life history of endangered humpback chub that spawn in the Little
Colorado River: variation in movement, growth, and survival. Ecology
and Evolution 4: 1006–1018.
Yard MD, Bennett GE, Mietz SN, Coggins LG Jr, Stevens LE, Hueftle S,
Blinn DW. 2005. Inﬂuence of topographic complexity on solar insolation
estimates for the Colorado River, Grand Canyon, AZ. Ecological Modelling 183: 157–172.
Yard MD, Coggins LG Jr, Baxter CV, Bennett GE, Korman J. 2011. Trout
piscivory in the Colorado River, Grand Canyon: effects of turbidity,
temperature, and ﬁsh prey availability. Transactions of the American
Fisheries Society 140: 471–486.
Young PS, Cech JJ Jr, Thompson LC. 2011. Hydropower-related pulsedﬂow impacts on stream ﬁshes: a brief review, conceptual model,
knowledge gaps, and research needs. Reviews in Fish Biology and
Fisheries 21: 713–731.

SUPPORTING INFORMATION
Additional supporting information may found in the online
version of the article at the publisher’s web site.

River Res. Applic. (2015)
DOI: 10.1002/rra