Effects of Electric Barrier on Passage and Physical Condition of Juvenile and Adult Rainbow Trout

Electric barriers have been used for nearly a century to inhibit the movement of targeted nuisance fish species (Baldwin 1921). However, the effects of electric barriers can be indiscriminant; nontarget species may also be affected (Ostrand et al. 2009). Barriers designed for movement inhibition of target species have been shown to also inhibit movement and cause stress, injury, and mortality to multiple species of nontarget fish (Swink 1999; Ostrand et al. 2009; Johnson et al. 2014). As a result, there is growing concern about the negative effects of electric barriers to nontarget species of concern (e.g., to migration patterns of native fishes such as salmonids, Fausch et al. 2009).

As electric barriers are increasingly used for fish control, the negative effects of these electric barriers are of concern for migratory fish. Movement among spatially discrete habitats is a life-history requirement for migratory fish during both juvenile and adult stages of development. In general, fish barriers can restrict or prevent the movement of migratory fish species into critical areas for development or spawning events and may greatly reduce genetic variation and fragment habitat (Wofford et al. 2005; Fausch et al. 2009). Further research into the ramifications of electric barriers to nontarget, migratory fish is needed to determine whether these structures impact critical spawning and rearing events by limiting migration to those areas or affect certain life stages such as juveniles (but see Roscoe and Hinch 2010).

The impact of electric barriers on nontarget, migratory native species is unknown. Concern over whether these structures limit critical spawning and rearing events, such as juvenile migrations, is growing. For example, south central Alaska managers are faced with the challenge of slowing the spread of invasive northern pike Esox lucius in the face of declining salmonid populations (Sepulveda et al. 2013) while maintaining safe passage of outmigrating salmonid smolts. Even if electric barriers can contain northern pike, they are likely to impede salmonid migratory events and induce physical injury.

The effects of electrical current on freshwater fishes have been well studied (Reynolds and Kolz 2012). However, the relationship between fish sizes, barrier configurations, and passage rates through electric barriers has received less attention, and little information exists on the implications of electrical barriers to fish in various life stages (but see Johnson and Miehls 2013; Johnson et al. 2014). The deterrence efficiency of an electric barrier is an interaction among multiple factors, including the electrical intensity (voltage gradient) of the field, waveform characteristics, type of barrier configuration (e.g., vertical electrodes [VEs] versus horizontal electrodes), fish size, physiology, and environmental conditions. Fish size is a critical factor affecting individual susceptibility to electricity (Reynolds and Kolz 2012). The total electrical intensity transferred to a fish increases with fish volume, and less electricity is often required to affect large-bodied fish relative to smaller bodied fish (Dolan and Miranda 2003; Reynolds and Kolz 2012). This inverse relationship suggests potential for electric barriers to effectively block large-bodied target fish while still allowing movement of smaller bodied fish, such as migrating juveniles. However, optimal electric barrier settings for blocking upstream movement of adult sea lamprey Petromyzon marinus (0.9–1.8 V cm⁻¹, 1.8-ms pulse width; Johnson et al. 2014) were similar to settings that inhibited passage and guided downstream movement of juvenile sea lamprey and nontarget juvenile rainbow trout (0.43–0.94 V cm⁻¹, 1.6-ms pulse width; Johnson and Miehls 2013). If electric barriers are to be used in systems with native fish species of concern, further research is needed to understand whether optimal electric settings for preventing passage of target fish affects passage of nontarget fish.

Rainbow trout is an important native migratory species to many regions of North America and is often used to assess the effects of fish barriers (Steel et al. 2004; Johnson and Miehls 2013; Johnson et al. 2014). Rainbow trout has also been introduced into drainages outside of its native range and is considered a nuisance species in many situations (Avenetti et al. 2006; Fausch et al. 2009). The range of electric barrier settings (voltage gradient, pulse width) that affect passage rates of rainbow trout is not known. Therefore, identifying barrier settings that influence passage rates of various life stages of rainbow trout may provide information for managers dealing with this species inside and outside of its native region.

The objectives of this study were to 1) identify voltage and pulse-width settings of a VE barrier that allowed passage of juvenile rainbow trout and inhibited adult rainbow trout passage; 2) assess whether behaviors such as unsuccessful attempts to pass the barrier (deflection) or loss of equilibrium after exposure to the barrier were related to voltage, pulse width, or fish size, or whether deflection affected fish passage; and 3) assess whether injury incidence varied with voltage, pulse width, fish size (fork length [FL]), or exposure time. This work expands on Johnson and Miehls (2013) and Johnson et al. (2014), studies that also evaluated the effects of a VE barrier on rainbow trout, but here the importance of varying standard pulse widths to fish passage and deflection using pulsed-direct waveform were investigated.

We conducted experiments at the U.S. Fish and Wildlife Service (USFWS) Bozeman Fish Technology Center, Bozeman, Montana, during January–February 2013. Experiments were conducted within a circular stream channel (20-m circumference, 1.2 m in depth × 1.55 m in width). We used a static environment (no flow) to assess passage and other behaviors, without the added effect of flow on behavior. However, static flows limit our study's applicability to field conditions in lotic environments.

The experimental channel was filled with water originating from cold and warm springs and circulated through a re-use system. Water depth was 0.9 m, ambient water conductivity was 408 μS cm⁻¹ (301–431 μS cm⁻¹), dissolved oxygen was 7.9 mg L⁻¹ (6.2–8.7 mg L⁻¹), and pH was 7.9 (7.72–8.01). Water quality was intended to mimic conditions of barrier applications in cold-water systems. The channel contained cobble substrate.

We used hatchery-reared juvenile (52–126-mm FL range) and adult (215–410-mm FL range) rainbow trout to represent nontarget and target subjects, respectively. These rainbow trout were held in circular tanks (1.9 m in diameter × 0.75 m in height) with recirculating water at a mean (range) temperature of 13.8°C (8.0–14.5°C) throughout the duration of the experiments. The care and maintenance of these rainbow trout were in compliance with Bozeman Fish Technology Center standards. Rainbow trout were tagged with 12-mm passive integrated transponder tags (Biomark, Inc., Boise, ID) 24–48 h before experiments. These tags allowed us to hold all tested fish in a shared circular tank, yet still determine whether injury incidence 0.5 h and 8 d postexposure varied with exposure conditions. Dip nets were used to minimize direct handling when transferring rainbow trout among circulating holding tanks, aquaria, buckets, and our artificial stream channel. Rainbow trout weight (grams) and FL (millimeters) were measured 0.5 h before and 8 d after experiments.

We tested rainbow trout passage through a pulsed-direct current VE barrier manufactured by Fishways Global, LLC (Livonia, MI) and Procom System (Wroclaw, Poland). The VE barrier was installed as a symmetrical vertical system. The barrier consisted of three electrode arrays that were placed across the width of the experimental stream channel. Each array was separated by 1.2 m of stream channel (Figure 1). The two outermost arrays were composed of three positive electrodes, each spaced 0.6 m apart, and the center array was composed of two negative electrodes spaced 0.7 m apart. This barrier design created an electrical field on either side of the center array. The electrodes were stainless steel pipes suspended from above the water surface and they extended to the bottom substrate. The barrier system was connected to a power supply with a 30-Hz pulsed-direct current output.

We assessed volitional rainbow trout movement in relation to the barrier as a function of voltage gradient, pulse width and FL. We chose a volitional approach because it provides insight about behavioral choices, whereas coercive methods (e.g., prodding) provide insight about physiological limits (Castro-Santos 2005). As such, results based on a volitional approach provide a conservative estimate of a rainbow trout's ability to pass through the barrier. We used three output voltages (51.4 ± 0.5 [mean ± SD], 69.0 ± 0.8, and 91.4 ± 1.7 V) at three pulse widths of 0.1 ms (0.3% duty cycle), 0.4 ms (1.3% duty cycle), and 0.7 ms (2.3% duty cycle) and a control (no electrical output), for a total of 10 electrical treatments (Table S1). Minimum and maximum output voltages and pulse widths were based on recommendations from Fishways Global, LLC, given the fixed width and conductive materials of the experimental stream channel. We then validated these recommendations using pilot trials, which identified VE settings where most juvenile and adult rainbow trout passed and where most juvenile and adult rainbow trout failed to pass. Intermediate settings were limited to the resolution of the VE barrier's dial tuning.

Eighty juvenile rainbow trout (98 ± 12 mm [mean ± SD]) and 53 adult rainbow trout (314 ± 31 mm) were exposed to the 10 barrier settings during each trial. Of the 53 adult rainbow trout, 28 were used twice and 25 were used once. Juvenile rainbow trout were used one time during the duration of the experiment because they were readily available; most adult rainbow trout were used twice because they were in limited supply. An individual juvenile or adult rainbow trout was randomly assigned to a single voltage × pulse width setting. A trial (experimental unit) consisted of all 10 possible voltage (0, 51, 69, and 91 V) × pulse width (0, 0.1, 0.4, and 0.7 ms) combinations randomly assigned to 10 juvenile rainbow trout or 10 adult rainbow trout. Eight trials were conducted on juvenile rainbow trout over three successive days. For adult rainbow trout, eight trials were conducted over 2 d, separated by 7 d of recovery. All adult rainbow trout were used once during four trials on day 1. These fish were then rested for 7 d before completing four additional trials on day 8.

We used an oscilloscope (TPS 2012, Tektronics, Beaverton, OR) and forked voltage probe (10-cm distance between probe ends) before each individual exposure to verify the extent and intensity of the electrical field and to measure voltage gradient (volts per centimeter). Voltage gradient was measured midway between the first positively charged array and the middle negatively charged array. To identify when a rainbow trout interacted with the barrier, the electric field extent was determined by detecting the electric field edge with the voltage probe.

Individuals were transferred from the holding tank to an aquarium (122 cm in width × 33 cm in depth × 51 cm in height), where we assessed physical condition, defined as an individual's general appearance and behavior. We noted any minor bruising, deformities, or lesions preexposure, and we only included individuals that demonstrated normal orientation and swimming behavior and responsiveness to stimuli (light probing on dorsal side). Rainbow trout were netted and then identified, weighed, and measured. Next, we placed the rainbow trout in 19-L buckets for 2 min before they were gently poured into the stream channel, 1 m away from the beginning of the electric field. An individual trial lasted for 5 min during which volitional behavior was observed and documented by one observer.

We recorded behavior in relation to the barrier as pass, deflect, or paralyzed. A pass referred to a rainbow trout moving completely through all three arrays. Deflect referred to a fish entering the electrical field, turning around, and leaving the barrier the same way it entered. Deflections or attempts to pass the barrier unsuccessfully could occur several times before successful passage; therefore, we recorded the total number of deflections during individual trials. Paralysis occurred when a fish entered the electrical field, lost equilibrium, and could not escape. If a fish passed the barrier, or was paralyzed, the barrier was immediately turned off and the fish was removed. If, after 5 min, a fish was only deflected or had not attempted a pass, the barrier was turned off and the fish was removed. We assessed physical condition, and mortality 0.5 h after exposure, and we also assessed individual weight, FL, physical condition, and mortality 8 d after exposure. Postexposure physical condition was assessed identical to preexposure physical condition.

To further test whether passage through the VE barrier can injure juvenile rainbow trout (objective 3), we passed individuals through the barrier for an increased exposure period. Based on data collected during experiment 1, most juveniles spent <15 s within the VE barrier. Therefore, we exposed juvenile rainbow trout to the electrical field for approximately 20 s to test an extreme. Individual juvenile rainbow trout (86.8 ± 13.0 mm [52–112 mm]) were exposed to the electrical field by using three voltage settings (51.4 ± 0.3, 69 ± 0.5, and 91.5 ± 0.6 V) at four pulse widths (0.1 ms [0.3% duty cycle], 0.4 ms [1.3% duty cycle], 0.7 ms [2.3% duty cycle], and 1.0 ms [3.3% duty cycle]) with controls (no electrical output). Because we observed few incidences of injury in experiment 1 (see Results) where pulse widths of 0.1–0.7 ms were used, we included a pulse width of 1.0 ms (3.3% duty cycle) in experiment 2. We conducted five replicates per treatment combination on a total of 65 juvenile rainbow trout; none of these fish were used in experiment 1. We used an oscilloscope before each individual exposure to verify the electrical field and voltage gradient. Individuals were placed in a mesh bag and pulled through the barrier at a constant velocity for a 20-s exposure time. Physical condition and mortality were assessed 0.5 h and 8 d after experiments. Bruising, bleeding, loss of equilibrium, or any other notable behaviors were recorded.

We used generalized linear mixed models to compare fish passage (binary response: pass or no pass) as a function of voltage, pulse width, and log FL (fixed effect terms). Experiment day and trial number were random effects. Voltage and pulse width were discrete variables, whereas log FL, experiment day, and trial number were continuous variables. We fitted generalized linear mixed models for random effects (experiment day and trial number) before fixed effects (pulse width, voltage, and FL). We ranked the global model and candidate models using Akaike's Information Criterion for small sample sizes (AIC_c; Burnham and Anderson 2002), where candidate models were all possible combinations of fixed effect terms (voltage, pulse width, and FL) with the random effects terms (experiment day and trial number). For the global and candidate models, we calculated change in AIC_c (ΔAIC_c) and AIC_c weights. We deemed models with ΔAIC_c < 4 as most supported (Burnham and Anderson 2002). All generalized linear mixed models were fitted with a binomial distribution by using the glmer function in lme4 package of R (R Studio, version 0.98.501; R Core Team 2014).

We did not include control levels (no electrical output) in generalized linear mixed models since it was not possible to compare 0 V cm⁻¹ to any other pulse width besides 0 ms; a voltage of 0.0 V cm⁻¹ is only encountered at 0.0 ms and no other pulse widths (e.g., 0.1 ms). We used analysis of variance (aov function in R) to test fish passage incidence between control and treatment voltages and between control and pulse widths. We used Tukey's honestly significant difference post hoc tests to make pairwise comparisons between voltage and pulse width levels (glht function in R). We used α = 0.05 as a significance threshold for all statistical analyses.

We used linear regression (lm function in R) to test whether the number of deflections (number of unsuccessful attempts to pass the barrier regardless of whether the individual ultimately passed the barrier) was a function of barrier settings (voltage or pulse widths) or FL. We then used a generalized linear model to test whether deflection rates and FL affected fish passage (glm function in R). We also used this generalized linear model to test whether the number of deflections affected whether a fish ultimately passed and to test for an interaction between deflection and FL on passage. Deflection rates and FL were log transformed before analyses. Control levels for pulse width and voltage were not included in these analyses since deflection could not occur when the barrier was off (control).

We also tested whether external signs of injury (binary response: bruising or no bruising) were related to barrier settings, deflections, or passages. We used an analysis of covariance to compare external signs of injury (binary response: bruising or no bruising) to voltage (0.24, 0.33, and 0.44 V cm⁻¹) and pulse width (0.1, 0.4, and 0.7 ms), deflection, and passage for rainbow trout in experiments 1 and 2. Rainbow trout FL was the covariate.

Finally, we tested the assumption that adult rainbow trout that were used twice provided independent results. We used analysis of covariance to test whether adult rainbow trout passage and physical condition on day 8 were a function of voltage, pulse width, number of deflections, passing, or bruising experienced on day 1 of experiments. Rainbow trout FL was the covariate.

Eighty juvenile rainbow trout and 40 adult rainbow trout were exposed to the barrier. Of these 120 rainbow trout, three juvenile rainbow trout (4%) and eight adult rainbow trout (10%) never approached the barrier's electric field. Four of these adult rainbow trout were control fish. These 11 rainbow trout were excluded from analyses.

Fish passage varied by pulse width (F_2,157 = 7.3; P < 0.01), but not by voltage (F_3,157 = 1.77; P = 0.15). The most supported models indicate a negative effect of increasing pulse width on rainbow trout passage (Table 1). Based on the most supported model, the probability of juvenile and adult rainbow trout passing the barrier at 0.1 ms was 43%, and the odds of passing at 0.1 ms were 2.7 times greater than at 0.4 ms (27% probability of passing) and 7.1 times greater than at 0.7 ms (12% probability of passing). There were two voltage × pulse width settings that inhibited 100% of adult rainbow trout: 0.33 ± 0.02 V cm⁻¹ and 0.4 ms (n = 7) and 0.32 ± 0.01 V cm⁻¹ and 0.7 ms (n = 6). Less than 30% of juvenile rainbow trout passed the barrier at these settings (Figure 2). The highest percentage of juvenile rainbow trout passing (63%) was at the lowest barrier setting (0.24 ± 0.01 V cm⁻¹ and 0.1 ms), at which 50% of adult rainbow trout also passed.

When comparing control levels to treatment levels, we found that passage was greater at 0 ms than at 0.7 ms (P < 0.01). Passage was similar for 0, 0.1, and 0.4 ms (P = 0.99). Results of post hoc tests also indicate that passage at 0 V cm⁻¹ did not differ from that at 0.24 V cm⁻¹ (P = 0.93), 0.33 V cm⁻¹ (P = 0.28), or 0.44 V cm⁻¹ (P = 0.38). When the barrier was turned off, five of eight juvenile rainbow trout (63%) and two of four adult rainbow trout (50%) swam through the entire length of the barrier.

Ninety-eight percent of juvenile rainbow trout (n = 70) that attempted to pass were deflected at least one time, whereas 84% of adult rainbow trout (n = 57) that attempted to pass were deflected at least one time (Table S1). Deflections increased with FL (β = −0.23, t = −2.8, P < 0.01) and rainbow trout exposed to 0.1 ms had fewer deflections than rainbow trout exposed to 0.4 ms (β = 0.25, t = 0.1, P < 0.05). We found no difference in deflections between 0.1 and 0.7 ms (β = 0.22, t = 1.8, P = 0.08) and deflections did not vary by voltage (β < 0.14, t < 1.11, P > 0.27). We also found that larger fish and fish that were deflected more by the barrier were less likely to pass (β = −1.1, z = −2.1, P < 0.05).

We observed little incidence of physical injury and no mortality for rainbow trout exposed to the VE. One juvenile rainbow trout (1%) and nine adult rainbow trout (16%) demonstrated loss of equilibrium 0.5 h after exposure (Table S1). External bruising was observed in 5 juvenile rainbow trout (7%) and 15 adult rainbow trout (21%) 0.5 h after exposure (Table S1). The sample size of bruised juvenile rainbow trout was too small to statistically evaluate. The number of rainbow adult trout bruised did not differ by voltage (F_2,60 = 0.34; P = 0.71), pulse width (F_2,60 = 1.31; P = 0.28), passage (F_1,64 = 0.33; P = 0.90), or the number of deflections (F_1,62 = 0.33; P = 0.57). Importantly, we also found no evidence that exposure of adult rainbow trout on day 1 influenced behavior of re-used adult rainbow trout on day 8. We observed seven adult rainbow trout with bruising on day 8 of the experiment (37% of rainbow trout, n = 7), but this was not correlated with voltage (F_3,12 = 0.06; P = 0.94), pulse width (F_2,13 = 1.9; P = 0.19), passing (F_1,15 = 0.21; P = 0.66), or number of deflections (F_1,15 = 0.05; P = 0.83) experienced on day 1 of experiments.

To test whether passage through the VE barrier caused injury to small rainbow trout, individuals were passed through the barrier for a prolonged exposure period (experiment 2). No juvenile rainbow trout died 0.5 h or 8 d after exposure. Four rainbow trout (6%) sustained external bruising at 0.34 ± 0.01 V cm⁻¹ and 0.1 ms (n = 1), 0.34 ± 0.01 V cm⁻¹ and 0.4 ms (n = 1), and 0.45 ± 0.01 V cm⁻¹ and 1.0 ms (n = 2).

The VE barrier system settings tested in this study did not inhibit all adult rainbow trout passage while still allowing all juvenile rainbow trout to pass. Although the VE barrier inhibited juvenile rainbow trout passage, it did not cause injury or mortality. We found that the probability of passing the barrier decreased as pulse width increased for juvenile and adult rainbow trout, regardless of FL. Output voltage had no effect on rainbow trout passage. Pulse widths and voltages used in the study were relatively low compared to settings tested in previous studies (Ostrand et al. 2009; Johnson and Miehls 2013; Johnson et al. 2014), yet the range of settings tested here inhibited passage of various sizes of rainbow trout (52–410 mm FL). However, higher incidences of juvenile passing occurred at lower pulse widths. Therefore, these VE barrier designs may be a practical approach for the containment of larger bodied, targeted cold-water fish species in systems where nontarget fish movement is not a concern. But in systems where nontarget fish movements are of concern, further refinement and continued testing of electric barrier settings are needed to ensure juvenile passage.

Johnson and Miehls (2013) and Johnson et al. (2014) evaluated the VE barrier as a tool to direct juvenile and adult sea lamprey into traps and also evaluated the effects of these barrier systems on passage of juvenile and adult rainbow trout. Not all juvenile rainbow trout were guided effectively through the barrier at optimal juvenile sea lamprey settings, and a proportion of those rainbow trout were paralyzed (Johnson and Miehls 2013). In addition, optimal barrier settings that inhibited passage of 100% of adult target lamprey also inhibited passage of all adult rainbow trout (Johnson et al. 2014). Direct comparisons between results of Johnson and Miehls (2013) and Johnson et al. (2014) and the results in this study are difficult. These referenced studies intended to use barriers as a guidance system for rainbow trout rather than a barrier, and they used a lower range of pulse widths than those in our study (Johnson and Meihls 2013, 1.6 ms; Johnson et al. 2014, dual-frequency pulsed-direct current waveform with 1.8 ms). Nevertheless, Johnson and Meihls (2013) and Johnson et al. (2014) and this study found barrier avoidance in both juvenile and adult rainbow trout. Voltage gradient may not have been a major factor in our study because we tested relatively low-voltage gradients (<0.45 V cm⁻¹) compared to most other electric barrier studies.

Body size is an important factor that influences susceptibility to electricity (Dolan and Miranda 2003; Reynolds and Kolz 2012). Increasing body size is positively linked to increasing incidence of spinal injuries and hemorrhaging after exposure to an electric field (Dolan and Miranda 2003). Pulsed-direct current and direct current waveforms have caused spinal injuries in juvenile salmonids (Atlantic salmon Salmo salar, Chinook salmon Oncorhynchus tshawytscha, and rainbow trout) and small-bodied species (slimy sculpin Cottus cognatus and blacknose dace Rhinichthys atratulus) with increased electrical intensities (voltage) or prolonged exposure (Clément and Cunjak 2010; Holliman et al. 2010). We observed little evidence of physical injury to rainbow trout that interacted with the VE barrier, even under prolonged exposure conditions. The bruising that was observed in adult rainbow trout was not linked to VE settings or passing through the barrier. Rather, bruises were more often observed in fish that came into direct contact with the electrodes rather than a particular setting or behavior. This kind of incidental injury may be minimized in lotic field applications, wherein current velocity may decrease the total time spent in the barrier and decrease the likelihood of coming in direct contact with the barrier (Johnson et al. 2014).

Behavioral responses to electricity have also been linked to fish size (Dolan and Miranda 2003). We observed that only three individual rainbow trout (two adults and one juvenile) were paralyzed within the barrier during experiments. We also observed more adult rainbow trout with loss of equilibrium than juvenile rainbow trout. These results suggest that acute exposure (20 s) may not elicit noticeable behavioral changes in juveniles. However, further research is needed to understand the long-term impacts of VE barrier exposure to growth and development of nontarget species or the long-term impacts barriers to migration and whether barriers cause movement inhibition to migratory species (Ainslie et al. 1998).

Since the impact of electric barriers to nontarget, migratory, native species is unknown, there is growing concern over whether these structures limit critical spawning and rearing events. In addition to impeding critical spawning and rearing events, electric barriers may also induce physical injury in nontarget fish. In these instances, electrical guidance systems may be a better option for preventing movement of invasive fish, while allowing for safer downstream passage by smolts. Field application of electric barriers should also consider the additional effect that flows may have on the ability of fish to move through barriers.

Please note: The Journal of Fish and Wildlife Management is not responsible for the content or functionality of any supplemental material. Queries should be directed to the corresponding author for the article.

Table S1. Trial conditions and responses of juvenile (98 ± 12 mm [mean ± SD] fork length) and adult (314 ± 31 mm fork length) rainbow trout Oncorhynchus mykiss exposed to the vertical electric barrier system at varying voltage gradients (volts per centimeter, mean ± SD) and pulse widths (milliseconds). Experiments were conducted at the U.S. Fish and Wildlife Service Bozeman Fish Technology Center in January–February 2013. Data include number of rainbow trout that passed (pass), number of rainbow trout deflections (deflected), mean (±SD) number of deflections (mean deflections), number of rainbow trout paralyzed (paralyzed), number of rainbow trout bruised (bruise), and number of rainbow trout with loss of equilibrium (equilibrium) at each of the 10 voltage × pulse width treatment combinations. The symbol “-” was used to indicate behaviors that could not occur when the barrier was off.

Found at DOI: http://www.nrmsc.usgs.gov/files/norock/research/Table_S1.docx. Also found at DOI: 10.3996/042015-JFWM-039.s1; (20 KB DOCX).

Table S2. Description of field heading in Tables S3 and S4.

Found at DOI: http://www.nrmsc.usgs.gov/files/norock/research/Table_S2.docx; Also found at DOI: 10.3996/042015-JFWM-039.s2; (16 KB XLSX).

Table S3. Movement, physical condition, and survival data for juvenile and adult rainbow trout Oncorhynchus mykiss exposed to a pulsed-direct current vertical electric barrier at varying voltage gradients (volts per centimeter) and pulse widths (milliseconds). Field headings are defined in Table S2. Experiments were conducted at the U.S. Fish and Wildlife Service Bozeman Fish Technology Center in January–February 2013.

Found at DOI: 10.3996/052015-JFWM-039.s3; Also found at DOI: 10.3996/042015-JFWM-039.s3; (36 KB XLSX).

Table S4. Physical condition and survival data for juvenile rainbow trout Oncorhynchus mykiss exposed for a prolonged duration (approximately 20 s) to a pulsed-direct current vertical electric barrier at varying voltage gradients (volts per centimeter) and pulse widths (milliseconds). Field headings are defined in Table S2. Experiments were conducted at the U.S. Fish and Wildlife Service Bozeman Fish Technology Center in January–February 2013.

Found at DOI: http://www.nrmsc.usgs.gov/files/norock/research/Table_S4.xlsx. Also found at DOI: 10.3996/042015-JFWM-039.s4; (17 KB XLSX).

Funding was provided by Alaska Sustainable Salmon Fund. We thank Robert Muth and Matt Toner of the U.S. Fish and Wildlife Service Bozeman Fish Technology Center for logistical support and equipment.

Any use of trade, product, or firm names is for descriptive purposes only and does not imply endorsement by the U.S. Government.