Laying the foundations for selective-fish guidance using electricity: multi-species response to pulsed direct currents

Introduction

With the global human population predicted to continue to grow, at least during the first half of the current century (Adam, 2021), fresh waters will be increasingly exploited, e.g., to generate energy and produce food. This will require the construction of large dams to supplement those that have already impacted nearly two-thirds of the world’s longest rivers (>1,000 km) (Grill et al., 2019) in the drive to further economic development (Shi et al., 2019). The rapid growth in exploitation of freshwater resources will cause further environmental and social-economic shocks associated with the modification and degradation of ecosystems and the services they provide (Kemp et al., 2022), e.g., through the disruption of fluvial connectivity and fragmentation of habitat. From the perspective of freshwater fisheries, dams impede movements between critical habitats of fish that might also be injured or killed if they enter intakes (e.g., to hydropower turbines or irrigation systems) (Kemp, 2015). The development of effective strategies to mitigate the negative impacts of river infrastructure is critical if environmental degradation is to be halted or reversed as society strives to meet sustainability targets (e.g., UN Sustainable Development Goals; United Nations, 2015).

As part of a programme to advance the sustainability of river engineering, fish passes and physical and mechanical screens can be built and installed at impoundments to preserve or restore migration routes and thus improve habitat connectivity (Clay, 1995) and prevent fish entering associated water intakes (Kemp, 2015). Behavioural stimuli such as acoustics (Jesus et al., 2019), light (Ford et al., 2018), bubbles (Flores Martin et al., 2021) and electric fields (Parasiewicz et al., 2016) have been tested as an alternative to (e.g., Noatch & Suski, 2012), or to enhance the effectiveness of (e.g., Deleau et al., 2020; Mussen & Cech Jr, 2019), physical and mechanical screens designed to block passage and/or guide fish away from dangerous areas.

Unfortunately, current mitigation strategies are not always fully effective, thus only providing partial solutions, referred to as “half-way” technologies by some (Brown et al., 2013). Furthermore, the mitigation can itself be damaging and have unforeseen consequences that are often overlooked or underappreciated (Mclaughlin et al., 2013). For example, fish passes that partially reconnect habitat critical for the completion of the life-cycle of desirable native species, such as those with high commercial, cultural, and conservation significance, may also facilitate range expansion of Aquatic Invasive Species (AIS) (Kerr et al., 2021). The introduction of AIS can have large negative consequences for recipient ecosystems, acting through predation (e.g., Weyl & Lewis, 2006; largemouth bass, Micropterus salmoides), parasitism (e.g., Patrick, Sutton & Swink, 2009; sea lamprey, Petromyzon marinus), resource competition (e.g., Baker & Levinton, 2003; zebra mussel, Dreissena polymorpha), habitat modification (e.g., Brown & Moyle, 1991; Sacramento squawfish, Ptychocheilus grandis), hybridisation (e.g., Ludwig et al., 2009; Siberian sturgeons, Acipenser baerii) and disease transmission (e.g., Alderman, Holdich & Reeve, 1990; signal crayfish, Pacifastacus leniusculus), as well as causing substantial economic impacts (Pimentel et al., 2000). Hence, trade-offs arise when mitigation strategies employed to benefit native species conflict with management decisions to control the spread of AIS (Mclaughlin et al., 2013), resulting in a “connectivity conundrum” (Zielinski et al., 2020). Consequently, there is a need to enhance conservation efforts that benefit native species, while reducing the risks posed by AIS. Developing selective environmental impact mitigation technologies could reduce the tensions between AIS control and native fish passage objectives (Rahel & McLaughlin, 2018).

Previous research to develop environmental impact mitigation technology and AIS control based on selecting specific traits exhibited by the target species has tended to focus on the use of physical structures to facilitate passage rather than guidance. Perhaps some of the earliest examples relate to sea lamprey control in the Laurentian Great Lakes (Zielinski et al., 2019). Multiple barrier designs, such as fixed-, seasonal- and adjustable-crest weirs and velocity barriers have been used to limit the movements of invasive sea lamprey during peak migration (McLaughlin et al., 2007), while allowing desirable families (such as the salmonids) to gain access to spawning streams. In another case, selective fish passage has been developed to conserve native Pacific lamprey (Entosphenus tridentatus) in the Pacific Northwest of the United States through the design and installation of lamprey passes that compensate for ineffective technical fishways designed for salmonids (Moser et al., 2011).

Behavioural deterrents designed to selectively block and/or guide fish away from dangerous areas to more preferred routes provide a management option to protect desirable species while deterring AIS (Noatch & Suski, 2012). Early examples of the use of low-voltage electric fields for sea lamprey control in the Great Lakes date back to the 1950s (Applegate, Smith & Nielsen, 1952; Erkkila, Smith & McLain, 1956; McLain, 1957). A more recent example of the use of electric barriers relates to that activated in the Chicago Sanitary and Ship Canal in 2002, the world’s largest such device, designed to prevent the interbasin transfer of AIS (in particular bighead carp, Hypophthalmichthys nobilis, and silver carp, Hypophthalmichthys molitrix), primarily from the Mississippi to the Great Lake catchments (Parker et al., 2015). Unfortunately, electric barrier and guidance devices too can have negative unintended environmental impacts, dating back to the early studies in which alternating current was used to block sea lamprey but resulted in excessive mortality of non-target species (Applegate, Smith & Nielsen, 1952; Erkkila, Smith & McLain, 1956). Despite improvements in understanding and design, this risk remains, as illustrated during the evaluation of the effectiveness of a portable seasonal electric barrier installed in a tributary of the Great Lakes for sea lamprey control, but that also blocked and killed hundreds of non-target fish (Johnson et al., 2021).

Even when electric barriers work, their efficiencies can be variable. Some demonstrate close to 100% barrier effectiveness (e.g., Swink, 1999 for invasive sea lamprey; Sparks et al., 2011 for common carp, Cyprinus carpio); whereas others show greater variability in efficiency if designed for guidance (e.g., Gosset & Travade, 1999, 5–28% for native Atlantic salmon smolts Salmo salar; Johnson & Miehls, 2013, 84% for sea lamprey). To advance selective electric barriers and fish guidance systems that respectively facilitate the control of AIS and conservation of native species, there is a need to develop fundamental knowledge of sensory capabilities (e.g., electrosensitivities) and behavioural response from a multi-species/fish community perspective. This includes quantifying intra- and interspecific response to electric fields exhibiting different characteristics, such as field strength, pulse frequency and width, that are known to influence the behaviour and physical condition (e.g., Larson, Meyer & High, 2014; Layhee et al., 2016) of different life-stages (e.g., Nutile, Amberg & Goforth, 2012 for zebrafish, Danio rerio, embryos; Miller et al., 2021; Miller, Sharkh & Kemp, 2022 for adult and juvenile European eel).

To inform the design of selective electric barriers, this experimental study used model species of high conservation concern (European eel) and those that are invasive (grass carp, Ctenopharyngodon idella, and common carp, Cyprinus carpio). The European eel is critically endangered throughout its range, having suffered declines in recruitment of 90–99% since the 1980s (ICES, 2017). Carp species represent a considerable risk in regions where they are non-native, e.g., through competition, habitat alteration, and spread of disease (Manchester & Bullock, 2000). Both species co-occur in many European rivers (e.g., Cooper & Wheatley, 1981; Lehtonen, 2002; Nunn et al., 2007; Nunn et al., 2011; Van der Veer & Nentwig, 2015). We quantified and compared threshold field strengths (V/cm) for three physiological responses (twitch, loss of orientation and tetany). The influence of the following parameters on threshold field strengths for the three physiological responses were assessed: (1) Pulsed Direct Current (PDC) waveforms differing with respect to pulse width and frequency (Objective 1); (2) species (grass and common carp) (Objective 2); and (3) family (cyprinid and anguillid) (Objective 3).

Materials & Methods

Experimental set-up

Experiments were conducted at the International Centre for Ecohydraulics Research (ICER) facility, University of Southampton, UK, using a clear glass (10 mm thick) walled rectangular tank (1.5 m long × 0.60 m wide × 0.23 m deep) (Fig. 1) (see Miller et al., 2021). Two aluminium plate electrodes (0.5 m wide × 0.35 m high × 2 mm thick) were placed at either end of the tank 1.42 m apart and an electrically insulating mesh screen (0.56 m wide × 0.23 m high × 2 cm deep, mesh diameter = 1 mm) was placed in front of each to prevent contact with the fish. The electrodes were connected to an ETS ABP-2 backpack electrofisher (ETS Electrofishing Systems LLC) modified as a pulse generator (200 W average output; 600 V/10 A maximum peak outputs), powered by a 12 V DC battery.

The electric field was mapped using a potential probe consisting of two-point conductors 27 mm apart connected to an oscilloscope (Gwinstek GDS-1052-U) via a differential probe module (Probemaster Model 4232). Measurements were taken in a grid at a spacing of 10 cm in the x and y direction and at two depths (5 and 10 cm depth from the water surface) to record peak-to-peak voltage. Electric field maps were generated for all output voltages and waveforms. Electric field strength was uniform across the tank and proportional to output voltage (i.e., for 7 V output voltage with 142 cm between electrodes: $\frac{7}{142}=0.05$ V/cm) (see Miller et al., 2021). Ambient water conductivity during mapping was 630 µS. cm⁻¹.

Four CCTV cameras (Swann 1080p; 1,920 ×1,080 pixel resolution) were used to monitor and record fish behaviour: two overhead (1 m directly above the tank rim), and two side-facing (34 to 39 cm perpendicular to the tank side). Two infrared lights (780–850 nm wavelength) were placed above the tank (70 cm from each camera) to provide additional illumination during periods of darkness.

Water (conditioned tap water) depth was maintained at 15 cm and obtained from the holding tank in which the experimental fish used that day were housed.

Experimental procedure

Trials using cyprinids were conducted from 10 August–4 September 2018. Grass carp are thought to have similar activity levels during the day and night (Mitzner, 1978), but common carp swimming activity is suggested to be higher at night (Rahman & Meyer, 2009), so trials were conducted between 17:00–06:00 to ensure direct comparison between species. Trials using adult European eel were conducted from 2–8 November 2017 between 17:00–02:00 (Miller et al., 2021) to replicate conditions experienced during the natural nocturnal downstream migration (Tesch, 2003). Ambient light levels during the trials were less than 0.01 lux.

The pulse generator was used to produce four different waveforms: (a) single pulse- 2 Hz, (b) double pulse- 2 Hz, (c) single pulse- 3 Hz and (d) double pulse- 3 Hz (Table 1). For the double pulse- 2 Hz and double pulse- 3 Hz waveforms, the time between the pulses in the set of two (i.e., pulse break) was 50 and 16.7 ms respectively. These waveforms were selected based on ethical considerations as PDC <15 Hz is suggested to reduce injuries in fish (Sharber et al., 1994). Comparisons between single and double pulse were performed as previous research has suggested this can elicit differences in behavioural responses (Bowen et al., 2003). Furthermore, similar frequencies to those tested in this study have also proved effective at guiding other species (2 Hz: white sturgeon, Acipenser transmontanus, Ostrand et al., 2009; 3 Hz: Coho salmon, Oncorhynchus kisutch, Raymond, 1956) and enabled a direct comparison with eel (Miller et al., 2021). To generate the desired field strength the input voltage on the pulser was divided by the distance between the electrodes and then verified using a custom-built probe connected to an oscilloscope (Gwinstek GDS-1052-U). Treatment waveforms were alternated across trials in a systematic design to ensure an even spread of treatments.

Table 1:

Characteristics of waveforms used to test electrosensitivity of cyprinid (grass and common carp) and anguillid (European eel) fish with sample sizes provided.

Waveform	Pulse width (ms)	Schematic (represents 1 s time frame)	Species tested and sample size (n)
Single pulse- 2 Hz	100		Grass carp (n = 15), Common carp (n = 15) and European eel (n = 17)
Double pulse- 2 Hz	50		Grass carp (n = 15), Common carp (n = 15) and European eel (n = 17)
Single pulse- 3 Hz	33.3		Grass (n = 15) and Common carp (n = 15)
Double pulse- 3 Hz	16.7		Grass (n = 15) and Common carp (n = 15)

DOI: 10.7717/peerj.17962/table-1

Prior to the start of each trial, a single individual fish was placed in the experimental area between the mesh screens (Fig. 1) and allowed to settle for 10 min (see Miller et al., 2021). This was followed by a 10 s control period (0 V/cm) and a 10 s treatment of 0.05 V/cm and subsequent 10 min recovery. The cycle of 10 s –10 s control-treatment followed by recovery was repeated with increasing field strength in increments of 0.05 V/cm until tetany was observed. The response (no response, twitch, loss of orientation, tetany) was recorded for each treatment interval.

Experimental tank temperature was measured before (mean temperature start [± SD] °C: cyprinid = 19.4 [± 0.78] °C, anguillid = 13.35 [± 0.61] °C) and after each trial (mean temperature end [± SD] °C: cyprinid = 19.5 [± 0.78] °C, anguillid = 13.44 [± 0.63] °C) and fish were weighed (mean mass [± SD] g: grass carp = 17.4 [± 5.3] g, common carp = 26.3 [± 7.45] g and anguillid = 334.4 [± 94.4] g) and measured when the trial was terminated (mean fork length [± SD] mm: grass carp = 110.3 [± 10.9] mm and common carp = 105.3 [± 16.7] mm; mean total length [± SD] mm: grass carp = 121.0 [± 10.8] mm, common carp = 118.4 [± 10.9] mm and eel = 558.7 [± 52.2] mm).

Results

Threshold field strengths for physiological responses across waveforms (Objective 1)

For grass carp, pulse width influenced the threshold field strengths for tetany (χ²(3) = 12.4, p = 0.006) but not twitch and loss of orientation (twitch: χ² (3) = 1.06, p = 0.79, loss of orientation: χ²(3) = 2.69, p = 0.44) (Fig. 2A). Post-hoc analyses revealed tetany was elicited at a lower field strength in the single pulse-2 Hz waveform treatment than the double pulse-3 Hz (Dunn’s Test: z = −3.21, p = 0.008). Pulse frequency did not affect threshold field strengths for any of the responses (twitch: χ²(1) = 0.22, p = 0.64, loss of orientation: χ²(1) = 0.15, p = 0.7 and tetany: χ²(1) = 1.57, p = 0.21).

For common carp, threshold field strengths for twitch, loss of orientation and tetany were not affected by pulse width (twitch: χ²(3) = 1.80, p = 0.62, loss of orientation: χ²(3) = 0.41, p = 0.94, tetany: χ²(3) = 4.55, p = 0.21) (Fig. 2B). Pulse frequency also had no effect on threshold field strengths for twitch (χ²(1) = 0.19, p = 0.67), loss of orientation (χ²(1) = 0.07, p = 0.79) or tetany (χ²(1) = 0.05, p = 0.82).

Threshold field strengths for physiological responses of two cyprinid species (Objective 2)

With the exception that higher field strengths were required to elicit loss of orientation for common carp than grass carp under the single pulse- 3 Hz treatment (χ²(1) = 5.51, p = 0.02), there was no difference in thresholds of twitch, loss of orientation or tetany under any of the treatments (p > 0.05) (Figs. 3A, 3B, 3C, 3D).

Threshold field strengths for physiological responses of carp and eel (Objective 3)

As there were no differences in the threshold responses between grass and common carp for the single pulse- 2 Hz and double pulse- 2 Hz (Figs. 3A, 3B), data was aggregated to compare carp (cyprinids) with European eel (anguillid) (Miller et al., 2021).

The field strength at which the eel exhibited threshold responses for twitch (single pulse- 2 Hz: χ²(1) = 4.34, p = 0.04; double pulse- 2 Hz: χ²(1) = 8.5, p = 0.004), loss of orientation (single pulse- 2 Hz: χ²(1) = 27.5, p < 0.0001; double pulse- 2 Hz: χ²(1) = 26.6, p < 0.0001), and tetany (single pulse- 2 Hz: χ²(1) = 28.9, p < 0.0001; double pulse- 2 Hz: χ²(1) = 28.1, p < 0.0001) was lower than for carp under both treatments (Figs. 4A, 4B).

Discussion

To inform the development of selective fish guidance systems using electric fields, this study determined the electrosensitivity to different pulsed direct current waveforms of two cyprinids, grass and common carp, and compared the results with those obtained for an anguillid species of high conservation concern, the European eel (Miller et al., 2021). For both cyprinid species the threshold field strengths at which three key physiological responses (twitch, loss of orientation and tetany) were elicited were largely unaffected by the electric field parameters (pulse frequency and width), with one exception; grass carp exhibited tetany at lower field strengths under the single pulse- 2 Hz than the double pulse- 3 Hz. Electrosensitivity was similar between cyprinids except in the single pulse- 3 Hz waveform treatment where loss of orientation occurred at a slightly higher field strength for common carp. The threshold field strengths for all physiological responses were higher for both cyprinid species (i.e., they exhibited lower electrosensitivity) than for adult European eel, presumably because of the latter’s longer body length. This study provides the foundations for future research to further develop selective guidance / deterrent systems using electric fields.

At field strengths higher than 0.4 V/cm both cyprinid species exhibited unvolitional tetany. At field strengths above this value effective guidance that requires the modification of a volitional response may be prohibited. Similar field strengths (0.2–0.4 V/cm) are known to be effective at inhibiting common carp movement in the laboratory (Kim & Mandrak, 2017). Conversely, in field application higher field strengths (0.79–0.91 V/cm) are employed to control cyprinid movement in the Chicago Sanitary Shipping Canal (Parker et al., 2015), with the justification that native species do not migrate through the area (Moy et al., 2011) and smaller silver carp are likely to have a lower electrosensitivity due to the voltage gradient measured across their body (anterior to posterior) (Dolan & Miranda, 2003; Parker et al., 2015). However, care must be taken when comparing results due to differences in the waveform parameters tested and environmental variables such as water conductivity as this will directly influence the power transferred to the fish.

Minimal differences in threshold field strengths of response were observed across waveforms with respect to pulse width and frequency. However, for grass carp, tetany occurred at lower field strengths in the single pulse-2 Hz treatment than for the double pulse-3 Hz waveform, presumably due to the longer pulse width in the former (100 ms compared to 16.7 ms). This would support observations related to current applications in which higher peak voltages (power) are needed to immobilise fish when shorter pulse widths are used in electric fishing (Dolan & Miranda, 2003). Furthermore, longer pulse widths reduce the probability of juvenile and adult rainbow trout (Oncorhynchus mykiss), passing an electric barrier (Layhee et al., 2016). Interestingly, pulse width did not have a similar and consistent influence on other responses, suggesting that the most extreme physiological response observed (tetany) is most sensitive to differences in pulse width.

Pulse frequency did not influence any of the threshold field strengths for the responses considered for either grass or common carp. The absence of an effect might be explained by the minimal difference in pulse frequency between the waveforms employed (2 vs. 3 Hz). We selected relatively low frequencies because: (1) previous observations indicate they are effective at modifying fish movements (e.g., Mesa & Copeland, 2009 for steelhead, Oncorhynchus mykiss and Pacific lamprey, Entosphenus tridentatus at 3 Hz and 2 Hz; Savino et al., 2001 for round Goby, Neogobius melanostomus, at 2 Hz); (2) the risk of injury is reduced compared to higher frequency waveforms (e.g., Culver & Chick, 2015; Miranda & Kidwell, 2010) and (3) low frequency waveforms can be generated using less power, under a fixed millisecond pulse width, and thus reduce economic costs and the carbon footprint compared to other guidance systems (Beaumont, 2016).

Both cyprinid species exhibited similar electrosensitivities with no differences in the response to the single pulse-2 Hz, double pulse-2 Hz and double pulse-3 Hz observed. However, the one exception was the loss of orientation exhibited by grass carp at lower field strengths than for common carp in the single pulse-3 Hz waveform treatment. Previous work indicates that the threshold field strength for immobilisation is negatively related to the length of the fish (Briggs et al., 2019). In this study, grass carp were longer than common carp, but the difference was slight (e.g., 5 mm in mean fork length). It may be possible that loss of orientation is more sensitive to differences in body length than the other physiological responses, but further work is needed to confirm this. An alternative explanation is that even among closely related species there is still the possibility for variation in physiological response under specific parameters that could result in slight differences in the effectiveness of different electric field treatments for selective guidance systems.

For all the physiological responses exhibited, lower threshold field strengths were required for adult eel than both cyprinid species (Miller et al., 2021). The most likely explanation is the considerably longer body length of the adult eel compared to carp resulting in a greater anterior-posterior voltage differential if parallel to field lines (Reynolds, 1996). The difference in electrosensitivity may also reflect physiological differences between the two families.

This study quantified the threshold field strengths at which physiological responses were exhibited for two carp species of considerable conservation concern due to their invasiveness, including in European Rivers (e.g., Lehtonen, 2002; Van der Veer & Nentwig, 2015), and compared these with previously collected data for European eel, a species that is critically endangered in Europe. Eel have a higher electrosensitivity than carp, while the two carp species largely responded to the electric field treatments in a similar way. The results may have value in informing the development of selective multi-species barriers and guidance devices that could employ electric fields where these species occur in sympatry (Haubrock et al., 2019) and negative impacts on eel might be expected. The differences in electrosensitivity exhibited between families is promising in application to advancing an integrated pest management programme in which there is an interest in the conservation of native species at risk of river infrastructure. Globally, this scenario applies to multiple species where both those that are invasive co-occur with those for which conservation programmes focus on re-establishing river connectivity. This approach will allow trapping of those deemed undesirable while facilitating the free movement of those of high conservation value. This study provides a first step in the design of future selective guidance systems. Field studies should be performed in the future to validate the results obtained and optimise the parameters of the electric fields used to account for site specific characteristics (e.g., water and sediment conductivity).

Supplemental Information