The main study area was a melon patch (about 8 ha) located immediately south of the Bison Range, about 3.5 km west of Ravalli, Flathead Indian Reservation, Montana, USA (Fig. 1). The melon patch was just north of MT Hwy 200, and just south of the Jocko River and the associated trees and shrubs in the riparian area. While the patch was dominated by different varieties of melons, there was also some corn planted along one edge of the field. An additional study site with a drive-over mat was located along US Hwy 93, just south of Ravalli (Fig. 1).

Figure 1.

The location of the melon patch, and the additional drive-over mat, Flathead Indian Reservation, Montana, USA.

A Non-Governmental Organization, People and Carnivores, built an electric fence around the melon patch in the summer of 2020 (Fig. 2). The fence was constructed before the harvest of the melons in 2020 with the intent of keeping black bears, and potentially also grizzly bears, out of the melon patch. In addition, electrified barriers were installed at 4 vehicle access points to the melon patch (Fig. 2). The electric fence consisted of 4 wires attached to composite fence posts made from polypropylene and wood (PasturePro®) (Fig. 3). This fence design is similar to those used by others to keep different bear species from accessing crops or other attractants (Huygens and Hayashi 1999; Otto and Roloff 2015; Khorozyan and Waltert 2020). Corner posts and braces, including at vehicle access points, were treated wood posts. The height of the fence was about 91 cm (Fig. 3). While the fence was designed to keep both black bears and grizzly bears out of the melon patch, it was not designed to keep other species from accessing the melon patch. Ungulates (e.g. white-tailed deer (Odocoileus virginianus), mule deer (O. hemionus), and elk (Cervus canadensis)) can easily jump this fence. The farmer sprayed herbicides along the fence to reduce voltage drop. At first, the 2^nd wire from the bottom was a designated ground wire, with the other 3 wires being hot (i.e., carrying current). After 27 August 2020, the 2^nd wire from the ground was made into a hot wire also, which meant that from then on, the current received by an animal depended on the contact points of the animal with the ground which varies with the conductivity of the animal itself, and the conductivity of the ground or vegetation. A solar panel, and associated battery and energizer powered the fence and all 4 access points. However, from 6 August 2021 onwards, access point 3 had the drive-over mat installed which was powered by its own solar panel and associated battery and energizer. The electric fence energizers are listed by the internationally recognized safety standards agencies (UL, CSA or IEC).

Figure 2.

The melon patch (roughly 450 m east-west and 180 m north-south). The electric fence (solid black line around the melon patch), the 4 vehicle access points, the 5 locations along the fence where fence peak DC voltage was measured, and the location of 3 trail cameras where we suspected black bears were digging under the fence.

Figure 3.

The approximate height of the non-modified electric fence and its 4 wires.

In 2020 and 2021, we evaluated 5 different electrified barrier designs at the 4 access points to the melon patch (Table 1, Fig. 4). The drive-through bump-gates were originally installed at 2 access points, but 1 of them was modified in 2020, and the drive-over wires barrier was replaced by a drive-over mat in 2021 (for exact dates see Table 1). In 2022, we continued to monitor only the swing gate at access point 2 and the drive-over mat at access point 3 (Table 1). We continued monitoring in 2022 to increase the sample size for the swing gate (from 9 in 2020-2021 to 23 in 2020-2022) and the drive-over mat (from 2 in 2021 to 3 in 2021-2022). Since the drive-over mat still had a very low sample size, data from a similar barrier design from the same manufacturer at a nearby site (US Hwy 93, just south of Ravalli) was added (Fig. 1). The electrified barrier along US Hwy 93 was monitored in 2022 and 2023 (Table 1). There were 2 black bears and 1 grizzly bear that approached this drive-over mat, bringing the total sample size for the drive-over mat design to 5 black bears and 1 grizzly bear (6 “bears” in total).

Figure 4.

The 5 electrified barriers at the 4 vehicle access points (a–e) and along US Hwy 93 (f) a The electrified swing gate (4.88 m wide, 1.37 m tall, about 18 cm gap between ground and bottom of gate). The wires are mounted at 15, 48, 97, and 140 cm above the ground. Installed at access point 2 b A drive-through bump-gate (not modified) (about 4.88 m wide, about 91 cm tall), with vertical electrified wires. The orange horizontal pole is metal and carries current. The green horizontal part is fiberglass and does not carry current. Installed at access points 1 and 4 c A modified drive-through bump-gate (with conductive netting) (about 4.88 m wide, about 91 cm tall), with vertical electrified wires and custom conductive netting (about 61 cm high) attached. Installed at access point 4 d The drive-over wires, about 4.34 m wide (post-post) and 3.10 m long. The 18 drive-over wires are about 18 cm above the ground and the gaps between the wires vary between 13-30 cm. There are “side-board” wires that angle toward the ground from the post along the sides of the barrier to reduce the likelihood of an animal bypassing the drive-over wires. However, these “side-boards” do not cover the full length of the barrier. Installed at access point 3 e The drive-over mat, about 4.34 m wide (post-post) and about 3.05 m long. On the far side, the habitat side, there is metal mesh on the ground, connected to a grounding rod (about 61 cm wide). This is followed by 8 sections of 25-28 cm wide expanded metal sheeting (alternating positive and negative (ground)) mounted on wooden planks. This drive-over mat is powered by its own solar panel, battery and energizer. Installed at access point 3 f The drive-over mat, about 7.36 m wide (post-post) and about 2.44 m long, installed at a gap in a wildlife fence for a low volume access road along US Hwy 93. The mat is positioned on a wildlife guard (bridge grate material). On the far side, the habitat side, the animals first encounter bridge grate material (about 66 cm wide) that is connected to a grounding rod. This is followed by a wooden ramp (about 20 cm wide) and 4 metal plates (each about out 51 cm wide, alternating positive and negative (ground)) mounted on rubber and wooden planks with about 6 cm gaps in between the metal plates, and another wooden ramp on the far side (about 20 cm wide). This drive-over mat is powered by its own solar panel, battery and energizer. Installed along US Hwy 93, just south of Ravalli (see Fig. 1).

The fence and the 5 electrified barriers at the 4 access points to the melon patch were modified during the study. The most important modifications to the electrified barriers, start and end dates of the melon picking seasons in 2020 and 2021, and the associated evaluation periods, are summarized in Table 2. Note that melons were also a potential attraction to animals before the first pick of a season and after the last pick of a season.

Each access point had a trail camera installed (Reconyx™ PC 900). The trail cameras fully covered the area up to 2 m in front of each access point. The 2 m distance from the access point was visible on each image based on the line between the trail camera’s viewpoint and a stick with reflective tape on the other end. This allowed us to consistently evaluate the behavior of large mammals that approached each access point within 2 m. We evaluated whether the animals succeeded in accessing the melon patch by crossing the electrified barriers. Some large mammals were also detected further away from the access point, but those animals were not included in the evaluation of the effectiveness of the electrified barriers at the 4 access points. We also detected or suspected that bears were digging under the electrified fence at 3 locations (Table 3, Fig. 2), and we monitored these locations with trail cameras. Interestingly, these locations were all adjacent to the riparian habitat along the Jocko River and not along the roadside of MT Hwy 200 or adjacent agricultural fields. We replaced the memory cards in all trail cameras about once a month. We replaced the batteries (Energizer® Ultimate Lithium™) about every 3 months.

We measured the peak DC voltage (Stafix digital volt meter) of the electrified fence on each of the 4 wires at 5 locations for most of the trail camera checks before, during, and after the melon harvest seasons in 2020 and 2021 (Figs 2, 3). We also measured the peak DC voltage of the electrified barriers at the 4 access points during most trail camera checks before, during, and after the harvest seasons in 2020 and 2021 (Figs 2, 4). For the bump-gate designs at access points 1 and 4, the peak DC voltage was measured for both the right and the left part of the gate. For the other barriers there was just 1 measurement. Though the peak DC voltage measurements only occurred during the trail camera checks - the voltage measurements were not continuous - the peak DC voltage readings showed whether the electric fence and electrified barriers were electrified at that moment. This indicated whether we could have expected the electric fence and electrified barriers at the access points to have discouraged bears from entering the melon patch since the last voltage readings.

When measured, the peak DC voltage on the fence was almost always 7–9 kV and very similar for the 4 fence wires (Fig. 5). The 2^nd wire from the bottom was a ground wire until 27 August 2020, hence the lack of voltage measurements before that date. In general, the 2 bump-gates (access points 1 and 4) had lower peak DC voltage (usually between 4–6 kV) than the barriers at the other 2 locations (usually between 7–10 kV) (Fig. 6). Note that the measurements in 2021 at access point 3 related to a drive-over mat with its own power source that was independent from that of the fence and the 3 other access points. Also note that the peak DC voltage at the end of November in 2021 was lower everywhere.

Figure 5.

The peak DC voltage on the 4 fence wires before, during, and after the harvest melon seasons in 2020 (point groups on the left) and 2021 (point groups on the right).

Figure 6.

The peak DC voltage at the 4 access points before, during, and after the melon harvest seasons in 2020 (point groups on the left) and 2021 (point groups on the right). The measurements in 2021 at access point 3 related to a drive-over mat with its own power source that was independent from that of the fence and the 3 other access points.

All the bears that were recorded at the melon patch were black bears; there were no observations of grizzly bears at this site. All the black bear events related to single animals; there were no events involving multiple black bears (e.g. a sow and cubs). Four out of the 5 electrified barrier designs for access points were an absolute (100%) or near absolute barrier (94.3%) for black bears (Fig. 7). However, the bump-gates that were originally designed for cattle were a poor barrier for black bears (48.4%). Based on the images from the trail cameras, the bears usually passed in between the vertical electrified strands, and thus minimized contact with the wires. After conductive netting was attached to the bump-gate at access point 4, the bears no longer passed through that bump-gate (100% barrier) and moreover, they did not even touch the electrified barrier. The same was true for the swing gate, drive-over wires, and drive-over mat; most bears that approached the barriers did not touch the electrified components at all. Two out of 35 black bears passed the drive-over wires above the ground (Fig. 7). Both intrusions involved a bear bypassing most of the wires by coming in or leaving from the side.

Figure 7.

The effectiveness of the different barriers in keeping black bears out of the melon patch. N = number of bears (sample size) approaching the barriers. ¹Sample size at melon patch was 3 black bears, but the data were supplemented by additional observations (2 black bears and 1 grizzly bear, all single animals) at a similar mat nearby (US Hwy 93, south of Ravalli, see Fig. 1).

At the 7 locations that were monitored with a trail camera, the number of black bear intrusions into the melon patch varied between 2.07 and 4.38 per day during the 2020 melon picking season (103 intrusions between 7 August 2020 and 2 October 2020) and was 0 per day during the 2021 melon picking season (0 intrusions between 30 July 2021 and 14 September 2021) (Fig. 8). Access point 4, the non-modified bump-gate adjacent to the Jocko River, and “fence west” accounted for most of the intrusions. After conductive netting was attached to the bump-gate at access point 4, black bear entries increased at “fence west” and at “fence middle” (Fig. 8). After fence modifications on 9 September 2020, almost no bears were able to enter the melon patch. There were only 2 black bears that entered after that date, and both did so through partially bypassing the drive-over wires at access point 3. We did not have any evidence that black bears entered the melon patch at any other location that was not monitored with a trail camera. The total number of black bears was highest in August through October (Fig. 9). The total number of black bear observations at any of the 7 locations monitored with a trail camera, regardless of the distance to the fence, vehicle access point or trail camera, and regardless of which side of the fence or electrified barriers the bears were on, was 95% lower in 2021 (N=24) than in 2020 (N=527).

Figure 8.

Black bear intrusions into the melon patch during different periods in 2020 and 2021. Note: No intrusions were reported at access point 2 (swing gate) and “fence east”. Note: Access point 3 had drive-over wires 10 July 2020 until 6 August 2021 and a drive-over mat after 6 August 2021.

Figure 9.

The total number of black bear observations, regardless of the distance to the fence, vehicle access point or trail camera, and regardless of which side of the fence or electrified barriers the bears were on, per month in 2020 and 2021.

In general, the peak DC voltage on the fence and at access points 2 and 3 was almost always 7–10 kV. However, the 2 bump-gates (access points 1 and 4) usually had lower peak DC voltage (usually between 4–6 kV), suggesting higher resistance of the materials or a short or voltage leak. The overall drop in peak DC voltage at the end of November in 2021 was most likely the result of moisture causing a voltage leak, or shorter days (not enough daylight to recharge the batteries) and lower temperatures (reduced capacity of the batteries). It is possible that there were additional voltage drops in between voltage readings, but if they were present at all, the peak DC voltage had recovered by the next voltage measurement.

Four out of the 5 electrified barrier designs for access points were an absolute or near absolute barrier for black bears. However, non-modified bump-gates that were originally designed for cattle were a poor barrier for black bears as they deterred only about half of the animals. Adding conductive netting to one of the bump gates made it into an absolute barrier, however. This is likely because the netting results in more contact points with an animal for a longer time when an animal tries to lift the fence material and pass under. Changing electrified wires to electrified netting also made a fence a much greater barrier to European wild rabbits (Oryctolagus cuniculus) (McKillop et al. 1992). The drive-over wires were partially bypassed by 2 bears, emphasizing the need for tight fences along the full length of the sides of the barrier. Interestingly, most bears that approached an absolute or near absolute barrier did not even touch the barrier. This is consistent with other studies that reported that black bears and other large mammal species tend to stay away from electrified barriers, apparently because they know about its potential impact (Huygens and Hayashi 1999; Fischer et al. 2011; Otto and Roloff 2015; Teixeira et al. 2017).

Through a step-by-step process, the weak points of the electrified barriers at the vehicle access points and the electric fence around the melon patch were addressed. The standard bump-gate at access point 4, and 2 fence locations (west and middle) were of particular concern. Interestingly, once the conductive netting was attached to the bump-gate at vehicle access point 4, the bears increasingly dug under the fence to enter the melon patch. However, from 9 September 2020 onwards, including during the melon picking season in 2021, the melon patch became almost inaccessible to black bears; the number of intrusions by black bear was reduced from 2.07–4.5 per day (2020 melon picking season) to 0 per day (2021 melon picking season). We found that black bears trying and succeeding to enter the melon patch at a particular location depended on how difficult it was to enter at other locations. This illustrates that fences and vehicle access points should be designed, operated, maintained, and monitored as a system rather than as individual features, regardless of whether the goal is to protect crops or to keep animals out of a fenced road corridor.

The total number of black bear observations at any of the 7 locations monitored with a trail camera, regardless of which side of the fence or electrified barriers the bears were on, was 95% lower in 2021 than in 2020. The farmer reported substantial reductions in melons lost to bears in 2021 compared to 2019 before the fence and electrified barriers were installed (100% reduction in losses; personal communication Cassie Silvernale). Although black bear activity in and adjacent to the melon patch was still relatively high in 2020, not nearly as much melon loss occurred because of bears in 2020 compared to 2019 (80% reduction in losses; personal communication Cassie Silvernale). The very substantial reductions in both black bear presence and melon losses after the installation of the fence and electrified barriers suggests that after the black bears were no longer able to enter the melon patch, they drastically reduced their presence and effort to try to access the melon patch. Apparently, the attraction of the melon patch and the associated habit of eating melons was broken. Note that based on direct observations of black bears by the authors, there was no indication of a population crash of black bears in 2021 in the immediate vicinity, or in the wider region (Montana Fish Wildlife & Parks 2024).

All 4 vehicle access points had manual switches allowing farm personnel to walk through the electrified barriers without shocking themselves. However, shortly after the electric barriers were activated, the farmer realized that the switches were sometimes accidentally left in the “off” position. From 29 August 2020 onwards, all switches were permanently taped in the “on” position. While relatively inexpensive, the bump-gates required custom conductive netting to become a substantial barrier to black bears. The netting is subject to tearing and needs to be adjusted and reattached regularly (e.g. with zip-ties). In addition, there are tensioners for the 2 horizontal poles of the bump-gates. These also need to be adjusted on a regular basis to ensure that the 2 horizontal poles align and do not leave a gap in the middle. The horizontal poles are also subject to breaking; one of the poles broke after it got stuck in a bumper or wheel well of a pickup truck.

There were no operation or maintenance issues with the electrified swing gate. A design problem of the drive-over wires barrier was that the side fences were too short. Both intrusions by black bears involved the animals bypassing most of the wires above the ground by accessing or leaving the barrier from one of the sides. Barriers or “side-fences” that run tight along the full length of the barrier would likely address this issue and force bears, if they try to access the crop, to walk on top of, or in between, the wires above the ground for the full length of the barrier. In contrast, the drive-over mat has full side barriers that have a tight connection to the mat, and they do run the full length of the mat. Here, no large mammals were able to bypass the mat by coming in from or leaving at one of the sides. We did observe that the drive-over mat can kill amphibians and small mammals. Between 6 August 2021 and 21 November 2021, we found 1 dead western toad (Anaxyrus boreas) and 1 dead deer mouse (Peromyscus maniculatus) on the mat. Such unintended side effects may be reduced through making the habitat immediately adjacent to the barrier less attractive or inaccessible to small species, e.g., through ABS screens attached to the side fences of the barriers. There could also be a sensor installed that would only turn the electricity on after a large animal has been detected that is approaching the barrier from the habitat side.

For locations that are accessible to the public, such as along highways, warning signs and “turn electricity off” buttons may be required. These buttons should be associated with a timer and an indicator light so that the electricity will automatically turn on again, e.g. after a minute or so, and that confirmation is visible to the public. Note that some of the tested electrified barrier designs are not a suitable barrier to ungulates (e.g. swing gate, bump-gates). Wider barriers such as the drive-over mat or, though to a lesser extent, drive-over wires, are or can be, at least in theory, substantial barriers to not only species with paws but also ungulates (Fischer et al. 2011; Gagnon et al. 2019).

The authors have declared that no competing interests exist.

No ethical statement was reported.

The research at the melon patch was funded by the following organizations: Alaska Department of Transportation and Public Facilities, Arizona Department of Transportation, California Department of Transportation, Iowa Department of Transportation, Michigan Department of Transportation, Minnesota Department of Transportation, Nevada Department of Transportation, New Mexico Department of Transportation, Ontario Ministry of Transportation, Oregon Department of Transportation, Parks Canada and Washington Department of Transportation through the Animal-Vehicle Collision (WVC) Reduction and Habitat Connectivity Transportation Pooled-Fund Project TPF-3 5(358). The research at the drive-over mat along US Hwy 93 was funded by the U.S. Department of Transportation, the Small Urban, Rural and Tribal Center on Mobility (SURTCOM), and the Montana Department of Transportation. Finally, we thank the US Fish and Wildlife Service for contributing funds for the electric fence.

Conceptualization: MPH. Data curation: SCG, MPH. Formal analysis: MPH. Funding acquisition: MPH. Investigation: MPH, SCG. Methodology: SCG, MPH. Project administration: MPH. Supervision: MPH. Validation: MPH, SCG. Visualization: MPH. Writing - original draft: MPH. Writing - review and editing: SCG, MPH.

Marcel P. Huijser https://orcid.org/0000-0002-4355-4631

Samantha C. Getty https://orcid.org/0000-0002-5610-0241

All of the data that support the findings of this study are available in the main text or Supplementary Information.