Trap evaluation to optimize wild pig removal efforts in remote settings

Introduction

Invasions by wild pigs (Sus scrofa) are underway on every continent except Antarctica (Barrios-Garcia and Ballari 2012; Hernández et al. 2018; Hegel et al. 2022), providing a source for zoonotic and livestock diseases, natural resource degradation, and destruction of agricultural and anthropogenic resources (Seward et al. 2004; Barrios-Garcia and Ballari 2012; Bevins et al. 2014; Snow et al. 2017). Wild pigs are highly adaptable and are expected to expand their range unless an aggressive and coordinated approach is used to suppress their population growth, emphasizing the need to develop and utilize more efficient control strategies (Snow et al. 2017: Hegel et al. 2022). Although wild pigs are challenging to manage across their introduced range, remote locations where wild pigs are often abundant are particularly challenging, especially in densely forested, highly diverse landscapes that characterize many island ecosystems (Hess et al. 2020; Risch et al. 2022). Additionally, damage caused by wild pigs to island ecosystems is often extensive considering the lack of co-existence with large herbivorous or omnivorous mammals prior to the introduction of wild pigs and other ungulate species (Nogueira-Filho et al. 2009; Wehr et al. 2018; Risch et al. 2021). Several factors, including limited accessibility and resource availability complicate wild pig management efforts in remote island settings (Anderson and Stone 1993; Bevins et al. 2014).

Lethal control of wild pigs is currently the most efficient way to reduce populations and mitigate damage (Davis et al. 2018; Snow et al. 2020; Kilgo et al. 2023; Treichler et al. 2023). Trapping followed by euthanasia is commonly used for wild pigs and an array of trap types exist (Wight and Boughton 2018; Ditchkoff and Bodenchuk 2020; Gaskamp et al. 2021; Lavelle et al. 2024). Capture effectiveness varies by trap type, including box traps that typically capture one or two pigs at a time, to corral traps that are large enough to capture entire sounders of wild pigs, although are labor and resource intensive (Williams et al. 2011; Long and Campbell 2012; Gaskamp et al. 2021; Kilgo et al. 2023; Snow et al. 2024; Westhoff et al. 2024). Traditionally, trapping has been a passive process, with trap gates being triggered by wild pigs with little control of how many wild pigs were in the trap before being triggered (Gaskamp et al. 2021; Kilgo et al. 2023). Recently, active trapping has become more popular, including launching or dropping nets on demand to target specific individuals or groups of wild pigs (Torres-Blas et al. 2020; Gaskamp et al. 2021; Conejero et al. 2022; Escobar-González et al. 2024). Most recently, camera and cellular phone technology has facilitated the development of ‘smart’ traps to improve trapping efficiency (Wight and Boughton 2018; Kenkel et al. 2019; Gaskamp et al. 2021; Snow et al. 2024; Taylor et al. 2025). Trapping with smart traps is an active process where a trapper can observe wild pigs in real time and trigger the gate remotely to drop once all targeted individuals have entered (Gaskamp et al. 2021; Lewis et al. 2022; Kilgo et al. 2023; Taylor et al. 2025). Remote monitoring and trap triggering have also enabled trappers to maximize efficiency by striving for capturing entire groups or sounders of wild pigs, also known as ‘whole sounder removal’ (WSR; Lewis et al. 2022; Kilgo et al. 2023). Although this seems ideal, it oftentimes results in a time-intensive strategy because of required user monitoring and input (Snow et al. 2024; Taylor et al. 2025). Further, current smart trap functionality is limited by cellular service coverage in the area, and therefore is not currently possible in areas where cellular service is lacking (e.g. remote, heavily vegetated, and rugged island settings). Access to these remote areas may also be physically challenging, thereby limiting the type or size of traps capable of being transported and utilized. Most recently, a highly portable, passive net trap (Pig Brig®, Field Engine Wildlife Research and Management, Moodus, CT, USA) has become available as another option for trapping wild pigs with potential to improve capture success (Jolly et al. 2024; Snow et al. 2024; Chalkowski et al. 2025; Taylor et al. 2025).

The Hakalau Forest National Wildlife Refuge (HFNWR), a remote and rugged area on the Island of Hawai‘i, was established to conserve endangered island ecosystems (Kendall et al. 2023). Conservation efforts have typically focused on mitigating the effects of wild pigs and other feral ungulates by removing and excluding them from fenced (i.e. protected) areas (Hone and Stone 1989; Katahira et al. 1993; Hess et al. 2006; Hess 2016; Wehr et al. 2018). Although exclusionary fencing is used to mitigate the impacts of wild pigs, fences are not entirely effective and routine maintenance and removal of invading wild pigs is necessary (Hone and Stone 1989; Katahira et al. 1993; Lavelle et al. 2011). Traditional means for controlling wild pigs within these fenced areas included hunting, snaring, and limited trapping, but more effective and efficient means are needed, especially in remote areas such as the HFNWR (Hone and Stone 1989; Hess et al. 2006, 2020; Wehr et al. 2018). Therefore, we evaluated three types of traps for removing wild pigs within a fenced management unit where limitations of cellular service precluded the use of smart trap technology. Our primary objective was to compare the effectiveness, or probability of capture and number of wild pigs per capture event and efficiency, or number of wild pigs caught per unit effort with corral traps, box traps, and novel passive net traps. Our goal was to inform and optimize strategies for wild pig trapping in remote locations with current and ‘not smart’ technologies.

Materials and methods

Study area

Our study was conducted within the Upper Maulua Unit of the HFNWR located on the east side of the Island of Hawai‘i (19°47′39″N, 155°19′14″W). The Upper Maulua unit was a fully fenced ~8.1 km² area in a tropical montane rainforest ranging from 1500 to 2000 m in elevation on the windward slope of the Mauna Kea volcano (Fig. 1). Climate data from 2002 to 2024 at the Hakalau Remote Automatic Weather Station (1950 m asl; Remote Automatic Weather Stations 2024) recorded a mean air temperature range of 10.1°C in February to 13.5°C in August, and an overall annual mean of 11.6°C. Mean monthly precipitation ranged from 64 mm in June to 238 mm in March, with a mean annual precipitation of 2066 mm. Although there were other free-ranging feral ungulates in the area (e.g. cattle, sheep), wild pigs were the only ungulates present within the fenced study area. From 2010 to 2015, 347 wild pigs were removed from the study area (Leopold et al. 2016), but management since then has been sporadic and wild pigs were present at unknown but suspected high densities. There were ~18.2 km of unmaintained roads in the unit, primarily along the boundary perimeter and across the middle, making travel difficult and limited to offroad and all-terrain vehicles (ATVs). Beyond existing roads, access was limited to foot travel and was difficult owing to dense vegetation, underlying basalt substrates, and steep topography. Cellular service was lacking in the Upper Maulua Unit during our evaluation; thus, use of smart traps was not an option. Additionally, lack of vehicular access in some areas limited which traps we could transport to certain sites.

Fig. 1.

Map of fenced study site for a comparative evaluation of effectiveness and efficiency of three trap types (passive net, corral, and box) on Hakalau Forest National Wildlife Refuge on the Island of Hawai‘i, USA, during 2021–2022. Coordinates: 19°52′25″ N 155°19′2″ W. Source data for map features are from ESRI ArcGIS Living Atlas (https://livingatlas.arcgis.com/en/home/); accessed 30 August 2024).

Trapping methodology

We used a total of 34 traps of three types, including corral, box, and passive net traps (Fig. 2). We had five circular corral traps ranging in diameter from 5 to 7 m, constructed of rigid 1.5-m-high wire mesh with 10 × 10-cm spacing and two constructed of 1.5-m-high rigid wire mesh with 5.1 × 10-cm spacing, secured to steel t-posts positioned every 1.5–3.0 m (Fig. 2a). All corral trap gates were set to be pig-activated by a trip wire that held the gate open until triggered. The trip wire height and tension could be adjusted to target specific sizes of wild pigs.

Fig. 2.

Types of traps (passive net, corral, and box) used in a comparative evaluation of effectiveness and efficiency during three removal efforts on Hakalau Forest National Wildlife Refuge on the Island of Hawai‘i, USA, during 2021–2022. Traps included (a) box traps, (b) passive net traps, and (c) corral traps. Photos: M. Lavelle.

We also evaluated nine small (0.8-m-tall × 0.9-m-wide × 1.5-m-long) and three large (0.9-m-tall × 1.4-m-wide × 2.4-m-long) box traps, all constructed of rigid-wire mesh with 10.2 × 10.2-cm spacing enclosed in a frame of 2.5-cm square steel tubing (Fig. 2b). Gates were guillotine-style consisting of similar framed rigid-wire mesh and were triggered by a trip wire extended to and across the rear of the trap. As with corral traps, the trip wire height and tension could be adjusted to target specific sizes of wild pigs. We also tested 15 Pig Brig passive net traps (Fig. 2c). The passive net traps were 6.1 m in diameter and 1.5 m high and were constructed by suspending the net within a circle of 9–10 equally spaced steel T-posts (1.8–2.1 m).

We implemented three removal efforts of wild pigs on 14 November−8 December 2021, 14 February−9 March 2022, and 11 April−8 May 2022, with the goal of removing as many wild pigs as possible during each removal effort. As a result of the removal efforts, the population was estimated to be reduced from 485 to 86, an approximate 82.3% decrease (Glow et al. 2023), on the basis of removal effort model estimates (Davis et al. 2022). All removal efforts were structured similarly and began with pre-baiting once a week for 2 weeks prior to trapping. We pre-baited with 23–46 kg of commercially available sweet feed consisting of corn, oats, and barley (COB; Producer’s Pride, Tractor Supply Co., Brentwood, TN, USA), dry COB, and/or locally sourced macadamia nuts (Macadamia spp.). We established a total of 53 bait and trap sites within 50 m of existing roads that were spaced an average distance of 247 m (s.e. = 12.63) apart. We attempted to ensure there was more than one trap available to every wild pig within the study area. With homeranges of wild pigs within the study site averaging 1.96 km² (Glow et al. 2023), all available wild pigs theoretically had multiple traps within their homeranges, increasing the probability that all were exposed to multiple opportunities for capture.

The corral traps and large box traps were present before the start of the study and remained in their pre-existing locations throughout the duration of the study. Locations for each passive net trap and small box trap were determined on the basis of accessibility (vehicle, ATV, or on foot), availability of a level and rock-free space, and to distribute trap types as evenly across accessible portions of the study area as possible, and where concentrated wild pig activity was evident (i.e. abundance of tracks, rooting, scat). T-posts for each passive net trap were installed prior to initiation of the first removal effort and left in place for ease of redeployment of nets throughout the study. We installed box traps and nets prior to each removal effort and, afterwards, retrieved and stored them until initiation of the next removal effort. To avoid accidental captures, all traps were secured open until we were ready to capture. Once most of the bait was consumed nightly within any specific trap, suggesting consistent-to-increasing visitation, traps were then set to capture. For passive net traps, we progressively lowered the net, starting at >1.5 m above the ground during pre-baiting. Following the 2-week pre-baiting phase and once visitation by wild pigs was evident, on the basis of bait consumption, we lowered the net to ~0.5 m for conditioning before fully lowering the net to ground level to set the trap to capture. After a successful capture event in a passive net trap, the net was returned to >1.5 m above the ground and progressively lowered as described above if additional wild pig visitation was evident.

Corral traps and the large box traps were reset after each capture and remained set each trap night for the entire duration of each removal effort. Additional bait sites were maintained so that if bait consumption ceased at a particular small box trap, we could relocate that trap to a bait site where consistent bait consumption was occurring. We checked trap status daily and refreshed bait to ensure that ≥11 kg was available nightly in each trap.

During all capture events, we completely enclosed traps with a fabric shroud as quickly and quietly as possible to reduce stress and movement of wild pigs within the trap and to facilitate precise shot placement for euthanasia (Lavelle et al. 2019). We euthanized the captured wild pigs via cranial gunshot with an integrally suppressed 0.22 long rifle firearm, following American Veterinary Medical Association guidelines (Leary et al. 2020). We collected demographic data and estimated ages of all wild pigs removed on the basis of tooth eruption (Halseth et al. 2018). All study procedures were approved by the USDA NWRC Institutional Animal Care and Use Committee (QA-3225).

Results

During our three removal efforts, we captured and euthanized a total of 435 wild pigs via trapping across 162 capture events, with 206, 79 and 150 wild pigs being removed during the first, second, and third removal efforts respectively (Table 1). We trapped over a total of 713 nights and averaged 0.29–1.15 (s.e. = 0.04–0.16) wild pigs captured in each trap per trap night. The mean number of wild pigs captured per capture event ranged from 1.60 to 3.92 (s.e. = 0.12–0.36), and we captured a maximum of 13 wild pigs in one capture event with passive net traps. As planned, we utilized a system of continuous pre-baiting at sites in addition to those that were initially outfitted with traps. Over the course of the removal efforts, we relocated two passive nets and more than five box traps to additional pre-baiting sites as visitation fluctuated.

We found that passive net traps were highly effective in that they had a higher probability of capture per trap night than did box traps (β = 0.61; 95% CI = 0.21–1.01), with a ~0.3 probability of capturing per trap night, significantly higher than the probability of capture in box traps (Fig. 3). We found no differences in the probability of capture between corral traps and passive net traps (β = 0.37; 95% CI = −0.09–0.85), and between corral traps and box traps (β = −0.24; 95% CI = −0.69–0.23). Passive net traps captured more wild pigs per capture event than did corral traps (β = 0.59; 95% CI = 0.31–0.90) or box traps (β = 0.91; 95% CI = 0.66–1.17). Model predictions indicated that passive net traps captured 1.77 times more wild pigs per capture event than did corral traps and 2.45 times more than did box traps. We found no difference in the count of wild pigs captured between corral traps and box traps (β = 0.32; 95% CI = −0.01–0.64). With an overall total of 70 captures involving single wild pigs, these captures consisted of 21, 50 and 64% solo captures in passive net, corral, and box traps respectively.

Fig. 3.

Model predictions and 95% prediction intervals for predicting the probability of capture per trap night, count of wild pigs captured per capture event, and count of wild pigs captured per effort hour during three removal efforts on Hakalau Forest National Wildlife Refuge on the Island of Hawai‘i, USA, during 2021–2022.

We spent a total of 389.1 person-hours, including 262.1 h building traps and 127.0 h deploying bait to trap wild pigs, resulting in ~1.11 wild pigs being captured per effort hour invested overall or ~53.67 mins of effort per wild pig captured. We found that passive net traps efficiently captured more wild pigs per effort hour than did box traps (β = 0.55; 95% CI = 0.08–1.05), with passive net traps capturing 2.38 times more wild pigs per effort hour. Corral traps were nearly as efficient as passive net traps (β = 0.27; 95% CI = −0.38–0.91), although captured a similar number of wild pigs per effort hour as box traps (β = −0.29; 95% CI = −0.94–0.33).

Overall, 56% (243 of the 430) wild pigs we removed were <12 months old (Table 2; Fig. 4). Further, the overall sex ratio was 1:1 for all trap types, although the ratio varied slightly among age classes by trap type. The majority (>0.50) of older (24–36 months and >48 months) wild pigs that were captured in corral traps and passive nets respectively, were females.

Fig. 4.

Age distribution by trap type (passive net, corral, and box) of all wild pigs removed during a comparative trap evaluation on Hakalau Forest National Wildlife Refuge on the Island of Hawai‘i, USA, during 2021–2022.

Discussion

We believe this is the first comprehensive evaluation comparing several commonly used types of wild pig-activated traps, including box and corral traps, against a new commercially available passive net trap. We found that all were valuable tools, with each serving a unique function in reducing wild pig populations in rugged and remote settings. We demonstrated that passive net traps were more effective than corral and box traps, as evidenced by an average of nearly 2–2.5-times the number of wild pigs removed per capture event with corral and box traps respectively. We also found passive net traps to be an efficient alternative to corral and box traps, although corral traps were similarly efficient. Overall, our calculated range of efficiency of ~1.11 wild pigs captured per effort hour was higher than other reported trapping efforts ranging from 0.14 to 0.9 (Gaskamp et al. 2021; Snow et al. 2024; Taylor et al. 2025). One difference in previously reported efficiencies of smart traps and this evaluation of wild pig-activated traps and passive net traps is likely to be due in part to the fact that other evaluations of efficiency included operator-activated traps, which rely on observation (remote via phone or direct) and can be extensive and burdensome (Gaskamp et al. 2021; Snow et al. 2024; Taylor et al. 2025).

Although not evaluated specifically in our study, passive net traps may be another valuable step towards accomplishing WSR, although without the additional need for time-intensive active observation and triggering by an operator (Snow et al. 2024; Taylor et al. 2025). Further, passive net traps improve the potential for achieving WSR by more effectively enabling all members of a sounder to enter the trap over time because there is essentially no endpoint associated with ‘closing the gate’ after which additional entries are not possible. The clustering of wild pigs, such as the presence of wild pigs in a trap, has the potential to attract additional wild pigs (Snow et al. 2021). Male wild pigs can be drawn to olfactory cues of receptive females (Choquenot et al. 1993; McIlroy and Gifford 2005) and females can be drawn to cues of males (Pearce and Hughes 1987; Dorries et al. 1991), potentially compounding the continuous catch nature of passive net traps further.

Nets were easily transported and installed by a single operator and an ATV, which was important considering limited accessibility throughout most of our study area. With passive net traps, once T-posts are initially installed, they effectively function as a permanent trap where nets are rapidly redeployed on already existing T-posts during successive removal efforts. Although we did not attempt it during our study, passive net traps can alternatively be suspended from existing trees at a trap site instead of from T-posts, which can further increase their portability and decrease costs and time for installation. Labor associated with the use of passive net traps is low owing to the ease of installation, although the training period to accustom wild pigs to enter does require several days and should not be rushed (Taylor et al. 2025). This increases both the number of visits by an operator as well as increased bait requirements (Taylor et al. 2025); yet, the level of efficiency of passive net traps was still high.

Although box traps were less effective and efficient than passive net traps, they still had several characteristics that made them valuable additions to our trapping strategy. They were portable and could be deployed or relocated easily by a single operator with an ATV. This enabled us to maximize captures by rapidly relocating box traps to active bait sites, which is especially important, given our short-duration removals. Once visitation waned at a trap site, rapidly relocating box traps to bait sites with consistent visitation was often successful, because we anecdotally observed quick (i.e. 1–2 nights following trap placement) habituation, followed by capture. Box traps were also advantageous for targeting single individuals, such as lone males or dispersing yearlings that may be more vulnerable to such traps (Escobar-González et al. 2024). The largest proportion of wild pigs removed across all trap types were <12 months old and this age demographic tends to be less risk averse than adults (Gray et al. 2020; Torres-Blas et al. 2020; Escobar-González et al. 2024).

Our overall equal sex distribution (215:215) of all wild pigs captured demonstrated that our capture strategy was not biased towards either sex. Generally, targeting adult females for removals is most efficient for establishing wild pig visitation to bait sites (Lavelle et al. 2018), and has the greatest effect on population control (Servanty et al. 2011; Pepin et al. 2017; Snow et al. 2025). Wild pigs in Hawai‘i have shown to have two birth pulses including one in October–November and another in February–March (Glow et al. 2023; Snow et al. 2024). This was also evident in our observation of capturing an abundance of young wild pigs shortly after farrowing, especially during our third removal effort when mobility and trap activity would likely have been increasing (Chinn et al. 2023; Glow et al. 2023; Snow et al. 2024). Ideally, timing of removals should be informed by these results and be scheduled to intensify before birth pulses, thus maximizing the effects by removing pregnant females before they give birth and while offspring from previous birth are accompanying the female (Pepin et al. 2017; Snow et al. 2024, 2025).

Our study had several limitations that should be considered for future management and research. First, incorporation of trail cameras used to document visitation to bait sites prior to, during, and after trapping would have enabled a comparison of the proportion of wild pigs captured from a sounder, which would assist in determining the extent WSR was achievable with passive or wild pig-activated traps. Second, this evaluation was limited by the availability of resources, including travel time and associated funds available, which determined the number, duration, and overall extent of removals we conducted. Next steps should focus on changes in effectiveness and efficiency over time as wild pig densities decline and potential for trap shyness increases (Fischer et al. 2020; Lewis et al. 2022; Kilgo et al. 2023; Snow et al. 2024). Last, exploring movement characteristics of wild pigs in remote and rugged areas would lead to a better understanding of optimal bait and trap site densities and distribution strategies, to maximize removal potential while minimizing overlapping effort (Saunders et al. 1993; Lavelle et al. 2017, 2018; McRae et al. 2020).

Conclusions

In rugged and remote island settings where resources, accessibility, and cellular service used to manage wild pig populations are often limited, optimizing removals by selecting the most effective and efficient tools is essential. Incorporation of multiple trap types allowed us to substantially reduce a population of wild pigs. Passive net traps were very portable and the passive, continuous capture style maximized capture potential throughout each trap night. Box and corral traps also captured wild pigs and small box traps could be rapidly redeployed during a removal to target additional wild pigs. In remote settings with difficult terrain and inability to utilize smart traps, we recommend passive net traps as the most efficient means of removing wild pigs. Portable box traps are also recommended for single animals and rapid deployment.