Influence of bait and habitat on site visitation by wild pigs ( Sus scrofa )

Abstract

BACKGROUND

Wild pigs (Sus scrofa), known for their impacts on ecosystems in both their native and invasive ranges, are commonly managed using lethal trapping or shooting methods reliant upon bait or scent lures. Previous studies evaluating the efficacy of attractants at improving wild pig visitations have yielded mixed results, which are likely to be a reflection of the generalist foraging strategies of wild pigs as well as their fine‐scale variability in space use.

RESULTS

We conducted experimental trials at 743 sites in South Carolina, USA, to quantify differences in wild pig visitation among a suite of bait, scent lures and bait + scent lure combinations to identify which maximized visitation. We monitored sites using a remote camera for 7 days and quantified visitation rates and time to visitation. Additionally, to identify habitat attributes that maximized wild pig visitation, we characterized a suite of habitat attributes at each location. Wild pigs visited more sites and had shorter times to visitation at sites with bait present (bait: 42.85%, 61.35 h; bait + scent: 47.99%, 60.98 h) than sites with scent lures alone (24.40%, 82.03 h), with similar results for groups of wild pigs and individuals, emphasizing the effectiveness of bait in increasing visitation. Our habitat modeling results suggest that wild pigs are more likely to locate bait sites in areas closer to water sources and in thick understory.

CONCLUSION

We recommend that managers maximize visitation of wild pigs by using high‐value baits, not relying on scent lures, and intentionally selecting sites based on habitat attributes where wild pigs have nearby access to water and cover. © 2025 The Author(s). Pest Management Science published by John Wiley & Sons Ltd on behalf of Society of Chemical Industry. This article has been contributed to by U.S. Government employees and their work is in the public domain in the USA.

Bait, a food item, is important for attracting wild pigs to a site for management and considering habitat, including distance to water and understory, increases likelihood of wild pig visitation.

1. INTRODUCTION

Over the last few decades, the geographic distribution and abundance of wild Sus scrofa (wild pigs in their invasive range and wild boar in their native range) ¹ have substantially increased, driven by changing human land use and climate, ² , ³ , ⁴ , ⁵ as well as through illegal translocations in their introduced range. ⁶ , ⁷ Growing populations of wild S. scrofa across their native and invasive ranges have led to increased destructive impacts to both native and anthropogenic habitats. ⁸ In particular, where wild pigs have been introduced and considered invasive, they outcompete native species, damage crops and habitat, pose a disease risk, and overall, negatively impact the ecosystem. ⁸ , ⁹ , ¹⁰ , ¹¹ , ¹² , ¹³ , ¹⁴

In order to reduce or eliminate their negative impacts, wild pigs and wild boar are managed extensively throughout their invasive and native ranges, respectively. Where they are invasive, typically, the goal is population reduction or eradication, which is often completed by lethal control through trapping and ground or aerial shooting. ¹⁵ Additionally, orally‐delivered toxic baits have been ¹⁶ , ¹⁷ and continue to be developed for reducing populations of wild pigs. ¹⁸ , ¹⁹ To utilize these lethal control methods, managers must attract wild pigs to sites and rely on attractants to do so including bait (food reward item) and scent lures (nonfood scent). Wild pigs are opportunistic omnivores that consume crops, grasses, legumes, hard mast, fungi, invertebrates, amphibians, reptiles, mammals and birds, amongst other items, ²⁰ , ²¹ and thus their diet and response to baits can vary spatially and temporally depending on local availability. ²² , ²³ , ²⁴ Whole kernel corn is commonly used by trappers as a bait owing to its attractiveness to wild pigs and its wide availability. ¹⁵ , ²³ , ²⁵ , ²⁶ However, given their wide dietary breadth, a variety of baits have been utilized to attract wild pigs, although previous research has shown mixed results in how different baits perform in attracting wild pigs. ²³ , ²⁷ , ²⁸ , ²⁹

In addition to bait, scent lures are commonly used to attract wild pigs to a site as a consequence of their advanced olfactory system. The olfactory bulb makes up 7% of a pig's brain size, ³⁰ compared to 0.01% in humans. ³¹ This advanced olfactory system plays an important role in their social interactions, breeding and foraging behaviors. ²³ , ³² In particular, wild pigs rely on their keen sense of smell to locate food ³³ as much of their diet is acquired from underground food sources such as invertebrate larvae, small fossorial vertebrates, tubers, rhizomes, corms and bulbs. ²⁰ , ³³ , ³⁴ In addition to food‐based lures, sow estrous scent has been shown to be effective at attracting wild pigs when population density is low, or when animals have become trap‐shy, ³⁵ although Choquenot et al. (1993) found that female estrous scent was not an effective lure for trapping wild pigs.

Despite scent being thought of as an important factor in attracting wild pigs, there have been mixed results in the performance of scent lures among studies. ²² , ³⁶ , ³⁷ , ³⁸ , ³⁹ For example, a study in the Great Smoky Mountains (USA) found control sites with no scent had higher visitation rates than sites with a scent lure. ⁴⁰ Likewise, Snow et al. (2022) found little evidence that any scent lure among the 28 that were evaluated in the categories of food, pheromone and curiosity performed better than control sites, and several scents had lower visitation than control sites. However, the majority of studies evaluating visitation to scent lures by wild pigs have not included bait within experimental trials. Given their enhanced olfactory abilities, an important limitation to these studies is that wild pigs were potentially attracted to the scent, but their response was not detected on remote cameras if they approached the station but stayed out of frame when they realized that there was no reward (bait). ⁴¹ Thus, there is a need for further research that integrates both bait and scent lures for optimizing wild pig visitation to bait sites. Furthermore, wild pigs are highly social ⁴² and management programs often focus on removal of entire social groups, ⁴³ , ⁴⁴ yet the response of groups of wild pigs (sounders) to attractants compared to individual wild pigs not been considered in previous studies.

Habitat also is likely to influence wild pig visitation of bait/scent locations. ⁴¹ Although wild pigs are habitat generalists, at a coarse scale wild pigs are typically associated with wetlands, streams, hardwoods and other similar habitats, ⁴⁵ the availability of which can influence home range size and shape. ⁷ In particular, wild pigs have limited thermoregulatory capabilities and thus often select for areas proximal to water and associated vegetation types. ⁴⁶ , ⁴⁷ , ⁴⁸ Within habitats where wild pigs are likely to occur, there are fine‐scale habitat features that may influence wild pig movements and ability to locate bait. For example, the spatial distribution of rooting, wild pig vehicle collisions and wild pig locations has been documented to be influenced by tree density, canopy cover, the presence of hard‐mast‐producing species, pine density, distance to swamp and creek habitat, and slope. ⁴⁹ , ⁵⁰ , ⁵¹ , ⁵² These studies suggest that both coarse‐ and fine‐scale habitat attributes may have a strong influence on wild pig visitation to bait sites.

Our goal was to compare wild pig visitation to bait, scent lure and bait + scent to determine optimal attractants for detecting individuals and groups of wild pigs, as well as the influence of habitat attributes on visitation rates and time to visitation. Our objectives were to (1) compare visitation rates of individuals and groups of wild pigs to sites across various baits and/or scent lures reported in previous literature as effective at attracting wild pigs, (2) compare time to initial visitation for sites that wild pigs visited to determine what treatment attracted individuals and groups of wild pigs in the shortest time, and (3) quantify the influence of habitat attributes on visitation rates of wild pigs at bait sites. Results from this research should help maximize the efficacy of wild pig control efforts by improving site visitation, decreasing the time to site visitation, and ultimately decreasing the cost of control programs.

2. METHODS

2.1. Study site

We conducted this study in South Carolina, USA on the Savannah River Site (SRS), a ~800 km² site managed by the United States Department of Energy (DOE). Most of the SRS is undeveloped lands or forested habitat (96%) that was managed for timber production and wildlife conservation by the United States Forest Service. At the time of our study the SRS consisted of ≈50% upland pine including loblolly pine (Pinus taeda), longleaf pine (P. palustris) and slash pine (P. elliottii), 25% bottomland hardwood including sweetgum (Liquidambar styraciflua), swamp gum (Eucalyptus ovata), red maple (Acer rubrum), water oak (Quercus nigra), bald cypress (Taxodium distichum) and tupelo gum (Nyssa sylvatica), 10% herbaceous/shrub areas, 8% upland hardwoods including white oak (Q. alba), black oak (Q. velutina), mockernut hickory (Carya tomentosa) and southern red oak (Q. falcata), and the rest was mixed forests and developed land. Over 20% of the SRS consisted of wetlands and freshwater systems including the Savannah River. Wild pigs have been present on the SRS before its establishment in the early 1950s, ⁵³ and are abundant and widely distributed across the site. ⁵⁴

2.2. Attractants

In order to quantify the effects of attractant type and habitat attributes on wild pig visitation to sites, we conducted experimental trials from January–April 2023 with various bait and scent lure combinations across the SRS and monitored wild pig visitation using remote cameras. We selected top‐performing bait from previous literature, also considering accessibility and cost of food bait. Whole kernel corn is the most common bait used for trapping wild pigs because of its availability and attractiveness, ²³ although peanut paste has been found to have similar visitation rate to corn. ²⁸ Therefore, we used whole kernel corn and whole dried peanuts as the two types of bait in the experimental treatments. In addition to bait, we selected three of the top performing scent lures from a previous study [Strawberry Flavor Oil (Olive Nation, Charlestown, MA, USA), Tuff Tusk Wild Hog Attractant (Razorback Outfitters LLC, McDade, TX, USA) and Tink's Specialty Power Pig Sow‐in‐Heat Estrous (Tink's®, Arcus Hunting LLC, Covington, GA, USA)] ⁴¹ representing disparate categories of lures (e.g. food, curiosity and pheromone) for this study (Table 1). We also selected a fourth scent lure, worm tea, based on the finding of greater consumption of earth worms compared to corn by Foster et al ²⁹ We included a control scent (i.e. sterile water) for comparative purposes. We made plaster scent tabs infused with each of our scent lures following the methods of Webster and Beasley. ⁵⁵ We placed each tab in a Ziplock bag with the scent lure for a minimum of 24 h before deployment in the field. We made worm tea by mixing Earthen Organics Worm Castings (Earthen Organics, Easley, SC, USA) with water in an 18.93‐L bucket in a ratio of 946 mL worm castings to 11.36 L water. We allowed worm tea to brew for a minimum of 48 h, then we poured 1.42 L over 9.46 L of either corn or peanuts to soak for a minimum of 12 h before being deployed.

Table 1.

Fifteen treatments performed January–April 2023 to quantify differences in wild pig visitation rates and time to visitation at scent, bait and bait + scent trials. We utilized lures corn and peanuts as bait as well as scents [Strawberry Flavor Oil (Olive Nation, Charlestown, MA, USA), Tuff Tusk Wild Hog Attractant (Razorback Outfitters LLC, McDade, TX, USA) and Tink's Specialty Power Pig Sow‐in‐Heat Estrous (Tink's®, Arcus Hunting LLC, Covington, GA, USA)] that represented disparate categories of attractants (food, curiosity and pheromone) and a control (i.e. sterile water)

Treatment	Bait	Lure	Treatment type
1	Control	Control	Control
2	Corn	Control	Bait
3	Peanut	Control	Bait
4	Corn	Sow in Heat	Bait + scent
5	Corn	Strawberry	Bait + scent
6	Corn	Tuff Tusk	Bait + scent
7	Corn	Worm	Bait + scent
8	Peanut	Sow in Heat	Bait + scent
9	Peanut	Strawberry	Bait + scent
10	Peanut	Tuff Tusk	Bait + scent
11	Peanut	Worm	Bait + scent
12	Control	Sow in Heat	Scent
13	Control	Strawberry	Scent
14	Control	Tuff Tusk	Scent
15	Control	Worm	Scent

2.3. Field trials

We used arcgis pro 3.1.1 ⁵⁶ to create random points in wetland and riparian habitats across the SRS for experimental trials including floodplain oak forests, floodplain sweetgum forests, mixed bottomland hardwoods, mixed composition floodplain hardwoods, bald cypress and water tupelo, bottomland hardwoods and cypress, marsh, and wetland scrub forests habitats. We focused on these landscapes because wild pigs utilize wetlands and riparian habitats extensively for shelter and proximity to water. ⁴⁵ , ⁵³ , ⁵⁷ Points were a minimum of 225 m apart to avoid site habituation and within 150 m of a road to facilitate accessibility (Supporting information Fig. S1). We conducted 12 rounds of testing, which consisted of 80–120 sites for each round. During each round, we selected points that were ≥1000 m apart to reduce the likelihood that individuals visited more than one site during a given round. Each site was monitored using a remote camera for 7 days, after which we moved and reset cameras at new locations until all trials were complete. Points received a random treatment among the 15 possible treatments (Table 1). Treatments consisted of a bait, scent lure, bait + scent, or no bait or scent (i.e. controls). Each of the 15 treatments (Table 1) had 50 replicates assigned randomly to the points established, resulting in trials at a total of 750 sites.

We deployed a remote camera (Hyperfire and Hyperfire 2 Professional Infrared; Reconyx®, Holmen, WI, USA) to monitor wild pig visitations at each site. We mounted cameras on trees ≈1 m off the ground. We set cameras to take three pictures per motion‐activated trigger with a 1‐s delay between pictures and a 3‐min quiet period. We placed a bait pile (corn, peanuts or control) 2.7–3.7 m in front of the camera. We placed a scent tab (sow in heat, strawberry, tuff tusk, or control) in plastic chicken wire caging and hung it with wire on the nearest tree branch ≥1.8 m above the bait pile to prevent wildlife tampering.

We processed images using the Colorado Parks and Wildlife Photo Database. ⁵⁸ For images with wild pigs, we delineated images into unique visits, with images that were >15 min from a previous image of a wild pig at that station recorded as a unique visit. We utilized a 15‐min period because the average length of a feeding bout was 14.2 min. ²⁸ For each of these visitations, we recorded whether an individual pig was visiting or a group of wild pigs (>1). We also recorded the age (adult or juvenile) and sex of each pig when possible, with juveniles and unidentifiable adults classified as unknown sex.

2.4. Attractant performance analysis

After removing seven sites where camera batteries failed prematurely, we quantified each of the 743 sites as having wild pigs visited or not. We used the 297 sites that had wild pig visitation and compared those locations to the 446 sites that did not have wild pig visitation using a used‐unused framework. ⁵⁹ We fitted general linear models (GLM) in R 4.2.3 using the ‘glm’ function in the stats package. ⁶⁰ We ran three models with the response variable as whether a wild pig visited a site (yes/no), an individual pig visited a site (yes/no), or a group of wild pigs visited a site (yes/no) which each used a binomial distribution. We first used the specific bait and lure combination (i.e. corn + strawberry, peanut + worm, control + sow in heat, etc.) as the fixed effect in each model. Finding no significant differences among the specific combinations (Figs S2–S5), we pooled these combinations into treatment types (bait, scent, bait + scent, control; Table 1) and used treatment type as the fixed effect for each model. From the 297 sites that wild pigs visited, we calculated the time to initial visitation in hours as the difference in time from when the bait station was deployed, and the time the first wild pig was captured on the remote camera. We fitted a GLM in R 4.2.3 using the ‘glm’ function in the stats package. ⁶⁰ The response variable was the time to visitation of a wild pig (hours) and we used an exponential distribution for the model. We did not develop separate models for groups and individual wild pigs due to small sample sizes. We first used the specific bait and lure combination as the fixed effect in the model. Finding no significant differences among the specific combinations, we combined these combinations into treatment types (bait, scent, bait + scent, control; Table 1) and used treatment type as the fixed effect for the model.

2.5. Effects of habitat attributes on wild pig visitation to bait sites

We selected a random subset of the 743 sites used for our attractant analysis to quantify the influence of habitat on wild pig visitation to bait sites. The selected sites consisted of 50 control and 25 each of our other treatments that included bait (bait and bait + scent; 300 sites). We excluded scent only sites owing to low visitation rates (24.40%; see results). At each of these sites we quantified the following biologically relevant fine‐scale habitat attributes identified in the literature as potentially important in influencing wild pig visitation: overstory composition, understory composition and percentage of hard‐mast‐producing species. Within a 15‐m diameter of the site, we measured diameter at breast height (DBH) and identified every tree with a DBH > 10.16 cm. We also estimated the percentage of hard‐mast‐producing trees in this area. We visually categorized the understory within this 15 m diameter as primarily consisting of saw palmettos (Serenoa repens), giant cane (Arundinaria gigantea), seedlings (woody stems) or bare/litter.

In addition to fine‐scale habitat attributes collected in the field, we also quantified additional site‐specific habitat attributes using GIS layers in arcgis pro 3.1.1. We characterized the location of streams and roads on SRS from existing geospatial layers. We classified primary roads as paved roads that are regularly traveled and secondary roads as unpaved roads that were gravel or logging roads with little vehicle travel. We determined specific forest types from the 2021 National Landcover Database (NCLD) raster layer (30 × 30 m resolution) ⁶¹ and used the reclassification tool in arcgis pro 3.1.1 to condense landcover into the following categories: open landscape, evergreens and upland hardwoods. We excluded bottomland hardwoods as a landcover type because the majority of our points were located in this landcover type. We used the Euclidean Distance tool in arcgis pro 3.1.1 to calculate distance to each of these habitat covariates from our bait trial locations. We used the NLCD 2021 USFS tree canopy raster (30 × 30 m resolution) to estimate the percentage canopy cover at each location. ⁶² We used lidar data provided by SRS Forest Service from 2018 to calculate normalized relative point density 0.15–2 m off the ground to represent understory density. ⁶³

2.6. Habitat analysis

We modeled wild pig visitation to bait sites following a used‐unused design in which sites that had wild pig visitation were considered ‘used’ and sites that did not have wild pig visitation were ‘unused’. We modeled the influence of distance to water, distance to landscape types, distance to primary roads, distance to secondary roads, canopy cover, hard‐mast‐producing trees, understory density, understory type and overstory type on the presence of wild pigs at a site. We fitted a GLM including all spatial explanatory variables as fixed effects in R 4.2.3 using the ‘glm’ function in the stats package. ⁶⁰ Habitat variables and treatment type (Table 1) were all included as fixed effects in our global model (Table 2). We used Pearson's correlation test to determine that there was no correlation between our continuous predictor variables (i.e. Pearson's |r| < 0.5) and all were included in the model. ⁶⁴ From our global model we used a backwards direction stepwise regression framework to identify the most important habitat attributes from which to conduct further model selection. Backwards stepwise regression uses an iterative process that removes variables from the model if doing so improves the model's Akaike information criterion (AIC), until a subset of the most influential variables are retained. We then used the ‘dredge’ function in the mumin package in R 4.2.3 on our top model produced from the stepwise regression analysis (Table 2) to compare all possible combinations of the remaining fixed effects. ⁶⁵ We ranked models using AIC to determine what spatial variables best explained wild pigs’ presence at a site. ⁶⁶ We used the mumin package in R 4.2.3 to model average our top models (< 2 ∆AIC) and calculate model averaged parameter estimates and 95% confidence intervals (CIs). To quantify the importance of each predictor variable, we then used the ‘sw’ function in the mumin package in R 4.2.3 to sum the Akaike weights of our top models to get importance values from 0 to 1.

Table 2.

Importance values and the number of models in which the variable occurred for all variables retained in the eight supported generalized linear mixed models evaluating the influence of habitat attributes on wild pig visitation to bait sites in South Carolina, USA. Distance to water is the standardized variable of distance to a water source in meters. Primary understory is the type of dominant understory which was categorized as seedlings, bare/litter, saw palmettos or giant cane. Treatment is determined by what was placed at the camera site to attract wild pigs which was bait only, scent only, bait + scent, or control. Understory density is the understory density at the location. Distance to open landscape and to evergreen habitat are the standardized variables of distance to an open landscape or evergreen habitat (respectively) in meters

Model variable	Distance to Water	Primary understory	Treatment	Understory density	Distance to open landscapes	Distance to evergreen habitat
Importance values	1.0	1.0	1.0	0.63	0.55	0.52
Number of containing models	8	8	8	4	4	4

3. RESULTS

3.1. Attractant performance

Of the 750 original trials, 743 had usable data resulting in >240 000 trail camera images with >33 000 photos of wild pigs. Wild pigs visited 297 sites (39.97% including controls, 38.41% excluding controls), groups of wild pigs visited 165 sites (22.21% including controls, 22.69% excluding controls) and individual wild pigs (1218 unique visits, 701 male visits, 165 female visits, 352 unknown visits) visited 278 sites (37.42% including controls, 37.16% excluding controls) (Fig. 1; Supporting informationTable S1).

Figure 1.

Percentage of sites with visitation for all wild pigs, groups of wild pigs (Sus scrofa) and individual pigs by treatment type. We used corn and peanuts for bait, and strawberry, Sow in Heat, Tuff Tusk sour mash and worm tea as scent lures. Control sites consisted of no bait or scent; site treatments were bait only, scent only and bait + scent. The visited bar represents any pig, single represents that an individual pig visited the site and group represents that more than one pig visited the site at a time. This study was conducted at the Savannah River Site, Aiken, SC, USA from January to April 2023.

We found evidence that wild pigs had greater visitation rates at sites with bait (42.85%; CI 33.03–53.25; P = 0.006) and bait + scent (47.99%, CI 43–53.02; P < 0.001) than sites with scent alone (24.40%; CI 18.86–30.91) (Figs 1 and S6; Table S2). Groups of wild pigs had greater visitation rates at sites with bait (26.53%; CI 18.11–36.41; P < 0.001) and bait + scent (30.90%; CI 26.29–35.69; P < 0.001) than sites with scent alone (5.26%; CI 2.66–9.22) (Figs 1 and S7; Table S2). Groups of wild pigs also had greater visitation rates at sites with bait + scent (30.90%; CI 26.29–35.69) than control sites (10.20%; CI 3.39–22.23; P = 0.022). Individual wild pigs had greater visitation rates at sites with bait (41.84%; CI 31.95–52.23; P < 0.001) and bait + scent (42.71%; CI 37.79–47.74; P < 0.001) than sites with scent alone (24.40%; CI 18.74–30.80) (; Table S2). None of the other relationships were significant (P < 0.05; Fig. 1).

We found quicker times to visitation for bait + scent sites (60.98 h; CI 54.45–67.52; P = 0.051) than sites with scent alone (82.03 h; CI 67.07–96.99), but surprisingly control sites (44.92 h; CI 25.33–64.51) also had quicker time to visitation than sites with scent alone (P = 0.019; Figs 2 and S9; Table S2). Groups of wild pigs had quicker time to visitation at control sites (22.04 h; CI −10.10‐54.17) than bait + scent (66.64 h; CI 58.67–74.61; P = 0.003), bait alone (64.44 h; CI 45.45–83.43; P = 0.008), and scent alone (83.25 h; CI 38.95–127.56; P = 0.002; Fig. 2; Table S2). There was no difference in time to initial visitation between treatment types for individual pigs (Fig. 2; Table S2).

Figure 2.

Average time to initial visitation in hours of a wild pig (Sus scrofa) at a site for all pigs, single pigs and groups of pigs for each of the treatment types for sites that had wild pig visitation with sample size below each bar representing the number of sites visited. We used corn and peanuts for bait, and strawberry, Sow in Heat, Tuff Tusk sour mash and worm tea as scent lures. Control sites consisted of no bait or scent; site treatments were bait only, scent only and bait + scent. The visited bar represents any pig, single represents that an individual pig visited the site, and group represents that more than one pig visited the site at a time. This study was conducted at the Savannah River Site, Aiken, SC, USA from January to April 2023.

3.2. Effects of habitat attributes on wild pig visitation to bait sites

Of the 284 sites where we measured fine‐scale habitat attributes in the field, 139 sites had wild pig visitation. Our backwards stepwise regression analysis on our global model resulted in a top model retaining the variables distance to water, distance to an open landscape, distance to evergreen habitat, understory density, primary understory type, and treatment. Model selection from this global model resulted in eight top models (Table S3). All supported models included distance to water, primary understory species, and treatment, indicating that these three variables had the greatest influence on the probability of wild pigs locating a bait site (relative importance = 1 for all three variables; Table 2). Understory density (relative importance = 0.63), distance to open landscape (relative importance = 0.55) and distance to evergreen habitat (relative importance = 0.52) were in four of the top models (Table 2).

Model averaged parameter estimates from the top‐performing models indicated that wild pig visitation to bait sites was positively influenced by proximity to water sources, when bait + scent was present, and when the site was not in an understory of seedlings (Fig. 3). Visitation also was positively influenced by sites closer to open landscapes, further from evergreen habitat, in areas of increased understory density, in an understory of saw palmettos, at locations with bait, and not in understories of seedlings or giant cane, although coefficients for these variables overlapped with 0 (Fig. 3). The odds of detecting a wild pig at a bait site decreased by 11.21% for every 100‐m increase in distance to a water source (Table S4; Fig. 4). Within sites with the understory species dominated by seedlings, the odds of detecting a wild pig at a site decreased by 59.1% (Table S4). The odds of detecting a wild pig at a site were 149.4% higher when bait + scent was present than control sites with no bait or scent present (Table S4).

Figure 3.

Estimated coefficients with 95% confidence intervals after model averaging top performing models for each of the habitat predictors for wild pig site visitation on the Savannah River Site, Aiken, SC, USA in January–April 2023. The dashed line corresponds to 0 and if the confidence interval overlaps, the habitat predictor does not have a significant effect. Distance variables (distance to evergreen habitat, open landscapes and water sources) that are <0 indicate that the sites had a higher probably of being detected closer to these features. All other habitat predictors [Primary Understory: Palmettos, Treatment: Bait + Scent, Treatment: Bait only, Understory Density, Primary Understory: Giant cane and Primary Understory: Seedlings(woody stems)] that are >0 indicate that the sites had a higher probably of being detected with these predictors.

Figure 4.

Predicted relative probability of wild pigs (Sus scrofa) being detected at a bait site in relation to distance to water sources in meters on the Savanah River Site near Aiken, SC, USA from January to April 2023. The solid line represents the mean predicted probability and the shaded region represents the 95% confidence interval. This is based on our top‐performing model and holds all other predictors constant (distance to evergreen, distance to open landscape, understory density and primary understory).

4. DISCUSSION

Collectively, our results demonstrate that use of bait and placing traps or cameras in habitats near water and in dense understory vegetation can maximize visitation of wild pigs and minimize their time to detection. Despite their widespread appeal and use in wild pig trapping, we found that using scent alone was insufficient for attracting both individuals and groups of wild pigs to a site. The addition of scent to bait sites did appear to enhance the detectability of wild pig social groups compared to our control sites, although our results combined with those of Snow et al. ⁴¹ and Wathen et al. ⁴⁰ suggest wild pigs may display an aversion to sites with certain novel scent lures, so managers should use caution when considering their application in management activities.

Scent lures have proven effective at attracting numerous wildlife species and thus are commonly used for trapping and monitoring (e.g. scent stations). ⁶⁷ , ⁶⁸ , ⁶⁹ , ⁷⁰ However, despite their advanced olfactory system our results add to a growing body of literature suggesting that scent alone is ineffective for attracting wild pigs to traps or camera monitoring stations. ³⁷ , ³⁹ , ⁴⁰ , ⁴¹ We also found that wild pigs visited over one third of control sites consisting of no bait or scent. The high visitation of wild pigs to our control sites is likely to reflect the placement of camera stations in our study near preferred habitats (riparian areas) of wild pigs. Interestingly, control sites had shorter times to visitation than sites with scent lures suggesting wild pigs may display an aversion to sites with certain novel scent lures. Bait/scent aversion has been documented for numerous species, including wild pigs owing to trapping pressure, nonlethal doses of toxicants/poisons, or novel items that may be perceived as ‘risky’. ⁷¹ , ⁷² , ⁷³ , ⁷⁴ , ⁷⁵ We recommend further research on novel scent lure aversion of wild pigs with GPS collars to determine how they may avoid these novel items.

A food reward is important for attracting wildlife and our results demonstrate that use of bait is critical to increase the probability of visitation and reduce times to initial visitation for wild pigs. Previous studies have demonstrated the use of bait is more efficient for attracting numerous species of wildlife than scent stations or nonbaited wildlife captures for monitoring and research. ⁷⁶ , ⁷⁷ For wild pigs, in our study the two evaluated baits (corn and peanuts) had similar visitation rates and time to initial visitation, suggesting a variety of food‐based baits may be effective for detecting wild pigs depending on local availability and wild pig preferences (Figs S2 and S3). Indeed, managers utilize a wide variety of baits across the world with success at attracting wild pigs and wild boar, with previous studies finding that wild pigs have similar visitation rates to various baits. ²³ , ²⁵ , ²⁷ , ²⁸ , ²⁹ One constraint to our study is that wild pigs are trapped on the SRS with bait and have the potential to exhibit bait aversion; therefore, in areas where wild pigs have not been exposed to bait from trapping there is the potential for higher visitation rates than were observed in our study. Although groups of wild pigs had shorter times to visitation at sites with bait and scent than control sites, in general we did not observe a difference between bait only sites and those where scent was added to the bait site. Further research with GPS collars could inform if wild pigs are attracted from further distances if a scent attractant is combined with bait.

Our results also demonstrate that habitat attributes play an important role in wild pig visitation rates, with distance to a water source and understory type having the greatest influence among the habitat attributes evaluated in our study. Wild pigs can survive in a diversity of environments enabling them to exist worldwide, but their habitat selection varies by location and season. ⁴⁵ , ⁷⁸ Wild pigs are dependent on water for foraging, wallowing and thermoregulation, so much so that the absence of available surface water has been suggested as the only habitat attribute that effectively excludes wild pigs from an area. ²⁶ , ⁷⁹ Indeed, home range placement and shape, vehicle collisions and rooting damage have all been shown to be related to water sources. ⁴⁸ , ⁵⁰ , ⁵² , ⁵⁷ , ⁸⁰ , ⁸¹ This is a consequence of the fact that wild pigs have few sweat glands and require water to thermoregulate. ⁸² As a result, wild pigs are often located near water and vegetation types associated with water including swamps, marshes, bottomland hardwoods and riparian areas. ⁴⁵ , ⁵⁷ Given the importance of water to wild pigs, it is unsurprising they are more likely to locate bait sites in close proximity to water sources.

Wild pigs also were more likely to locate bait sites in areas with specific understories. Understory habitat and understory density can be a driving factor in habitat selection for many species as this is the portion of their habitat that often provides food, shelter and breeding sites. ⁸³ , ⁸⁴ , ⁸⁵ For wild pigs specifically, dense areas can provide important shade for this species that is a poor thermoregulatory, and cover for bedding and nesting sites. ⁸² , ⁸⁶ , ⁸⁷ , ⁸⁸ In our study area, we found wild pigs less likely to locate sites with seedlings as the primary understory type, representing a less dense understory type compared to giant cane and saw palmettos. In particular, among dense understory types in our study area, wild pigs tended to locate more sites in areas with saw palmettos. Saw palmetto is a native clumping, shrubby palm that typically grows 1.5–3.0 m tall and spreads 1.2–3.0 m wide and is considered a dense understory type. ⁸⁹ Wild pigs have been documented consuming the apical meristem material by pulling new leaves out with their mouth and consuming the base. ²⁴ , ⁹⁰ Wild pigs also have been documented constructing nests or bedding sites by cutting and piling palmettos to protect their young ²⁷ (Fig. S10). On the SRS, wild pigs may have located bait sites with palmettos as the primary understory more frequently because palmettos can provide cover and are usually located in floodplains near water sources at our study site. However, we had a relatively low sample size of sites with an understory dominated by saw palmettos, and thus future research investigating microhabitat preferences of wild pigs would be useful to better elucidate the importance of saw palmettos for wild pigs.

Wild pigs are highly social animals, with groups typically consisting of 1–3 or more adult females and their offspring (i.e. sounders) or small groups of related subadult males, ⁴² whereas adult males are typically solitary and only temporarily assimilate within sounders for breeding. Given their polygynous mating strategy and high reproductive capacity, ⁹¹ most wild pig management goals are focused on removing sounders to reduce population sizes. ¹⁵ Furthermore, wild pigs are semi‐territorial, ³² , ⁴⁵ , ⁹² , ⁹³ , ⁹⁴ and thus targeting groups and conducting whole sounder removal has been proven an effective method to reduce wild pig populations. ⁴³ , ⁴⁴ Sounders visited fewer control and scent only sites than sites with bait, signifying that bait is important for targeting these groups. Individual wild pigs were primarily males, and those individuals visited more control sites than sounders. Males tend to have larger home ranges and move more than females ⁷ , ⁷⁸ ; therefore, in the present study they may have encountered more control sites by chance while moving throughout the landscape.

5. CONCLUSION

We recommend that managers focus on using baits, or foodstuffs that are locally available food resources to maximize wild pig visitation to traps or camera locations. Managers should avoid using only scent lures for attracting wild pigs unless they are used to enhance a bait, but when using scent should be careful because novel scents can act as deterrents. We also recommend managers optimize locations for bait sites by focusing on the shelter and water requirements of wild pigs within a given landscape. Given that both use of bait and certain habitat attributes (i.e. proximity to water and understory) were likewise important in influencing whether wild pigs visited a site, both of these factors should be considered when establishing management programs. Finally, our data suggest that if wild pigs do not visit bait sites within a reasonable time frame (e.g. 1–2 weeks), shifting the bait site to a new location or use of alternative baits is likely to be more effective in detecting wild pigs than deployment of novel scent attractants. Collectively, these recommendations should lead to increased efficiency and efficacy of wild pig management programs.

DISCLAIMER

This report was prepared as an account of work sponsored by an agency of the United States Government. Neither the United States Government nor any agency thereof, nor any of their employees, makes any warranty, express or implied, or assumes any legal liability or responsibility for the accuracy, completeness. Or usefulness of any information, apparatus, product, or process disclosed, or represents that its use would not infringe privately owned rights. Reference herein to any specific commercial product, process, or service by trade name, trademark, manufacturer, or otherwise does not necessarily constitute or imply its endorsement, recommendation, or favoring by the United States.

Supporting information

Data S1. Supporting Information.

DATA AVAILABILITY STATEMENT

The data that support the findings of this study are available from the corresponding author upon reasonable request.