Proceedings of the Vertebrate Pest Conference

South Africa has approximately 8,000 commercial small livestock farms and 5,800 communal/subsistence farmers throughout the country. Reported rates of small livestock loss to predation range from 3-13% and 0.5-19% from communal farming areas. A range of predators exist on the African continent, but in southern Africa major livestock losses are primarily due to black-backed jackal and caracal. South Africans have been managing caracals and jackals for over 300 years with no elimination of predation. During the aforementioned time frame, producers have used and/or developed a number of techniques including lethal, nonlethal, and integrated predator damage management to address predation losses. In the Karoo area of South Africa, one producer decided that a new way needs to be developed after losing over 60 lambs in a month, while practicing continuous removal of caracal and black-backed jackal. His integrated predator damage management system includes using a prototype nonlethal collar system for sheep and lambs. The collars are used to train dominant pairs of predators to avoid predation while maintaining their territories and keeping transient predators out of the area. The system has now gone into production in South Africa and is being distributed by its inventor.

Use of UAVs (unmanned aerial vehicles) in wildlife applications has been increasing in recent years as system costs have come down and regulations regarding their use have become more well-defined. Medium and larger UAVs can accommodate sophisticated payloads, allowing for missions using LIDAR to obtain measurements of vegetation height and fine-scale elevation data; high resolution video and thermal imaging for surveying wildlife; remote spraying for control of exotic plants; and broadcasting audio calls for hazing wildlife at oil spills. We have been developing some additional capabilities for potential use in wildlife research. The first is using UAV platforms as a means to remotely deliver anesthetic darts into larger wildlife species. This capability would allow for anesthetizing free-ranging deer, elk, bison, moose, etc. without the restriction of being close enough to use traditional rifle-based darting. Other drugs that could be delivered include those for immunocontraception and disease inoculation. We are also developing a drone-based remote net launcher system to allow for capture of both birds and mammals. Use of UAVs to aid in wildlife management activities that previously required more expensive aerial assets (e.g., airplanes or helicopters) or were not possible due to other restrictions, may allow managers to be more efficient and expand capabilities beyond what are currently available.

Those who have been in the field of wildlife damage management very long probably have a file drawer full of half-baked ideas and ill-fated research projects that never should have seen the light of day. This paper will be a tongue-in-cheek look at the scientific method and saving grace of pilot studies. A pilot study is a small-scale test of the procedures to be used in a large-scale study. The goal of pilot work is not supposed to be the testing of hypotheses, but sometimes researchers just can’t help themselves. Beware of small sample sizes and the potential for false negative and false positive results. I have been involved in more pilot studies than I care to admit that ended up being a total bust, but there have been some that led to well-informed modifications of study designs and results that were immediately publishable because the variation in resultant data was low and results were clear cut. Pilot studies have expanded my knowledge of systems, study design, methodology, and the behavior of individual animals. I encourage the use of pilot studies in research associated with vertebrate pest management.

All current vertebrate pest control and detection devices require a lure to be effective. Even advanced multi-kill and multi-species technologies that can work remotely for extended periods without maintenance require lures to be effective. We have developed, over the past five years, several sustained release, long-life lures for controlling and monitoring rats. The lures contain blends of chemical compounds that we identified as attractive to rats. Our early prototype lures were effective at kill-trapping wild, free-ranging ship rats without the need for replenishment for six months. We have since engineered other lures to suit different applications (e.g., kill-trapping and monitoring) and users (e.g., pest control and conservation) and are also testing our lures as bait consumption motivators. We are currently transforming our prototypes into products for international markets.

The 0.005% diphacinone rodenticide pellets used in this study have been proposed for use in field applications to control introduced rodents on conservation lands in the state of Hawaii. Introduced rodents (especially Rattus spp.) cause a wide array of conservation problems in the Hawaiian Islands and on other islands. We assessed the acceptability and toxicity (should the pellets be consumed) of two rodenticide baits to Canada geese, a surrogate species for the endangered Hawaiian goose. Based on these trials with captive, wild Canada geese, it appears that neither the whole nor the chopped (to simulate broken or weathered baits) pellets pose a significant risk to the Hawaiian goose, a species considerably smaller than the Canada goose. The pellets (whole or chopped) were not accepted by the Canada geese during this study despite their having only a small amount of green grass sod as an alternative food. There were no mortalities of geese during the feeding trials and all geese remained healthy, based on body weights and packed blood cell volumes. The endangered status of the Hawaiian goose precluded using it as the target study species.

Nest predation often limits recovery of threatened and endangered birds, especially ground-nesting species. Accordingly, a variety of techniques are used to reduce the impact of nest predation on listed species. We examined the efficacy of conditioned taste aversion, a nonlethal technique designed to induce avoidance behavior in predators after being exposed to prey items that have been treated, usually with a chemical emetic that causes predators to become ill within minutes of consumption. We used carbachol (carbamyl choline chloride) as a taste-aversive agent to condition corvids responsible for high levels of nest predation on two federally listed species [the western snowy plover (Charadrius nivosus nivosus) and California least tern (Sternula antillarum browni)] breeding at Marine Corps Base Camp Pendleton, California. Carbachol is tasteless, colorless, and odorless, which makes it a preferred aversive agent as predators are unable to detect the chemical and therefore associate their resultant sickness with consumption of the prey item. We conducted two separate experiments in 2013 and 2014, during which we deployed 772 artificial nests during the first experiment and 760 artificial nests during the second experiment. Both experiments were conducted prior to the onset of egg laying for plovers and terns (i.e., Feb-Mar) and each artificial nest contained three quail (Coturnix sp.) eggs. During the first stage of both experiments all of the nests only contained untreated quail eggs, and nest predation was high with >90% of nests failing within 1-2 days of deployment. In subsequent stages, we deployed carbachol-treated eggs in increasing proportion. We used nest survival models to evaluate daily survival rates (DSR) of artificial nests in all stages of both experiments. During both experiments, DSR increased concomitant with a greater proportion of carbachol-treated eggs. Common ravens (Corvus corax) accounted for 98.1% (n = 471) of all artificial nest predations in Experiment 1, and 95.6% (n = 498) of all artificial nest predations in Experiment 2. Using carbachol as a taste-aversive agent was effective at reducing predation on artificial nests as illustrated by increased DSR (0.47 to 0.98 in the first experiment and 0.00 to 0.99 in the second experiment); however, transferability of this technique to plover and tern nests was not fully realized.

Invasive ship rats (Rattus rattus) are the major threat to the native species and ecosystem of Goat Island (9.3 ha), New Zealand. The island is only 100 m away from the mainland, which imposes a risk of incursions by rats swimming over. Accessibility depends on weather and tide times which makes regular trap servicing complicated. In 2016 we extended an existing trapping grid of 8 kill traps (DOC200; Department of Conservation, New Zealand) with 10 self-resetting traps (Goodnature A24s; Goodnature Limited, Wellington, New Zealand) to improve current management and ideally achieve eradication. Before our study started, DOC200s on the island had not been serviced for six months. Rats were active even during the day and rat numbers were assumed to be high. The DOC200 kill traps were lured with an egg and A24s were lured with Goodnature automatic lure pumps (ALP) baited with chocolate formula for rats. The A24s were equipped with Goodnature digital strike counters to document the number of rats killed by the self-resetting traps. All devices were on average checked every 49 days from August 2016 to October 2017. DOC200s were reset after triggering or after three months, whichever occurred first. ALPs were replaced in January and July 2017. Gas cartridges of the A24s were replaced when the strike counter showed 20 or more. A substantial number of rats were killed on the island (242 by A24s and 27 by DOC200s) in 8,838 uncorrected trap nights between August 2016 to October 2017 (Table 1). The initial number of individuals killed by A24s in the first month after deployment in August 2016 was high. The number of rats removed by A24s remained at a high level from November 2016 until June 2017. The number of A24 kills varied widely between the beginnings of the breeding seasons across the two years, with 46 individuals killed between September and November 2016 but only 4 kills between August and October 2017. Even though initial trapping success was high, eradication could not be achieved and the self-resetting traps did not perform better than traditional kill traps once rat abundance was low. The Goodnature A24 has shown the potential to work effectively for initial knock-down when rat numbers and activity were high. The advantage over traditional single kill traps, like the DOC200, was the low need for servicing. However, once the population density was reduced to a lower level, this advantage vanished. Most kills by A24s happened during the breeding season in 2016 in the first third of the study. In the last three months of the project the kill numbers did not differ meaningfully from the DOC200 kill numbers. Even though trapping numbers were low, rat abundance was still assumed to be high. After our study had finished rat control on the island using these devices was continued by local community volunteers. A further 357 rats were caught between June 2018 and March 2020, indicating no decline in rat captures. On 1 September 2019 and 1 March 2020 tracking tunnel indices were 100% (K. Tricklebank pers. commun.). In March 2020 initial cost for a DOC200 was NZD$145.00. The purchase costs for a Goodnature A24 with lure and gas for 6 months was NZD$169.00 (excluding digital strike counter). For the first six months the differences in costs were moderate. However, after six months, the material costs for servicing a DOC200 was NZD$0.55 (one freerange egg), or less when using peanut butter. The servicing cost for a Goodnature A24 is NZD$19.00 (gas cartridge + ALP). On Goat Island, servicing costs for a six months period were approximately NZD$9.00 for eight DOC200s and NZD$190.00 for ten A24s. Higher costs for A24 purchase and servicing compared to DOC200 were not compensated by noteworthy higher kill numbers once rat numbers were reduced. However, A24s performed well during peak times and labour costs were not considered in our study. Self-resetting devices at one per hectare did moderately reduce rat numbers in an area where kill trap maintenance was time and cost intensive but suppressing to very low rat numbers or achieving eradication requires additional investment in the system (e.g., a combination of different tools including toxins or a higher density of devices).

Feral cat (Felis catus) predation has negative impacts on native species, especially in island ecosystems (Vitousek 1988, Dowding and Murphy 2001, Bonnaud et al. 2011, Doherty et al. 2016). Feral cats are prolific opportunistic carnivores that prey upon whatever food source is most abundant and preferred (Parsons et al, 2019). They are the cause of extinctions of a plethora of species world-wide. Feral cats on San Clemente Island (SCI) are non-native predators to various endemic species that include the San Clemente Island deer mouse (Peromyscus maniculatus), Island night lizard (Xantusia riversiana), San Clemente Island Bell’s sparrow (Artemisiospiza belli clementeae), and the San Clemente Island loggerhead shrike (Lanius ludovicianus mearnsi) (Biteman et al 2015). Feral cats on SCI also compete for prey items with a subspecies of the Island fox (Urocyon littoralis clementae) that is naturalized to San Clemente. Understanding the feral cat’s activity patterns and estimating their density are important to improve the efficacy of feral cat management efforts. Camera traps are used to quantify activity patterns and population sizes of a variety of species world-wide, but have seldom been used to address these questions with feral cats, and no study has been done on SCI. We used camera traps on SCI to quantify activity patterns, identify high density areas, and determine feral cat population size. Our study took place from November 2018 to March 2019. We had 75 stations with two cameras at each (Figure 1). We checked each trap every two weeks to maintain bait, batteries, and SD cards. Cameras at a station were placed 5 m apart and were facing each other. Stations were always set on a game trail. We used two bait types (i.e. feline and general mammalian predator) at each station, with one piece of bait in front of each camera. Photos from SD cards were manually sorted; photos with no wildlife in them were deleted. Photos were uploaded to a camera trap photo processor called Timelapse2, which extracted date and time, and gave a unique identifier to each photo. We manually entered the species seen in each photo. We used N-mixture models to calculate density, activity, and population data for feral cats. We used program R to run these statistics. We found 114 individual cats identified on 75 cameras over the 4 months of testing. We estimated the population at 319 to 331 cats (95% CI). Activity peaks were found at 1:30 pm and 7:30 pm daily (Figure 2). Spatial density varied but most of the dense areas were located far from roads (or at least outside of shooting distance) and also in exclusion zones (Figure 1). We think multiple factors contributed to the daily patterns. Winter climate in southern California is rainy and rainy conditions likely forced cats into more daytime activity for better thermoregulation. We have been conducting nighttime control of feral cats on the island for 20+ years; nighttime hunting pressure may have shifted their activity toward a daytime peak. Both of the activity peaks we calculated are during hours, when we have not had full firearm access (e.g., limited to 40 yards and closer), and our usual method of identifying cats (by spotlighting) is ineffective. Spatial density is another factor: San Clemente is a valuable and heavily used training ground for the U.S. Navy, so there are areas on the island we cannot access at certain times, or ever. Due to firearms range, terrain, and access time, efforts have been focused along roads, but cat “hot spots” are far off the roads, in or around exclusion zones, or a combination of the two. All this results in large areas where predator control cannot be carried out. More efforts to hike and hunt are being made to extend our reach, giving better data and the ability to remove cats efficiently from high density areas where they have been unchecked. Efforts to hunt during the day should be considered. Other control efforts elsewhere have used air rifles (which do not qualify as firearms), that could be used during the day. We could also use sit-and-wait methods (i.e., use attractants to lure cats in; then remove the cats). We conclude that our efforts could be doubled by adding day hunting in high density areas off the road.

Rodents cause devastating damage to both agriculture and ecosystems worldwide. Invasive rodents are commonly found on islands, historically free of these animals, and have significant negative impacts on both native plant and animal species. Rodents are exceptionally well adapted to their environments and therefore, quite challenging to control. Current control strategies often include large scale applications of toxicants, which have potential adverse effects on non-target wildlife. In island ecosystems, these adverse effects are a major hurdle to eradication efforts. The time and financial resources required to minimizing risks to non-target species and performing post-eradication exposure monitoring can limit the number of islands from which rodents are successfully eradicated. Therefore, the development of new species-specific rodenticides would be a valuable advancement in the effort to control these pest species, especially for island eradications. To that end, USDA Wildlife Services is investigating the use of interfering RNA (RNAi) as a novel way to control rodent species. RNAi reduces the amount of a specific protein that is made by a cell. This is done through post-transcriptional gene-silencing. The RNAi pathway is initiated when a small section of double stranded RNA is introduced into the cytoplasm of a cell. This double stranded RNA comprises a guide strand and a passenger strand. In the first step of the RNAi pathway, the foreign RNA is incorporated into an enzyme complex called RNA-induced silencing complex (RISC) at which time the passenger strand is degraded. The RISC/RNA complex then finds the complementary mRNA made by the cell and binds to it. Subsequently, the RISC complex degrades the complementary mRNA. This breaking down of the mRNA prevents the synthesis of the corresponding protein. The reduction in protein synthesis is the benchmark of RNAi and is how it will be used to elucidate lethal physiological changes in pest species. The species specificity of RNAi depends on the selection of portions of genes that are unique to the target animal species. By screening the rodent genome and comparing sequences of rodent genes to non-target species, we can choose sections of genes that are present in the pest species and absent in the non-target species. Previous research has established guidelines for both the nucleic acid composition of RNAi sequences and their location in the corresponding mRNA that facilitate maximum inhibition of protein synthesis while maintaining species specificity. Therefore, unlike current rodenticides, if non-target species consume the RNAi they will not be affected. This specificity could allow for the eradication of rodents from islands that have historically not been feasible. For RNAi to be useful as a rodenticide, it will likely have to be formulated in a bait for oral consumption. This is a significant hurdle: most RNAi based therapeutics for human use are formulated for intravenous or subcutaneous injection. Like all rodenticide baits, an RNAi-based bait will have to be stable in a wide range of environmental conditions and have a long shelf life. However, once consumed by the target animal, RNAi-based baits will have to be significantly more complex than traditional toxicant baits. The double stranded segments of RNA will have to be protected from the changing pH of the digestive tract and absorbed into the systemic circulation while maintaining viability. Once in circulation, RNAi molecules must be delivered to the target tissues at concentrations high enough to elucidate a physiological response. Advances in bioengineering have given researchers products that both hide RNAi inside stable exterior shells and direct these carrier molecules to the site of action. These cutting-edge technologies make the development of a RNAi-based rodent bait a feasible option. The use of RNAi for rodent control shows promise because of its species specificity and low non-target impact. RNAi sequences are selected to sections of the rodent genome that are significantly different from non-target species genes and therefore do not bind to and degrade the non-target mRNA. Formulating RNAi into oral baits presents challenges but recent advances in bioengineering have led to the development of mechanisms for delivery useful for this application of RNAi. RNAi-based baits will be a great benefit to efforts to eradicate rodents off islands because they will reduce the time and funding necessary to mitigate risks to non-target species and the environment.

The small Indian mongoose is an invasive pest species and rabies reservoir in Puerto Rico and other islands in the Caribbean. In the United States and Europe, rabies in wild carnivores is largely controlled through oral rabies vaccination (ORV), but no ORV program for mongooses exists. The oral rabies vaccine currently licensed for use in wild carnivores in the United States has not been reported as immunogenic for mongooses. A mongoose-specific bait has been developed but field-based bait flavor preference trials have not been performed in Puerto Rico. We evaluated removal of egg-flavored (treatment) vs. unflavored (control), water-filled placebo ORV baits in a subtropical dry forest in southwestern Puerto Rico from 2014-2015. During six trials at four plots we distributed 350 baits (175 treatment and 175 control) and monitored baits for five days or until at least 50% of baits had been removed or were rendered unavailable to mongooses due to inundation by fire ants. The estimated overall probability of bait removal within five days was 85% (95% CI 75-91%) and 45% (95% CI 35-55%) for treatment and control baits, respectively. Removal rate estimates in the spring were 95% (95% CI 86-98%) and 63% (95% CI 49-76%) for treatment and control baits, respectively. Removal rate estimates in autumn were 68% (95% CI 58-77%) and 30% (95% CI 22-39%) for treatment and controls, respectively. Model estimates suggest that treatment and season were more influential on bait removal rates than diel period or experimental day, although bait removal rates were higher at night than during the day, suggesting non-target bait removal by nocturnal rodents. Our results suggest that egg-flavored baits were preferred by mongooses over unflavored baits. During operational ORV bait application, non-target bait removal should be taken into consideration when calculating bait application rates.

The reintroduction of the Mexican gray wolf to the southwest U.S. has been controversial because of documented wolf-livestock conflict (and fear of potential conflicts). Wolf-livestock interactions can lead to economic losses for ranchers directly from depredation and indirectly through physiological impacts on livestock such as weight loss. Ranchers report that, in addition to economic losses, they face additional management costs due to the presence of wolves. Relying on a survey of Arizona ranchers, this study explores ranchers’ attitudes toward wolf reintroduction, identifies and estimates the costs of management practices implemented by ranchers to avoid or reduce wolf-livestock conflicts, and examines how spending on preventative management practices, including the value of ranchers’ time, compares with net returns per cow under three different price scenarios: a low-price, mid-price, and high-price year. Building upon literature that finds ranchers are motivated by lifestyle and other non-monetary benefits of ranching, we posit that factors beyond profit maximization influence ranchers’ decision to implement management practices to limit wolf-livestock conflicts. We find that spending on preventive management practices can be large relative to net returns. We also find that negative attitudes toward wolves are not well correlated with experiences with or losses from wolf depredation. These results illuminate the complexity of rancher attitudes and management decisions, with implications for predator coexistence and conservation efforts.

The common vampire bat apparently is expanding its range northwards in Mexico and seems poised to enter the southern United States. Climate models predict suitable habitat in the U.S. in south Texas and parts of southern Arizona. While vampire bats’ northward range expansion is not unexpected, the fact that this species brings a strain of rabies that impacts livestock and people warrants a strategic response. Annual economic damages from bats are estimated between $7M and $9M, largely associated with deaths of livestock from rabies. To prepare for the emerging rabies issue, USDA Wildlife Services programs in Texas and Arizona have begun training employees to recognize symptoms and respond to bat presence. Surveillance of livestock at sale barns and on ranches is designed to maximize the opportunity to detect bat bites in livestock. Outreach on the issue, via one-on-one training and a DVD handout to landowners along both sides of the border, has been initiated. This paper details the extent of preparations for an emerging disease; quantifies expenditures necessary for a responsive program; and discusses some issues associated with the proximity of vampire bats to the U.S.-Mexico border.

Beginning in 2009, the Army’s Natural Resource Program on O’ahu implemented the first of three ecosystem-scale rat trapping grids of traditional snap traps in the Waianae Mountains using the model outlined in The New Zealand Department of Conservation’s current best practices for kill-trapping rats. Traps were generally checked every two weeks, but bait often remained palatable for just a few nights due to slug interference. Because of the amount of labor required for single set traps, trials with Goodnature A24 self-resetting traps were conducted from 2014-2016. Early findings showed that traps were malfunctioning at a rate of ~25% and there were major deficiencies with the bait and bait-delivery system. In 2016 the bait system was improved when Goodnature developed the automatic lure pump that continually releases fresh bait for ~4-6 months. Other improvements were also made to the A24 trap to decrease the malfunction rate. In 2017, we replaced more than 1,300 snap traps at all ecosystem-scale grids with 1,000 A24s. Tracking tunnels were used as an independent monitoring system to determine rat control effectiveness. At all sites, rat activity measured in the tracking tunnels has been low (less than 15%) for over 18 months. In this paper we discuss the results of the transition from single- to multi-set rat traps, highlights some successes and obstacles, and describes grid spacing and applicability to other sites.

Since 1995, the Army’s Natural Resources Program on O‘ahu has been controlling rodents in O‘ahu’s forests to protect native plants, invertebrates, and birds. Bait longevity and attractiveness are keys to successful rodent trapping. Success is impeded when slugs interfere with bait intended for rodents. Slugs can consume all or a portion of the bait, make it less attractive to rodents via their slime, and large slugs can trigger the traps. The goal of this study was to determine whether food grade citric acid (up to 5% concentration) added to common rodent bait (i.e., commercial peanut butter and Goodnature lure) would repel slugs. We conducted several trials including where captive slugs were offered both a test (0.5-5% citric acid added) and control bait. We found slugs significantly preferred the control bait (i.e., bait without any citric acid) in the two-choice feeding experiment when the bait was Goodnature rat + mouse lure. Goodnature rat + mouse lure is the standard bait used in A24 self-resetting rat traps. The addition of citric acid may improve the longevity and attractiveness of bait to rodents in slug-inhabited environments, thereby having the potential to aid in rodent control programs.

The Berry College campus has a population of white-tailed deer habituated to the presence of humans. One area of the campus contains residential houses, and apartments for faculty and staff. In this area, there have been numerous anecdotal reports, as well as photographic and video evidence, of white-tailed deer exhibiting aggressive behavior, particularly toward dogs being walked on leashes. In addition, deer following individuals or circling humans at a distance, making them uncomfortable, have been reported. The objective of this study was to document and establish the locations, frequency, and types of behaviors exhibited by deer in response to humans walking with and without leashed dogs. During the summer of 2019, a total of 13 instances of aggressive behavior were encountered by residents of the area and the investigators. In a majority of the events, (n = 12), deer would follow individuals with their dogs an average distance of 212.7 m ± 24.8 (100%), circle in front of the individuals, and stop in the walking path (50%), forcing them to turn another direction or be subjected to warning snorts or grinding teeth (33%), and/or pawing behavior (33%). We had proposed to administer a negative reinforcement stimulus in the form of impact of a needleless paint-marking dart delivered by a tranquilizer dart gun. It was hypothesized that the negative reinforcement might reduce the likelihood of individual deer repeating the aggressive behavior. However, no deer presented the aggressive behavior when researchers were carrying the tranquilizer dart gun. The investigators are planning on continuation of this project in 2020.

Integrated Pest Management (IPM) strategies increasingly incorporate natural predators of pest species to reduce the abundance and persistence of pest species in agricultural settings. Specialist predators and parasitoids have been demonstrated to be successful tools to reduce the damage from invertebrate pest species, but less research has focused on the effectiveness of generalist predators to reduce the abundance of vertebrate pest species. To investigate this, we employed a case study of a globally used IPM tactic: the use of barn owls (Tyto alba) for rodent biocontrol. When used as a biocontrol agent, barn owls are typically recruited to the area through the installation of nest boxes in agricultural fields. Anecdotal evidence suggests that barn owls forage within the agricultural fields and reduce rodent pest populations, but no replicated studies that monitor both rodent and owl populations exist to date. We developed models of this system, parameterized using published data on barn owls depredating pocket gophers (Thomomys spp.) and voles (Microtus spp.) in California agricultural fields. Using the model, we sought to understand how effectively generalist predators (i.e., predators that are not numerically responsive to the density of a single prey species) might control vertebrate pest species. In this case, the generalist predator density is determined by the land manager through installation of nest boxes and not dependent on the density of pest species on the landscape. We found that the equilibrium density of the pest population is determined by three key parameters. The first is the scaled half-saturation constant, which describes the density of the pest species when the predator’s feeding rate is half the maximum. The second is the scaled predator density, an index of how many predators are present and preying on the target pest species. The third controls the shape of the functional response of the pest species and describes the strength of the predator’s preference for more abundant prey. To determine how the existence and local stability of pocket gopher or vole population equilibria varied across parameter space, we used bifurcation analysis. This revealed that, depending on predator abundance and efficiency, the pest species can persist at stable equilibria at high abundance, low (or zero) abundance, or both. Bi-stability could occur in biologically realistic parameter values, suggesting that barn owls may sometimes control rodent populations, but it depends on initial conditions. These findings suggest that generalist predators can be an effective tool in IPM strategies, but their success depends on both the predator density and the efficiency of the predator.

To mitigate the economic and environmental costs of rodent pests, winegrape producers in Napa Valley, California, have installed nest boxes to attract barn owls to their properties, but their effectiveness to control rodent pests in vineyards has not yet been thoroughly tested. A rigorous estimate of the number of rodents that barn owls remove from the landscape is a necessary step, and this study aimed to produce estimates of rodent removal and prey species composition by using remote nest box cameras. Results indicate that each barn owl chick received 170.2 ± 8.92 rodents before dispersing from the nest box. Combined with the average number of chicks fledged (3.62 ± 1.40), this finding indicates adults deliver on average 616 rodents per nest box, with low and high estimates ranging from 358 to 899 rodents. With conservative assumptions of owl survival and consumption during the non-breeding season, we estimate a barn family could remove 3,466 rodents in a full year (estimates ranged from 1,821 to 7,563). An analysis linking videography to owl telemetry data suggested that 43% of rodents killed were taken from vineyard habitat, which nearly matches the availability of vineyard habitat around the monitored nest boxes (46%). In contrast, more prey were captured from riparian habitat and fewer from grassland habitat than expected given their availabilities. Our results suggest barn owl nest boxes could contribute meaningfully to integrated pest management. Future research should involve rodent surveys in vineyards without and without barn owl nest boxes.

Coyotes are among the most successful carnivores in urban and suburban environments, which has increasingly led to conflicts with pets and people in southern California. One possible contributor to high coyote population densities and human-coyote conflicts is an abundance of free-roaming domestic cats subsidized by backyard feeding and trap-neuter-release (TNR) programs. To determine if coyotes regularly eat free-roaming cats, we identified prey items in the stomachs of 311 coyotes between 2015 and 2018; specimens were either road-killed coyotes or coyotes taken as nuisance animals. We used two methods to estimate coyote diet: visual identification of stomach contents and molecular polymerase-chain reaction (PCR) analysis of prey remains in stomachs. A total of 245 coyotes stomachs contained identifiable food items, including 200 (43%) that contained mammalian prey (based on hair, bones, and tissue); 178 of these had sufficient tissue from which DNA could be extracted. Combining the two methods, we found cat remains in 35% (n = 86) of stomachs with identifiable mammalian remains. This makes cats the most common mammalian prey item identified, surpassing rabbits and small rodents, and means cats are likely a more common prey than has been reported previously, including in other areas of southern California. We used a GIS approach to compare landscape characteristics associated with known locations of coyotes that ate cats to the same characteristics around locations determined to be TNR cat colonies based on public shelter records. These characteristics included amount and intensity of urban development, coverage of grassland and shrubland, building density, and the distance to the nearest natural vegetation; these were entered into a principal component analysis (PCA) to create composite variables that described the degree of urbanization around coyote and TNR colony locations. Logistic regression of PCA variables revealed that cat-eating coyotes were significantly associated with landscapes that were more intensively developed, had little natural or altered green space, and a higher building density than coyotes that did not have cats in their stomachs. Locations of TNR cat colonies had similar landscape characteristics, with colonies often located in intensively developed areas such as apartment complexes or industrial or commercial zones that are relatively far from natural areas. The subset of coyotes associated with TNR colonies were also highly likely to have consumed cats. Coyotes that had been removed (vs. roadkill) tended to be cat-eaters, suggesting that consumption of pets may have led to targeting these coyotes for lethal removal. The high frequency of cats in coyote diets, combined with the concordance of landscape characteristics associated with TNR colonies and cat-eating coyotes, support the argument that high cat densities and supplemental feeding attract coyotes. Effective mitigation of human-coyote conflicts in southern California may require a ban on outdoor feeding of cats and wildlife, and the removal of TNR colonies that coyotes apparently exploit as an abundant source of food.

Non-native, commensal rodents damage property and represent a significant public health hazard through the spread of diseases. Pest management professionals commonly use second-generation anticoagulant rodenticides, usually placed in tamper-proof bait stations, to control commensal rodents in urban areas; however, there are significant concerns about poisoning of native wildlife, especially through secondary exposure of predators and scavengers that consume prey killed by rodenticides. Behavioral responses of commensal and native rodents to bait stations are not well understood, especially at the urban-wildland interface. The goals of this project were to 1) determine temporal patterns of bait station use by commensal rodents; 2) identify factors that may influence bait station visitation by native species; and 3) devise simple mitigation techniques to reduce wildlife exposure to rodenticides. We deployed Reconyx™ PC800 digital cameras to monitor bait stations placed in 90 residential yards across Orange County, California. Two bait stations, armed with non-toxic bait, were monitored continuously in each yard for approximately 30 consecutive days from December 2017 to August 2018 (Session 1). A subset of 64 of these yards were surveyed again from September 2018 to March 2019 (Session 2). One bait station was placed on the ground, whereas the other was elevated 1-1.5 m in a tree or along a fence to determine if non-target exposure to rodenticides could be reduced by elevating bait stations. Roof rats (Rattus rattus) were the only commensal rodents detected in Orange County yards, and were present at 80% of sites, with average activity ranging from 0 to 9 hours each night. Bait stations were discovered quickly in yards with a lot of rat activity, especially at boxes placed directly on the ground, where rats were recorded in nearly 90% of yards within 10 days of placement. Time to discovery did not differ significantly, however, between ground and elevated stations, with mean times ranging from 7 to 10 days across both sampling sessions. Mean time to enter the bait station also did not differ between elevated and ground stations, varying from 10 days in Session 1 and 11-12 days in Session 2. Rats entered only 59-70% of the bait stations they visited, suggesting some degree of neophobia. Native rodents, including California ground squirrels (Otospermophilus beecheyi), woodrats (Neotoma sp.), and deer mice (Peromyscus sp.) were relatively rare among our 90 sites (13% of yards), and were recorded at elevated stations much less frequently than ground stations. Woodrats, deer mice, and, in one instance, a juvenile Virginia opossum (Didelphis virginiana), were photographed entering bait stations, and non-native fox squirrels (Sciurus niger) were able to access bait by chewing through the top and walls of the bait stations. Yards visited by native rodents tended to be closer to areas of natural vegetation and to have no significant barriers to entry (e.g., solid fences or walls) from nearby open spaces. We identified several ways that our results may help improve the management of commensal rodents at the urban-wildland interface in southern California, where reducing non-target exposure is a critical concern. First, the high level of rat activity in some yards suggests that bait may be rapidly depleted, which could reduce effectiveness. Rats had a behavioral response to bait depletion and replenishment, so it may be useful to monitor bait consumption frequently during the first week of bait application and adjust levels accordingly. We caution, however, that our stations contained bait but lacked rodenticides, so we do not know how the presence of toxicant or the deaths of other rats might alter behavioral responses to bait stations. Second, the fact that rats were photographed entering only a fraction of the bait stations suggests either that cameras missed some of these events or that rats showed some reluctance to enter bait stations, even in the absence of rodenticide. Even in yards where rats eventually entered bait stations, it was 7 or 8 days until their first direct exposure to bait. Because the first mortalities from rodenticide might not occur for more several days, pest management professionals should be prepared to communicate the possibility of delays to customers to prevent them from becoming impatient and taking more drastic (and potentially illegal) measures if results are not immediate. Lastly, given the mobility of many wild predators and scavengers in southern California, if rodenticides are the only option available for effective pest control, special efforts should be made to search for and remove rat carcasses quickly, especially in yards that are accessible to native wildlife or adjacent to natural areas. In open, accessible yards close to natural areas, integrated pest control approaches should first be attempted to minimize risk to non-target wildlife species. If rodenticides must be used to control roof rats, bait stations should be elevated to try to prevent native rodents from consuming bait; this approach may also reduce exposure to children and pets.

Since the year 1500, islands have been home to over 75% of known bird, mammal, amphibian, and reptile extinctions. The majority of these have been caused by introduced species, particularly vertebrates such as rats, mice, cats, and ungulates. Arguably, the most damaging vertebrate taxon on island ecosystems is the rodents. Mice and rats have been implicated in around half of all bird and reptile extinctions. Rodents have been introduced now to over 80% of the world’s islands. Over the last 70+ years, conservationists around the world have been working to recover species and island ecosystems from the impacts of invasive species, particularly rodents, developing systematic approaches and techniques that are guided by principles of island invasive species eradication. The eradication of rodents from islands is not only possible, but has been completed on over 600 islands, from small offshore rocks to 400,000-ha South Georgia Island, with hundreds of native species protected from the threat of extinction. Rodent eradication is becoming a mainstream tool used by managers worldwide. However, there are limits to current technologies and approaches, and, globally, we can only reach a relatively small number of islands and threatened island species. To protect and recover threatened species, conservationists must increase the scale, scope, and pace of eradication of invasive species from islands, and focus on innovation of new tools, techniques, and strategies to be allow restoration on larger and more complex islands. One of the biggest challenges to success is ensuring that the public is supportive and allows pest management tools to be used for conservation purposes. New technologies are on the horizon to improve invasive species eradications, including genetic tools and species-specific toxicants.

Since 2000, the Alameda County Vector Control Services District has received only 10 requests for service regarding bobcats. Seven of those cases have occurred in the past three years. Of those, only one case has involved the bobcats creating a den on a residential property. This is the case study of a mother bobcat and four kittens denning in the backyard of a suburban home in Fremont, California. The eviction of the bobcats from the deck of the vacant home was completed in May and June of 2019. During a four-week period, a wide range of eviction techniques were used, including noise, light, and water harassment, along with eviction fluid and human presence. This experience gave us the opportunity to observe and record the behaviors of bobcats living in a suburban environment and to test various eviction methods that had previously been untested in Alameda County. After four weeks of continuous humane harassment, the eviction was successful.

The impacts of wildlife (especially birds) on agriculture cannot be underestimated, and result in human-wildlife conflicts. This study identified avian pests of rice and maize in two agro-ecological zones of Ondo State, Nigeria. Information gathered from informal interviews and questionnaires administered to FADAMA farmers in local communities, together with direct observations made over exploratory transects, were used to survey avian pests. A total of 27 avian pest species of rice and maize, belonging to nine families and three orders, were recorded in the two agro-ecological zones in Ondo State. In the savanna agro-ecological zone, a total of 18 avian pest species belonging to nine families and three orders were recorded, whereas in the rainforest agro-ecological zone, a total of 21 avian pest species belonging to eight families and two orders were recorded. Farmers revealed that bird damage is a serious problem which usually leads to reduction in yield, harvest quality, and interest in production. Human bird scarers and scarecrows were reported as the most effective control measures against birds’ infestation and damage. There should be a synergized effort towards Integrated Pest Management involving farmers, crop scientists, and ornithologists to ensure the survival of birds in agrarian areas as well as improve crop yield. This approach will not only help farmers but also conservationists.

California ground squirrels (Otospermophilus spp.) are one of the most damaging vertebrate species in California agricultural systems, including rangelands. One of the primary tools used to mitigate damage caused by ground squirrels is rodenticide application. First-generation anticoagulants, such as diphacinone, are the most commonly used rodenticide for California ground squirrels and are applied in one of three methods: 1) spot treatments, 2) broadcast applications, and 3) bait stations. Spot treatments involve spreading bait very thinly by hand around a burrow entrance. Spot treatments tend to be time consuming, so they are generally only used to treat a small number of burrow systems. Broadcast applications require the use of a calibrated seed spreader. This allows easier application of bait to much larger areas. Bait stations house bait within a centralized location. They are used to exclude non-target access to bait by animals larger than the entrance of the bait station. Bait stations are generally believed to have the lowest risk of primary non-target exposure (i.e., potential of rodenticide exposure from direct feeding on toxicant) given this exclusionary advantage, but they have been postulated to lead to higher secondary exposure risk (i.e., potential of exposure to predators or scavengers from feeding on intoxicated animals) if repeated feedings at bait stations result in higher residual concentrations in dead and dying ground squirrels. Other factors can influence secondary exposure risk as well, including the amount of bait applied, the time from bait application until death, and the proportion of carcasses that are exposed aboveground. Anticoagulant rodenticides have come under intense scrutiny over the last several decades, given the potential for secondary exposure concerns associated with their use. Therefore, we established a study in rangelands in central California using radiotransmittered individuals to address how bait application strategy (i.e., spot treatment, broadcast application, and bait station) influences diphacinone (0.005% concentration for spot treatments and bait stations, primarily 0.01% for broadcast applications, although 0.005% was used in one application period; see Baldwin et al. 2020 for additional details) secondary exposure risk in California ground squirrels. Specifically, we addressed: 1) differences in amount of bait applied across application strategies, 2) differences in residual concentrations of diphacinone across application strategies, 3) potential variability in time from application until death across application strategies, and 4) proportion of ground squirrels that died belowground. Bait stations resulted in the greatest amount of bait applied ( = 18.6 kg ha-1), followed by spot treatments ( = 3.2 kg ha-1), and broadcast applications ( = 1.0 kg ha-1). We believe much of the bait that was removed from bait stations was cached, but this was not documented, nor was the potential impact of this caching behavior on non-target species known. This could be an area of additional research in the future. Average time from bait application until death did not vary across application strategies ( = 9.1 days), suggesting that application strategy has little effect on this potential exposure risk. The vast majority of ground squirrels died belowground (82-91%), substantially reducing scavenging concerns. Likewise, most carcasses were severely decayed within three days post-mortality, further reducing scavenging risk. We detected no difference in residual concentrations of diphacinone across the three application strategies ( = 1,399 ppb), although our results from broadcast applications may have been confounded by our primary use of a higher-concentration diphacinone bait. Our use of this higher-concentration bait was in accordance with the label at the time of this study, but it is possible that the use of a lower concentration bait could yield similar efficacy as compared to the higher concentration product while yielding lower residual diphacinone concentrations. This is another area of potential exploration. We did observe substantially lower residual concentrations of diphacinone in ground squirrels that survived the bait application period ( = 112 ppb), suggesting that the greatest risk of secondary exposure lies with scavengers and predators that consume squirrels that would have died from lethal exposure. Collectively, this study helps fill in knowledge gaps into the effect of application strategy on secondary exposure risks. Following proper application protocols, combined with daily carcass searches, should substantially reduce secondary exposure concerns associated with first-generation anticoagulants.

Firearms are often used in lethal control of invasive vertebrates, but safety and regulatory aspects limit the circumstances under which they can be used. During August 2016 at the Brown Treesnake Project laboratory on Guam, we evaluated hobby‐grade Airsoft Electric Guns (AEGs; a lower powered, less‐hazardous, and less‐regulated alternative to firearms) for capture and control of small animals, with specific emphasis on invasive brown treesnakes (Boiga irregularis). Tests of AEGs differing in power with ammunition (plastic pellets) masses ranging from 0.20 to 0.39 g, fired at gelatin blocks from distances of 4, 8, and 12 m, showed that heavy ammunition is of overriding importance for maximizing lethality: 0.39‐g pellets penetrated more deeply at 12 m than did 0.20‐g pellets at 4 m. Inspection of tissue damage in brown treesnake carcasses subjected to fire with the 0.39‐g ammunition from the same distances suggested that injuries sustained by a direct hit from 12 m would often be lethal, and snakes would be unlikely to survive multiple hits from automatic fire discharged at approximately 17/s. Limited trials with live snakes helped us to understand behavioral responses in a snake hit by ≥1 pellets, including distance traveled over time. Based on these factors, we assessed the risk that a snake injured by pellet fire might evade subsequent capture by rapid responders in the proximity. We also discuss ethical considerations and regulatory advantages of using AEGs.

Two fundamental processes in predator-prey interactions have never been exploited in pest control. First, predators constantly make foraging decisions to maximize their energy intake, thereby ignoring unrewarding food cues; and second, predators and prey frequently use chemical mimicry to avoid being recognized. Given olfaction is the primary sense of most mammalian species, we tested whether we could deceive generalist mammalian predators into ignoring the odor cues of secondary prey by repeatedly providing prey odor with no food reward until predators gave up and searched elsewhere. We habituated predators in the wild to bird odor by exposing them to odor extracted from commercially available bird species before native birds (double-banded plover, wrybill, pied oystercatchers) arrived for nesting. We continued the odor treatment during nesting to camouflage the birds’ real odor from that of the extracted odor. Chick production over a 25- to 35-day period in treated areas was 1.7 times greater compared to non-treated areas, after which treatment effects disappeared. These results suggest that filtering of rewarding and unrewarding food cues by generalist predators is powerful enough as a conservation management tool to protect secondary prey species temporarily without removing a single predator.

New Zealand biodiversity is highly endemic and suffers from extreme loss due to habitat destruction and invasive mammalian predators. Building on New Zealand’s expertise in mammal eradications on offshore islands, the New Zealand government recently announced NZ$28 million to initiate the process to eradicate invasive rats, possums, and mustelids from the New Zealand mainland by 2050. Predator Free 2050 Ltd was established to distribute these funds to groups who could demonstrate local eradication and landscape scale suppression of predators, and lever additional funding to achieve their goals. Not surprisingly, this programme has raised a few eyebrows, enthralled people, angered others, or kick-started a tranche of new biodiversity protection projects across the country. Whatever one’s perspective, the bar has been set very high, and consequently has engaged a lot of people in the thinking, planning, and on-the-ground action. This is one of the most exciting times in New Zealand’s history to be either managing predators or studying them. The challenge is huge, and the opportunities tremendous, but we won’t get there unless we address a number of important knowledge gaps. Manaaki Whenua Landcare Research is one of many New Zealand research providers helping to fill these gaps. Achieving eradication of predators at a national scale requires quantum leaps in pest management, including: 1) new tools and strategies for removing predators; 2) more accurate methods of detecting predators at very low abundance; 3) statistical methods for declaring success; and 4) public co-operation and involvement in the programme. Manaaki Whenua scientists have recently developed a long-life novel lure using kairomones emitted by ferrets. Adding ferret odour to regular bait increased stoat detection rates by 200%, and in kill traps, ferret odour increased stoat kills by 150%. Responses were not limited to stoats, as weasel and hedgehog detections increased by similar amounts with ferret odour. We have also developed a free decision-support tool (TrapSim) to simulate the effects of varying trapping and poisoning regimes. The tool is being further developed to include predator reinvasion and the variation between individuals in their interaction rates with control devices. We have also developed software to help managers estimate the minimum amount of surveillance needed to achieve the required sensitivity and the target probability of eradication. Declaring local eradication reliably is a critical part of moving from one eradication zone to the next. Lack of detections does not necessarily mean eradication, as this depends on the detection network and the probability that a predator can be detected if it is present. We modified software created for determining proof-of-freedom from animal diseases for proof-of-eradication of pests. This allows managers to declare the probability of eradication based on Bayesian statistical methods, given zero detections. Motion-triggered cameras are excellent devices to detect pests at very low densities, but they suffer from the time required to process thousands of images. We are co-developing artificial intelligence that automatically culls out images with no animals and learns to identify animal species when animals are present. This will be a huge cost saver. A key component of the success of this programme is demonstrating the biodiversity, economic, and social benefits. We conduct repeated surveys of urban and rural public to gauge the extent to which the programme affects people’s lives in terms of fewer encounters with pests, more encounters with native biodiversity, perverse outcomes, and greater awareness and understanding about their local biodiversity. Regardless of whether the programme succeeds in the next 30 years, the research and technological advances that are already happening will be hugely beneficial for pest management. Technical challenges, however, are only part of the solution. Working in a populated and agricultural landscape, maintaining the initial enthusiasm we are currently experiencing, and ensuring the majority of the population remains on board, are all enormous challenges.

The Utah wild turkey management plan identifies insufficient winter habitat (limited food resources) and low-quality breeding and summer habitat as high and medium priority concerns, respectively, for wild turkey management in Utah. Along the Escalante River, over 70 miles of riparian areas were treated to remove Russian olive and tamarisk, which could improve breeding and summer habitats. A challenge to wild turkey management is predicting how turkeys may adapt to habitat changes. We studied the seasonal habitat use of wild turkeys in Escalante, Utah to determine their basic habitat use and movements, and their use of treated areas. Turkeys had large summer home range sizes (females = 151.1 ±41.5 km2, males = 147.4 ±60.7 km2), suggesting fragmented habitats and limited resources. Summer habitat use showed two distinct patterns: use of lowland habitats along the Escalante River or use of higher elevation ponderosa, aspen, and pinyon habitats. Male and female turkeys frequently used treated pinyon-juniper, treated developed areas (farms, rural housing), and treated agricultural areas in the winter. The proximity of treated natural areas and available roost sites to agricultural fields could lead to conflicts with landowners in the region, particularly if agricultural crops are used by turkeys to replace Russian olive as a food source in the winter.

California ground squirrels have been implicated in causing damage to anthropogenic structures, critical infrastructure, sensitive wildlife species, and agricultural areas in California. Current methods employed to reduce the abundance of California ground squirrels include trapping, shooting, exclusion, fumigation, filling of burrows, natural predation, habitat modification, and use of rodenticides. Recent technological advances in rodent traps provide an opportunity to test CO2-powered, self-resetting traps to reduce California ground squirrel abundance. Goodnature A24 automatic rat+stoat traps deployed in three 80 × 80 m trapping arrays reduced the relative abundance of California ground squirrels on average by 84.8% over a period of nine days. When trapping arrays were compared to control arrays, A24 automatic traps also significantly reduced the relative abundance of California ground squirrels. Inspection of California ground squirrel carcasses indicated that A24 automatic traps successfully controlled adult male and adult and juvenile female California ground squirrels. Although these data are preliminary, A24 self-resetting traps show promise as an effective and efficient means to reduce California ground squirrel abundance, potentially reducing the need to implement less efficient methods and methods that pose a risk to non-target wildlife in higher trophic levels, including rodenticides. Logistical issues, non-target wildlife effects, human safety concerns, and future directions of this research are also discussed.

Possums, stoats, and rats introduced into previously mammal-free New Zealand (NZ) seriously impact our native flora and fauna. As a result, considerable research effort has focused on their control, with excellent success in the eradication of mammals from offshore islands. Unfortunately, we have run out of defendable, non-human occupied islands and the current focus is the NZ mainland, with a new government goal called Predator Free NZ 2050. In 2010, the Centre for Wildlife Management and Conservation started a research programme investigating new control and monitoring tools that could be used on the NZ mainland. More recently (2015) a privately funded research entity called Zero Invasive Predators Ltd (ZIP; both based at Lincoln University) was established with the goal of developing technologies to remove predators from large mainland areas and then defending them from reinvasion. ZIP has since demonstrated that a modified delivery technique for aerial 1080 can achieve near eradication of rodents and possums at two study sites. ZIP was then able to defend both sites in the short term using a virtual barrier of traps and/or geographical features such as rivers. The CWMC (with the Taranaki Mounga Project) investigated the use of self-resetting traps as a ground-based rodent control tool. The traps were unable to maintain average rodent tracking rates below 5% (avg. 11.5%) without regular trap servicing. To maintain low rodent numbers required a trap service every 3-4 months and this is not cost effective compared with alternatives. To improve detection rates the CWMC (with Cacophony Project) have shown that that thermal cameras are 3.6 times more sensitive than trail cameras for detecting possums. These cameras will quickly find survivors and/or reinvading animals using species recognition software combined with wireless communication. As Predator Free NZ 2050 scales up, more pest control will take place near urbanised areas. As such, the next focus should be the development of control tools with higher social acceptance. Surveys indicate the preferred control options are trapping and species-specific toxins. What this feedback highlights is that future research needs to reduce the cost of trapping so that it is affordable for community groups. In addition to this, funding is required for the registration of species-specific toxins, which have higher public acceptance.

Recent failures of tropical island rat eradications have spurred a re-appraisal of the approach required to increase the success rate. A review of some failures implies that direct transfer of the temperate island eradication strategy to tropical islands, based on a ‘seasonal vulnerability’ model, may be erroneous, and particularly so for ‘wet’ tropical islands with year-round breeding by rodents. A new model based on intra-specific competition is proposed, with suggested research avenues to tease out aspects of rodent population dynamics that may make rodents vulnerable to eradication techniques in tropical islands.

The 2019 rodent eradication on 1,455-ha Lord Howe Island was the second and largest attempted on a permanently inhabited island. With 350 residents, it presented numerous novel challenges, resulting in an operation best summarised in four words: Compromise, Commensal, Complexity, and Cost. A ground-based operation was conducted across the built-up portion of the island, some 300 ha, with aerial bait applied on forested higher ground (1,200 ha); brodifacoum was used exclusively for this attempted eradication. Initial community resistance and the presence of mice meant that almost 19,000 external bait stations were established within the Settlement, on a 10-m grid. The intensive grid was expected to result in numerous bait stations within rodents’ home ranges. An additional 3,500 internal bait stations were put in all buildings and 9,500 hand-broadcast points overlapped the aerial and bait station boundaries. Over 60 field staff were employed locally, from Australia and overseas, to run the toxic baiting operation for 5.3 months. On-going resistance from a small community group resulted in two legal challenges early in the operational stage, including one in the Australian Supreme Court. Additional complications included initial active opposition to private land access; unaccepted personnel by the private land owners; resistance to livestock removal requiring novel bait station infrastructure; possible significant bait loss to invertebrates; and a small proportion of rats apparently avoiding bait stations. A fundamental aim of future operations on inhabited islands should be that they are community-led, which is likely to take several years to mature to the operational stage. Eradication practitioners should prepare for a significantly more complex operation with a concomitant increase in resourcing and planning.

Interest in beaver-related restoration is growing in the western U.S. but understanding the basic ecology of American beaver and their population dynamics is often overlooked when integrating beaver into stream restoration goals. Our study investigated the spatial-temporal distribution of beaver colonies and their damming activities to better inform stream restoration projects in the West Fork Cow Creek Basin of the Umpqua Watershed in southwestern Oregon. During fall 2017, we conducted beaver activity surveys at 144 randomly selected reaches predicted to be either suitable or unsuitable for damming, but suitable for beaver occupation. We categorized beaver use at each reach using assessments of their activities and time of last use. We recorded dam structure and impoundment characteristics at all identified dams. Evidence of beaver activity was documented at 57% of locations suitable for dam establishment and 48% of unsuitable dam sites. Beaver dams were found only in reaches identified as suitable for damming and were concentrated throughout two tributaries located on private ownership. Our beaver activity observations will be combined with other data collected in the Umpqua Watershed, and used to construct a probability of use model that will identify dam and non-dam habitat associations. This work will provide novel insights into the landscape ecology of beaver, and inform critical decisions involving trade-offs of ecological benefits and human-beaver conflicts in freshwater systems of interest.

Rodent eradications have contributed to the recovery of many threatened species, but challenges often exist for campaigns that occur on tropical islands when compared to more temperate regions. A post-operational review of a rat eradication operation on Wake Atoll indicated that certain areas, such as those with high alternative food abundance, may have contributed to the failure to remove all Polynesian rats. We conducted a nontoxic bait uptake trial to evaluate whether the maximum prescribed bait application rate for Brodifacoum-25W rodenticide pellets was sufficient to expose all rats to a lethal dose at three sites on Wake Atoll, including around a solid waste aggregation area (SWAA), which was previously identified as “high risk.” We monitored bait persistence and condition throughout the treatment period as well as rat movement via radio tracking. Bait uptake by rats was also assessed by trapping and examination of rat orifices and gastrointestinal contents for pyranine biomarker incorporated into the bait pellets. The rate of bait disappearance differed by site, with bait disappearing the fastest in vicinity of the SWAA. Rat movement also varied by site, with rats observed traveling greater distances around the SWAA, sometimes exceeding 300 m. The SWAA was the only site at which we observed rats negative for biomarker exposure. We suggest that these negative observations resulted from lack of bait availability or movement of rats into the core trapping area from outside the treatment area. However, we cannot rule out preferential selection of alternative food sources over bait pellets and suggest that this possibility should receive further attention. Based on our results, we conclude that, of the three sites, the maximum bait application rate prescribed on the product label was not high enough to provide every rat an opportunity to encounter bait at and around the SWAA. Given the rapid disappearance of bait and the regular immigration of rats from distant habitat, we recommend that an even greater application rate be prescribed and that the heavier treatment be extended over a much larger area surrounding the SWAA.

Understanding residents’ perceptions and expectations surrounding wildlife species that some may consider “vertebrate pests” is an important element in developing socially acceptable, yet ecologically appropriate and scientifically sound, management strategies. Coyotes are a native wildlife species that are sometimes viewed as vertebrate pests. Human-coyote conflicts in southern California illustrate the importance of incorporating the social sciences, particularly knowledge of human behavior, communication and education, in a coyote management strategy. The California Department of Fish and Wildlife has developed Wildlife Watch as a community-based approach to coyote management across eight cities in Los Angeles and Orange Counties. Wildlife Watch (based on the Neighborhood Watch national crime prevention program) uses conservation-oriented principles to empower local communities, agencies, and residents to remove wildlife attractants and to exclude or deter coyotes from neighborhoods. Here, we outline the main components of Wildlife Watch and use case studies of successful programs to identify three common components: 1) multiple methods for residents to report human-coyote encounters and/or sightings (e.g., online, phone), 2) a clearly written policy that all incident reports receive an acknowledgement or response from the city, and 3) strong support from the city’s police department. An adaptive community-based program, like Wildlife Watch, offers a valuable toolkit to managers for navigating the diverse array of human perceptions, values, and attitudes regarding vertebrate pest species and human-wildlife conflicts.

A common adage is “you can’t manage what you can’t measure.” How applicable this saying may be to wildlife conservation and management is debatable; however, understanding the “where,” “who,” and “why” of human-wildlife conflict can help managers evaluate and prioritize incident response and conflict mitigation efforts. It is critical to note that no tracking or reporting system is capable of effectively capturing all human-wildlife incidents. The format and functionality of the tracking system, how the system is advertised to the public, and who manages the system are all important factors in the accessibility, utility, and success of each tracking system. Here, we examine three different systems for tracking reported human-coyote incidents and encounters in California: 1) the Wildlife Incident Reporting system, operated by the California Department of Fish and Wildlife; 2) Coyote Cacher, operated by the University of California - Agricultural and Natural Resources Cooperative Extension; and 3) iNaturalist, a citizen science initiative, operated by a non-governmental organization. We find that because each system offers different incentives to the public (and poses different potential challenges or barriers to reporting), each receives a significantly different volume of coyote reports. Each system provides a unique perspective of reported human-coyote conflicts in California. Understanding these differences and being cognizant of the inherent or potential limitations of a reporting system are crucial for integrated, scientifically defensible, and robust wildlife management, effective policy development, and informed decision-making.

On Aotea/Great Barrier Island, New Zealand, two invasive rat species (Pacific rats and ship rats) pose risks to the ecosystems and challenge the management in two sanctuaries. At Glenfern Sanctuary (83 ha) an eradication has successfully removed ship rats and a predator-proof fence prevents reinvasion. However, Pacific rats persist in low abundance. At Windy Hill Sanctuary (770 ha) intensive rodent control maintains both species at low abundance despite ongoing reinvasion. A capture-mark-recapture study was conducted between February and April in 2016 and repeated between July and September 2017 to determine population densities, confirm species composition, and analyse the effects of time, population density, and interspecific competition on rat behaviour. Live traps were monitored with camera traps to analyse behaviour of rats around traps. Population density and detection probability of Pacific rats varied between times reflecting seasonality in food abundance and rat reproduction. The detection probability of Pacific rats also differed between sites, being higher at Glenfern Sanctuary than at Windy Hill Sanctuary, presumably due to interspecific competition with ship rats. Where Pacific rats were the sole species they were captured in traps in the first night. However, in coexistence with ship rats, Pacific rat detection was delayed by at least ten days. Population density influenced the number of trap encounters and interactions but did not significantly influence the capture rate. Interspecific competition was identified as problematic for monitoring, controlling, and eradicating Pacific rats.

Predator urine is sold commercially and marketed as a deterrent for nuisance wildlife. Previous studies have shown mixed support for this application. We assessed the potential application of coyote urine as a mesocarnivore deterrent at the Ozark Research Field Station in south-central Missouri. The field station is a 4-ha residential university property bordered by state conservation land and national forests. In Fall of 2019, bait stations were deployed at eight sites at the field station. A bait station consisted of one game camera and one bait pile (protein). Each bait station was deployed for 21 consecutive nights (eight sites × 21 nights = 168 trap nights). From days 7-14, coyote urine was deployed at all bait stations. Bait piles were weighed and refilled daily. Camera traps were assessed for battery charge and card storage daily. Bait removal, diversity, species composition, occurrence, activity, and abundance were compared among treatments. Raccoons were the most abundant and active species at all bait stations, and Virginia opossum was the second most abundant. Raccoon occurrence and bait removal decreased during urine treatment; however, raccoon abundance and activity did not change. Bait removal was highest during and post-urine treatment. Our study concludes that coyote urine has limited effects as a raccoon deterrent at our study site.

Bear damage to western larch trees on intensively managed public and private forest lands of the Intermountain West continues to be a problem for forest managers. Bark stripping and subsequent cambium feeding by bears commences upon den emergence in the spring, when foraging options are most restrictive. Various damage reduction methods are often controversial (snaring, hunting, supplemental feeding), or do not always adequately resolve the problem (silvicultural strategies); hence, a need exists for the development of alternative nonlethal techniques to reduce damage. We tested the efficacy of three candidate repellents (Hot Sauce®, Tree Guard™, and grizzly bear feces) to reduce spring/summer bear damage to western larch trees on reforestation units in northern Idaho. Plots were laid out and treated in early June and examined for damage four months later. Thirty-four of 300 (11.3%) treated trees were damaged by black bears. Of the newly damaged trees, the highest damage rates were on the control plots: 15 of 75 (20%) trees. Damage levels to trees on treatment plots were 8-9.3%. Mean area of damage, or bark removal, on newly damaged trees only, was 452.8 cm2. Almost half (47%) of the bear-damaged trees were completely girdled. Chi-square contingency table analysis showed that the damage rate of treated trees was less than of control trees. No difference in mean surface area of damage was detected among treatments. Further testing is necessary to reveal the true potential of chemical repellents for reducing black bear damage to conifers in the spring. A wide array of chemicals should be evaluated for their potential.

Collecting data with wildlife can present several challenges to researchers, especially when the species is a bird nesting at heights requiring a ladder. These challenges can include using time to efficiently to visit large quantities of nesting sites, the costs to purchase new technology, and minimizing disturbances to wild animals. In order to meet these parameters, our team came up with a solution to allow us to do head counts and note general nesting behavior of barn owls, a biological control agent of voles. Our solution cost just $60 to provide equipment for a trained volunteer or researcher to collect this data on their own. Each team member uses a wireless endoscope, a small LED flashlight, a painter’s pole, and a smart phone to collect data in the field without a need for cell phone service. This solution allows more data collectors to participate for less money and time. Otherwise, this would require more costly equipment and the setup and usage of ladders. Usage of said technology can also be an affordable way to advocate for wildlife by mounting permanent cameras to boxes that do not require a wi-fi or an ongoing cellular connection. This setup could also be used to monitor other wildlife in hard-to-reach places with minimal disturbance for a low price.

A 2012 attempt to remove two rat species (Rattus tanezumi and R. exulans) from Wake Atoll was partially successful. R. tanezumi was eradicated from all three islands (Wake, Wilkes, and Peale), and R. exulans was eradicated from Peale. However, R. exulans remained on Wake and Wilkes and have since recovered to very high densities. In 2013, a panel of experts reviewed the eradication operation and offered a list of possible causes of the partial failure. Since that time, further research has been conducted to address several of the issues identified in the review. In this paper, we conduct a current review of the remedial studies, identify remaining knowledge gaps, and make recommendations for ensuring the feasibility of a future operation to remove R. exulans from Wake and Wilkes Islands.

Although coyotes are a natural component of southern California ecosystems, they are sometimes considered a nuisance because their opportunistic habits and tolerance for urban and suburban environments bring them into conflict with people. Recent attacks on people and pets have increasingly led to lethal control of nuisance animals, yet it is unclear whether the demographic distribution of nuisance individuals is representative of the coyote population as a whole. We used two methods, cementum annuli analysis and tooth wear, to estimate the age of coyotes collected as nuisance or road-killed animals in southern California. Age estimates based on tooth wear, a non-lethal method, were broadly similar to those from cementum annuli analysis, although tooth-wear estimates were highly variable and tended to overestimate age, especially for younger individuals. The demographic structure of coyotes collected as nuisance animals was biased toward young adults and males, which suggests that this demographic class may be more likely to killed by control efforts, possibly because their behavior creates greater opportunities for conflict with people.

Invasive rats are among the most damaging animals to native species on many island ecosystems including those in Hawaii. On Oahu Island, U.S. Army Garrison Natural Resources Program currently manages invasive rat populations to protect natural resources by using grids of A24 automated traps, and previously snap-trap grids and rodenticide bait stations. Despite these control efforts generally suppressing rats, some lands with natural resources that are at risk to rat predation are not easily accessible for implementing these traditional rat control methods. In a 430-ha mesic forest on Oahu where ungulates are excluded and site access is limited due to military training and presence of live ordnance, we tested the efficacy of aerial application of anticoagulant rodenticide bait pellets (Diphacinone-50 Conservation), applied in two applications at a rate of 12.82 kg/ha per application. We measured the effectiveness of the rodenticide bait application by deploying tracking tunnels (inked and baited cards to identify rat presence) before, during, and after applications within treated and nearby untreated areas. Due to restricted access, we failed to estimate nest success of an endangered bird; yet previous research showed rat control increases this bird’s population. We also measured diphacinone residues in stream water at the treatment site to determine this method’s risk level to the aquatic ecosystem. The aerial application resulted in immediate and sustained reduction in the rat population, as evidenced by rat activity decreasing from ~44% to 3.8% during the first three months after bait application and maintained <20% rat activity for 10 months. Trail cameras and recovered rat carcasses also highlighted effectiveness. One of 34 stream samples analyzed had detectable diphacinone residues and this single sample was taken one week after application and it had very low levels of diphacinone (below levels quantifiable). Aerial application of diphacinone appears to be an efficient and effective rat suppression technique for natural resource protection in complex landscapes.

The rose-ringed parakeet has been introduced to >40 countries, gaining its status as the most widely introduced parrot in the world. Although regarded as a strikingly beautiful bird by many people, this species is a severe agricultural pest that establishes and experiences a lag time prior to exponential population growth. In addition to agricultural damage, these birds cause noise and fecal pollution, aggregate in large night roosts near human structures, may be involved in disease transfer, and, in some cases, they may help spread invasive plants. In the U.S., rose-ringed parakeets have been reported in several southern states, but established populations are in Hawaii, Florida, and California. Escapees from the pet trade probably account for most introductions, and parakeets have been reported occurring in California as early as 1956. The estimated population in Bakersfield was 3,000 birds in 2012, and additional smaller populations have been reported in San Diego, Anaheim, Santa Cruz, Malibu, and Pasadena. Much of California, excluding expansive natural areas unoccupied by humans, is potentially at risk of rose-ringed parakeet colonization, and this species represents a conceivably important threat to California agriculture. Rose-ringed parakeets are known to consume and damage crops such as rice, sunflower, safflower, rapeseed, and citrus in native India; almonds in rural Italy; and corn and fleshy fruit (e.g., guava, mangos, lychee, papaya) in Hawaii. In 1975, the California Department of Agriculture estimated that the potential crop losses due to a well-established rose-ringed parakeet population could reach US$735,000 annually, a value that resulted from an estimate of their damaging 0.1% of crops grown in the area. Rose-ringed parakeets are a serious agricultural pest in several areas outside of California, and all areas where they thrive are in human-altered landscapes. An update is needed on this species’ distribution, population changes, and impacts to agriculture within California. Furthermore, we recommend establishing an annual monitoring plan of the established California populations, as well as efforts to prevent their spread and reduce their threat to California agriculture.

Heron streamers/droppings, which are electrical conductors, have caused power outages at electrical substations and transmission towers. It is necessary to understand the effectiveness of techniques for scaring herons to prevent electrical outages caused by their streamers. Here, we focused on bird-management techniques using visual deterrents and evaluated the deterrent effects of light-emitting diodes (LED), lasers, and a robotic approach. Grey herons were observed on the outdoor steel structure of the High Power Testing Laboratory of the Central Research Institute of the Electric Power Industry in Yokosuka, before and after dawn. LED lights were attached to the uppermost steel structure of the laboratory to evaluate the deterrent effects. A green laser was installed 60 m north of the steel structure. The lights and laser were fired manually when herons landed on the steel structure and we evaluated the response of the birds to the stimuli. We also installed a commercial mobile robot, which was programmed to move underneath the steel structure after dawn. To investigate the deterrent effect of the size of the robot, we put a scarecrow on the robot and compared the proportions of herons deterred by the robot with and without the scarecrow. Both the LED and laser were effective before dawn, but their deterrent effects decreased significantly as the ambient illuminance increased. In the robot experiment, more than 78% of the herons were deterred from the steel structure when the scarecrow was attached, while less than 6% of the herons were deterred when we used the robot without the scarecrow. Our findings suggest the effectiveness of visual scaring techniques for herons. Bird-management strategies combining LED, lasers, and robots may be useful for deterring herons during both the day and night.

The only pesticide currently registered for mongoose control is a product developed for rats that consists of a hard-cereal bait block. Although the active ingredient (diphacinone) is known to be highly effective for mongoose, previous studies indicate that carnivorous and omnivorous mongooses do not readily consume the hard bait matrix designed for gnawing rodents. A palatable bait matrix with a consistency more appropriate to mongoose dentition and feeding behavior will be required to develop a more effective mongoose pesticide. We evaluated the acceptance and consumption of nontoxic versions of four candidate bait matrices: FOXECUTE® and FOXSHIELD® (Animal Control Technologies, Australia; ACTA); HOGGONE® (ACTA); and a potted pork shoulder loaf containing artificial dead mouse scent developed by WS-NWRC as a bait for invasive brown treesnakes (hereafter ‘BTS bait’). We offered test groups of six mongooses one of the candidate bait matrices alongside dry dog kibble dog food as a challenge diet for five days. Because the potential active ingredients para-aminopropiophenone and sodium nitrite require accumulation of the toxicant within a relatively brief period of time to affect lethal toxicity before they are metabolized, we conditioned mongooses to feeding within only a four-hour window rather than slowly sampling the bait throughout the day. We estimated rate and amount of consumption through review of time-lapse photography of feeding trials and measured total consumption by weighing uneaten portions of bait. From the first day offered, most mongooses readily consumed ample amounts of all four bait matrices and consumed almost no challenge diet. Overall, consumption was highest and most consistent with the BTS bait. Although this trial did not clearly discriminate an optimal bait matrix, this result is highly encouraging in that we have multiple palatable options. The final selection will be based on other characteristics of the bait matrix such as longevity in the field, compatibility with the selected toxicant, and ease of manufacture, storage, and use. We provide an overview of some of these characteristics for each candidate bait type.

The economically and ecologically catastrophic introduction of invasive brown treesnakes to the Pacific Island of Guam has long served as a cautionary tale about the dangers of invasive species and the seeming impossibility of their management on a landscape scale. USDA Wildlife Services and federal and private partners have engineered a system for the automated manufacture and aerial delivery of toxic baits for landscape-scale suppression of brown treesnakes in large and remote forest plots. The helicopter-borne dispensing module can launch four bait cartridges per second, and a single payload of 3,600 cartridges can treat 30 ha of forest at 120 baits/ha in 15 to 30 minutes depending on flightpath efficiency. In this paper we recap the research, development, testing, and implementation of the system, including the procedures for monitoring biological responses to bait applications during an experimental suppression within a 55-ha forest plot surrounded by a snake-proof barrier.

Use of M-44 sodium cyanide devices has been opposed by various groups contending that M-44s threaten human health and safety and endanger non-target animals. In Montana, M-44 sodium cyanide devices may be used by non-USDA-Wildlife Services individuals licensed by the Montana Department of Agriculture. This paper summarizes the use data submitted by these non-federal applicators between 2006-2019. The data includes use records, take (both target and non-target), and livestock loss reports. It is hoped that this information provides additional data and context to inform the debate over this controversial predator management tool.

Wildlife control operators have observed that animals caught in plastic-walled traps during summer conditions were showing signs of heat stress. To verify the plausibility of these anecdotes, three temperature and humidity sensors were used to monitor local environmental conditions as well as those inside a covered cage trap and a plastic-walled trap. The study lasted for more than 365 days. Average temperature/humidity for ambient air was 10.8°C/72%; for covered cage traps it was 12.0°C/79%; and for plastic-walled traps it was 12.75°C/78.4%. Results are consistent with anecdotal observations that animals caught in plastic traps would experience higher temperatures. This information can be used to improve animal welfare during cage and box trapping activities.

Predator issues related to interactions with livestock and poultry represent a state, national, and international issue. Predators are important for healthy ecosystems; however, with increased interactions between predators and agricultural animals there is also an increased likelihood of predator depredation. The 4-H Youth Development Program is a national nonformal youth education program. Annually, over 1.5 million youth participate in 4-H Animal Science projects, mainly agricultural animals. 4-H youth who raise agricultural animals face issues of predator-livestock interactions. A 4-H curriculum was developed to help youth learn about predator-livestock interactions and strategies to mitigate these issues through improved animal husbandry. Pilot research on the use of the curriculum revealed improved knowledge among participating youth, as well as the application of mitigation strategies to youths’ own practice and outreach to their communities through a service-learning project.

A non-toxic liquid fertility control bait for rats has recently become commercially available (ContraPest® from SenesTech, Inc.). This product contains two chemicals, both of which impair spermatogenesis in male and reduce ovulations in female rats. We tested the efficacy of this bait in wild-caught adult black rats from the island of Hawai’i in a short-term laboratory trial. A control group (n = 25) was offered placebo bait and the treatment group (n = 25) was offered fertility control bait, both ad libitum, during a 15-day introduction period and during the first of four breeding rounds, for a total of 58 days of exposure. After treatment, all rats were provided placebo bait for the remainder of the study and randomly paired with mates from within their treatment groups for two additional breeding cycles. Treatment and control groups comprised 10 breeding pairs each, with random re-pairings between breeding rounds. The treatment group produced no litters during the first and second breeding rounds, while 70% of the control females produced litters. In the third breeding round, 70 days after stopping treatment, the treatment group produced three litters (six pups) compared to seven litters (24 pups) in the control group. During a fourth and final breeding round, control rats were crossed with treated rats, producing six litters (27 pups) from treated dams and nine litters (40 pups) from control dams, indicating no apparent infertility effect 99 days after cessation of treatment. This study demonstrates that the reproduction rate of wild-caught black rats can be chemically suppressed if provided ad libitum access to the fertility control bait under laboratory conditions.

Island rodent eradication is often a prerequisite for ecological restoration. These operations have been scaling up in size and complexity, and typically revolve around the thorough distribution of rodenticides in bait stations, by hand broadcast, by helicopter-borne spreading buckets, or by combinations of these methods. Many of the requirements of an eradication can be met by simply purchasing the right materials and following published best practices. However, intangible factors such as training and mindset of personnel are equally critical, and less commonly understood. We briefly review these factors and highlight issues such as the limited pool of experienced eradication practitioners, the increasing complexity of eradication projects (in scale, number of species to be eradicated, nontarget species, and integration with resident human populations), and potential for human error. We close by making recommendations for addressing some of these less-tangible factors and conclude that the “little things” can influence the outcomes of rodent eradication projects.

Rangeland livestock operations in California and elsewhere are increasingly turning to livestock guardian dogs (LGD) to protect their herds from predators. LGD success depends on a variety of factors including social bonding, environmental and operational context, and individual behaviors. Observation and first-hand experience with LGD on foothill rangeland, Sacramento Valley cropland, and Sierra Nevada/Great Basin rangeland can provide practical evaluation of historic and current research regarding LGD efficacy, breed differences, and economic costs versus benefits. However, little is known about the relationship between LGD and livestock behavior and forage utilization. This article synthesizes current research, experiential knowledge from practitioners, and new frontiers for LGD research. An updated understanding of the principles of bonding LGD pups to livestock will improve LGD success rates and reduce costs for producers.

Camera traps provide an unobtrusive means to monitor wildlife presence and behavior. Yet there is a steep learning curve associated with their deployment. Camera model, settings and position, target behavior, and technicians’ skill greatly influence the success of camera trapping. Furthermore, data storage and management are complex, as copious photos occupy considerable storage space. Finally, evaluating large numbers of digital images is time-consuming for low frequency events; in each of our trials we amassed 10,000-50,000 photos, of which 6-20% were target animals. The application of artificial intelligence (AI) to digital image datasets can greatly increase efficiency, but few existing algorithms have been trained on small animals. We embarked on a camera trapping project to assess interactions of target (rodent) and non-target (bird) species with 125 GoodNature A24 rat traps deployed in rainforest sites on Kauai, Hawaii, following several observations of non-target mortality. While our long-term goal was to use camera trap data to suggest modifications to traps that would maintain target kills while minimizing bycatch, the short-term goal presented in this manuscript focused on perfecting our camera trapping program and AI to classify photos of small animals. Specifically, we described lessons learned regarding 1) the performance of several camera models, 2) camera placement, 3) data management, and 4) artificial network training and development. First, we report on field studies assessing Bushnell TrophyCam HD, Bushnell HD, Reconyx HyperFire, and Reconyx HyperFire2 models on a variety of settings, distances, and angles with respect to the traps. Camera model and placement at traps are critical to capturing images amenable to AI development, as is variation in the training dataset. Second, we outline our data management and sharing protocols. Third, we describe the development of preliminary AI models to review and sort camera trap data. Early models reduced the workload of reviewing camera trap data by correctly classifying photos of rats, birds, humans, pigs, and empty frames. We expect these results to further improve with more training data. These results will greatly enhance the efficacy of several camera trapping studies that we have recently undertaken and help us modify traps to avoid bycatch.

Zinc phosphide has been recently approved in Europe as a vole control product. Currently, only one formulation (lentils/pellets) is marketed with 0.8% Zn3P2. It is applied with a delivery device for burrow baiting. In many instances, zinc phosphide poisoning has been confirmed in non-target species (primary poisoning). In order to be prepared for potential non-target poisoning incidents in wildlife, the SAGIR network, FREDON Franche-Comté, and University of Franche-Comté conducted a field study on common voles to test the sampling method and storage impact under realistic field conditions on the detection of zinc phosphide. The toxicology laboratory of Vetagro Sup, member of SAGIR, worked on the improvement of the World Health Organization WHO technique in order to lower the Limit of Quantification (LOQ) and to validate the technique for the correct identification of field cases. The specificity was tested on 20 gastric content samples (100%), and the LOQ was established at 0.01 g/l (i.e., a 100-fold decrease as compared with the 1995 WHO technique). Zinc content was measured by Flame Atomic Absorption Spectroscopy and non-poisoned animals were tested to check baseline values and to estimate recovery of spiked samples (94-102%). Quantification of zinc in the liver of poisoned versus control animals was also performed. A total of 30 voles were collected in treated and control fields and submitted for analysis. Technicians were not aware of the poisoning status of the animals when performing analytical investigations. Twenty-one individuals were trapped in the control area and nine in the treated fields. Phosphine and high levels of Zn could be detected in eight of the nine intoxicated individuals. Phosphine was not detected in the control voles. Zinc concentrations in the gastric content were significantly different between negative control and exposed animals, but liver concentrations of zinc were similar.

Field trials are reported in this paper on a new bait containing 1% norbormide. Two separate field trials were recently completed on commercial chicken farms in South Auckland, New Zealand. Norway rats were abundant both inside the farm sheds and around the surrounding farmland. Monitoring undertaken before toxic baiting recorded high levels of rat activity. Post-treatment monitoring found no rat paw prints in any of the tracking tunnels from Site 1, and in only one tunnel at Site 2. The decrease of 100% and 96%, respectively, represented a marked reduction in the Norway rat population at both sites.

Wild birds cause significant damage to dairy farms through the consumption and spoilage of cattle feed. A survey of Washington State dairy farmers revealed approximately $14 million in bird damage losses for the Washington State dairy industry, annually. Furthermore, farms that reported the presence of more than 10,000 birds per day were more likely to report the presence of Salmonella spp. or Johne’s disease (caused by Mycobacterium avium subsp. paratuberculosis). Over the course of three years, we assessed the impact of bird populations on the presence of bacteria in bird feces and the nutritional composition of cattle feed. Five dairies were enrolled into the study and visited to collect bird fecal samples and cattle feed samples. Several pens were monitored on each dairy. Bird fecal samples were analyzed for three bacterial populations. Fresh and bird-depleted feed samples were analyzed for dry matter, total digestible nutrients, protein, crude fiber, ash, fat, and net energy. The prevalence of bacterial populations in bird fecal samples did not differ among farms, but Campylobacter jejuni, a bacterial strain known for causing abortions in cattle, was discovered in one location. The number of birds observed at the feed bunk and the percentage of nutritional loss in cattle feed differed among pens. Understanding where birds prefer to feed on dairies may improve the effectiveness of bird deterrent management techniques. A variety of bird deterrent methods are available for dairy farmers but, at best, the most commonly used methods were considered only “somewhat effective” by farmers. The use of more sustainable methods, such as attracting native birds of prey to dairies, may be beneficial to dairy cattle well-being as well as dairy farmer economic sustainability.

Since the Pocatello Supply Depot (PSD) was created in 1936, the PSD has gone through many transformation and organizational changes. The most recent of these changes occurred in 2014, when the PSD transitioned from a cooperative service agreement between the U.S. Department of Agriculture (USDA) and the Greater Pocatello Chamber of Commerce, to a fully federal facility within USDA. Despite the many organizational changes, the purpose of the PSD has remained the same. The PSD manufactures and provides specialized wildlife damage management materials and services that are not readily available from commercial sources, for use by USDA Wildlife Services (WS), other Federal and non-Federal government entities, and the public. The products produced and distributed by the PSD have changed over time to meet the needs of those managing wildlife damage. The PSD produces and/or distributes gas cartridges, zinc phosphide and strychnine grain baits, zinc phosphide concentrate, DRC-1339, synthetic fatty acid lures, Neutroleum Alpha deodorizer, M-44 capsules and components, sylvatic plague vaccine baits, and warning signs. Besides products that are distributed directly from the PSD, the PSD processes orders for other Animal and Plant Health Inspection Service (APHIS) products such as brodifacoum conservation rodenticide baits; diphacinone conservation rodenticide bait; acetaminophen tablets and acetaminophen mouse baits for brown treesnake control; GonaCon immunocontraceptive vaccines for deer and horses; and livestock protection collars. The PSD works closely with WS operations, the WS National Wildlife Research Center, and APHIS Environmental and Risk Analysis Service to maintain its products’ pesticide registrations with the Environmental Protection Agency and state pesticide regulatory agencies.

Island rodent eradication operations have been remarkably successful at eliminating damages caused by these harmful vertebrate pests. As efforts increase in scale and complexity, so does risk of eradication failure. In this paper we present the example of a partially successful rodent eradication project to highlight how best practices and lessons learned are being integrated to reduce risk of failure during a second attempt. In 2012 the U.S. Air Force (USAF) commissioned an attempted eradication of two rat species from Wake Atoll in the Western Pacific. Asian house rats were successfully eradicated, but it was soon confirmed that some Polynesian rats survived; population numbers have since soared. A panel of outside experts was asked to review the project and identify factors that may have contributed to failure. The USAF and Wildlife Services National Wildlife Research Center (NWRC) have used this report as a road map for further studies addressing issues including bait delivery strategies, bait application rates, and alternate bait formulations. A subsequent data gap analysis conducted by USAF, NWRC, and Island Conservation documented technical advances in the intervening years that address risk factors identified in the original review, and highlighted remaining needs including development of a community outreach component and refinement of baiting strategies for inhabited areas and abandoned structures. This exchange of knowledge and expertise among cooperating organizations is helping to refine feasibility assessments and address lingering knowledge gaps. These efforts include a review of other failed rodent eradications that were redone successfully. Ongoing studies continue to resolve areas of uncertainty, and results are being integrated into operational planning for a subsequent eradication effort on Wake Atoll. This process highlights the importance of ongoing refinement of best practices, incorporation of lessons learned, and transfer of knowledge to the wider eradication community.

Homeless encampments are a persistent feature in the city of Oakland, California. Unsanitary conditions in these camps can contribute to large populations of Norway rats and associated vector-borne disease. Alameda County Vector Control Services District has developed a surveillance program for safe and efficient data collection in these encampments. This program includes outreach to residents, effective live-trap ping, ectoparasite collection, vector species suppression, and coordinating with other agencies. The District’s operations around Oakland’s pilot “Tuff Shed” homeless shelter strategy are presented as a case study for Norway rat and Oriental rat flea risk assessment and control.

Habitat manipulation is an important technique that can be used for controlling rodent damage in agricultural ecosystems. It involves intentional manipulation of vegetation cover in habitats adjacent to active burrows of rodents to reduce shelter and food availability and to increase predation pressure. The current study was conducted in the Pothwar Plateau region of Pakistan during respective non-crop periods of wheat-groundnut (post-harvested and un-plowed/non-crop fallow lands). The pur­pose was to assess the impact of reduction in vegetation height of adjacent habitats (field borders) on rodent richness and abun­dance. The study area was divided into two sites: treated and non-treated. At the treated sites, habitat manipulation was carried out by removing crop cache and non-crop vegetation over 10 cm in height to a distance of approximately 20 m from the fields. The trapping sessions carried out at both treated and non-treated sites adjacent to wheat-groundnut fields were significantly different (F2,6 = 13.2, P = 0.001) from each other, with the maximum number of rodents captured from non-treated sites. There was a significant difference in the overall abundance of rodents (P < 0.05) between crop stages and between treatments in both crops. The manipulation effect was observed with respect to damage to crops and yield production, significantly reducing damage within the associated croplands (P < 0.05). The outcomes of this study indicated a significant reduction of rodent population at treated sites due to changes in vegetation height and cover, which directly affect habitat and behavior attributes (e.g., food, shelter, movements, increased risk sensitivity, and feeding behavior) for rat. Rodents apparently were unable to reach levels where they could cause significant crop damage. This method is recommended as a cost-effective and easy application.

Round-tailed ground squirrels are common residents of natural areas throughout most of the desert southwest region of North America. They live in colonies of several adults, subadults, and young, and are diurnal during the active season that ranges from March to September. They are well adapted to desert life and live in burrows they excavate in the ground, but will also modify and occupy burrows created by other animals. Round-tailed ground squirrels are frequently seen in many human community environments. Their burrowing is usually not a significant cause of concern, nor do they cause severe damage to humans or their property. However, they very often cause concerns due to human-wildlife interactions that may include the squirrels themselves, but also their predators such as rattlesnakes, coyote, feral dogs, and other large mammals. Another cause for concern is that round-tailed ground squirrels can be hosts for fleas and other parasites and could vector plague or other diseases during human interactions. An integrated pest management strategy was successfully used at an archaeological site in Arizona to manage activity of these rodents, using a combination of monitoring, trapping, and repellents.

In Modoc County, located in northeastern California, there is a high elevation sage-steppe rangeland ecosystem heavily populated by wild horses and managed primarily by the United States Forest Service (USFS) called the Devil's Garden Plateau. Wild horses have significantly exceeded (roughly 2,000 horses) appropriate management levels (206-402 horses) in recent years and expanded their range outside of the designated territory (258,000 acres) and onto private and tribal lands (nearly 500,000 acres). Increased pressure from wild horses on the multiple use mandate of Forest Service lands have put strains on livestock, wildlife, and the local rural economy. The Modoc National Forest has decreased grazing by roughly 5,000 AUMs (animal unit months) on the Devil’s Garden Plateau due to excessive wild horse use. Each lost AUM results in a decrease of $57.43 - $144.70 of income to Modoc County. Three helicopter gathers in recent years have removed over 1,500 horses from the Devil’s Garden Plateau. Due to the extensive collaboration between the USFS, Modoc National Forest, and local partners, many of these horses have found new homes. This was done by using a tiered pricing approach and offering sales with limitations for older and unadoptable horses. The addition of a robust social media campaign run by volunteers has created a brand for the Devil’s Garden Plateau horses and educated the public on the declining range condition and horse health. Although good progress is being made to gather and place horses in private care, more diverse management solutions may be needed in the future.

The high toxicity and effectiveness of anticoagulant rodenticides have led to their widespread use for controlling rodent pests; however, significant concerns remain about the potential exposure of non-target wildlife species at the urban-wildland interface. Such species can be exposed by consuming toxic baits directly, or indirectly, by scavenging rodenticide-killed prey (secondary exposure). To investigate opportunities for secondary exposure, we used Reconyx digital game cameras to quantify the fates of 20 rat carcasses placed in residential backyards in Orange County, California. We anchored rat carcasses to the ground and then followed their fates for seven days or until carcasses were removed. We also recorded yard characteristics (e.g., vegetation density, permeability of exterior barriers, presence of pets, water, and anthropogenic foods) to help explain variation in carcass removal rates between yards. Rats were discovered fairly quickly, with 35% of carcasses visited within 24 hours. Thirteen carcasses (65%) were removed within seven days, with Virginia opossums and corvids removing the most carcasses (9/13). Coyotes, free-roaming cats, striped skunks, and black rats also consumed rat carcasses, which, by the end of the trials, had attracted scavenging arthropods that then also appeared to be eaten. Yards from which carcasses were routinely removed had relatively low vegetation density; pets, water sources, and anthropogenic foods; and barrier types that permitted movement by wildlife into the yard. Our results improve our understanding of the routes by which native carnivores and scavengers are exposed to rodenticides in suburban settings and can be used to improve pest management practices.

Urban coyotes are commonly exposed to rodenticides used to control non-native commensal rodents, but these rodents are rare in published accounts of their diets. An alternative source of rodenticide exposure is through the consumption of mesocarnivores that have themselves eaten either toxic bait directly or poisoned rodents or invertebrates. Carcasses of 311 nuisance and road-killed coyotes from suburban and urban areas of southern California were collected from 2016-2018. Stomachs were dissected and prey items were identified visually. Stomach contents containing tissue from suspected mammalian prey (N = 178) were homogenized and DNA was extracted. Genus-specific primers (123-366 bp) were designed for Virginia opossums, raccoons, and striped skunks, regionally common species that are known to be consumed by coyotes. PCR was performed for each primer pair, and presence of PCR products of particular amplicon lengths were determined by gel electrophoresis. Coyote stomachs containing a PCR product of the appropriate size were considered to contain that prey item. Land use data were used to assess landscape factors that are associated with the consumption of mesocarnivores. Combining both techniques, mesocarnivores were detected at low frequencies: opossums (8%) were more common than raccoons (2%) and skunks (2%). Some 72% of meso-carnivores present in stomachs were detected by molecular methods, while 66% were identified by morphological methods. Opossums were associated with increased development and anthropogenic land use, while skunks were associated with large natural areas, and raccoons used all habitat types. The extent to which mesocarnivores themselves eat poisoned prey remains unknown, although they may be a potential source of exposure for coyotes. Additionally, landscape factors do not appear to be related to raccoon consumption but may influence presence, and therefore consumption, of skunks and opossums.

The introduction of rodents to islands poses a threat to native fauna, which often have no adaptation to defend their offspring or themselves from predation. To combat predation of nests and brooding females, the Kauai Forest Bird Recovery Project (KFBRP) has deployed 425 Goodnature A24 self-resetting rat traps at two field sites where high densities of native forest birds remain. One site is fenced to exclude invasive ungulates. KFBRP conducts routine trap checks every four months to assess bait and trap function and count carcasses. Typically, we find 0-3 rat or mouse carcasses at a trap, but in November 2018, we found a dead bird under a trap at the fenced site. We assume that traps kill more animals than indicated by carcass counts, because 75% of traps have counters that record when traps fire, and counter tallies exceed carcass counts. Thus, we hypothesize that some carcasses are scavenged or decompose between trap checks, and as a result we are a) underestimating target mortality with carcass counts and b) failing to detect non-target mortality. To test our hypotheses, we placed 30 non-native bird carcasses on transects in the fenced trapping grid in early December 2018. Carcasses were surveyed every at 10, 20, 45, 90, and 130 days after deployed. At the end of survey period, 19 (63%) carcasses could be easily detected, suggesting that we are detecting most carcasses after four months unless they are scavenged. Furthermore, our findings suggest that we would detect non-target mortality if it was prevalent.

Unmanaged cottontail rabbit populations can cause significant damage to drip irrigation tubing. Common integrated pest management strategies to reduce damage include trapping, exclusion, and repellent use. Trapping and exclusion, while effective at managing cottontail rabbits, are impractical when applied to large scale habitat restoration projects. To evaluate repellent use under these conditions, we conducted a preliminary conditioned avoidance field trial using anthraquinone applied to drip irrigation tubing installed in a riparian habitat undergoing restoration in Silverado, CA. The postingestive repellent, anthraquinone, was selected due to prior laboratory research indicating its effectiveness in inducing conditioned avoidance feeding behaviors in cottontail rabbits. Following a complete repair of the irrigation system, alternating sections of the irrigation tubing were treated. After the first treatment, there was an estimated 50% reduction in damaged tubing between the treated and control sections. An estimated 0.18% of the total tubing surveyed was damaged after the second treatment. Between the first and second treatments, we observed an estimated 99.5% decrease in total damaged tubing. Our results suggest that anthraquinone may be successful in reducing cottontail rabbit damage by inducing conditioned avoidance to drip irrigation line. As a preliminary study, these findings are promising and warrant future field trials to validate the use of anthraquinone as a repellant to reduce damage by cottontail rabbits.

The City of Oakland is the largest city within Alameda County, the 8th largest city in California, and the 45th largest in the United States. Due to various socioeconomic factors, the number of homeless encampments within Oakland has been increasing over the past few years. A recently completed survey showed that there has been a 47.45% increase in the number of homeless living within the city limits. Approximately 4,071 people are now living in various encampments around the city, primarily concentrated underneath freeway/infrastructure overpasses and on adjoining lands. Surveillance by our staff found that several of these encampments also had active Norway rat populations as indicated by active burrows within and adjacent to the camps, as well as resident reports. Beginning in the fall of 2017, District biologists began live-trapping at a few of the larger encampments to try and ascertain the composition and load of ectoparasites on corresponding Norway rat populations. We specifically looked at flea abundance and species composition, as they are vectors of diseases, such as Murine typhus (Rickettsia typhi), flea-borne typhus (Rickettsia felis), and plague (Yersinia pestis). We trapped at four different camps in Oakland over a nine-month period and found that the flea, mite, and louse abundance, along with species composition, varied between the camps. To date, we have yet to determine the causes of these differences, but fleas tested at two of the four camps have come back as positive for R. felis. We are continuing to trap at these camps and are expanding our trapping program to include additional camps in an effort to determine what variables affect the ectoparasite composition on Norway rat populations within the City of Oakland.

As of January 2019, a survey of homeless people in Alameda County, CA, documented 8,022 homeless individuals countywide, including 6,312 unsheltered people. Dozens of homeless encampments exist throughout the County, and most lack sanitation facilities (e.g., rodent-proof garbage storage and weekly waste removal; sanitary toilets; and running water for hand washing, showering, or laundry). This situation represents a breakdown of the sanitation interventions that can lead to the outbreak of vector-borne disease (e.g., 2020 outbreaks of flea-borne typhus in southern California). Some characteristics of surveyed encampments make them more or less suitable for Norway rat population establishment and the exhibition of problems associated with rodent ectoparasites. Many urban encampments are in parts of Oakland that have old infrastructure including sewers that support an endemic population of Norway rats and where there may be undetected sewer breaks that allow these rats into these homeless encampments. Anecdotally, before the introduction of unmanaged leftover food sources (garbage), rats would go virtually unnoticed due to their low population, but with a regular supply of food, their populations apparently skyrocketed. Unmanaged rodent populations at homeless encampments are a major public health concern for local public health, particularly considering that ectoparasites such as fleas are potential rodent-borne disease vectors.

Fleas infesting urban wildlife have been epidemiologically linked to the transmission of flea-borne rickettsial pathogens in urban and suburban areas of Orange County, California. To understand the prevalence of flea-borne rickettsioses caused by either Rickettsia felis or R. typhi, a survey of fleas from wildlife was conducted to determine the flea species composition of host animals and prevalence of rickettsial pathogens in fleas on host animals. This study reports flea abundance, species composition, and infestation intensity on unowned domestic cats and wildlife (i.e., coyotes, opossums, rabbits, skunks, squirrels, raccoons, and commensal rodents) collected in urban neighborhoods of Orange County. The survey revealed presence of the northern rat flea on eastern fox squirrels, and widespread distribution of the human flea on skunks and coyotes in Orange County. The flea index and prevalence of flea-bone rickettsioses in fleas has been used by the Orange County Mosquito and Vector Control District to guide decisions regarding risk management and intervention strategies to reduce and prevent the transmission of flea-borne pathogens. The prevalence of R. felis and R. typhi in fleas in Orange County was 8.94% and 0.39%, respectively. Roof rats, eastern fox squirrels, and striped skunks had the highest diversity of flea species, while the Virginia opossum had the lowest, as determined by the Simpson’s Diversity Index. The sticktight flea was found to have the highest diversity of mammal hosts. It is not known how flea species composition on hosts impacts the maintenance and persistence of rickettsial and other pathogens in fleas from urban wildlife in Orange County.