Wildlife Damage Management

Contributed Paper
ROOM: Room 130 – Cimarron

1:10PM Civil Airports from a Landscape Perspective: Implications for Reducing Damaging Bird Strikes
Morgan B. Pfeiffer; Tavis L. DeVault
Collisions between birds and aircraft are a global problem that jeopardize human safety and cause economic losses. Preventing bird strikes includes deterring wildlife from areas of high traffic with habitat manipulation and predicting areas of high avian use in the landscape. However, no landscape level analysis of factors that influence bird strikes has been conducted. We investigated the effects of landscape structure on the rate of damaging bird strikes at 270 Part-139 certificated airports in the conterminous United States. Landscape level drivers were identified at three spatial scales (3, 8, and 20 km). Number of damaging bird strikes to air carriers, as reported to the U.S. Federal Aviation Administration, per air carrier movements were calculated from 2010-2015 for each airport. Landscape structure (heterogeneity, connectivity, and edge habitat) and composition (percentage of landscape, number of patches, and distances between patches) were calculated within the three spatial scales from the airports. The importance of the landscape variables was evaluated using factor analysis and linear models. Variables associated with landscape heterogeneity were the strongest drivers of bird strikes. Higher damaging strike rates occurred at airports with high landscape heterogeneity at all three spatial scales. Airports surrounded by large amounts of agricultural crops and fields reported more damaging birds strikes. Bird strikes are influenced by the landscape structure of the airport and surrounding areas. Since land beyond the airport property influences bird strikes, our results offer a guide that can be used when working with land owners. These results can be used in airport planning to build an environment that is less attract to wildlife.
1:30PM Efficacy of Translocation as a Management Tool for Urban Mule Deer in Utah
Channing R. Howard; David N. Koons; Mark K. Brunson; Mary L. Connor
Urbanized areas throughout much of western North America have been expanding and especially along the Wasatch Front in Utah. Wildlife managers have spent increasing amounts of time and money to respond to escalating complaints and conflicts with mule deer (Odocoileus hemionus) in these urbanized areas. This is driven in part by increasing expansion of communities into mule deer habitat as well as increasing use of these communities by mule deer, leading to situations of overabundant urban mule deer. Traditional urban deer management techniques, such as sharpshooting, can be viewed as dangerous or immoral, so wildlife managers are investigating translocation as an alternative non-lethal option. By determining if translocation of overabundant urban mule deer is a feasible method to reduce urban complaints and issues, we could expand the tools available to managers; however, little research exists on this topic. We examined the annual survival rate of 217 translocated urban mule deer and the associated costs of capture in Utah from 2014 to 2016. We additionally investigated whether the survival rate differed across sexes, release sites and years, and compared results to annual survival for wild deer from neighboring sites and wild mule deer translocation studies. Annual survival for urban translocated adult female mule deer was higher than that for adult males, but no differences were detected across release sites and years. Annual survival of urban translocated adult females was nevertheless lower than that for wild adult females in neighboring sites, but was similar to the annual survival of wild translocated mule deer. Capture costs were $242 to $282 per deer. Our results highlight the potential of translocation as a management tool to address overabundant urban mule deer when lethal control methods are deemed inappropriate.
1:50PM Spatio-Temporal Strategies for Efficient Planning of Feral Swine Population Management
Kim M. Pepin; Amy J. Davis; Kurt C. Vercauteren
Understanding effectiveness of control strategies of pest species is fundamental for planning efficient and cost-effective management programs. In addition to culling rates, there are many potential factors that determine effectiveness of different management strategies, including demographic processes such as immigration rates, birth dynamics and social structure. We developed a stochastic, data-based simulation model of feral swine population dynamics which accounted for social dynamics in space. We tested the impacts of different spatio-temporal management strategies (i.e., culling rates, timing of culling during the year, spatial pattern of culling and strength of a barrier to immigration) on population response and efficiency. We found that for the same level of moderate culling effort, prioritization of culling during the low-birthing period resulted in fewer pigs relative to prioritization during the high-birthing period, or spreading the work over a year period. An annual culling proportion of 40% drove the population to extinction within 5 years when there was no immigration and zonation or high-density targeting were used as spatial strategies, while a spatially random culling strategy under the same conditions didn’t. As culling intensity increased (70% of target population annually) and the target population reached low density (<5% of original density), the importance of spatial strategy became less pronounced relative to the importance of immigration barrier. Thus, knowledge of population status during management is important for updating management strategies in response to current population conditions. We also predicted that up to 54% more pigs would need to be removed using the random or high-density targeted strategy relative to a wave-like (zoning) spatial culling strategy, suggesting that the latter could be the most cost-effective. Our results highlight that consideration of birth and social dynamics in the development of management plans can decrease wasted effort on “mowing the grass” when the goal is maximum population control.
2:10PM Testing the Effectiveness of Bear-Resistant Containers to Reduce Black Bear-Human Conflicts
Heather Johnson; David Lewis; Stacy Lischka; Stewart Breck
Human-black bear conflicts within urban environments have been increasing throughout North America with the main driver being the availability of garbage. It is believed that significantly “bear-proofing” a town should dramatically reduce conflict but little research has been conducted to test this idea. We conducted an experiment in Durango, CO from 2011-2016 in which we distributed 1,110 bear proof trash containers, enhanced education, and increased enforcement to all residents in two treatment areas and compared results to equivalent control areas. We assessed a variety of ecological and social outcomes including 1) were human-bear conflicts reduced in treatment areas; 2) what rate of compliance (i.e., properly locking bear-resistant containers) was needed to reduce rates of conflict; 3) whether the effectiveness of bear-proofing increased over the course of the study; and 4) if the distribution of bear-resistant containers changed public attitudes about residents satisfaction with management, support for bear-proofing ordinances, or perception about future risk of conflicts. To assess these questions, we collected data on the frequency and locations of trash-related conflicts; whether individual residences were in compliance (i.e., containers were properly locked); and surveyed residents multiple times throughout the study. We found that in treatment areas, trash conflicts were significantly lower; compliance increased by 40%; both compliance and conflict decreased over time; and favorable perceptions about management significantly increased. Furthermore, we found the relationship between conflict and compliance was non-linear indicating that compliance over 60% was an important milestone for reducing conflict, matching published theoretical predictions. We will provide details of the experiment and the results and discuss implications of our study for managing human black-bear conflicts in developed environments.
2:30PM Changes in Feral Swine Activity in Response to Prescribed Burning in the William B. Bankhead National Forest, Alabama.
Patience Knight; Helen Czech; William Stone
Feral swine (Sus scrofa) are a serious threat to biodiversity, especially in forested and riparian habitats. Since other invasive species tend to thrive after habitat disturbance and many other wildlife species abound in areas that are periodically burned, it is important to understand how prescribed burning affects feral swine habitat use. This study investigated whether or not feral swine activity changed after prescribed burning in upland, terrestrial sites and riparian sites in the William B. Bankhead National Forest. Using line transect sampling, 8 prescribed-burn and 9 control terrestrial, upland forest stands were surveyed for pig sign before and after prescribed burning. Riparian transects within low-order streams, were similarly surveyed for pig sign in or near those same prescribed-burn and control stands. Although control forest stands had overall higher pig activity than burn stands, the most recent prescribed burn did not alter pig sign activity in those stands. Feral swine activity did not change over time in burn terrestrial and riparian sites. Four habitat variables, understory density, understory species richness, percentage bare ground, and percentage course woody debris cover, were significantly and positively correlated with pig activity. It is important to note that these variables are also significantly affected by prescribed burning. Therefore, the overall fire history and ecology may cause the differences in feral swine activity, rather than the individual prescribed burning applications. This research may be a useful starting point for making better informed decisions in feral swine management.
2:50PM Refreshment Break
3:20PM Predicting Spatial Factors Associated with Livestock Depredations by the Mexican Wolf in Arizona and New Mexico
Reza Goljani A; Jennifer K. Frey; David L. Bergman; Stewart W. Breck; James W. Cain; John Oakleaf; Julia B. Smith
The Mexican wolf (Canis lupus baileyi) is an endangered species that is being reintroduced into its native range in the American Southwest. However, predation on livestock is one of the primary concerns in the Mexican wolf recovery program and most tools to reduce depredation are reactive or relatively ineffective. As an alternative, spatial risk models predict the spatial distribution of depredation locations, thereby providing opportunity for managers to avoid conflicts before they occur. The objective of this study was to develop a spatial risk model of cattle depredation by Mexican wolf in Arizona and New Mexico. We used a presence-only maximum entropy modeling approach (Maxent) to develop the risk model based on 120 confirmed depredation incidents occurring on public lands. Predictor variables included eight landscape and human features, in addition to models we developed for potential livestock and natural prey abundances. The model for potential livestock abundance was created using linear regression of Animal Unit Month data from public grazing allotments, while models for abundance of natural prey (elk, mule deer, white-tailed deer) were created using Maxent. Results indicated an increase the risk of livestock depredation by Mexican wolves in areas with high elk abundance, more diverse canopy cover, more grasslands, more herb canopy cover type, lower slope, and located further from developed areas. The risk map revealed areas with relatively high or low potential for cattle depredations that can inform future expansion of Mexican wolf distribution. In addition, knowledge about locations of predicted hotspots for cattle depredation may allow use of proactive measures to prevent depredation such as prioritizing use of active non-lethal methods, especially during periods when cattle are particularly vulnerable to depredation. Further, minimizing the attractiveness of depredation hotspots to elk, and thus wolves (e.g., such as by increasing human activity), could decrease the incidence of depredations.
3:40PM A Call for Science-Based Metrics of Success for Management Actions Addressing Invasive Species
Amy Davis; Bruce Leland; Michael Bodenchuk; Kurt VerCauteren; Kim Pepin
Management of invasive species is a constant challenge. Resources are typically limited and management planning is not always science-based, but more driven by logistical constraints. Invasive wild pigs (Sus scrofa) cause significant damage to agriculture, infrastructure, and native ecosystems, and spread disease to livestock and wildlife. Wild pigs have expanded their range across the United States in the last three decades; consequently, the management efforts have ramped up accordingly. Methods for measuring management effects are often based on body counts which may not be indicative of impacts on population viability or damage reduction. We examined this idea in a pilot study where repeated removal events by aerial gunning were conducted on three consecutive days in three study sites. Using a removal estimator, we calculated the proportion of the population removed by flight. We found that capture rates vary by habitat (0.08 per hour in open habitats and 0.04 in shrubby habitats) and personnel (one pilot/gunner team was considerably better). We used simple population modeling to show the time it would take for a population to recover if only one, two, or three flights were conducted. We used three possible damage/density curves to show what impact on damage reduction would result from different management efforts. Our results show that populations are typically reduced by ~35% for the first flight, ~62% for the second flight, and ~79% for the third flight. For a typical damage curve this would represent a reduction in damage of 2%, 19%, and 60% respectively. Therefore, we argue that it is critical to measure the impact to the population and not just the raw number removed to demonstrate a scientifically sound management practice. Further, we emphasize the importance of monitoring damage impacts to provide clear justification for, and optimization of, resources to this critical type of management work.
4:00PM The Causal Impact of Changes in Deer Abundance on Deer-Vehicle Collisions
Jennifer L. Raynor
The purpose of this study is to estimate the causal impact of changes in deer abundance on roadway collisions and associated economic losses. An estimated one million deer-vehicle collisions (DVCs) occur every year in the United States, causing 20,000 injuries, 200 fatalities, and $9 billion in economic losses—DVC losses are more than half of annual spending on big game hunting equipment and trips. Consistent with intuition, previous research shows that DVCs are correlated with deer abundance, but the causal effect is unknown. I use ordinary least squares and instrumental variables regressions to identify the effects of white-tailed deer (Odocoileus virginianus) abundance, traffic volume, land cover, and weather on the frequency of county-level DVCs in Wisconsin between 1998 and 2013. All regressions include county and year effects that capture time-invariant county characteristics (e.g., road density and speed limits) and state-wide trends (e.g., gasoline prices and highway spending). A 1 percent increase in deer abundance leads to a 0.3 percent increase in DVCs. The effects of traffic volume and precipitation are of a similar magnitude, with elasticities of 0.4 and 0.2, respectively. Forest decreases and farmland increases DVCs. All estimated coefficients have the expected sign and are robust across specifications. One additional deer above the county-year average causes $30 in economic losses from DVCs each year. Equivalently, 301 additional deer cause one more DVC per year. Achieving the latest posthunt population goal would reduce DVC losses by an estimated $23.3 million annually, roughly equal to hunting license and permit revenues. Looking forward, DVCs losses likely will continue to grow. Cultural shifts in driving or development patterns are unlikely, so only deer abundance provides a plausible leverage point for broad-scale DVC mitigation. The results suggest that a relatively small decrease in deer abundance yields an economically significant reduction in DVCs.
4:20PM Anthraquinone as a Vole Repellent: Not Just for the Birds?
Roger A. Baldwin; Ryan Meinerz; Gary W. Witmer; Scott J. Werner
California voles (Microtus californicus) cause extensive damage to tree crops through girdling of young stems. Recent laboratory trials have indicated substantial repellency (up to 84%) of anthraquinone (a post-ingestive repellent) to voles on treated grain. Therefore, we established a study to test the efficacy of anthraquinone applications to tree stems to reduce girdling damage from voles. We also assessed the impact of vegetation around the base of trees to determine the impact of cover on girdling activity. In Fresno County, CA, during summer 2016, we established twenty 3.35 m × 2.44 m bins (hereafter mesocosms) where we evenly spaced eight 1-yr old clementine orange trees. Cover crops including various grasses and forbs were planted on randomly selected halves of each mesocosm. All trees were treated with anthraquinone in half of the mesocosms; trees were left untreated in the remaining half. We captured 40 voles and released 2 individuals into each mesocosm and tracked girdling damage once weekly for 5 weeks. We observed a significant reduction in girdling damage on anthraquinone-treated trees (control: = 1.79 cm2; anthraquinone: = 0.10 cm2; P < 0.001). The lack of vegetation around the base of trees further reduced girdling damage for anthraquinone-treated trees (vegetation: = 0.18 cm2; no vegetation: = 0.00 cm2; P = 0.002), although vegetation did not significantly impact damage for untreated trees (vegetation: = 2.11 cm2; no vegetation: = 1.47 cm2; P = 0.213). Damage in control mesocosms generally increased over the 5-wk period (Friedman’s test, χ2 = 13.5, P = 0.009), but we did not see such an increase in damage in anthraquinone-treated mesocosms (Friedman’s test, χ2 = 4.4, P = 0.350) indicating that anthraquinone maintained its repellency during the duration of this study. Anthraquinone was an effective repellent for voles; removal of vegetative cover around trees further increased efficacy.
4:40PM Hunter and Nonhunter Disease Risk Perceptions over Time
Craig A. Miller; Jerry J. Vaske
We compared 2004 and 2012 Illinois deer hunters’ and nonhunters’ disease risk sensitivity relative to six diseases and illnesses (West Nile Virus, Lyme disease, Chronic Wasting Disease (CWD), Bovine Spongiform Encephalopathy (BSE), E. coli¸ Salmonella). We hypothesized: (a) perceptions of specific risks would decline over time (i.e., 2004 vs. 2012) and (b) structure of general risk perceptions (i.e., no, slight, moderate risk) would remain the same over the two time periods. Data were obtained from two mailed surveys 1) Illinois resident hunters and Nonhunters conducted in 2004 (n = 1,879, response rate = 65%), 2) Illinois hunters conducted in 2012 (n = 5,569, response rate = 53%). Each of the dependent variables (e.g., West Nile, CWD, BSE) was coded on a 4-point scale (1 = no risk, 2 = slight risk, 3 = moderate risk, 4 = high risk). Hypothesis 1 tested year (2004 vs. 2012) as the independent variable. Hypothesis 2 used K-means cluster analysis that grouped respondents into “No Risk,” “Slight Risk,” or “Moderate Risk” clusters. Results supported the first hypothesis; all six t-tests were statistically significant (p < 0.002). Means for the specific beliefs were in the predicted direction for four of the variables. Two exceptions were becoming ill from Salmonella and E. coli. K-means cluster analysis supported hypothesis two. Findings suggested the three cluster solution provided best fit to the data. Distribution of respondents in each of the three clusters remained relatively stable over the two time periods (i.e., 19% vs. 22% for No Risk, 54% vs. 55% for Slight Risk, 27% vs. 23% for Moderate Risk; 2004 vs. 2012 respectively). Distributions differed statistically (χ2 = 12.50, p = 0.002), but the difference was attributable to the large sample size, given less than minimal effect size (Cramer’s V = 0.042). Implications of these findings are discussed.


Contributed Paper
Location: Albuquerque Convention Center Date: September 25, 2017 Time: 1:10 pm - 5:00 pm