Robust mapping of human–wildlife conflict: controlling for livestock distribution in carnivore depredation models

Identifying the Risk Regions of Wild Boar (Sus scrofa) Incidents in China

The objectives of this study were to identify the risk regions of wild boar incidents in China and to draw a risk map. Risk maps can be used to plan the prioritization of preventive measures, increasing management effectiveness from both a short- and a long-term perspective. We used a web crawler (web information access technology) to obtain reports of wild boar incidents from China's largest search engine (Baidu) and obtained 196 valid geographic locations of wild boar incidents from the reports. Subsequently, a system of environmental variables-with climate, topography, landscape, and human disturbance as the main variable types-was constructed, based on human-land-system thinking. Finally, the Maxent model was applied to predict the risk space of wild boar incidents in China by integrating the geographic location information for wild boar incidents with the environmental variables. We observed that the types of environmental variables that contributed to wild boar incidents were in the descending order of climate (40.5%) > human disturbance (25.2%) > landscape (24.4%) > topography (9.8%). Among the 14 environmental variables, annual precipitation, the GDP index, and the mean annual temperature were the main environmental variables. The distance from woodland, distance from cultivated land, and elevation were the secondary environmental variables. The response curves of the environmental variables demonstrated that the highest probability of wild boar incidents occurred when the annual average temperature was 16 °C, the annual precipitation was 800 mm, and the altitudes were 150 m and 1800 m. The probability of wild boar incidents decreased with an increase in the distance from cultivated and forested land, and increased sharply and then levelled off with an increase in the GDP index. Approximately 12.18% of China was identified as being at a high risk of wild boar incidents, mainly on the eastern side of the Huhuanyong Line.

Planning for wolf-livestock coexistence: landscape context predicts livestock depredation risk in agricultural landscapes

Extensive pastoral livestock systems in Central Europe provide multiple ecosystem services and support biodiversity in agricultural landscapes but their viability is challenged by livestock depredation (LD) associated with the recovery of wolf populations. Variation in the spatial distribution of LD depends on a suite of factors, most of which are unavailable at the appropriate scales. To assess if LD patterns can be predicted sufficiently with land use data alone at the scale of one federal state in Germany, we employed a machine-learning-supported resource selection approach. The model used LD monitoring data, and publicly available land use data to describe the landscape configuration at LD and control sites (resolution 4 km * 4 km). We used SHapley Additive exPlanations to assess the importance and effects of landscape configuration and cross-validation to evaluate the model performance. Our model predicted the spatial distribution of LD events with a mean accuracy of 74%. The most influential land use features included grassland, farmland and forest. The risk of livestock depredation was high if these three landscape features co-occurred with a specific proportion. A high share of grassland, combined with a moderate proportion of forest and farmland, increased LD risk. We then used the model to predict the LD risk in five regions; the resulting risk maps showed high congruence with observed LD events. While of correlative nature and lacking specific information on wolf and livestock distribution and husbandry practices, our pragmatic modelling approach can guide spatial prioritisation of damage prevention or mitigation practices to improve livestock-wolf coexistence in agricultural landscapes.

Patterns of livestock loss associated with a recolonizing wolf population in Germany

Predation on livestock presents a daunting challenge for human–carnivore coexistence in agricultural landscapes. In Germany, the recolonization of wolves is ongoing and its consequences are insufficiently understood. Knowledge about which livestock species are susceptible to wolf predation, which farm types are predisposed to attacks by wolves, and when predation on livestock occurs is valuable for mitigating stakeholder conflicts. To this end, we analyzed 14 years of monitoring data and assessed the livestock prey spectrum, identified correlates between predation on livestock, farm type and livestock category, and described temporal patterns of livestock loss caused by a recolonizing wolf population in the state of Brandenburg (Germany). Among a total of 1387 recorded cases, 42% were unequivocally attributed to wolves (SCALP criteria C1 and C2) and 12% of cases were not caused by wolves. The number of head of livestock killed during a single wolf attack was mediated by farm type and livestock species; losses per event were greater in full-time farms vs. other farm types and greater in sheep, farmed deer and other livestock species, compared to cattle. While sheep were the most commonly killed livestock species, the increase in wolf territories over the investigation period was associated with a widening of the domestic prey species spectrum. Count regression models provided evidence for the increasing frequency of predation events over the 14-year period, along with an exponential increase in wolf territories. Predation on livestock occurred throughout the year, yet seasonality of events was evident and differed across livestock categories. Predation on sheep peaked in the fall, coinciding with the post-weaning period of wolf offspring. Predation on cattle peaked in the spring, coinciding with the cattle calving period. These results call for renewed investment in the implementation of prevention methods for all susceptible domestic species, particularly during times of elevated predation risk.

Integrated species distribution models reveal spatiotemporal patterns of human-wildlife conflict

To mitigate human-wildlife conflict it is imperative to know where and when conflict occurs. However, standard methods used to predict the occurrence of human-wildlife conflict often fail to recognize how a species distribution likely limits where and when conflict may happen. As such, methods that predict human-wildlife conflict could be improved if they could identify where conflict will occur relative to species' underlying distribution. To this end, we used an integrated species distribution model that combined presence-only wildlife complaints with data from a systematic camera trapping survey throughout Chicago, Illinois. This model draws upon both data sources to estimate a latent distribution of species; in addition, the model can estimate where conflict is most likely to occur within that distribution. We modeled the occupancy and conflict potential of coyote (Canis latrans), Virginia opossum (Didelphis virginiana), and raccoon (Procyon lotor) as a function of urban intensity, per capita income, and home vacancy rates throughout Chicago. Overall, the distribution of each species constrained the spatiotemporal patterns of conflict throughout the city of Chicago. Within each species distribution, we found that human-wildlife conflict was most likely to occur where humans and wildlife habitat overlap (e.g., featuring higher-than-average canopy cover and housing density). Furthermore, human-wildlife conflict was most likely to occur in high-income neighborhoods for Virginia opossum and raccoon, despite the fact that those two species have higher occupancy in low-income neighborhoods. As such, knowing where species are distributed can inform guidelines on where wildlife management should be focused, especially if it overlaps with human habitats. Finally, because this integrated model can incorporate data that have already been collected by wildlife managers or city officials, this approach could be used to develop stronger collaborations with wildlife management agencies and conduct applied research that will inform landscape-scale wildlife management.