Abstract Context Exclusion fences are increasingly used to prevent interactions between predators (introduced and native) and assets such as endangered species or livestock. However, challenges remain in identifying when exclusion fences are an optimal investment and the intended outcome is likely to be achieved. The level of association with complementary methods of control that is needed is also unclear. Aims We aimed to quantify the interactions among factors that affect fencing efficiency, including the size of the fenced area, the fence permeability, the initial density of the predator population, and its survival of complementary control methods. Methods Using a spatially explicit, individual-based model, we simulated wild dog (dingo) populations as a proxy for describing predator dynamics inside a fenced area under different management practices and fence designs. We then fit a generalised linear model to the model outcomes to assess the effects of the four factors mentioned above. Key results Lethal control had a strong effect on wild dog density when the survival of control was lower than 0.5. Fences generally had an effect on wild dog density only when their permeability was lower than ~1% and their effect was most noticeable when the initial density was very low (<2 dogs per 100 km2), or when survival of control was very low (<0.5). Conversely, when the initial density was very high (~12 dogs per 100 km2), a fence with a low permeability (<1.5%) caused the paradoxical effect that wild dog density could be higher than that obtained with a more permeable fence. Wild dog eradication was possible only when survival of control was 0.25 or lower, except when either initial density or fence permeability were extremely low (<2 dogs per 100 km2 and <0.1% respectively). Conclusions Our results demonstrated that large exclusion fences can be an effective aid in managing predator populations. We recommend that this tool should be used as a preventive measure before predators establish a population inside the area targeted for exclusion, in tandem with lethal control, or when an initial marked reduction of predator density can be achieved. We also demonstrated that eradication can be achieved only when a narrow combination of parameters is met. Implications Land managers should carefully evaluate when and at what scale control tools should be deployed to control wild dog populations. Landscape application of exclusion fences faces the challenge of high maintenance to ensure low permeability, coupled with very high sustained suppression of wild dog density, which are unlikely to be feasible options in the long term. Conversely, the same control techniques could provide efficient asset protection at a smaller scale where fence maintenance and sufficient control effort can be sustained.

中文翻译：

---

捕食者排除栅栏什么时候效果最好？空间显式建模方法

摘要 上下文排除栅栏越来越多地用于防止捕食者（引进的和本地的）与濒危物种或牲畜等资产之间的相互作用。然而，在确定排除栅栏何时是最佳投资以及可能实现预期结果方面仍然存在挑战。与所需的补充控制方法的关联程度也不清楚。目的我们旨在量化影响围栏效率的因素之间的相互作用，包括围栏区域的大小、围栏渗透性、捕食者种群的初始密度以及补充控制方法的存活率。方法 使用空间明确的、基于个体的模型，我们模拟了野狗 (dingo) 种群，作为描述不同管理实践和围栏设计下围栏区域内捕食者动态的代理。然后，我们将广义线性模型拟合到模型结果中，以评估上述四个因素的影响。主要结果 当控制的存活率低于 0.5 时，致死控制对野狗密度有很强的影响。围栏通常只有在其渗透性低于~1% 时才会对野狗密度产生影响，并且当初始密度非常低（每 100 平方公里 < 2 只狗）或对照的存活率非常低时，它们的影响最为明显（ <0.5）。相反，当初始密度非常高（每 100 平方公里约 12 只狗）时，围栏的渗透性较低（<1. 5%) 导致野狗密度可能高于使用更具渗透性的围栏获得的密度的矛盾效应。只有当控制的存活率为 0.25 或更低时才有可能根除野狗，除非初始密度或围栏渗透率极低（分别为 <2 只狗/100 平方公里和 <0.1%）。结论 我们的结果表明，大型隔离栅栏可以有效地帮助管理捕食者种群。我们建议，在捕食者在目标区域内建立种群之前，与致命控制相结合，或当捕食者密度可以实现初始显着降低时，应将此工具用作预防措施。我们还证明，只有在满足参数的狭窄组合时才能实现根除。影响 土地管理者应仔细评估何时以及以何种规模部署控制工具来控制野狗种群。隔离围栏的景观应用面临着高维护以确保低渗透性的挑战，再加上对野狗密度的持续抑制非常高，从长远来看这不太可能是可行的选择。相反，相同的控制技术可以在可以维持围栏维护和足够控制工作的较小规模下提供有效的资产保护。从长远来看，这不太可能是可行的选择。相反，相同的控制技术可以在可以维持围栏维护和足够控制工作的较小规模下提供有效的资产保护。从长远来看，这不太可能是可行的选择。相反，相同的控制技术可以在可以维持围栏维护和足够控制工作的较小规模下提供有效的资产保护。