Populations of mountain pine beetle (Dendroctonus ponderosae) frequently persist in low-density endemic states in which they attack low-value compromised trees. Under favorable conditions beetles may be able to overcome the defenses of vigorous, resource-rich trees triggering a local population increase that can develop into a regional-scale epidemic. The spatiotemporal dynamics of epidemics are influenced by climatic factors such as overwinter temperatures and drought, as well as human activities such as thinning and fire suppression and in recent years, the spatial extent of epidemics has increased across western North America. Many climatic and anthropogenic factors affect the transitions between endemic and epidemic states, and land managers often aim to reduce the likelihood and/or effects of an epidemic outbreak within individual forest stands.
We created an agent-based model of beetle dispersal and reproduction to contrast the conditions that lead to stable endemic populations, stable endemic/epidemic cycles, and population crashes in individual forest stands. In a series of simulations, we varied three aspects of the forest environment: beetle fertility, tree density, and the proportion of compromised trees. These three model components represent attributes of a forest stand influenced by numerous environmental factors, both climatic and anthropogenic.
We found boundaries in parameter space, at which the long-term behavior of the system transitioned between stable endemic populations, endemic/epidemic cycles, or population crashes. Breakpoints between different behaviors from small changes in model parameters highlight the fact that we cannot understand long-term patterns in real beetle population behavior within forest stands from local conditions alone. Within a forest stand, year-to-year variation in climatic factors may rapidly shift beetle populations among basins of attraction to different behaviors, while longer-term regional climate patterns could provide a stabilizing influence at larger spatial scales.
Recurring epidemics are common in forests, but the majority of our simulations of isolated forest stands resulted in population crashes. The population crashes and breakpoints that we found suggest that spatiotemporal heterogeneities such as asynchronous endemic/epidemic cycles in neighboring stands are crucial for understanding beetle populations. Of particular interest is how management decisions at multiple spatial scales might influence the likelihood of epidemics in individual, high-value forest stands within regional-level outbreaks. Our model provides both a framework in which to explore the consequences of different assumptions about beetle populations’ responses to environmental conditions and a tool to refine hypotheses for future studies of real forests.