Assessing landscape scale risk of bark beetle infestation: methods and experience with Mountain Pine Beetle

T.L. Shore, A. Fall, W.G. Riel, J. Hughes, and M. Eng

Canadian Forest Service, Gowlland Technologies Ltd., Consultant, and British Columbia Ministry of Forests and Range

Several bark beetle species, mostly in the family Scolytidae, have the potential for dramatic population increases under favorable forest and climate conditions which can result in landscape scale mortality to the host tree species. For example, the mountain pine beetle (Dendroctonus ponderosae Hopk.) has killed between 20% and 30% of mature lodgepole pine over 10 million ha in British Columbia in recent years. This level of mortality has widespread implications for current and future forest management.

Landscape-scale risk assessment of bark beetle infestation aims to quantify the spatial and temporal likelihood of attack extent and severity. We have developed and applied a range of methods from structural risk (i.e. strictly assessing patterns) to functional risk (i.e. assessing interactions and feedbacks between pattern and process).

Susceptibility and risk rating systems classify each stand or grid cell of a landscape according to local characteristics (e.g. stand age, distance to nearest attack).  As such, these approaches are temporally static with limited spatial accounting, but have the benefit of limited data requirements and ease of application. Although likely pathways and interactions with management cannot be identified, the mountain pine beetle (MPB) susceptibility and risk rating system remains one of the most widely used tools.

Spatial connectivity assessment increases the spatial dimension from rating systems. Our approach to connectivity assessment uses spatial graphs to analyze scales at which patches of susceptible hosts are well-connected, in particular with existing attack. These methods are relatively easy to apply and data requirements are fairly modest (although information on movement cost/impedance is required). Although still static in nature, likely pathways can be identified, and large areas can still be processed efficiently.

Empirical information on infestation progression can be used to increase the temporal dimension with semi-Markovian projection models. These methods require moderate effort to develop and data requirements, requiring historical time series information. A key assumption is that future outbreak dynamics will mimic past trends. However, they allow trends to be identified and interaction with management scenarios to be explored.

Dynamic population models make a shift to a more process-oriented approach to modeling outbreak dynamics by explicitly capturing demographic changes in space and time with processes of mortality, birth, dispersal, etc. Such approaches require substantial effort to develop and have fairly high data requirements, in particular the need for a reasonable understanding of beetle biology and interactions with hosts at relatively fine scales. The advantage is a closer match with the ecological process, and greater ability to assess interactions with management.

Individual-based dispersal models allow more detailed exploration of how beetles may interact functionally with a landscape. However, such approaches are generally prohibitive at the landscape scale due to lack of sufficient data (both for details of landscape pattern and beetles) as well as computing power (especially for a large outbreak).

We have developed and applied the above methods, with the exception of the latter, for assessing MPB risk at landscape scales in pine forests of British Columbia, Canada. Susceptibility/risk rating has proven useful for a quick overview of landscape state and general patterns. Connectivity assessments have been useful in areas with limited attack, and to provide a more comprehensive assessment of the spatial pattern of hosts and likely pathways of attack. Empirical outbreak projection methods have been useful to assess very broad scale dynamics (e.g. at provincial scales) and potential interactions with management. We have used a population modeling approach for more process-based assessments of outbreak development. Together, these methods form a suite of tools useful to assess risk of bark beetle attacks at broad spatial scales.

Thursday Morning Plenary

corresponding author:

Dr. Terry L. Shore
Natural Resource Canada
Canadian Forest Service
506 W. Burnside Rd.
Victoria BC Canada V8Z 1M5
250 363-0666
tshore@pfc.forestry.ca