Mountain Pine Beetle Outbreak Dynamics

    • Case study of Cypress Hills outbreaks
      Mountain pine beetle larvae

      Mountain pine beetle larvae
      Photo by Mélodie Kunegel-Lion

      • In this project, we used the highly-detailed mountain pine beetle data collected in the Saskatchewan portion of the Cypress Hills interprovincial park to examine the impact of ecological and environmental factors and management strategies on mountain pine beetle population change during the course of an outbreak. Several modelling methods are applied to explore this question under different perspectives: individual-based and mechanistic models, logistic regressions, and boosted classification and regression trees. Related Publications
    • Large-scale dynamical models with spatial autocorrelation
      • Investigations into population dynamics over large, densely sampled, geographical areas are often complicated by the presence of autocorrelation among sample points. We developed mathematical shortcuts for simplifying the intensive computations involved in likelihood-based approaches to such models. Related Publications
    • Machine learning methods for predicting insect outbreaks
      Adult mountain pine beetle encased in pine tree resin

      Adult mountain pine beetle encased in pine tree resin
      Photo by Mélodie Kunegel-Lion

      • Planning forest management relies on predicting insect outbreaks such as mountain pine beetle, especially in the intermediate-term future, e.g., five years. Machine-learning algorithms are potential solutions due to their record of successes across a variety of prediction tasks. There are, however, many challenges in applying them: identification of the best learning models and the best subset of available features (including time lags) and how to properly evaluate the models to avoid misleading performance-measures. We aim at systematically addressing these questions in the context of predicting the chance of a mountain pine beetle outbreak in the Cypress Hills area and seek models with the best performance at predicting infestations many years in the future. Related Publications