| Abstract | When the mountain pine beetle populations (Dendroctonus ponderosae) reach epidemic levels, large areas of old growth forests are destroyed. In British Columbia alone, the mountain pine beetles have affected more than 10 million ha of land. Recently the pine beetle infestation successfully crossed the Rocky mountain range and is now threatening to spread across the Prairie provinces. As a consequence there is an urgent need to predict potential outbreaks. Here we use a Markov process logistic regression model that integrates ecological, climatic, and topographic information to determine the probability of a mountain pine beetle outbreak one year ahead. The resulting model maximizes predictive ability by accurately identifying high-risk regions, is simple to implement, uses information that is readily available, and can easily be validated. We develop recommendations for how this model can support strategic planning by identifying high-risk areas for control and management. |
|---|
