A hybrid modelling approach to estimating establishment and growth of advanced regeneration following mountain pine beetle attack

A hybrid model that links modules of sortie-nd and Prognosisbc is described. The impetus to develop this model has come in response to the need for a dynamic management tool that can provide reasonably accurate estimates of growth for stands that have been attacked by mountain pine beetle (mpb), but will go unsalvaged. The rapid change in stand structure following disturbance events such as mpb infestation challenges our ability to use traditional growth and yield models. Generally, these models rely heavily on historical field measurements to establish empirically based relationships describing diameter distributions and tree growth. Models of this type run into problems when a limited amount of historical data is available, and rapidly changing and variable conditions within the stand make it difficult to establish empirical relationships that provide acceptable levels of prediction accuracy. Linking modules of sortie-nd with Prognosisbc brings together empirical and process-based methods. The hybrid (linked) model operates on two main assumptions: (1) the relationships related to adult tree growth described using empirical means remain relatively intact following disturbance; and (2) seedling establishment and growth are best described through a process-based approach. The model is evaluated by comparing projections of stand density, diameter distribution, and species composition against actual measurements from stands that were attacked by mpb roughly 25 years ago. Overall, prediction accuracy for the hybrid model was better than simulations run using sortie-nd and Prognosisbc independently. Improved model accuracy is discussed with specific reference to changes in the crown architecture in sortie-nd.