Abstracted models for assessing reforestation stocking in spruce-aspen mixedwoods

作者：Farnden Craig^*

来源：Forestry, 2012, 85(5): 621-628.

DOI：10.1093/forestry/cps073

Individual tree growth models are often constructed with much more complexity than is required for many of the tasks to which they might be applied. In many cases, they also have detailed requirements for describing initial conditions, which may necessitate costly data collection. In cases where a large number of model predictions are required for which there is a mismatch between available data and model requirements, an abstracted low-dimension predictive model (LDPM) that uses alternate input variables and that accurately mimics the outcomes of the more complex model for the specified problem may be an attractive option. In this simulation study, several sets of LDPMs are developed as possible replacements for individual tree models in assessing regeneration stocking implications for mixtures of white spruce (Picea glauca) and trembling aspen (Populus tremuloides). These models can accurately mimic a selected set of outcomes of the parent model, but add a small degree of uncertainty in the process. Limitations introduced by the abstraction process include an overall loss of information and a large reduction in the variability of conditions to which the model can be applied. Benefits introduced by the process include a much better efficiency of data collection and model application efforts towards a specific management problem.

• 出版日期2012-12