To examine the relationships between trembling aspen (Populus tremuloides Michx.) productivity, environmental attributes, and foliar nutrients and to make accurate predictions of trembling aspen productivity, we sampled 60 naturally established, fire-originated, and even-aged trembling aspen stands in northern British Columbia. Trembling aspen site index significantly varied with latitude, elevation, aspect, slope position, edatopes, some forest floor and mineral soil physical and chemical properties, and concentrations of some foliar nutrients. To predict site index, we developed multiple linear regression models using climatic variables, topographic properties, edatopes, soil physical and chemical properties, or foliar nutrients as predictors. Model accountability for variation of site index differed in decreasing order from soil model, climatic model, forest floor model, foliar nutrient model, edatope model, topographic model, to mineral soil model. Examined by the test data set, all models were unbiased, but they had different levels of precision in prediction in decreasing order from edatope model, soil model, forest floor model, mineral soil model, foliar nutrient model, climatic model, to topographic model. The soil and foliar nutrients models may provide insight into ecosystem processes, but the models using climatic variables and topographic properties or edatopes as predictors are recommended for predicting trembling aspen site index.
Fine root dynamics of trembling aspen in boreal forest and aspen parkland in central Canada
