Modeling white spruce height growth of three ecosite phases in the Alberta Boreal Mixedwood ecological area

摘要

White spruce height growth data from three ecosite phases in the Alberta Boreal Mixedwood ecological area were fitted by logistic and Chapman-Richards models using the generalized nonlinear least squares algorithm. Both models characterized well whitespruce height growth pattern in three ecosite phases; likelihood ratio tests rejected the hypothesis that a combined model adequately representing white spruce height growth in all three ecosite phases. Initially white spruce grows faster in the aspen dominating BMd1 ecosite phase than in the white spruce predominant BMd3 and the wet and nutrient rich BMf3 ecosite phases. The asymptotic height is shorter in the BMf1 than in the BMd3 and the BMf3 due possibly to a suppression effect of the overtopping aspen in the BMd1 ecosite phase. Incorporating a power variance function of height stabilized heterogeneity in residuals errosof the height model; nevertheless, residuals from weighted models remained large. An alternative mixed-effects modeling paradigmis suggested to reduce variability in white spruce height growth within ecosite phases and increase the prediction precision.