A Basal Area Increment Model for Individual Conifers in the Northern Rocky Mountains

A basal area increment model is developed and evaluated for individual trees of 11 conifer species. This distance-independent model is applicable over much of the northern Rockies, an extensive region where stands have considerable variation in species composition, age structure, and past management. The model represents the diverse ecological requirements of the various species and responds appropriately across the observed range of predictor variables. Furthermore, it is well behaved even when used for conditions found in the region but not represented in the large calibration data set (44,086 trees). In addition, it possesses desirable statistical properties such as homogeneous residual variance, minimal multicollinearity, and linear parameters. The model is validated against independent data from two sources: a research study of young managed stands (2,881 trees) and a collection of long-term permanent research plots (6,594 trees). Predictions for small suppressed trees from the permanent plots are generally too large, but there are no significant relationships between residuals and other predictor variables in either data set. An unusual feature of the model is the intentional omission of site index and age, which is motivated by the large number of irregular stands in the region. Validation tests using the permanent research plots indicate, however, that residuals are not correlated with site index and age. Furthermore, long-term projections of stand attributes are unbiased. For. Sci. 36(4):1077-1104.