• Mayer, Spencer R.
  University of Maine School of Forest Resources
  

Leaf area prediction models are derived for red spruce and balsam fir growing in mostly even-aged stands in Maine. The branch summation method was used to calculate total tree leaf area for 43 balsam fir and 46 red spruce trees. A comprehensive suite of model forms using both sapwood and allometric variables was screened to evaluate their predictive power and reliability across species, crown class, and geographic region of Maine. Best-fit models were chosen based on a modified likelihood criterion, residual analyses, and corrected R2. For both balsam fir and red spruce, models based on area inside the bark and basal area (fir and and spruce, respectively) performed best; however, the predictive power varied according to the secondary predictors if crown length and modified live crown ratio. Testing several forms of crown length and modified live crown ratio. Testing several forms of crown length and crown ratio as secondary predictors revealed little difference between definitions and we conclude that in conjunction with basal area, the best secondary predictor is modified live crown ratio. Best-fit models accounted for 95 and 89% of the variation for fir and spruce, respectively. Sapwood models performed nearly as well as the best-fit models and accounted for some variation in age structure apparent within some stands. Overall, leaf area predictions showed less variation  for balsam fir than for red spruce. Neither geographic location nor site index was a significant factor of model performance for either basal area or sapwood area models.