Leaf area prediction models for Tsuga canadensis in Maine

Abstract

Tsuga canadensis (L.) Carr. (eastern hemlock) is a common species throughout the Acadian forest. Studies of leaf area and growth efficiency in this forest type have been limited by the lack of equations to predict leaf area of this species. We found that sapwood area was an effective leaf area surrogate in T. canadensis, though adding crown length to the sapwood equations improved model performance. Prediction bias was observed at the upper end of our data for the best sapwood equation. Sapwood area at crown base did not predict leaf area as well as sapwood area at breast height. Equations using crown length or crown volume alone were the least effective of all models tested. Models using stem cross-sectional area inside the bark or tree basal area with a modified live crown ratio produced results comparable with those of the best sapwood-based model and were unbiased across the range of our data. There findings verify the value of nonsapwood-based approaches to T. canadensis leaf area prediction.