Leaf Area, Stemwood Volume Growth, and Stand Structure in a Mixed-Species, Multi-Aged Northern Conifer Forest

Tree- and stand-level leaf area (LA) - stemwood volume growth relationships were explored in mixed-species, multi-aged northern conifer stands dominated by Tsuga canadensis (L.) Carr. (eastern hemlock), Abies balsameu (L.) Mill. (balsam fir), and Piceu rubens Sarg. (red spruce). A T. cunudensis LA model was developed to supplement published equations for A. bulsumeu and P. rubens. Sapwood area was an effective T. cunudensis LA surrogate, though nonsapwood-based models using stem cross-sectional area and modified live crown ratio produced results comparable to sapwood-based models. Equations for predicting tree-level stemwood volume increment (AVINC) were derived, and implied growth efficiencies (GE) were used to assess magnitude and pattern of GE change between strata and species. AVINC was effectively modeled using LA alone, and the slightly peaking GE - LA relationships approximated those previously conceptualized for shade-tolerant species. A. bulsamea and trees in upper strata were most efficient. T. canadensis maximum GE occurred at a higher LA than A. balsumea or P. rubens, and A. bulsameu maximum GE was less than that previously observed in even-aged stands. Age was not a significant predictor of tree-level GE or stemwood increment. Multi-aged, stand-level stemwood volume increment - leaf area index (LAI) relationships were assessed on 0.02-ha plots dominated by T. cunudensis, A. bulsumeu, and P. rubens. Structures conceptually associated with even-, two-, and multi-aged stands were documented. Inter-plot structural variability was attributed to incomplete stand-level conversion to a balanced condition, and a hybrid single-tree and group selection regeneration method. Plot-level stemwood volume increment (PAVINC) was modeled for all plots combined using LAI and proportion of midstory leaf area (LAC). Two clusters of plots with different basal area, LAI, LAC, and PAVINC were identified. Average diameter distributions within each cluster were typical of multi-aged stands. Within-cluster analysis indicated that quantified structural variables did not improve estimation of PAVINC from LAI, though additional study is warranted due to small sample size. LAI -relative density relationships suggested that plot-level stocking and PAVINC could be increased. This study extends physiologically based silviculture research to mixed-species, multi-aged stands of shade-tolerant species, and confirms the usefulness of LA - growth relationships for describing and evaluating stand dynamics