Abstract
Tree-ring data collected from balsam fir (Abies balsamea [L.] Mill.) in 31 spruce (Picea spp.)/fir-dominated stands in northern New Brunswick were used to evaluate how radial growth reductions during the last spruce budworm (Choristoneura fumiferana Clem.) outbreak (1972–1992) varied according to the hardwood content of surrounding forest landscapes, insecticide spray history and elevation. Stepwise regression analysis was used to build 12 models describing relationships between mean percent growth reduction (i.e., deviation from mean tree growth averaged over the lifespan of all trees in a stand) and the explanatory variables for years when >80% of stands had growth reductions (1976–1978, 1982–1984, 1990, 1991, and 1994) and for three growth reduction periods (1972–1980, 1981–1985, and 1986–1993). The hardwood content of surrounding forest landscapes entered into all significant regression models (P < 0.05), explaining 12–41% of the variance in budworm-caused growth reductions. Differences in growth reductions among stands were greatest during the initial years of the budworm outbreak (1972–1980), particularly in 1976 when mean reductions were 40% in stands with <50% hardwood and content and 20% in stands in which surrounding hardwood content was >50%. Relationships between variation in growth reductions among stands and elevation or insecticide spray history were weak and less consistent than with landscape hardwood content. Our findings suggest that increasing the hardwood content of forest landscapes could help to reduce the impact of outbreaks on timber supplies.