Landscape-level variability in the age underestimation of understory black spruce in the northern boreal forest of Quebec

This study examines the variability of the potential aging error for saplings (height <=1.5 m) of black spruce (Picea mariana (Mill.) B.S.P.) in mature fire-dominated stands (n = 14 stands) of the northern boreal forest of northwestern Quebec. Age underestimation was determined by counting the number of underground bud scars of saplings. The magnitude and variability of age underestimation was compared within and among stands. The relationship between corrected ages (the sum of age underestimation and the number of growth rings at ground level) and ground-level ages was strongly linear and was therefore described with simple linear regression models. To evaluate landscape-level variability in this relationship, the regressions were compared among stands using mixed models. Despite high variability in age underestimation at the stand level, the relationship between corrected and ground-level age was highly significant (p < 0.0001) for all stands except one. However, there were many significant statistical differences between these regressions, indicating high landscape-level variability. The magnitude of age underestimation was found to be highly site specific (means of 7-26 years), the phenomenon being more marked and variable in older stands. Given high landscape-level variability, age underestimation of understory black spruce saplings in northern boreal stands must be documented for every sampled stand to apply a valid age correction in studies that involve multiple stands.