A total of 110 white spruce (Picea glauca (Moench.) Voss) trees were systematically sampled by 30, 40 and 50 cm diameter-at-breast height (DBH) classes from three natural stands in Saskatchewan located near Big River, Candle Lake and Hudson Bay. Mean ages of the tree samples were 120, 110 and 107 years respectively. Based on sample trees, site indices at breast-height age 50 were 18.5, 18.1 and 18.3 respectively. Wood basic relative density at breast height was determined for each sample tree by X-ray densitometry. Mean values for each tree sample were 0.372, 0.369 and 0.361 respectively. ANOVA of basic relative density on DBH class and stand (R2 = 0.32) revealed that differences in mean density between stands were not significant. The effect of DBH class (rate-of-growth) was significant (p < 0.0001). Consequently, mean relative density values were determined for the 30, 40 and 50 cm DBH classes for the three tree samples combined. These were 0.384, 0.368 and 0.349 respectively compared to the species average of 0.354. Pith-to-bark density trends were inversely related to ring-width trends, consistent with expectations for white spruce.
Density trends observed in Saskatchewan coincided with those obtained from white spruce trees sampled similarly from three stands in northeastern British Columbia and three stands in north central Alberta. In the BC study trends in breast-height wood density were reflected in similar and more significant trends in bending modulus of rupture (MOR) and modulus of elasticity (MOE), MOE in compression, and ultimate compression strength (UCS) of small clear specimens. The combined results of the two previous studies demonstrated robustly that for stands of similar age and site index, wood density and related structural wood properties of white spruce are influenced primarily by rate-of-growth. Pooled results for the three Saskatchewan stands confirmed this wood density/growth-rate relationship.
Considered within each stand, wood density generally declined significantly (a = 0.05) as diameter class increased. A notable exception occurred at Candle Lake. In that stand, although not significantly different, the mean wood density in the 30 cm DBH class was slightly lower than that of the 40 cm class. On review, a similar lack of significant difference in mean density occurred between the 30 and 40 cm DBH classes in two previous samples, one in BC and one in Alberta, but in those stands the density hierarchy remained as expected. One plausible explanation is a deleterious effect of greater competition combined with low site index. The lower than expected density values for small diameter trees coincided with the three lowest site indices of the nine samples.
Wood density of Saskatchewan white spruce was higher than that observed in BC and Alberta with even the 50 cm DBH class showing no significant difference from the species average. This suggests that faster growth can be pursued in Saskatchewan before encountering a detrimental reduction in average wood density. Pronounced increases in annual growth rate that occurred after cambial age twenty in the 40 cm and 50 cm trees at the Big River stand coincided with pronounced declines in breast-height wood density. This was consistent with results observed in two of the previous six samples studied, and strengthens evidence that natural events that result in release will reduce white spruce wood density. Silvicultural interventions that result in similar release can be expected to have a similar effect.