Regeneration patterns on some old-growth and clearcut sites in the Mountain Hemlock zone of southern British Columbia

Patterns of tree regeneration change with elevation in old-growth forest stands on British Columbia's southern coast. At lower elevations, where snow is infrequent, regeneration tends to be centred in canopy gaps caused by the death of one or more trees (the gap model). At higher elevations, where snow usually remains until early summer, regeneration is restricted to elevated microsites that emerge earliest from the snow (the tree-island model). Little is yet known about regeneration patterns in old-growth forest stands between these two systems, those within the forested Mountain Hemlock (MH) biogeoclimatic subzone. Our lack of knowledge has become more of a concern a s logging progresses into higher elevations of the subzone where there are ever-deeper snowpacks. To investigate these patterns, I established 12 study sites in the Tetrahedron Range near Sechelt, 50 km northwest of Vancouver. Six sites were in old-growth forest stands, and six were in naturally-regenerated clearcuts that had been logged 1.1-12 years prior to sampling. Six sites were steep ( - 5 0% slope) and six were flat (~ 2 5% slope). Elevations were slightly higher for old-growth sites (1080-1195 m) than clearcut sites (1060-1100 m). Old-Growth Sites: Trees were very slow-growing and took an average of almost 500 years to enter the canopy layer. Regeneration was most successful on mounds and near to a canopy tree. It was unaffected by overhead canopy cover (that is, the presence or absence of a canopy gap), apparently because of the prevalence of low-angle, diffuse light. In contrast to most forested ecosystems, almost all trees were growing on the undisturbed forest floor rather than on decaying wood or mineral soil. Overall, regeneration patterns were more consistent with the tree-island model than the gap model of regeneration since trees were most likely to survive on mounds and close to a canopy tree. Still, the presence of some regeneration in gaps, especially on steep sites, showed that the study sites occupied a transition between the gap and tree-island models. The tree-island model was best expressed on late-snowmelt sites that were most similar to high-elevation sites. It was also more apparent in the regeneration patterns of Chamaecyparis nootkatensis (Alaska yellow-cedar), a species near the upper limit of its elevational range, than those of Tsuga mertensiana (mountain hemlock), a species in the middle of its range. Clearcut Sites: Almost all trees >150 cm tall were A. amabilis (Pacific silver fir) which had been present in the previous old-growth stand before cutting. A surprisingly high proportion of trees (45%) established within a 3-year window from 1 year before logging through 1 year after logging, more than half of which were C. nootkatensis. Only 2 0% of regeneration established more than one year after logging, and none established >8 years after logging. This limited ingress likely resulted from the absence due to clearcutting of nearby seed-producing trees. There was much more friable forest floor and coarse woody debris (from logging slash) than in adjacent old-stands, but almost all regeneration was still found on undisturbed forest floor. Regeneration was less common on mounds in clearcuts than on mounds in adjacent oldgrowth stands, apparently because mounds were disturbed during logging more than other microsites. There was no evidence that Vaccinium spp. (blueberries and huckleberries) impeded regeneration since 8 4% of trees and seedlings were growing below or amidst Vaccinium and establishment and survival was higher where it was present. The stands that develop on these clearcuts will remain for many centuries dramatically different from the old-growth forest stands they replaced. Where cutting is appropriate, such negative features could be avoided by leaving an adequate seed source, retaining live and dead canopy trees, and protecting sub-canopy trees during cutting. Sites would then also retain many of the old-growth characteristics required by wildlife and other, non-timber values. Results from old-growth sites highlight the site-specific nature of regeneration patterns and the abruptness of the transition to tree-island patterns. Yet low-elevation cutting methods, especially clearcutting, are still used within this transition even where regeneration requires the protection of an overhead canopy. Any presence of regeneration patterns matching the treeisland model should warn forest managers of potential regeneration problems. In such areas, the decision to cut should not be automatic, especially given the slow growth and high non-timber values of these forests. Where cutting does occur, it should leave a s much of the subcanopy and canopy layers as possible. A s snow increases further and there is a greater presence of regeneration patterns matching the tree-island model, any cutting is inappropriate. The relationship between regeneration patterns and snow depths could provide an ecological basis for managing forests within the MH zone. Specifically, the presence of regeneration patterns that match the tree-island model (even if discrete tree islands are not present) is a reflection of severe growing conditions and potential regeneration problems. Simple measures of the relative abundance of tree-island patterns could be added during standard site diagnosis to determine the severity of growing conditions, e.g., the proportion of understory and sub-canopy trees that are growing near a canopy tree or on mounds. Such a classification would be applicable regardless of management objective.Forestry, Faculty ofGraduat