VEGETATION RESPONSES TO EDGE ENVIRONMENTS IN OLD-GROWTH DOUGLAS-FIR FORESTS

Forest edges created by dispersed-patch clear-cutting have become a conspicuous landscape feature in western North America, but the effects of edge on forest structure and function are still poorly understood. In this paper we describe responses of stocking density. growth, mortality, and regeneration for three conifer species from the clear-cut edge into the interior of old-growth forest patches adjacent to 10-15 yr old clearcuts in southern Washington and central Oregon. The significance of edge effects for each variable was tested through a single-factor (distance) analysis of' variance (F test). Relationships between these variables and depth-of-edge influence (i.e., edge width) on old-growth forest were characterized by nonlinear regression models. Near the edge (forest-clearcut boundary line), the old-growth forest has (1) reduced stocking density, as measured by canopy cover, number of stems per hectare. and basal area; (2) increased growth rates of dominant Douglas-fir (Pseudotsuga menziesii) and western hemlock (Tsuga heterophylla), as calculated by an index of relative growth rate; (3) elevated rates of tree mortality, as measured by standing dead and down trees (snags and logs); and (4) greater numbers of Douglas-fir and western hemlock seedlings (less-than-or-equal-to 100 cm tall) and saplings (101-200 cm) but fewer of Pacific silver fir (Abies amabilis). The depth-of-edge influence, when calculated as the point along the clearcut-forest gradient at which a given variable has returned to a condition representing 2/3 of the interior forest environment, ranged from 1 6 to 13 7 m for variables related to distance from the edge. The amount of a square forest patch affected by edge decreased as patch size increased and varied greatly with the depth-of-edge influence. With increasing concerns about organisms and processes that require interior forest habitat, determining the area of residual forest influenced by adjacent clearcut is critical to current and future resource management. Responses of additional biological variables must be explored and information on edge phenomena should be extended to the scale of landscapes.