Eighteen years of tree mortality and structural change in an experimentally fragmented Norway spruce forest

Long-term experimental forest fragmentation studies remain uncommon, despite their critical role in the advancement of ecological theory and conservation planning. In 1986 five circular forest fragments (1/16-1 ha) were exposed through clearcutting within an old-growth Norway spruce (Picea abies) forest in northern Sweden. Initial responses to fragmentation (1986-1991) showed very high tree mortality and structural degradation of the fragments. In the present study we re-inventoried these fragments to evaluate tree mortality patterns and structural changes occurring over a longer time period (1991-2004). The fragments can readily be viewed as harvest retention patches or 'woodland key habitats' (i.e., set-aside patches of high conservation value), allowing us to make inferences about the effectiveness of these novel conservation tools. Tree mortality rates dropped markedly (to 1.2-3.9%/year) compared to the initial. responses, yet remained elevated over those of control plots in the nearby unfragmented forest (0.7%). Mortality increased with tree diameter, resulting in smaller-diameter, more homogenous stands. Mortality also generally increased with decreasing fragment size and was dependent of tree location within fragments. Standing death (45% of dead trees, 19912004) replaced uprootings (71%, 1986-1991) as the dominant mode of mortality. Numbers of dying and standing dead trees increased during the second sampling period, further adding to structural change and reduced stand density. Elevated tree mortality resulted in uncharacteristically high volumes of coarse woody debris. Results clearly show that adverse edge-related changes to forest structure and function persist up to two decades after fragmentation. Fragments of this size largely fail as remnants intended to maintain forest interior conditions and late-successional forest structure. However, when embedded within a harvested landscape, they: (1) provide abundant coarse woody debris and snags for deadwood-dependent species that risk extirpation in the surrounding matrix and (2) retain important structures for the developing stands. (C) 2007 Elsevier B.V. All rights reserved.