What is the distance of edge influence on the structure and understorey composition at 16–17-yr-old cut edges in black spruce boreal forest? How do these edges compare with more recent 2–5-yr-old cut edges in the same region? Northwestern Quebec, Canada. Forest structure and understorey composition were sampled along transects perpendicular to ten 16–17-yr-old clear-cut edges, and compared to published results from 2–5-yr-old cut edges. We used randomization tests to assess the magnitude and distance of edge influence, and to compare edge influence between different edge ages. Black spruce forest next to the 16–17-yr-old cut edges was structurally and compositionally very similar to interior forest, with little edge influence from harvesting beyond 5 m into the forest. Edge influence on the understorey was weak (low magnitude) and not very extensive (short distance) at these edges, with no significant edge influence on the abundance of individual species. Logs peaked in abundance on the forest side of the edge, with values higher than in either adjacent ecosystem. Overall, 16–17-yr-old cut edges in black spruce forest showed little evidence of further structural change compared to the 2–5-yr-old cut edges. Structural development of these edges as well as regeneration of the disturbed areas also resulted in reduced edge influence on the understorey. Instead, clear-cut edges in black spruce forest may experience more forest influence on the regenerating disturbed area.

Boreal forest landscapes are dynamic with stands in different stages of development following stand-replacing disturbances such as fire and insect outbreaks. Forest edges are an important component of these heterogeneous landscapes but there have been few studies on intermediate-aged forest edges which are needed for a comprehensive perspective on the spatiotemporal dynamics of forest edges. We described the structure, composition and extent of edge influence at 13, 25 and 39-year old fire edges in black spruce boreal forest in northwestern Québec and northeastern Ontario to characterize their structural development and to assess effects of edge development on the understorey. Forest structure and understorey composition were sampled along transects perpendicular to edges of the fires. Edge influence was assessed using randomization tests. Black spruce forest was relatively unaffected by edge influence beyond 5 m into the forest at all ages of edges studied. Edge influence on the understorey was weak and not extensive at intermediate-aged edges with few consistent responses of individual species. Less decayed snags and logs at 13 and 25-year old edges peaked in abundance at or near the edge with values higher than in either adjacent ecosystem. Overall, intermediate-aged fire edges in black spruce forest showed little evidence of further changes in canopy structure with time. Structural development of these edges as well as the regeneration of the disturbed areas also resulted in reduced edge influence on the understorey. A new insight from our study is that intermediate-aged forest edges may contribute unique structural features to landscapes such as a reservoir of deadwood that may be important for wildlife species.

The simultaneous influence from multiple edges on remnant forest patches (such as wildlife corridors, protection buffers, small unharvested remnants or corners of larger patches) in harvested forest landscapes could impair, or possibly enhance, their effectiveness. When multiple edges are in close proximity, there may be interactions of edge influence such that the observed response is greater or less than would result from the influence of either edge alone. We examined possible ways in which two nearby forest edges of similar or different types might interact in terms of their influence on forest structure. We present an ‘edge interaction’ model for three possible types of interaction of edge influence: (1) no interaction, edge influence is limited to the strongest influence from either edge, (2) positive interaction, observed edge influence is greater than from either edge alone; (3) negative interaction or resistance, the influence from both edges is less than from a single edge (e.g., an older edge is resistant to effects from a younger edge). Empirical data for forest structure at the edges of cutblocks (harvested areas) and water bodies were entered into the models to predict edge influence in narrow forest corridors assuming the null hypothesis of no interaction. Randomization tests were used to compare predictions to observed edge influence on recently-fallen logs and Populus spp. (P. tremuloides Michx. and P. balsamifera L.) sapling density in lakeshore buffers in boreal mixedwood forest as well as on canopy cover and log, tree and snag abundance in riparian buffers and forested corridors separating cutblocks in Picea mariana (Mill.) BSP. forest. In lakeshore buffers, there was evidence of both positive and negative interaction at different locations within the buffer or at different times since buffer creation for both the abundance of logs and Populus sapling density. Trends suggested positive interaction for canopy cover and live tree density in riparian buffers and cutblock separators, and for snag density near the cut edges of riparian buffers. Testing hypotheses arising from our model of interaction of edge influence could lead to a clearer understanding of edge influence in fragmented landscapes.

Abstract

Questions: 1. How does the spatial structure of the organic layer affect tree sapling physiology? 2. Are the organic layer and Picea mariana height growth spatially structured at different scales? 3. Does microtopography influence the accumulation of organic matter and does organic layer thickness affect height growth?

Locations: Picea mariana forests, northwestern Quebec, Canada.

Methods: We assessed the spatial pattern of each variable in one wildfire site and one harvest site using semivariograms and correlograms. We measured the cross-correlation between relative elevation and organic layer thickness, and between organic layer thickness and growth using cross-correlograms.

Results: Picea mariana height growth was autocorrelated to a greater extent in the wildfire site (103 m) than in the harvest site (43 m). The spatial structure of organic layer thickness was similar in both sites. Deeper depressions in the harvest site, as illustrated by spatial variance in relative elevation at short distances (ca. 50 m), and by high autocorrelation values, increased the accumulation of organic matter within 20 m.

Conclusions: The interaction between microtopography and organic matter accumulation led to paludification and poor growth of Picea mariana at the harvest site. Paludification at the wildfire site was independent of microtopography and was probably a result of stand development.

Questions: How do gap abundance and the spatial pattern of trees and snags change throughout stand development in Picea mariana forests? Does spatial pattern differ among site types and structural components of a forest?

Location: Boreal forests dominated by Picea mariana, northern Quebec and Ontario, Canada.

Methods: Data on the abundance, characteristics and spatial location of trees, snags and gaps were collected along 200 m transects at 91 sites along a chronosequence. Spatial analyses included 3TLQV, NLV and autocorrelation analysis. Nonparametric analyses were used to analyse trends with time and differences among structural components and site types.

Results: Gaps became more abundant, numerous and more evenly distributed with time. At distances of 1-4 m, tree cover, sapling density and snag density became more heterogeneous with time. Tree cover appeared to be more uniform for the 10- 33 m interval, although this was not significant. Patch size and variance at 1 m were greater for overstorey than for understorey tree cover. Snags were less spatially variable than trees at 1 m, but more so at intermediate distances (4 - 8 m). Few significant differences were found among site types.

Conclusions: During stand development in P. mariana forest, gaps formed by tree mortality are filled in slowly due to poor regeneration and growth, leading to greater gap abundance and clumping of trees and snags at fine scales. At broader scales, patchy regeneration is followed by homogenization of forest stands as trees become smaller with low productivity due to paludification.

We investigated stand development along a chronosequence on organic, clay and sand sites in black spruce boreal forest in northwestern Quebec, Canada. Our objectiveswere: (1) to describe trends and stages of structural development following fire; (2) to compare trends and stages of development both in isolation from and in conjunction with species replacement. We tested the hypothesis that although trends in structural development are similar among site types, productivity and composition affect the timing of developmental stages. Data on live trees, snags and logs were collected at 91 sites. Trends with time since fire were analyzed using segmented piecewise linear regression. On organic sites, tree basal area and density increased continuously with time since fire, while deadwood abundance decreased and then increased. Live tree basal area, tree density and deadwood abundance generally followed expected S-, N- and U-shaped trends, respectively, on clay sites, but often with decreases in later stages due to paludification. Fewer trends were significant on sand sites, although tree basal area decreased likely due to a change in species composition. Older forests on all site types weremore structurally diverse. To estimate the timing of the stages of structural development, we introduce a newanalysis technique which uses the breakpoints of the piecewise regressions.On organic sites, only three stages of stand development were evident, whereas a four-stage stand development model was appropriate for both clay and sand sites.We found that local conditions affected not only the timing of developmental stages, but also the number of stages and the trends themselves.We attributed these differences to changes in species composition and productivity. We refine the theory of structural development by representing patterns in both live and deadwood as two-stage trends with two possible outcomes for each stage. Our new method of determining the timing of the developmental stages using empirical data can be used to develop management practices that emulate structural development in order to conserve biodiversity on a landscape scale. © 2005 Elsevier B.V. All rights reserved.

We compared structure and composition at forest edges created by wildfire and clear-cutting in black spruce (Picea mariana (Mill.) BSP) dominated boreal forest in northwestern Quebec. Forest structure and plant species composition were sampled along transects perpendicular to eight 3- to 4-year-old fire edges and eight 2- to 5-year-old cut edges. Significance of edge influence was assessed by comparing mean values at different distances from the edge to the range of variation in interior forest. The influence of clearcut edges was minimal, generally extending only 5 m from the edge, and included greater log density and different species composition, compared with interior forest. At fire edges, prominent responses to edge creation included increased snag density and lower moss cover, compared with interior forest, extending up to 40 m into the forest. This initial structural change was likely due to partial burning extending into the forest. Overall, fire edges had more snags and a different species composition than cut edges. Our hypothesis that edge influence is more extensive at fire edges than at cut edges was supported for overstory and understory structure, but not for species composition. We suggest that there is a need for management to consider the cumulative effect of the loss of fire edges on the landscape.

Nous avons comparé la structure et la composition des bordures de feu et de coupe dans la forêt Boréale du Nord-Ouest du Québec dominée par l'épinette noire (Picea mariana (Mill.) BSP). La structure et la composition ont été échantillonnées le long de transects perpendiculaires à huit bordures de feu de 3 à 4 ans et à huit bordures de coupes totales de 2 à 5 ans. L'influence de la bordure a été évaluée en comparant les valeurs moyennes aux différentes distances de la bordure à l'étendue de la variation à l'intérieur de la forêt. L'influence de la bordure des coupes totales était minimale, s'étendant généralement jusqu'à seulement 5 m de la bordure. Ces changements comprenaient une augmentation de la densité des arbres morts au sol et une composition en espèces différente de celle observée à l'intérieur de la forêt. Aux bordures de feux, une densité accrue des chicots et une diminution de la couverture de mousses, comparativement à l'intérieur de la forêt, s'étendaient jusqu'à 40 m en forêt. Ces changements structuraux étaient probablement dus au brûlage partiel qui s'est étendu dans la forêt. De façon générale, les bordures de feu avaient plus de chicots et une composition différente d'espèces comparativement aux bordures de coupes. Notre hypothèse voulant que l'influence de la bordure soit plus grande pour les bordures de feu comparativement aux bordures de coupe est supportée pour la structure des étages supérieur et inférieur de la canopé mais non pour la composition des espèces. Puisque l'influence des bordures de coupe est limitée dans la pessière noire boréale, les aménagistes forestiers devraient davantage tenir compte des effets cumulatifs de la perte d'hétérogéneité structurale des bordures issues de feu dans les paysages aménagés.©2004 NRC Canada