Modelling the influence of different harvesting methods on forest dynamics in the boreal mixedwoods of western Quebec, Canada

Highlights

• Harvesting spatial pattern had a greater effect on stand dynamics than cut intensity.

• Stands of different development stages responded differently to harvest treatments.

• Clear-cuts and gap cuts favoured high regeneration and basal area increment of aspen.

• Dispersed partial cuts favoured the recruitment and growth of coniferous species.

• Clear-cuts in aspen dominated stands performed similar to stand replacing wildfires.

Abstract

Forest management aims to better understand effects of natural disturbance regimes on forest dynamics and use this knowledge to formulate guidelines in forest planning, thereby narrowing gaps between managed and unmanaged forest landscapes. Using forest simulators to reconstruct forest dynamics in relation to ecosystem processes, including disturbances, could help forest managers to better understand harvesting effects on forest dynamics. Using SORTIE-ND, a spatially explicit forest simulator, we generated stand dynamics for 100 years following simulated clear-cut and partial harvests (dispersed vs aggregated, with 30% and 60% basal area removal). Based on the hardwood: conifer basal area ratio, we grouped post-fire stands into three stand types corresponding to natural post-fire succession (deciduous, mixed deciduous, and mixed coniferous) and assessed long-term effects of clear-cutting and partial harvesting on each. Our results suggest that spatial configurations of harvested and residual trees had a greater effect on stand dynamics than did tree removal intensity. Following dispersed partial harvesting, both deciduous and mixed deciduous stands had species composition and structure similar to unharvested stands of the next successional stage. In these same stand types, aggregated harvests and clear-cutting favoured increased regeneration and basal area increments of aspen, which set succession back to aspen dominance, as has been observed after wildfire. Dispersed partial harvests (both 30% and 60%) and 30% aggregated cuts, in mixed coniferous stands, maintained recruitment and dominance of conifers to levels comparable with unmanaged stands. Clear-cutting in all stand types greatly modified stand compositional and structural attributes, and, when conducted in stands where aspen was abundant, performed as a stand-replacing disturbance, setting succession back to early developmental stages, i.e., to aspen dominance. We conclude that partial harvesting, which emulates gap dynamics similar to undisturbed stands, can maintain natural stand dynamics.

Introduction

Forest management emphasises an understanding of the effects of natural disturbance regimes on forest dynamics, which can be used as templates for forest management and help to reduce the gaps between managed and unmanaged forest landscapes (Grumbine, 1994, Bergeron and Harvey, 1997). The past several decades have witnessed a gradual evolution in forest ecology from static descriptions of forest community distributions along environmental gradients, such as climate and site characteristics (Whittaker, 1975), to the reconstruction of forest dynamics in relation to biotic and abiotic disturbances and ecosystem processes (Bergeron et al., 2002a). In the prolonged absence of large-scale disturbances such as wildfire, canopy gaps control stand dynamics (Kneeshaw and Bergeron, 1998). In this context, efforts have been devoted to developing forest gap models for various forest types where individual trees followed to simulate successional dynamics of forest ecosystems at different spatial scales (Larocque et al., 2011, Elzein et al., 2020). Such models could support forest management (Gauthier et al., 2009) to better identify and minimise the influence of silvicultural practices on managed forests, thereby maintaining forest dynamics that are comparable to what is observed in unmanaged stands (Seymour and Hunter, 1999). Gap models can assess the effects of different silvicultural treatments on stand dynamics over various temporal scales, whilst providing detailed information on the residual composition and structure of managed stands (Coates et al., 2003, Papaik et al., 2010). These models are highly flexible in terms of integrating conditions that are not easily or accurately obtained from empirical growth and yield tables. Implementing these models might also improve forest policy and planning (Landsberg, 2003, Taylor et al., 2009).

Mixedwoods are important sources of wood (Penner, 2008) and are the most diverse and productive forest types within the boreal biome of North America (Chen and Popadiouk, 2002). The dynamics of boreal mixedwoods are broadly regulated by natural disturbances, such as catastrophic wildfire (Payette, 1992, Bergeron, 2000) and insect outbreaks (Morin, 1994), together with anthropogenic disturbances, such as timber harvesting. Following a stand-replacing wildfire, the boreal mixedwoods of eastern Canada typically undergo replacement of shade-intolerant deciduous species by shade-tolerant conifer species (Bergeron, 2000). More specifically, stand development can be described as moving through three successional stages, from hardwood-dominated to mixedwoods, which culminate in conifer-dominated stands (Bergeron and Dubuc, 1989, Bergeron, 2000). The transitional development of the boreal mixedwoods is at the core of a three-cohort model. This tool was developed by Bergeron et al. (2002b), and guides forest management in emulating natural forest dynamics. The three-cohort model is an ecological concept that is based primarily upon stand composition and structure, which vary in both time and space depending upon factors, such as the availability of seed and regeneration sources, secondary disturbances, and the fire cycle (Bergeron, 2000, Bergeron et al., 2002b). For instance, stands experiencing frequent wildfires are dominated by species that are adapted to fire, such as trembling aspen (Populus tremuloides Michaux), paper birch (Betula papyrifera Marshall), black spruce (Picea mariana [Miller] BSP), and jack pine (Pinus banksiana Lambert),. (Bergeron, 2000, Kurkowski et al., 2008). Where longer fire return intervals prevail, abundance of shade-tolerant species increases, especially for balsam fir (Abies balsamea (L.) Miller), white spruce (Picea glauca [Moench] Voss), and eastern white cedar (Thuya occidentalis L.). Increasing proportions of fir and spruce leads to the increased probability of spruce budworm (Choristoneura fumiferana [Clemens]; SBW) outbreaks. Gap dynamics following canopy tree mortality modulate both species composition and structure of post-fire stands throughout stand development (de Römer et al., 2007, Kneeshaw et al., 2011). Species with different ecological requirements, growth rates, and modes of regeneration compete to fill gaps and recruit into the upper canopy (Taylor and Chen, 2011, Colford-Gilks et al., 2012). Nevertheless, species replacement dynamics within gaps is a complex process that depends upon various allogenic and autogenic factors, such as gap size, site conditions, disturbance regime, and stand composition and structure prior to disturbance (Bergeron, 2000, Colford-Gilks et al., 2012).

In recent decades, harvesting with variable tree retention and harvesting patterns (spatial configurations) has been proposed to reproduce the composition and structure of unmanaged stands of the boreal mixedwoods (Gauthier et al., 2009). Within the framework of forest ecosystem management (Grumbine, 1994, Christensen et al., 1996), clear-cuts and other even-aged harvesting systems are used to emulate stand-replacing wildfire, whereas selective and partial harvesting are used to emulate partial disturbances, such as insect outbreaks and gap dynamics (Bergeron et al., 2002b, Gauthier et al., 2009). Moreover, partial harvesting that retains greater residual forest structure is conducted to maintain stand attributes and ecosystem functions similar to those of mid- and late-successional stages by accelerating the replacement of shade-intolerant hardwoods with shade-tolerant conifers (Gauthier et al., 2009). Nevertheless, emulating stand dynamics by harvesting strongly depends upon the size and spatial distribution of gaps that are created as growing spaces for new tree cohorts (Harvey et al., 2002, Bose et al., 2015). For example, high-severity (dispersed or aggregated) harvesting with low tree retention creates large gaps and, accordingly, leads to greater variability in tree size classes through aspen recruitment (McCarthy, 2001), whilst promoting conifer sapling growth when these are present (Brais et al., 2013). Conversely, low-severity dispersed harvesting has been shown to maintain a greater number of structural attributes of even-aged stands (Bose et al., 2015). To consider not only the forest ecosystem management perspectives, but also timber supply and economic needs, the effects of harvesting practices on stand dynamics require long-term evaluation (Ruel et al., 2013). These effects could be considered positive if the growth and survival of residual trees and recruitment are successful over long periods of time, thereby ensuring the continuous supply of timber and ecosystem services (Thorpe and Thomas, 2007).

In this study, we used SORTIE-ND, which is a spatially explicit, individual-based stand dynamics model, to simulate 100 years successional dynamics of three stand types (deciduous, mixed deciduous, and mixed coniferous) in western Quebec, Canada. SORTIE-ND simulates small gaps and, thus, is a particularly suitable model for studying the dynamics of mixed stands related to partial disturbances (Coates et al., 2003). This model has been used to predict and assess the long-term effects of various silvicultural treatments, which could not be obtained by empirical studies (Vanderwel et al., 2011, Bose et al., 2015). Using the simulated outputs, (i) we evaluated the changes in stand dynamics, species growth and recruitment in the three stand types following silvicultural treatments with harvesting intensities ranging from partial harvesting (i.e., 30% and 60% basal area removal) to clear-cuts (93% basal area removal). We hypothesized that both pre-harvest stand composition and harvest intensity define post-harvest changes in managed stands, in which clear-cuts and high harvest intensities in deciduous and mixed deciduous stands would generally induce higher abundances of trembling aspen (H1). Conversely, we anticipated that low harvesting intensities, regardless of stand composition, would favour conifer recruitment to reproduce stand characteristics (in terms of composition and structure) similar to unharvested, natural stands (H2). For partial harvesting, (ii) we also assessed the contribution of tree removal patterns to changes in stand dynamics by comparing dispersed harvesting to aggregated (gap) harvesting with similar levels of basal area removal. We anticipated that for similar basal area removal rates, aggregated cuts would generate larger gaps that trigger recruitment of shade-intolerant species, whereas dispersed tree removal can successfully emulate natural gap dynamics that lead to gradual conifer dominance during stand succession (H3). The objective of the prevailing forest ecosystem management model that has been imposed the study region (Bergeron et al., 2002b, Harvey et al., 2002), is directed towards maintaining the composition and structure of the natural forest mosaic (Bergeron and Harvey, 1997, Brais et al., 2004). Therefore, (iii) we determined which harvesting treatment best emulates the natural dynamics of stands over 100 years of succession following harvesting.