Under what circumstances can the forest sector contribute to 2050 climate change mitigation targets? A study from forest ecosystems to landfill methane emissions for the province of Quebec, Canada

Meeting climate change mitigation targets by 2050, as outlined in international pledges, involves determining optimal strategies for forest management, wood supply, the substitution of greenhouse gas‐intensive materials and energy sources, and wood product disposal. Our study quantified the cumulative mitigation potential by 2050 of the forest sector in the province of Quebec, Canada, using several alternative strategies and assessed under what circumstances the sector could contribute to the targets. We used the Carbon Budget Model of the Canadian Forest Sector to project ecosystems emissions and sequestration of seven alternative and one baseline (business‐as‐usual [BaU]) forest management scenarios over the 2018–2050 period. Three baskets of wood products were used in a Harvested Wood Products model to predict wood product emissions. The mitigation potential was determined by comparing the cumulative CO2e budget of each alternative scenario to the BaU. The proportion of methane emissions from landfills (RCH4%) and the required displacement factor (RDF) to achieve mitigation benefits were assessed both independently and jointly. The fastest and most efficient way to improve mitigation outcomes of the forest sector of Quebec is to reduce end‐of‐life methane emissions from wood products. By reducing methane emissions, the RDF for achieving mitigation benefits through intensification strategies can be reduced from 1.2–2.3 to 0–0.9 tC/tC, thus reaching the current provincial mean DF threshold (0.9). Both a reduction and an increase in the harvested volume have the potential to provide mitigation benefits with adequate RCH4% and RDF. Increased carbon sequestration in ecosystems, innovations in long‐lived wood products, and optimal substitution in markets offer potential avenues for the forest sector to contribute to mitigation benefits but are subject to significant uncertainties. Methane emission reduction at the end of wood product service life is emerging as a valuable approach to enhance mitigation benefits of the forest sector.

from landfills (RCH 4 %) and the required displacement factor (RDF) to achieve mitigation benefits were assessed both independently and jointly. The fastest and most efficient way to improve mitigation outcomes of the forest sector of Quebec is to reduce end-of-life methane emissions from wood products. By reducing methane emissions, the RDF for achieving mitigation benefits through intensification strategies can be reduced from 1.2-2.3 to 0-0.9 tC/tC, thus reaching the current provincial mean DF threshold (0.9). Both a reduction and an increase in the harvested volume have the potential to provide mitigation benefits with adequate RCH 4 % and RDF. Increased carbon sequestration in ecosystems, innovations in long-lived wood products, and optimal substitution in markets offer potential avenues for the forest sector to contribute to mitigation benefits but are subject to significant uncertainties. Methane emission reduction at the end of wood product service life is emerging as a valuable approach to enhance mitigation benefits of the forest sector.

| INTRODUCTION
The carbon budget of the forest sector of a given jurisdiction can be defined as the addition and removal of carbon from ecosystems, harvested wood products (HWP) and the substitution effect that HWP have in markets that affect emissions in other sectors Pukkala, 2017;Skytt et al., 2021;Smyth et al., 2014;Xu et al., 2018). To effectively contribute to greenhouse gas (GHG) emission reduction and climate change mitigation, actions must be taken to improve this sector's carbon budget relative to the business-as-usual (BaU) scenario by increasing forest carbon sequestration, minimizing the emissions caused by HWP use and improving the market penetration of HWP to meet needs that would otherwise be met by fossil fuel-based GHG-intensive materials and energy sources . These actions need to go above and beyond the BaU scenario; only the net differential in GHG emissions between an alternative scenario/course of action and the BaU scenario count as a carbon mitigation benefit in the evaluation of mitigation potential (Smyth et al., 2014). In this study, the carbon budget is characterized by carbon dioxide (CO 2 ) and methane (CH 4 ) emissions, expressed in terms of carbon dioxide equivalent (CO 2 e).
However, there remains uncertainty in the optimal management strategies to implement in forest ecosystems to maximize GHG reductions: should the aim be to maximize wood production (i.e., intensifying management) or to reduce it (i.e., increasing conservation) (Berndes et al., 2018)? In the context of international pledges to attain net-zero emissions by 2050, such as the Canadian Net-Zero Emissions Accountability Act (Government of Canada, 2021), defining the optimal course of action for the forest sector to contribute to these pledges is a challenge. Assessing the mitigation potential of the forest sector is not a straightforward task. One must consider the complexity and interplay of various ecological and industrial factors, often tailored to specific local or regional conditions, that determine the outcomes of climate change mitigation strategies (Chen, Ter-Mikaelian, Ng, et al., 2018;Gustavsson et al., 2017;Lundmark et al., 2014;Pukkala, 2018;Skytt et al., 2021;Smyth et al., 2014).
The province of Quebec, in eastern Canada, has a large commercial forest territory that covers a climatic gradient ranging from northern temperate forests in the south to boreal forests in the north (Saucier et al., 1998). The province has a mature and well-established forest industry that provides HWP for domestic and international markets. In a previous Canadian study, by Smyth et al. (2014), a strategy that involved coupling a reduction in wood harvesting levels with an increase in the proportion of harvested wood going to the manufacturing of long-lived wood products was predicted to provide the highest mitigation potential by 2050 for Quebec's predominant ecozone (Boreal Shield East). Conversely, increased harvest rates (i.e., an intensification strategy) were expected to provide high mitigation potential for other Canadian boreal ecozones as well as some parts of the Boreal Shield East ecozone, which highlights the difficulty of generalizing a single strategy to all regional forest conditions (Moreau et al., 2022;Smyth et al., 2014). The operational implementation of strategies such as forest management intensification or conservation can also be multifaceted. For example, longer harvest rotations (i.e., the period between two harvests), afforestation/ reforestation without additional harvesting, and generally lower harvest rates can all contribute to conservation strategies. Setting aside areas for protection can also contribute to conservation strategies (although we did not test it in our study). On the other hand, intensification strategies can include increased clearcut harvesting, increased procurement of low-quality trees that are part of the allowable annual cut but have been left standing because their fiber does not meet the requirements of sawmills and pulp mills (Durocher et al., 2019), increased commercial thinning rates and increased removal of logging residues Smith et al., 2014).
The fate of harvested wood as it undergoes processing and manufacturing in the forest industry and is then delivered to markets as finished products is a key aspect of the CO 2 e budget for a given forest strategy. Notably, the share of long-lived wood products (e.g., sawnwood) versus short-lived ones (e.g., pulp and paper) that are sourced from harvested wood has been shown to have a significant influence on GHG emissions (Chen, Ter-Mikaelian, Ng, et al., 2018;Smyth et al., 2014;Xie et al., 2021Xie et al., , 2023. Furthermore, bioenergy from forest biomass can be used in place of fossil fuels (Smyth et al., 2017), although the overall carbon benefits of forest bioenergy are affected by multiple factors (Laganière et al., 2017). The organic carbon in wood products is ultimately partially re-emitted to the atmosphere as carbon dioxide (CO 2 ) and methane (CH 4 ) molecules, depending on the type of landfill and the end use considered. The

K E Y W O R D S
biogenic carbon, forest management, forest sector, methane emissions, mitigation potential, substitution, wood products global warming potential (GWP) of CH 4 is 25 for a 100year horizon (GWP100) (Forster et al., 2007). CH 4 emissions are therefore important in the calculation of the carbon balance of the forest sector. These emissions may be lost to the atmosphere, captured for energy or flared to convert CH 4 to CO 2 and thus reduce the GWP associated with wood product decay.
The substitution of GHG-intensive materials and energy sources with wood products is key for the forest sector to contribute to emission reduction and climate change mitigation (Lemprière et al., 2013;Nabuurs et al., 2007Nabuurs et al., , 2022Seppälä et al., 2019). For the purpose of calculating carbon budgets in mitigation analyses, substitution is estimated using displacement factors (DFs) for the different types of wood products considered. A DF is an index that quantifies the amount of emission reduction achieved per unit of wood used compared to a functionally equivalent non-wood material. It is reported in tonnes of carbon avoided per tonne of carbon present in the wood product (tC/tC) (Sathre & O'Connor, 2010). Meta-analyses and literature reviews on the subject of the substitution effect of wood products in Western developed countries give very variable results but seem to be decreasingly optimistic over the years. Indeed, while in 2010 Sathre and O'Connor reported an average DF for the construction sector of 2.1 tC/tC, this factor was reduced to 1.2 tC/tC in Leskinen et al. (2018) and to around 0.55 tC/tC in Hurmekoski et al. (2021) and Smyth et al. (2017) for Canada. Larger DFs can also be found in the literature but they are often used for specific products or uses and, therefore, cannot be easily applied on a larger scale (Cordier et al., 2021;Xu et al., 2018). The estimated average DF for the Quebec province was 0.9 tC/ tC in the current study. Yet, CO 2 e budgets remain very sensitive to DF-related assumptions. Budgets are especially fraught with uncertainty related to the projected evolution of processes and emissions of the products that wood is meant to displace in the future (Harmon, 2019;Howard et al., 2021;Leturcq, 2020).
Considering the wide range of mitigation actions found in the literature, often tailored to different conditions, the implementation of a mitigation strategy for the forest sector (forest ecosystems, products and substitution) requires a regional assessment of the sector's mitigation potential and how it can be achieved by 2050. This study sought to quantify the cumulative mitigation potential of Quebec's forest sector by 2050 for several alternative strategies. We also sought to determine under what circumstances the forest sector could contribute to climate change mitigation targets by 2050. In particular, we explored, through sensitivity analyses, the impact of substitution effects and methane emission pathways on the forest sector's mitigation potential.

| Case study area
The study area corresponds to the province of Quebec's managed forest lands as used in Canada's National Inventory Report (NIR) for annual reporting to the United Nations Framework Convention on Climate Change (Environment and Climate Change Canada, 2020). It includes all managed forest ecosystems below the northern limit of the commercial forest as defined by the provincial government (Jobidon et al., 2015), regardless of the forest land tenure (public or private). Quebec forests have been subject to forest management for several decades. Forests in Quebec that are under public tenure are now managed according to the province's Sustainable Forest Development Act (Gouvernement du Québec, 2013). This act promotes ecosystem-based forest management and aims to reduce the differences between natural and managed ecosystems. It also incorporates regulations governing the protection of soil, water, and biodiversity (Gouvernement du Québec, 2013). All carbon fluxes associated with the forest ecosystems in the study area, the HWP sourced from those ecosystems (whether used domestically or exported), and the substitution effect of these HWP on domestic and international markets were considered according to the methods described in the following section.

| Forest ecosystems
Simulations of carbon emission and removal from forest ecosystems were conducted using the core model of the National Forest Carbon Monitoring, Accounting and Reporting System (NFCMARS), namely the Carbon Budget Model of the Canadian Forest Sector (CBM-CFS3). CBM-CFS3 uses an age-based yield curve approach to simulate stand dynamics, which is well suited for the large portions of the Canadian boreal forest that develop after stand-initiating disturbances (see Kurz et al., 2009;Stinson et al., 2011 for a complete description of NFCMARS and CBM-CFS3). This landscape-level modelling framework was used to project the dynamics of forest carbon stocks and stock changes in biomass and dead organic matter pools over time according to our forest management scenarios (Kurz et al., 2009). The forest data corresponding to our study area were based on a modified 2018 National Inventory Report database (Environment and Climate Change Canada, 2020) for the Quebec region and was provided by Natural Resources Canada. All ecological parameters (decay, growth, etc.) from this data set were left unchanged. We have set the annual harvesting volume and silvicultural operations in accordance with the practices established by the province. The practice of burning residues was not included as it is uncommon in Quebec. It should be noted that the species-specific growth curves (that quantify the relationship between stand volume and age) are two hundred years long, beyond which the model considers the increment to be null, meaning there is no increase or decrease in volume. After 100 years, the majority of growth curves, (i.e., 70%) exhibit a declining trend. A stable trend is observed in 29% of the cases, while approximately 1% demonstrate an increasing trend.
The main natural disturbances of Quebec's managed forest are wildfires and spruce budworm (SBW) outbreaks. Their occurrences were included up until 2018. For the 2019-2050 projections, we applied a constant yearly average area disturbed by wildfire and SBW based on historical values. There was no feedback between changes in forest age-class structure and disturbance risks. After a disturbance, the default assumption in the model is that the new stand will replicate the old one with the same leading species and growth curve. However, if reforestation was set to take place, it would accelerate the rate of regrowth.

| Wood products and substitution
The annual harvested carbon projection from CBM-CF3, that is, the proportion of aboveground biomass classified as merchantable and harvested, was used as an input in a custom HWP model (QC HWP v1). This model was created by modifying the 2018 CBM-HWP model to include Quebec-specific parameters (Moreau et al., 2022). HWP model simulations were run using ANSE v0.9 developed by Natural Resources Canada. This tool tracks harvested carbon stocks and emissions through exportation, wood product use (both locally and internationally), and their end of life in landfills based on the specifications of a basket of wood products. For exported wood products, we considered only trade with the United States of America (U.S.), that is, 38.5% of total sawnwood production, 55.7% of pulp and paper, and 42% of panel-based products, based on expert inputs. The balance remained in the provincial industrial framework, that is, it was not exported elsewhere. We tracked all CO 2 and CH 4 emissions from decay for all products, regardless of their geographical location (the U.S. or Quebec). Additional N 2 O and CH 4 emissions from wood combustion were also considered in our CO 2 e budgets.
In our HWP model, the emissions produced by wood products during their service life are based on a decay function with half-life values. We used the default values suggested by the Intergovernmental Panel on Climate Change (IPCC, 2014)-2 years for pulp and paper products, 25 years for wood-based panels and 35 years for sawn products. For bioenergy, the carbon content of the biomass is considered to be emitted in the year of harvest. We did not consider emissions from wood products harvested prior to 2018 in any of our scenarios; the HWP stock was therefore considered to be empty at the beginning of our simulations. It should be noted that since this initial stock is the same for all scenarios, it cancels itself out in our mitigation potential calculation.
The disposal of wood products was simulated following the same framework used in Smyth et al. (2014). A large part (99%) of wood products was assumed to be sent to landfills, while the rest was assumed to be incinerated (without energy capture). In landfills, 40% of solid wood products (sawnwood and panels) was assumed to be not degradable, while the other 60% was assumed to be degradable with a half-life of 11.7 years (IPCC, 2006;Rüter et al., 2019). For pulp and paper, the proportions were 93% to landfills and 7% to incineration. In landfills, 40% was assumed to be not degradable, while the remaining 60% was assumed to be degradable with a half-life of 11.7 years (IPCC, 2006;Rüter et al., 2019). Landfill carbon emissions were assumed to comprise 50% CO 2 and 50% CH 4 for all products. CO 2 emissions were not captured in any way and were thus simulated as going directly into the atmosphere. CH 4 emissions from wood product decay can be left unrecovered and released to the atmosphere, flared or used as an energy carrier; the fate of CH 4 was different depending on the location considered, that is, the U.S. or Quebec. Quebec landfill carbon emission settings allocated 65.94% of CH 4 emissions to the unrecovered CH 4 stock (100% going to the atmosphere in CH 4 form), 16.83% to the flared stock (99.7% going to the atmosphere in CO 2 form, and 0.3% in CH 4 form), and 17.23% to energy (99.7% going to the atmosphere in CO 2 form, and 0.3% in CH 4 form). Thus, for Quebec landfills, around 33% of the carbon emissions from wood decay released to the atmosphere were in CH 4 form (and 67% were CO 2 ). U.S. landfill carbon emission settings allocated 32.6% of CH 4 emissions to the unrecovered CH 4 stock (90% going to the atmosphere in CH 4 form, and 10% in CO 2 form), 10.6% to the flared stock (99% going to the atmosphere in CO 2 form, and 1% in CH 4 form), and 56.8% to energy (99% going to the atmosphere in CO 2 form, and 1% in CH 4 form). Thus, for U.S. landfills, about 15% of the wood decay carbon emissions were simulated to be released to the atmosphere in the form of CH 4 (and 85% were CO 2 ). In the end, in the context of our study, the carbon emissions in CH 4 form from wood product decay in landfills accounted for about 25% of total carbon emissions (the other 75% were carbon in CO 2 form). The long-term effect of the CO 2 /CH 4 emissions proportion can be analyzed through the GWP using the CO 2 e unit, with a GWP100 of 25 for CH 4 . After converting carbon (C) into CO 2 or CH 4 (calculation based on molar mass) and calculating the GWP100, it was found that CH 4 emissions accounted for 75% of the total annual warming potential of landfills (CO 2 e), despite representing only 25% of C emissions.
The amount of GHG emission reduction that can be achieved in markets by substituting wood products for other materials and energy sources was calculated using DF that compare the fossil fuel emissions reported in product lifecycle analyses (Sathre & O'Connor, 2010). As in Moreau et al. (2022), we utilized DF values from Beauregard et al. (2019) for solid wood (0.91) and composite wood panels (0.77). Based on unpublished data provided by some of the authors of Beauregard et al. (2019), we used updated DF values of 0.76 for bioheat and 1.2 for biofuels in the Province of Québec. This resulted in a mean provincial DF of 0.9 tC/tC for products under substitution, which may vary slightly depending on the alternative scenario being considered. We made the assumption that substitution occurred only if additional wood products were brought to market relative to the BaU scenario. Only the substitution effect of this additional wood was taken into account, and it was accounted for as a carbon sequestration (in tonnes of carbon dioxide equivalent [tCO 2 e]) during the year of wood harvest. Because the DF values already account for fossil fuel emissions from harvesting, transportation, manufacturing of wood products, and also emissions from bioheat production to meet internal needs within the wood industry, we did not calculate them in the study to avoid any double counting. In instances in which a reduction in the amount of wood products brought to market was projected, that is, for strategies that include a reduction in harvest levels, it was assumed that functionally equivalent, more GHG-intensive non-wood products would be used instead of wood products. This was accounted for as a positive value of emissions released to the atmosphere (in tonnes of CO 2 e) using the same DF mentioned above (Moreau et al., 2022).

| Mitigation potential
We used the IPCC simple decay approach (Rüter et al., 2019) for carbon estimation. Thus, we considered both emission and sequestration fluxes between the ecosystem and the atmosphere, as well as emissions from wood products sourced from the case study area, whether they were used domestically or exported to other countries (Rüter et al., 2019). Additionally, DF values were associated with each type of wood product to determine the substitution effect relative to the BaU in markets.
We added carbon fluxes from forest ecosystems and wood products, and from the substitution effect of wood products in markets, and calculated the forest sector's cumulative carbon budget for the 2018-2050 period in CO 2 e. To calculate CO 2 e, we used a GWP of 25 for CH 4 and 298 for N 2 O for a 100-year horizon (Forster et al., 2007).
The mitigation potential of a given alternative scenario was computed as the difference between the CO 2 e budget of the alternative scenario and that of the BaU scenario. With this approach, uncertainty is reduced as all scenarios use the same starting conditions; possible common uncertainties cancel each other out, and the focus is exclusively on the changes due to mitigation actions (Smyth et al., 2014;Xu et al., 2018). Thus, accounting for emissions that occur at the reference level and in each scenario (e.g., bioheat production used for internal needs within the wood industry) becomes unnecessary in a mitigation potential analysis framework because they cancel each other out. A negative mitigation potential value indicates lower emissions or higher atmospheric CO 2 removal than the BaU scenario, while a positive value indicates higher emissions or lower CO 2 removal.

| Forest management
A BaU scenario that reflects current trends in forest management activity across the commercial forest and alternative scenarios with varying activity levels were projected for the 2018-2050 period. The BaU scenario is a modified version of the NIR 2018 dataset (Environment and Climate Change Canada, 2020) for Quebec's forests. The NIR 2018 database includes assumptions about provincial forest landscape characteristics (species, age, volume, etc.), ecological dynamics details (growth curve, decomposition rates, disturbances, etc.) and harvest rate data (harvested volume, rotation cycle, etc.).
In the BaU scenario, the harvest rate is predicted to increase from 30 million cubic meters of wood per year (Mm 3 /year) in 2018 to 31.6 Mm 3 /year in 2050, with a rotation cycle of about 60 years (Beauregard et al., 2019). Over this period, the commercial thinning rate is predicted to stay at about 10,000 ha/year. The recovery of logging residues for bioenergy is currently uncommon in Quebec and therefore applies to only about 1% of the area harvested annually by clearcutting in the BaU scenario (with only 35% of the total biomass of residues from a given clearcut site recovered), which represents about 90,000 oven dry tonnes per year (odt/ year). In Quebec, around 20% of the area harvested by clear-cutting (50,000 ha/year) is replanted to ensure adequate regeneration. The balance (80%) is naturally regenerated.
Alternative scenarios were modelled to reflect future realistic management activities based on either intensification of the BaU scenario-"Plant more"/"Harvest more"/"Harvest low-quality stands"/"Recover logging residues"-or a conservation strategy-"Harvest less"/"Extend rotation"/"Plant more" (Table 1). We simulated an incremental set of management activities (with each scenario adding a new action to the previous scenario) to isolate the effect of each individual mitigation action (Table 1). Each increase (or decrease) in activity was simulated to occur linearly over time.
For the intensification scenarios, the volume of wood harvested was increased over time relative to the BaU scenario from 0 to 4.5 Mm 3 /year by 2050 in the "Harvest more" (I-Harv) scenario and from 0 to 10 Mm 3 /year by 2030 in the "Harvest low-quality stands" (I-HLQS) scenario, the latter of which was in addition to the volume already harvested in the I-Harv scenario. Other activities were also included in our simulations: additional reforestation through the planting of harvested areas (+15,000 ha/year or +30%) in the "Plant more" scenario (I-Refo), additional commercial thinning (+10%) in the I-HLQS scenario, and increased procurement of clear-cut harvesting residues for bioenergy production (+1.2 Modt/year, with 50% of the total biomass of residues from a site being salvaged) in the "Recover logging residues" scenario (I-Resi) as shown in Table 1. For conservation scenarios, the volume of wood harvested was decreased relative to the BaU scenario from 0 to −10 Mm 3 /year by 2030 in the "Harvest less" (C-Harv) scenario. The following actions were also included starting in 2018 (Table 1): longer clear-cut harvest rotation (+15 years) in the "Extend rotation" (C-Rota) scenario and additional reforestation through the planting of harvested areas (+15,000 ha/year or +30%) in the "Plant more" (C-Refo) scenario.
Thus, eight forest management scenarios-that is, one business as usual, four intensification and three conservation scenarios-were included in our simulations (Table 1).

| Wood products
Emissions from the degradation of wood products during their lifecycle were projected using three distinct portfolios of products manufactured from harvested wood. The "Default" (01) basket refers to the province's current basket of products (Beauregard et al., 2019). Two alternative baskets were also considered (Table 2; Figure 1): the "Bioenergy" (02) basket, which included harvest residues and low-quality trees (new bioenergy was entirely from additional harvest) for institutional heat (1.5%), and transport and other biofuel production (98.5%) (Beauregard et al., 2019), and the "Bioenergy + Long-lived products" basket (03), which coupled the "Bioenergy" basket with a higher proportion of long-lived products made from harvested wood. We simulated an increase in the proportion of sawnwood products and panels at the expense of pulp and paper (+4%, +1.41% and −4.25%, respectively, by 2050, see Figure 1 for a complete flow diagram). Finally, the forest management and HWP scenarios were combined to create 16 alternative scenarios, which are presented in Table 2.

| Required displacement factor
We conducted a sensitivity analysis of the impact of the substitution effect of wood products on the forest sector's mitigation potential. We did so by calculating the required displacement factor (RDF) in tonnes of carbon per tonne of carbon (tC/tC), as described by Seppälä et al. (2019). The RDF is the minimum annual DF required to achieve a mitigation benefit relative to the BaU scenario for the 2018-2050 period. We calculated the RDF for all scenarios considering cumulative forest ecosystem emissions, cumulative wood decay emissions and the cumulative change in wood product volume relative to the BaU scenario for the 2018-2050 period.

| Required proportion of methane emissions from landfills
We also conducted a sensitivity analysis of the impact of methane emissions from wood product decay in landfills on the forest sector's mitigation potential. We calculated the proportion of carbon emissions in the form of methane from wood product decay in landfills, hereafter referred as the required proportion of methane emissions (RCH 4 %), needed to achieve a mitigation benefit relative to the BaU scenario (with default carbon emission values being 25% in CH 4 form/75% in CO 2 form) for the 2018-2050 period. The default values used for carbon emissions from wood decay in landfills are set out in Section 2.2.2. The calculation accounted for cumulative CO 2 emissions and removals from forest ecosystems, wood product decay (with default landfill settings) and substitution in markets for the BaU scenario by 2050. At the same time, cumulative CO 2 emissions and removals from forest ecosystems and market substitution for an alternative scenario by 2050 were calculated. As a result, the only variable that could lead to a mitigation benefit compared to the BaU scenario by 2050 was the emissions from wood product decay in the alternative scenario. This allowed us to determine the RCH 4 % value for each alternative scenario. x Yes x I-HLQS_02 x x Yes I-HLQS_03 x Yes x I-Resi_02 Abbreviations: BaU, business-as-usual; CBM, Carbon Budget Model; HWP, harvested wood products.
Note: X indicates a not applicable situation.
The RCH 4 % values for the intensification scenarios should be read as the minimum proportion target for which a given scenario provides a mitigation benefit (sink) relative to the BaU scenario. The RCH 4 % values for the conservation scenarios should be read as the maximum target above which a given scenario would no longer provide a mitigation benefit and thus become a source relative to the BaU scenario. We calculated the RCH 4 % for all scenarios considering cumulative forest ecosystem emissions, cumulative substitution emissions, cumulative bioenergy emissions and the variation in cumulative wood decay emissions from landfills (CH 4 /CO 2 ratio) relative to the BaU scenario by 2050 (see Section 2.2.2). We calculated the RDF as an average value.

| Required displacement factor depending on the methane emissions from landfills
The variations in RDF were at least partially attributable to variations in methane emissions from wood product decay in landfills. Thus, we jointly analyzed the RDF and RCH 4 % values to identify potential interactions between the two factors and determine how a change in RCH 4 % F I G U R E 1 Breakdown of harvested wood among product categories for the three baskets of wood products. (a) "Default" (01) basket of wood products. (b) "Bioenergy" (02) basket of wood products. (c) "Bioenergy + Long-lived products" (03) basket of wood products. relative to the BaU scenario could impact the RDF for each of our scenarios.
Thus, we calculated, for each alternative scenario and for each RCH 4 % value, from 0%CH 4 /100%CO 2 to 40%CH 4 /60%CO 2 , the corresponding minimal annual DF required (RDF) to achieve a mitigation benefit relative to the BaU scenario for the 2018-2050 period. Due to the prevailing political commitment to diminish methane emissions from landfills (Environment and Climate Change Canada, 2022a;Gouvernement du Québec, 2021), it is unlikely that there will be a significant rise in such emissions in the future. As a result, the calculation of RCH 4 surpassing 40% was considered unnecessary.

| Forest management actions
Under the assumptions of the "Default" (01) wood product basket (Figure 2), the main intensification-related forest management action that caused variations in the CO 2 e budget was increasing the harvested volume. Doing so resulted in overall higher CO 2 e emissions ( Figure 2).
The increase in harvested volume associated with the I-HLQS (+10 Mm 3 /year by 2030) and I-Harv (+6 Mm 3 / year by 2050) scenarios strongly impacted the cumulative CO 2 e budget relative to the BaU scenario, increasing emissions by +214 and +40 TgCO 2 e, respectively, by 2050. On the other hand, planting more (the I-Refo scenario) and recovering logging residues (the I-Resi scenario) had little to no effect on the cumulative CO 2 e budget; the differentials in their cumulative CO 2 e budgets compared with the BaU scenario were −3.4 and +0.1 TgCO 2 e, respectively, by 2050.
For the conservation-related forest management actions, the main one that influenced the total CO 2 e budget was directly linked to decreasing the harvested volume ( Figure 2). A harvest decrease of −10 Mm 3 /year by 2030 (the C-Harv scenario) was associated with a net cumulative reduction in CO 2 e emissions of −179 TgCO 2 e by 2050. Conversely, planting more (the C-Refo scenario) and extending the harvest rotation (the C-Rota scenario) had little impact on the cumulative CO 2 e budget compared to the BaU scenario-+2.8 and −5.4 TgCO 2 e, respectively, by 2050 ( Figure 2).

| Basket of wood products
The basket of wood products considered affected the carbon flux attributable to both wood decay in landfills and substitution (Figure 3). A CO 2 e sink effect was observed for all our alternative forest management scenarios when the "Bioenergy" basket (02) or the "Bioenergy + Long-lived products" basket (03) was used instead of the "Default" basket (01) (Figure 3). Using the "Bioenergy" basket (02) instead of the "Default" (01) basket, for a given management scenario, generated a cumulative sink effect of −36 and −52 TgCO 2 e for the I-HLQS and I-Resi forest management scenarios, respectively. This was mainly due to the fact that in the "Bioenergy" basket, more biomass was allocated to biofuel (higher DF) than to bioheat, which created a stronger substitution effect (Figure 3). The "Bioenergy + Long-lived products" basket (03), for its part, produced the largest cumulative sink of the baskets considered. It caused a reduction in emissions of between −23 and −78 TgCO 2 e for the intensification scenarios and around −16 TgCO 2 e for all conservation scenarios relative to the "Default" (01) basket (Figure 3). The cumulative emission reduction of the alternative baskets was directly linked to the harvest level, with higher emission reductions associated with higher volumes of wood allocated to producing products.

| Overall mitigation potential and carbon flux distribution
According to the breakdown of the cumulative CO 2 e flux for the 2018-2050 period in our study framework, the intensification scenarios led to an increase in emissions (relative to the BaU scenario) from forest ecosystems and wood product decay over the period considered ( Figure 4). The only exception is the "Plant more" (I-Refo) scenario, in which forest ecosystems remained a CO 2 e sink at the end of the period considered. The substitution effect, which is recorded as sequestration, was only able to partially offset the overall increase in emissions induced by ecosystems and wood product decay in the intensification scenarios (Figure 4). Therefore, in this analysis, the intensification of forest management lead the forest sector to be a cumulative CO 2 e source in comparison to the BaU scenario over the period considered (Figure 4). The increase in harvested wood volume associated with the intensification strategy tended to increase the forest sector's emissions relative to the BaU scenario over the entire period considered . The difference in carbon emissions ranged from +37 TgCO 2 e for I-Harv_01 and +251 TgCO 2 e for I-Resi_01 (Figure 4).
Conversely, the conservation scenarios lead ecosystems to be a cumulative CO 2 e sink and caused lower emissions from wood product decay compared with the BaU scenario ( Figure 4). However, since fewer wood products were generated in the conservation scenarios than in the BaU scenario, it was assumed that the material and energy needs in markets would be met by functionally equivalent, more GHG-intensive non-wood products (negative substitution), thereby causing an increase in emissions. Yet, these increased emissions were not sufficient to completely offset the overall emission reduction caused by lower harvest levels ( Figure 4). All the conservation scenarios exhibited a similar degree of harvest reduction (Table 1) and a similar cumulative mitigation potential in 2050, the latter of which ranged from −200 TgCO 2 e for C-Rota_03 to −179 TgCO 2 e for C-Harv_01. These results suggested that strategies focused on reducing the annual volume of forest harvests can make the forest sector a cumulative CO 2 e sink in comparison to the BaU scenario over the period considered ( Figure 4).
The basket of wood products considered also had an important impact on the forest sector's mitigation potential. For a given forest management strategy, the "Bioenergy + Long-lived products" (03) basket of products generated lower net emissions, or higher carbon removals, than the "Default" (01) and "Bioenergy" (02) baskets, as is shown, for example, by I-Resi_01 with +251 TgCO 2 e, I-Resi_02 with +199 TgCO 2 e and I-Resi_03 with +173 TgCO 2 e in Figure 4.

| Required displacement factors for wood products
A sensitivity analysis of the impact of the substitution effect of wood products on the forest sector's mitigation potential made it possible to determine RDF F I G U R E 3 Cumulative mitigation effect of alternative wood product baskets relative to the "Default" (01) basket for each forest management scenario and for the 2018-2050 period. CO 2 e budgets include emissions and removals from wood product decay and substitution. Positive values represent a CO 2 e source to the atmosphere, and negative values represent a CO 2 e sink. The results represent the difference between each alternative wood product basket and the "Default" (01) basket of the given management analysed. Due to minimal differences in the volume of wood harvested in the conservation strategies, all the conservationrelated scenarios overlap in this figure (C-All).

Wood product basket:
Bioenergy + Long-lived products (03) Bioenergy (02) values for the scenarios considered. The RDF is the mean value of DF needed each year to achieve a mitigation benefit for the entire 2018-2050 cumulative budget. Because the differences in harvested volume between the I-Refo scenarios and the BaU scenario were small, no RDF values were calculated for the I-Refo scenarios.
The RDF values at which the intensification scenarios could have a mitigation benefit relative to the BaU scenario ranged from 1.2 to 2.3 by 2050 depending on the combination of forest management strategy and wood product basket considered (Table 3). This RDF should be compared to the current average DF value of approximately 0.9, which may vary slightly depending on the alternative scenario being considered. Overall, when a higher proportion of long-lived wood products was assumed to be generated from harvested wood [such as in the "Bioenergy + Long-lived products" basket (03)], the substitution effect required from wood products was lower (i.e., the estimated RDF was lower). With a change of basket, the RDF could be reduced by almost 50% (e.g., in the I-Harv scenario). We also projected lower RDF values with increased harvested volume (Table 3). Conversely, the conservation scenarios that reduce harvesting levels did not provide any mitigation benefit by 2050 relative to the BaU scenario when the DF value for wood products was between 2.8 and 3.4 (Table 3). Moreover, the RDF value was higher in all cases in which we assumed a basket of wood products with a larger share of long-lived products and biofuels [i.e., the "Bioenergy + Long-lived products" (03) basket]. It should be noted that a range of RDF values-those between 1.2 and 3.4-made it possible for both a reduction and an increase in the harvested volume to provide mitigation benefits relative to the BaU scenario by 2050 (Table 3).

| Required proportion of methane emissions from landfills
A sensitivity analysis of the emissions from wood product decay in landfills made it possible to determine values for the required proportion of methane emissions from F I G U R E 4 Quebec forest sector's cumulative carbon mitigation potential and carbon flux distribution for alternative scenarios relative to the BaU scenario for the 2018-2050 period. (a) Intensification scenarios. (b) Conservation scenarios. CO 2 e budgets include carbon emissions and removals from forest ecosystems, wood product decay and substitution in markets. Values are calculated relative to the BaU scenario, and plots are ranked from highest (top left) to lowest (bottom right) cumulative emissions in 2050. Positive values represent a carbon source to the atmosphere, and negative values represent a carbon sink. 01 indicates the use of the "Default" basket of wood products, 02 the use of the "Bioenergy" basket, and 03 indicates the use of the "Bioenergy + Long-lived products" basket. landfills (RCH 4 %). Because the differences in harvested volume between the I-Refo scenarios and the BaU scenario were small, no RCH 4 % values were calculated for the I-Refo scenarios. We found the RCH 4 % values for the intensification scenarios ranged from 14% to 24% compared to the average value of 25% in the BaU scenario. The values determined for the I-Harv_01 scenario (23%) and the I-Harv_03 scenario (24%) in particular were very close to the default CH 4 emission value of 25% of the BaU scenario (Table 3).
A higher proportion of long-lived wood products induced a higher RCH 4 % value and thus reduced the need for CH 4 management (Table 3). Six out of the eight intensification scenarios could enhance the mitigation benefit relative to the BaU scenario if the level of all CH 4 emissions from wood product decay in landfills were reduced to the level used in the U.S. landfill emission settings (15% CH 4 /85% CO 2 ) ( Table 3). The conservation scenarios would not provide a mitigation benefit relative to the BaU scenario if the RCH 4 % value were to increase and exceed the Quebec landfill emission settings for CH 4 , that is, 35% to 34% (Table 3).

| Combined effect of substitution and methane emission management
Overall, a slight variation in RCH 4 % over our time horizon had a significant effect on the RDF for strategies based on a reduction or an increase in the harvested volume ( Figure 5; Table 4). A reduction in RCH 2 % relative to an unaltered BaU scenario could benefit all scenarios ( Figure 5; Table 4).
For intensification scenarios, a decrease in CH 4 emissions, as indicated by a lower RCH 4 % value, had the potential to significantly reduce the RDF ( Figure 5). Decreasing the percentage of RCH 4 from 25% to 15%, which is the average landfill emission settings for CH 4 in this study, to the U.S. landfill emission settings resulted in a decline of RDF from a range of 1.2 to 2.3 to a range of 0 to 0.9 ( Figure 5; Table 4). An increase in RCH 4 % values, on the other hand, led to higher RDF values, especially for the scenarios with a smaller proportion of long-lived products ( Figure 5). All intensification scenarios could provide a mitigation benefit, even without considering the substitution effect, with an RCH 4 % of approximately 6%, relative to the BaU with default settings (Figure 5).
For conservation scenarios, an increase in RCH 4 % led to a decrease in RDF ( Figure 5). A decrease in the RCH 4 % value, on the other hand, led to higher RDF values ( Figure 5). These RDF values were therefore much higher than the DF currently found in the literature. As such, conservation scenarios could also benefit from a reduction in CH 4 emissions from landfills by providing mitigation potential even if high RDF values were reached in the future ( Figure 5; Table 4).

| DISCUSSION
Our study revealed that strategies that aim to reduce overall forest harvesting levels (i.e., conservation strategies) can provide a higher cumulative mitigation potential by 2050 than strategies with higher harvesting levels without needing to change the substitution potential or the management of methane emissions from landfills relative to the BaU scenario. We also demonstrated that these conclusions are extremely sensitive to the underlying assumptions about the proportion of carbon that may be emitted as methane at the end of wood product service life in landfills relative to the current BaU settings. Our estimate of the substitution effect of wood products over the 2018-2050 period was insufficient to offset CO 2 e emissions from ecosystems and wood product decay. If markets could be more efficient at displacing GHG-intensive products with wood products, or at reducing methane emissions from landfills, the forest T A B L E 3 Sensitivity analyses of the RDF or RCH 4 % value needed to reach a mitigation benefit by 2050 for all scenarios considered. Note: RDF and RCH 4 % represent values for the 2018-2050 period; either the RDF or the RCH 4 % value is needed to reach a mitigation benefit relative to the BaU scenario. Because the differences in harvested volume between the I-Refo scenarios and the BaU scenario were small, the I-Refo scenarios were excluded from these sensitivity analyses.
sector might see all conservation and intensification forest management strategies provide mitigation benefits (i.e., be CO 2 e sinks relative to the BaU scenario). Climate change feedback, particularly on natural disturbances, the potential for carbon leakage to other jurisdictions from reduced wood harvests in Quebec, and the saturation of ecosystem carbon stocks, have not been considered in this assessment.  9.44 9.44 9.44 9.44 9.44 9.44 9.44 9.44 9.44 9.44 9.44 9.44 9.44 9.44 9.44 9.44 9.44 9.44 9.44 9.44 9.44 9.44 9.44 9.44 9.44 9.44 9.44 9.44 9.44 9.44 9.44 9.44 9.44 9.44 9.44 9.44 9.44 9.44 9.44 9.44 9.44 9.44 9.44 9.44 9.44 9.44 9.44 9.44 9.44 9.44 9.44 9. 0.9 0.9 0.9 0.9 0.9 0.9 0.9 0.9 0.9 0.9 0.9 0.9 0.9 0.9 0.9 0.9 0.9 0.9 0.9 0.9 0.9 0.9 0.9 0.9 0.9 0.9 0.9 0.9 0.9 0.9 0.9 0.9 0.9 0.9 0.9 0.9 0.9 0.9 0.9 0.9 0.9 0.9 0.9 0.9 0.9 0.9 0.9 0.9 0.9 0.9 0.9 0.9 0.9   The RDF values that we estimated for the intensification scenarios (1.2-2.3 tC/tC) are currently deemed unattainable, as they fall outside the range of DF values recently reported for wood products both globally (Hurmekoski et al., 2021;Leskinen et al., 2018), nationally (Beauregard et al., 2019;Smyth et al., 2017), as well as within our own study (with a mean DF of 0.9 tC/tC). However, the reliance on emission reduction from substitution can be drastically reduced (0-0.9 tC/tC) if the CH 4 emissions from wood product decay in Quebec landfills are reduced to a level that is already achieved in other jurisdictions. The mitigation potential could be even greater if an increased share of long-lived wood products were generated from harvested volumes and used in markets. With optimal methane emission management and product substitution on markets, a reduction or an increase in the harvested volume may be applied simultaneously to different regions of our case study area to optimize the substitution potential of wood products. Our study highlights the need to adequately consider both the substitution potential of wood products on markets and the methane emissions from their decay in landfills at the end of their service life to be able to reduce uncertainty and maximize mitigation potential when designing alternative mitigation strategies for the forest sector. Methane emission management, including the diversion of wood products from landfills, is notably emerging as a valuable tool to increase the forest sector's mitigation potential.

| Forest sector mitigation potential
Several other studies have also suggested that conservation-focused boreal forest management strategies are more adequate for emission mitigation than increasing harvesting (Chen, Ter-Mikaelian, Ng, et al., 2018;Kalliokoski et al., 2020;Pukkala, 2018;Schulte et al., 2022;Skytt et al., 2021;Ter-Mikaelian et al., 2021). Those studies that conclude that an intensification-focused forest management strategy provides greater mitigation benefits are usually based on either the assumption that substitution effect of wood products on markets on markets is higher than what we estimated using regional DF values or different final methane emission pathways (Chen, Ter-Mikaelian, Yang, et al., 2018;Gustavsson et al., 2017;Lundmark et al., 2014;Petersson et al., 2022;Poudel et al., 2012;Smyth et al., 2018). Differences in results can also be explained by variations in the assumptions regarding the half-life of wood products, in cascading use or in accounting and time frame parameters. For example, a time frame that exceeds 2050 might reveal that intensive forest management has longer-term potential benefits over unmanaged old-growth forests in the form of faster regrowth and carbon sequestration. However, Soimakallio et al. (2022) show, based on 45 studies on temperate and boreal forests, that an increase in harvested volume may negatively affect carbon stocks over the short, mid-, and long terms. Yet, evidence suggests that intensification and conservation strategies can coexist at a national and provincial scale to maximize the forest sector's mitigation potential (Moreau et al., 2022;Smyth et al., 2014). Some ecosystems may indeed benefit from anthropogenic disturbances that release stand growth potential by enabling the regeneration of new vegetation cohorts (Landry et al., 2021;Moreau et al., 2022).

| Forest management
Our results may be mitigated by the fact that all harvesting in our simulations was performed by clear-cutting. In boreal and northern temperate stands, this type of logging has been shown to create a period of net emissions to the atmosphere in the years following harvesting due to the emissions from decomposition being significantly higher than annual photosynthesis-driven sequestration for 10-30 years Paradis et al., 2019;Senez-Gagnon et al., 2018). Thus, other silvicultural regimes such as continuous-cover forestry may provide greater mitigation potential by limiting the net emissions attributable to harvesting (Laiho et al., 2011;Paradis et al., 2019) while maintaining timber production (Pukkala, 2014). These alternatives could make it possible to develop forest management strategies that synergize the maximization of carbon sequestration and timber production.
While conservation strategies may limit emissions from forest ecosystems, forest carbon stocks can be vulnerable to natural disturbances. Our projections were based on only the yearly average area disturbed by wildfire and SBW in the province, without taking into account the vulnerability of stands and landscapes to natural disturbances that may emerge due to management strategies or climate change. Strategies that involve reducing the annual harvested volume might be more at risk of carbon inversion caused by natural disturbances due to the abundance of mature and old-growth forests that such strategies create (Sharma et al., 2013). The significant uncertainty related to the carbon sink capacity of those stands over the long term (Gao et al., 2018;Gundersen et al., 2021;Smyth et al., 2020) may also be a concern, as ecosystem carbon stocks may eventually become saturated (Harel et al., 2021). Yet, uncertainty regarding the net carbon sink potential of mature and old-growth stands should not overshadow their role as carbon reservoirs (Harel et al., 2021). On the other hand, expected changes in natural disturbance regimes could negatively impact the productivity and regeneration capacity of forest ecosystems (Boulanger et al., 2017) and thus jeopardize the sustainability of wood production (Bouchard et al., 2022;Boucher et al., 2018;Brecka et al., 2020). Changes in natural disturbance regimes could become a major issue beyond the period covered by our study  and result in additional future uncertainty. Therefore, additional research is needed to ensure that the combined impacts of natural disturbances and forest management on the carbon balance are captured, especially in the context of a changing climate that is expected to dramatically alter forest disturbance dynamics (Metsaranta et al., 2023;Seidl et al., 2017).
Our study did not take into account the effect of climate change on forest ecosystems. In their study, Moreau et al. (2022) indicate that the carbon budget of the boreal forest sector is expected to remain relatively unaffected by climate change under RCP 4.5. However, the northern temperate forest sector located south of the province may experience a greater impact. It could potentially become a carbon source under RCP 4.5, even in the absence of changes in forest management or wood product use. Additional sources of uncertainty include changes in surface albedo and biogenic volatile organic compounds that will need to be assessed and taken into account in future mitigation assessments (Williams et al., 2021).
Our scenario that involved reducing the annual harvested volume did not capture potential carbon leakage to areas where forest management is less regulated and potentially less sustainable. Carbon leakage, that is, the relocation of carbon emissions from an emission-constrained region to unregulated areas, is a major concern for the forest sector around the world (Nabuurs et al., 2007. Carbon leakage can significantly compromise the effectiveness of Quebec's forest-based climate policies, as a wood harvest reduction in one area can trigger an intensification of forest management in other regions due to domestic or international linkages in forest product markets (Gan & McCarl, 2007). The province of Quebec faces a potential carbon leakage risk caused by changes in its wood harvesting practices. This risk, associated with both increasing and decreasing the harvested volume (Kallio & Solberg, 2018), could account for 60%-100% of the volume variation resulting from the adoption of an alternative harvesting approach (Hu et al., 2014;Kallio & Solberg, 2018;Nepal et al., 2013). Forecasts of an upsurge in demand for bio-based resources in the forthcoming decades (FAO, 2022;Hassegawa et al., 2022;Hurmekoski et al., 2023) substantiate the notion that the carbon leakage hazard could be amplified, especially in the case of reduced harvesting. Thus, uncertainty about carbon leakage at the provincial level should not be ignored when implementing climate policies that are based on a conservation-focused forest management strategy. Based on these results, efforts must be made to anticipate and understand potential market responses to regional wood harvest variations.

| Impact of methane emission pathways
Although managing methane emissions from landfills is an important topic and a viable tool to incorporate in mitigation policies (Chen et al., 2014;Environment and Climate Change Canada, 2022b), alternative methane emission pathways from landfills are rarely assessed in the forestry literature. Our study demonstrates the potentially large impact such emission pathways can have on the design of mitigation strategies for the forest sector. Some RCH 4 % may also be considered in our uncertainty value range, and such intensification (I-Harv) scenarios could therefore already be CO 2 e sinks relative to the BaU scenario by 2050. In fact, a slight reduction in the simulated methane emissions from wood product decay in landfills can yield large increases in mitigation potential. The short-term uncertainty regarding methane emissions from landfills is low, and technical innovations are already available to reduce unrecovered CH 4 and maximize flaring and energy uses. Reducing CH 4 emissions relative to the BaU scenario would greatly benefit all forest management strategies. Eventually, biogenic methane capture and use as an energy carrier might replace some fossil fuels and create additional substitution effects. Improved management of landfill methane emissions through diversion of organic material, methane capture or flaring should be a high priority for climate change mitigation in the forest sector. Canada's national and provincial governments are currently pushing for protocols to increase landfill methane recovery and destruction, especially in the Canadian Greenhouse Gas Offset Credit System Regulations (Environment and Climate Change Canada, 2022a;Gouvernement du Québec, 2021). Additional mitigation efforts can be aimed at diversion of wood products from landfills through increased recycling and diversion to incineration facilities that reduce methane emissions and capture energy.
Disregarding the temporal aspect of GHG emissions in the forest sector may result in significant misinterpretations of the sector's actual potential for mitigation and its impact on global warming (Brandão et al., 2013;Breton et al., 2018;Levasseur et al., 2010;Moreau et al., 2023). The use of static GWP values introduces inconsistencies between the chosen GWP time horizon (100 years) and the study period (30 years). Any analysis based on biogenic carbon strategies should consider such temporal issues, as emphasized in previous literature (Brandão et al., 2013). Although the use of a uniform static GWP for wood products is a prevalent practice in the literature (Smyth et al., 2014(Smyth et al., , 2017, employing dynamic assessments can offer a more precise representation of the effects of biogenic carbon emissions and removals from forest ecosystems, emissions resulting from wood products decay and substitution effects (Breton et al., 2018;Moreau et al., 2023). Using a single GWP100 may not fully capture the effects of methane emissions on climate change; including a GWP20 to account for the short-term warming potential of methane may be needed in future studies (Ocko et al., 2017). By doing so, the urgent need to manage methane emissions in the forestry sector and mitigate their impact on climate change can become even more evident (Moreau et al., 2023).

| Impact of substitution and related uncertainty
In the absence of improved management of methane emissions from landfills, the current estimate of the substitution effect of wood products is insufficient to offset CO 2 e emissions from ecosystems and wood product decay in landfills over our time horizon.
The range of RDF values that we estimated (1.2-2.3 tC/ tC) for which intensification of forest sector management can provide a mitigation benefit relative to the BaU scenario (without a change in CH 4 emissions) is in line with RDF values estimated in a similar study from Finland (1.3-2.4 tC/tC) (Seppälä et al., 2019) and higher than the ones calculated for Germany (<1.5 tC/tC) (Christian & Stefan, 2022).
It therefore appears that Quebec's forest sector strong dependence on the substitution effects to reach mitigation targets can be drastically reduced through improved landfill management. The forest sector may have realistic mitigation potential from substitution that could be achieved in the near future, if and only if new methane management measures are put in place (RDF 0-0.9 tC/tC). Otherwise, it is unrealistic to consider that average values between 1.2 and 2.3 tC/tC can be reached without the need for significant technological breakthroughs (Hurmekoski et al., 2021;Leskinen et al., 2018;Smyth et al., 2017). Methane emissions from wood product decay in landfills appear to be a key element to transform a forest sector management strategy into an efficient climate change mitigation tool. It should also be noted that new wood-based products are being developed and may represent significant opportunities for replacing fossil fuel-based products (textiles, liquid biofuels, platform chemicals, plastics, and packaging) in existing and emerging markets (FAO, 2022; Hassegawa et al., 2022;Hurmekoski et al., 2018). New substitution potential needs to be assessed .
We identified a range of RDF values (1.2-3.4 tC/tC at 25% RCH 4 % and 0-4.9 tC/tC at 15% RCH 4 %) at which both a reduction or an increase in the harvested volume would provide a mitigation benefit relative to the BaU scenario by 2050. This suggests that the two strategies could be used together to optimize Quebec's mitigation potential further, as each strategy may be individually relevant depending on the initial forest characteristics (Moreau et al., 2022;Smyth et al., 2014), ownership and land management objectives. It is important to note that our study only evaluated the negative substitution effects in specific product categories such as sawnwood and panels. This overlooks the possible substitution of paper and paper-based packaging products with alternatives that may have higher GHG emissions, which could be a significant limitation of the negative substitution assessment (Gustavsson et al., 2022). This limitation is expected to become more pronounced over the next decades due to increased demand for packaging, despite the projected decline in demand for graphic paper (FAO, 2022).
Any future improvements in the GHG footprint of industries that produce carbon-intensive products (e.g., cement and steel) would reduce the forest sector's climate change mitigation potential due to wood products having a lower substitution effect (Harmon, 2019). Many other assumptions used in determining DF values are only partially supported by the literature, such as assumptions related to market dynamics, that is, pricing, carbon leakage, and rebound effects (Harmon, 2019;Howard et al., 2021;Leturcq, 2020). Furthermore, using high DF values for wood products may imply that international mitigation targets will not be reached by the most carbon-intensive sectors in the coming decades. Thus, due to sources of considerable uncertainty, it does not seem appropriate for the forest sector to rely solely on the substitution effect of wood products as a way to provide mitigation benefits. However, considering both substitution and alternative CH 4 emission pathways can reduce the uncertainty by decreasing the need for high DF values.

| Impact of wood products
Our study supports the idea that the basket of wood products considered has a strong impact on the forest sector's mitigation potential. One way in which it does so is by increasing the substitution effect when pulp and paper, that is, products with no or low substitution effect, are replaced by sawnwood, panels, or bioenergy (Chen, Ter-Mikaelian, Ng, et al., 2018;Chen, Ter-Mikaelian, Yang, et al., 2018;Hurmekoski et al., 2020;Nabuurs et al., 2007;Smith et al., 2014;Xie et al., 2021). Another way is by delaying decay emissions when the half-life of a substitute wood product is longer than that of the product it replaced, which increases the carbon stock in the wood product pool and reduces decay emissions (Chen, Ter-Mikaelian, Ng, et al., 2018;Chen, Ter-Mikaelian, Yang, et al., 2018;Dugan et al., 2018;Parobek et al., 2019;Smith et al., 2014). Furthermore, a small change in the proportions of the products in the basket has a significant impact on mitigation potential: our "Bioenergy + Long-lived products" (03) basket was characterized by a relatively minor change in sawnwood products (+4%), panels (+1.41%), and pulp and paper(−4.25%). Larger shifts to long-lived products would further increase the mitigation benefits (Xie et al., 2021(Xie et al., , 2023. However, our results suggest that the emission reduction generated by improving the proportions of the wood products in the basket would occur regardless of the forest management strategy (intensification or conservation) implemented in the case study area.
Innovation in wood products should be conducted with end-of-life disposal in mind. If new long-lived products can be easily recycled or used as energy carriers, substitution may increase and methane emissions from wood decay in landfills may be reduced. There is inherent uncertainty in innovation, however, as economic parameters and consumer demand are not always predictable. The question that remains, which was not addressed in our study, concerns the volume of products with high DF values that markets can absorb (Xie et al., 2021(Xie et al., , 2023. Specific products with high mitigation potential can be designed, but their final substitution effect will depend on their effective use. This suggests that the forest sector's longer-term role in a low-carbon economy will need to evolve over time and adapt to emerging needs. Policies such as "wood first" or changes in building codes that encourage mass timber uses or requirements to account for embodied energy in building construction can further help increase market uptake of long-lived wood products. We did not focus on the cascading use of wood in our study. Thus, almost all retired products were sent to landfills (the most common practice in Quebec at the moment) in our simulations. The cascading use of wood can have a direct mitigation benefit relative to the BaU scenario by increasing how long carbon is stored in products and by creating new opportunities to replace GHG-intensive products in markets (Brunet-Navarro et al., 2018). For example, in Scandinavia, up to 70% of wood products and 30% of paper products are used for bioenergy production at the end of their service life (Zubizarreta-Gerendiain et al., 2016), which suggests that there is no major technological barrier for this alternative in Quebec. Increasing the cascading use of wood, especially the diversion of wood and paper products from landfills, may represent a significant mitigation potential for the forest sector.

| CONCLUSION
Our study confirms the complexity and interacting factors that determine the outcomes of climate change mitigation strategies based on the forest sector. The most important and novel finding of this study is that because end-oflife disposal of wood products in Quebec involves predominantly landfilling with limited methane emissions management, increases in wood harvest and production of wood products result in net increases in GHG emissions: a substantial proportion of CO 2 removed from the atmosphere through tree growth is released back to the atmosphere as methane (CH 4 ) with a 25-fold higher GWP than CO 2 . Conducting measurements of CH 4 emissions using the 20-years GWP (GWP20 of 86) or dynamic radiative forcing would emphasize the urgent requirement for implementing alternative practices to manage emissions from landfills. Thus, the fastest and most efficient way to improve mitigation outcomes for the forest sector relative to the current reference scenario is to reduce end-of-life methane emissions from wood products.
Our results show that even small reductions in end-oflife methane emissions can turn a strategy that increases harvest rates into a climate-effective strategy relative to the current reference scenario. This approach will also lessen the requirement for a challenging and almost unattainable substitution effect for these scenarios. Moreover, the required reductions of methane emissions for Quebec are already achieved in other jurisdictions through the greater diversion of wood products from landfills and better flaring or capture of methane from landfills. Other aspects of wood product use-in particular, the proportion of wood harvest allocated to products with longer life spans (e.g., mass timber and other engineered wood products)-and the displacement of emissions-intensive materials can further increase climate mitigation benefits of wood product uses.
A portfolio of mitigation actions, focused on regionallydifferentiated timber harvest variation, improved utilization of wood, and a greater allocation to long-lived products with high displacement factors, can contribute to climate change mitigation. However, they remain subject to considerable uncertainty and temporal issues. Reduction in methane emissions at the end-of-life disposal of wood products, which should present little technical uncertainty in the short term, is emerging as a valuable tool to increase the mitigation potential of the forest sector.
Further research should address factors not considered here, including reduced carbon sinks due to saturation of forest ecosystem carbon stocks in aging forests, leakage of harvest to other jurisdictions, opportunities for improvements to silvicultural methods, temporal dynamic assessments, real substitution benefits as other sectors decarbonize their production or changes in disturbance regimes with forest management and climate change impacts.

ACKNO WLE DGE MENTS
We are thankful to NRCan for the provision and training of the HWP and CBM modelling tools.