Sustainability in Boreal Forests: Does Elevated CO₂ Increase Wood Volume?

Abstract

While boreal forests constitute 30% of the Earth’s forested area and are responsible for 20% of the global carbon sink, there is considerable concern about their sustainability. This paper focuses on the role of elevated CO₂, examining whether wood volume in these forests has responded to increased CO₂ over the last 60 years. To accomplish this, we use a rich set of wood volume measurement data from the Province of Alberta, Canada, and deploy quasi-experimental techniques to determine the effect of elevated CO₂. While the few experimental studies that have examined boreal forests have found almost no effect of elevated CO₂, our results indicate that a 1.0% increase in lifetime exposure to CO₂ leads to a 1.1% increase in aboveground wood volume in these boreal forests. This study showcases the value of research designs that use natural settings to better account for the effects of prolonged exposure to elevated CO₂. Our results should enable improved delineation of the drivers of historical changes in wood volume and carbon storage in boreal forests. In addition, when combined with other studies, these results will likely aid policymakers in designing management or policy approaches that will enhance the sustainability of forests in boreal regions.

1. Introduction

Boreal forests, which cover 30% of the Earth’s forest area, are a critical component of the global carbon budget, contributing 20% of the global carbon sink [1,2]. While increases in atmospheric CO₂ likely have already enhanced the amount of carbon stored in these forests [3], some research suggests that these forests have immense potential to store even more carbon [4] through carbon fertilization, a process where elevated CO₂ results in greater average wood volume due to an increase in the efficiency of photosynthesis. However, in recent years, bug infestations and forest fires have slowed the accumulation of carbon in boreal forests [5,6]. Worries about these emerging trends have long been present [7], but only over the past couple of decades have data supported modeling projections.

Carbon fertilization might have an important role to play in offsetting these negative trends and in increasing carbon storage, but its effects, especially in boreal forests, have not been well quantified. One recent study found that increased fire-related carbon emissions were strongly correlated with carbon fertilization effects [8]. However, many studies have argued that there are stronger drivers of wood volume than elevated CO₂, such as nutrient availability, species-specific traits, and local growing conditions [9]. Soil nutrient limitations, particularly from phosphorus competition among microbes, have been shown to significantly constrain the growth response to elevated CO₂ [10,11]. It has also been found that physiological acclimation may dampen the initial positive effects of elevated CO₂ on tree growth [12,13]. Other research has found that, while elevated CO₂ results in increased water-use efficiency, this efficiency does not consistently translate into enhanced biomass accumulation [14,15]. Thus, the earth system may be transitioning from a carbon fertilization-dominated system to one dominated by climate change itself, given limitations in other important nutrients [16].

In light of accelerating climate change, understanding whether elevated atmospheric CO₂ has encouraged increased wood volume in boreal forests, and if so to what degree, is a pressing research priority. Experiments in temperate forests have demonstrated a strong and significant relationship between elevated CO₂ levels and increased tree volume [17,18,19]. In boreal regions, the impact of elevated CO₂ on tree volume remains relatively unexplored in the literature. Where it has been studied, primarily in experimental settings, no significant carbon fertilization effect has been found [15,20,21,22]. However, these results may have been driven by their experimental design. Experimental research is generally feasible only across a few years, and this may be problematic for identifying impacts in slow-growing boreal forests where changes are not rapidly manifested. This may be because: (1) elevated CO₂ may precipitate an effect that is likely to be larger than that found in experiments and tree-ring studies when exposure exists for a longer term, and (2) a sudden increase in CO₂ might have a different impact than the impact from a long-term, slow increase in CO₂ [23].

Recently, advances have been made in quantifying the impacts of elevated CO₂ using observational data. For example, using quasi-experimental techniques and U.S. Forest Service Forest Inventory and Analysis (USFS-FIA) data, researchers were able to demonstrate the positive impact of elevated CO₂ in temperate forests [24]. This study builds on that work and thus contributes to the existing literature by allowing us to test these hypotheses utilizing observational, forest inventory data for Alberta, Canada.

We employ a quasi-experimental doubly robust approach that combines propensity-score matching and regression analysis [24,25] to generate results that research has shown are robust even if there is misspecification in the regression model or the matching process [26,27,28,29]. Our approach works by disentangling the direct effect of elevated CO₂ from other drivers like increasing temperatures, changes in precipitation patterns, fires, and insect infestations that also influence forest biomass.

Our results suggest that elevated CO₂ does have an overall positive and significant effect on wood volume in boreal forests. In our study, a 1.0% increase in lifetime exposure to CO₂ leads, on average, to a 1.1% increase in wood volume. These results are significant, as they suggest that these forests may be able to play a larger role in offsetting carbon emissions and in meeting net-zero sustainability targets.

2. Materials and Methods

2.1. Methods

In order to estimate the impact of elevated CO₂ on wood volume, we employed the doubly robust econometric approach [24,28]. This approach first employs propensity-score matching to match trees from the control period (observations before 1990) with trees from the treatment period (observations from 1990 onward) that grew under similar conditions. This balancing of the observations from the two periods helps reduce any bias in inference that may happen due to differences in characteristics between the control and treatment groups to better approximate a randomized controlled trial. Matches were formed based on criteria that included age, lifetime average seasonal temperature, lifetime average seasonal precipitation, elevation, latitude, fire events, and regions (categories used by the Canadian Forest Service to detail the land type upon which the trees grew).

The matching was performed following the procedures outlined in [28] that include estimating a standard logit, predicting a probability (called a propensity score), and then comparing the conditional probabilities of the treatment period trees with the control trees to find matches (We found the nearest matches on the common support within a caliper distance of 0.5) that enable two comparable groups to be formed [30]. Note that matching in our study was performed without replacement. Matching was conducted for softwood observations within two distinct age groups: (1) those aged 1 to 100 years and (2) those above 100 years. This process was then repeated for hardwoods but only for those aged 1 to 100 years, as there were insufficient observations of hardwoods above 100 years of age to enable robust matching.

With balanced datasets created and bias minimized, we then ran regressions using the following generalized form shown in Equation (1):

$$Ln\left(Volum{e}_{it}\right)=c+X_{it}^{\prime }\beta +\alpha Ln\left(Lifetime {CO}_{2t}\right)+{\tau }_t+{ϵ}_{it}$$

(1)

where our outcome variable $Ln\left(Volum{e}_{it}\right)$ is the natural logarithm of the aboveground wood volume of each tree measured in cubic meters. This volume is assumed to be a function of $X_{it}$, which contains controls such as age, the lifetime mean temperature, and lifetime mean precipitation exposures for each season, the regions, the elevation, the latitude, and whether a tree $i$ had experienced a fire event before time $t$. $c$ denotes the constant term and ${ϵ}_{it}$ the error term, which is clustered at the plot level.

The main variable of interest is $Ln\left({Lifetime {CO}_2}_t\right)$, which is the natural logarithm of the cumulative lifetime CO₂ exposure up to time $t.$ To clearly identify the impact of elevated CO₂ requires, in part, isolating the impact of temperature and precipitation, which is relatively easy as these variables vary both temporally and spatially. As elevated CO₂ varies only temporally, we must also take advantage of the fact that trees exist over a longer period than, for example, agricultural crops, which often are planted and harvested within the same year. The benefit of this is derived from the fact that the multi-year life of trees results in two different-aged trees measured at the same time having different lifetime CO₂ exposures. Similarly, two trees of the same age measured at different times would also have different exposures. This variation helps us to isolate the impact of elevated CO₂ apart from other positively trended drivers like improvements in seed technologies or forest management practices. We also employ time fixed effects, using decadal time dummies, to account for other time-varying, episodic factors, such as nitrogen deposition and insect infestations.

Importantly, research has shown that this doubly robust econometric approach of first balancing the data through matching and then running a multivariate regression produces robust results even if there is a misspecification in the regression model or the matching process [26,27,28,29].

2.2. Data

This study uses Canadian Forest Service inventory data, which were collected in Alberta, Canada from 1960 to 2019 and contain information on factors such as each tree’s aboveground measured wood volume in cubic meters, its age, and the environment in which it grew (

Table A1. Definitions of key variables.

Variable	Definition
Volume (m3)	Aboveground wood volume of an observed tree
Ln (Volume)	Logarithmic transformation of the aboveground wood volume of a tree
Ln (Lifetime CO2)	Logarithmic transformation of the sum of yearly atmospheric CO2 exposure over the stand’s lifetime
1/Stand Age	Inverse of the stand age (years)
Lifetime Mean Seasonal Temperature (°C)	Mean seasonal (i.e., spring, summer, autumn, and winter) temperature over the stand’s lifetime
Lifetime Mean Seasonal Precipitation (mm)	Mean seasonal (i.e., spring, summer, autumn, and winter) precipitation over the stand’s lifetime
Natural Region	Dummy variables for location in Forest, Foothills, or Rocky Mountains
Elevation (m)	Distance above sea level a stand is located
Latitude (°)	Latitude of the stand
Fire Events in Lifetime	Dummy variable indicating whether stand experienced fire in its lifetime

). A few observations that were deemed erroneous were excluded (e.g., those that had a large, recorded wood volume but a forest age less than or equal to zero). Data for the main variable of interest, atmospheric CO₂ concentrations, was obtained from the Carbon Dioxide Information Analysis Center at Oak Ridge National Laboratory and the National Oceanic and Atmospheric Administration (NOAA) of the U.S. Department of Commerce [31,32]. To this, data from the government of Alberta was added that recorded whether the trees had experienced fire events from 1931 to 2021 [33].

Finally, historical weather data from 1901 to 2021 [34] was added to control for the impacts of precipitation and temperature over time at each site. This was accomplished by matching plot locations to weather data. ClimateNA (v7.42), a standalone MS Windows software application [34], was used to extract and downscale the gridded (4 × 4 km) monthly climate data to scale-free point locations for the reference period (1961–1990) from PRISM [35] and WorldClim [36]. This was handled through a combination of bilinear interpolation and dynamic local elevational adjustment. The resulting scale-free data was then used as the baseline to downscale the historical climate data for the years between 1901 and 2019. With this process complete, the weather variables were constructed by calculating the mean temperature and precipitation each plot was exposed to over its lifetime. For some observations, this required an extra step because weather data prior to 1901 was unavailable. In these cases, we used the average temperature and precipitation data from 1901 to 1910 as a proxy for the relatively stable climate patterns that existed before 1901.

For the matching process, we selected the year 1990 as the threshold for the categorization into treatment and control groups, as it represented the midpoint of our dataset and because the slope of the atmospheric CO₂ concentration curve steepens after this year. To attain the best matches, matching was conducted separately for hardwoods and softwoods due to the heterogeneity in their average wood volume. To understand the types of trees that were included in each forest group, please see

Table A2. Composition of hardwood and softwood forest groups.

Forest Group	Strata	Description
Hardwoods	Hw	Hardwoods
	HwSw	Hardwoods (leading) interspersed with White Spruce
	HwSb	Hardwoods (leading) interspersed with Black Spruce
Softwoods	P	Pine
	PHw	Pine (leading) interspersed with Hardwoods
	Sb	Black Spruce
	SbHw	Black Spruce (leading) interspersed with Hardwoods
	Sw	White Spruce
	SwHw	White Spruce (leading) interspersed with Hardwoods

. The softwoods were further split into two groups before matching: those aged 1 to 100 years and those aged above 100 years. For hardwoods, only observations aged 1 to 100 years were utilized as there were not sufficient observations of hardwoods above 100 years to conduct robust matching. For each of the three groups, there were sizeable differences in their average characteristics before matching, but once matching was conducted, the differences between the control and treatment groups for all three groups were significantly minimized (

Table 1. Quality of matches of softwoods aged 1 to 100 years.

	Unmatched	Mean
Variable	Matched	Treated	Control	%bias
Age	U	44.38	64.38	−64.3
	M	57.29	62.43	−16.5
Lifetime Mean Winter Temperature	U	−9.61	−10.77	40.4
	M	−10.15	−10.40	8.7
Lifetime Mean Spring Temperature	U	2.39	2.03	35.7
	M	2.21	2.13	8.2
Lifetime Mean Summer Temperature	U	13.03	12.95	9.8
	M	12.95	12.98	−3.8
Lifetime Mean Autumn Temperature	U	2.30	2.34	−4.4
	M	2.39	2.41	−1.7
Lifetime Mean Winter Precipitation	U	79.56	81.40	−6.1
	M	78.89	79.50	−2.0
Lifetime Mean Spring Precipitation	U	121.84	117.02	16.3
	M	120.73	121.94	−4.1
Lifetime Mean Summer Precipitation	U	265.62	257.83	22.2
	M	263.11	262.28	2.3
Lifetime Mean Autumn Precipitation	U	114.13	109.65	22.1
	M	111.30	110.91	2.0
Foothills	U	0.43	0.36	13.2
	M	0.40	0.37	6.2
Rocky Mountains	U	0.06	0.06	1.7
	M	0.05	0.07	−8.9
Elevation	U	1149.70	1104.20	15.5
	M	1143.60	1145.50	−0.7
Latitude	U	53.49	53.76	−15.2
	M	53.47	53.40	3.5
Fire Events or Not in Lifetime	U	0.10	0.01	42.7
	M	0.01	0.02	−4.0

Note: %bias = 100 × (mean^treated − mean^control)/((variance^treated + variance^control)/2)^0.5.

,

Table 2. Quality of matches of softwoods of ages >100 years.

	Unmatched	Mean
Variable	Matched	Treated	Control	%bias
Age	U	136.21	136.46	−0.9
	M	137.15	137.56	−1.3
Lifetime Mean Winter Temperature	U	−12.18	−13.44	32.5
	M	−12.49	−12.74	6.3
Lifetime Mean Spring Temperature	U	1.54	1.10	33.2
	M	1.44	1.34	7.7
Lifetime Mean Summer Temperature	U	12.91	12.85	8.1
	M	12.87	12.85	2.6
Lifetime Mean Autumn Temperature	U	1.85	1.41	33.0
	M	1.73	1.64	7.0
Lifetime Mean Winter Precipitation	U	79.19	79.07	0.7
	M	79.18	79.23	−0.3
Lifetime Mean Spring Precipitation	U	111.89	105.78	23.5
	M	108.92	108.68	0.4
Lifetime Mean Summer Precipitation	U	251.47	246.44	9.7
	M	250.08	249.01	2.1
Lifetime Mean Autumn Precipitation	U	105.37	101.86	19.3
	M	104.19	103.66	2.9
Foothills	U	0.28	0.26	4.7
	M	0.30	0.29	2.3
Rocky Mountains	U	0.07	0.06	7.3
	M	0.06	0.07	−1.7
Elevation	U	1036.70	973.93	18.3
	M	1012.30	1003.90	2.5
Latitude	U	54.47	55.15	−33.7
	M	54.73	54.81	−3.8
Fire Events or Not in Lifetime	U	0.08	0.04	17.2
	M	0.06	0.05	4.7

Note: %bias = 100 × (mean^treated − mean^control)/((variance^treated + variance^control)/2)^0.5.

and

Table 3. Quality of matches of hardwoods aged 1 to 100 years.

	Unmatched	Mean
Variable	Matched	Treated	Control	%bias
Age	U	34.17	55.05	−67.3
	M	43.34	48.04	−15.2
Lifetime Mean Winter Temperature	U	−13.97	−13.24	−22.7
	M	−13.14	−13.31	5.4
Lifetime Mean Spring Temperature	U	2.34	2.19	14.4
	M	2.45	2.32	11.9
Lifetime Mean Summer Temperature	U	14.51	13.88	84.0
	M	14.14	14.07	9.9
Lifetime Mean Autumn Temperature	U	1.45	1.92	−35.3
	M	1.83	1.80	2.1
Lifetime Mean Winter Precipitation	U	67.43	76.56	−44.0
	M	71.44	71.67	−1.1
Lifetime Mean Spring Precipitation	U	90.31	96.57	−31.2
	M	94.28	94.01	1.3
Lifetime Mean Summer Precipitation	U	234.07	232.40	5.2
	M	239.36	238.39	3.0
Lifetime Mean Autumn Precipitation	U	96.12	98.38	−17.4
	M	97.49	97.07	3.2
Foothills	U	0.00	0.00	5.3
	M	0.00	0.00	0.0
Rocky Mountains	U	0.00	0.01	−15.2
	M	0.00	0.00	0.0
Elevation	U	683.33	773.32	−40.8
	M	755.09	759.51	−2.0
Latitude	U	55.79	55.22	33.5
	M	55.12	55.25	−2.3
Fire Events or Not in Lifetime	U	0.18	0.01	62.6
	M	0.00	0.01	−3.9

Note: %bias = 100 × (mean^treated − mean^control)/((variance^treated + variance^control)/2)^0.5.

).

3. Results

Having minimized bias across the control and treatment groups through matching, we next attempted to assess the average impact of elevated CO₂ on wood volume in boreal forests. To do this, all three sets of matched data were first aggregated into a single group. Then, regression analysis was conducted. The only difference relative to the general form detailed in Equation (1) was that a dummy variable was added to account for the heterogeneity in average volume between hardwood and softwood trees.

Using this approach, we estimated the impact of elevated CO₂ on tree volume, and our results show that elevated CO₂ has had a significant (p < 0.01), positive effect on tree volume. On average, across all the forest groups in our study, every 1.0% increase in lifetime exposure to CO₂ leads to a 1.122% increase in wood volume (Column 1 of

Table 4. Average impact of elevated CO₂ on the wood volume in boreal forests.

	Natural Logarithm of Wood Volume
	All Ages	Ages 1–100 Years
Natural log of lifetime CO2	1.122 ***	1.723 ***
99% Confidence Interval	(0.812, 1.431)	(1.261, 2.185)
Adjusted R2	0.877	0.883
Observations	4870	2532

Note: Matching was first conducted to select similar trees between control (pre-1990) and treatment (≥1990) groups in terms of their age, region, elevation, latitude, lifetime seasonal weather patterns, fire events, and natural regions. Matching was conducted for softwoods aged 1 to 100 years, softwoods aged above 100 years, and hardwoods aged 1 to 100 years. There were not sufficient hardwoods above 100 years of age to conduct matching. For the all-ages result, all three matched groups were aggregated into a single group, and regression analysis was performed. The ages 1–100 years result is derived from a regression run on softwoods aged 1 to 100 years. For simplicity, all covariates other than the treatment variable are omitted from this table, but they are presented in Table A3 and Table A4. *** p < 0.01.

; See

Table A3. Average impact of elevated CO₂ on the wood volume in boreal forests of hardwood and softwood stands of all ages.

	Natural Logarithm of Wood Volume
	(1)	(2)	(3)
Ln(Lifetime CO2)	1.122 ***	1.289 ***	1.376 ***
	(0.12)	(0.13)	(0.124)
1/Age	−39.13 ***	−34.78 ***	−33.13 ***
	(2.777)	(3.096)	(3.352)
Lifetime Mean Spring Temperature (°C)	−0.545 **	−0.163	−0.0204
	(0.247)	(0.261)	(0.17)
Lifetime Mean Spring Temperature squared	−0.242 ***	0.0263
	(0.0531)	(0.0414)
Lifetime Mean Spring Temperature cubed	0.0820 ***
	(0.0123)
Lifetime Mean Summer Temperature (°C)	49.08 ***	−0.955	0.501 ***
	(10.4)	(0.913)	(0.119)
Lifetime Mean Summer Temperature squared	−3.779 ***	0.0573 *
	(0.813)	(0.0336)
Lifetime Mean Summer Temperature cubed	0.0975 ***
	(0.021)
Lifetime Mean Autumn Temperature (°C)	−0.430 **	−0.184	0.016
	(0.182)	(0.226)	(0.151)
Lifetime Mean Autumn Temperature squared	0.337 ***	0.0673
	(0.0723)	(0.0501)
Lifetime Mean Autumn Temperature cubed	−0.0457 ***
	(0.0161)
Lifetime Mean Winter Temperature (°C)	0.343	−0.109	0.0863 *
	(0.449)	(0.108)	(0.0498)
Lifetime Mean Winter Temperature squared	0.0407	−0.00824 *
	(0.0328)	(0.00492)
Lifetime Mean Winter Temperature cubed	0.00168 **
	(0.000784)
Lifetime Mean Spring Precipitation (mm)	0.118 ***	0.0553 ***	0.0112 ***
	(0.0444)	(0.015)	(0.00305)
Lifetime Mean Spring Precipitation squared	−0.000650 **	−0.000165 ***
	(0.000268)	(0.0000468)
Lifetime Mean Spring Precipitation cubed	0.00000119 **
	(0.000000522)
Lifetime Mean Summer Precipitation (mm)	−0.184 ***	0.00749	−0.00575 ***
	(0.0474)	(0.00683)	(0.00138)
Lifetime Mean Summer Precipitation squared	0.000713 ***	−0.0000265 **
	(0.00018)	(0.0000127)
Lifetime Mean Summer Precipitation cubed	−0.000000920 ***
	(0.000000222)
Lifetime Mean Autumn Precipitation (mm)	−0.0374	−0.0146	−0.00183
	(0.0683)	(0.0141)	(0.00344)
Lifetime Mean Autumn Precipitation squared	0.000272	0.0000489
	(0.000497)	(0.000052)
Lifetime Mean Autumn Precipitation cubed	−0.000000625
	(0.00000115)
Lifetime Mean Winter Precipitation (mm)	−0.0692 ***	−0.0323 ***	−0.00917 ***
	(0.0266)	(0.00862)	(0.00212)
Lifetime Mean Winter Precipitation squared	0.000474 **	0.000121 ***
	(0.000219)	(0.0000354)
Lifetime Mean Winter Precipitation cubed	−0.00000104 **
	(0.000000523)
Foothills	0.115	0.0395	0.0363
	(0.0778)	(0.0737)	(0.0727)
Rocky Mountains	0.184	−0.00308	0.073
	(0.148)	(0.128)	(0.106)
Elevation	−0.000282	0.000881 *	0.000608
	(0.000515)	(0.000525)	(0.000415)
Latitude	0.169 *	0.188 **	0.063
	(0.0879)	(0.0822)	(0.0624)
Fire Events or Not in Lifetime	−0.0557	−0.0967	−0.154 **
	(0.0754)	(0.0739)	(0.0723)
Hardwood	0.452 ***	0.408 ***	0.473 ***
	(0.0904)	(0.092)	(0.0863)
Constant	−212.0 ***	−17.05 **	−16.61 ***
	(43.03)	(7.278)	(4.825)
Adjusted R2	0.877	0.866	0.863
Decadal Fixed Effects	Yes
Observations	4870

Note: Softwoods include ages 1 to 200 years. Hardwoods include only ages 1 to 100 years due to insufficient observations above that threshold. Standard errors are presented in parentheses. * p < 0.10, ** p < 0.05, *** p < 0.01.

for the full regression results). This result is important as it contradicts the findings of many experimental studies, which argued that elevated CO₂ has no impact on boreal forests [21,22]. Moreover, this result supports the findings generally found for temperate forests (e.g., [24]). Thus, this result is suggestive of a potentially significant role for boreal forests in meeting sustainability goals tied to net-zero emissions.

Next, to understand if the impact of elevated CO₂ is stronger on younger stands, a regression of the form detailed in Equation (1) was run while limiting the observations only to those of ages 1 to 100 years. This result is also significant (p < 0.01), and the impact is even larger than the all-ages result, with every 1.0% increase leading to a 1.723% increase in wood volume (Column 2 of Table 4; See

Table A4. Average impact of elevated CO₂ on the wood volume in boreal forests of hardwood and softwood stands of ages 1 to 100 years.

	Natural Logarithm of Wood Volume
	(1)	(2)	(3)
Natural logarithm of lifetime CO2	1.723 ***	2.107 ***	2.214 ***
	(0.179)	(0.187)	(0.152)
1/Age	−33.26 ***	−25.59 ***	−22.29 ***
	(3.005)	(3.076)	(3.35)
Lifetime Mean Spring Temperature (°C)	−0.33	−0.0598	0.162
	(0.286)	(0.336)	(0.242)
Lifetime Mean Spring Temperature squared	−0.291 ***	0.0988 **
	(0.0761)	(0.0459)
Lifetime Mean Spring Temperature cubed	0.105 ***
	(0.0167)
Lifetime Mean Summer Temperature (°C)	98.47 ***	−4.114 ***	0.315 *
	(19.94)	(1.509)	(0.183)
Lifetime Mean Summer Temperature squared	−7.518 ***	0.165 ***
	(1.508)	(0.0541)
Lifetime Mean Summer Temperature cubed	0.190 ***
	(0.0378)
Lifetime Mean Autumn Temperature (°C)	−0.927 ***	−0.345	−0.233
	(0.228)	(0.264)	(0.208)
Lifetime Mean Autumn Temperature squared	0.324 ***	0.107 *
	(0.0718)	(0.0582)
Lifetime Mean Autumn Temperature cubed	−0.0187
	(0.0182)
Lifetime Mean Winter Temperature (°C)	1.002	−0.25	0.0948
	(0.663)	(0.16)	(0.0902)
Lifetime Mean Winter Temperature squared	0.108 **	−0.0115
	(0.0506)	(0.00704)
Lifetime Mean Winter Temperature cubed	0.00359 ***
	(0.00123)
Lifetime Mean Spring Precipitation (mm)	0.0209	0.0481 **	0.0201 ***
	(0.0513)	(0.02)	(0.00386)
Lifetime Mean Spring Precipitation squared	0.00000978	−0.000135 **
	(0.000308)	(0.0000649)
Lifetime Mean Spring Precipitation cubed	−2.65 × 10−7
	(0.000000599)
Lifetime Mean Summer Precipitation (mm)	−0.0609	−0.0156	−0.0017
	(0.0739)	(0.0117)	(0.00188)
Lifetime Mean Summer Precipitation squared	0.000249	0.0000243
	(0.000285)	(0.0000224)
Lifetime Mean Summer Precipitation cubed	−3.32 × 10−7
	(0.00000036)
Lifetime Mean Autumn Precipitation (mm)	0.0874	0.0217	−0.00359
	(0.09)	(0.0192)	(0.00483)
Lifetime Mean Autumn Precipitation squared	−0.000672	−0.0000811
	(0.000637)	(0.0000713)
Lifetime Mean Autumn Precipitation cubed	0.00000165
	(0.00000142)
Lifetime Mean Winter Precipitation (mm)	−0.0876 ***	−0.0484 ***	−0.00987 ***
	(0.0336)	(0.0109)	(0.00277)
Lifetime Mean Winter Precipitation squared	0.000569 **	0.000176 ***
	(0.000276)	(0.0000429)
Lifetime Mean Winter Precipitation cubed	−0.00000108
	(0.000000667)
Foothills	0.375 ***	0.288 ***	0.204 *
	(0.11)	(0.111)	(0.11)
Rocky Mountains	0.651 **	0.4	0.419
	(0.264)	(0.289)	(0.259)
Elevation	−0.00253 ***	0.0000077	−0.000955
	(0.000867)	(0.000829)	(0.000703)
Latitude	−0.0266	0.161	0.00951
	(0.113)	(0.108)	(0.0902)
Fire Events or Not in Lifetime	−0.167	−0.0462	−0.112
	(0.16)	(0.165)	(0.16)
Hardwood	0.407 ***	0.396 ***	0.396 ***
	(0.109)	(0.111)	(0.112)
Constant	−427.2 ***	−0.133	−19.55 ***
	(84.97)	(12.36)	(6.638)
Adjusted R2	0.883	0.873	0.867
Decadal Fixed Effects	Yes
Observations	2532

Standard errors are presented in parentheses. * p < 0.10, ** p < 0.05, *** p < 0.01.

for the full regression results). This suggests that the impact of elevated CO₂ is larger for younger stands than for those over 100 years of age. Another interesting finding is that this result for boreal forests is larger than the average effect estimated in [24], which is 1.15%. The differences in results may be driven by the composition of trees in the respective studies or differences between temperate and boreal forests.

Next, we turned to a differential analysis of softwoods and hardwoods across different age groups using the Equation (1) regression format. We began to examine the impact across all ages for softwoods. The process was then repeated for two subgroups of softwoods: (1) those aged 1 to 100 years and (2) those aged >100 years. For hardwoods, only one regression was run, as all matched observations were between 1 and 100 years of age. The results thus are the same for hardwoods for both the all-ages column and the ages-1–100-years column.

Our results show a difference in the effect of elevated CO₂ on both softwoods and hardwoods and across age groups (

Table 5. Average impact of elevated CO₂ on the wood volume of hardwoods and softwoods by age.

	Natural Logarithm of Wood Volume
	All Ages	Ages 1–100 Years	Ages >100 Years
A. Softwoods
Natural log of lifetime CO2	0.432 **	1.113 ***	0.722 *
Confidence Intervals 1	(0.048, 0.816)	(0.421, 1.845)	(0.016, 1.428)
Adjusted R2	0.891	0.902	0.224
Observations	4126	1788	2338
B. Hardwoods
Natural log of lifetime CO2	2.191 ***	2.191 ***
99% Confidence Interval	(1.505, 2.877)	(1.505, 2.877)
Adjusted R2	0.897	0.897
Observations	744	744

Note: Matching was first conducted to select similar trees between control (pre-1990) and treatment (≥1990) groups in terms of their age, region, elevation, latitude, lifetime seasonal weather patterns, fire events, and natural regions. Matching was conducted for softwoods aged 1 to 100 years, softwoods aged above 100 years, and hardwoods aged 1 to 100 years. There were not sufficient hardwoods above 100 years of age to conduct matching. With matches created, regression analysis was performed. For simplicity, all covariates other than the treatment variable are omitted from this table, but they are presented in Table A5, Table A6, Table A7 and Table A8. ¹ Softwood confidence intervals are: 95% for all ages, 99% for ages 1–100 years, and 90% for ages >100 years. * p < 0.10, ** p < 0.05, *** p < 0.01.

). For softwoods (Panel A in Table 5; See

Table A5. Average impact of elevated CO₂ on the wood volume in boreal forests of softwood stands of all ages.

	Natural Logarithm of Wood Volume
	(1)	(2)	(3)
Natural logarithm of lifetime CO2	0.432 **	0.527 **	0.469 **
	(0.196)	(0.209)	(0.224)
1/Age	−60.96 ***	−60.31 ***	−58.56 ***
	(5.389)	(5.876)	(6.119)
Lifetime Mean Spring Temperature (°C)	−1.086 ***	−0.806 ***	0.142
	(0.188)	(0.192)	(0.148)
Lifetime Mean Spring Temperature squared	−0.118 **	0.210 ***
	(0.0465)	(0.0341)
Lifetime Mean Spring Temperature cubed	0.0853 ***
	(0.0127)
Lifetime Mean Summer Temperature (°C)	15.64 *	2.455 ***	0.179
	(9.372)	(0.695)	(0.131)
Lifetime Mean Summer Temperature squared	−1.082	−0.0760 ***
	(0.716)	(0.0246)
Lifetime Mean Summer Temperature cubed	0.0257
	(0.0183)
Lifetime Mean Autumn Temperature (°C)	0.452 **	0.677 ***	−0.152
	(0.203)	(0.193)	(0.139)
Lifetime Mean Autumn Temperature squared	0.0757	−0.153 ***
	(0.0675)	(0.0341)
Lifetime Mean Autumn Temperature cubed	−0.0305 *
	(0.0177)
Lifetime Mean Winter Temperature (°C)	−0.352	−0.0959	0.128 ***
	(0.518)	(0.115)	(0.0436)
Lifetime Mean Winter Temperature squared	−0.0206	−0.00703
	(0.0373)	(0.00539)
Lifetime Mean Winter Temperature cubed	−0.0000104
	(0.000849)
Lifetime Mean Spring Precipitation (mm)	0.180 ***	0.0696 ***	−0.00543 ***
	(0.0517)	(0.0167)	(0.00158)
Lifetime Mean Spring Precipitation squared	−0.000970 ***	−0.000213 ***
	(0.000311)	(0.0000543)
Lifetime Mean Spring Precipitation cubed	0.00000166 ***
	(0.000000597)
Lifetime Mean Summer Precipitation (mm)	−0.105 **	−0.00586	0.0011
	(0.048)	(0.00595)	(0.00363)
Lifetime Mean Summer Precipitation squared	0.000383 **	−0.00000317
	(0.00018)	(0.0000108)
Lifetime Mean Summer Precipitation cubed	−0.000000481 **
	(0.000000215)
Lifetime Mean Autumn Precipitation (mm)	−0.0767	−0.0289 **	0.00455
	(0.0643)	(0.0133)	(0.00279)
Lifetime Mean Autumn Precipitation squared	0.000529	0.000132 ***
	(0.000487)	(0.0000511)
Lifetime Mean Autumn Precipitation cubed	−0.00000106
	(0.00000114)
Lifetime Mean Winter Precipitation (mm)	−0.105 ***	−0.0455 ***	−0.00794 ***
	(0.0265)	(0.00808)	(0.00234)
Lifetime Mean Winter Precipitation squared	0.000693 ***	0.000170 ***
	(0.000223)	(0.0000346)
Lifetime Mean Winter Precipitation cubed	−0.00000138 ***
	(0.000000521)
Foothills	0.0066	−0.0638	−0.156 **
	(0.0667)	(0.0625)	(0.0625)
Rocky Mountains	−0.149	−0.277 ***	−0.0924
	(0.0991)	(0.0973)	(0.0923)
Elevation	0.000458	0.00111 **	0.000089
	(0.000449)	(0.000442)	(0.0004)
Latitude	0.346 ***	0.257 ***	0.0976 *
	(0.0843)	(0.0672)	(0.0557)
Fire Events or Not in Lifetime	−0.0466	−0.0988	−0.105 *
	(0.064)	(0.0645)	(0.062)
Constant	−86.83 **	−32.65 ***	−3.339
	(40.15)	(7.189)	(4.604)
Adjusted R2	0.891	0.883	0.876
Decadal Fixed Effects	Yes
Observations	4126

Standard errors are presented in parentheses. * p < 0.10, ** p < 0.05, *** p < 0.01.

,

Table A6. Average impact of elevated CO₂ on the wood volume in boreal forests of softwood stands of ages 1 to 100 years.

	Natural Logarithm of Wood Volume
	(1)	(2)	(3)
Natural logarithm of lifetime CO2	1.113 ***	1.417 ***	1.417 ***
	(0.276)	(0.315)	(0.338)
1/Age	−48.99 ***	−45.12 ***	−43.90 ***
	(5.413)	(5.792)	(6.529)
Lifetime Mean Spring Temperature (°C)	−1.017 ***	−1.277 ***	0.0244
	(0.29)	(0.296)	(0.248)
Lifetime Mean Spring Temperature squared	−0.225 **	0.357 ***
	(0.113)	(0.0484)
Lifetime Mean Spring Temperature cubed	0.116 ***
	(0.0228)
Lifetime Mean Summer Temperature (°C)	71.00 ***	2.603 *	−0.116
	(20.88)	(1.368)	(0.229)
Lifetime Mean Summer Temperature squared	−5.328 ***	−0.0900 *
	(1.577)	(0.0512)
Lifetime Mean Summer Temperature cubed	0.133 ***
	(0.0395)
Lifetime Mean Autumn Temperature (°C)	−0.00468	1.156 ***	0.0689
	(0.407)	(0.397)	(0.231)
Lifetime Mean Autumn Temperature squared	0.132	−0.166 ***
	(0.129)	(0.0632)
Lifetime Mean Autumn Temperature cubed	−0.00457
	(0.0287)
Lifetime Mean Winter Temperature (°C)	−0.347	−0.283	0.148
	(0.753)	(0.212)	(0.102)
Lifetime Mean Winter Temperature squared	−0.00696	−0.0112
	(0.0568)	(0.00897)
Lifetime Mean Winter Temperature cubed	0.000671
	(0.00136)
Lifetime Mean Spring Precipitation (mm)	0.144 **	0.0757 ***	−0.00443 **
	(0.0669)	(0.0204)	(0.00204)
Lifetime Mean Spring Precipitation squared	−0.000734 *	−0.000240 ***
	(0.000399)	(0.0000634)
Lifetime Mean Spring Precipitation cubed	0.00000112
	(0.000000783)
Lifetime Mean Summer Precipitation (mm)	−0.11	−0.0384 ***	0.00217
	(0.0735)	(0.00993)	(0.00523)
Lifetime Mean Summer Precipitation squared	0.000373	0.0000624 ***
	(0.00028)	(0.0000181)
Lifetime Mean Summer Precipitation cubed	−0.000000426
	(0.000000346)
Lifetime Mean Autumn Precipitation (mm)	−0.0336	−0.0149	0.0124 ***
	(0.0813)	(0.0179)	(0.00444)
Lifetime Mean Autumn Precipitation squared	0.000237	0.000105 *
	(0.000592)	(0.0000633)
Lifetime Mean Autumn Precipitation cubed	−0.000000351
	(0.00000132)
Lifetime Mean Winter Precipitation (mm)	−0.0919 **	−0.0657 ***	−0.0105 ***
	(0.0385)	(0.0107)	(0.00298)
Lifetime Mean Winter Precipitation squared	0.000507	0.000242 ***
	(0.000331)	(0.0000417)
Lifetime Mean Winter Precipitation cubed	−0.000000803
	(0.000000811)
Foothills	0.158	0.12	−0.0752
	(0.103)	(0.102)	(0.108)
Rocky Mountains	0.249	0.0294	0.222
	(0.245)	(0.223)	(0.248)
Elevation	−0.0013	0.000995	−0.00145 *
	(0.000943)	(0.000739)	(0.000783)
Latitude	0.273 **	0.361 ***	0.0615
	(0.119)	(0.0969)	(0.0899)
Fire Events or Not in Lifetime	0.053	0.0542	−0.0129
	(0.169)	(0.159)	(0.163)
Constant	−325.1 ***	−45.45 ***	−6.511
	(89.75)	(11.37)	(7.108)
Adjusted R2	0.902	0.897	0.885
Decadal Fixed Effects	Yes
Observations	1788

Standard errors are presented in parentheses. * p < 0.10, ** p < 0.05, *** p < 0.01.

and

Table A7. Average impact of elevated CO₂ on the wood volume in boreal forests of softwood stands of ages above years.

	Natural Logarithm of Wood Volume
	(1)	(2)	(3)
Natural logarithm of lifetime CO2	0.722 *	0.537	−0.479
	(0.429)	(0.423)	(0.42)
1/Age	113.1 **	90.41	−36.39
	(55.48)	(55.22)	(55.19)
Lifetime Mean Spring Temperature (°C)	−0.136	0.0158	−0.123 *
	(0.167)	(0.127)	(0.0648)
Lifetime Mean Spring Temperature squared	0.0368	0.0962 ***
	(0.0419)	(0.0232)
Lifetime Mean Spring Temperature cubed	0.0229 *
	(0.0127)
Lifetime Mean Summer Temperature (°C)	−2.309	−0.643	−0.0518
	(11.77)	(0.512)	(0.0728)
Lifetime Mean Summer Temperature squared	0.188	0.0274
	(0.942)	(0.0204)
Lifetime Mean Summer Temperature cubed	−0.00487
	(0.025)
Lifetime Mean Autumn Temperature (°C)	0.406 ***	0.308 ***	0.137
	(0.098)	(0.0777)	(0.0921)
Lifetime Mean Autumn Temperature squared	−0.0929 *	−0.174 ***
	(0.0499)	(0.0227)
Lifetime Mean Autumn Temperature cubed	−0.0238
	(0.0145)
Lifetime Mean Winter Temperature (°C)	0.365	0.289 ***	0.0302
	(0.52)	(0.0725)	(0.028)
Lifetime Mean Winter Temperature squared	0.0209	0.0109 ***
	(0.0371)	(0.00326)
Lifetime Mean Winter Temperature cubed	0.000363
	(0.000867)
Lifetime Mean Spring Precipitation (mm)	0.0778 **	0.0456 ***	−0.00280 ***
	(0.0307)	(0.0103)	(0.000672)
Lifetime Mean Spring Precipitation squared	−0.000392 **	−0.000139 ***
	(0.000189)	(0.0000346)
Lifetime Mean Spring Precipitation cubed	0.000000619 *
	(0.000000372)
Lifetime Mean Summer Precipitation (mm)	−0.0783 *	0.0147 ***	−0.00389 **
	(0.0405)	(0.00533)	(0.00172)
Lifetime Mean Summer Precipitation squared	0.000302 **	−0.0000357 ***
	(0.000144)	(0.00000941)
Lifetime Mean Summer Precipitation cubed	−0.000000399 **
	(0.000000168)
Lifetime Mean Autumn Precipitation (mm)	0.0966	−0.0202	−0.00106
	(0.0966)	(0.0123)	(0.00195)
Lifetime Mean Autumn Precipitation squared	−0.000881	0.0000412
	(0.000787)	(0.0000494)
Lifetime Mean Autumn Precipitation cubed	0.00000233
	(0.00000211)
Lifetime Mean Winter Precipitation (mm)	−0.127 ***	−0.0302 **	0.00285
	(0.0397)	(0.0145)	(0.00193)
Lifetime Mean Winter Precipitation squared	0.00120 ***	0.000175 **
	(0.00038)	(0.000078)
Lifetime Mean Winter Precipitation cubed	−0.00000338 ***
	(0.00000115)
Foothills	−0.127 **	−0.117 **	−0.0880 *
	(0.0537)	(0.0504)	(0.0511)
Rocky Mountains	−0.254 ***	−0.250 ***	−0.109
	(0.0839)	(0.0819)	(0.0906)
Elevation	0.000671	0.000528	0.0000857
	(0.000466)	(0.000396)	(0.000263)
Latitude	0.166 ***	0.139 ***	0.00826
	(0.0594)	(0.0531)	(0.0343)
Fire Events or Not in Lifetime	−0.0759	−0.0849	−0.0624
	(0.0556)	(0.054)	(0.0583)
Constant	1.687	−4.999	12.68 **
	(48.44)	(6.395)	(5.228)
Adjusted R2	0.224	0.213	0.148
Decadal Fixed Effects	Yes
Observations	2338

Standard errors are presented in parentheses. * p < 0.10, ** p < 0.05, *** p < 0.01.

for the full regression results), we observe for all ages a positive and statistically significant (p < 0.05) impact of lifetime CO₂ exposure on wood volume, with every 1.0% increase in lifetime exposure to CO₂ engendering a 0.432% increase in wood volume. We also find there is a stronger (1.113%) and more significant effect (p < 0.01) on trees up to 100 years of age than those over 100 years of age (0.722%; p < 0.10).

For hardwoods as well (Panel B in Table 5; See

Table A8. Average impact of elevated CO₂ on the wood volume in boreal forests of hardwood stands of ages 1 to 100 years.

	Natural Logarithm of Wood Volume
	(1)	(2)	(3)
Natural logarithm of lifetime CO2	2.191 ***	2.412 ***	2.483 ***
	(0.266)	(0.306)	(0.293)
1/Age	−25.54 ***	−16.72 ***	−14.60 ***
	(5.327)	(4.299)	(3.732)
Lifetime Mean Spring Temperature (°C)	0.148	0.0353	−0.587
	(0.349)	(0.408)	(0.415)
Lifetime Mean Spring Temperature squared	−0.219 **	0.0193
	(0.103)	(0.0669)
Lifetime Mean Spring Temperature cubed	0.0756 ***
	(0.0244)
Lifetime Mean Summer Temperature (°C)	300.0 **	−9.994 **	0.71
	(123.9)	(4.779)	(0.583)
Lifetime Mean Summer Temperature squared	−21.61 **	0.371 **
	(8.737)	(0.175)
Lifetime Mean Summer Temperature cubed	0.517 **
	(0.205)
Lifetime Mean Autumn Temperature (°C)	−0.760 *	−0.454	−0.393
	(0.402)	(0.392)	(0.336)
Lifetime Mean Autumn Temperature squared	0.132 *	−0.0285
	(0.0753)	(0.0853)
Lifetime Mean Autumn Temperature cubed	−0.0264
	(0.0256)
Lifetime Mean Winter Temperature (°C)	0.422	0.347	0.247
	(1.878)	(0.527)	(0.177)
Lifetime Mean Winter Temperature squared	0.0563	0.0117
	(0.126)	(0.0183)
Lifetime Mean Winter Temperature cubed	0.00166
	(0.00277)
Lifetime Mean Spring Precipitation (mm)	−0.131	−0.0146	0.0132 ***
	(0.152)	(0.0417)	(0.00488)
Lifetime Mean Spring Precipitation squared	0.000908	−0.0000687
	(0.00112)	(0.000145)
Lifetime Mean Spring Precipitation cubed	−0.00000209
	(0.00000252)
Lifetime Mean Summer Precipitation (mm)	0.265	−0.0383	0.0119
	(0.296)	(0.0457)	(0.0115)
Lifetime Mean Summer Precipitation squared	−0.0011	0.000107
	(0.00127)	(0.0000947)
Lifetime Mean Summer Precipitation cubed	0.00000157
	(0.0000018)
Lifetime Mean Autumn Precipitation (mm)	−0.342	0.0017	−0.018
	(0.356)	(0.06)	(0.016)
Lifetime Mean Autumn Precipitation squared	0.00315	0.000055
	(0.00295)	(0.000244)
Lifetime Mean Autumn Precipitation cubed	−0.00000883
	(0.00000782)
Lifetime Mean Winter Precipitation (mm)	−0.106	0.0107	0.0146
	(0.0757)	(0.0244)	(0.0117)
Lifetime Mean Winter Precipitation squared	0.000881	0.0000503
	(0.000695)	(0.0000952)
Lifetime Mean Winter Precipitation cubed	−0.00000195
	(0.00000189)
Elevation	−0.00417 **	−0.00215	−0.00252
	(0.00194)	(0.00216)	(0.00226)
Latitude	−0.678 ***	−0.498 ***	−0.363 **
	(0.194)	(0.181)	(0.165)
Fire Events or Not in Lifetime	−0.335 *	−0.365 **	−0.484 **
	(0.188)	(0.181)	(0.23)
Constant	−1358.8 **	84.60 **	−5.224
	(583.6)	(40.3)	(17)
Adjusted R2	0.897	0.889	0.885
Decadal Fixed Effects	Yes
Observations	744

Standard errors are presented in parentheses. * p < 0.10, ** p < 0.05, *** p < 0.01.

for the full regression results), increased exposure to elevated CO₂ has a very positive and statistically significant (p < 0.01) effect on tree volume, with a 1% increase in lifetime CO₂ exposure leading to a 2.191% increase in wood volume. As the results for both softwoods and hardwoods are especially robust for trees up to age 100 years, this is suggestive of a potential role for planting and forest management in expanding sequestration outcomes.

4. Discussion and Conclusions

This study develops empirical estimates of the impact of elevated CO₂ on the wood volume of boreal trees. Because of the difficulties in finding long-term datasets that enable the use of quasi-experimental techniques such as those deployed here, there has been limited work using observational data like that developed in [24]. Our study thus contributes to this research area through the use of observational, forest inventory data for Alberta, Canada, and doubly robust econometric techniques that combine propensity score matching with multivariate regression to better identify the effect of elevated CO₂ on tree volume in boreal forests.

Our results suggest that elevated CO₂ has had a positive impact on wood volume in boreal forests, with variations observed across tree species and age groups. To our knowledge, this is the first study to establish this relationship for boreal forests. These findings help improve understanding of the effects of climate change on forest ecosystems and underscore the importance of considering species and age-specific responses when assessing the impact of environmental factors on forest dynamics.

Moreover, by increasing wood volume—and presumably carbon storage—elevated CO₂ has important implications for sustainability in boreal forest systems. As policymakers increasingly attempt to distinguish between carbon accumulation resulting from direct human actions (e.g., due to forest management activities) and those resulting from passive gains (e.g., due to carbon fertilization) when setting net-zero sustainability targets, being able to identify the effects of carbon fertilization independently from other factors—as accomplished in this study—is indispensable.

Our results may also help address the considerable discrepancy among carbon flux estimates for Canada in the literature. Inversion models and other satellite approaches [2,37] find that Canada’s forests have been a net sink over the last couple of decades. These results vary from on-the-ground inventory numbers [1] and the official Canadian carbon inventory, which both estimate that these forests have become a net source of carbon in Canada. Because they are reliant on models instead of observations or widespread plot data, the latter numbers may reflect existing experimental results, which have suggested that there is little effect of elevated CO₂ in boreal forests [21,22]. Research suggests that this is because long-term exposure to elevated CO₂ seems to engender a larger effect than that precipitated by a sudden, short-term increase in CO_2, as is common in experimental research [23]. Our results, which examine a much longer period, contest the experimental findings and provide some support for the inventory numbers of the satellite-driven models.

The accuracy of those models, though, is still uncertain as even if elevated CO₂ is increasing wood volume, other factors, such as infestations by the Mountain Pine Beetle and increases in forest fire activity, may be reducing it, potentially at a rate greater than the increases from carbon fertilization. This has increased concerns about the sustainability of boreal forests. Furthermore, while fire activity is separately influenced by climate variation (e.g., temperature and precipitation changes), increased wood volume due to carbon fertilization may also play a role in fire events [8] and contribute to the increasing fire trends in the region [6].

Another important finding from our study is that the effects of elevated CO₂ appear strongest in younger stands, in this case, in stands under the age of 100 years. To make the most of the forest carbon sink, Canadian policymakers may thus wish to focus on accelerating regeneration in areas that have been impacted by insects and/or forest fires. That, however, does not mean that older-age forests should be devalued. Despite the generally stronger effect in younger stands, results for softwoods suggest that carbon fertilization continues to increase wood volume in stands older than 100 years, a result that highlights the value of protecting old-growth forests, both from harvesting timber and fires.

While our results are limited to Alberta, the region does contain a wide range of typical boreal forest trees. Still, it is possible that these results are limited only to this region or to those with similar forests and climatic variation. Future research should seek to encompass more of the boreal forest, both in Canada and in other regions like the Nordic nations. Research should also seek to quantify the relative impact of each of the major drivers of wood volume change (e.g., temperature, precipitation, disturbances, age, and forest management). Another avenue for future study could involve investigating the impact of prioritizing the conservation of specific boreal forest types on timber markets, land-use changes, and carbon emissions.