DRIVERS OF BOREAL FOREST GROWTH 1
Taxonomy, together with ontogeny and growing conditions, drives needleleaf
species’ sensitivity to climate in boreal North America
Running head: DRIVERS OF BOREAL FOREST GROWTH
Marchand, W.1, 2, Girardin, M.P. 1, 2, Hartmann, H.3, Gauthier, S.1, 2, Bergeron, Y.2
1. Natural Resources Canada, Canadian Forest Service, Laurentian Forestry Centre,
1055 du P.E.P.S, P.O. Box 10380, Stn. Sainte-Foy, Québec, QC, G1V 4C7,
2. Centre d’étude de la forêt, Université du Québec à Montréal, C.P. 8888, succ.
Centre-ville, Montréal, QC, H3C 3P8, Canada; and Forest Research Institute,
Université du Québec en Abitibi-Témiscamingue, 445 boul. de l’Université,
Rouyn Noranda , QC, Canada, J9X 5E4
3. Max-Planck Institute for Biogeochemistry, Department of Biogeochemical
Processes, Hans-Knöll Str. 10, 07745 Jena, Germany
Corresponding author: William Marchand. Email: firstname.lastname@example.org;
Tel.: (1) 418-809-4690
Paper type: Primary research article
DRIVERS OF BOREAL FOREST GROWTH 2
Currently, there is no consensus regarding the way that changes in climate will affect boreal
forest growth, where warming is occurring faster than in other biomes. Some studies
suggest negative effects due to drought-induced stresses, while others provide evidence of
increased growth rates due to a longer growing season. Studies focusing upon the effects
of environmental conditions on growth-climate relationships are usually limited to small
sampling areas that do not encompass the full range of environmental conditions; therefore,
they only provide a limited understanding of the processes at play. Here, we studied how
environmental conditions and ontogeny modulated growth trends and growth-climate
relationships of black spruce (Picea mariana) and jack pine (Pinus banksiana) using an
extensive data set from a forest inventory network. We quantified the long-term growth
trends at the stand scale, based upon analysis of the absolutely-dated ring-width
measurements of 2266 trees. We assessed the relationship between annual growth rates and
seasonal climatic variables, and evaluated the effects of various explanatory variables on
long-term growth trends and growth-climate relationships. Both growth trends and growth-
climate relationships were species-specific and spatially heterogeneous. While the growth
of jack pine barely increased during the study period, we observed a growth decline for
black spruce which was more pronounced for older stands. This decline was likely due to
a negative balance between direct growth gains induced by improved photosynthesis
during hotter-than-average growing conditions in early summers and the loss of growth
occurring the following year due to the indirect effects of late-summer heatwaves on
accumulation of carbon reserves. For stands at the high end of our elevational gradient,
frost damage during milder-than-average springs could act as an additional growth stressor.
DRIVERS OF BOREAL FOREST GROWTH 3
Competition and soil conditions also modified climate sensitivity, which suggests that
effects of climate change will be highly heterogeneous across the boreal biome.
Keywords: boreal forest, Canada, climate change, climate-induced stress, dendroecology,
The figure displays the effects (red=negative, blue=positive) of explanatory variables on
tree sensitivity to climate, and the resulting 1970-2005 growth trends. Old-growth boreal
black spruce stands exhibited a more negative response to previous summer temperature,
identified as the primary climatic driver of growth trajectories for this species. This finding
suggests an exacerbated effect of heat-induced stresses, which resulted in more negative
long-term growth trends for old-growth stands, especially when combined with late-frost
damage. Other explanatory variables, such as regional climate, competition and soil
conditions, modified tree sensitivity to climate.
DRIVERS OF BOREAL FOREST GROWTH 4
The boreal biome is warming faster than other regions of the world (IPCC, 2013). As a
result of a 35 % increase in atmospheric CO2 concentrations relative to pre-industrial
conditions, mean annual air temperature has risen by 0.5 to 3 °C in boreal North America
and an additional increase of 4-5 °C is expected by 2100 (Price et al., 2013). Climate
change threatens the ecological, social and economic services that boreal forests provide
(Gauthier, Bernier, Kuuluvainen, Shvidenko, & Schepaschenko, 2015). It is also unclear
whether boreal forests will continue to act as a carbon sink or will shift to become a carbon
source (Kurz et al., 2013), thereby renewing the scientific focus on these ecosystems and
on their ability to cope with, and to mitigate, global warming. As a proxy for tree vigour,
secondary growth can be used to study the response of trees to a changing climate and,
thus, to assess current and future trajectories of the boreal forest.
In the Northern Hemisphere, evidence of increased mortality rates and decreases in tree
growth and forest productivity is accumulating, not only for the interior of the boreal forest
(Cahoon et al., 2018; Girardin et al., 2016; Hember, Kurz, & Coops, 2016; Zhu et al., 2016),
but also at its northern edge (D’Arrigo et al., 2004). These ‘negative’ trends were linked,
amongst other factors, to heat and hydric stresses resulting from rising temperatures and
more frequent, longer-lasting, and harsher drought episodes (Barber, Juday, & Finney,
2000; Girardin et al., 2016; Juday & Alix, 2012; Nicklen et al., 2018; Trugman, Medvigy,
Anderegg, & Pacala, 2018; Zhang et al., 2008). In contrast, other studies provided strong
evidence for increased growth rates and higher stand productivity (Beck et al., 2011;
Girardin et al., 2011; Hember, Kurz, & Coops, 2017). These ‘positive’ trends, which were
observed mainly for the northernmost forested area, namely, the taiga, were likely due to
DRIVERS OF BOREAL FOREST GROWTH 5
higher rates of carbon assimilation and a longer growing season (Gennaretti, Arseneault,
Nicault, Perreault, & Bégin, 2014; Ju & Masek, 2016). These contrasting observations
demonstrate uncertainties regarding the persistence of the existing structure, composition
and function of the boreal biome under future warmer and dryer climatic conditions.
Tree sensitivity to climate is highly variable across the globe and is modulated by
environmental drivers that vary at local to global scales (Babst, Poulter, et al., 2012;
Girardin et al., 2016). Amongst these drivers, topography creates spatially heterogeneous
macroclimatic conditions and can modify the way that trees respond to changes in regional
climate (Hasler, Geertsema, Foord, Gruber, & Noetzli, 2015; Matías, Linares, Sánchez-
Miranda, & Jump, 2017). For example, in Central Europe, water-limited lowland forests
are more sensitive to drought, whereas forests in the upland portion of the elevational
gradient are primarily temperature-limited (Altman et al., 2017) and can benefit from
stronger and faster warming, which is expected for mountainous areas (Pepin et al., 2015).
More specifically, higher mean summer temperatures could improve the growth of trees at
the high end of the elevational gradient by increasing the number of wood cells that are
produced annually through improved xylogenetic processes and hydraulic performance
(Castagneri, Petit, & Carrer, 2015; Dulamsuren, Hauck, Kopp, Ruff, & Leuschner, 2017).
In contrast, some studies have observed decreased growth rates, even for trees growing in
mountainous sites in both central Europe and North America (Dittmar, Zech, & Elling,
2003; McLaughlin, Downing, Blasing, Cook, & Adams, 1987; Piovesan, Biondi, Filippo,
Alessandrini, & Maugeri, 2008), which questions the capacity of high-elevation forested
ecosystems to better cope with climate change (Austin & Niel, 2011).
DRIVERS OF BOREAL FOREST GROWTH 6
The annual growth performance of a tree is linked to its ability to access optimal
amounts of water, light and nutrients (Fritts, 1971), the availability of which is primarily
controlled by site-specific abiotic factors, such as soil conditions (e.g., Hember et al.,
2017). Soil structure, drainage and thickness of the organic layer determine soil water-
holding capacity (Minasny & McBratney, 2017) and drive nutrient cycling (e.g., Cavard,
Bergeron, Paré, Nilsson, & Wardle, 2018). In addition to its direct effects on tree growth,
soil moisture content alters microclimate and, thus, evapotranspiration rates, which may
modulate growth-climate relationships (Manrique-Alba et al., 2017). By modifying
resource availability, inter-individual competition can exacerbate tree sensitivity to harsh
climatic conditions (e.g. Buechling, Martin, & Canham, 2017; Ford et al., 2016; Gleason
et al., 2017; Jiang et al., 2018; Nicklen et al., 2018), or buffer growth gains from favourable
periods (Cortini, Comeau, & Bokalo, 2012). Ultimately, the capacity of a tree to efficiently
use resources will also dictate its response to climate (e.g., Carrer & Urbinati, 2004). Apart
from genotype-driven differences, ontogeny-related changes in a tree’s physiological needs
and in the efficiency of its hydraulic network (Ryan, Phillips, & Bond, 2006) can modify
its sensitivity to climate (e.g., Altman et al., 2017).
The high spatial variability in growing conditions that is encountered in boreal forests,
together with the multiplicity of interacting effects and feedbacks of environmental
variables that are present, hinder our understanding of the response of boreal forest trees to
climate. In regions with geographically limited and sparsely replicated sample networks
(Gewehr, Drobyshev, Berninger, & Bergeron, 2014), assessing climate effects on tree
growth is very difficult (but see Girardin et al., 2016), given that field-based measurements
do not cover the full range of variation in growing conditions. Some studies in western
DRIVERS OF BOREAL FOREST GROWTH 7
boreal North America and boreal Europe have examined variations in growth-climate
relationships along latitudinal and longitudinal gradients (Lloyd, Bunn, & Berner, 2011) or
between sites with different slope orientations (i.e. north vs south facing sites; Johnstone,
McIntire, Pedersen, King, & Pisaric, 2010; Walker & Johnstone, 2014) and moisture
conditions (Walker & Johnstone, 2014; Wilmking & Myers-Smith, 2008). However,
studies testing the effect of elevation gradient on the trees sensitivity to climate are lacking,
particularly in the eastern boreal North America. Furthermore, most past studies have
focused upon the direct effects of abiotic or biotic factors on tree growth, while the
feedback effects of environmental conditions on growth-climate relationships are still
rarely decribed (But see Nicklen et al., 2018; and Nicklen, Roland, Ruess, Schmidt, &
Lloyd, 2016 for the Pacific Coast of North America).
Here, we used an extensive and well-replicated provincial inventory network that
provides absolutely dated and annually resolved tree-growth data, as well as site-specific
environmental information for unmanaged forests in eastern boreal North America. This
network is located at the boundary between the interior boreal forest and the taiga, and
includes sample plots characterized by highly contrasting growing conditions. Our overall
objective was to examine whether the potential impacts of recent changes in climate varied
as functions of explanatory variables with respect to the growth of two needleleaf species
that are broadly distributed across North America, black spruce (Picea mariana (Miller)
B.S.P.) and jack pine (Pinus banksiana Lambert). We first quantified the recent growth
trends for the two species which, given the high variability in growing conditions, were
expected to be heterogeneous across the study zone. Then, we determined the climate
sensitivity of the two species, i.e., the relationship between inter-annual variation of
DRIVERS OF BOREAL FOREST GROWTH 8
secondary growth rates and fluctuations in seasonal values of mean temperature and total
precipitation over the period 1970-2005. We hypothesized that the growth of both species
would be negatively impacted by higher-than-average temperature during summer and
positively affected by higher-than-average temperature during spring and by higher-than-
average precipitation during summer. Finally, we assessed how explanatory variables (e.g.
climate, competition and soil conditions) affected spatial variability in growth-climate
relationships. We hypothesized that the negative effect of hotter- and dryer-than-average
summers, as well as the positive effect of high spring temperature on tree growth, would
be exacerbated in stands in the upper portion of the elevational gradient. We also
hypothesised that old stands, as well as trees growing in a highly competitive environment
and in well-drained sites, would respond more negatively to summer heatwaves.
2. MATERIALS AND METHODS
2.1 Sampling area
Our sampling network covered three degrees of latitude (50.25-53.25°N) and nearly
extended across the entire Province of Quebec (Canada) from east to west (57.5-78.25°W).
It was located in the boreal biome, which is characterised by needleleaf-dominated forests
(Robitaille, Saucier, Chabot, Côté, & Boudreault, 2015). Some regional patterns of climatic
conditions, dominant vegetation and natural disturbance regimes make it possible to divide
this wide biome into bioclimatic domains (Ansseau et al., 1997). In the north portion of the
region, the spruce-lichen bioclimatic domain is characterised by a harsh, cold and dry
climate, resulting in an open black spruce-dominated forest with a lichen mat, i.e., the taiga
vegetation subzone. South of the 52nd parallel, continuous boreal forest that is composed
mostly of pure black spruce stands covers the spruce-moss bioclimatic domain. The latter
DRIVERS OF BOREAL FOREST GROWTH 9
is subdivided into western and eastern zones based on precipitation patterns and fire cycles.
The western part is drier and, consequently, more prone to wildfire than the eastern zone
(Gouvernement du Québec, 2003). Within these three main bioclimatic domains, hereafter
referred to as “Boreal West,” “Boreal East” and “Taiga” (Figure 1), lower-level landscape
units are defined based upon the recurrent arrangements of the main permanent ecological
and vegetation features (48 landscape units are present in our sampling area), which in turn
are divided into ecological districts (284 ecological districts within our sampling area) that
are based upon their geological and physiographic features (Ansseau et al., 1997). Please
refer to the Figure 1 (B) for examples of geographical units mentioned throughout the
DRIVERS OF BOREAL FOREST GROWTH 10
FIGURE 1 (A) Forest inventory plot network. The pink squares and blue triangles
represent black spruce and jack pine temporary sample plots, respectively. The three main
bioclimatic domains encompassing the sample network are also delineated. The
background colour gradient represents the elevation gradient. (B) Geographical units
involved in statistical analyses, from the broader global scale of the province of Quebec to
the finer scale of the sample plot.
2.2 Tree-ring material
DRIVERS OF BOREAL FOREST GROWTH 11
The data that we used for this study were acquired from a sampling program of 400-m2
randomly distributed temporary circular sample plots (n = 875 plots), which was
established by the Ministère des Ressources naturelles et de la Faune du Québec (MRNFQ)
from 2005 to 2009 (Programme d’inventaire écoforestier nordique; Létourneau et al.,
2008). In each sample plot, the diameter at breast height (DBH, 1.3 m) of all living and
dead stems (DBH > 9 cm) was measured and environmental and stand-level conditions
were recorded. Disks were collected for stem analysis from one to three dominant or co-
dominant trees per species according to the provincial normative sampling protocol
(Ministère des Ressources Naturelles du Québec, 2008). We used only black spruce and
jack pine data since these species represented most (76 % and 15 %, respectively) of the
sampled trees. We selected 1-m-height stem-disks as a trade-off between basal ring
distortion and the number of visible rings (DesRochers & Gagnon, 1997). A total of 1914
black spruce and 352 jack pine disks with each having a minimum of 20 visible rings,
representing 812 sample plots, were retained for subsequent analyses. Sample disks were
processed using standard dendrochronological procedures for acquisition of ring-width
measurement series across four radii per disk (Ministère des Ressources Naturelles du
Québec, 2008). For each ring-width series, cross-dating and measurements were
statistically verified using the program COFECHA (Holmes, 1983). No major anomaly in
these tree-ring measurements was observed, and therefore all were retained for subsequent
2.3 Climate data and explanatory variables
For each plot, daily maximum and minimum temperatures (°C) and total precipitation
(mm) were obtained for the period of 1970-2005 using thin plate spline smoothing
DRIVERS OF BOREAL FOREST GROWTH 12
algorithms (ANUSPLIN), which interpolate site-specific estimates at a 0.08° x 0.08°
spatial resolution from historical weather observations, as described in Price et al. (2011).
Mean temperatures, which were computed as the average of monthly maximum and
minimum temperature values, were averaged and precipitation was summed at a seasonal
scale (meteorological seasons according to Trenberth, 1983: spring = March-May, summer
= June-August, autumn = September-November, winter = December-February). Readers
should refer to Supporting Information S1 for an overview of the trends in seasonal climate
in the study area.
Stand and environmental conditions were acquired from the plot survey conducted
by the Ministère de la Faune, des Forêts et des Parcs du Québec (Table 1, MFFPQ;
Robitaille et al., 2015). We considered the proportion of clay, sand and silt in the mineral
soil, organic layer thickness (OLT) and hydrological conditions of the sample plot assessed
as drainage classes. Elevation and slope were extracted for our sample plots from the
SRTM 90m Digital Elevation Database v4.1 (Jarvis, Reuter, Nelson, & Guevara, 2008).
For stand-level demographic features, stand age was defined as the age of the oldest
sampled tree in the plot. Stand basal area (BA) was computed as the sum of basal areas of
all trees with DBH > 9 cm within the plot, on a per-hectare basis. A tree-level competition
index (CI) was computed as the number of trees that were taller than the focal tree, divided
by the total number of trees within the plot, to assess assymetric competition (Ford et al.,
2016), following Weber et al. (2008). To do so, the height of all trees within a plot was
estimated from DBH using the allometric equations of Fortin et al. (2009). Individual CI
values were averaged at the plot level to ensure consistency with the working scale of the
growth-climate analyses. Temperature and precipitation normals were computed for the
DRIVERS OF BOREAL FOREST GROWTH 13
1970-2005 period to account for the west-to-east (continental-to-oceanic) climate gradient.
For brevity’s sake, these plot-level characteristics will be refered hereafter to as
TABLE 1 Plot-level statistics for the studied explanatory variables, by bioclimatic domain
range (min | max)
range (min | max)
range (min | max)
0 | 79
0 | 18
0 | 47.9
0 | 99.5
0 | 99.5
0 | 96.5
0 | 52
0 | 53.9
0 | 72
1 | > 100
0 | > 100
0 | > 100
1 | 6
1 | 6
1 | 6
96 | 637
100 | 860
113 | 912
0.13 | 13.15
0.13 | 19.58
0.13 | 12.36
28 | 294
28 | 331
30 | 309
BA (m2 ha-1)
0.78 | 49.88
0.90 | 55.39
0.59 | 35.44
0.07 | 1.00
0.02 | 0.98
0.07 | 1
685.56 | 927.26
775.83 | 1174.67
-3.07 | -0.31
-4.02 | 0.49
-4.35 | -0.93
Note: Clay = percentage of clay within the soil; Sand = percentage of sand within the soil; Silt = percentage of silt within the soil; OLT = organic layer thickness;
Drainage = drainage classes: from 1: rapid drainage to 6: poor drainage ; Elevation = altitudinal gradient; Slope = terrain’s slope, in degrees; Age = stand age (age
of the oldest tree in a plot, computed as the number of years between the calendar year of the oldest ring and the calendar year of the most recent ring recorded for
a tree); BA = basal area; CI = competition index; Prec. = average annual precipitation over the 1970-2005 period; Temp. = average mean annual temperature over
the 1970-2005 period.
DRIVERS OF BOREAL FOREST GROWTH 14
2.4 Statistical procedures
To test our working hypotheses, we applied a 3-step statistical procedure involving
different spatio-temporal scales (see workflow diagram in Supporting Information S2).
Step 1 : Trend analysis
Ring-width measurements of the four radii were averaged (arithmetic mean statistics, see
Supporting Information S3 for descriptive statistics of the raw series), and the mean ring-
width series were converted into basal area increments (
) using the
function bai.out in the R-package dplr (Bunn, 2008). We assumed the cross-sections were
perfectly circular in shape, and used these as a proxy for secondary growth to provide an
accurate quantification of wood production with ever-increasing tree diameter (Biondi &
Qeadan, 2008). Rings that were formed during the first 10 years were then eliminated,
given that they usually exhibit an atypical response to environmental drivers compared
with more mature rings (Loader, McCarroll, Gagen, Robertson, & Jalkanen, 2007). Next,
BAI were detrended using Generalised Additive Mixed Models (GAMM) to remove the
remaining ontogeny-induced (i.e., tree age and size) trends. One model was constructed for
each species and ecological district (See Supporting Information S4 for information about
the BAI chronologies and diagnostic plots of the GAMM models). Organic layer thickness
was added as a fixed term to account for the spatially-heterogeneous and mostly time-
independent effect of site quality on tree growth (Lavoie, Harper, Paré, & Bergeron, 2007).
BAI values were log-transformed to improve the normality of their distributions. The
structure of the GAMM model is as follows:
DRIVERS OF BOREAL FOREST GROWTH 15
where i represents the individual tree, j represents the species, k represents the plot, l
represents the ecological district, and t represents the year. BAI is the basal area increment
of tree i at specific year t, BA is the basal area of tree i at specific year t (computed as the
sum of BAI of previous years), OLT is the organic layer thickness of plot k, and AgeC is
the cambial age (1-m height ring count) of tree i at year t. An autoregressive term, AR1
(autoregressive order p = 1, moving average order q = 0), was added to account for temporal
autocorrelation. We tested the significance of a nested random effect (tree nested in plot)
by conducting ANOVAs and likelihood ratio tests. Because it did not improve the model’s
fit and led to the same results (data not shown), we discarded the random term of the plot
from the final model and kept only the random effect of the tree (TreeID).
Annual Growth Changes (GC), which were expressed as the percent deviation from
predicted values of the GAMM model, were then computed following Girardin et al.
(2016). GC values were aggregated by year, plot and species using the median statistics for
computation of GCmedian chronologies (robust statistics; Huber, 2005). Because, for several
trees, the 2005 growth-ring was the last whole growth-ring, the upper temporal limit of the
analyses was fixed to 2005 to ensure consistency between chronologies. From the GCmedian
chronologies, growth trends were examined over two time periods: 1950-2005 and 1970-
2005. These periods were marked by significant increases in mean annual temperatures of
the area (Price et al., 2013) and characterized by the highest number of tree rings per
calendar year (i.e. the highest sample depth, see Supporting Information S3.2). Linear
regressions were applied (GCmedian ~ year), and the estimated regression slope was used as
a proxy for the long-term growth trend. Trend significance was assessed following the
statistical procedure described by Yue et al. (2004). This method corrects the p-value of
DRIVERS OF BOREAL FOREST GROWTH 16
the non-parametric Mann-Kendall trend test with the effective sample size of the time
series to reduce the influence of serial correlation (function mkTrend in the R-package
fume; Santander Meteorology Group, 2012). Even if there were trend reversals for a few
plots (Figure 2), the 1970-2005 and 1950-2005 trends were globally highly correlated (see
Supporting Information S5.1). For the purposes of comparison, 1970-2005 GAMM-based
trends were compared with trends that were estimated from the application of two other
commonly used detrending methods, namely modified negative exponential models and
regional curve standardisation (See Supporting Information S5.2).
Step 2 : Growth-climate relationships
Since weather station data availability and, therefore, climate data accuracy, is better for
the most recent time periods (Ols, Girardin, Hofgaard, Bergeron, & Drobyshev, 2017), we
decided to retain data from the shorter and most recent period, i.e., 1970-2005, for climate-
growth analyses. Linear mixed models were fitted by plot and species, which included
residuals of GAMM-detrended BAI as response variables, together with the set of
seasonally aggregated climatic variables as fixed terms, and tree identity as a random term.
Mean seasonal temperature and total precipitation of the year of ring formation were
considered as explanatory variables. Since trees can allocate carbohydrates that were
acquired in the previous growing season to the biomass production of the year of ring
formation (Granda & Camarero, 2017), climate data from summer and autumn of the
previous year were also considered as fixed terms, leading to a total of ten climatic
variables (please refer to the Supporting Information S6.1 for the list of climate variables
used in linear mixed models). The structure of the global model is as follows:
DRIVERS OF BOREAL FOREST GROWTH 17
where i represents the tree, j represents the species, k represents the plot and t represents
the year. (TreeID) is a random term that accounts for the variability between individual
trees. An error term with an AR1 (p = 1, q = 0) correlation structure was added to the model
which accounts for the serial correlation. Collinearity amongst climatic variables was low,
with the mean of pairwise Pearson correlations between variables below a stringent
threshold value of 0.4 (Supporting Information S6.1; maximum value of |0.37|; Dormann
et al., 2013). Multi-model selection based upon the Akaike information criterion corrected
for small sample size (AICc), was performed for this global model using the package
MuMIn (Bartoń, 2018). A 95% confidence set of models was selected for multi-model
inference as models whose cumulative Akaike weight is below 0.95 (Burnham &
Anderson, 2002). Readers can consult Supporting Information S6.3 for AICc values of all
of the 1024 evaluated models, along with Akaike weights of the best model and the number
of models used for multi-model inferences. Weighted averages of parameter estimates were
derived from this set of best approximating models, and Student’s t-statistics were
computed as the ratio between the average model estimate and its corresponding standard
error. These statistics provide both the general direction of the growth-climate relationship
(i.e., negative or positive slope), and the strength of this relationship (the farther from zero
the t-value is, the stronger is the effect), weighted by the model’s predictive capacity and
based upon the selected climatic variables. The 95 % adjusted confidence intervals of the
t-statistics were also computed, together with Pearson correlations between residuals from
the GAMM models and predicted values from the climate models (Supporting Information
DRIVERS OF BOREAL FOREST GROWTH 18
S6.2) as an additional means of assessing the model’s predictive capacity. Results of
growth-climate analyses that were based upon residuals from the two additional detrending
methods are provided in Supporting Information S7.
Slopes from the linear regressions and t-statistics from the climate-growth mixed
models were interpolated across the whole area using the Empirical Bayesian Kriging
algorithm of the Geostatistical Analyst tool in ArcGIS v.10.4 (input options: empirical
transformation of the data, semi-variogram model = exponential-type, search radius = 1°,
smoothing factor = 0.2). The output raster maps permitted visual examination of
geographical patterns in long-term growth trends and climate sensitivity.
Step 3 : Link with explanatory variables
The relationships between explanatory variables (listed in Table 1) and tree sensitivity to
climate were assessed by conducting redundancy analyses (RDA) using Canoco software
v.4.5 (ter Braak & Smilauer, 2009). Because tree sensitivity to climate and environmental
conditions are highly variable from east to west (see Figure 4 and Table 1), site conditions
might affect growth-climate relationships depending upon the location of the plot (Wu et
al., 2018). If averaged over the whole gradient, the effect of these conditions could cancel
each other out. Consequently, one RDA was conducted per bioclimatic domain as a trade-
off between data aggregation and ecological relevance, as recommended by Ols et al.
(2018). The t-statistics from the climate mixed models were considered as response
variables (i.e., the “species” data matrix) and explanatory variables were considered as
independent variables (i.e., the “environment” matrix). Climate normals were also included
as independent variables, together with a dummy variable accounting for the species
DRIVERS OF BOREAL FOREST GROWTH 19
identity of the sampled tree, i.e., the difference in sensitivity to climate between jack pine
(the reference level) and black spruce. Please refer to Supporting Information S8.3 for the
list of independent variables considered in RDA analyses. Latitude, longitude and the
average distance to the four nearest weather stations (ranging from 3.8 km to 153.1 km, see
Supporting Information S8.1) were added as conditioning variables to remove the effects
of spatial non-independence of the plots and the potential lack of accuracy in the climate
data set prior to analysis. Independent variables were transformed to improve the normality
of their distributions, and then standardised prior to analysis (R package rcompanion;
Mangiafico, 2017; Tukey’s ladder of powers; Tukey, 1977). Forward selection of
independent variables was done using Monte-Carlo permutation tests (n = 9999
permutations under the reduced model; α = 0.05). Growth trends were included passively
in the RDA in order to examine these in context with climate-environmental relationships
(such supplementary ‘passive’ variables do not influence the ordination). To summarise
the information that was displayed by the ordination plots (Supporting Information S8.2),
modified t-tests accounting for spatial autocorrelation were conducted between each of the
RDA-selected independent variables and response variables (i.e., tree sensitivity to
climate). The function modified.ttest of the R package SpatialPack was used (Osorio,
Vallejos, Cuevas, & Mancilla, 2018); α = 0.05).
Significant variables were grouped into six sets according to the ecological process they
represent: stand maturity, competition, altitudinal gradient, soil conditions, regional
climate, and species identity (also see Supporting Information S8.3). Variation partitioning
was then conducted to identify common and unique contributions to the total percentage
of variation in the matrix of response variables (adjusted R2) explained by the model and
DRIVERS OF BOREAL FOREST GROWTH 20
better address the question of relative influences of the six sets of independent variables
that were considered in the model (Peres-Neto, Legendre, Dray, & Borcard, 2006). The
forward selection procedure used in the RDA led to up to five sets of independent variables
by bioclimatic domain. The variation partitioning algorithm (varpart) in the R-package
vegan was used (9999 permutations, partitions computed from adjusted R2; Oksanen et al.,
2018), which only allows a maximum of four sets of independent variables to be considered
in a same analysis. To overcome this limitation, we determined the unique and common
contributions of stand maturity, competition, altitudinal gradient, soil conditions and
regional climate. Next, we determined the contribution of the taxonomic identity of the tree
(selected in each of the three bioclimatic domains) by comparing it to the contribution of
all other independent variables grouped together. The dummy species variable in RDAs
allowed the quantification of the variability in growth-climate relationships resulting from
the difference between the two species without splitting the data by species, which would
have lowered the number of sample plot by analysis and consequently the statistical power,
i.e. the likelihood to correctly reject the null hypothesis. Analyses by species were also
tested and results of these analyses are provided as Supporting Information S8.4.
3.1 Growth trends are spatially heterogeneous and species-specific
When averaged over the sample plots, dissimilar long-term growth trends appeared
between species (Supporting Information S9). Growth rates of black spruce decreased, with
a trend estimated at -0.35 % y-1 ± (std) 1.61 % y-1 from 1950 to 2005. For the 1970-2005
period, the trend in the annual growth rate was -0.14 % y-1 ± 2.44 % y-1. For jack pine, both
the 1950-2005 and 1970-2005 periods were characterised by an annual increase in growth
DRIVERS OF BOREAL FOREST GROWTH 21
of 0.21 % y-1 ± 3.31 % y-1 and 0.21 % y-1 ± 3.37 % y-1, respectively. However, species-
specific growth trajectories were not homogeneous across the study region (Figures 2 and
3; Supporting Information S9). Growth of black spruce increased in the western part of the
Boreal West and in the central part of the Boreal East between 1970 and 2005, but
decreased elsewhere (Figure 2). Growth of jack pine increased regardless of bioclimatic
domain between 1950 and 2005, but decreased in the western parts of Boreal West and
Taiga between 1970 and 2005 (Figures 2 and 3).
FIGURE 2 (A) Growth trends for black spruce and jack pine, for the 1950-2005 and 1970-
2005 periods, shown as slope coefficients of the plot-scaled regression models of detrended
BAI values against calendar years. Empirical Bayesian kriging was applied to interpolate
plot-based trends across the entire area. Dots highlight significant trends (α = 0.1). The
DRIVERS OF BOREAL FOREST GROWTH 22
proportion of significant trends is shown at the bottom of each map. (B) Distributions of
growth trend slopes by species and bioclimatic domain (boxplots). Black dots represent the
mean value for the specific species and bioclimatic domain. Black lines inside the boxplots
are median values, and error bars represent the lower and upper whiskers (representing the
variability outside the upper and lower quartiles). The dotted line represents a value of zero,
i.e., no trend in long-term growth.
FIGURE 3 Median chronologies (red curves) of black spruce (left panels) and jack pine
(right panels) detrended BAI (growth change) per bioclimatic domain (upper row: Boreal
West; middle row: Boreal East; lower row: Taiga). Yellow shading and dotted lines delimit
the bootstrapped 95 % confidence intervals, with LOESS smoothing shown by the blue
lines (span = 0.2). Violet box and blue shading highlight the two time intervals (1950-2005
and 1970-2005, respectively). Black dashed lines denote a zero effect, i.e., no deviation
compared to the value predicted by the GAMM model.
3.2 Sensitivity to climate is dissimilar across the landscape
DRIVERS OF BOREAL FOREST GROWTH 23
Growth–climate response patterns were estimated for the two tree species to identify the
key climate factors that were driving observed variability in growth (Figure 4). Summer
temperature of the year preceding growth and spring precipitation in the year of growth
had significant negative relationships with black spruce growth, while winter precipitation
and winter temperature had a positive influence. The importance of these variables was not
limited to particular regions but extended across vast areas (Figure 5). Black spruce tree
sensitivity to other climate variables was more spatially heterogeneous (Figures 4 and 5).
A high level of precipitation during previous-year summers had a significant positive effect
upon the growth of black spruce within the Boreal East and Boreal West; this effect was
not statistically significant in the Taiga (Figure 4). Excess-heat and high precipitation
during previous-year autumns negatively affected spruce growth in the Boreal West and
Taiga but had no significant effect in the Boreal East. Within the Boreal East and Taiga,
the growth of black spruce was increased by hotter-than-average summers occurring during
the year of ring formation and was decreased by milder-than-average springs. These
relationships were mostly the opposite of what was observed within the Boreal West.
The response of jack pine to climate was less statistically significant and often
opposite to that of black spruce. Regardless of bioclimatic domain, jack pine growth was
increased by previous-year warm autumns and current-year summer warmth, but it was
decreased by high winter precipitation (Figures 4 and 5). Current- and previous-year wet
summers significantly increased the growth of jack pine within the Boreal East and Boreal
West (Figure 4). Jack pine growth was positively correlated with mild and wet springs
within the Taiga and with mild winters within the Boreal West, but was negatively
impacted by wet springs within the Boreal East (Figure 4).
DRIVERS OF BOREAL FOREST GROWTH 24
FIGURE 4 Arithmetic means (black dots) and bootstrapped 95 % confidence intervals
(rectangles, R=10000 replications) of t-statistic values per bioclimatic domain for black
spruce and jack pine, for each of the seasonal climatic variables. “T” and “P” at the
beginning of a variable’s name denote temperature and precipitation, respectively.
Uppercase letters denote climatic variables for the current growing season (winter, spring,
DRIVERS OF BOREAL FOREST GROWTH 25
summer), and lowercase letters denote climatic variables of the previous growing season
(previous summer, previous autumn). Blue and red rectangles indicate a significant (95 %
confidence interval excluding zeroes) positive and negative effect, respectively, of the
climatic variable at the scale of the bioclimatic domain, and grey rectangles are for non-
DRIVERS OF BOREAL FOREST GROWTH 26
FIGURE 5 Kriging-interpolated growth response significance (based on t-statistics, 1970-2005) to seasonal climatic variable (left:
temperature, right: precipitation) for black spruce (left panel) and jack pine (right panel). Green-to-blue colours denote a negative effect
of the climate variable on tree growth, yellow colour means no impact of the climate variable on tree growth, and orange-to-red colours
DRIVERS OF BOREAL FOREST GROWTH 27
denote a positive effect. Dots display significant values, i.e., plots for which the 95 % confidence interval of the t-statistics excludes
DRIVERS OF BOREAL FOREST GROWTH 28
3.3 Plot-level features had low but significant effects on growth-climate relationships
Sensitivity to climate differed between the two species, especially within the Boreal West,
where species identity of the sampled trees alone accounted for 15 % of variation in
growth-climate relationships (Figure 6). Such taxononomic variability in growth sensitivity
to climate can be readily noted in Figures 4 and 5. Contributions of the sets of explanatory
variables stand maturity, competition, altitudinal gradient, soil conditions, and regional
climate to the climate sensitivity variance were much lower. The elevational gradient
explained the highest proportion of variation in tree response to climate within the Boreal
East and Taiga (5 % and 9 %, respectively; Figure 6). Stand maturity, alone or in
combination with other explanatory variables, accounted for 7 %, 1 % and 5 % of the
variation in growth-climate relationships within the Boreal West, Boreal East and Taiga,
respectively. For competition, these values were respectively 3 %, 2 % and 2 %.
Stands that were composed mainly of old black spruce trees exhibited growth that
was more negatively correlated with previous-year summer and autumn temperatures, but
more positively correlated with winter precipitation compared to recently regenerated
stands (Figures 6 and 7, and Supporting Information S10.1). These old-grown black spruce
stands also exhibited the steepest declines in growth rates during 1970-2005 (Supporting
Information S11). The positive effect of warmer-than-average autumns, winter and springs
on the growth of jack pine was lower for stands that were composed of old trees in
comparison with more recently regenerated stands (Supporting Information S8.4). Snowy
and mild winters increased the growth of black spruce more than that of jack pine, but
black spruce growth was more negatively correlated with wet and warm springs and with
DRIVERS OF BOREAL FOREST GROWTH 29
excess-heat during autumns of the previous years than that of jack pine. Previous-year wet
summers and current-year mild springs decreased the growth of stands in the upper portion
of the elevational gradient (i.e., above 500 m a.s.l.), while excessively high temperatures
during current-year summers increased their growth more strongly than for stands at lower
elevations (Figures 6 and 7, and Supporting Information S10.1 and S10.2). Similarly,
growth in stands that were composed of taller trees (higher CI) was more negatively
affected by excess-heat during previous-year summers than those stands that were
composed of smaller-sized trees. Tree growth in more densely populated stands (higher
BA) was also more positively correlated with winter temperature, but less positively
correlated (within Boreal East and Taiga) or more negatively correlated (within Boreal
West) with current-year summer precipitation than stands of lower densities (Figures 6 and
7 and Supporting Information S9.1).
The effect of other explanatory variables on tree sensitivity to climate was restricted
to a specific region, such as soil conditions within the Boreal West and the continental-to-
oceanic climate gradient within the Boreal East, which accounted for 3 % and 5 % of the
variation in growth-climate relationships, respectively (Figure 6).
DRIVERS OF BOREAL FOREST GROWTH 30
FIGURE 6 Left: The sets of independent variables used in variation partitioning (Please
refer to Table 1 for variable ranges). Middle: Effect of each explanatory variable on tree
sensitivity to climate, based on autocorrelation-corrected Pearson correlations. The
relationship between long-term growth trends and climatic variables is also shown
(uppercase letters: current year; lowercase letters: previous year). Red and blue shadings
are for negative and positive relationships, respectively, that are significant at α = 0.05.
Gray shadings denote non-significant relationships. Significant relationships common to at
least two bioclimatic domains are emphasised with a dot. Right: Proportion of variance
DRIVERS OF BOREAL FOREST GROWTH 31
explained by each set of independent variables, alone or in combination with other sets
(Venn diagrams), by bioclimatic domain. The proportion of variance explained only by the
species (pine or spruce) and the proportion of variance unexplained by the selected
variables are shown below the diagrams.
FIGURE 7 Effect of topographic position on sensitivity to (A) spring temperature, (B)
summer temperature, (C) previous sumer precipitation, and (D) effect of the age of the
stand on the sensitivity to previous summer temperature. Blue and orange dots are observed
values for jack pine and black spruce, respectively. Also shown are the Spearman’s rho
coefficient (r) and p-value of the modified t-test corrected for the effect of spatial
DRIVERS OF BOREAL FOREST GROWTH 32
Using a dendroecological dataset from a randomly distributed forest inventory network
that consisted of 812 plots and 2266 trees, we provided an overview of the response of two
major boreal needleleaf species to recent climate change, across explanatory variables that
include stand maturity, competition, elevational gradient, soil conditions, and regional
climate within eastern boreal North America. Our results highlighted spatial heterogeneity
in long-term growth trends across the studied forest: in some areas growth rates decreased,
while in others growth increased over the last few decades. Tree sensitivity to climate was
also highly spatially heterogeneous. Our study underscores the utility of employing broadly
distributed datasets for assessing the complexity of climate change effects on a forest
ecosystem (Klesse et al., 2018; Nicklen et al., 2018).
The species identity of the tree explained a greater proportion of variation in
growth-climate relationships than did all other explanatory variables in the Boreal West.
Further, we observed contrasting growth trends between the two species. Our analyses
suggest that sensitivity to climate is determined primarily by a species-specific set of
functional traits. Black spruce and jack pine usually occupy sites with different soil
structures (Balland, Bhatti, Errington, Castonguay, & Arp, 2006) and have very different
root system architectures and physiological efficiencies (Blake & Li, 2003; Strong & La
Roi, 1983), which could explain differences in climate sensitivity.
We identified previous-year summer temperature and, to a lesser extent, previous-
autumn temperature as two of the climatic factors with the greatest negative effects on
annual growth rates of black spruce trees. Excess-heat during late summer and autumn may
lead to declines in carbohydrate reserve accumulation at the end of the growing season,
thereby negatively affecting spring growth, as was previously observed in both boreal
DRIVERS OF BOREAL FOREST GROWTH 33
North America and northern Europe (Girardin et al., 2016; Ols et al., 2018). Decreased
reserve formation of heat-stressed trees can result from accelerated respiration, which leads
to higher and more rapid use of photosynthates that otherwise would be available for
storage (Anderegg & Anderegg, 2013; Granda & Camarero, 2017; Sala, Woodruff, &
Meinzer, 2012). The steepest declines in observed growth for individuals that were the
most greatly affected by above-average temperature during summer of the year prior to
ring formation suggest that hot extremes are one of the primary determinants of growth
trajectories for boreal black spruce forests, as has been observed for white spruce seedlings
in plantations by Benomar et al. (2018). The deeper root system of jack pine trees could
have allowed them to access additional water resource in deeper soil layers, and their
greater resource use efficiency could have prevented them from an overuse of carbon
reserves during heatwaves, potentially leading to an uninterrupted carbohydrates storage
during hotter-than-average late growing seasons. The resulting higher amount of
photosynthates available the following spring could explain the positive correlation
between jack pine growth and previous autumn temperature. Our results also suggest that
in addition to the effect of species-specific traits, some variation originated from spatially
varying site features like stand maturity, position along the elevational gradient, regional
climate and soil conditions. Yet, these site-level features explained a lower proportion of
the variance in patterns of growth-climate relationships within bioclimatic domains.
As predicted from our main working hypothesis, position along the elevational
gradient explained a low but significant proportion of the variation in tree response to
spring and summer temperature and to previous summer precipitation. This finding
illustrates diverging climatic constraints, from water-limited trees at low elevation to trees
DRIVERS OF BOREAL FOREST GROWTH 34
constrained by cold temperatures during the growing season at the high end of the
elevational gradient. Contrary to our expectations, growth of black spruce was more
negatively affected by mild springs when located in the upper portion of our elevational
gradient. This counter-intuitive effect of mild springs was recently observed elsewhere
(Babst, Carrer, et al., 2012; Ols et al., 2017) and could result from earlier onset of
physiological activity and growth in response to warming (Gu et al., 2008; Richardson et
al., 2018; Vitasse, Signarbieux, & Fu, 2017). Late-frost events generally occur more
frequently at higher elevations and in cold regions such as the Boreal East and Taiga than
at lower elevation or in relatively warmer regions such as the Boreal West. These events
can damage early formed shoots and roots, thereby reducing total seasonal growth (Kidd,
Copenheaver, & Zink-Sharp, 2014; Montwé, Isaac-Renton, Hamann, & Spiecker, 2018).
Summer warmth had a contrasting effect on tree growth, depending upon its
occurrence. While previous-year high temperatures had negative effects on growth during
the following growing season, hotter-than-average summers had an immediate and positive
effect on growth rates in the year of occurrence. The latter relationship, which had been
observed in the Boreal East and Taiga, was more pronounced for stands at upper elevation
sites than at lower elevations and could be linked to a decrease in the risk of late-frost
damage and faster snowmelt in early summer (Vitasse et al., 2017). Hot summers are also
correlated with high solar radiation and, consequently, with higher rates of photosynthesis,
especially in sites where water is not a factor limiting to tree growth, such as stands at the
high end of our elevational gradient in the Boreal East and Taiga. Temperature generally
decreases with elevation, so an increase in summer temperature can lead to a greater net
beneficial effect on tree growth at higher elevations (see Supporting Information S12).
DRIVERS OF BOREAL FOREST GROWTH 35
However, the resulting growth gain would have been outweighted by the growth decline
due to late frost damage, which could explain that growth trends in stands within the central
hilly area were more negative than in the westernmost stands.
Effects of other variables on climate sensitivity, such as stand maturity and
competition, were generally consistent accross regions. Excess-heat during previous-year
summers and autumns had a significantly greater negative impact on black spruce growth
in older stands compared to more recently regenerated stands. The increase in climate
sensitivity with age has been extensively documented (e.g., Altman et al., 2017; Schuster
& Oberhuber, 2013), and was linked to ontogeny-related morphophysiological changes
(Ryan et al., 2006) leading to a decrease in hydraulic conductance (Magnani, Mencuccini,
& Grace, 2000). During drought, hydraulic conductance may decrease more strongly in old
and tall trees because of greater path resistance (Ryan & Yoder, 1997); the resulting
decreases in stomatal conductance and photosynthesis may entail, along with greater
metabolic demand in tall trees (Hartmann, 2011), depletion of carbohydrate reserves in
older stands. This response is a potential explanation for the negative relationship between
1970-2005 growth trends and stand age (Supporting Information S11; see also Chen et al.
(2016); Girardin et al. (2014)).
Competition pressure also significantly modulated the growth-climate
relationships. Growth of trees in densely vegetated (high BA) stands that were composed
of taller individuals (high CI) was more negatively correlated with excess-heat during
previous-year summers and less positively correlated to current-year wet summers than in
a less competitive environment. These relationships could have originated from lower
carbon assimilation and carbohydrate reserve formation originating from reduced water
DRIVERS OF BOREAL FOREST GROWTH 36
availability (Gleason et al., 2017). In contrast, black spruce trees responded more positively
to mild winters in densely vegetated compared to more sparsely populated stands, which
may be due to the stabilizing effect of a dense canopy on local-scale hydrothermal
properties (Gu et al., 2008; Vaganov, Hughes, Kirdyanov, Schweingruber, & Silkin, 1999),
similar to the effect of high structural diversity (Aussenac et al., 2017).
Soil conditions accounted for a significant proportion of the variation in growth-
climate relationships, but this was true only for black spruce in the Boreal West. This region
is characterised by a contrasting physiography spanning comparatively flat landcapes with
a high proportion of peatlands in the west to hilly terrain with sandy-loam soils in the east
(Robitaille et al., 2015; also see Figure 1 and Table 1), together with resulting differences
in soil hydrology. During hot summers, the water table of soils with a high proportion of
organic material is lowered, and in combination with the high degree of dessication of the
peat substrate (Gewehr et al., 2014; Voortman et al., 2013), may have reduced water
availability and exacerbated summer heat stress, particularly for trees with shallow rooting
systems such as black spruce. Paradoxically, an excess of water during consecutive wet
springs and summers also reduced growth of black spruce in Boreal West (Figure 5), most
likely because of hypoxic stresses resulting from elevation of the water table in poorly
drained sites (Zobel, 1990). A positive correlation between growth of western black spruce
trees and the annual area burned, which is a proxy for litter and deep organic layer dryness
(Molinari et al., 2018), adds credibility to the assumption that tree sensitivity to
precipitation was strongly modulated by soil hydrology in the Boreal West (Supporting
DRIVERS OF BOREAL FOREST GROWTH 37
Overall, we identified mostly negative growth trends for black spruce and only
barely positive trends for jack pine during the 1970-2005 period, which confirms the
absence of climatically-induced stimulation of tree vigour that was previously observed for
the boreal forest (Girardin et al., 2016; Hember et al., 2016; Ju & Masek, 2016; Zhu et al.,
2016). However, forest growth trends were spatially heterogeneous, and the productivity
of some areas increased over the last few decades. Variability in growth-climate
relationships that was explained by the set of variables considered in our analysis remained
low (< 25 %), as is the case in many studies focusing on ecological processes. Our random
sampling strategy implies that many factors, which are potentially involved in growth-
climate relationships, were not considered and could not be controlled for, such as the
effect of non-tree vegetation, insect epidemics, or nutrient cycling. In addition, the genetic
diversity of the species under study surely induced different responses to climate between
populations (Housset et al., 2016; but see Avanzi et al., 2019). Based upon our results, we
suggest that the warming threshold beyond which the productivity of the boreal forest will
shift from positive to negative (~ +2 °C; D’Orangeville et al., 2018) is likely very
heterogeneous across the boreal biome, but may already have been reached in many of our
black spruce stands.
This research was conducted as part of the International Research Group on Cold Forests.
This study was made possible thanks to the financial support that was provided by the
Strategic and Discovery programs of NSERC (Natural Sciences and Engineering Research
Council of Canada), and a MITACS scholarship co-funded by NSERC and Ouranos.
Additional financial support was provided by the Canadian Forest Service and the UQAM
DRIVERS OF BOREAL FOREST GROWTH 38
Foundation (De Sève Foundation fellowship and TEMBEC forest ecology fellowship). We
thank Dan McKenney and Pia Papadopol (Canadian Forest Service) for providing the
ANUSPLIN climate data. Many thanks to XiaoJing Guo (Canadian Forest Service) for the
initial version of the R-scripts and helpful advice during statistical analyses, to Claire
Depardieu for relevant comments on the manuscript, and to Isabelle Lamarre and W. F. J.
Parsons for language editing.
Altman, J., Fibich, P., Santruckova, H., Dolezal, J., Stepanek, P., Kopacek, J., … Cienciala, E. (2017).
Environmental factors exert strong control over the climate-growth relationships of
Picea abies in Central Europe. Science of the Total Environment, 609, 506–516.
Anderegg, W. R. L., & Anderegg, L. D. L. (2013). Hydraulic and carbohydrate changes in
experimental drought-induced mortality of saplings in two conifer species. Tree
Physiology, 33(3), 252–260. https://doi.org/10.1093/treephys/tpt016
Ansseau, C., Bélanger, L., Bergeron, J. F., Bouchard, A., Brisson, J., De Grandpré, L., … Vasseur, L.
(1997). Écologie forestière. In Manuel de foresterie (1st Edition, pp. 134–279). Sainte-
Foy, Québec: Presses de l’Université Laval.
Aussenac, R., Bergeron, Y., Mekontchou, C. G., Gravel, D., Pilch, K., & Drobyshev, I. (2017).
Intraspecific variability in growth response to environmental fluctuations modulates the
stabilizing effect of species diversity on forest growth. Journal of Ecology, 105(4), 1010–
Austin, M. P., & Niel, K. P. V. (2011). Improving species distribution models for climate change
studies: variable selection and scale. Journal of Biogeography, 38(1), 1–8.
DRIVERS OF BOREAL FOREST GROWTH 39
Avanzi, C., Piermattei, A., Piotti, A., Büntgen, U., Heer, K., Opgenoorth, L., … Leonardi, S. (2019).
Disentangling the effects of spatial proximity and genetic similarity on individual growth
performances in Norway spruce natural populations. Science of the Total Environment,
650, 493–504. https://doi.org/10.1016/j.scitotenv.2018.08.348
Babst, F., Carrer, M., Poulter, B., Urbinati, C., Neuwirth, B., & Frank, D. (2012). 500 years of
regional forest growth variability and links to climatic extreme events in Europe.
Environmental Research Letters, 7(4), 045705.
Babst, F., Poulter, B., Trouet, V., Tan, K., Neuwirth, B., Wilson, R., … Frank, D. (2012). Site- and
species-specific responses of forest growth to climate across the European continent.
Global Ecology and Biogeography, 22(6), 706–717. https://doi.org/10.1111/geb.12023
Balland, V., Bhatti, J., Errington, R., Castonguay, M., & Arp, P. A. (2006). Modeling snowpack and
soil temperature and moisture conditions in a jack pine, black spruce and aspen forest
stand in central Saskatchewan (BOREAS SSA). Canadian Journal of Soil Science,
86(Special Issue), 203–217. https://doi.org/10.4141/S05-088
Barber, V. A., Juday, G. P., & Finney, B. P. (2000). Reduced growth of Alaskan white spruce in the
twentieth century from temperature-induced drought stress. Nature, 405(6787), 668–
Bartoń, K. (2018). MuMIn: Multi-Model Inference (Version 1.40.4). Retrieved from
Beck, P. S. A., Juday, G. P., Alix, C., Barber, V. A., Winslow, S. E., Sousa, E. E., … Goetz, S. J. (2011).
Changes in forest productivity across Alaska consistent with biome shift. Ecology Letters,
14(4), 373–379. https://doi.org/10.1111/j.1461-0248.2011.01598.x
Benomar, L., Lamhamedi, M. S., Pepin, S., Rainville, A., Lambert, M.-C., Margolis, H. A., …
Beaulieu, J. (2018). Thermal acclimation of photosynthesis and respiration of southern
DRIVERS OF BOREAL FOREST GROWTH 40
and northern white spruce seed sources tested along a regional climatic gradient
indicates limited potential to cope with temperature warming. Annals of Botany, 121(3),
Biondi, F., & Qeadan, F. (2008). A theory-driven approach to tree-ring standardization: defining
the biological trend from expected basal area increment. Tree-Ring Research, 64(2), 81–
Blake, T. J., & Li, J. (2003). Hydraulic adjustment in jack pine and black spruce seedlings under
controlled cycles of dehydration and rehydration. Physiologia Plantarum, 117(4), 532–
Buechling, A., Martin, P. H., & Canham, C. D. (2017). Climate and competition effects on tree
growth in Rocky Mountain forests. Journal of Ecology, 105(6), 1636–1647.
Bunn, A. G. (2008). A dendrochronology program library in R (dplR). Dendrochronologia, 26(2),
Burnham, K. P., & Anderson, D. R. (2002). Model Selection and Multimodel Inference: A Practical
Information-Theoretic Approach (2nd ed.). Retrieved from
Cahoon, S. M. P., Sullivan, P. F., Brownlee, A. H., Pattison, R. R., Andersen, H.-E., Legner, K., &
Hollingsworth, T. N. (2018). Contrasting drivers and trends of coniferous and deciduous
tree growth in interior Alaska. Ecology, 99(6), 1284–1295.
Carrer, M., & Urbinati, C. (2004). Age‐dependent tree‐ring growth responses to climate in Larix
decidua and Pinus cembra. Ecology, 85(3), 730–740. https://doi.org/10.1890/02-0478
DRIVERS OF BOREAL FOREST GROWTH 41
Castagneri, D., Petit, G., & Carrer, M. (2015). Divergent climate response on hydraulic-related
xylem anatomical traits of Picea abies along a 900-m altitudinal gradient. Tree
Physiology, 35(12), 1378–1387. https://doi.org/10.1093/treephys/tpv085
Cavard, X., Bergeron, Y., Paré, D., Nilsson, M.-C., & Wardle, D. A. (2018). Disentangling effects of
time since fire, overstory composition and organic layer thickness on nutrient
availability in Canadian boreal forest. Ecosystems, 1–16.
Chen, H. Y. H., Luo, Y., Reich, P. B., Searle, E. B., & Biswas, S. R. (2016). Climate change-
associated trends in net biomass change are age dependent in western boreal forests of
Canada. Ecology Letters, 19(9), 1150–1158. https://doi.org/10.1111/ele.12653
Cortini, F., Comeau, P. G., & Bokalo, M. (2012). Trembling aspen competition and climate effects
on white spruce growth in boreal mixtures of Western Canada. Forest Ecology and
Management, 277, 67–73. https://doi.org/10.1016/j.foreco.2012.04.022
D’Arrigo, R. D., Kaufmann, R. K., Davi, N., Jacoby, G. C., Laskowski, C., Myneni, R. B., & Cherubini,
P. (2004). Thresholds for warming-induced growth decline at elevational tree line in the
Yukon Territory, Canada. Global Biogeochemical Cycles, 18(3).
DesRochers, A., & Gagnon, R. (1997). Is ring count at ground level a good estimation of black
spruce age? Canadian Journal of Forest Research, 27(8), 1263–1267.
Dittmar, C., Zech, W., & Elling, W. (2003). Growth variations of Common beech (Fagus sylvatica
L.) under different climatic and environmental conditions in Europe—a dendroecological
study. Forest Ecology and Management, 173(1), 63–78. https://doi.org/10.1016/S0378-
DRIVERS OF BOREAL FOREST GROWTH 42
D’Orangeville, L., Houle, D., Duchesne, L., Phillips, R. P., Bergeron, Y., & Kneeshaw, D. (2018).
Beneficial effects of climate warming on boreal tree growth may be transitory. Nature
Communications, 9(1), 3213. https://doi.org/10.1038/s41467-018-05705-4
Dormann, C. F., Elith, J., Bacher, S., Buchmann, C., Carl, G., Carré, G., … Lautenbach, S. (2013).
Collinearity: a review of methods to deal with it and a simulation study evaluating their
performance. Ecography, 36(1), 27–46. https://doi.org/10.1111/j.1600-
Dulamsuren, C., Hauck, M., Kopp, G., Ruff, M., & Leuschner, C. (2017). European beech responds
to climate change with growth decline at lower, and growth increase at higher
elevations in the center of its distribution range (SW Germany). Trees, 31(2), 673–686.
Ford, K. R., Breckheimer, I. K., Franklin, J. F., Freund, J. A., Kroiss, S. J., Larson, A. J., …
HilleRisLambers, J. (2016). Competition alters tree growth responses to climate at
individual and stand scales. Canadian Journal of Forest Research, 47(1), 53–62.
Fortin, M., Bernier, S., Saucier, J.-P., & Labbé, F. (2009). Une relation hauteur-diamètre tenant
compte de l’influence de la station et du climate pour 20 espèces commerciales du
Québec (No. 153; p. 41). Québec (Province): Direction de la Recherche Forestière.
Fritts, H. C. (1971). Dendroclimatology and Dendroecology. Quaternary Research, 1(4), 419–449.
Gauthier, S., Bernier, P., Kuuluvainen, T., Shvidenko, A. Z., & Schepaschenko, D. G. (2015). Boreal
forest health and global change. Science, 349(6250), 819–822.
DRIVERS OF BOREAL FOREST GROWTH 43
Gennaretti, F., Arseneault, D., Nicault, A., Perreault, L., & Bégin, Y. (2014). Volcano-induced
regime shifts in millennial tree-ring chronologies from northeastern North America.
Proceedings of the National Academy of Sciences of the United States of America,
111(28), 10077–10082. https://doi.org/10.1073/pnas.1324220111
Gewehr, S., Drobyshev, I., Berninger, F., & Bergeron, Y. (2014). Soil characteristics mediate the
distribution and response of boreal trees to climatic variability. Canadian Journal of
Forest Research, 44, 487–498. https://doi.org/10.1139/cjfr-2013-0481
Girardin, M. P., Bernier, P. Y., Raulier, F., Tardif, J. C., Conciatori, F., & Guo, X. J. (2011). Testing
for a CO2 fertilization effect on growth of Canadian boreal forests. Journal of
Geophysical Research: Biogeosciences, 116(G1), G01012.
Girardin, M. P., Bouriaud, O., Hogg, E. H., Kurz, W., Zimmermann, N. E., Metsaranta, J. M., …
Bhatti, J. (2016). No growth stimulation of Canada’s boreal forest under half-century of
combined warming and CO2 fertilization. Proceedings of the National Academy of
Sciences of the United States of America, 113(52), E8406–E8414.
Girardin, M. P., Guo, X. J., De Jong, R., Kinnard, C., Bernier, P., & Raulier, F. (2014). Unusual
forest growth decline in boreal North America covaries with the retreat of Arctic sea ice.
Global Change Biology, 20(3), 851–866. https://doi.org/10.1111/gcb.12400
Gleason, K. E., Bradford, J. B., Bottero, A., D’Amato, A. W., Fraver, S., Palik, B. J., … Kern, C. C.
(2017). Competition amplifies drought stress in forests across broad climatic and
compositional gradients. Ecosphere, 8(7), e01849. https://doi.org/10.1002/ecs2.1849
Gouvernement du Québec. (2003). Vegetation Zones and Bioclimatic Domains in Québec.
DRIVERS OF BOREAL FOREST GROWTH 44
Granda, E., & Camarero, J. J. (2017). Drought reduces growth and stimulates sugar
accumulation: new evidence of environmentally driven non-structural carbohydrate use.
Tree Physiology, 37(8), 997–1000. https://doi.org/10.1093/treephys/tpx097
Gu, L., Hanson, P. J., Post, W. M., Kaiser, D. P., Yang, B., Nemani, R., … Meyers, T. (2008). The
2007 Eastern US spring freeze: Increased cold damage in a warming world. BioScience,
58(3), 253–262. https://doi.org/10.1641/B580311
Hartmann, H. (2011). Will a 385 million year-struggle for light become a struggle for water and
for carbon? – How trees may cope with more frequent climate change-type drought
events. Global Change Biology, 17(1), 642–655. https://doi.org/10.1111/j.1365-
Hasler, A., Geertsema, M., Foord, V., Gruber, S., & Noetzli, J. (2015). The influence of surface
characteristics, topography and continentality on mountain permafrost in British
Columbia. The Cryosphere, 9(3), 1025–1038. https://doi.org/10.5194/tc-9-1025-2015
Hember, R. A., Kurz, W. A., & Coops, N. C. (2016). Relationships between individual-tree
mortality and water-balance variables indicate positive trends in water stress-induced
tree mortality across North America. Global Change Biology, 23(4), 1691–1710.
Hember, R. A., Kurz, W. A., & Coops, N. C. (2017). Increasing net ecosystem biomass production
of Canada’s boreal and temperate forests despite decline in dry climates. Global
Biogeochemical Cycles, 31(1), 2016GB005459. https://doi.org/10.1002/2016GB005459
Holmes, R. L. (1983). Computer assisted quality control in tree-ring dating and measurement.
Retrieved from https://www.scienceopen.com/document?vid=a110a27a-36fa-41dd-
DRIVERS OF BOREAL FOREST GROWTH 45
Housset, J. M., Carcaillet, C., Girardin, M. P., Xu, H., Tremblay, F., & Bergeron, Y. (2016). In situ
comparison of tree-ring responses to climate and population genetics: The need to
control for local climate and site variables. Frontiers in Ecology and Evolution, 4.
Huber, P., J. (2005). Generalities. In Robust Statistics (pp. 1–19).
IPCC. (2013). Climate Change 2013 : The physical science basis. Contribution of Working Group I
to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change. (p.
Jarvis, A., Reuter, H., Nelson, A., & Guevara, E. (2008, January 1). Hole-filled seamless SRTM data
v4. Retrieved from http://srtm.csi.cgiar.org
Jiang, X., Huang, J.-G., Cheng, J., Dawson, A., Stadt, K. J., Comeau, P. G., & Chen, H. Y. H. (2018).
Interspecific variation in growth responses to tree size, competition and climate of
western Canadian boreal mixed forests. Science of The Total Environment, 631–632,
Johnstone, J. F., McIntire, E. J. B., Pedersen, E. J., King, G., & Pisaric, M. J. F. (2010). A sensitive
slope: estimating landscape patterns of forest resilience in a changing climate.
Ecosphere, 1(6), art14. https://doi.org/10.1890/ES10-00102.1
Ju, J., & Masek, J. G. (2016). The vegetation greenness trend in Canada and US Alaska from
1984–2012 Landsat data. Remote Sensing of Environment, 176, 1–16.
Juday, G. P., & Alix, C. (2012). Consistent negative temperature sensitivity and positive influence
of precipitation on growth of floodplain Picea glauca in Interior Alaska. Canadian Journal
of Forest Research, 42(3), 561–573. https://doi.org/10.1139/x2012-008
DRIVERS OF BOREAL FOREST GROWTH 46
Kidd, K. R., Copenheaver, C. A., & Zink-Sharp, A. (2014). Frequency and factors of earlywood
frost ring formation in jack pine (Pinus banksiana) across northern lower Michigan.
Écoscience, 21(2), 157–167. https://doi.org/10.2980/21-2-3708
Klesse, S., DeRose, R. J., Guiterman, C. H., Lynch, A. M., O’Connor, C. D., Shaw, J. D., & Evans, M.
E. K. (2018). Sampling bias overestimates climate change impacts on forest growth in
the southwestern United States. Nature Communications, 9(1), 5336.
Kurz, W. A., Shaw, C. H., Boisvenue, C., Stinson, G., Metsaranta, J., Leckie, D., … Neilson, E. T.
(2013). Carbon in Canada’s boreal forest — A synthesis. Environmental Reviews, 21(4),
Lavoie, M., Harper, K., Paré, D., & Bergeron, Y. (2007). Spatial pattern in the organic layer and
tree growth: A case study from regenerating Picea mariana stands prone to
paludification. Journal of Vegetation Science, 18(2), 213–222.
Létourneau, J. P., Matejek, S., Morneau, C., Robitaille, A., Roméo, T., Brunelle, J., & Leboeuf, A.
(2008). Norme de cartographie écoforestière du Programme d’inventaire écoforestier
nordique. Québec: Ministère des Ressources naturelles et de la Faune du Québec.
Lloyd, A. H., Bunn, A. G., & Berner, L. (2011). A latitudinal gradient in tree growth response to
climate warming in the Siberian taiga. Global Change Biology, 17(5), 1935–1945.
Loader, N. J., McCarroll, D., Gagen, M., Robertson, I., & Jalkanen, R. (2007). Extracting climatic
information from stable isotopes in tree rings. In Stable Isotopes as Indicators of
Ecological Change: Vol. 1. Terrestrial Ecology (pp. 25–48).
DRIVERS OF BOREAL FOREST GROWTH 47
Magnani, F., Mencuccini, M., & Grace, J. (2000). Age-related decline in stand productivity: the
role of structural acclimation under hydraulic constraints. Plant, Cell & Environment,
23(3), 251–263. https://doi.org/10.1046/j.1365-3040.2000.00537.x
Mangiafico, S. (2017). rcompanion: Functions to Support Extension Education Program
Evaluation (Version 1.11.1). Retrieved from https://cran.r-
Manrique-Alba, À., Ruiz-Yanetti, S., Moutahir, H., Novak, K., De Luis, M., & Bellot, J. (2017). Soil
moisture and its role in growth-climate relationships across an aridity gradient in
semiarid Pinus halepensis forests. Science of The Total Environment, 574, 982–990.
Matías, L., Linares, J. C., Sánchez-Miranda, Á., & Jump, A. S. (2017). Contrasting growth forecasts
across the geographical range of Scots pine due to altitudinal and latitudinal differences
in climatic sensitivity. Global Change Biology, 23(10), 4106–4116.
McLaughlin, S. B., Downing, D. J., Blasing, T. J., Cook, E. R., & Adams, H. S. (1987). An analysis of
climate and competition as contributors to decline of red spruce in high elevation
Appalachian forests of the Eastern United states. Oecologia, 72(4), 487–501.
Minasny, B., & McBratney, A. B. (2017). Limited effect of organic matter on soil available water
capacity. European Journal of Soil Science, 69(1), 39–47.
Ministère des Ressources Naturelles du Québec. (2008). Norme d’inventaire écodendrométrique
nordique (p. 245). Direction des inventaires forestiers.
DRIVERS OF BOREAL FOREST GROWTH 48
Molinari, C., Lehsten, V., Blarquez, O., Carcaillet, C., Davis, B. A. S., Kaplan, J. O., … Bradshaw, R.
H. W. (2018). The climate, the fuel and the land use: Long-term regional variability of
biomass burning in boreal forests. Global Change Biology, 0(0).
Montwé, D., Isaac-Renton, M., Hamann, A., & Spiecker, H. (2018). Cold adaptation recorded in
tree rings highlights risks associated with climate change and assisted migration. Nature
Communications, 9(1), 1574. https://doi.org/10.1038/s41467-018-04039-5
Nicklen, E. F., Roland, C. A., Csank, A. Z., Wilmking, M., Ruess, R. W., & Muldoon, L. A. (2018).
Stand basal area and solar radiation amplify white spruce climate sensitivity in interior
Alaska: Evidence from carbon isotopes and tree rings. Global Change Biology, 0(0), 1–16.
Nicklen, E. F., Roland, C. A., Ruess, R. W., Schmidt, J. H., & Lloyd, A. H. (2016). Local site
conditions drive climate–growth responses of Picea mariana and Picea glauca in interior
Alaska. Ecosphere, 7(10), e01507. https://doi.org/10.1002/ecs2.1507
Oksanen, J., Blanchet, F. G., Friendly, M., Kindt, R., Legendre, P., McGlinn, D., … Wagner, H.
(2018). vegan: Community Ecology Package (Version 2.5-2). Retrieved from
Ols, C., Girardin, M. P., Hofgaard, A., Bergeron, Y., & Drobyshev, I. (2017). Monitoring climate
sensitivity shifts in tree-rings of eastern boreal North America using model-data
comparison. Ecosystems, 1–16. https://doi.org/10.1007/s10021-017-0203-3
Ols, C., Trouet, V., Girardin, M. P., Hofgaard, A., Bergeron, Y., & Drobyshev, I. (2018). Post-1980
shifts in the sensitivity of boreal tree growth to North Atlantic Ocean dynamics and
seasonal climate. Global and Planetary Change, 165, 1–12.
DRIVERS OF BOREAL FOREST GROWTH 49
Osorio, F., Vallejos, R., Cuevas, F., & Mancilla, D. (2018). SpatialPack: Tools for assessment of the
association between two spatial processes (Version 0.3). Retrieved from
Pepin, N., Bradley, R. S., Diaz, H. F., Caceres, E. B., Forsythe, N., Fowler, H., … Yang, D. Q. (2015).
Elevation-dependent warming in mountain regions of the world. Nature Climate
Change, 5, 424–430.
Peres-Neto, P. R., Legendre, P., Dray, S., & Borcard, D. (2006). Variation partitionning of species
data matrices: estimation and comparison of fractions. Ecology, 87(10), 2614–2625.
Piovesan, G., Biondi, F., Filippo, A. D., Alessandrini, A., & Maugeri, M. (2008). Drought-driven
growth reduction in old beech (Fagus sylvatica L.) forests of the central Apennines, Italy.
Global Change Biology, 14(6), 1265–1281. https://doi.org/10.1111/j.1365-
Price, D. T., Alfaro, R. I., Brown, K. J., Flannigan, M. D., Fleming, R. A., Hogg, E. H., … Venier, L. A.
(2013). Anticipating the consequences of climate change for Canada’s boreal forest
ecosystems. Environmental Reviews, 21(4), 322–365. https://doi.org/10.1139/er-2013-
Price, D. T., McKenney, D. W., Joyce, L. A., Siltanen, R. M., Papadopol, P., & Lawrence, K. (2011).
High-resolution interpolation of climate scenarios for Canada derived from general
circulation model simulations. Information Report NOR-X-421. Edmonton, AB.: Natural
Resources Canada, Canadian Forest Service, Northern Forestry Centre. 104 P. Retrieved
Richardson, A. D., Hufkens, K., Milliman, T., Aubrecht, D. M., Furze, M. E., Seyednasrollah, B., …
Hanson, P. J. (2018). Ecosystem warming extends vegetation activity but heightens
DRIVERS OF BOREAL FOREST GROWTH 50
vulnerability to cold temperatures. Nature, 560(7718), 368–371.
Robitaille, A., Saucier, J.-P., Chabot, M., Côté, D., & Boudreault, C. (2015). An approach for
assessing suitability for forest management based on constraints of the physical
environment at a regional scale. Canadian Journal of Forest Research, 45(5), 529–539.
Ryan, M. G., Phillips, N., & Bond, B. J. (2006). The hydraulic limitation hypothesis revisited. Plant,
Cell & Environment, 29(3), 367–381. https://doi.org/10.1111/j.1365-3040.2005.01478.x
Ryan, M. G., & Yoder, B. J. (1997). Hydraulic Limits to Tree Height and Tree Growth. BioScience,
47(4), 235–242. https://doi.org/10.2307/1313077
Sala, A., Woodruff, D. R., & Meinzer, F. C. (2012). Carbon dynamics in trees: feast or famine?
Tree Physiology, 32(6), 764–775. https://doi.org/10.1093/treephys/tpr143
Santander Meteorology Group. (2012). fume: FUME package (Version R package version 1.0).
Retrieved from https://CRAN.R-project.org/package=fume
Schuster, R., & Oberhuber, W. (2013). Age-dependent climate-growth relationships and
regeneration of Picea abies in a drought-prone mixed-coniferous forest in the Alps.
Canadian Journal of Forest Research, 43, 609–618. https://doi.org/10.1139/cjfr-2012-
Strong, W. L., & La Roi, G. H. (1983). Root-system morphology of common boreal forest trees in
Alberta, Canada. Canadian Journal of Forest Research, 13(6), 1164–1173.
ter Braak, C., & Smilauer, P. (2009). Canoco for Windows Version 4.56. Biometris - Plant
DRIVERS OF BOREAL FOREST GROWTH 51
Trenberth, K. E. (1983). What are the Seasons? Bulletin of the American Meteorological Society,
64(11), 1276–1282. https://doi.org/10.1175/1520-0477(1983)064<1276:WATS>2.0.CO;2
Trugman, A. T., Medvigy, D., Anderegg, W. R. L., & Pacala, S. W. (2018). Differential declines in
Alaskan boreal forest vitality related to climate and competition. Global Change Biology,
24(3), 1097–1107. https://doi.org/10.1111/gcb.13952
Tukey, J. W. (1977). Exploratory data analysis (Addison-Wesley). Reading, MA.
Vaganov, E. A., Hughes, M. K., Kirdyanov, A. V., Schweingruber, F. H., & Silkin, P. P. (1999).
Influence of snowfall and melt timing on tree growth in subarctic Eurasia. Nature,
400(6740), 149–151. https://doi.org/10.1038/22087
Vitasse, Y., Signarbieux, C., & Fu, Y. H. (2017). Global warming leads to more uniform spring
phenology across elevations. Proceedings of the National Academy of Sciences of the
United States of America, 201717342. https://doi.org/10.1073/pnas.1717342115
Voortman, B. R., Bartholomeus, R. P., Bodegom, P. M. van, Gooren, H., Zee, S. E. A. T. M. van
der, & Witte, J.-P. M. (2013). Unsaturated hydraulic properties of xerophilous mosses:
towards implementation of moss covered soils in hydrological models. Hydrological
Processes, 28(26), 6251–6264. https://doi.org/10.1002/hyp.10111
Walker, X. J., & Johnstone, J. F. (2014). Widespread negative correlations between black spruce
growth and temperature across topographic moisture gradients in the boreal forest.
Environmental Research Letters, 9(6), 064016. https://doi.org/10.1088/1748-
Weber, P., Bugmann, H., Fonti, P., & Rigling, A. (2008). Using a retrospective dynamic
competition index to reconstruct forest succession. Forest Ecology and Management,
254(1), 96–106. https://doi.org/10.1016/j.foreco.2007.07.031
DRIVERS OF BOREAL FOREST GROWTH 52
Wilmking, M., & Myers-Smith, I. (2008). Changing climate sensitivity of black spruce (Picea
mariana Mill.) in a peatland–forest landscape in Interior Alaska. Dendrochronologia,
25(3), 167–175. https://doi.org/10.1016/j.dendro.2007.04.003
Wu, X., Li, X., Liu, H., Ciais, P., Li, Y., Xu, C., … Zhang, C. (2018). Uneven winter snow influence on
tree growth across temperate China. Global Change Biology.
Yue, S., & Wang, C. (2004). The Mann-Kendall test modified by effective sample size to detect
trend in serially correlated hydrological series. Water Resources Management, 18(3),
Zhang, K., Kimball, J. S., Hogg, E. H., Zhao, M., Oechel, W. C., Cassano, J. J., & Running, S. W.
(2008). Satellite-based model detection of recent climate-driven changes in northern
high-latitude vegetation productivity. Journal of Geophysical Research: Biogeosciences,
113(G3), G03033. https://doi.org/10.1029/2007JG000621
Zhu, Z., Piao, S., Myneni, R. B., Huang, M., Zeng, Z., Canadell, J. G., … Zeng, N. (2016). Greening
of the Earth and its drivers. Nature Climate Change, 6(8), 791–795.
Zobel, M. (1990). Soil oxygen conditions in paludifying boreal forest sites. Suo, 41(4–5), 81–89.