Drivers of forest change in the Greater Yellowstone Ecosystem

Abstract

Questions
Global climate change is predicted to cause widespread shifts in the distribution and composition of forests, particularly in mountain environments where climate exerts strong controls on tree community arrangement. The upslope movement of vegetation has been observed in association with warming temperatures and is especially evident in ecotones—the transition zones between vegetation types. We explored the role of drought and tree mortality on recent changes in high‐elevation forests.

Location
Greater Yellowstone Ecosystem, USA.

Methods
We established 19 forest demography plots along an elevational gradient spanning dominant high‐elevation vegetation types.

Results
Tree establishment dates indicated the upslope movement of Pinus albicaulis (whitebark pine) treeline and ecotone shift from meadow to forest starting in the 1950s. An expansion of the growing season likely contributed to the upward expansion of the treeline. Comparisons between overstory and understory tree composition suggested ongoing succession in the absence of fire at lower elevations, namely the replacement of Pinus contorta (lodgepole pine) by Abies lasiocarpa (subalpine fir). P. contorta seedlings were distributed at higher elevations than overstory trees of the same species, suggesting some potential for upslope movement with warming conditions; P. albicaulis seedlings, conversely, were distributed throughout all elevations of the transect. Significant tree mortality occurred in Pinus spp. and disproportionately affected P. albicaulis , as a result of a regional Dendroctonus ponderosae (mountain pine beetle) outbreak (2008–2012). Mortality events were strongly associated with drier than average conditions 2–3 years prior to tree death.

Conclusion
Rising sensitivity to arid conditions in the mid‐20th century amid already dense, aging forests appears to have increased susceptibility to beetle‐induced mortality during the most recent drought. Tree species in the study area responded individually to global change stressors, which acted on these forests in complex ways and led to both ecotone shifts and stability. This work highlights the interplay between succession, forest disturbances and climate‐related growth responses in driving forest compositional change in subalpine and treeline environments.

Full text

J Veg Sci. 2022;33:1–14.
|
1
Journal of Vegetation Science
wileyonlinelibrary.com/journal/jvs
Received: 15 December 2021
|
Revised: 25 May 2022
|
Accepted: 2 June 2022
DOI : 10.1111/j vs.13141
RESEARCH ARTICLE
Drivers of forest change in the Greater Yellowstone Ecosystem
Erika M. Blomdahl1 | James H. Speer2 | Margot Kaye3 | Nicole E. Zampieri4 |
Maegen Rochner5 | Bryce Currey6 | Denise Alving3 | Gabriel D. Cahalan7 |
Ben Hagedorn8 | Hang Li2 | Rose Oelkers9 | Lissa Pelletier10 | Ichchha Thapa2 |
Kevin Willson11 | Brian D. Woodward12 | R. Justin DeRose1
© 2022 International Ass ociation for Vegetation Science.
1Depar tment of Wildland Resou rces and
Ecology Center, Uta h State University,
Logan, U tah, USA
2Depar tment of Eart h and Environmental
Systems, Indiana State Uni versit y, Terre
Haute, In diana, USA
3Depar tment of Ecosystem Scien ce and
Managem ent, Penn State University,
University Park, Pennsylvania, USA
4Depar tment of Geography, Florida State
University, Tallahass ee, Florida, US A
5Depar tment of Geographic an d
Environmental S cience s, Universit y of
Louisville, Lo uisville, Kentu cky, USA
6Depar tment of Land Resources and
Environmental S cience s, Mont ana State
University, Bozeman, Mont ana, USA
7The Nature Conse rvan cy, Arlington,
Texas, USA
8Washington State D epar tment of N atural
Resources, Oly mpia, Washington, USA
9Lamont- Doherty Earth Observatory,
Columbia Univer sity, Palisades, New York,
USA
10Depa rtme nt of Sust ainab le Resources
Managem ent, SU NY College of
Environmental S cience a nd Fores try,
Syracus e, New York, USA
11Depar tment of Biolog y, University of
New Mexico, Albuquerqu e, New Mexico,
USA
12Depar tment of Ecosystem Scien ce and
Sustainability, Colorado State University,
Fort Collins, Colorado, USA
Correspondence
Erika M. B lomda hl, Dep artm ent of
Wildland Resou rces and Ecology Center,
Utah State University, 5230 Old Main Hill,
Logan, U tah 84 322- 5230.
Email: erika.blomdahl@usu.edu
Abstract
Questions: Global climate change is predicted to cause widespread shifts in the dis-
tribution and composition of forests, particularly in mountain environments where
climate exerts strong controls on tree community arrangement. The upslope move-
ment of vegetation has been observed in association with warming temperatures and
is especially evident in ecotones— the transition zones between vegetation types. We
explored the role of drought and tree mortality on recent changes in high- elevation
forests.
Location: Greater Yellowstone Ecosystem, USA.
Methods: We established 19 forest demography plots along an elevational gradient
spanning dominant high- elevation vegetation types.
Results: Tree establishment dates indicated the upslope movement of Pinus albicaulis
(whitebark pine) treeline and ecotone shift from meadow to forest starting in the
1950s. An expansion of the growing season likely contributed to the upward expan-
sion of the treeline. Comparisons between overstory and understory tree composi-
tion suggested ongoing succession in the absence of fire at lower elevations, namely
the replacement of Pinus contorta (lodgepole pine) by Abies lasiocarpa (subalpine fir).
P. contorta seedlings were distributed at higher elevations than overstory trees of the
same species, suggesting some potential for upslope movement with warming condi-
tions; P. albicaulis seedlings, conversely, were distributed throughout all elevations of
the transect. Significant tree mortality occurred in Pinus spp. and disproportionately
affected P. albicaulis, as a result of a regional Dendroctonus ponderosae (mountain pine
beetle) outbreak (2008– 2012). Mortality events were strongly associated with drier
than average conditions 2– 3 years prior to tree death.
Conclusion: Rising sensitivity to arid conditions in the mid- 20th centur y amid already
dense, aging forests appears to have increased susceptibility to beetle- induced mor-
tality during the most recent drought. Tree species in the study area responded indi-
vidually to global change stressors, which acted on these forests in complex ways and
led to both ecotone shifts and stability. This work highlights the interplay between

2
|
Journal of Vegetation Science BLOMDAHL et al.
1 | INTRODUC TION
Ongoing climate change has resulted in novel temperature
gradients, modified resource availability, and altered distur-
bance regimes in forested systems across the world (Anderegg
et al., 2013; Allen et al., 2015). In high- elevation mixed- conifer
forests of North America, mortality events have occurred with
increased magnitude and frequency in recent decades (Loehman
et al., 2018). Although multiple hypotheses exist about the causes
of this mor tality (Trugman et al., 2021), the specific drivers are
likely a complex interac tion among temperature stress, moisture
stress, and disturbance agents (e.g., insect outbreaks, fire) acting
on older and denser forests (Allen et al., 2010; Rocca et al., 2 014).
Mortality that results from these interactions ranges in intensit y
from individual trees to entire stands, which can alter forest dy-
namics depending on stand structure and composition. Increased
mortality can also create opportunities for regeneration, migra-
tion, and colonization of forest species acro ss spatial scales (Brice
et al., 2 019).
When new colonization opportunities are presented (e.g.,
disturbances, tree mortality), a reshuffling of tree species com-
position could occur (Bell et al., 2014). Models have predicted veg-
etation shifts poleward and up elevational gradients (Iverson &
McKenzie, 2013), indicating that many vegetation t ypes may experi-
ence type conversions as temperatures warm. Examples of poleward
and upslope movement of vegetation have already been observed
in some forested systems (Johnstone & Chapin, 2003; Beckage
et al., 2008; Breshears et al., 2008; Smithers et al., 2018). Given that
forests are slowly but continuously changing (Christensen, 2014),
identif ying climate- driven community shifts of long- lived species is
a challenging task.
Shifts in forest composition are thought to be especially evident
in ecotones— transition zones between vegetation types (Hufkens
et al., 2009)— particularly in mountainous environments where cli-
mate can act as a strong control on tree community arrangement
(Smith et al., 2009). In the central Rocky Mountains, the Greater
Yellowstone Ecosystem (GYE) includes thousands of hectares of
subalpine forest and alpine environments. These systems are char-
acterized by a suite of ecologic al legacies driven by climate (Krause
& Whitlock, 2017), disturbance dynamics (Romme & Despain, 1989;
Hatala et al., 2010), and competitive interactions (Tomback
et al., 2001a) that act as strong filters for tree species composition.
Ecotone shifts have already been observed in some lower- elevation
forests of the GYE (Donato et al., 2016), but the role of climate
change and disturbance in these shifts remains unresolved.
The principal disturbance agents in the GYE are bark beetle out-
breaks, wildfire and drought. In the early 2000s there was a wide-
spread Dendroctonus ponderosae (mountain pine beetle) outbreak in
the region, affecting Pinus albicaulis (whitebark pine) in particular,
with nearly half of the GYE population estimated to have severe tree
mortality (Macfarlane et al., 2013). Although large D. ponderosae o u t-
breaks are a cyclical occurrence in the GYE, the extent and severity
of this most recent outbreak was likely amplified by climate change.
In high- elevation environments, warming temperatures interact
with bark beetle dynamics by increasing overwintering survival and
larval development rates (Bentz et al., 2010 ). Water- stressed trees
are less able to defend against beetle attack via two main mecha-
nisms: (1) less allocation of secondary metabolites to defense, and (2)
less hydraulic pressure to pitch out beetles (Franceschi et al., 2005;
Anderegg et al., 2015). The period from 2000 to 2010 has been
termed a “mega- drought” of likely unprecedented severity in the
Upper Missouri River Basin, reflecting more arid conditions and re-
duced snowpack in its headwaters, the Rocky Mountains (Martin
et al., 2020). In addition to amplifying bark beetle ac tivit y, increased
aridity interacts with wildfire frequency and severity. Historical fire
regimes in subalpine forests of the interior Rockies are infrequent
and mixed- to high- severity, driven by periods of prolonged drought
sufficient enough to dry long- term fuel accumulations (Schoennagel
et al., 2004). Fire suppression has altered more historically mixed-
severity forest types in the GYE and has likely contributed to the
decline of P. albicaulis forests (Tomback et al., 2001a).
In this study, we investigated changes in overstory and under-
story forest composition and structure (i.e., for tree species only)
across a 50 0- m elevational gradient of common forest types in
the GYE to determine how and why high- elevation ecotones have
changed over the past several decades. Our objective was to char-
acterize the role of drought and recent beetle- caused mortalit y on
possible changes in tree species composition and structure, and
determine whether those changes reflect ecotone shifts, succes-
sional change or some combination of both. We analyzed forest and
dendrochronological data characterizing species- specific demogra-
phy to identify species distribution changes and ecotone shifts, ex-
pecting upslope movement across all elevations and species, and in
particular among more drought- sensitive species. We further inves-
tigated climate- growth relationships and patterns in tree mortality
to help explain observed compositional shifts.
Funding information
Utah State University ; Utah A gricultural
Exper iment St ation; National Science
Foundation
succession, forest disturbances and climate- related growth responses in driving for-
est compositional change in subalpine and treeline environments.
KEYWORDS
climate change, dendrochronology, ecotone shift, mountain pine beetle, whitebark pine
Co- ordinating Editor: Ant onio Gazol

|
3
Journal of Vegetation Science
BLOMDAHL et al.
2 | METHODS
2.1 | Study area
We se lecte d a fo r este d slo p e on th e sout hwe s t aspec t of Sou t h Bi rd
Mountain in the Shoshone National Forest in northwest Wyoming,
as our study area (Figure 1). The climate in the region is character-
ized by a mean annual temperature of 1.2°C, with mean minimum
and maximum monthly temperatures of −14.5°C and 22.5°C re-
spectively, and a mean annual precipitation of 775.1 mm (extrac ted
from climateWNA, Wang et al., 2016). The predominant tree spe-
cies in the study area are Pinus contorta (lodgepole pine) and Pinus
albicaulis, with a lesser component of Abies lasiocarpa (subalpine fir),
Picea engelmannii (Engelmann spruce), and Pseudotsuga menziesii
va r. glauca (Douglas fir). The transect s sampled during the study
spanned the following ecotones from lowest (2,558 m) to highest
(3,028 m) elevation: sagebrush steppe, P. contorta- dominated for-
est, P. albicaulis- dominated forest, and alpine meadow. P. contorta-
dominated forests initiated in the early- to- mid- 19th century, and
could be described as late- successional with complex structure
and advanced regeneration of shade- tolerant A. lasiocarpa in the
understory. P. albicaulis- dominated forests initiated in the early
19th century and progressively earlier with higher elevations, with
early- successional forests above 3,000 m.
FIGURE 1 Location of the study transect in relation to Grand Teton National Park (GTNP), Yellowstone National Park (YNP), and the
Greater Yellowstone Ecosystem (GYE). The transect spans a 50 0- m elevational gradient and was sampled over a 3- year period, from 2017 to
20 19.

4
|
Journal of Vegetation Science BLOMDAHL et al.
The alpine meadows were sparsely populated with large, dead
remnant P. albicaulis and P. engelmannii stems dating to many centu-
ries ago (Rochner et al., 2021).
2.2 | Study design and data collection
To assess changes in forest composition, we established two par-
allel study transec ts on South Bird Mountain. We used aerial im-
agery to identify a transec t location that appeared representative
of the subalpine forests of the GYE and included an elevational
gradient of multiple ecotones. We selected a living forested area,
thus our scope of inference does not include the post- disturbance
dynamic s of those “ghost” forests severely impacted by the bark
beetle outbreaks in the early 200 0s. The first transect consisted
of six plots, ranging in elevation from 2,866 to,3006 m, whereas
the second transect consisted of 13 plot s, ranging in elevation
from 2,561 to 3,020 m (Figure 1). Plots were spaced 250 m apar t
across the elevational gradient within each transect. We sampled
the fir st trans ec t in 2017 and the second in 2018 an d 2019. Of th e
19 plots, two were in meadows at the elevational extremes of the
transect.
At each forested plot, we measured two forest demographic
groups: (1) trees, and (2) seedlings and saplings. Trees were defined
as stems >5 cm diameter at coring height (DCH) and >1.37 m in
height. Seedlings and saplings were defined as stems <5 cm DCH,
with saplings >1.37 m in height. We applied N- tree distance sam-
pling (Moore, 195 4), in which we sampled the 10 trees (live or dead)
nearest to the plot center. We established each plot radius by mea-
suring from plot center to half the distance between the 10th and
11th trees. Across the dataset, the average forested plot radius was
5.3 m, with radii ranging from 3.8 m to 7.4 m. We identified spe-
cies and status (e.g., live or dead) for all trees sampled, and recorded
observations for canopy position (dominant, co- dominant, and sup-
pressed) and tree condition (e.g., evidence of bark beetles, fungal
fruiting bodies, physical damage). We used increment borers to col-
lect two cores per tree at a coring height of 3 0 cm. For dead trees
that were not sound enough to core, we collected cross- sections.
We identified and destruc tively sampled all seedlings and saplings
within the plot radius determined as above.
2.3 | Sample preparation and tree- ring
chronologies
We processed increment cores and cross- sections from overstory
trees according to standard dendrochronological methods de-
scribed by Stokes and Smiley (1968) and Speer (2010). Cores were
mounted and sanded using progressively finer grit (40, 120, 220,
320, and 40 0) and finished with 30, 15, and 9 μm sanding film
until cell structure was discernible. We developed skeleton plots
for a subset of individual series to identify marker years for each
species. We then used the memorization method to cross- date the
remaining cores (Douglass, 1941). Tree- ring widths were meas-
ured (resolution: 0.001 mm) using a Velmex TA Measuring Machine
with J2X software, and via scanned images (1,20 0 dpi) processed
with CooRecorder (Cybis Elektronik, 2010). We statistically vali-
dated the visually cross- dated cores with COFECHA software
(Holmes, 1983). Of the five species present on the study transect,
increment core sample depth was large enough to develop final
tree- ring chronologies for P. albicaulis (142 series from 89 trees)
and P. contorta (121 series from 66 trees). The expressed popu-
lation signal for the chronologies used in climatic analyses were
0.92 and 0.89 for the P. albicaulis and P. contorta, respectively
(Appendix S1). We used an age- specific smoothing spline with a
fixed stiffness of 30 years to detrend the ontogenetic growth pat-
terns (Klesse, 2021), and autoregressive modeling to remove tem-
poral autocorrelation. All chronology building was done in the R
package dplR (Bunn, 2008).
To age each of the harvested seedlings and saplings >30 cm tall,
we prepared two cross- sections from each sample, one at the base
(0 cm height) and another at 30 cm stem height . Seedlings smaller
than ~1 cm in diameter were cut with a razor blade and the rings
were visually counted under a microscope. Cross- sections of seed-
lings 1– 5 cm in diameter were sanded, rings were counted under the
microscope, and the memorization method was used to cross- date
tree rings when possible.
2.4 | Analytical approach
2.4.1 | Forest demography
Live and dead trees >5 cm DCH (i.e., overstory) were assigned a plot
scaling factor based on the radius calculated from the N- tree design.
Live and dead total basal area, trees per hectare, QMD, and stand
density index (SDI; Vacchiano et al., 2013) were calculated. Trees per
hectare for the understory trees were calculated on a per- species
basis using the N- tree scaling factor.
Overstory tree cores that intersected the pith were noted,
otherwise the number of rings to the pith were estimated using
pith locators developed by Applequist (1958). We used the seed-
ling and sapling cross- sections taken at 0 and 30 cm to develop an
equation for extrapolating the number of years for a seedling to
grow from 0 to 30 cm. We estimated tree establishment year by
subtracting the modeled number of years to grow to 30 cm from
the pith year measured at 30 cm measurement height. The estab-
lishment yea rs of se edlings wer e the pith yea r of the cross- s ec ti on
at 0 cm.
Dates of tree death were assumed to be the last calendar year in
which a standing dead tree formed a tree ring. In cases in which the
latest calendar year did not match across the two core samples for
a given tree, the date of death was assigned the most recent year.
In addition, bark beetles (assumed Dendroctonus ponderosae) were
ascribed as a factor associated with death if blue- stain fungus (as-
sumed one of, Grosmannia clavigera Robinson- Jeffrey and Davidson,

|
5
Journal of Vegetation Science
BLOMDAHL et al.
Ophiostoma montium Rumbold or Leptographium longiclavatum S.W.
Lee, J.J. Kim & C. Breuil) was present in the sapwood of increment
cores of standing dead trees.
2.4.2 | Forest change
We compared patterns of establishment by tree species and status
(e.g., live or dead) graphically. To aid graphical interpretation, we
binned establishment years into decades, and forested plots into
five elevational groups. Elevation bands were assigned by round-
ing to the nearest 100 m (e.g., “2,800 m” includes plots ranging from
2,750 to 2,849 m in elevation). To assess possible changes in the dis-
tribution and composition of overstory trees relative to understory
seedlings and saplings of the same species, we used Non- metric
Multidimensional (Distance) Scaling (NMDS; vegan package in R ver-
sion 2.4- 5; R Core Team, R Foundation for Statistical Computing,
Vienna, AT) on tree density (stems/ha) by species and form for the
17 forested plots. NMDS results were assessed for the overall re-
duction of stress. Ordination bi- plots were assessed graphically,
and the relationship between NMDS axes (expressed as MDS1 and
MDS2) and associated environmental variables were calculated. We
then plotted the relative amount of change for each species over the
NMDS axes.
2.4.3 | Climate response
To assess relationships between climate variables and the residual
chronologies we used response function analyses in the R pack-
age treeclim (Zang & Biondi, 2015). We explored monthly, seasonal
and water- year responses of tree- ring widths to precipit ation,
minimum temperature, maximum temperature, and the Palmer
Drought Severit y Index (PDSI) extracted from the Parameter-
elevation Regressions on Independent Slopes Model (PRISM Climate
Group, 2020). After preliminary testing, we settled on presenting
results for growth responses to monthly minimum and maximum
temperature and PDSI using a moving correlation analysis with a
35- year window. We settled on species- specific climate- grow th
patterns grouped across all elevations because they did not dif-
fer substantially when separated into upper and lower- elevation
groups, perhaps because the majority of stems for each species oc-
curred in a smaller elevational range than the whole transect at large
(Appendices S2 and S3).
To examine possible correspondence of tree mortality with
year- to- year drought variabilit y we conducted a superposed epoch
analysis (SEA) using tree death dates and historical time series
(1895– 2019) of summer seasonal drought (June– August) developed
from gridded PDSI data (PRISM Climate Group, 2020). We identified
16 unique mortality event years and used the “sea” function in the
R package burnr (Malevich et al., 2018) to analyze each focal year in
relation to the 10 years before and after to test the null hypothesis
that drought conditions in the years surrounding a mortality event
do not significantly differ from the mean drought conditions over
the period tested (1934– 2017).
3 | RESULTS
3.1 | Forest demography
We determined establishment dates for 91 Pinus albicaulis trees,
91 Pinus contorta trees, 5 Abies lasiocarpa trees, 2 Picea engelman-
nii trees, and 1 Pseudotsuga menziesii tree. Establishment dates of
live and dead overstory trees revealed clear successional pat-
terns of species recruitment at lower elevations along our study
transect, with stand initiation dominated by P. contorta typically
followed by establishment of P. albicaulis (largely after 1879;
Figure 2; Appendix S4). Basal area and stem densit y of P. con -
torta ranged from 2.9 to 51.3 m2/ha and from 163 to 720 stems/
ha respectively, with the greatest basal area around 2,70 0 m
and the greatest stem density around 2,800 m (Tables 1 and 2;
Appendices S5and S6). Basal area and stem density of P. albicau-
lis ranged from 1.9 to 23.9 m2/ha and from 29 to 992 stems/
ha respectively, with the greatest basal area around 2,90 0 m
and the greatest stem density around 3,0 00 m (Tables 1 and 2;
Appendices S5and S6). Stocking of P. albicaulis and P. contorta at
all plots was high, with SDI values ranging from 313.7 to 864.5
(Table 3; Appendix S7). A. lasiocarpa, P. engelmannii, and P. me n-
ziesii were only minor components of the overstor y at all but the
highest elevations (Tables 1 and 2). Tree establishment year was
positi ve ly re lated to e le vation, wi th th e old est tree s at the lowest
elevations and predominantly younger trees at higher elevations
(Figure 2; Appendix S4).
3.2 | Forest change and potential drivers
3.2.1 | Ecotone shift
Age structure and tree species composition of the forest across dif-
ferent elevation bands highlighted areas on the transect where un-
derstory communities did not reflect overstory composition. Across
the P. contorta- dominated stands, 8. 5% of total seedling density was
P. contorta, whereas 45.1% was A. lasiocarpa and 44. 2% was P. al -
bicaulis, suggesting the potential for ecotone shif t if heterospecific
seedlings and saplings accede to the overstory (Figure 2; elevation
bands: 2,60 0– 2,800 m). P. albicaulis- dominated stands had similar
overstory and understory communities, suggesting relative stability
in composition (Figure 2; elevation bands 2,900– 3,000 m).
NMDS ordination comparing densities of overstory trees and
understory populations of the same species suggested both sta-
bility and the potential for ecotone shift. These populations sepa-
rated most strongly along the axes of elevation (MDS1, r = −0.85)
and total basal area (r = 0.58; Appendix S8), with a stress of
0.02. Comparisons of MDS1 axis scores between overstory and

6
|
Journal of Vegetation Science BLOMDAHL et al.
understory trees of the same species suggested an upslope shift
in the establishment of P. contorta and a downslope shif t in P. a lb i-
caulis (Figure 3). Slight downward shifts were suggested for A. la-
siocarpa, P. engelmannii, and P. menziesii; however, these results
must be viewed in light of limited tree densities in most plot over-
stories and some plot understories.
We found clear evidence of an ecotone shift in plots greater
than 3,000 m in elevation, from meadow to closed- canopy forest.
Virtually all of the understory recruitment at the highest elevations
occurred in the 20th century, and a substantial majority occurred
after 1950 (Figure 4). P. albicaulis dominated the understory recruit-
ment signal, in particular at high elevation where it was nearly the
only species recruiting.
3.2.2 | Pine mortality
Tree mor talit y across all elevations ranged from 0.4 to 27.2 m2/ha,
was concentrated exclusively in Pinus spp. (Table 1; Appendices S5-
S7). At higher elevations, 48.0% of the total forest basal area was
P. albicaulis standing dead, and at lowest elevations, 37.9% of the
total forest basal area was P. contorta standing dead (Table 1;
Appendix S5). Mortality on the transect disproportionately occurred
in P. albicaulis (62% of total, compared with 38% of P. contorta),
where P. albicaulis standing dead also tended to be older on average
(p < 0.05; Figure 5).
Drought and D. ponderosae were the main drivers of mortal-
ity patterns in the study area. Indeed, 42% of mortality events fell
FIGURE 2 Species- specific tree densities (stems/ha) displayed by estimated establishment year decade, grouped by 100 - m elevation
bands for: (a) live and dead (designated by black hatch marks) overstor y trees; and (b) understor y seedlings. ABLA: Abies lasiocarpa; PIAL:
Pinus albicaulis; PICO: Pinus contorta; PIEN: Picea engelmannii; PSME: Pseudotsuga menziesii.

|
7
Journal of Vegetation Science
BLOMDAHL et al.
within the period 2000– 2010, characterized as the recent “mega-
drought”, and fully 77% died post- 2000. Mortality of Pinus spp. was
largely attributed to the D. ponderosae outbreak in 2008– 2012. Of
the 26 dead trees for which we were able to assign a year of death,
58% had evidence of blue- stain fungus, 19% did not have blue- stain
fungus, and the remainder were undetermined. The average DCH of
beetle- killed P. contorta was 29.0 cm, whereas the average DCH of
beetle- killed P. albicaulis was 20.5 cm. Finally, the timing of mortal-
ity for the dead trees encountered on the transect was significantly
(p < 0.01) associated with drier than average drought conditions
2– 3 years prior to tree death (Figure 6), refuting the null hypothesis
that deaths occurred randomly over temporal variability of summer
drought.
3.2.3 | Climate and tree growth
Ring- width growth relationships with climate variables suggested
both a strengthening of limiting temperature conditions and an
emerging and strengthening response to monthly PDSI (Figure 7).
Throughout the record, P. albicaulis grow t h ha d a po s i t i v e re lat i o n s h ip
with cool season maximum temperatures. The streng th of this rela-
tionship was variable but consistent (positive) over time (Figure 7a).
By contrast, P. contorta grow th had a strong negative response to
previous warm season (August and July) monthly maximum tem-
peratures throughout the historic al record, indicative of the lagged
effect of previous growing season climate conditions on the fol-
lowing growing season, caused by determinate growth (Figure 7b).
TAB LE 1 Average live and dead basal area (m2/hect are) for all trees in the study transect >5 cm diameter at coring height, grouped by
elevation band
Elevation band
Live basal area (m2/ha) (% of total live)
Dead basal area (m2/ha) (%
of total live + dead)
ABLA PIAL PICO PIEN PSME Tota l PIAL PICO
3,000 0 (0) 9 (77.6) 2.6 (22.4) 0 (0) 0 (0) 11.6 10.9 (48) 0. 2 (0.9)
2,90 0 0 (0) 23.9 (80.2) 4.7 (15.8) 1.2 (4) 0 (0) 29. 8 13.5 (31.2) 0 (0)
2,800 0 (0) 4.1 (9.6) 36.6 (86.1) 1.2 (2.8) 0.6 (1.4) 42. 5 0 (0) 9.8 (18.7)
2,700 0 (0) 2.8 (5.2) 51.3 (94.8) 0 (0) 0 (0) 54.1 0 (0) 0.4 (0.7)
2,600 5.7 (12.8) 1.9 (4.3) 36 .9 (82 .9) 0 (0) 0 (0) 44.5 0 (0) 27.2 (37.9)
Average 1.1 8.3 26.4 0.5 0 .1 36.5 4.9 7. 5
ABLA: Abies lasiocarpa; PIAL: Pinus albicaulis; PICO: Pinus contorta; PIEN: Picea engelmannii; PSME: Pseudotsuga menziesii.
TAB LE 2 Average live and dead stem densities (trees/ha) for all trees in the study transect >5 cm diameter at coring height, grouped by
elevation band
Elevation band
Live stem density (trees/ha) Dead stem density (trees/ha)
ABLA PIAL PICO PIEN PSME To t al PIAL PICO Tot a l
3,000 0992 169 0 0 1,160 271 23 294
2,90 0 0775 163 38 01,002 343 0343
2,800 0337 720 32 36 1,124 085 85
2,700 0346 626 0 0 972 048 48
2,600 388 29 586 0 0 1,004 0388 388
Average 78 496 453 14 71,0 52 123 109 232
ABLA: Abies lasiocarpa; PIAL: Pinus albicaulis; PICO: Pinus contorta; PIEN: Picea engelmannii; PSME: Pseudotsuga menziesii.
Elevation band SDI QMD
Live
BA Dead BA
Live
density
Dead
density
3,000 313.7 12 11.6 11 .2 1,160 294
2,90 0 614.1 19.7 29.9 13.5 1,0 02 343
2,800 828.9 21.8 42 .4 9. 8 1,124 85
2,700 961 24 .4 54.2 0.4 972 48
2,600 864.5 25.5 4 4.4 27.2 1,004 388
Average 664.2 19. 6 33 .1 12.1 1,065 241
TAB LE 3 Average stand density index
(SDI), quadratic mean diameter (QMD),
live and dead basal area (BA; m2/hect are),
and live and dead stem density (trees/ha)
for all trees in the study transect >5 cm
diameter at coring height, grouped by
elevation band

8
|
Journal of Vegetation Science BLOMDAHL et al.
P. contorta exhibited a positive and temporally stable relationship to
late- growing season temperature (Figure 7b). Both species exhibited
a strengthening positive relationship to previous growing season
and fall PDSI in recent decades. The growth response of P. albicaulis
shifted from a negative to a positive relationship with PDSI during
the mid- 20th century (Figure 7a). P. contorta exhibited a negative re-
lationship to current year growing season PDSI in recent decades
(Figure 7b).
FIGURE 3 Non- metric Multidimensional (Distance) Scaling (NMDS) ordination of species distributions. Overstory trees are represented
by upper case species codes, whereas understory trees (seedlings and saplings) are represented by lowercase species codes. ABLA : Abies
lasiocarpa; PIAL: Pinus albicaulis; PICO: Pinus contorta; PIEN: Picea engelmannii; PSME: Pseudotsuga menziesii. (a) Two- axis NMDS ordination
of density (stems/ha) by species and form (overstory trees vs. seedlings and saplings) at each sampling plot (stress = 0.02), plotted along an
elevational gradient. Open circles represent plot locations, black circles represent centers of species forms (trees, seedlings and saplings),
and contour lines are at 50- m altitude intervals. (b) Difference and magnitude in MDS scores for overstory and understory species
represented by plot ting MDS1 scores x−1 (for ease of interpretation). A negative slope (red) in panel b suggests the center of the distribution
of species seedlings is located at lower elevations than mature trees of the same species, and a positive slope (green) suggests an upward
shift in establishment by a species.
FIGURE 4 Estimated establishment
year of overstory trees by plot elevation.
Regression lines are shown for the two
most prevalent species on the transect.
ABLA: Abies lasiocarpa; PIAL: Pinus
albicaulis; PICO: Pinus contorta; PIEN:
Picea engelmannii; PSME: Pseudotsuga
menziesii.

|
9
Journal of Vegetation Science
BLOMDAHL et al.
4 | DISCUSSION
Demographic changes and elevational movement in long- lived trees
can be challenging to decipher. Pairing forest demographic transects
with dendroecological data offers insight into the past and possible
future trajectories of the forest when long- term, repeated obser-
vations are not available. We measured overstory and understory
composition as well as live and dead age structure across a transect
spanning 500 m in elevation to assess the potential for ecotone shif t in
the high- elevation forests of the GYE. Trees at the highest elevations
(>3,000 m) established after 1950, indicating ecotone shift from high-
elevation meadow to Pinus albicaulis- dominated forest. We attribute
this shift to an upward advance of P. albicaulis rather than regeneration
following a mortality event, due to the paucity of dead stems of snags
in plots above 3,000 m (Appendices S 5 - S 7 ). This shift was synchro-
nous with and positively related to warming winter maximum temper-
atures, suggesting a relaxing of environmental conditions previously
limiting to P. albicaulis establishment. Comparisons between overstory
and understory tree composition at different elevational bands sug-
gested compositional stability at higher elevations and ongoing suc-
cessional patterns in the absence of fire at lower elevations, although
our results should be interpreted with the understanding that stochas-
tic effects could infl uence obse rved demo graphy at the plot level. Our
analysis also revealed that Pinus contorta seedlings were distributed
at higher elevations than trees of the same species, which suggested
some potential for expec ted upslope movement with warming condi-
tions. Conversely, P. albicaulis seedlings were distributed at lower el-
evations than P. albicaulis overstory trees, possibly a combined result
of seed caching and canopy gaps due to extensive tree mortality. At
lower elevations dominated by P. contorta, about one- third of the total
bas al are a were sn ag s that we re like ly create d via dro ug ht- and beetle-
caused mortality exacerbated by warming temperatures and higher
stand densities. At higher elevations mortality on the transect was
greatest in P. albicaulis- dominated st ands, despite lower stand densi-
ties. This combined with snag age structure reflects a disproportion-
ate level of mort ality amongst old P. albicaulis trees.
4.1 | Vegetation change and potential drivers
4.1.1 | Ecotone shift
Differences between overstory and advanced regeneration composi-
tion may approximate ecotone shifts in GYE subalpine forests. If cur-
rent disturbance and climate trends continue, we expect understory
FIGURE 5 (a) Snag mortality year (n = 26) by plot elevation. Of the five species present on the transect, tree mortality was concentrated
exclusively in Pinus contorta (PICO; n = 10) and Pinus albicaulis (PIAL; n = 16). We assessed tree core samples for blue- st ain fungus as an
indicator of Dendroctonus ponderosae presence (Y: yes; N: no; U: undetermined). Mortality year had a positive relationship with elevation in
P. contorta (p = 0.047) and no relationship in P. albicaulis (p = 0.443). (b) Snag mortality year by estimated tree age. Recent mortality events
tended to af fect older P. albicaulis trees on average (p = 0.003).
FIGURE 6 Superposed epoch analysis results of drought
conditions surrounding tree mortality events (n = 26) in the study
area from 1895 to 2019. A composite of all mortality event dates
and their associated Palmer Drought Severity Index (PDSI) values
(year “0”) is presented with a ±10- year lag with 95% (dashed lines)
and 99% (solid lines) boot strapped confidence intervals. Bars
represent mean departures in PDSI for the years surrounding a
mortality event from mean conditions across the entire time series,
with a significant departure exceeding the confidence intervals.
Two to three years prior to the mortality event years, conditions
were much drier than average (p < 0.01).

10
|
Journal of Vegetation Science BLOMDAHL et al.
seedling composition to estimate future overstory composition as
fire exclusion, warmer temperatures, greater aridity and bark beetle
attacks shape recruitment into the canopy. Aging P. contorta stands,
which typically establish after stand- replacing disturbances such as
wildfire, are being predictably replaced by Abies lasiocarpa and P. al -
bicaulis, and to a lesser extent, Picea engelmannii and Pseudotsuga
menziesii (Figure 2). We speculate that high stand densities created
conditions less favorable for shade- intolerant P. contorta seedlings
and more favorable for shade- tolerant A. lasiocarpa seedlings, as
predicted by successional theory (Clements, 1910). This is consist-
ent with the findings of Chai et al. (2019) and Brice et al. (20 19), who
demonstrated the replacement of shade- intolerant pioneer species
by shade- tolerant species in the absence of disturbance via a differ-
ent study design involving permanent plots with multiple censuses.
The distribution of P. contorta also appears to be moving upslope
(Figure 3), coincident with elevated levels of P. albicaulis mortality
that created light gaps in the canopy at higher elevations (Table 1).
By contrast, P. albicaulis may establish in denser forests at lower
elevations, in large part due to dissemination from Clark's nut-
cracker. Goeking and Izlar (2018) found that the majority of P. al -
bicaulis stems in the western USA occur in forest types dominated
by other species, including the forest types: P. contorta, spruce fir,
A. lasiocarpa, P. menziesii, non- stocked (<10% of full stocking of live
trees), and P. engelmannii. However, P. albicaulis seedlings are less
shade- tolerant than A. lasiocarpa (Minore, 1979), so long- term sur-
vivorship may be expected to be lower in most areas on our study
transect except for the young, sparse, leading edge of P. albicaulis
stands. These successional processes, as well as climate trends and
disturbances such as bark beetle outbreaks and fire will likely dictate
which understor y trees accede to the canopy in future forest s.
The upward expansion of P. albicaulis forest into high- elevation
meadows is likely a consequence of changing climate and fire sup-
pression. We posit that a general warming trend that started around
the mid- 20th century has allowed for P. albicaulis establishment.
Increasing winter temperatures were also positively correlated with
P. albicaulis grow th , po ssibl y due to acc e l e r a t e d snowm elt and ex pan-
sion in the length of the growing season. This is consistent with other
studies that have related the expansion of forest treeline to warming
climate (Klasner & Fagre, 2002; Millar et al., 2004; Kullman, 2016).
However, Schrag et al. (2008) modeled a decrease in treeline P. a l-
bicaulis under climate change scenarios of a 4.5°C increase in tem-
perature and a 35% increase in precipitation, suggesting a climate
FIGURE 7 Correlation coefficients of
Pinus albicaulis (a) and Pinus contorta (b)
annual ring- width growth relationships to
monthly maximum temperature and the
Palmer Drought Severity Index (PDSI),
across a moving window of 35 years.
Months labeled in upper case letters are
from the year prior to that of the ring-
width measurement. Asterisks indicate
significance (p < 0.05) using the 95%
percentile range method (Dixon, 2001).

|
11
Journal of Vegetation Science
BLOMDAHL et al.
envelope that would eventually inhibit expansion at treeline. We
also suspect that 20t h cen tury fire suppression played a role in limit-
ing fire in the region that encouraged greater seedling est ablishment
rates in our transect s during this period (Brown et al., 2020). The lack
of fire may have also played a role in the ecotone shift we observed
because young trees are unlikely to survive fire of any severity.
4.1.2 | Pine mortality
Patterns in stand densit y generally coincided with areas of high
mortality across the t wo transects. Observed densities suggested
imminent density- dependent mortality at all but the highest el-
evations (i.e., stand density of P. contorta stands >420 [McC arter &
Long, 1986]; stand density of P. albicaulis stands >370 [Shaw, 2017]).
High densities create stressful competitive environments that in-
crease the risk of spread of biotic disturbance agents such as bark
beetles (Perkins & Roberts, 2003; Das et al., 2011). Mortality, largely
due to Dendroctonus ponderosae, was high at the lowest elevations
characterized by a P. contorta overstory and high SDI (Table 3). In
lower- density, higher elevation P. albicaulis stands, mort ality was
better explained by drought stress in conjunction with D. pondero-
sae activity. Despite seemingly better growing conditions, areas
converted from alpine meadow to P. albicaulis forest experienced
mortality in nearly half the total stand basal area (Table 1).
The results of the SEA suggest a drought- driven mortalit y spiral
for the Pinus spp. in this study (Figure 6). Although the “fading re-
cord” of long- dead trees that decomposed prior to this study limit s
our inference to more recent decades, it is notable that nearly half of
the mortality events we obser ved fell within the period 200 0– 2010,
characterized as the recent “mega- drought” (Martin et al., 2020).
Given that D. ponderosae has a 1- year life cycle and takes several
years to build to epidemic levels (Bentz & Powell, 2014), the 2– 3- year
lag between extremely dry conditions and mortality suggests that
these trees died from the combined pressure of drought and D. pon-
derosae attack. Drought can act in a dual capacity to enable tree
death: (1) reducing host vigor because of increased vapor pressure
deficit and lower soil moisture available to trees; and (2) by increas-
ing the population of the ultimately poikilothermic bark beetles.
This lag in timing between tree death and environmental variability
has been found for other D. ponderosae hosts (Boutte et al., 2016),
and also for other Dendroctonus spp. host species like P. engelmannii
(Mast & Veblen, 1994; DeRose & Long, 2012; DeRose et al., 2017 ).
4.1.3 | Climate and tree growth
Warming temperatures and drought both played a role in forest
change in the GYE. Correlations between radial growth and climate
suggested an increasingly positive relationship with monthly maxi-
mum temperature in P. albicaulis, coincident with an expansion of
the P. albicaulis treeline since 1950. This observation follows expec-
tations because tree growth in subalpine elevations was historically
constrained by the length of the growing season and snow cover
(Peterson, 1998). Mid- twentieth century, P. albicaulis growth
switched from a negative to a positive relationship with PDSI (nega-
tive values correspond to drier conditions), suggesting a switch to
a more limiting, arid environment following an especially long, cool
period (Rochner et al., 2021). For both P. albicaulis and P. contorta,
the positive relationship to previous growing season PDSI strength-
ened in recent decades, likely exacerbating Pinus spp. mortalit y dur-
ing the 2008– 2012 D. ponderosae outbreak. P. contorta growth also
had a negative relationship to current growing season PDSI in re-
cent decades, a counterintuitive finding that merits further inquiry.
Interestingly, the high- elevation changes in P. albicaulis that we have
observed on our transect may be a relatively short- term snippet of a
multi- centennial shift in ecotones that often occurs in harsh environ-
ments. Upslope of our transects, extensive P. albicaulis and P. engel-
mannii forest existed until the middle- to- late part of the Little Ice
Age (Rochner et al., 2021). Recruitment of these forests occurred
nearly 1,000 years ago, but experienced widespread die- off in the
mid- 1800s during the coolest conditions of the last millennium, pu-
tatively due to climatic causes (Rochner et al., 2021).
4.2 | Implications for P. albicaulis decline
The decline in P. albicaulis in the GYE has been the cause of much
concern in recent decades (Tomback et al., 2001b; Keane et al., 2017;
Goeking & Izlar, 2018), leading the US Fish and Wildlife Service to
propose it be listed as threatened under the Endangered Species Act
in December 2020. P. albicaulis is considered a keystone species be-
cause it promotes biodiversit y by providing habitat to many species
and is a central food source via its large, nutritious seeds (Tomback
& Kendall, 2001). Given recent declines and ongoing climate change,
studies that assess P. albicaulis stability and upslope movement are
important to help managers target restoration efforts. We found
a dispropor tionate level of mortality in P. albicaulis in our study
area, with virtually all mortalit y event s attributable to D. pondero-
sae. Although we did not make explicit comparisons (e.g., genetics)
between living and dead P. albicaulis, death dates suggested that
climate change- driven drought played a role in creating conditions
that resulted in elevated beetle- related mortality (Six et al., 2018).
Despite the elevated levels of mortality in mature trees, there was
substantial P. albicaulis in the understor y, some of which wa s mak ing
its way into the high- elevation meadows. The slow march toward
higher elevations could portend the return of P. albicaulis to a niche
it realized prior to the Little Ice Age cooling (Rochner et al., 2021).
5 | CONCLUSIONS
We observed a forest undergoing compositional change across
different elevations, mediated by an interplay of climate- related
stressors, bark beetle outbreak and successional processes. Non-
stable temperatures and increased sensitivity to aridity during the

12
|
Journal of Vegetation Science BLOMDAHL et al.
mid- 20th century, in combination with increasing stand density as-
sociated with aging forests, likely created conditions of increased
susceptibility to beetle- induced mortalit y during the most recent
two decades. In the context of these compound stressors, forest
age, structure, and composition sampled across an elevational gradi-
ent suggested evidence for both ecotone shift s for some species and
stability for others. Our assessment of tree species movement in the
context of climate change and disturbance history advances under-
standing of the drivers of tree community change in the subalpine
forests of the GYE.
AUTHOR CONTRIBUTIONS
EMB, JHS, MK and MR conceived of the research idea and method-
ology; all authors collected the data; EMB, JHS, NEZ, BC, GDC, BH,
HL, RO, LP, IT, KW and RJD performed statistical analyses; EMB, RJD
and BDW wrote the paper with contributions from all KW and DA;
all authors discussed the results and commented on the manuscript.
ACKNOWLEDGMENTS
This paper was prepared, in part, as participants in the North
American Dendro- Ecological Fieldweek (2017, 2018 and 2019) and
we would like to thank the additional students who contributed
to field data collection. We wish to acknowledge the Shoshone
National Forest with a special thanks to Amy Haas for coordinating
access for our sampling area.
FUNDING INFORMATION
The field week was supported by the National Science Foundation
(Grant No. 1759694). This research was supported by the Utah
Agricultural Experiment Station, Utah State Universit y, and ap-
proved as journal paper number 9453.
DATA AVA ILAB ILITY STATE MEN T
The data that support the findings of this study are available at
htt ps://doi.or g/10 .5281/zenodo.67 74751.
ORCID
Erika M. Blomdahl https://orcid.org/0000-0002-2614-821X
James H. Speer https://orcid.org/0000-0003-1188-0552
Nicole E. Zampieri https://orcid.org/0000-0003-1990-9153
Maegen Rochner https://orcid.org/0000-0002-2340-5428
Bryce Currey https://orcid.org/0000-0001-9794-9906
Gabriel D. Cahalan https://orcid.org/0000-0003-3445-4900
Hang Li https://orcid.org/0000-0002-0348-9812
Ichchha Thapa https://orcid.org/0000-0002-4869-7228
Kevin Willson https://orcid.org/0000-0001-9233-0557
R. Justin DeRose https://orcid.org/0000-0002-4849-7744
REFERENCES
Allen, C.D., Breshear s, D.D. & McDowell, N.G. (2015) On underestima-
tion of global vulnerabilit y to tree mor tality and forest die- of f from
hotter drought in the Anthropocene. Ecosphere, 6(8), 1– 55. ht t p s : //
doi.org/10.1890/ES15- 00203.1
Allen, C.D., Macalady, A.K ., Chenchouni, H., Bachelet, D., McDowell, N.,
Vennetier, M. et al. (2010) A global overview of drought and heat-
induced tree mortalit y reveals emerging climate change risks for
forest s. Forest Ecology and Management, 259(4), 660– 684. h t tp s : //
doi.org/10.1016/j.foreco.2009.09.001
Anderegg, W.R.L ., Hicke, J.A., Fisher, R.A., Allen, C.D., Aukema, J., Bentz,
B. et al. (2015) Tree mort ality from drought, insects, and their in-
teractions in a changing climate. New Phytologist, 20 8(3), 674– 683.
https ://doi.or g/10.1111/n ph .13 477
Anderegg, W.R.L., Kane, J.M. & Anderegg, L.D. (2013) Consequences
of widespread tree mor tality triggered by drought and tem-
perature stress. Nature Climate Change, 3(1), 30– 36. h t tp s : //d o i .
org /10.103 8/nclim ate1635
Appl equist, M. (1958) A simp le pit h loc ator for use with of f- center incre -
ment cores. Journal of Forestry, 56(1), 141.
Beckage, B., Osborne, B., Gavin, D.G., Pucko, C., Siccama, T. & Perkins,
T. (2008) A rapid upward shift of a forest ecotone during 40 years
of warming in the Green Mountains of Vermont. Proceedings of
the National Academy of Sciences, 105(11), 4197– 4202. ht t p s : //d o i.
org /10.1073/pnas.070 89 21105
Bell, D.M., Bradford, J.B. & Lauenroth, W. (2014) Earl y indicators of
change: divergent climate envelopes bet ween tree life stages
imply r ange shifts in the western United States. Global Ecology and
Biogeography, 23(2), 168– 180. htt ps: //doi. org/10 .1111/ge b.12109
Bentz, B.J. & Powell, J.A. (2014) Mountain pine beetle seasonal tim-
ing and constraints to bivoltinism (a comment on Mitton and
Ferrenberg, ‘mountain pine beetle develops an unprecedented
summer generation in response to climate warming ’). The A merica n
Naturalist, 184(6), 787– 796. h ttp s: //do i.o rg /10.1086/678 4 05
Bentz, B. J., Regniere, J., Fettig, C .J., Hansen, M., Hayes, J. L., Hicke, J.A .
et al. (2010) Climate change and bark beetles of the western United
States and Canada: direct and indirect effects. Bioscience, 60(8),
60 2– 61 3. https://doi.org/10.1525/bio.2010.60.8.6
Boutte, P.C., Hicke, J.A., Preisler, H.K., Abatzoglou, J.T., Raffa, K.F. & Logan,
J.A. (2016) Climate influences on whitebark pine mortality from
mountain pine beetle in the greater Yellowstone ecosystem. Ecological
Applications, 26(8), 2507– 2524. https://doi.org/10.1002/eap.1396
Breshears, D.D., Huxman, T.E., Adams, H.D., Zou, C.B. & Davison, J.E.
(2008) Vegetation synchronously leans upslope as climate warms.
Proceedi ngs of the National Academy of Sciences, 105(33), 11591–
11592 . ht tps://doi.o rg/10.1073/pnas.08 065 79105
Brice, M.H., Cazelles , K., Legendre, P. & Fortin, M. J. (2019) Disturbances
amplif y tre e community responses to climate change in the
temperate- boreal ecotone. Global Ecolog y and Biogeography, 28(11),
1668– 1681. ht tps://doi.org /10.1111/geb.12 971
Brown, S.R., Baysinger, A., Brown, P.M., Cheek, J.L., Diez, J.M., Gentr y,
C.M. et al. (2020) Fire history across forest types in the southern
Beartooth Mountains, Wyoming. Tree- Ring Research, 76(1), 27– 39.
https://doi.org/10.3959/ TRR20 18- 11
Bunn, A.G. (20 08) A dendrochronology program library in R (dplR).
Dendrochronologia, 26(2), 115– 124. ht tps://doi.o rg /10.1016/j.
dendro.2008.01.002
Chai, R.K., Andrus, R. A., Rodman, K., Har vey, B.J. & Veblen, T.T. (2019)
Stand dynamics and topographic setting influence changes in live
tree biomass over a 34- year permanent plot record in a subal-
pine forest in the Colorado front range. Canadian Journal of Forest
Research, 49, 1256– 1264. https://doi.org/10.1139/cjfr- 2019- 0 023
Christensen, N.L. (2014) An historical per spective on fores t succession
and it s relevance to ecosystem restoration and conser vation prac-
tice in North America. Forest Ecology and Management, 330, 312–
322. https://doi.org/10.1016/j.foreco.2014.07.026
Clements, F.E. (1910). The life histor y of lodgepole burn forests. USDA For.
Serv. Bull. 79: 7– 56.
Cybis Elektronik (2010) CDendro and CooRecorder. Available at: ht t p: //
www.cybis.se/forfu n/dendr o/index.htm [Acce sse d 18t h July 20 19] .

|
13
Journal of Vegetation Science
BLOMDAHL et al.
Das, A., Battles, J., Stephenson, N.L. & van Mantgem, P.J. (2011) The
contribution of competition to tree mor tality in old- growth conif-
erous forest s. Forest Ecology and Management, 261(7), 1203– 1213.
https://doi.org/10.1016/j.foreco.2010.12.035
DeRose, R.J., Bekker, M.F. & Long, J.N. (2017) Traumatic resin ducts
as indic ators of bark beetle outbreak s. Canadian Journal of
Forest Resources, 47, 1168– 1174. https://doi.org/10.1139/
c j f r - 2 0 1 7 - 0 0 9 7
DeRose, R.J. & Long, J.L. (2012) Drought- driven disturbance his-
tory characterizes a sout hern Rocky Mountain subalpine forest.
Canadian Journal of Forest Resources, 42, 1649– 1660. ht t p s :// do i .
o r g / 1 0 . 1 1 3 9 / X 2 0 1 2 - 1 0 2
Dixon, P.M. (2001) Bootstrap resampling. In: El– Shaarawi, A.H.,
Pie gorsch, W.W. (Eds .), The Encyclopedia of Environmetrics. New
York: Wiley.
Donato, D.C., Harvey, B .J. & Turner, M.G. (2016) Regeneration of mon-
tane forests 24 years af ter the 1988 Yellowstone fires: a fire-
catalyzed shift in lower treelines? Ecosphere, 7(8), e01410. h t t p s: //
doi.org/10.10 02/ecs2 .1410
Douglass, A.E. (1941) Crossdating in dendrochrology. Journal of Forestr y,
39(10), 825– 831. https://doi.org/10.1093/jof/39.10.825
Franceschi, V.R., Krokene, P., Christiansen, E. & Krekling, T. (2005)
An at om ic al and chem ic al def en ses of coni fe r bar k ag ai nst ba rk be e-
tles and other pests. New Phytologist, 167(2), 353– 376. h t t p s: //d o i .
org /10.1111/j .1469- 8137.20 05.014 36 .x
Goeking, S.A. & Izlar, D.K. (2018) Pinus albicaulis Engelm. (whitebark
pine) in mixed- species stands throughout its US range: broad- scale
indicators of extent and recent decline. Forests, 9(3), 131. h t t p s ://
doi.org/10.3390/f9030131
Hatala, J.A., Crabtree, R.L., Halligan, K.Q. & Moorcroft, P.R. (2010)
Landscape- scale patterns of fores t pest and pathogen damage in
the greater Yellowstone ecosystem. Remote Sensing of Environment,
114(2), 375– 384. https://doi.org/10.1016/j.rse.2009.09.008
Holmes, R.L. (1983) Computer- assisted quality control in tree- ring dating
and measurement. Tree- Ring Bulletin, 43, 69– 78.
Hufkens, K., Scheunders, P. & Ceulemans, R. (20 09) Ecotones in
vegetation ecology: methodologies and definitions revisited.
Ecological Research, 24, 977– 986. https://doi. org /10.10 07/s1128
4 - 0 0 9 - 0 5 8 4 - 7
Iverson, L.R . & McKenzie, D. (2013) Tree- specie s range shifts in a chang-
ing climate: detecting, modeling, assisting. Landscape Ecology, 28(5),
879– 889. h t t p s : / / d o i . o r g / 1 0 . 1 0 0 7 / s 1 0 9 8 0 - 0 1 3 - 9 8 8 5 - x
Johnstone, J.F. & Chapin, F.S. (2003) Non- equilibrium succession
dynamics indicate continued northern migration of lodge-
pole pine. Global Change Biology, 9(10), 1401– 1409. ht t p s :// do i .
org/10.1046/j.1365- 2486.2003.00661.x
Keane, R., Holsinger, L., Mahalovich, M. & Tomback, D. (2017) Restoring
whitebark pine ecosystems in the face of climate change. General
tec hnical re port RMRS- G TR- 361. Fo rt Collin s, CO: US De pa rtme nt of
Agriculture, Forest Ser vice, Rocky Mountain Research Station, 123.
Klasner, F.L. & Fagre, D.B. (2002) A half centur y of change in alpine
Treeline pat terns at glacier National Park, Montana, U.S.A .
Arctic, Antarctic, and Alpine Research, 34(1), 49– 56. h t t p s :// do i .
org/10.1080/15230 430.2002.12003468
Klesse, S. (2021) Critical note on the application of the “two- third”
spline. Dendrochronologia, 65, 125786. ht tps ://doi.org/10.1016/j .
dendro.2020.125786
Krause, T.R. & Whitlock, C. (2017) Climatic and non- climatic controls
shaping early postglacial conifer history in the northern greater
Yellowstone ecosystem, USA . Journal of Quaternary Science, 32(7),
1022– 1036. https://doi.org/10.1002/jqs.2973
Kullman, L. (2016) Climate change and primary birch Forest (Betula pu-
bescensssp. Czerepanovii) succession in the Treeline ecotone of the
Swedish Scandes. Inter national journa l of research in Geogra phy, 2(2),
36 – 4 7. https://doi.org/10.20431/ 2454- 8685.0202004
Loehman, R. A., Bentz, B.J., DeNit to, G.A., Keane, R .E., Manning, M.E .,
Duncan, J.P. et al. (2018) Effects of climate change on ecological
disturbance in the northern Rockies. In: Climate change and Rocky
Mountain ecosystems. Cham: Springer, pp. 115– 141.
Macfarlane, W.W., Logan, J.A. & Kern, W.R. (2013) An innovative ae-
rial assessment of greater Yellowstone ecosystem mountain pine
beetle- caused whitebark pine mortality. Ecological Applications,
23(2), 421– 437. https://doi.org/10.1890/11- 1982.1
Malevich, S.B., Guiterman, C.H . & Margolis, E.Q. (2018) Burnr: fire his-
tory analysis and graphics in R. Dendrochronologia, 49, 9– 15. ht t p s: //
doi.org/10.1016/j.dendro.2018.02.005
Martin, J.T., Pederson, G.T., Woodhouse, C. A., Cook, E.R., McCabe, G.J.,
Anchukaitis, K. J. et al. (2020) Increased drought severity tracks
warming in the United States' largest river basin. Proceedings of the
National Academy of Sciences, 117(21), 11328– 11336. ht t p s: //d o i .
org /10.1073/p nas.19162 08117
Mast , J.N. & Vebl en, T.T. (1994) A dend rochron ological method of stu dy-
ing tree mortality pattern. Physical Geography, 15(6), 529– 542.
htt ps://doi.or g/10.1080/02723 64 6.1994 .106 42533
McCar ter, J.B. & Long, J.N. (1986) A lodgepole pine density management
diagram. Western Journal of Applied Forestry, 1(1), 6– 11. h t t p s :// do i .
org /10.1093/wjaf/1.1. 6
Millar, C.I., Westfall, R.D., Delaney, D.L., King, J.C. & Graumlich, L.J. (2004)
Response of subalpine conifers in the Sierra Nevada, California,
U.S.A., to 20th- centur y warming and decadal climate variability.
Arctic, Antarctic, and Alpine Research, 36(2), 181– 200. ht t p s :// d oi .
org/10.1657/1523- 0430(20 04)036[0181:ROSCI T]2.0.CO;2
Minore, D. (1979) Comparative autecological characteristics of nor thwest-
ern tree species: a literature review. General Technic al Report PNW-
87. Portland, Oregon: US Department of Agriculture, US Forest
Service, Pacific Nor thwest Forest and Range E xperiment Station.
Moore, P. (1954) Spacing in plant populations. Ecology, 35(2), 222– 227.
htt ps://doi.org /10.23 07/193112 0
Oksanen, J., Blanchet, F.G., Friendly, M., Kindt, R., Legendre, P., McGlinn,
D., Minchin, P.R., O’Hara, R .B., Simpson, G.L., Solymos, P., Stevens,
M.H.H., Szoec s, E., Wagner, J. (2018) Vegan: Community Ecology
Package. R package version 2 . pp. 4– 6.
Perkins, D. & Rob ert s, D. (2003) Predictive models of whitebark
pine mort ality from mountain pine beetle. Forest Ecolog y and
Management, 174(1– 3), 495– 510. https://doi.org/10 .1016/S0378
- 1 1 2 7 ( 0 2 ) 0 0 0 6 6 - X
Peterson, D.L. (1998) Climate, limiting factors and environmental change
in high- altitude forests of Western North America. In: The impacts
of climate variabilit y on fores ts. Berlin, Heidelb erg (Lecture Notes in
Earth Sciences): Springer, pp. 191– 208.
PRISM Climate Group (2020) Climatologic al normals, 1981- 2010. ht t p s ://
prism.orego nstate.edu/norma ls/ [Accessed 14 May 2020].
Rocca, M.E., Brown, P.M., MacDonald, L.H. & Carrico, C.M. (2014)
Climate change impacts on fire regimes and key ecosystem ser vices
in Rocky Mountain forests. Forest Ecology and Management, 327,
290– 305. https://doi.org/10.1016/j.foreco.2014.04.005
Rochner, M.L ., Heeter, K.J., Harley, G.L., Bekker, M.F. & Horn, S.P. (2021)
Climate- induced treeline mort ality during the termination of the lit-
tle ice age in the greater Yellowstone ecoregion, USA. The Holocene,
31(8), 1288– 1303. https ://doi .or g/10.117 7/09596 83621 1011656
Romme, W.H. & Despain, D.G. (1989) Historical perspective on the
Yellowstone fires of 1988. Oxford Journals, 39(10), 695– 699.
htt ps://doi.org /10.23 07/1311000
Schoennagel, T., Veblen, T.T. & Romme, W.H. (20 04) The interac-
tion of fire, fuels, and climate across Rocky Mountain fore sts.
Bioscience, 54(7), 661– 676. https://doi.org/10.1641/0006-
3568(2004)054[0661:TIOFF A]2.0.CO;2
Schrag , A.M., Bunn, A.G. & Graumlich, L.J. (2008) Influence of biocli-
matic variable s on tree- line conifer distribution in the greater
Yellowstone ecosystem: implications for species of conservation

14
|
Journal of Vegetation Science BLOMDAHL et al.
concern. Journal of Biogeography, 35, 698– 710. ht t p s :// d oi .
org /10.1111/j .136 5- 2699. 20 07.01815. x
Shaw, J.D. (2017) National determination of maximum stand density index
from Fores t inventory and analysis plots. Ogden, UT: unpublished
report on file at: U.S. Department of Agriculture, Forest Service,
Rocky Mountain Research Station.
Six, D.L., Vergobbi, C. & Cut ter, M. (2018) Are survivors dif ferent?
Genetic- based selection of trees by mountain pine beetle during
a climate change- driven outbreak in a high- elevation pine for-
est. Frontier s in Plant Science, 9, 993. https://doi.org/10.3389/
fpls.2018.00993
Smith, W.K., G ermino, M.J., Johnson, D.M. & Reinhardt, K. (2009) The
altitude of alpine treeline: a bellwether of climate change effects.
The Botanical Review, 75, 163– 190. https://doi.org/10.1007/s1222
9 - 0 0 9 - 9 0 3 0 - 3
Smithers, B.V., North, M.P., Millar, C.I. & Latimer, A .M. (2018) Leap frog
in slow motion: divergent responses of tree species and life stages
to climatic warming in Great Basin subalpine forests. Global Change
Biology, 24(2), e442– e457. htt ps ://doi.or g/10.1111/gcb.13881
Speer, J.H. (2010) Fundamentals of tree- ring research. Tucson,
AZ:Universit y of Arizona Press.
Stokes, M. A. & Smiley, T.L. (1968) An introduction to tree- ring dating.
Tucson, AZ: University of Chicago Press.
Tomback, D.F., A nderies, A.J., Carsey, K.S., Powell, M.L. & Mellmann-
Brown, S. (2001a) Delayed seed germination in whitebark pine
and regeneration patterns following the Yellowstone fires.
Ecology, 82(9), 2587– 2600. https://doi.org/10.1890/0012-
9658(2001)082[2587:DSGIW P]2.0.CO;2
Tomback, D.F., Arno, S.F. & Keane, R.E. (Eds.). (20 01b) Whitebark pine
communities: Ecology and re storation. Washin gton, DC: Island Press.
Tomback, D.F. & Kendall, K.C . (2001) Biodiversity losses: The downward
spiral. In: Tomback, D.F., Arno, S.F. & Keane, R. E. (Eds.) Whitebark
pine communities: Ecology and restoration. Washington, DC: Island
Press, pp. 243– 262.
Trugman, A.T., Anderegg, L.D., Anderegg, W.R., Das, A.J. & Stephenson,
N.L. (2021) Why is tree drought mortality so hard to predict? Tren ds
in Ecology & Evolution, 36(6), 520– 532. https://doi.org/10.1016/j.
tree.2021.02.001
Vacchiano, G., Derose, R.J., Shaw, J.D., Svoboda, M. & Mot ta, R. (2013)
A density management diagram for Nor way spruce in the temper-
ate European montane region. European Journal of Fores t Research,
132(3), 535– 549. h t t p s : / / d o i . o r g / 1 0 . 1 0 0 7 / s 1 0 3 4 2 - 0 1 3 - 0 6 9 4 - 1
Wan g, T., Ha ma nn , A., Spitt le ho us e, D. & Ca rroll, C. (2016) Loc ally down-
scaled and spatially customizable climate data for historical and fu-
ture periods for North America. PLoS One, 11(6), e0156720. ht t p s: //
doi.org/10.1371/journ al.pone.0156720
Zang, C. & Biondi, F. (2015) Treeclim: an R pack age for the numerical cal-
ibration of proxy- climate relationships. Ecography, 38 (4), 431– 436.
https://doi.org/10.1111/ecog.01335
SUPPORTING INFORMATION
Additional supporting information can be found online in the
Suppor ting Information section at the end of this article.
Appendix S1. Tree- ring chronology statistics.
Appendix S2. Correlation coefficients of climate variables monthly
average ring- widths of Pinus contorta separated into t wo elevational
groups (above and below 2,8 00 m) to explore whether growth
responses differed across high and low portions of the transec t.
Appendix S3. Correlation coefficients of climate variables monthly
average ring- widths of Pinus albicaulis sepa rated into two el ev at io na l
groups (above and below 2,8 00 m) to explore whether growth
responses differed across high and low portions of the transec t.
Appendix S4. Species composition and age struc ture of all trees and
seedlings displayed by elevation.
Appendix S5. Average live and dead basal area (m2/hectare) for all
trees in the study transect >5 cm diameter at coring height . Live and
dead stem densities (trees/hectare) for all trees in the study transect
>5 cm diameter at coring height.
Appendix S6. Live and dead stem densities (trees/hec tare) for all
trees in the study transect >5 cm diameter at coring height.
Appendix S7. Stand density index (SDI), quadratic mean diameter
(QMD), live and dead basal area (BA , m2/hect are), and live and dead
stem densit y (stems/hectare) for all trees in the study transect > 5
cm diameter at coring height.
Appendix S8. Correlation matrix of MDS scores and environmental
variables.
How to cite this article: Blomdahl, E.M., Speer, J.H., Kaye, M.,
Zampieri, N.E., Rochner, M. & Currey, B. et al. (2022) Drivers
of forest change in the Greater Yellowstone Ecosystem.
Journal of Vegetation Science, 33, 1– 14. Available from:
https://doi .org /10.1111/j vs .13141