Treeline advances along the Urals mountain range – driven by improved winter conditions?

Abstract

High-altitude treelines are temperature-limited vegetation boundaries, but little quantitative evidence exists about the impact of climate change on treelines in untouched areas of Russia. Here, we estimated how forest-tundra ecotones have changed during the last century along the Ural mountains. In the South, North, Sub-Polar, and Polar Urals, we compared 450 historical and recent photographs and determined the ages of 11 100 trees along 16 altitudinal gradients. In these four regions, boundaries of open and closed forests (crown covers above 20% and 40%) expanded upwards by 4 to 8 m in altitude per decade. Results strongly suggest that snow was an important driver for these forest advances: (i) Winter precipitation has increased substantially throughout the Urals (~7 mm decade-1), which corresponds to almost a doubling in the Polar Urals, while summer temperatures have only changed slightly (~0.05 °C decade-1). (ii) There was a positive correlation between canopy cover, snow height and soil temperatures, suggesting that an increasing canopy cover promotes snow accumulation and, hence, a more favorable microclimate. (iii) Tree age analysis showed that forest expansion mainly began around the year 1900 on concave wind-sheltered slopes with thick snow covers, while it started in the 1950s and 1970s on slopes with shallower snow covers. (iv) During the 20th century, dominant growth forms of trees have changed from multistemmed trees, resulting from harsh winter conditions, to single-stemmed trees. While 87%, 31%, and 93% of stems appearing before 1950 were from multistemmed trees in the South, North and Polar Urals, more than 95% of the younger trees had a single stem. Currently, there is a high density of seedlings and saplings in the forest-tundra ecotone, indicating that forest expansion is ongoing and that alpine tundra vegetation will disappear from most mountains of the South and North Urals where treeline is already close to the highest peaks.

Full text

Treeline advances along the Urals mountain
range –driven by improved winter conditions?
FRANK HAGEDORN
1
, STEPAN G. SHIYATOV
2
,VALERIYS.MAZEPA
2
,NADEZHDAM.
DEVI
2
,ANDREYA.GRIGOR’EV
2
,ALEXANDRA.BARTYSH
3
,VALERIYV.FOMIN
3
,
DENIS S. KAPRALOV
3
,MAXIMTERENT’EV
2
, HARALD BUGMAN
4
,ANDREASRIGLING
1
and PAVEL A. MOISEEV
2
1
Swiss Federal Institute for Forest, Snow and Landscape Research WSL, Z€
urcherstrasse 111, Birmensdorf CH-8903, Switzerland,
2
Institute of Plant and Animal Ecology, 8 Marta St. 202, Yekaterinburg 620144, Russia,
3
Ural State Forest Engineering
University, Sibirskii trakt 37, Yekaterinburg 620100, Russia,
4
ETH, Zurich CH-8903, Switzerland
Abstract
High-altitude treelines are temperature-limited vegetation boundaries, but little quantitative evidence exists about
the impact of climate change on treelines in untouched areas of Russia. Here, we estimated how forest-tundra eco-
tones have changed during the last century along the Ural mountains. In the South, North, Sub-Polar, and Polar
Urals, we compared 450 historical and recent photographs and determined the ages of 11 100 trees along 16 altitudi-
nal gradients. In these four regions, boundaries of open and closed forests (crown covers above 20% and 40%)
expanded upwards by 4 to 8 m in altitude per decade. Results strongly suggest that snow was an important driver
for these forest advances: (i) Winter precipitation has increased substantially throughout the Urals (~7 mm decade
1
),
which corresponds to almost a doubling in the Polar Urals, while summer temperatures have only changed slightly
(~0.05 °C decade
1
). (ii) There was a positive correlation between canopy cover, snow height and soil temperatures,
suggesting that an increasing canopy cover promotes snow accumulation and, hence, a more favorable microclimate.
(iii) Tree age analysis showed that forest expansion mainly began around the year 1900 on concave wind-sheltered
slopes with thick snow covers, while it started in the 1950s and 1970s on slopes with shallower snow covers. (iv) Dur-
ing the 20th century, dominant growth forms of trees have changed from multistemmed trees, resulting from harsh
winter conditions, to single-stemmed trees. While 87%, 31%, and 93% of stems appearing before 1950 were from mul-
tistemmed trees in the South, North and Polar Urals, more than 95% of the younger trees had a single stem. Cur-
rently, there is a high density of seedlings and saplings in the forest-tundra ecotone, indicating that forest expansion
is ongoing and that alpine tundra vegetation will disappear from most mountains of the South and North Urals
where treeline is already close to the highest peaks.
Keywords: Betula pubescens subsp. tortuosa, climate change, forest-tundra ecotone, Larix sibirica, microclimate, mountain ecosys-
tem, Picea obovata, snow, tree establishment
Received 3 February 2014 and accepted 11 March 2014
Introduction
Global temperatures have risen during the last century,
with the largest and most rapid changes occurring at
high altitudes and latitudes (Arctic Climate Impact
Assessment, 2005; Stocker et al., 2013). This warming
affects the world’s vegetation, particularly in cold tem-
perature-limited ecosystems (e.g., Hudson & Henry,
2009; Pauli et al., 2012). One of the most striking vegeta-
tion boundaries is the upper treeline, where tree
growth is thought to be primarily limited by low tem-
peratures (K€
orner, 2012). Worldwide, upper treelines
are confined to altitudes with mean growing season
temperatures of 5–8°, strongly suggesting that the
upright growth of trees is impaired if summer condi-
tions are too cold (K€
orner & Paulsen, 2004). Due to this
growth limitation by low temperature, treeline position
is a highly responsive bio-indicator (Kullman &
€
Oberg,
2009). Advances of woody vegetation have been
reported for various arctic and alpine treeline ecotones
around the world (e.g., Payette & Filion, 1985; Shiyatov
et al., 2007; Harsch et al., 2009). There are a number of
key factors and ecological processes that change
abruptly within just a few altitudinal meters at the
treeline: wind, snow cover, albedo, soil temperatures,
plant productivity, biodiversity, soil development, and
carbon and nutrient cycling (e.g., Holtmeier, 2003;
Kammer et al., 2009; Loranty et al., 2014). As a conse-
quence, treeline advances can have a tremendous
impact on ecosystems and their functions.
Treeline shifts are usually related to climatic warm-
ing, mainly to increased temperatures during the
Correspondence: Frank Hagedorn, tel. +41 44 739 2463, fax +41 44
739 2215, e-mail: hagedorn@wsl.ch
1©2014 John Wiley & Sons Ltd
Global Change Biology (2014), doi: 10.1111/gcb.12613
Global Change Biology

growing season, because trees can adapt to very low
winter temperatures (K€
orner, 2012). In a global meta-
analysis, however, Harsch et al. (2009) found that tree-
lines have advanced more strongly when they have
experienced greater winter than summer warming and
attributed this observation to an amelioration of harsh
winter conditions. However, the most important winter
‘stress’ is not temperature but the exposure of trees to
wind abrasion, snow and ice damage, and winter desic-
cation (Holtmeier, 2003). So far, winter stress has been
mainly discussed as a potential growth constraint on a
local scale and for individual trees (Lavoie & Payette,
1992; Weisberg & Baker, 1995; Holtmeier, 2003).
Recently, although, Kullman &
€
Oberg (2009) and Aune
et al. (2011) observed that, at a landscape scale, treeline
changes in the Scandes during the last century were
greatest in regions with the highest winter precipitation
and on wind-protected, concave slopes. In these loca-
tions, snow can accumulate and effectively remove
winter stress by insulating plant tissues and the rooting
zone, suggesting that snow conditions are at least an
important codriver for treeline advances. In support of
this ‘snow hypothesis’, a positive feedback between
increases in snow fall and enhanced plant growth has
been suggested as an important mechanism for the
observed large-scale increases in shrub abundance in
Northern Alaska and Siberia (e.g., Jia et al., 2003; Sturm
et al., 2005; Frost & Epstein, 2014). In these regions, a
larger snow pack leads to higher soil temperatures in
winter, thereby increasing nutrient availability and
plant growth, which in turn promotes the accumulation
of more snow.
The fact that the rate of treeline advance differs
strongly among regions as well as within a single area
provides support for the idea that, although treeline
formation is primarily caused by low temperature,
other factors can determine the rates of change. While
treelines in the Alps, interior Labrador, Alaska, and
North-Western Canada have hardly advanced (Lloyd,
2005; Gehrig-Fasel et al., 2007; Payette, 2007), treelines
in the Scandes have risen by up to 200 m in altitude
(Kullman &
€
Oberg, 2009; Aune et al., 2011). The reasons
proposed for this large variation in the rate of change
despite similar climatic warming include permafrost
depth, topography, wind exposure, tree species, soil
development, geomorphic processes, and snow cover
(e.g., Holtmeier, 2003; Lloyd, 2005; Kullman &
€
Oberg,
2009; Wilmking et al., 2012; Macias-Fauria & Johnson,
2013). However, most of these assessments were single
case studies and, when several sites were included, rel-
atively few attempts were made to systematically
assess in situ conditions such as topography, air and
soil temperature, and snow height and distribution. In
addition, climatic-driven treeline changes have been
confounded with shifts due to land abandonment in
many regions, especially in European mountain ranges,
making it difficult to disentangle the drivers for treeline
advances (Bolli et al., 2007; Gehrig-Fasel et al., 2007;
Aune et al., 2011).
Here, we synthesized regional surveys and new data
from remote regions of the Ural mountains to quantify
how and how fast the forest-tundra ecotone and its
boundaries have changed during the last centuries and
to identify the processes underlying these changes. Our
approach was (i) to compare the current position of the
forest-tundra ecotone with historical photographs and
maps in four regions of the Urals along a 1500 km
south-north gradient, (ii) to reconstruct the advances of
the forest-tundra ecotone and the changes in tree
growth form by determining sizes and ages of more
than 10 000 trees on 16 altitudinal study plot series
installed on slopes differing in topography and domi-
nant tree species, (iii) to relate these changes to climate
records from the last century collected at nearby
weather stations, and (iv) to identify the effects of
microclimatic conditions, such as air and soil tempera-
tures but also the heights and distribution of snow
within today’s forest-tundra ecotone. Our hypothesis
was that the forest-tundra ecotone has expanded
upwards along the whole Ural mountain range and
that snow cover has played a key role in these
advances.
Materials and methods
Study sites
The treeline ecotone is a diffuse vegetation boundary between
forest and alpine tundra (K€
orner, 2012) that is unique for each
slope and therefore does not allow a strict definition of a forest
line. Here, we distinguish between four upper limits (bound-
aries) according to Shiyatov et al. (2007) and Moiseev (2011):
(i) ‘species line’, with trees growing in creeping and shrubby
forms (multi- or single-stemmed) with heights reaching the
average snow surface; (ii) ‘treeline’, with individuals or tree
islands of multi- and single-stemmed trees with heights of
more than 2 m, distances between trees from 20 to 60 m and a
total crown cover of 5–10%; (iii) ‘open forest’ line, with distances
between trees from 7 to 30 m and total crown cover of
20–30%; and (iv) ‘closed forest’ line, continuous forest with dis-
tances between trees below 7–10 m and a total crown cover
above 40–50%.
Our treeline study was conducted along a 1500 km latitu-
dinal gradient in the Ural mountains, spanning the South
Urals (massif Iremel’), North Urals (Konzhakovsky Kamen’),
Sub-Polar Urals (massif Neroika), and Polar Urals (massif
Rai-Iz, Tchernaya, Slancevaya) (see Fig. 1, Table 1). The
altitude of the forest-tundra ecotone increases from about
170–320 m a.s.l. at the Polar Circle to 1225–1375 m a.s.l. in the
©2014 John Wiley & Sons Ltd, Global Change Biology, doi: 10.1111/gcb.12613
2F. HAGEDORN et al.

South Urals. Dominant tree species are Siberian spruce [(Picea
obovata (Ledeb.))] and white birch (Betula pubescens Ehrh. ssp.
tortuosa (Ledeb.) Nyman) in the Southern Urals; spruce, birch,
Siberian larch (Larix sibirica Ledeb.), and Siberian stone pine
[Pinus cembra ssp. sibirica (Du Tour) E. Murray] in the North
Urals; and larch and birch in the Sub-Polar and Polar Urals.
More details on the altitudinal ranges and climatic conditions
are given in Table 1.
Comparisons with historical photographs and mapping of
treeline shifts
Changes in the forest-tundra ecotone of the Ural mountains
were assessed by comparing present-day photographs with
historical photographs and maps, as well as by analyzing
stand age structure. Historical photographs were taken by
A.A. Mitenkov in 1900–1905 (South Urals), L.N. Tyulina in
1929–1930 (South), K.N. Igoshina in 1955–1962 (South, Polar),
L.D. Dolgushin in 1938–1956 (Sub-Polar), P.L. Gorchakovsky
in 1951–1961 (South, North, Sub-Polar), and by S.G. Shiyatov
in 1956–1983 (South, North, Sub-Polar, Polar). At the same
locations, we made more than 1000 photographs between 1999
and 2011 (Table S1). The locations of the original photographs
were identified in the field by seeking and matching the most
noticeable landmarks (rocks, hills, summits) on multiple hori-
zon lines (Moiseev & Shiyatov, 2003). On both the historical
and recent photographs, we marked the treeline and the
boundaries of the closed and open forest and estimated their
changes in altitudinal and horizontal distance using Adobe
Photoshop v8.0 software. Historical topographic and field-
derived thematic maps showing the distribution of different
types of forest-tundra vegetation communities were available
for the 1910s (Polar Urals), 1950s (Sub-Polar and North Urals),
1960–1970s (South Urals) and 2000s (all parts of Urals) (Shiya-
tov et al., 2005, 2007; Kapralov et al., 2006, 2007; Fomin et al.,
2007). On Russian topographic maps (1:250000, 1:500000),
Table 1 Characteristics of the four study areas in the Ural mountains
Study area in the Urals
South North Sub-Polar Polar
Local name of mountain Iremel’ Konzhakovsky
Kamen’
Neroika Rai-Iz
Geographical coordinates
N54°300–54°34059°300–59°40
0
64°300–64°35066°470–66°510
E58°490–58°54059°000–59°20059°300–59°35065°260–65°380
Max. altitude, m a.s.l. 1586 1569 1608 1236
Altitudinal range, treeline
ecotone m a.s.l.
1225–1375 875–1025 575–725 200–300
Geology Quartzite,
carbon-clay shale
Pyroxenit, gabbroid Metamorphized
shale, granite
Ultramafic rock
Mean June air temperature 16–19°C15–18°C14–16°C12–14°C
Mean January air temperature 15 to 17°C17 to 19°C19 to 23°C21 to 24°C
Annual precipitation*, mm 600–900 800 400–900 450–820
Maximal snow depth, cm 80–150 100–200 150–300 150–250
Dominant tree species†Po, Bp Ls, Po, Bp, Ps Ls, Bp Ls, Bp
*measured at nearest metero stations with altitudes of 457 and 1102 m a.s.l. in the South, 464 m in the North, 29 and 432 m in the
Sub-Polar, and 16 and 890 m in the Polar Urals.
†Po - Picea abies ssp. obovata,Bp- Betula pubescens ssp. tortuosa., Ls–Larix sibirica;Ps–Pinus sibirica.
Fig. 1 The Ural mountains span about 1500 km, from 54°to
66°N. The four study regions are ‘Iremel’ in the South Urals,
‘Konzhakovsky Kamen’ in the North Urals, ‘Maly Chender’ and
‘Neroika’ in the Sub-Polar Urals, and ‘Rai-Iz’ and ‘Chernaya’ in
the Polar Urals.
©2014 John Wiley & Sons Ltd, Global Change Biology, doi: 10.1111/gcb.12613
TREELINE ADVANCES ALONG THE URAL MOUNTAINS 3

forests are defined as areas with a crown cover above 20% and
average tree heights above 4 m. Historical maps were com-
pared with online-available satellite images with a spatial res-
olution of less than 1 m (www.maps.yandex.ru; www.google.
ru/maps; www.bing.com/maps). Based on the historical pho-
tographs, old topographic maps and old geo-botanical maps,
we assessed the altitudinal shifts of the upper boundaries of
open (20–30% tree cover) and closed forests (>40–50% tree
cover) quantitatively using the ARC/INFO geographic infor-
mation system (GIS; ESRI Inc., USA) with the AML language.
The boundary lines reflecting the positions of trees stands
with different cover values in former decades and in the 2000s
were converted into a raster format (each line was represented
by a set of cells 10 910 m in size) and superimposed on a
10-m digital elevation model of the study region in the GIS.
Thus, information on altitudinal position was obtained for
each raster cell. On this basis, we plotted histograms showing
the distributions of upper altitudinal boundaries of open and
closed forests at the beginning and end of the study period.
Stand structure analysis
In 2002–2008, we established 16 altitudinal study plot series
(transects) along different slopes of the forest-tundra ecotones
in all four regions of the Ural mountains (Table S1). Each alti-
tudinal transect consisted of three altitudinal levels: the tree-
line, the open forest line and the closed forest line. At each
altitudinal level, we established 3–6 plots of 20 920 m. In
each of these plots, all saplings taller than 20 cm and all trunks
of single- or multi-stemmed trees were recorded (total n=
20 600). We mapped the location of each stem and measured
its height, diameter at the base and breast height, and the area
covered by the crown. The age structure of all plots was deter-
mined by dendrochronological methods as follows. From trees
with diameters exceeding 3–5 cm at their base, we took a tree
core at heights from 0 to 30 cm from every second single-
stemmed living tree and from every fourth stem of multi-
stemmed trees. From every second tree taller than 0.2 m but
less than 3–5 cm in diameter, we sampled stem disks at the
root collar.
All cores were mounted on wooden strips. Cores and stem
disks were both cleaned with a paper knife and a shaving
blade. After enhancing ring boundary contrasts with white
powder, samples with narrow annual rings were measured on
the linear table LINTAB-V (F. Rinn S.A., Heidelberg) to a pre-
cision of 0.01 mm and were cross-dated using the computer
program TSAP-3.0 (Rinn, 1998) and Cofecha (Holmes, 1995).
Samples with wide rings were visually cross-dated, paying
special attention to frost and light rings. The dates of tree ger-
mination (single-stemmed trees) or the starting of upright
growth of individual trunks of multistemmed trees were
estimated by correcting for the number of years required to
grow to the height of sampling and for the number of years to
the pith when the core missed the inner ring. For cores hitting
the pith, the distance to the center of the tree was estimated by
fitting a circular template to the innermost curved ring
(Braeker, 1981). The number of years it took for a stem to grow
to the core height was determined from a regression of tree
age with height established for all seedlings and saplings at
each study site. Age and height were significantly related to
each other with an exponential relationship at all sites
(R
2
>0.6, P<0.001). In total, we determined the ages of
11 100 trees. The differences among slope types (concave, uni-
form, convex) were tested by fitting mixed-effects models
using maximum likelihood [lme function from the nlme pack-
age, R 2.10.1, R Development Core Team (2010)]. The model
included effects of region (n=4), altitude (treeline, open, and
closed forest) and slope type sequentially, and we separated
our data set into 50-year intervals (e.g., 1801–1850, 1851–1900).
Microclimatic measurements
Temperatures at 2 m height and at 10 cm soil depth were
recorded hourly with StowAway TidbiT TBI32-20 +50 tem-
perature data loggers (Onset Computer Corp., Bourne, MA,
USA) from September 2003 until August 2005 at treeline and
at open and closed forest lines in the South, North, and Polar
Urals (total n=146). Loggers for measuring air temperature
were installed in tree crowns at 2 m height and shielded with
tree bark and branches to prevent direct exposure to sun; soil
temperature loggers were buried at 10 cm depth both under
and outside of tree crowns. To study winter conditions in
greater detail, we buried additional temperature loggers at
10 cm soil depth across the forest-tundra ecotone in the South
and North Urals for 5 days during additional measurement
campaigns in March. These soil temperatures represent mini-
mum values, as soil temperatures continuously decrease
throughout the cold season and reach their minimum in
March. Snow heights were measured during sampling cam-
paigns in March from 2006 until 2008 in all plots used for
stand structure analyses in the four study regions.
Climate records
Historical climatic data for the Ural mountains were taken
from weather stations on the eastern and western side along
the Ural mountain range: Ufa (54°N450,56°E000; since 1888),
Zlatoust (55°N110,59°E410, since 1837), Kazan’ (55°N470,
49°E110; since 1812), Krasnoufimsk (56°N370,57°E450; since
1926), Ekaterinburg (56°N480,60°E380; since 1832), Perm
(57°N570,56°E120; since 1883), Biser (58°N310,58°510; since
1889), Karpinsk (59°N450,60°E010; since 1838), Cerdyn
(60°N240,56°E310; since 1890), Njaksimvol (62°N260,60°E520;
since 1888), Troicko–Pecerskoe (62°N420,56°E120; since 1888),
Ust’-Sugor (64°N160,57°E370; since 1896), Saran–Paul
(64°N170,60°E530; since 1888), Pecora (65°N070,57°E060; since
1888), Petrun’ (66°N280,60°E450; since 1888), and Salekhard
(66°N320,66°E320; since 1883). All precipitation data were
adjusted according to KNMI Climate Explorer http://cli-
mexp.knmi.nl and www.meteo.ru. Temporal changes in sum-
mer (June–August) and winter (November–March)
temperatures and precipitation were estimated by linear
regression for the period 1930–2008 where climate data were
available from all weather stations. For longer periods, we
used data from the weather station at Zlatoust (457 m a.s.l.;
90 km north-east of Iremel) for the South Urals. For the North
©2014 John Wiley & Sons Ltd, Global Change Biology, doi: 10.1111/gcb.12613
4F. HAGEDORN et al.

Urals, historical climatic data were taken from the weather sta-
tion Karpinsk, 45 km east of Konzhakovsky Kamen’ at 228 m
a.s.l.. Monthly average temperatures were available for 1838–
2008 and monthly precipitation for 1837–2004. For the Polar
Urals, we took climatic data from the weather station Salek-
hard, 55 km east of the study sites at 16 m a.s.l., where
monthly average temperatures were recorded from 1881–2008
and monthly precipitation from 1899–2000.
Results
Reconstructing treeline advances using historical
photographs and maps
The comparisons of historical with recent landscape
photographs, of old topographic maps with modern
satellite photos, and of historical with recent tree stand
surveys clearly documented advances of the forest-tun-
dra ecotone along the Urals in the second half of the
20th century (Figs. 2, 3; Table 2). The woody vegetation
has expanded upwards in all of the four regions and on
all slopes, with formerly open areas becoming signifi-
cantly more forested. Table 2 represents a synthesis of
all data, showing upward shifts of the upper treeline,
open and closed forest lines by 4.3–8.4 m per decade
during the last 33–57 years. The historical photographs
demonstrate that the forest advances depended upon
substrate and relief. For instance, on the south-west fac-
ing gentle and concave slope of Maly Iremel in the
South Ural (Fig. 2), the closed forest line (40–50% can-
opy cover) shifted by 80 m in altitude and by 600 m in
horizontal distance during the last 80 years. The cover
of the formerly open forest increased from 20% to 60–
70%. By comparison, on the southern stony slope of
Maly Iremel (Fig. 2; on the right hand side in the back-
ground), the boundaries of the open and closed forest
only moved upwards by 20–40 m and horizontally by
100–300 m. In the North Urals on the south-western
slope of the Serebryansky Kamen’, the upper limits of
closed and open larch forests moved by 30–40 m in alti-
tude and horizontally by up to 200–300 m over last
50 years on the slopes with developed soil (Fig. 2).
However, on the steep boulder fields, the forest expan-
sion was only 10–20 m in altitude and 50–150 m in
horizontal distance. Similar changes occurred in the
Sub-Polar Urals (Maly Chender; 750 m a.s.l.), where the
canopy cover of the forest-tundra ecotone in the valley
became more dense between 1939 and 2008 (Fig. 3). In
the Polar Urals, on the south-eastern slope of the Rai-Iz
Massif at about 300 m a.s.l., the canopy cover increased
from 20–30% in 1962 to 50–70% in 2004 (Fig. 3), but the
positions of the uppermost tree individuals on paired
photos hardly changed. During the same period, the
average tree height increased from 8 to 12 m. This pair
of photos also documents that not only the trees but
also the cover and height of the alder layer has
increased significantly in the Polar Urals; young indi-
viduals and patches of Alnaster fruticosus are now grow-
ing on many formerly grass-covered open sites.
Tree age structure
The age structure analysis based on 11 100 trees was
carried out on the same slopes where the historical
photographs documented an upward shift of the for-
est-tundra ecotone. Therefore, the age distributions of
the living trees reaching 300 years back reflect the
development of forest-tundra ecotone since the Little
Ice Age (Fig. 4; Figures. S1, S2, and S3). The analysis
1929
2009
1953
2009
Fig. 2 Treeline advances in the South Urals (‘Maly Iremel’, left) and North Urals (‘Konzhakovsky Kamen’, right) documented by the
comparison between historical and recent photographs.
©2014 John Wiley & Sons Ltd, Global Change Biology, doi: 10.1111/gcb.12613
TREELINE ADVANCES ALONG THE URAL MOUNTAINS 5

of tree ages shows a decreasing mean tree age from
the closed forest line to the treeline. In the South,
North, Sub-Polar, and Polar Urals, mean tree ages
decreased by 50, 100, 50, and 100 years, respectively,
across 100 m in altitude. Moreover, tree ages changed
from mostly even-aged structures in the lower parts
of the forest-tundra ecotone to left-skewed age distri-
butions dominated by trees younger than 70 years in
the upper part of the ecotone. Nine of 11 sites at the
treeline had a predominance of younger trees, and
only one snow-rich convex site in the South Urals
had an even age structure. The age structures show
three main recruitment phases, the first starting
around 1840 and peaking in 1900, the second peaking
in the 1940s, and the third peaking in the 1970s
(Fig. 4). However, the advances of the forest-tundra
ecotone differed strongly among the various altitudi-
nal transects, with slope topography playing a signifi-
cant role in the rate of advance (P<0.01; Fig. 4,
Figures. S1, S2, and S3). On wind-sheltered concave
slopes with thick snow covers, treeline reached higher
altitudes and tree establishment started earliest, fol-
lowed by transects with uniform slopes, and the more
wind-exposed slopes with a shallow snow cover
showed the smallest and latest advances. This pattern
occurred in all four regions (see Figures. 4, S1, S2,
and S3).
In the North Urals, slope topography also deter-
mined tree species composition (Fig. 4): Larix sibirica
grew primarily on convex and uniform slopes, Picea
obovata on concave slopes, and Betula pubescens on very
concave slopes. Our tree age analysis indicated that
these patterns changed during the last century: young
Picea obovata trees started to appear on uniform slopes,
and Betula pubescens trees started to appear on uniform
as well as on concave slopes. In the Sub-Polar Urals,
Betula pubescens additionally started to grow on uni-
form and concave slopes during the last century
(Figure. S2). Our analysis of tree ages showed that the
appearance of the two tree species indeed resulted
from the establishment of ‘new’ tree species and did
not simply reflect shorter regeneration cycles, as both
Betula and Picea can reach much greater ages (at least
200 years) under favorable growing conditions on
other transects.
Changes in growth forms
Trees in the forest-tundra ecotone have two main
growth forms: single-stemmed trees with one upright
stem and multistemmed trees with several upright
stems per tree individual (Fig. 5). The contribution of
these two growth forms changed across the forest-tun-
dra ecotone, with higher contributions of multi-
stemmed trees at the treeline than in the closed forest
(data not shown; for Polar Urals see Devi et al., 2008).
In addition, the contribution of multistemmed trees
was smaller on transects with a concave slope and thick
snow cover than on transects with uniform or convex
slopes (15% vs. 60%; data not shown). Our analysis of
tree ages showed a changing contribution of multi- vs.
single-stemmed trees during the last centuries (Fig. 5).
The multistemmed trees are substantially older than
the single-stemmed ones: 87%, 31%, and 93% of the
stems appearing before 1950 were from multistemmed
trees in the South, North, and Polar Urals respectively.
Thereafter, more than 95% of the stems were from sin-
gle-stemmed trees.
1939
2009
1962
2004
Fig. 3 Treeline advances in the Sub-Polar Urals (‘Maly Chender’) and Polar Urals (‘Rai-Iz’) documented by the comparison between
historical and recent photographs.
©2014 John Wiley & Sons Ltd, Global Change Biology, doi: 10.1111/gcb.12613
6F. HAGEDORN et al.

Microclimate in the forest-tundra ecotone
Microclimatic conditions varied strongly across the for-
est-tundra ecotone during winter but only slightly in
summer. Mean annual air temperatures differed by less
than 1 °C between the three altitudinal levels within
the forest-tundra ecotone. At 10 cm soil depth, mean
temperatures during the growing season were between
8.0 and 11.7 °C at the treeline in the South and Polar
Urals (Table 1). Across forest-tundra ecotones, growing
season temperatures in soils hardly changed, with less
than 1 °C higher temperatures in the closed forest than
at treeline. In contrast with the summer, soil tempera-
tures during winter were less uniformly distributed
within the forest-tundra ecotone. They were around
1°C in the closed forest of all regions but 2°Cto
12 °C at the treeline (Fig. 6). The strong decrease in
soil temperature from the closed forest to the treeline
was related to the strong decline in snow heights in all
transects (Fig. 6), which additionally depended upon
the canopy cover and wind exposure of the transects.
Averaged across all transects and all measurement
campaigns in the South and North Urals, the mean
snow height increased slightly from 77 2 cm in the
closed forest to 91 3 cm at the upper boundary of the
open forest but decreased strongly to 30 2cm
toward the treeline. At treeline, the snow height was
spatially highly heterogeneous. Snow accumulation
was often more than 100 cm on the leeward side of tree
clusters but was less than 25 cm on the windward side
and in open areas. Snow height also differed among
slope types. At treeline, snow heights were less than
30 cm on convex slopes but above 100 cm on concave
slopes (data not shown).
Climate records at weather stations
Climate records at weather stations along the Ural
mountains documented a particularly strong change
in winter conditions during the 20th century (Fig. 7).
For the period 1930–2009, there were no significant
increases in summer temperature at any of the 16
weather stations. Winter temperatures, however,
increased at eight of the stations, and all of these sta-
tions were located south of 59°N (linear regression;
P<0.05; Figure. S4). Precipitation increased signifi-
cantly at 14 of the 16 stations during winter but at
only five stations during summer. The increase in
winter precipitation was greater on the western side
than on the eastern side of the Urals (10 vs. 4 mm
decade
1
; Fig. 7). In relative terms, this increase in
winter precipitation corresponded on average to
50 8% higher values compared to the period from
1930–1940.
Table 2 Altitudinal shifts of treeline, open and closed forest lines on slopes of the Ural mountains in the second half of the 20th century (Means SD). The boundary shifts
were estimated by comparisons of historical with recent landscape photographs (regular font), old topographic maps with modern satellite photos (italic font) and data of histori-
cal and recent tree stand surveys (bold font). More details on measured parameters can be found in Table S1
Study region
South (massif Iremel’) North (Konzhakovsky Kamen’) Sub-Polar (massif Neroika) Polar (surroundings Mnt. Chernaya)
Upper boundary Span year
Altitude shift (m)
Span year
Altitude shift (m)
Span year
Altitude shift (m)
Span year
Altitude shift (m)
Absolute Per decade Absolute Per decade Absolute Per decade absolute per decade
Тreeline 1973–2006*14.1 13.5 4.6 4.1 ––––––1962–2004 28 9.3 6.4 2.2
Оpen forests line 1952–2009 47.9 25.7 8.4 4.5 1956–2005‡42 25.5 8.4 4.9 1970–2009 31 19.5 7.9 3.8 1962–2004§31 12.5 7.1 2.9
Сlosed forests
line
1955–1985†24.6 24.3 8.1 8.1 1956–2009 41 25.0 7.7 4.7 1956–2009 41 22.9 7.7 4.4 1962–2004§29 11.6 6.7 2.7
*Kapralov et al. (2007), stand survey.
†Fomin et al. (2007), historic maps vs. satellite photos.
‡Kapralov et al. (2006) stand survey.
§Shiyatov et al. (2005, 2007), historical vs. recent photographs combined with mapping.
others, this study.
©2014 John Wiley & Sons Ltd, Global Change Biology, doi: 10.1111/gcb.12613
TREELINE ADVANCES ALONG THE URAL MOUNTAINS 7

Tree establishment showed a closer relationship with
winter precipitation than with summer temperatures.
During the first recruitment phase from 1840 until 1925,
mean summer temperatures did not change signifi-
cantly (Fig. 8). Winter precipitation, however, increased
between 70% and 100% from 1840 until 1925 (P<0.05)
and correlated significantly with tree establishment
(r=0.76; P<0.001 for the North Urals; r=0.62,
P<0.01 for the South Urals). In the Polar Urals, the
precipitation record of Salekhard started in 1895 and,
hence, did not allow such an analysis. After the first
recruitment wave, there was no direct relationship
between climatic parameters and tree establishment.
However, the greatest winter precipitation occurred in
years after 1960.
Discussion
Large scale treeline advances
Our results show that the forest-tundra ecotones along
the whole Ural mountain range have changed substan-
tially during the last century. The comparisons of his-
torical and recent photographs, old topographic maps,
and the tree-age analysis along 16 transects all indicate
upward shifts of the upper treeline as well as the open
and closed forest lines in the four different regions of
the Ural mountains. Areas that had a sparse tree cover
of less than 20% at the beginning of the 20th century
are now covered by a dense forest, and thus, the for-
ested areas have advanced upwards. Koshkina et al.
(2008) observed high numbers of seedlings and sapling
within the open-forest areas of the South and Polar
Urals, which implies that the forest expansion is ongo-
ing and that further changes can be expected in the
near future. The photographic comparisons, however,
show that tree species lines have hardly shifted
upwards; the establishment of trees without shelter in
the open tundra is apparently much slower than in
areas where existing trees provide protection against
extreme climatic conditions. We therefore conclude that
the major changes in the forest-tundra ecotone have
been a densification of formerly open forests and an
increase in biomass of existing trees.
In the Polar Urals, there are thousands of 1500 year
old subfossil trees, both standing upright and lying hor-
izontally, at and above the current treeline (Mazepa,
2005). These trees all died between the 7th and the 19th
century and their wood decomposes extremely slowly
in the dry climate with extremely cold winters. The tree
Fig. 4 Tree age structure within the treeline ecotone on slopes with different topographies in the Northern Urals (‘Konzhakovsky Ka-
men’). Columns represent the number of living trees established in a specific 5 years interval (e.g., 1901–1905).
©2014 John Wiley & Sons Ltd, Global Change Biology, doi: 10.1111/gcb.12613
8F. HAGEDORN et al.

remnants show that past climatic changes caused
upward and downward shifts of the forest-tundra eco-
tone (Mazepa, 2005; MacDonald et al., 2007). In addi-
tion, the subfossil wood indicate that the current tree
generation has not reached its former growth limit and
that it lags behind ongoing climatic changes. Moreover,
the presence of subfossil trees implies that the ages and
the spatial distribution of the current living tree genera-
tion mostly reflect the expansion of a ‘new’ forest in
space and time. For the other more southern regions
such as the South and North Urals, where tree rem-
nants do not occur, the historical photographs provide
evidence of a spreading of trees into formerly open
areas.
Our analysis of historical photos and tree-age struc-
tures reveal that the rates of forest expansion have
been very site dependent, with the highest upward
shifts occurring on gentle concave slopes with well-
developed soils. The measured mean elevational shifts
of the treeline ecotone of 4.6–8.4 m per decade are
similar to shifts reported in Siberian mountains further
east. For the Putorana Mountains in Northern Siberia,
a tree-age analysis along an altitudinal gradient indi-
cated an upward shift of the forest-tundra ecotone by
30–50 m during the last century (Kirdyanov et al.,
2012). In Southern Siberia, Kharuk et al. (2010) used
archived maps and satellite images from the Senigelen
Ridge to estimate that the treeline advanced by about
60 m in altitude between 1960 and 2000. Comparable
treeline advances have also been reported for the Scan-
des, where Kullman &
€
Oberg (2009) observed mean
treeline advances of 70–90 m in altitude over the past
century, but only on ‘protected’ concave slopes and
not under wind-exposed topoclimatic conditions. Com-
pared to the rapid forest expansion, the historical pho-
tographs of the Urals did not indicate substantial shifts
of the positions of the uppermost trees and thus, of the
species line (Figs. 2,3). Also, Koshkina et al. (2008) did
not find tree seedlings beyond the current species line
in the South and Polar Urals, strongly suggesting that
the uppermost trees will not move upwards in the
near future. Greater amounts of viable seeds and facili-
tated growth of trees emerging in the shelter of already
established trees are likely reasons for the faster
upward spread of the forested zones than of the spe-
cies lines. Trees in forested zones are less exposed to
wind, are covered under more snow, and grow in war-
mer soils than the uppermost tree individuals that
have to survive in open terrain. At the Alaskan tree-
line, for instance, Lloyd (2005) estimated that the time-
span for an open forest to develop from the first tree
seedling is 150 years, i.e. the time period since the
Little Ice Age.
Snow cover may drive forest expansion
What drives the large scale upward expansion of the
forest-tundra ecotone? Although treeline sites in the
Urals are in rather remote areas, changing pressure by
browsers might have contributed to this change, as
reported for the Subarctic Scandes and the European
Alps (Gehrig-Fasel et al., 2007; Aune et al., 2011; Van
Bogaert et al., 2011). In the Urals, however, hunting
pressure had been consistently intense for centuries: for
fur trading until the 19th century and in organized
hunting during the Soviet time. Hunting has decreased
slightly only in recent years after the breakdown of the
Soviet Union (Korytin, 2011). In the Polar Urals, the
intensity of reindeer herding by nomadic Nentsy peo-
ple has also been continuous for centuries, with a slight
increase during the last decades (Stammler, 2005).
Given that treeline advances occurred along the whole
Ural mountain range, we presume that pressure by
browsers did not change substantially during the last
Fig. 5 Appearance of living multistemmed and single-stemmed
trees in the treeline ecotone of the Ural mountains.
©2014 John Wiley & Sons Ltd, Global Change Biology, doi: 10.1111/gcb.12613
TREELINE ADVANCES ALONG THE URAL MOUNTAINS 9

century and that treeline advances were rather driven
by climatic change.
Treeline advances have often been attributed to cli-
matic warming, particularly summer warming because
temperatures during the vegetation period are thought
to be the main factor controlling treeline formation
(Holtmeier & Broll, 2005; Kharuk et al., 2010). In the
Urals, however, climate records show only very small
increases in summer temperatures, i.e. less than 0.05 °C
per decade (Fig. 7). Our study rather suggests that
changes in winter conditions, particularly increases in
snow fall, might have been the primary cause of the
Fig. 6 Relationship between crown cover and snow height, as well as soil temperatures at 10 cm depth. Measurements of snow heights
and soil temperatures were carried out in March 2006–2008.
Fig. 7 Decadal changes in summer and winter temperature and precipitation from 1930–2000 at weather stations along the Ural moun-
tain range. Open symbols represent climate records from weather stations on the western side of the Urals and filled symbols indicate
stations on the eastern side of the mountain range.
©2014 John Wiley & Sons Ltd, Global Change Biology, doi: 10.1111/gcb.12613
10 F. HAGEDORN et al.

forest advances in the Urals: (i) Warming observed dur-
ing winter was double that observed during summer,
but winter precipitation showed the largest changes
(Fig. 7). In the Ural mountains, winter precipitation has
increased on average by 7 mm per decade. This change
corresponded almost to a doubling of winter precipita-
tion during the 20th century in the Polar Urals. (ii)
Snow cover played an important role in the spatio-tem-
poral dynamics of treeline advances in the Urals. First,
forest-tundra ecotones reached higher altitudes on con-
cave slopes covered with a thick snow pack during
winter (80–200 cm) than on wind-exposed ridges with
a shallow snow cover of less than 30 cm (Fig. 4). Sec-
ond, forest-tundra ecotones experiencing a thick snow
cover mainly started to shift upwards around 1900 AD
while forests on slopes with medium and shallow snow
covers started to expand in the 1950s and 1970s (Fig. 4).
Our conclusion that increasing winter precipitation is
an important driver for the observed shift of the forest
tundra-ecotone is in agreement with a recent satellite
image based survey from the lowlands of Northern
Siberia by Frost & Epstein (2014), which showed that
canopy expansion rates of tall shrubs and trees during
the last decades correlate more closely with winter pre-
cipitation than with summer temperature. A thick snow
cover protects young trees from frost, wind damage,
and abrasion (Holtmeier, 2003). Moreover, snow insu-
lates soils effectively. Our in situ measurements show
higher winter soil temperatures under thick snow
packs, either on wind-protected slopes, in the closed
forests, or on the leeward side of tree clusters than
under shallow snow cover (Fig. 6). Warmer soils are
associated with less damage of the fine root system
(Sveinbj€
ornsson et al., 1996; Groffman et al., 2001) and
can lead to higher nitrogen availability (Sturm et al.,
2005). In the South Urals, Kammer et al. (2009) showed
that much more nitrogen is mineralized in soils in
snow-protected forests than in soils in tundra above the
treeline. As trees affect snow accumulation, the increase
in snow fall, especially in the late 19th and early 20th
centuries, might have induced a positive feedback.
More snow allowed trees to grow better and larger
trees promoted the accumulation of additional snow,
thereby providing more favorable conditions for the
establishment of a new tree generation (Fig. 9). Our
measurement campaigns within the forest-tundra eco-
tone in late winter support such a feedback: when
crown covers were below 20%, snow heights in
between trees were low and soil temperatures reached
as low as 10 °C. However, when tree crowns covered
more than 20% of the area, then snow heights exceeded
50–70 cm and soil temperatures in winter remained
above 2°C. This snow feedback might be accentuated
further when higher summer temperatures (e.g.,
between 1920 and 1940 and during the last two dec-
ades) stimulate the growth of trees thereby enhancing
their role as barriers for drifting snow. The decoupling
of tree establishment with climate after the first recruit-
ment phase might also be due to a feedback between
snow and tree growth (Fig. 8). While the establishment
of the first tree generation correlated positively with
Fig. 9 Feedback between climatic changes and forest expan-
sion.
Fig. 8 Age distribution of established trees in the forest-tundra
ecotone (grey bars) and climate records in the Urals. Summer
temperatures are indicated by the red line, which shows the
moving average over 20 years, and by the red circles, which
represent 5 year averages. Winter precipitation is indicated by
the blue line and blue triangles. Climate records are taken from
nearby weather stations: Salekhard in the Polar Urals, Karpinsk
in the North Urals and Zlatoust in the South Urals.
©2014 John Wiley & Sons Ltd, Global Change Biology, doi: 10.1111/gcb.12613
TREELINE ADVANCES ALONG THE URAL MOUNTAINS 11

winter precipitation (1840–1925), the establishment of
the following tree generation was less directly depen-
dent on the amount of snow fall, probably because
snow accumulating between existing trees facilitated
the survival of seedlings and trees.
Increased winter precipitation is also indirectly
related to winter temperature and permafrost depth,
two other factors discussed in the literature as potential
drivers of treeline advances. In their global meta-analy-
sis, Harsch et al. (2009) found, contrary with their
expectations, that treeline advance was more strongly
associated with winter rather than with summer warm-
ing. They explained this finding by the removal of win-
ter stress. Winter precipitation was not included in
their set of variables, but warmer winter temperatures
usually occur at lower air pressures and are therefore
generally associated with more precipitation. Findings
from our study suggest that increases in snow cover
will have a much larger impact on tree survival than
increases in temperature by a few degrees under very
cold conditions because a thick snow cover effectively
removes winter stress (Holtmeier, 2003 and discussion
above). Permafrost depth was identified as one of the
key factors for the magnitude of treeline advances in
Alaska and Northern Canada (Lloyd, 2005; Payette,
2007). For instance, Suarez et al. (1999) did not detect
treeline advances in tundra soils with less than 0.5 m
deep active layers. Similarly, at the latitudinal treeline
in North-Western Siberia, Wilmking et al. (2012)
observed a lack of tree establishment after the 1980s,
which they attributed to a strong decrease in the active
layer depth toward the treeless tundra. Snow cover is
one of the key drivers of permafrost depth (Zhang,
2005). Hence, increases in winter precipitation and
snow height will very likely promote permafrost melt,
thereby improving conditions for tree establishment.
Reconstructions of climatic conditions indicate that
the 20th century was the wettest period during the last
millennium in Eurasia (Treydte et al., 2006), and instru-
mental records show an increasing precipitation on
land at mid and high latitudes in the Northern Hemi-
sphere during the last century (New et al., 2001). More-
over, climate change scenarios for Eurasia predict
higher precipitation in the coming century, particularly
in winter (Stocker et al., 2013). We therefore suggest
that that increases in snow fall have been and will con-
tinue to be at least as important for treeline advances as
increasing winter temperatures in the Urals and proba-
bly also in other regions of Eurasia.
Changes in growth forms and tree species
The growth form of treeline trees is a response to their
harsh environment, in particular to winter stress (Lavoie
& Payette, 1992; Pereg & Payette, 1998; Holtmeier, 2003;
Harsch & Bader, 2011). At the upper limit of tree
growth, trees often form clusters where several stems
belong to one tree individual. These multistemmed
trees result from winter stress such as wind abrasion,
snow and ice damage, and winter desiccation (Pereg &
Payette, 1998). For the Polar Urals, Devi et al. (2008)
showed that these multistemmed trees grew in a creep-
ing form for centuries but that vertical stems emerged
when conditions improved at the beginning of the 20th
century. In our large scale study, we found multi-
stemmed trees at the treeline in three of four regions in
the Urals. Moreover, the dominant growth form chan-
ged during the last century in the forest-tundra ecotone
of the South, North, and Polar Urals (Fig. 5). While
multistemmed trees were the predominant growth
form 100 years ago, the majority of trees in the new
generation in all regions were single stemmed. We con-
sider these high contributions of multistemmed trees in
the first phase of forest expansion and the recent estab-
lishment of single-stemmed trees as additional evi-
dence that improving winter conditions has been an
important driver for the observed treeline advances.
Another bio-indicator for increasing snow heights
during the last century is the changing tree species
composition in the North and Sub-Polar Urals (Fig. 4).
Betula pubescens, characterized by flexible wood, pri-
marily grows at sites that are protected by a thick snow
cover (Kullman &
€
Oberg, 2009), i.e. on concave slopes.
However, in the mid 20th century Betula also started to
appear on uniform slopes, very likely as a result of
improving snow conditions.
The occurrence of multistemmed trees resulting from
winter stress implies that treelines occur below the
thermal growth limit of trees (Harsch & Bader, 2011). In
our study, this assumption is supported by relatively
high soil temperatures during the growing season.
While the thermal limit for wood formation is approxi-
mately 6 °C (Rossi et al., 2007) and ‘typical’ treelines
occur at soil temperatures of 5.4–7.8 °C during the veg-
etation period (K€
orner & Paulsen, 2004), measured
summer soil temperatures ranged between 8 and
11.7 °C at treeline in the Urals and hence above critical
temperatures. The consequences of relatively high and
hence nonlimiting summer temperatures are that the
forest-tundra ecotone can advance rapidly when other
restricting factors, such as winter conditions, improve.
In addition, our findings suggest that the rate of forest
expansion is particularly fast for treelines with multi-
stemmed trees, where vertical stems develop from
existing but so far creeping trees. In comparison, the
establishment of ‘new’ trees will likely occur more
slowly because these trees must successfully emerge
from seeds and compete with other plants.
©2014 John Wiley & Sons Ltd, Global Change Biology, doi: 10.1111/gcb.12613
12 F. HAGEDORN et al.

In summary, historical photographs and the analy-
sis of tree-age structures in four regions throughout
the 1500 km-long Ural mountain range show that the
forest-tundra ecotone has moved upwards signifi-
cantly during the last century. The closed and open
forest lines have expanded by as much as 50 m in
altitude during the second half of the 20th century
and dominant growth forms of trees have changed
from creeping, shrubby and multistemmed trees to
trees with a single stem. The positions of the upper-
most tree individuals, however, have hardly changed,
presumably because tree establishment is hampered
by the harsher environment and lack of viable seeds
above the species line. Our results suggest that
improving winter conditions, probably in combina-
tion with increasing summer temperatures and
longer vegetation periods, contributed to the large
changes observed in the forest-tundra ecotone. First,
climate records indicate increases in winter precipita-
tion during the last 150 years and second, the forest
expansion started earlier on snow-rich and thus, pro-
tected sites. High seedling and sapling densities in
the open forest indicate that the forest expansion is
ongoing. The likely consequences of treeline advances
are that the newly forested areas will absorb more
solar radiation, particularly during winter (Loranty
et al., 2014; de Wit et al., 2014), and C sequestration
in plant and soils will be altered (Devi et al., 2008;
Kammer et al., 2009). The ongoing forest expansion
also indicates that alpine tundra vegetation will dis-
appear from the South Urals and large parts of the
North Ural, where the treeline is already close to the
highest peaks.
Acknowledgement
This work was performed within the framework of joint projects
conceived by the Institute of Plant and Animal Ecology of the
Ural Branch of the Russian Academy of Science (IPAE), the Ural
State Forest Engineering University (USFEA) and the Swiss Fed-
eral Institute for Forest, Snow, and Landscape Research (WSL).
The project was supported by the following grants: INTAS 01-
0052, RFBR 05-04-48466, RFBR 08-04-00208, RFBR 10-05-00778,
RFBR 11-04-00623, ERA.Net RUS STProject-207. We are thankful
to M. Dawes for her linguistic review and her comments on the
manuscript.
References
Arctic Climate Impact Assessment - Scientific Report (2005) Cambridge University
Press. 1046p. Available at: www.acia.uaf.edu/pages/scientific.html (accessed 5
December 2013).
Aune S, Hofgaard A, S€
oderstr€
om L (2011) Contrasting climate- and land-use-driven
tree encroachment patterns of subarctic tundra in northern Norway and the Kola
Peninsula. Canadian Journal of Forest Research,41, 437–449.
Bolli J, Rigling A, Bugmann H (2007) Regeneration dynamics of Norway spruce (Picea
abies L.) on a subalpine meadow near the treeline in Sedrun, Kt. Graub €
unden Swit-
zerland. Silva Fennica,41,55–70.
Braeker OU (1981) Der Alterstrend bei Jahrringdichten und Jahrringbreiten von
Nadelh€
olzern und sein Ausgleich. Mitteilungen Forstliche Bundes-Versuchsanstalt
Wien,142,75–102.
Devi N, Hagedorn F, Moiseev P et al. (2008) Expanding forests and changing growth
forms of Siberian larch at the Polar Urals treeline during the 20th century. Global
Change Biology,14, 1581–1591.
Fomin VV, Kapralov DS, Terent’ev MM et al. (2007) Spatio-temporal dynamics of
upper tree line in the Southern Urals in the second half of XXth century. Geoinfor-
matika,1,56–61 (in Russian).
Frost GV, Epstein HE (2014) Tall shrub and tree expansion in Siberian tundra eco-
tones since the 1960s. Global Change Biology,20, 1264–1277.
Gehrig-Fasel J, Guisan A, Zimmermann NE (2007) Tree line shifts in the Swiss
Alps: climate change or land abandonment? Journal of Vegetation Science,18,
571–582.
Groffman PM, Driscoll CT, Fahey TJ et al. (2001) Colder soils in a warmer world: a
snow manipulation study in a northern hardwood forest ecosystem. Biogeochemis-
try,56, 135–150.
Harsch MA, Bader MY (2011) Treeline form –a potential key to understanding tree-
line dynamics. Global Ecology and Biogeography,20, 582–596.
Harsch MA, Hulme PE, McGlone MS, Duncan RP (2009) Are treelines advancing? a
global meta-analysis of treeline response to climate warming. Ecology Letters,12,
1040–1049.
Holmes RL (1995) Dendrochronological Program Library (computer program). The Univer-
sity of Arizona, Laboratory of Tree Ring Research, Tucson, Arizona.
Holtmeier F-K (2003) Mountain Timberlines. Ecology. Patchiness, and Dynamics. Kluwer,
Dordrecht.
Holtmeier F, Broll G (2005) Sensitivity and response of northern hemisphere altitudi-
nal and polar treelines to environmental change at landscape and local scales.
Global Ecology and Biogeography,14, 395–410.
Hudson JMG, Henry GHR (2009) Increased plant biomass in a high arctic heath com-
munity from 1981 to 2008. Ecology,90, 2657–2663.
Jia G, Epstein HE, Walker DA (2003) Greening of arctic Alaska. Geophysical Research
Letters,30, 2067.
Kammer A, Hagedorn F, Shevchenko I et al. (2009) Treeline shifts in the Ural moun-
tains affect soil organic matter dynamics. Global Change Biology,15, 1570–1583.
Kapralov DS, Shiyatov SG, Moiseev PA, Fomin VV (2006) Changes in the composi-
tion, structure, and altitudinal distribution of low forests at the upper limit of their
growth in the Northern Ural mountains. Russian Journal of Ecology,37, 367–372.
Kapralov DS, Shiyatov SG, Fomin VV, Shalaumova YV (2007) Spatio-temporal tree-
line dynamics on the South Urals. Izvestiya Saint Petersburg State Forest Technical
Academy, Issue 180, 59–68.
Kharuk VI, Ranson KJ, Im ST, Vdovin AS (2010) Spatial distribution and temporal
dynamics of high-elevation forest stands in southern Siberia. Global Ecology and
Biogeography,19, 822–830.
Kirdyanov AV, Hagedorn F, Knorre AA et al. (2012) 20th century tree-line advance
and vegetation change along an altitudinal gradient in the Putorana Mountains,
northern Siberia. Boreas,41,56–67.
K€
orner C (2012) Alpine Treelines. Functional ecology of the global high elevation tree limits.
Springer, Basel.
K€
orner C, Paulsen J (2004) A world-wide study of high altitude treeline temperatures.
Journal of Biogeography,31, 713–732.
Korytin NS (2011) Changes in the numbers of predatory mammals in the Middle
Urals caused by anthropogenic factors. Russian Journal of Ecology,42, 231–235.
Koshkina NB, Moiseev PA, Goryaeva AV (2008) Reproduction of the Siberian spruce
in the timberline ecotone of the Iremel’ Massif. Russian Journal of Ecology,39,83–91.
Kullman L,
€
Oberg L (2009) Post-little ice age treeline rise and climatic warming in the
Swedish Scandes: a landscape ecological perspective. Journal of Ecology,97,
415–429.
Lavoie C, Payette S (1992) Black spruce growth forms as records of a changing winter,
Quebec, Canada. Arctic and Alpine Research,24,40–49.
Lloyd AH (2005) Ecological histories from Alaskan treelines provide insight into
future change. Ecology,86, 1687–1695.
Loranty MM, Berner LT, Goetz SJ, Jin Y, Randerson JT (2014) Vegetation controls on
northern high latitude snow-albedo feedback: observations and CMIP5 model sim-
ulations. Global Change Biology,20, 594–606.
MacDonald GM, Kremenetski KV, Beilman DW (2007) Climate change and the north-
ern Russian treeline zone. Philosophical Transactions of the Royal Society Biological
Sciences,363, 2285–2299.
Macias-Fauria M, Johnson EA (2013) Warming-induced upslope advance of subal-
pine forest is severely limited by geomorphic processes. Proceedings of the National
Academy of Sciences of the United states of America,110, 8117–8122.
©2014 John Wiley & Sons Ltd, Global Change Biology, doi: 10.1111/gcb.12613
TREELINE ADVANCES ALONG THE URAL MOUNTAINS 13

Mazepa VS (2005) Stand density in the last millennium at the upper timberline eco-
tone in the Polar Ural Mountains. Canadian Journal of Forest Research,35, 2082–2091.
Moiseev PA (2011) Structure and dynamics of woody vegetation at the upper limit its
growth on the Urals mountains. Abstract of Double Doctor Dissertation. Ekaterin-
burg, 42 p (in Russian).
Moiseev PA, Shiyatov SG (2003) Vegetation dynamics at the tree-line ecotone in the
Ural Highlands, Russia. Ecological Studies,167, 423–435.
New M, Todd M, Hulme M, Jones P (2001) Review. Precipitation measurements and
trends in the 20th century. International Journal of Climatology,21, 1899–1922.
Pauli H, Gottfried M, Dullinger S et al. (2012) Recent plant diversity changes on Eur-
ope’s mountain summits. Science,336, 353–355.
Payette S (2007) Contrasted dynamics of northern Labrador treelines caused by cli-
mate change and migration lag. Ecology,88, 770–780.
Payette S, Filion L (1985) White spruce expansion at the tree-line and recent climatic
change. Canadian Journal of Forest Research,15, 241–251.
Pereg D, Payette S (1998) Development of black spruce growth forms at tree-line.
Plant Ecology,138, 137–147.
Rinn F (1998) TSAP V 3.5: Computer program for tree-ring analysis and presentation.
Frank Rinn Distribution, Germany.
Rossi S, Deslauriers A, Anfodillo T, Carraro V (2007) Evidence of threshold tempera-
tures for xylogenesis in conifers at high altitudes. Oecologia,152,1–12.
Shiyatov SG, Terent’ev MM, Fomin VV (2005) Spatiotemporal dynamics of forest-tun-
dra communities in the Polar Urals. Russian Journal of Ecology,36,69–75.
Shiyatov SG, Terent’ev MM, Fomin VV, Zimmermann NE (2007) Altitudinal and hor-
izontal shifts of the upper boundaries of open and closed forests in the polar urals
in the 20th Century. Russian Journal of Ecology,38, 243–248.
Stammler F (2005) Reindeer nomads meet the market. Culture, property and global-
ization at the “End of the Land”. LIT Verlag M€
unster 379 pp.
Stocker TF, Qin D, Plattner G-K et al. 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. Cambridge University Press, Cam-
bridge, UK and NY, USA.
Sturm M, Schimel J, Michaelson G et al. (2005) Winter biological processes could help
convert Arctic tundra to shrubland. BioScience,55,17–26.
Suarez F, Binkley D, Kaye MW (1999) Expansion of forest stands into tundra in the
Noatak National Preserve, northwest Alaska. Ecoscience,6, 465–470.
Sveinbj€
ornsson B, Kauhanen H, Nordell O (1996) Treeline ecology of mountain birch
in the Tornetrask area. Ecological Bulletin,46,65–70.
Treydte KS, Schleser GH, Helle G et al. (2006) The 20th century was the wettest per-
iod in northern Pakistan over the past millennium. Nature,440, 1179–1182.
Van Bogaert R, Haneca K, Jonasson C et al. (2011) A century of tree line changes in
sub-Arctic Sweden shows local and regional variability and only a minor influence
of 20th century climate warming. Journal of Biogeography,38, 907–921.
Weisberg PJ, Baker WL (1995) Spatial variation in tree seedling and krummholz
growth in the forest-tundra ecotone of Rocky Mountain National Park, Colorado.
Arctic and Alpine Research,27,40–49.
Wilmking M, Sanders TGM, Zhang Y et al. (2012) Effects of climate, site conditions
and seed quality on recent treeline dynamics in NW Russia: permafrost and lack
of reproductive success hamper treeline advance. Ecosystems,15, 1053–1064.
de Wit HA, Bryn A, Hofgaard A, Karstensen J, Kvalev
ag MM, Peters GP (2014) Cli-
mate warming feedback from mountain birch forest expansion: reduced albedo
dominates carbon uptake. Global Change Biology. doi: 10.1111/gcb.12483
Zhang T (2005) Influence of the seasonal snow cover on the ground thermal regime:
an overview. Reviews of Geophysics,43, RG4002.
Supporting Information
Additional Supporting Information may be found in the
online version of this article:
Figure S1. Tree age structure of the forest-tundra ecotone in
the South Urals (‘Maly and Bolschoi Iremel’).
Figure S2. Tree age structure of the forest-tundra ecotone in
the Sub-Polar Urals (‘Maly Chender’).
Figure S3. Tree age structure of the forest-tundra ecotone in
the Polar Urals (‘Rai-Iz’).
Figure S4. Summer and winter temperatures and precipita-
tion in the South, North, and Polar Urals. Linear trends are
indicated when significant (P<0.05). Climate records are
taken from nearby weather stations: Salekhard in the Polar
Urals, Karpinsk in the North Urals and Zlatoust in the South
Urals.
Table S1. Overview of measured parameters in this treeline
study in the Ural mountains.
©2014 John Wiley & Sons Ltd, Global Change Biology, doi: 10.1111/gcb.12613
14 F. HAGEDORN et al.