TYPE Original Research

PUBLISHED 14 June 2023

DOI 10.3389/ffgc.2023.1187724

Elevation, aspect, and slope

OPEN ACCESS influence woody vegetation
EDITED BY
Hubert Hasenauer,
University of Natural Resources and Life
structure and composition but
Sciences, Austria
REVIEWED BY
not species richness in a
Romà Ogaya,
Ecological and Forestry Applications Research
Center (CREAF), Spain
human-influenced landscape in
María Leticia Monge-Gonzáelz,
Instituto de Ecología (INECOL), Mexico northwestern Yunnan, China
*CORRESPONDENCE
Zhongqian Cheng
[email protected] Zhongqian Cheng1,2,3*, Tuomas Aakala3 and Markku Larjavaara4*
Markku Larjavaara 1
Key Laboratory for Earth Surface Processes of the Ministry of Education, Peking University, Beijing,
[email protected]
China, 2 Institute of Ecology, College of Urban and Environmental Sciences, Peking University, Beijing,
China, 3 Faculty of Science and Forestry, School of Forest Sciences, University of Eastern Finland,
RECEIVED 16 March 2023 Joensuu, Finland, 4 Department of Forest Sciences, Faculty of Agriculture and Forestry, University
ACCEPTED 19 May 2023 of Helsinki, Helsinki, Finland
PUBLISHED 14 June 2023

CITATION
Cheng Z, Aakala T and Larjavaara M (2023)
Elevation, aspect, and slope influence woody
vegetation structure and composition but not Elevation-for-latitude substitution offers a tool for studying the influence
species richness in a human-influenced of temperature and precipitation variability on vegetation structure and
landscape in northwestern Yunnan, China.
Front. For. Glob. Change 6:1187724. composition. Understanding how elevation, aspect, and slope influence
doi: 10.3389/ffgc.2023.1187724 vegetation patterns may help in predicting how climate change influences
COPYRIGHT human forest usage and in developing strategies for ensuring the sustained
© 2023 Cheng, Aakala and Larjavaara. This is an
provision of ecosystem services. However, most ecological studies have been
open-access article distributed under the terms
of the Creative Commons Attribution License carried out in protected areas, leaving forest areas used by humans to lesser
(CC BY). The use, distribution or reproduction attention. Therefore, we asked how elevation, aspect, and slope impact the
in other forums is permitted, provided the
original author(s) and the copyright owner(s) vegetation on a human-influenced mountain. We measured woody vegetation
are credited and that the original publication in size, richness, and composition on a mountain with plots set systematically in
this journal is cited, in accordance with
accepted academic practice. No use, four cardinal directions at 100-m elevational intervals from the peak, from 1900
distribution or reproduction is permitted which to 4200 m above sea level, in the Hengduan Mountains in eastern Himalaya,
does not comply with these terms.
southwestern China. We quantified how tree maximum height, basal area,
aboveground biomass (AGB), tree and shrub species richness, and woody species
composition changed with elevation, aspect, and slope. Based on generalized
linear models, the maximum tree height, tree basal area, and woody species
AGB followed a unimodal trend along elevational gradients, with tree height and
basal area peaking at 3100 m, while AGB was highest at 3300 m and somewhat
higher on the southern slope. Basal area increased with slope degree. Neither
tree nor shrub species richness was influenced by elevation, aspect, or slope.
According to canonical correspondence analysis and TWINSPAN classification,
elevation and north-south orientation of the slope were major factors influencing
woody species compositions, and vegetation was classified into five types of
communities. Our results indicated that the influences of elevation, aspect, and
slope on woody vegetation structure were similar in a human-influenced forested
mountain area as in protected mountain landscapes based on the literature.
However, as forests in this area are used more intensively at low and middle

Frontiers in Forests and Global Change 01 frontiersin.org

Cheng et al. 10.3389/ffgc.2023.1187724

elevations of the southern and western slopes, where aridity restricts tree size
and AGB, climate change is likely to challenge traditional harvesting practices and
place pressure on moving forest usage to higher altitudes.

KEYWORDS

elevational gradients, aspect, slope degree, vegetation structure, vegetation composition

(Acharya et al., 2011), and living biomass (Lin et al., 2012; Wang
1. Introduction et al., 2014).
Species richness at the community level may vary with elevation
Understanding how vegetation structure and dynamics depend
in four ways: a monotonic decrease, a plateau before descent, and
on climate is a fundamental basis for predicting their phenotypes
hump-shaped and inverse hump-shaped curves. These patterns
among various environmental conditions (Moritz and Agudo,
commonly result from a combination of decreasing temperature
2013), and their fate under a changing climate (Bonan, 2008; Xu
and mountain-specific precipitation patterns along elevational
et al., 2019). Such predictions are important for understanding
gradients (Tang and Fang, 2004; Tang et al., 2012; Yang et al., 2016a).
species range shifts and biodiversity more generally (Elsen and
In addition to the effect of elevation, the slope degree controls
Tingley, 2015). These changes can have detrimental consequences,
vegetation size, composition, and distribution by influencing wind
such as decreasing forest productivity and increasing risk of
speed, soil water and nutrient content, solar radiation intensity,
natural disturbances, for human communities that rely on forest
and seed dispersal distance (Moeslund et al., 2013) and more
ecosystems.
indirectly by influencing fire severities and frequencies (Rogeau
Studying how climate impacts vegetation by field sampling
and Armstrong, 2017; Gowda et al., 2019). On steeper slopes,
large areas using standardized field techniques is arduous and
higher wind speeds lead to shorter trees (Gardiner et al., 2016).
is often hampered by variation in confounding environmental
Deep-rooted species are potentially better adapted to steeper slopes
factors (Sanders and Rahbek, 2012). An alternative is to set such
because of wind conditions but also due to soil stability issues
studies in smaller areas with elevational variation, i.e., substituting
(Moore et al., 2018). Litter depth changes with slope degree due to
elevation for latitude (Jump et al., 2009). As an example, a
sliding caused by wind and gravity, leading to spatial variations in
5.2–6.5◦ C decrease in mean annual temperature every 1000 km soil moisture, nutrient status, and soil temperature (Dulamsuren
toward a higher latitudinal in the temperate region corresponds and Hauck, 2008), which again indirectly influence fire regimes
to a mere 1000 m upslope elevation shift (Colwell et al., 2008). (Luo et al., 2021).
As this elevation-for-latitude substitution allows covering climate Aspect influences solar radiation, precipitation, and wind,
gradients over short distances (Sanders and Rahbek, 2012; Rahbek which in turn impact vegetation composition, structure, and
et al., 2019a), many confounding factors, such as biogeographical growth (Burnett et al., 2008; Moeslund et al., 2013; Qin et al., 2021).
history, time since glaciation, and soil formation, which covary In the temperate region of the Northern Hemisphere, equator-
along latitudinal gradients may change less along elevational facing aspects between the south and southwest experience a
gradients (Körner, 2007). Mountains could consequently be treated remarkably warmer microclimate than other slopes (Perring, 1959).
as relatively uniform and continuous entities at a regional scale This temperature difference has been demonstrated to influence
(Moeslund et al., 2013). Mountain topography can thus serve as species composition and diversity at various scales in several
an excellent substitution of a global environment for studying mountain ranges (Yang et al., 2016b; Heydari et al., 2021; Qin
vegetation distributions and functions in the context of climate et al., 2021). The dryness of equator-facing slopes increases fire
change (Sanders and Rahbek, 2012). severity, consequently influencing vegetation (Beaty and Taylor,
Topography, specifically elevation, slope, and aspect, influence 2001). These could either change the vegetation structure and
vegetation structures and dynamics (Grytnes, 2003; Sanders and species richness (Gowda et al., 2019) or alter fire return intervals
Rahbek, 2012). Of these, the effect of elevation on vegetation is (Rogeau and Armstrong, 2017).
the most studied topographic factor. Atmospheric pressure and Studies on vegetation patterns show that less-human-impacted
air temperature consistently decrease while radiation under a study sites are typically preferred to avoid the confounding factors
cloudless sky and the fraction of ultraviolet-B uniformly increase caused by human influence. However, completely intact forest
with elevation, whereas other factors, such as precipitation, hours landscapes only covered 22% of forested land globally in 2000
of sunshine, growing season length, geology, and human land use, (Potapov et al., 2017). Although many researchers have already
are mountain-specific variables (Körner, 2007). emphasized the importance of understanding human impacts on
A combination of environmental factors at each elevational forest ecosystems at local (Monge-González et al., 2019), regional
belt shapes the vegetation structure and composition by controlling (Qu et al., 2020; Ge et al., 2021), and global scales (Isbell et al., 2017),
the temperature, soil moisture, and light availably (Sanders and most ecological studies were nonetheless conducted in protected
Rahbek, 2012; Moeslund et al., 2013). A low temperature at high areas (Martin et al., 2012). Because of this systematic bias, studies
elevations limits tree growth (Fan et al., 2009; Wang et al., 2014, in unprotected landscapes are especially valuable for understanding
2021; Gaire et al., 2020), height (Kunwar et al., 2021), basal area the current state of forested ecosystems of the world. More

Frontiers in Forests and Global Change 02 frontiersin.org

Cheng et al. 10.3389/ffgc.2023.1187724

importantly, the management of managed landscapes benefits from mountain represented the region. Ultimately, we selected a
evidence-based guidelines that can ensure the sustained provision mountain (27.66–27.73◦ N, 99.45–99.54◦ E) peaking at 4231 m. Its
of ecosystem services in the changing world. base reaches the Jinsha River, a tributary of the Yangtze River,
To set up our research, we chose Wujing county, a sparsely at 1920 m above sea level to the west, while Wujing town is
populated (population density 12.2 people km−2 ) region in located nearby. The majority of settled villages are situated at
northwestern Yunnan, China. As in most of the world, most approximately 3000 m on the western and southern slopes of
vegetation ecological research in northwestern Yunnan has been the mountain. Although the land is owned by the state, the local
carried out in protected areas, but we selected a mountain where community can use the forest with permission.
local people still utilize wood from the forests. We arranged the Before the Natural Forest Protection Program was launched in
sample plots systematically and objectively along cardinal compass 1998 (Hua et al., 2018), large-scale logging operations were carried
directions and contours, thereby avoiding a potential bias from out in the region including the studied landscape. Significant tree
restricting the sampling close to roads or trails. Specifically, we harvesting has probably been carried out, and in the decades
studied how three topographic factors, i.e., elevation, aspect, and before this project, logs have been floating in the Yangtze river for
slope, influence (1) vegetation structure, represented by maximum transportation to eastern China (Isbell et al., 2017). Unfortunately,
height, basal area, and aboveground biomass (AGB); (2) woody at the time of interviewing the local people and discussing the
species richness; and (3) woody species composition in this history of the mountain and residents’ livelihoods, we lacked
human-impacted landscape. For each of these, we hypothesized scientific data on this harvesting history and on other impacts by
that (1) vegetation structure and richness would follow a hump- people on these forests. Nowadays, villagers can apply for cutting
shaped pattern along elevational gradients due to more frequent permits from the local state-owned forest farm to log timber for
disturbances and a more xeric climate at lowest elevations, and building houses or to use as firewood. Meanwhile, when funding is
cold temperatures at highest elevations that constrain vegetation available, the local forest farm calls on villagers to afforest the open
growth; (2) these vegetation attributes would show a linear increase land areas and along the roads, by seeding or planting seedlings.
from cooler to warmer aspects along both north-south and east- Villagers also used to grow crops at some distance from the village
west orientations; and (3) these attributes would also show an among the forests, but this practice was abandoned after the Grain
increase with slope steepness due to less human activity at steeper for Green Project launched in 1999 (Hua et al., 2016). Three ethnic
slopes; (4) Vegetation composition would be driven by elevation, minorities live in the area. Naxi and Lisu peoples mainly cultivate
aspect, and slope degree. lower-elevation land, while Tibetans additionally raise yaks and
cattle at elevations above 3500 m.
During 1958–2021, the mean annual air temperature (MAT)
of the closest meteorological station (28.02◦ N; 99.73◦ E) at
2. Materials and methods 3290 m was 5.2◦ C and the mean annual precipitation (MAP)
was 702 mm, with >90% of the precipitation falling from April
2.1. Location and geography to October according to climatic data derived from the National
Meteorological Information Center of China.1 Soils at altitudes
We chose Shangri-La, Diqing Tibetan Autonomous Prefecture, between 2000 and 4000 m can be classified as yellow cinnamon
in the northwestern corner of Yunnan Province in China as soil, mountain dark brown soil, mountain brown coniferous forest
our study area. It is part of the eastern Himalaya-Hengduan soil, or alpine meadow soil (Huang et al., 2020). In terms of IUCN
Mountain region, which is characterized by dramatic topographical Global Ecosystem Typology (Keith et al., 2020), the study area is
differences. This area is a globally recognized hotspot for tropical or subtropical lowland rainforests, boreal and temperate
biodiversity conservation (Myers et al., 2000; Sloan et al., 2014; high montane forests and woodlands, deciduous temperate forests,
Rahbek et al., 2019b) and a high-priority area for restoration and pastures.
efforts (Strassburg et al., 2020). For example, the area has the
highest number of gymnosperm species in the world (Barthlott
et al., 2007). The climate is controlled by the Indian southwest 2.2. Study design
monsoon in the southwest and the southeast Pacific monsoon in the
southeast (Li et al., 2011; Shen, 2016). Three parallel rivers, Jinsha To determine the sampling plots, we utilized the Advanced
River, Nu River, and Lancang River, running from north to south, Land Observing Satellite (ALOS) DEM with 30-m resolution to
are separated by the Hengduan Mountains. The area has distinct locate the contours. The mountain’s highest point was identified as
dry and wet seasons, with rainy summers and dry winters. The the primary reference point (Peak in Figure 1). Four transects were
dramatic topography results in broad climatic gradients ranging drawn at the cardinal north, east, south, and west of the peak, which
from subtropical to polar along elevational gradients (Liu et al., continued until they intersected with a river or a significant incline
2016). where the subsequent plot would be higher than the previous plot.
We narrowed down our study area to Wujing County Circular plots were placed at the intersections of the contours and
in Shangri-La by balancing accessibility and remoteness, and transects, except in the presence of a village, cropland, or road.
by utilizing the digital elevation model (DEM) to locate the In total, we established 50 plots: eight plots at 3500–4200 m in
greatest elevational difference. To explore the area, we interviewed the north transect, four plots at 3900–4200 m in the east transect,
local people and conducted field visits to various mountains in
Wujing. We tried to avoid recent major construction projects in
the mountains and protected areas, ensuring that the selected 1 http://data.cma.cn/

Frontiers in Forests and Global Change 03 frontiersin.org

Cheng et al. 10.3389/ffgc.2023.1187724

FIGURE 1
Research area location and plot arrangement.

twenty plots at 2200–4200 m in the south transect (excluding situated on the edge of a riverbank that we wished to exclude.
2300 m due to a cliff), and eighteen plots at 2000–4200 m in the west Therefore, we reduced the plot radius to 6 m. Plot S2600 was located
transect (excluding 2100, 2200, 2300, 2600, and 2700 m due to the on abandoned cropland that was surrounded by planted trees. To
presence of villages). North and east transects were shorter because exclude the trees, we set the plot radius to 2.5 m, extending to the
the mountain is asymmetric with longer slopes to deep valleys in border of the planted area. Plot S3200 was located at a cliff edge, and
the south and west. We derived plot information on elevation, we reduced the plot radius to 2.5 m to accommodate the terrain.
aspect in azimuth degree, and slope degree from the DEM. All Starting from the north direction, trees with diameter at
the procedures were carried out using ArcGIS 10.6 software (ESRI, 1.3 m height (DBH) of more than 3 cm were counted clockwise,
USA). mapped, and measured. Trees that were forked below 1.3 m were
counted as separate individuals. Each tree was numbered with
an aluminum tag. Tree mapping was performed by recording
2.3. Field sampling the azimuth (measured by Garmin GPSMAP 66i) and distance
(measured by Haglöf Vertex Laser Geo, Haglöf Ltd., Sweden) of
To maintain a relatively consistent number of trees across the tree from the plot center. The DBH of all trees were measured
all plots, we calculated the radius of each circular plot based with a DBH tape. To reduce the amount of fieldwork, the heights
on the number of trees with a diameter at breast height of the first five trees (measured with a Vertex Laser Geo or a 20-
(DBH) greater than 3 cm found within a 5-m radius plot m fishing rod) were measured for each species in the plots, and
centered around the plot’s midpoint. If more than 25 trees subsequently the height was measured whenever the DBH of a tree
were counted, the plot radius was set at 5 m. If the plot was either larger or smaller than the first five trees. Heights for
contained between 11 and 25 trees, a 10-m radius was used. the remaining trees were estimated based on plot- and- species-
If the plot contained between 6 and 11 trees, a 15-m radius specific height–DBH relationships. Dead standing trees with DBH
was used. If the plot contained less than 6 trees, a 20-m >10 cm were recorded if they were in the early stage of decay,
radius was used. their heights were measured only if the remaining stem was more
The plots were named based on a combination of the compass than half of their final height. Cut stumps were mostly shorter than
direction of each transect and the elevation of the plot. However, 1.3 m, and they are not considered. The species, height, diameter,
for three of the 50 plots, the standard radius could not be used. coverage of the plot in percentage, number, and distribution was
Plot S2200, located at the southernmost point of the transect, was recorded for shrubs with more than one main stem growing from

Frontiers in Forests and Global Change 04 frontiersin.org

Cheng et al. 10.3389/ffgc.2023.1187724

the ground in the 2.5-m radius subplot with the same center. For on individual DBH, plot area, and the number of trees. These
herbaceous vascular plants, their species, height, cover, and number values were then converted to hectares. Shrub species richness
were recorded in the 1-m radius subplot with the same center. was determined by analyzing the data collected from a subplot
Species identification and their scientific name were checked using with a 2.5-m radius.
the Flora of China online edition (Institute of Botany, Chinese Plot aspects in azimuth degrees were sine transformed to
Academy of Sciences). provide information of each plot’s east–west orientation (aspect
EW), from west (sine = −1), north or south (sine = 0), to east
(sine = 1). Similarly, they were cosine transformed to represent
2.4. Statistical analysis north–south orientation (aspect NS), from south (cosine = −1),
east or west (cosine = 0), to north (cosine = 1) (Joly and Gillet,
All statistical data analyses were computed in R language using 2017).
R Studio (Crawley, 2013). Canonical correspondence analysis (CCA) was used to identify
Non-linear regression analysis was carried out to estimate the relationships with species and elevation, aspect EW, aspect NS, and
heights of unmeasured trees. The relationship between height and slope. The proportions of each species out of the total AGB were
DBH was expressed as the function (Zhang et al., 2020) (Eq. 1): used to estimate the importance of each species and was log (1 + x)
transformed. ANOVA and Monte Carlo significance tests were
log (y) = a + b log (x) (1) used to check whether CCA ordination for both axes and for all
where x is DBH, y is height, respectively; a, b are the environmental factors were significant. Multicollinearity between
estimated parameters. environmental factors was evaluated with variance inflation
To obtain comparable values for plots of varying sizes for factors (VIFs). We used variation partitioning to distinguish the
species richness and maximum height, we used the area from 5-m contribution proportion of these four environmental factors to
radius plots, namely 78.5 m2 , as the basis for these two attributes. the vegetation distribution pattern. TWINSPAN classification was
We used the actual measured value when plot radius was 5 m. implemented for grouping plots. Jaccard similarity index was used
Otherwise, we used the relationship between richness and area to classify vegetation community groups. A height 1.5 m was
as the basic model, which also followed the same form as Eq. 1, used where the hierarchical cluster of woody species shall be cut.
where x was area computed based on the distance of the tree to Ordination was carried out with the vegan package (Oksanen
the plot center and y was either the tree species richness or the et al., 2022), and classification was performed with packages vegan
maximum height, respectively (Scheiner, 2003; Lyman and Ames, (Oksanen et al., 2022) and twinspanR (Zeleny, 2021).
2007). All non-linear parameter analyses were carried out with the Generalized linear models (GLM) were used to detect the
nls function in the dplyr package (Hadley et al., 2023). However, this responses of AGB per ha, the estimated maximum tree height
function in R requires more than three observations. Thus, if a plot in the 5-m radius, basal area per hectare, estimated tree species
had two tree species, we recorded the number of species within a 5- richness in the 5-m radius, and shrub species richness in the 2.5-m
m radius. For plots that had only one species, the estimated number radius to elevation, aspect EW, aspect NS, and slope. Considering
was one. that the response variable to elevation could either be monotone
For calculating the AGB of tree species (Mg ha−1 ), we first or a second-order polynomial (Rowe and Lidgard, 2009), the
searched for species- and genus-specific allometric equations. testing environmental independent variables were second-order
However, for Quercus and Juglans these were parameterized with polynomial elevation, elevation, aspect EW, aspect NS, slope,
much smaller trees than what were growing on our plots. Therefore, interaction of elevation and aspect EW, interaction of elevation
instead of extrapolating, which could potentially have led to severe and aspect NS, interaction of elevation and slope, interaction of
biases, we used a pantropical AGB model from Equation 4 in Chave aspect EW and slope, and interaction of aspect NS and slope. We
et al. (2014): used a log-transformed value with Gaussian distribution for AGB,
maximum tree height in the 5-m radius, and basal area per hectare,
AGB = 0.0673 × (ρD2 H)0.976 , (2) while we log transformed the number of species and used the
Poisson distribution for tree and shrub species richness. Akaike
where ρ is wood-specific gravity in g cm−3 derived from a wood
information criterion (AIC) (Akaike, 1973) was used to select the
density dataset in the BIOMASS package; D is DBH of a tree in
best model. χ2 was defined as null deviance minus the residual
cm; H is height in m. This calculation was carried out with the
deviance of the best model. The models were validated by checking
BIOMASS package (Rejou-Mechain et al., 2017).
whether the assumptions are met. The residuals plot for regression
We used species-, genus-, or community-specific allometric
models were presented in Supplementary Figure 1. GLM was
equations for calculating the AGB of shrub species (Zhang and Liu,
carried out using the lme4 package (Bates et al., 2015), and AIC
2010; Wang J. et al., 2012; Xie et al., 2018). We first investigated
evaluation was calculated with the MuMIn package (Bartoń, 2022).
species-specific allometric equations, and if no matching ones were
found, we used equations from the same genus. If no such equations
were available, we used general equations from the vegetation
community located in a region close by that contained similar 3. Results
species. The allometric equations for shrubs are in Supplementary
Table 1. There were 42 plots with trees, including young forests,
In addition to estimating the maximum height of tree species, secondary forests after disturbance, and mature forests. The eight
AGB, and tree species richness, BA per hectare was calculated based treeless plots were all located at least 3900 m above sea level in

Frontiers in Forests and Global Change 05 frontiersin.org

Cheng et al. 10.3389/ffgc.2023.1187724

the west and south transects, except plot S2600, which was on environmental gradients, except for a potential unimodal trend
abandoned cropland. We tagged 1225 trees, 34 of which were of tree species richness along elevational gradients (p = 0.0529;
standing dead trees. The mean tree density of all plots was 1548 Figure 2 and Table 1).
trees ha−1 . Stand density was greatest in a pure Pinus yunnanensis Thirty-nine of the 50 plots contained shrubs. The number
forest in W2400 (6875 trees ha−1 ), followed by 5603 trees ha−1 of species per plot ranged from one to three (Figures 2D1–
in S3200, where Quercus aquifolioides and Pinus yunnanensis were D4). Tree plot E4200 had three shrub species: Rhododendron
the dominant tree species. The third highest density was found in beesianum, Rhododendron rupicola var. chryseum, and Berberis
N4200, 4170 trees ha−1 , with Rhododendron beesianum the only yunnanensis. Fargesia spathacea, Rhododendron beesianum, and
species. Tree density was very low in W3600 and W3700, only 64 Asparagus miscanthus were found in S3000, and Fargesia spathacea,
and 96 trees ha−1 . We recorded 76 woody species, including 59 Rhododendron beesianum, and Cotoneaster franchetii grew in
tree species, 22 shrub species, and 5 species that were either trees S3300. Three shrub species found in treeless plot S4100 were
or shrubs. The five species were: Camellia yunnanensis, Juniperus Rhododendron rupicola var. chryseum, Berberis yunnanensis, and
squamata, Quercus aquifolioides, Rhododendron anthosphaerum, Juniperus squamata. Like tree species richness, shrub species
and Rhododendron beesianum. Their shrub form is defined as richness did not show any clear trends along environmental
lacking a clear main stem. gradients (Figure 2 and Table 1).
In the 42 tree plots, the estimated maximum height in the 5- Elevation, aspect EW, aspect NS, and slope showed significant
m radius plot ranged from a 3.5-m Rhododendron beesianum in effects on woody vegetation community composition in the study
N4200 to a 27.8 m Quercus aquifolioides in S3300 (Figures 2A1– area, explaining 51.9% of the variation in vegetation composition
A4). The tallest tree for which height was measured was an data (F = 2.0669, p = 0.001). Elevation explained 27.9% of the
Abies georgei, 31.8 m in N3900. Elevation strongly influenced the variation, while aspect EW, aspect NS, and slope explained 11.2%,
estimated maximum height in the 5-m radius with a hump-shaped 19.3%, and 10.0% of the variation, respectively. The interactions
trend along the elevational gradient (Table 1 and Figure 2A1). between these topographic factors have less significant impacts than
The BA for all 42 tree plots ranged from 1.5 to 85.4 m2 ha−1 their individual effects (see Supplementary Figure 3).
(Figures 2B1–B4). The smallest BA occurred in W2000, which was The eigenvalue and proportion of explanatory variables in
just above the main road and the Yangtze River, dominated by the selected CCA1 and CCA2 were 0.870 and 6.40% (F = 3.408,
Pistacia weinmanniifolia. The largest BA was recorded in S3200, p = 0.001), and 0.563 and 4.14 (F = 2.205, p = 0.002), respectively.
with Quercus aquifolioides as the dominant species. Both elevation The first two axes of CCA accumulate explained 67.9% of
and slope had significant effects on BA (Table 1). BA was higher at the relationship between environmental factors and vegetation
medium elevations than at low and high elevations (Figure 2B1). composition. Elevation and aspect EW were negatively correlated,
According to the AIC score of the GLM model selection, slope and aspect NS and slope were positively correlated to CCA1.
showed one of the lowest AIC scores despite being insignificant in Elevation, aspect EW, and aspect NS were positively connected,
the full model (p = 0,122; Table 1), and BA increased with slope and slope was negatively connected to CCA2 (Table 2). CCA1 was
degree (Figure 2B4). mainly constituted by elevation and slope, CCA2 was composed
The AGB of woody species for all plots ranged from <0.1 to by aspect NS and aspect EW. Elevation and aspect NS were
656.3 Mg ha−1 , with a mean value of 128.8 Mg ha−1 (Figures 2C1– the most relevant variables in predicting vegetation community
C4). For the eight treeless plots, the total woody species AGBs compositions, accounting for 9.88 and 7.14% of the variation,
were <1.0 Mg ha−1 . The lowest AGB in the tree plots was W2000 respectively.
(3.1 Mg ha−1 ). The highest AGBs were found in S3200, S3300, The ordination of 50 sample plots indicated that all studied
and W3400 (480.3, 602.9, and 656.3 Mg ha−1 , respectively), which plots can be divided into five groups based on their woody
were all dominated by Quercus aquifolioides. The AGBs of the species biomass proportion (Figure 3). Both W2000 (green color
two deadwood-dominated plots, W3600 and W3700, were 34.4 in Figure 3), located close to the Yangtze River, and S2600 (sky
and 23.1 Mg ha−1 , respectively. AGB was significantly influenced blue color in Figure 3), which was on abandoned cropland, were
by elevation (Table 1). The trend of AGB along elevational isolated plots, as they did not correlate with the other plots. S2200–
gradients followed a hump-shaped curve, which peaked at 3300 m 2500, S2700–3300, W2400–2800, and W3000–3100 (red color in
(Figure 2C1). Even though the residuals of AGB against aspect Figure 3) were positively correlated, and they were mainly tree
NS showed heteroscedasticity, AGB decreased slightly from the plots at medium to low elevation in the west and south transects.
south to the north aspect (Figure 2C3). Furthermore, AGB W2900, W3300–3900, S3400–3800, E3900–4100, and N3500–4200
was significantly influenced by their interaction (Table 1 and (black color in Figure 3) illustrated positive correlations among
Supplementary Figure 2). each other and were in the medium to high elevation. S3900, S4000,
W3000 had the highest estimated tree species richness with S4100, S4200, W4000, W4100, W4200, and E4200 (navy blue in
12 tree species. S2700, S2800, and W3200 had 12, 10, and 11 Figure 3) were positively correlated, representing alpine shrub and
tree species within the 10-m radius, which resulted in estimated grassland in the south and west transects. Specifically, W2900,
5-m radius plot values of 4.3, 6.0, and 6.6 species, respectively W3300–3600, and S3400 were placed in between two communities
(Figures 2D1–D4). There were only 0.24 tree species in the 5- (red and black ones), meaning they shared some similar woody
m radius in W3300, yet the plot’s actual tree species richness is 7 species compositions. Similarly, E4200 was located close to the tree
dues to the plot radius being 20 m. Similarly, the estimated tree plots at a high elevation in the north and east transects because it is
richness in the 5-m radius was 0.36 species in W3700, while the in forested part of the tree and alpine grassland ecotone.
measured number in the 20-m radius plot was 5. We observed The ordination of 76 sample species illustrated five vegetation
no clear pattern of tree species richness in the 5-m radius along community types (Supplementary Figure 4, same below) with four

Frontiers in Forests and Global Change 06 frontiersin.org

Cheng et al. 10.3389/ffgc.2023.1187724

FIGURE 2
Relationship between elevation, aspect east–west orientation (EW), aspect north-south orientation (NS), slope, and (A1–A4) estimated maximum
tree heights in the 5-m radius (m), (B1–B4) basal area (BA, m2 ·ha−1 ), (C1–C4) aboveground biomass (AGB; Mg·ha−1 ), (D1–D4) estimated tree
species richness in the 5-m radius, and (E1–E4) shrub richness in the 2.5-m radius. Smoothing lines for one given environmental variable to the
response variables were fitted by GLMs (p < 0.05), and median values of other variables were used.

TABLE 1 Summary of GLM regression.

Response variables Estimated regression Coefficients T-value P of χ2 P-value of R2 *

parameters parameters model*
Estimated maximum height in Intercept −8.400 −2.699 0.011* 333.7 <0.001*** 0.330
the 5-m radius (m)
Elevationˆ2 < − 0.001 −3.618 0.001***

Elevation 0.007 3.537 0.001**

BA (m2 ·ha−1 ) Intercept −12.200 −1.999 0.053 5164 <0.001*** 0.307

Elevationˆ2 < − 0.001 −2.430 0.020*

Elevation 0.009 2.426 0.020*

Slope 0.023 1.580 0.122

AGB Mg·ha−1 Intercept −23.248 −0.244 0.808 589888 <0.001*** 0.555

Elevationˆ2 −485.138 −3.517 0.902

Elevation 17.252 0.123 0.001***

AspectNS −10.252 −3.111 0.003**

AspectNS: Elevation 0.004 3.129 0.003**

*Coefficient of determination calculated based on the likelihood-ratio test (R_LR2 ). ***P < 0.001, **P < 0.01, *P < 0.01. BA denotes basal area, AGB denotes aboveground biomass.

Frontiers in Forests and Global Change 07 frontiersin.org

Cheng et al. 10.3389/ffgc.2023.1187724

TABLE 2 Conditional effects table from canonical correspondence analysis (CCA) ordination showing the partial effect of each environmental factor,
representing the F-value (and its significance) on woody vegetation community composition, its variance inflation factor (VIF), and its correlations
with two CCA axes.

Environmental variables Variance (%) F P VIF CCA1 CCA2

Elevation 10.83 3.348 0.001*** 1.374 -0.985 0.057
Aspect EW 1.16 1.514 0.026* 1.099 -0.233 0.121
Aspect NS 8.57 2.164 0.001*** 1.126 0.217 0.953
Slope 3.28 1.242 0.170 1.192 0.487 −0.051
EW, east-west orientation; NS, north-south orientation. ***P < 0.001, *P < 0.01.

distinct vegetation communities. Pseudotsuga, Padus, Camellia,

Pinus, and Cinnamomum occurrences (green in Supplementary
Figure 4) correlated positively with each other at medium elevation
on a moderate slope. Castanopsis, Sophora, and Malus (navy blue
in Supplementary Figure 4) had positive associations and were
located at lower altitudes than the previous community, with a
more southern or western aspect. The physical environment of
Cornus, Acer, Machilus, Toxicodendron, etc. (red in Supplementary
Figure 4) was quite similar to Castanopsis, Sophora, and Malus, but
could still be divided into two distinct communities. Tilia, Tsuga, FIGURE 3
Rhododendron, Crataegus, Taxus, etc. (light blue in Supplementary Canonical correspondence analysis (CCA) ordination of 50 plots
Figure 4) were located at medium elevation on the southern or classified into four vegetation communities using WA scores
(calculate based on plot, same below) classified by TWINSPAN
western aspect on a medium slope. The remaining species were classification.
positively connected to each other regardless of variations in their
environmental factors, which could possibly be explained by their
wide distributions in the study area.
vary greatly, temperature becomes the most limiting climate factor
in AGB. In northern Pakistan, for instance, forest AGB presented
a monotonous decline from 850 to 8600 m, where mean annual
4. Discussions precipitation is consistently lower than at 200 m while the mean
annual temperature ranges from 40 to −10◦ C (Ullah et al., 2021).
Our study demonstrated the influence of elevation, aspect, and
Both natural and anthropogenic disturbances influence
slope on the woody vegetation in a human-impacted landscape.
vegetation patterns in unprotected mountain landscapes. Villages
Consistent with previous research (Wang et al., 2014; Ullah
are mainly located below 3000 m and activities such as wood
et al., 2021), we found that elevation is a significant predictor of
harvesting, grazing, and igniting fires, both intentional and
vegetation size. Specifically, the estimated maximum height, BA,
unintentional, have likely decreased tree size and AGB in lower
and AGB followed a unimodal trend along the elevational gradient,
elevations where human presence is more frequent, making it
peaking at 3100 m for maximum tree height and BA, and at
3300 m for AGB, before decreasing at higher elevations. However, occasionally difficult to distinguish the climatic effects of elevation
we did not observe a clear relationship between species richness from human use of the forest. As an example, Acharya et al.
and any topographical attributes. Additionally, BA increased with (2011) found a unimodal pattern of BA from 300 m to 4700 m
slope steepness, while AGB decreased with a south to north aspect. in the eastern Himalayas to be due to human disturbances
We also generated separate visualizations of the five vegetation from surrounding lower elevation villages. Occasionally ancient
attributes along elevation for each of the four transects (see human-use of the forest can be inferred from vegetation structure
Supplementary Figure 5). The compositions and distributions of and composition (Foster et al., 1992). In our study area, the
vegetation communities were primarily determined by elevation shrub growth form of Quercus aquifolioides at low elevations was
and north–south orientation. potentially one such indicator.
Vegetation size and biomass largely influenced by climatic The range of elevation sampled may have influenced observed
factors, e.g., mean annual temperature, total annual precipitation, patterns of vegetation structure along elevational gradients. Our
and annual solar radiation that determine vegetation patterns along findings differ from those of Alves et al. (2010), who found AGB
environmental gradients (Šímová et al., 2018; O’Brien et al., 2000). to increase with elevation up to 1100 m in a coastal Atlantic
We found maximum tree height, BA, and AGB to each follow a forest, and from those of Wang Y. et al. (2012) and Wang et al.
unimodal pattern along the elevational gradients. The climate in (2014), who found that maximum tree height and AGB of Abies
our research area is reportedly dry at lower elevations of the valley georgei decreased along elevational gradients, peaked at 3800 m and
in the west transect facing the Yangtze River (Shen, 2016; Yang 3100 m, respectively, in the Tibetan Plateau. Our study covered an
et al., 2016a), making aridity one of the constraint factors limiting elevation range of 2000–4200 m above sea level, with structural
tree size (Munne-Bosch, 2018) and biomass accumulation (Muller- attributes peaking around 3000 m and declining as elevation
Landau et al., 2021) at lower elevations. If precipitation does not increased. This emphasizes the importance of sampling the entire

Frontiers in Forests and Global Change 08 frontiersin.org

Cheng et al. 10.3389/ffgc.2023.1187724

elevational gradient to accurately describe the relationship between (Moeslund et al., 2013). We found elevation and aspect to be
vegetation and the environment. the two most relevant variables defining vegetation composition,
We observed a slightly higher AGB at the southern aspect than with slope degree playing a smaller role. Similarly, Fadl et al. (2021)
the northern aspect, which was consistent with a result focusing on reported the distribution of vascular plant species to be strongly
the effects of elevation on vegetation, where a warmer climate in the affected by elevation and NS aspect at an elevation range from 1060
southern aspect is conducive to AGB accumulation (Pepin et al., to 1240 m in Sarawat Mountain, Saudi Arabia. Bai et al. (2021)
2017). demonstrated elevation to be the primary driving factor of vascular
The influence of elevation on aboveground biomass (AGB) can plant distribution in a Larix gmelinii forest in northeast China,
vary depending on the north-south orientation of aspect, and vice while aspect was the second most important factor and slope had
versa. In our study, we observed that AGB tends to be higher at only minor importance.
mid-elevation on south-facing slopes. Our study had limitations that either constrained the external
In addition, BA increased slightly with slope degree, which is use of the results or called for further research. Firstly, our sampling
less straightforward to explain. However, one possible explanation design, based on contours, led to a slight overrepresentation of
is that steep slopes may reduce fire occurrence and spread due to a steep slopes. Secondly, our research was based on landscape-
lack of fuel, allowing trees to grow taller and have larger basal area scale sampling, and the results should not be assumed to hold
compared to areas with more frequent fires (Morandini et al., 2002; in other regions. Thirdly, earlier studies have shown that forests
Gowda et al., 2019). in southwestern China are commonly disturbed by fire (Han
Although woody species richness appears to reach its et al., 2015) and potentially other major natural disturbances,
highest values at mid-elevations, none of the trends of woody which should optimally be included in forest structure studies.
species richness along environmental gradients encompassed in Future research on disturbance history would therefore be valuable.
the topographic variables were statistically significant. This is Topographic attributes indirectly affected vegetation via different
inconsistent with many studies conducted in protected forests climatic variables at each site. Detecting climatic factors at each
located close to our research area (Wang et al., 2007; Xu et al., 2016; plot to link with the topographic attributes would be helpful for
Yang et al., 2016a). However, this finding is consistent with earlier understanding the influence of the local climate on vegetation
findings concerning the absence of β-diversity trends for tree and in mountainous regions. Despite these problems, our results are
shrub species along elevational gradients in 46 mountains of China valuable for understanding vegetation structure changes along
(Tang et al., 2012) that were well protected only from 1997 to 2008. environmental gradients, but the picture would improve further
Topographic heterogeneity is strongly connected with species
with more information on these listed problems.
diversity (Stein et al., 2014). Other researchers have concluded
Our study highlights that differences in vegetation structure,
from nearby protected areas that vascular plant richness along
richness, and composition were correlated with elevation, aspect,
elevational gradients, ranging from either 3100–4300 m or from
and slope degree, which can be used as a baseline for predicting
2000 to 4300 m in the Hengduan Mountain presented a unimodal
AGB shifting under climate change. At higher elevations, close
shape, with highest richness in the mid- elevational range (Wang
to the peak of our research area, AGB decreased with increasing
et al., 2007; Yang et al., 2016b). We did not find a clear trend of
elevation, suggesting an increase under a warming climate. While
woody species richness changes along EW or NS orientations. Yang
at elevations of 2000–3300 m in this study, warmer temperatures
et al. (2016b) found 300 vascular plant species on the western aspect
are likely to reduce AGB via increased disturbances, assuming that
compared with 501 vascular plant species on the eastern aspect of
precipitation is not changing.
Baima Snow Mountain Reserves, very close to our study area. Our
Our results on tree size and biomass could be used to
results demonstrated that tree species richness tended to increase
provide recommendations for designing forest management plans
with slope degree, although this was not statistically significant.
including carbon sequestration objectives. As most settlements in
Elevation and NS aspect were the two dominant topographical
this area are below 3000 m and harvesting Pinus and Quercus trunks
factors determining woody species distributions in our study area.
currently occurs mainly at this range of elevation, increasing aridity
Slope had a relatively minor effect on these factors. The impact
caused by climate change is likely to cause challenges, eventually
of the NS aspect was much greater than that of the EW aspect.
Woody vegetation in the 50 sample plots could be grouped into leading to a pressure to move human activities upward.
five types: hot-dry valley shrub-like forests, abandoned cropland,
mid-low warm aspect forests, high elevation forests, and alpine
shrub and grassland. When considering the indicator woody Data availability statement
species in the area, all woody species classified into five vegetation
communities, four of which placed in their distinct environmental The original contributions presented in this study are included
conditions. Pseudotsuga, Pinus, and Cinnamomum communities in the article/Supplementary material, further inquiries can be
were placed at medium elevation, both the Acer community and directed to the corresponding authors.
Castanopsis community were at the medium-to-low elevation on a
steeper slope, Tilia, Tsuga, Taxus, and Picea communities were in
the medium elevation at a warm aspect. The remaining species,
though grouped into one, presented little information of their Author contributions
environmental preferences.
Previous studies have indicated that topographic factors ZC participated in the collection of field data, performed the
are major drivers of vascular plant distribution patterns statistical analyses, prepared the figures, and wrote the original

Frontiers in Forests and Global Change 09 frontiersin.org

Cheng et al. 10.3389/ffgc.2023.1187724

manuscript. TA provided insight into the data used and the Publisher’s note
mathematical aspects of the methods. ML conceived the idea of
the manuscript, applied for funding, planned the experiments, and All claims expressed in this article are solely those
participated in the collection of field data. All authors discussed the
of the authors and do not necessarily represent those of
results, thoroughly revised the manuscript, contributed critically to
their affiliated organizations, or those of the publisher,
the drafts, and gave final approval for publication.
the editors and the reviewers. Any product that may be
evaluated in this article, or claim that may be made by
its manufacturer, is not guaranteed or endorsed by the
Funding
publisher.
This research was supported by the National Natural Science
Foundation of China (No. 32171539).

Conflict of interest Supplementary material

The authors declare that the research was conducted in the The Supplementary Material for this article can be found online
absence of any commercial or financial relationships that could be at: https://www.frontiersin.org/articles/10.3389/ffgc.2023.1187724/
construed as a potential conflict of interest. full#supplementary-material

References
Acharya, B. K., Chettri, B., and Vijayan, L. (2011). Distribution pattern of trees Fadl, M. A., Al-Yasi, H. M., and Alsherif, E. A. (2021). Impact of elevation and slope
along an elevation gradient of Eastern Himalaya, India. Acta Oecol. 37, 329–336. aspect on floristic composition in wadi Elkor, Sarawat Mountain, Saudi Arabia. Sci.
doi: 10.1016/j.actao.2011.03.005 Rep. 11:16160. doi: 10.1038/s41598-021-95450-4
Akaike, H. (1973). “Information theory and an extension of the maximum Fan, Z.-X., Bräuning, A., Cao, K.-F., and Zhu, S.-D. (2009). Growth–climate
likelihood principle,” in International symposium on information theory, eds B. N. responses of high-elevation conifers in the central Hengduan Mountains,
Petrov and F. Csaki (Budapest: Akademiai Kiado), 267–281. southwestern China. For. Ecol. Manag. 258, 306–313. doi: 10.1016/j.foreco.2009.04.017
Alves, L. F., Vieira, S. A., Scaranello, M. A., Camargo, P. B., Santos, F. A. M., Joly, Foster, D. R., Zebryk, T., Schoonmaker, P., and Lezberg, A. (1992). Post-settlement
C. A., et al. (2010). Forest structure and live aboveground biomass variation along an history of human land-use and vegetation dynamics of a Tsuga canadensis (Hemlock)
elevational gradient of tropical Atlantic moist forest (Brazil). For. Ecol. Manag. 260, woodlot in central New England. J. Ecol. 80, 773–786. doi: 10.2307/2260865
679–691. doi: 10.1016/j.foreco.2010.05.023
Gaire, N. P., Fan, Z.-X., Bräuning, A., Panthi, S., Rana, P., Shrestha, A., et al.
Bai, X., Sadia, S., and Yu, J. (2021). Community composition and structure along (2020). Abies spectabilis shows stable growth relations to temperature, but changing
environmental gradients of Larix gmelinii forest in northeast China. Pakistan J. Bot. response to moisture conditions along an elevation gradient in the central Himalaya.
53, 1845–1850. doi: 10.30848/PJB2021-5(24) Dendrochronologia 60:125675. doi: 10.1016/j.dendro.2020.125675
Barthlott, W., Hostert, A., Kier, G., Küper, W., Kreft, H., Mutke, J., et al. (2007). Gardiner, B., Berry, P., and Moulia, B. (2016). Review: wind impacts on plant
Geographic patterns of vascular plant diversity at continental to global scales. growth, mechanics and damage. Plant Sci. 245, 94–118. doi: 10.1016/j.plantsci.2016.
Erdkunde 61, 305–315. doi: 10.3112/erdkunde.2007.04.01 01.006
Bartoń, K. (2022). MuMIn: multi-model inference. R package version 1.46.0. Available Ge, W., Deng, L., Wang, F., and Han, J. (2021). Quantifying the contributions of
online at: https://CRAN.R-project.org/package=MuMIn (accessed January 20, 2022). human activities and climate change to vegetation net primary productivity dynamics
in China from 2001 to 2016. Sci. Total Environ. 773:145648. doi: 10.1016/j.scitotenv.
Bates, D., Maechler, M., Bolker, B., and Walker, S. (2015). Fitting linear mixed-effects
2021.145648
models using lme4. J. Stat. Softw. 67, 1–48. doi: 10.18637/jss.v067.i01
Gowda, J. H., Tiribelli, F., Mermoz, M., Kitzberger, T., and Morales, J. M.
Beaty, R. M., and Taylor, A. H. (2001). Spatial and temporal variation of fire regimes
(2019). Fragmentation modulates the response of dichotomous landscapes to
in a mixed conifer forest landscape, Southern Cascades, California, USA. J. Biogeogr.
fire and seed dispersal. Ecol. Modell. 392, 22–30. doi: 10.1016/j.ecolmodel.2018.
28, 955–966. doi: 10.1046/j.1365-2699.2001.00591.x
10.014
Bonan, G. B. (2008). Forests and climate change: Forcings, feedbacks, and the Grytnes, J. A. (2003). Species-richness patterns of vascular plants along seven
climate benefits of forests. Science 320, 1444–1449. doi: 10.1126/science.1155121 altitudinal transects in Norway. Ecography 26, 291–300. doi: 10.1034/j.1600-0587.
Burnett, B. N., Meyer, G. A., and McFadden, L. D. (2008). Aspect-related 2003.03358.x
microclimatic influences on slope forms and processes, northeastern Arizona. Hadley, W., Romain, F., Lionel, H., Kirill, M., and Davis, V. (2023). dplyr: a grammar
J. Geophys. Res. 113:F03002. doi: 10.1029/2007JF000789 of data manipulation. R package version 1.1.0. Available online at: https://CRAN.
Chave, J., Rejou-Mechain, M., Burquez, A., Chidumayo, E., Colgan, M. S., Delitti, Rproject.org/package=dplyr (accessed January 20, 2022).
W. B., et al. (2014). Improved allometric models to estimate the aboveground Han, J., Shen, Z., Ying, L., Li, G., and Chen, A. (2015). Early post-fire regeneration
biomass of tropical trees. Glob. Change Biol. 20, 3177–3190. doi: 10.1111/gcb. of a fire-prone subtropical mixed Yunnan pine forest in Southwest China: effects of
12629 pre-fire vegetation, fire severity and topographic factors. For. Ecol. Manag. 356, 31–40.
Colwell, R. K., Brehm, G., Cardelus, C. L., Gilman, A. C., and Longino, J. T. (2008). doi: 10.1016/j.foreco.2015.06.016
Global warming, elevational range shifts, and lowland biotic attrition in the wet Heydari, M., Cheraghi, J., Omidipour, R., Mirab-balou, M., and Pothier, D. (2021).
tropics. Science 322, 258–261. doi: 10.1126/science.1162547 Beta diversity of plant community and soil mesofauna along an elevational gradient in
Crawley, M. J. (2013). The R Book, 2nd Edn. Chichester: John Wiley & Sons Ltd. a mountainous semi-arid oak forest. Commun. Ecol. 22, 165–176. doi: 10.1007/s42974-
021-00046-7
Dulamsuren, C., and Hauck, M. (2008). Spatial and seasonal variation of climate
on steppe slopes of the northern Mongolian mountain taiga. Grassl. Sci. 54, 217–230. Hua, F., Wang, X., Zheng, X., Fisher, B., Wang, L., Zhu, J., et al. (2016).
doi: 10.1111/j.1744-697X.2008.00128.x Opportunities for biodiversity gains under the world’s largest reforestation
programme. Nat. Commun. 7:12717. doi: 10.1038/ncomms12717
Elsen, P. R., and Tingley, M. W. (2015). Global mountain topography and the
fate of montane species under climate change. Nat. Clim. Change 5, 772–776. doi: Hua, F., Xu, J., and Wilcove, D. S. (2018). A new opportunity to recover native forests
10.1038/nclimate2656 in China. Conserv. Lett. 11:e12396. doi: 10.1111/conl.12396

Frontiers in Forests and Global Change 10 frontiersin.org

Cheng et al. 10.3389/ffgc.2023.1187724

Huang, B., Lu, S., and Zhang, G. (2020). Soil series of China: central and western in controlling mountain climate. Glob. Planet. Change 157, 244–258. doi: 10.1016/j.
regions vol., Yunnan vol. Beijing: Science press. gloplacha.2017.08.006
Isbell, F., Gonzalez, A., Loreau, M., Cowles, J., Diaz, S., Hector, A., et al. (2017). Perring, F. (1959). Topographical gradients of chalk grassland. J. Ecol. 47, 447–481.
Linking the influence and dependence of people on biodiversity across scales. Nature doi: 10.2307/2257376
546, 65–72. doi: 10.1038/nature22899
Potapov, P., Hansen, M. C., Laestadius, L., Turubanova, S., Yaroshenko, A., Thies, C.,
Joly, D., and Gillet, F. (2017). Interpolation of temperatures under forest cover et al. (2017). The last frontiers of wilderness: tracking loss of intact forest landscapes
on a regional scale in the French Jura Mountains. Int. J. Climatol. 37, 659–670. from 2000 to 2013. Sci. Adv. 3:e1600821. doi: 10.1126/sciadv.1600821
doi: 10.1002/joc.5029
Qin, Y., Feng, Q., Adamowski, J. F., Zhu, M., and Zhang, X. (2021). Community
Jump, A. S., Matyas, C., and Penuelas, J. (2009). The altitude-for-latitude disparity level response of leaf stoichiometry to slope aspect in a montane environment: a
in the range retractions of woody species. Trends Ecol. Evol. 24, 694–701. doi: 10.1016/ case study from the Central Qilian Mountains, China. Glob. Ecol. Conserv. 28:e01703.
j.tree.2009.06.007 doi: 10.1016/j.gecco.2021.e01703
Keith, D. A., Ferrer-Paris, J. R., Nicholson, E., and Kingsford, R. T. (2020). ). Qu, S., Wang, L., Lin, A., Yu, D., Yuan, M., and Li, C. A. (2020). Distinguishing
The IUCN global ecosystem typology 2.0: descriptive profiles for biomes and ecosystem the impacts of climate change and anthropogenic factors on vegetation dynamics in
functional groups. Gland: IUCN. doi: 10.2305/IUCN.CH.2020.13.en the Yangtze River Basin, China. Ecol. Indic. 108:105724. doi: 10.1016/j.ecolind.2019.
Körner, C. (2007). The use of ‘altitude’ in ecological research. Trends Ecol. Evol. 22, 105724
569–574. doi: 10.1016/j.tree.2007.09.006 Rahbek, C., Borregaard, M. K., Antonelli, A., Colwell, R. K., Holt, B. G., Nogues-
Kunwar, S., Wang, L. Q., Chaudhary, R., Joshi, P. R., Ali, A., and Ward, D. (2021). Bravo, D., et al. (2019a). Building mountain biodiversity: geological and evolutionary
Stand density of co-existing species regulates above-ground biomass along a local-scale processes. Science 365, 1114–1119. doi: 10.1126/science.aax0151
elevational gradient in tropical forests. Appl. Veg. Sci. 24:e12577. doi: 10.1111/avsc. Rahbek, C., Borregaard, M. K., Colwell, R. K., Dalsgaard, B., Holt, B. G., Morueta-
12577 Holme, N., et al. (2019b). Humboldt’s enigma: what causes global patterns of mountain
biodiversity? Science 365, 1108–1113. doi: 10.1126/science.aax0149
Li, Z., He, Y., Wang, C., Wang, X., Xin, H., Zhang, W., et al. (2011). Spatial
and temporal trends of temperature and precipitation during 1960–2008 at the Rejou-Mechain, M., Tanguy, A., Piponiot, C., Chave, J., and Herault, B. (2017).
Hengduan Mountains. China. Quat. Int. 236, 127–142. doi: 10.1016/j.quaint.2010. BIOMASS: an R package for estimating above-ground biomass and its uncertainty in
05.017 tropical forests. Methods Ecol. Evol. 8, 1163–1167. doi: 10.1111/2041-210X.12753

Lin, D., Lai, J., Muller-Landau, H. C., Mi, X., and Ma, K. (2012). Topographic Rogeau, M.-P., and Armstrong, G. W. (2017). Quantifying the effect of elevation and
variation in aboveground biomass in a subtropical evergreen broad-leaved forest in aspect on fire return intervals in the Canadian rocky mountains. For. Ecol. Manag. 384,
China. PLoS One 7:e48244. doi: 10.1371/journal.pone.0048244 248–261. doi: 10.1016/j.foreco.2016.10.035

Liu, Y., Li, P., Xu, Y., Shi, S., Ying, L., Zhang, W., et al. (2016). Quantitative Rowe, R. J., and Lidgard, S. (2009). Elevational gradients and species richness: do
classification and ordination for plant communities in dry valleys of Southwest China. methods change pattern perception? Glob. Ecol. Biogeogr. 18, 163–177. doi: 10.1111/j.
Biodivers. Sci. 24, 378–388. doi: 10.17520/biods.2015241 1466-8238.2008.00438.x

Luo, Y., Zhao, X., Li, Y., Liu, X., Wang, L., Wang, X., et al. (2021). Wind disturbance Sanders, N. J., and Rahbek, C. (2012). The patterns and causes of elevational
on litter production affects soil carbon accumulation in degraded sandy grasslands diversity gradients. Ecography 35, 1–3. doi: 10.1111/j.1600-0587.2011.07338.x
in semi-arid sandy grassland. Ecol. Eng. 171:106373. doi: 10.1016/j.ecoleng.2021. Scheiner, S. M. (2003). Six types of species-area curves. Glob. Ecol. Biogeogr. 12,
106373 441–447. doi: 10.1046/j.1466-822X.2003.00061.x
Lyman, R. L., and Ames, K. M. (2007). On the use of species-area curves to detect Shen, Z. (2016). Plant diversity in the dry valleys of Southwest China: spatial
the effects of sample size. J. Archaeol. Sci. 34, 1985–1990. doi: 10.1016/j.jas.2007.01.011 deviation and determinants for flora and plant communities. Biodivers. Sci. 24,
363–366. doi: 10.17520/biods.2016049
Martin, L. J., Blossey, B., and Ellis, E. (2012). Mapping where ecologists work: biases
in the global distribution of terrestrial ecological observations. Front. Ecol. Environ. Šímová, I., Violle, C., Svenning, J. C., Kattge, J., Engemann, K., Sandel, B., et al.
10:195–201. doi: 10.1890/110154 (2018). Spatial patterns and climate relationships of major plant traits in the New
World differ between woody and herbaceous species. J. Biogeogr. 45, 895–916. doi:
Moeslund, J. E., Arge, L., Bøcher, P. K., Dalgaard, T., and Svenning, J.-C. (2013).
10.1111/jbi.13171
Topography as a driver of local terrestrial vascular plant diversity patterns. Nordic J.
Bot. 31, 129–144. doi: 10.1111/j.1756-1051.2013.00082.x Sloan, S., Jenkins, C. N., Joppa, L. N., Gaveau, D. L. A., and Laurance, W. F. (2014).
Remaining natural vegetation in the global biodiversity hotspots. Biol. Conserv. 177,
Monge-González, M. L., Craven, D., Krömer, T., Castillo-Campos, G., Hernández-
12–24. doi: 10.1016/j.biocon.2014.05.027
Sánchez, A., Guzmán-Jacob, V., et al. (2019). Response of tree diversity and
community composition to forest use intensity along a tropical elevational gradient. Stein, A., Gerstner, K., Kreft, H., and Arita, H. (2014). Environmental heterogeneity
Appl. Veg. Sci. 23, 69–79. doi: 10.1111/avsc.12465 as a universal driver of species richness across taxa, biomes and spatial scales. Ecol.
Lett. 17, 866–880. doi: 10.1111/ele.12277
Moore, J., Gardiner, B., and Sellier, D. (2018). “Tree mechanics and wind loading,”
in Plant biomechanics, eds A. Geitmann and J. Gril (Berlin: Springer), 79–106. doi: Strassburg, B. B. N., Iribarrem, A., Beyer, H. L., Cordeiro, C. L., Crouzeilles, R.,
10.1007/978-3-319-79099-2_4 Jakovac, C. C., et al. (2020). Global priority areas for ecosystem restoration. Nature
Morandini, F., Santoni, P. A., Balbi, J. H., Ventura, J. M., and Mendes-Lopes, J. M. 586, 724–729. doi: 10.1038/s41586-020-2784-9
(2002). A two-dimensional model of fire spread across a fuel bed including wind Tang, Z., and Fang, J. (2004). A review on the elevational patterns of plant species
combined with slope conditions. Int. J. Wildland Fire 11, 53–63. doi: 10.1071/WF diversity. Biodivers. Sci. 20, 20–28. doi: 10.17520/biods.2004004
01043
Tang, Z., Fang, J., Chi, X., Feng, J., Liu, Y., Shen, Z., et al. (2012). Patterns of plant
Moritz, C., and Agudo, R. (2013). The future of species under climate change: beta-diversity along elevational and latitudinal gradients in mountain forests of China.
resilience or decline? Science 341, 504–508. doi: 10.1126/science.1237190 Ecography 35, 1083–1091. doi: 10.1111/j.1600-0587.2012.06882.x
Muller-Landau, H. C., Cushman, K. C., Arroyo, E. E., Martinez Cano, I., Anderson- Ullah, F., Gilani, H., Sanaei, A., Hussain, K., and Ali, A. (2021). Stand structure
Teixeira, K. J., and Backiel, B. (2021). Patterns and mechanisms of spatial variation determines aboveground biomass across temperate forest types and species mixture
in tropical forest productivity, woody residence time, and biomass. New Phytol. 229, along a local-scale elevational gradient. For. Ecol. Manag. 486:118984. doi: 10.1016/j.
3065–3087. doi: 10.1111/nph.17084 foreco.2021.118984
Munne-Bosch, S. (2018). Limits to tree growth and longevity. Trends Plant Sci. 23, Wang, G., Ran, F., Chang, R., Yang, Y., Luo, J., and Jianrong, F. (2014).
985–993. doi: 10.1016/j.tplants.2018.08.001 Variations in the live biomass and carbon pools of Abies georgei along an elevation
Myers, N., Mittermeier, R. A., Mittermeier, C. G., da Fonseca, G. A., and Kent, J. gradient on the Tibetan Plateau. China. For. Ecol. Manag. 329, 255–263. doi:
(2000). Biodiversity hotspots for conservation priorities. Nature 403, 853–858. doi: 10.1016/j.foreco.2014.06.023
10.1038/35002501 Wang, J., Ou, G., Tang, J., Zhenlong, L., Li, H., and Hui, X. (2012). Biomass
O’Brien, E. M., Field, R., and Whittaker, R. J. (2000). Climatic gradients estimation model of shrub community at Jatropha curcas growing area in Lincang of
in woody plant (tree and shrub) diversity: water-energy dynamics, residual Yunnan. J. West China For. Sci. 41:167.
variation, and topography. Oikos 89, 588–600. doi: 10.1034/j.1600-0706.2000.8 Wang, Y., Cufar, K., Eckstein, D., and Liang, E. (2012). Variation of maximum
90319.x tree height and annual shoot growth of Smith fir at various elevations in the Sygera
Oksanen, J., Simpson, G. L., Blanchet, F. G., Kindt, R., Legendre, P., Minchin, P. R., Mountains, southeastern Tibetan Plateau. PLoS One 7:e31725. doi: 10.1371/journal.
et al. (2022). vegan: Community Ecology Package. R package version 2.6-4. Available pone.0031725
online at: https://CRAN.R-project.org/package=vegan (accessed December 20, 2021).
Wang, Z., Liu, X., Wang, H., Zheng, K., Li, H., Wang, G., et al. (2021). Monitoring
Pepin, N. C., Pike, G., Schaefer, M., Boston, C. M., and Lovell, H. (2017). A vegetation greenness in response to climate variation along the elevation gradient
comparison of simultaneous temperature and humidity observations from the SW in the three-river source region of China. ISPRS Int. J. Geo Informat. 10:193. doi:
and NE slopes of Kilimanjaro: the role of slope aspect and differential land-cover 10.3390/ijgi10030193

Frontiers in Forests and Global Change 11 frontiersin.org

Cheng et al. 10.3389/ffgc.2023.1187724

Wang, Z., Tang, Z., and Fang, J. (2007). Altitudinal patterns of seed plant richness three parallel rivers region, northwestern Yunnan. Biodivers. Sci. 24, 440–452. doi:
in the Gaoligong Mountains, south-east Tibet, China. Divers. Distribut. 13, 845–854. 10.17520/biods.2015361
doi: 10.1111/j.1472-4642.2007.00335.x
Yang, Y., Shen, Z., Han, J., and Zhongyong, C. (2016b). Plant diversity along the
Xie, Z., Wang, Y., Tang, Z., and Xu, W. (2018). Handbook of biomass allometric eastern and western slopes of Baima Snow Mountain. China. Forests 7:89. doi: 10.3390/
models of shrubs in China. Beijing: Science press. f7040089
Xu, Y., Li, P., Liu, Y., Zhang, W., and Qin, S. (2016). Spatial patterns Zeleny, D. (2021). twinspanR: TWo-way INdicator SPecies ANalysis (and its
and determinants of species richness of alien and native plants in the modified version) in R. R package version 0.22. Available online at: https://github.com/
Nujiang River valley. Biodivers. Sci. 24, 389–398. doi: 10.17520/biods.201 zdealveindy/twinspanR (accessed January 10, 2022).
5243
Zhang, W., Zhao, L., Larjavaara, M., Morris, E. C., Sterck, F. J., and
Xu, Y., Shen, Z., Ying, L., Zang, R., and Jiang, Y. (2019). Effects of current climate, Wang, G. (2020). Height-diameter allometric relationships for seedlings
paleo-climate, and habitat heterogeneity in determining biogeographical patterns of and trees across China. Acta Oecol. 108:103621. doi: 10.1016/j.actao.2020.
evergreen broad-leaved woody plants in China. J. Geogr. Sci. 29, 1142–1158. doi: 103621
10.1007/s11442-019-1650-x
Zhang, Y., and Liu, Y. (2010). Datasets of China ecosystem research station. Forest
Yang, Y., Han, J., Liu, Y., Zhongyong, C., Shi, S., Sina, C., et al. (2016a). A comparison Ecosystems Vol. Ailaoshan Station in Yunnan Province: 2003˜2007. Beijing: China
of the altitudinal patterns in plant species diversity within the dry valleys of the agriculture press.

Frontiers in Forests and Global Change 12 frontiersin.org