Zhangetal. Plant Methods (2021) 17:5
New eld wind manipulation methodology
reveals adaptive responses ofsteppe plants
toincreased andreduced wind speed
Shudong Zhang1,2,3, Guofang Liu1, Qingguo Cui1, Zhenying Huang1*, Xuehua Ye1*
and Johannes H. C. Cornelissen3
Background: Wind strongly impacts plant growth, leaf traits, biomass allocation, and stem mechanical properties.
However, whether there are common whole-plant wind responses among diﬀerent plant species is still unclear.
We tested this null hypothesis by exposing four eudicot steppe species to three diﬀerent wind treatments in a ﬁeld
experiment: reduced wind velocity using windbreaks, ambient wind velocity, and enhanced wind velocity through a
novel methodology using wind-funneling baﬄes.
Results: Across the four species, wind generally decreased plant height, projected crown area, and stepwise bifurca-
tion ratio, and increased root length and stem base diameter. In contrast, the response patterns of shoot traits, espe-
cially mechanical properties, to wind velocity were idiosyncratic among species. There was no signiﬁcant diﬀerence in
total biomass among diﬀerent treatments; this might be because the negative eﬀects on heat dissipation and photo-
synthesis of low wind speed during hot periods, could counteract positive eﬀects during favorable cooler periods.
Conclusions: There are common wind response patterns in plant-size-related traits across diﬀerent steppe species,
while the response patterns in shoot traits vary among species. This indicates the species-speciﬁc ways by which
plants balance growth and mechanical support facing wind stress. Our new ﬁeld wind manipulation methodology
was eﬀective in altering wind speed with the intended magnitude. Especially, our ﬁeld wind-funneling baﬄe system
showed a great potential for use in future ﬁeld wind velocity enhancement. Further experiments are needed to reveal
how negative and positive eﬀects play out on whole-plant performance in response to diﬀerent wind regimes, which
is important as ongoing global climatic changes involve big changes in wind regimes.
Keywords: Biomass and allocations, Ecological method, Mechanical properties, Mu Us Sandland, Plant size, Wind
© The Author(s) 2021. This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing,
adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and
the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material
in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material
is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the
permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creat iveco
mmons .org/licen ses/by/4.0/. The Creative Commons Public Domain Dedication waiver (http://creat iveco mmons .org/publi cdoma in/
zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data.
Almost all terrestrial plants face wind stress, especially
in mountain and coastal ecosystems, plateaus and other
inland ecosystems where wind has suﬃcient open space
or natural funneling to gain in force [1–4]. Wind is a
major source of mechanical loading and leaf tempera-
ture regime on plants, by which it has a major impact
on plant growth, morphology, physiology, dispersal and
ecology [3, 5–7]. Almost all plants have to balance ﬁve
major requirements throughout their lifetime: photo-
synthesis, water transport, growth, reproduction, and
mechanical support . Wind may aﬀect all these ﬁve
requirements. First, in the natural environment, wind
is the main source of mechanical perturbation. e
mechanical signals that plants perceive imposed by wind
*Correspondence: email@example.com; firstname.lastname@example.org
1 State Key Laboratory of Vegetation and Environmental Change, Institute
of Botany, Chinese Academy of Sciences, Beijing 100093, People’s
Republic of China
Full list of author information is available at the end of the article
Page 2 of 16
Zhangetal. Plant Methods (2021) 17:5
can induce thigmomorphogenesis, which may alter plant
growth patterns and cause lower stature, thicker stem,
and smaller shoots [5, 9–13]. Additionally, mechanical
stress can reduce plant growth [14, 15], increase the root/
shoot ratio [13, 16], and change leaf properties including
a reduction in the number of leaves, individual leaf area
or dry mass and petiole length, and increase leaf thick-
ness and ﬂexibility [5, 7, 17–19]. Second, besides creating
stressful drag force, wind also inﬂuences photosynthesis
and transpiration of plants in diﬀerent ways depending
on plant traits, wind intensities and ambient tempera-
tures. Photosynthetic rates will decrease at (very) low
wind speed due to the increase in the leaf boundary layer
and the consequent reduction in the diﬀusive resistance
for carbon dioxide; at high ambient air temperatures a
thick boundary layer may also cause excessive leaf tem-
peratures that inhibit photosynthesis. At (very) high wind
speed and low ambient air temperature, photosynthesis
may be reduced due to below-optimal leaf temperatures
and stomatal conductance, and to leaves rolling up [5,
16, 18]. ird, in some speciﬁc areas (e.g. coastal dunes,
inland arid and semi-arid dunes), sand movement caused
by strong wind is a common environmental agent aﬀect-
ing plants [20–24]. In these habitats, plants may face
stress caused by soil losses or sand burial under wind
erosion. In extreme cases, wind denudation may cause
soil water loss, reduction in nutrient content or greater
exposure of the upper roots of plants to high temperature
[24–28]. In contrast, moderate sand burial may improve
soil water and nutrient conditions and moderate the tem-
perature ﬂuctuation around plant roots, which may ben-
eﬁt plant growth .
Plants have developed various adaptations to cope with
strong wind exposure, e.g. through trait variation in their
branches  and leaves , biomass allocation ,
root structure , and/or stem mechanical properties
[18, 34]. However, these traits of diﬀerent plant organs
are subject to trade-oﬀs (e.g. carbon distribution) or
coordination (e.g. allometric relationships). erefore, it
is hard from most previous studies focusing on particu-
lar plant organs and particular structural, morphological,
and/or physiological traits, to infer whole-plant strate-
gies in terms of adaptive response to wind. Few previous
studies have compared wind responses among several
species, even though plant species are known to vary in
their capacity to resist wind, and to recover from or oﬀ-
set the eﬀects of wind damage . In addition, research
ﬁelds study diﬀerent aspects of plant response to wind.
Forestry researchers tend to observe storm-damaged for-
est and often focus on the growth versus shade-tolerance
trade-oﬀ of tree species [17, 36]. Agricultural research
has focused on lodging resistance and yield-related traits
[8, 15]. us, the scientiﬁc community is dispersed in the
speciﬁc variables they measure in response to wind .
Research methods applied for wind manipulation also
vary greatly, from ﬁeld observation to ﬁeld experiment to
glasshouse controlled-environment experiment.
Two main experimental approaches have so far been
used to impose wind disturbance eﬀects on plants. One
is artiﬁcially creating mechanical perturbation by shak-
ing and brushing plants [5, 12, 19, 37–41], which has
focused on the mechanical perturbation caused by wind
and has been conducted mostly in controlled environ-
ments in glasshouses in the absence of wind. Another has
been the use of wind tunnels or electric fans to enhance
wind force [5, 16, 18, 32, 42, 43]. ese methods depend
on power supply and indoor facilities, which makes them
hard to be widely applied under ﬁeld conditions. Because
of these limitations, we still understand poorly whether
there are general syndromes of whole-plant response to
wind exposure across organs, i.e. wind-coping strategies,
or whether responses to wind are idiosyncratic among
species without consistent responses among organs and
their various traits.
Here we aimed to overcome the above limitations in a
unique ﬁeld experiment in order to test the null hypoth-
esis that there are common whole-plant wind responses
among diﬀerent plant species. We assigned four contrast-
ing vascular species to three treatments (reducing wind
velocity, ambient wind velocity, enhancing wind velocity)
in the steppe area of north China. To increase wind veloc-
ity we employed a novel methodology using connectivity
modiﬁers (baﬄes) to converge wind. Connectivity modi-
ﬁers, a patch-scale manipulation, can eﬀectively change
the size of connected pathways for wind or water under
ﬁeld conditions without directly aﬀecting other abiotic
and biotic factors. is methodology was previously used
in some studies to collect foliar litter and seeds [44, 45],
but has, to our knowledge, not been used to study wind
eﬀects on living plants. is methodology allows the
wind regimes in this study to reﬂect the real environmen-
tal conditions that plants may encounter outdoors, where
wind stress or disturbance may vary with weather condi-
tions and the local properties of the terrain.
e experiment was carried out at the Ordos Sandland
Ecological Station of the Chinese Academy of Science
(OSES, 39° 29′ N, 110° 11′ E, 1296m a.s.l.), located in the
northeastern Mu Us Sandland in Inner Mongolia, China.
Mean annual precipitation is 350mm, with inter-annual
ﬂuctuations from 161 to 664 mm. Mean annual tem-
peratures range from 5.0 to 8.5°C, and mean potential
evaporation is 2300 mm (OSES weather station, 2005–
2015). On average 140 days per year have maximum
Page 3 of 16
Zhangetal. Plant Methods (2021) 17:5
wind speed > 6bft (> 10.8m/s), and the wind predomi-
nantly comes from the north-west (Additional ﬁle1: Fig.
S1). During the study year (2017), the total precipitation
was 312.6 mm, most of which (280.4 mm) fell during
the experimental period in the growing season (May to
October; Additional ﬁle1: Fig. S2A). Monthly mean wind
velocities ranged from 1.6 to 2.31 m/s and maximum
wind velocity (14.4m/s) were reached in May (Additional
ﬁle1: Fig. S2B, C; all the wind data were collected from
the OSES weather station at the standard 10m height).
e distribution of wind direction of maximum wind
speed was consistent with these data (Additional ﬁle 1:
Fig. S2D). e landscape in this area is characterized by
mobile, semi-ﬁxed and ﬁxed sand dunes. As water avail-
ability is low, the area is dominated by steppe or desert
vegetation with low height and sparse cover.
Two predominant shrubs, i.e. Artemisia ordosica Krasch
(Asteraceae), Caragana intermedia Kuang (Fabaceae)
and two locally common herbs, i.e. Agriophyllum squar-
rosum (L.) Moq. (Chenopodiaceae) and Salsola ruthenica
Iljin (Chenopodiaceae), were selected for the experiment.
A. ordosica is a dominant shrub in (semi-)ﬁxed sand
dunes, approximately 0.5–1.0m tall with plumose, lin-
early lobate leaves. Its lateral roots are mainly distributed
in the upper 30cm of the sand soil proﬁle, while its pri-
mary roots may reach 1–3m deep . C. intermedia is
a deciduous pinnate-leaved shrub approximately 1.5–2m
tall, widely used to mitigate desertiﬁcation in north
China . A. squarrosum and S. ruthenica are annual
herbs, up to 1 m tall, with erect stems and basal side-
shoots. A. squarrosum is an important pioneer on mov-
ing and semi-ﬁxed sand dunes, widely distributed in arid
and semiarid regions of central Asia [47, 48] and widely
naturalized in many other parts of the world.
Seeds of the four species were collected near OSES in
2015. On 16 September 2016, A. ordosica and C. inter-
media seeds were germinated on wet ﬁlter paper; on 23
September 2016 the seedlings were transplanted into
fabric bags (10cm diameter, 15cm height) and grown in
the greenhouse of OSES through the winter. On 18 May
2017, 18 seedlings of each species with similar height
(28.5 ± 4.07cm for C. intermedia and 10.74 ± 0.89cm for
A. ordosica) and stem base diameter (1.91 ± 0.42mm for
C. intermedia and 4.00 ± 0.86 cm for A. ordosica,) were
chosen and randomly transplanted into a 50 L (0.4m
diameter, 0.4m high) pot ﬁlled with sand without drain-
age holes in the bottom. On 15 May 2017, A. squarro-
sum and S. ruthenica seeds were germinated on wet ﬁlter
paper. After 16days, 18 seedlings per species were trans-
planted into the pots (see above). All pots were buried in
the quadrats of the three diﬀerent treatments on 19 May,
2017. During the experimental period, all plants received
2L of water once a week and 1L of 1g/L nutrient solu-
tion (Peters Professional 20-20-20 General Purpose, the
Scotts Company, Ohio, USA; N:P:K = 1:0.83:0.44, plus
microelements) once a month.
We imposed three diﬀerent wind treatments on the four
plant species: decreased wind velocity (D), ambient wind
velocity (CK), and increased wind velocity (I); see Fig.1
for methodological details. In brief, treatment D was
implemented through a wind shield (Fig.1a). e design
of using transparent plastic sheets was intended to mini-
mize the inﬂuence of the wind shield on the light regime.
Treatment I was implemented through artiﬁcial air baf-
ﬂes that funneled the ambient wind towards the target
plant (Fig.1c). To the best of our knowledge, this repre-
sents a new method for wind enhancement with minimal
experimental artefact (see under “Results”). e wind
velocity increasing method was based on an ideal model
(Fig.2a) coming from the ﬂuid continuity equation:
where ρ was a constant representing the density of air
ﬂow. A1 was the interface area of the wind ﬂow coming
in and A2 was the interface area of the wind ﬂow going
out. V1 was the initial wind velocity when air ﬂow came
in and V2 was the wind velocity when air ﬂow went out.
In the ideal model, air should ﬂow through a closed tun-
nel. e basic continuity equation, states that the amount
of air ﬂowing in on one side must equal the air ﬂowing
out on the other side. In this case, we could change the
velocity of air ﬂow by changing the ratio of A1 and A2. In
our design, we used air baﬄes to create an air tunnel that
had a larger area where the wind ﬂow entered than the
area where the wind ﬂow exited. According to the conti-
where Vout represents the wind velocity at the exit point
of the wind funnel, and Vin is the initial wind velocity
before treatment. s is the length of the short side of the
wind funnel, l the length of the wind baﬄe. However, the
above formula assumes a closed funnel at the top. As our
wind funnel was open at the top, the actual wind veloc-
ity at the exit point of the wind funnel should be smaller
than Vout. Increasing the baﬄe height would make the
funnel closer to the sealed status, i.e. it could make the
ﬁnal wind velocity closer to Vout.
Both treatments CK (Fig.1b) and I had a large-mesh
fence around the plant to protect it from herbivory. In
total, there were 72 plants: 4 species × 3 treatments × 6
Page 4 of 16
Zhangetal. Plant Methods (2021) 17:5
Fig. 1 Experimental design and images of wind treatments. a The Decreased wind velocity treatment was implemented through a wind shield,
which was 1 m × 1 m × 0.75 m (0.75 m high) and built with four steel tubes and transparent plastic sheets. The plastic chambers did not have
a roof, so that ambient air could mix with that in the cubicles freely. b There was no additional treatment for the ambient treatment, apart from
a large-mesh, 0.6 m high fence to avoid herbivory by wild rabbits (In a pilot test, the fence had a negligible inﬂuence on wind velocity). c The
Increased wind velocity treatment was implemented through experimental wind baﬄes. We placed 1.5 m long and 1 m high iron sheets at each
side directed towards the North-East, North-West, South-East and South-West, respectively from each pot, to converge and increase wind velocity.
These baﬄes were at 0.7 m distance from the plant pot to avoid a shading eﬀect
Page 5 of 16
Zhangetal. Plant Methods (2021) 17:5
replicates (Fig. 1). e experimental site was very ﬂat,
and treatments were far away from each other and taller
surrounding vegetation. For avoiding obstruction of the
air ﬂow to other treatments, the D section was laid out
at the south-west, the CK section was set out at the east
and I section at the west of the experimental site. Dur-
ing the experiment period, sand burial happened due to
the increased wind velocity in treatment I despite the
protection of the plastic pot. Extra sand was periodically
removed from the pots using a brush when burials hap-
pened to minimize the inﬂuences of the sand burial on
Environment parameters measurements
Wind speed, air temperature and relative humidity
were measured at 0.8m above the ground (i.e. close to
maximum plant height), using AZ9671 Anemometer
Loggers (Shenzhen frank electronics co. LTD., China).
e loggers recorded data with 90-s intervals. Because
the AZ9671 Anemometer Loggers can only quantify air
movement in one direction, we chose westerly as our
main direction (due to prevailing strong wind coming
from this direction; see Additional ﬁle1: Fig. S1) and
all the loggers were positioned toward the west dur-
ing the whole experiment period. Soil volumetric water
content and temperature were measured 10cm below
the soil surface in the middle of the buckets by Em50
series data loggers (Decagon Devices, Inc., USA). e
loggers collected data at 2-min intervals. In each treat-
ment section, three AZ9671 and Em50 loggers were
installed in random positions. We collected data from
1 June to 7 September 2017. Due to the limited of the
power supplies and data storage of AZ9671 loggers, we
had to retrieve the loggers from the ﬁeld for replacing
batteries and download data regularly. Additionally, the
anemometer had a high rate of malfunction in the rain.
us we also avoid the continuously rainy days. In this
case, the data set of wind speed, air temperature and
relative humidity had some breaks in time during the
whole experimental period, but we tried our best to
make sure that the measurements covered most of the
available time. e speciﬁcations of all the instruments
used in this study can be found in Additional ﬁle 1:
Fig. 2 Diagrammatic sketch of wind velocity increasing methodology and eﬀects of wind treatments. a Wind velocity increasing method. Ain is
the interface area of the wind ﬂow coming in and Aout is the interface area of the wind ﬂow going out. Vin is the initial in-coming wind velocity and
Vout is the wind velocity of air ﬂowing out. s is the length of the short side of the wind funnel, and l is the length of the wind baﬄe. b Diurnal mean
wind velocity pattern (mean ± SE) over 24 h under wind treatments throughout the experimental period. c Diurnal maximum wind velocity pattern
(mean ± SE) over 24 h under wind treatments throughout the experimental period. d Eﬀects of wind treatments. The black dashed line with the
) represents the ideal wind velocity that could be reached by acceleration. The black dashed line with the equation (y = x)
represents the wind velocity in the ambient treatment (CK). Red dots and line represent the relationship between the temporally matched hourly
maximum wind velocity in I and CK. Blue dots and line represent the relationship between the temporally matched hourly maximum wind velocity
in D and CK. Regression equations and R2 are given. D means decreased wind velocity treatment, CK means ambient wind velocity treatment, and I
means increased wind velocity treatment
Page 6 of 16
Zhangetal. Plant Methods (2021) 17:5
Plant trait measurements
All plants were harvested within nine days from 7 Sep-
Plant height, stem base diameter (SBD), crown traits,
leaf traits and branching pattern were measured before
harvest. Plant crown length (L) and width (W) were
measured, and projected crown area (PCA) was calcu-
For each plant, 15 fresh leaves were scanned and leaf
length (LL) and leaf area (LA) were measured. e leaves
were dried at 80°C for 48h, then weighed. Speciﬁc leaf
area (SLA) was calculated as fresh leaf area per unit leaf
e branching pattern was measured following Strahler
method . Numbers of ﬁrst to third degree branches
were counted, then overall bifurcation ratio (OBR) and
stepwise bifurcation ratio (SBR) were calculated as
where numbers of branches are represented by Nt for all
segments, Ns for the highest-order branches, N1 for the
ﬁrst-degree branches, while Ni the branch number of i
After harvest, each plant was divided into root and
shoot. We measured root length (RL) of the longest root.
For shoots, 15cm long primary stem sections were sepa-
rated from the bottom of the plant shoot for determina-
tion of mechanical properties.
Young’s modulus (E) of the main stems was measured
with an electromechanical device (Type 5540; Instron,
Norwood, MA, USA), applying the three-point bending
technique . E is a measure of stiﬀness of a material
. Vertically applied forces (F; N) and resulting deﬂec-
; m) were recorded. E was calculated as
where L is the length between the supports (m) and I the
second moment of area (m4). I was calculated as
where r = stem radius in m. is index is a measure of the
geometric contribution to rigidity . Flexural stiﬀness of
stems is a measure of the rigidity of a material, calculated
as the product of E and I (EI, N/m2; ).
Mechanical measurements of each plant were com-
pleted within 15min after cutting. After these tests, all
parts of the plants were dried at 80°C for at least 72h.
en, root and shoot biomass were weighed and root/
shoot ratio was calculated. After weighing, the main
stems used for mechanical measurements were ground
by ball mill; stem lignin content (SLC) was determined
as acid insoluble (Klason) lignin. Stem cellulose con-
tent (SCC) was determined colorimetrically with the
anthrone reagent .
Because the data set of wind speed, air temperature and
relative humidity had breaks in time, only days with the
completed data were chosen for the analysis. In total
there were 31 completed days that covered the whole
experimental period. en, the average hour-by-hour val-
ues of each parameter in each treatment throughout this
period were calculated. Based on the hourly maximum
wind velocities in consecutive days, the hour-by-hour
mean hourly maximum wind velocity was calculated. To
evaluate the results of wind velocity changes, the hourly
maximum wind velocity in the treatment CK and I, and
Treatment CK and D were temporally matched, then
the linear regression lines were generated by the for-
) in R software (v3.3.0, R core team, 2015).
Because of the limitation of the AZ9671 Anemometer
Logger, wind velocity data were collected from only one
direction (west). e wind velocity data collected from
the directions perpendicular to the logger were expected
to be underestimates. For checking the accuracy of our
wind velocity data, we used the wind data collected from
OSES weather station as a reference. e OSES weather
station recorded the hourly average wind direction dur-
ing the whole experimental period. We ﬁrst extracted the
time intervals during which the wind came from the west
(from 225° to 315°) during the experimental period. en,
we paired these time points with the data collected from
the AZ9671 Anemometer Logger, extracting the time
intervals each day when the wind was mainly coming
from the west during our experimental period. By com-
paring the average hour-by-hour wind velocities between
the westerly wind data set with those in the whole data
set, we could evaluate the accuracy of our whole wind
velocity data set collected from AZ9671 Anemometer
Logger, and especially the robustness of the diﬀerences
found among wind treatments. Two-way ANOVAs
were used to analyze the (interactive) eﬀects of diﬀerent
species and wind treatments on each plant trait. One-
way ANOVA was used to test diﬀerences in plant traits
among the three wind treatments for each species sepa-
rately, followed by Tukey’s HSD tests for multiple com-
parisons. Data were log10(x + 1)-transformed if necessary
to improve the equality of variance distributions among
treatments. Statistical analyses were performed using
SPSS18 (SPSS Inc., USA2009). A redundancy analysis
Page 7 of 16
Zhangetal. Plant Methods (2021) 17:5
was carried out in R software (v3.3.0, R core team, 2015)
using the package Vegan to explore the covariance struc-
ture of the diﬀerent trait responses to the wind treat-
ments across the species comprehensively. e plant
traits were used as the respective response variables.
en, the response matrix was standardized by a scale
function. e explanatory matrix was deﬁned by the
plant species and wind treatments. For testing the sig-
niﬁcance of the variation in response matrix explained
by explanatory variables, a Monte Carlo permutation
test was used (Anova function with permutation options,
permutations = 999). e adjusted R2 was computed,
after the permutation test.
Environmental variables indierent wind treatments
e wind shields (treatment D) reduced daily ambient
wind velocity by 71% and maximum wind velocity by 67%
on average (Fig.2b, c). It increased ambient air tempera-
ture on average by 2.1°C, and reduced ambient relative
humidity by 3%, ambient soil temperature by 0.4°C and
ambient soil volumetric water content by 14% on average
(Additional ﬁle1: Fig. S3). e experimental wind baf-
ﬂes (treatment I) increased daily ambient wind velocity
by 56% and maximum wind velocity by 114%, decreased
ambient air temperature by 0.6°C and ambient soil tem-
perature by 1.7°C on average, while they increased ambi-
ent relative humidity and soil volumetric water content
each by 2% (Fig.2b, c, Additional ﬁle1: Fig. S3). e aver-
age hour-by-hour wind velocity between the westerly
wind data set and the whole data set (Additional ﬁle1:
Fig. S4) showed similar patterns when comparing the
mean wind velocities among diﬀerent wind treatments.
is means that, although the limitation of the AZ9671
Anemometer Logger caused some inaccuracy in absolute
wind velocity values, the whole data set still reﬂected the
real wind patterns in the diﬀerent treatments during our
e temporally matched relationship between the
hourly maximum wind velocity in the CK and I showed
that our wind funnel design had potential to accelerate
wind ﬂow reaching the velocity calculated by the ideal
model (Fig. 2d). According to the regression equation
y = 1.51x, the actual accelerated hourly maximum wind
velocity of our design can be predicted by the hourly
maximum wind velocity in the ambient condition. e
temporally matched relationship between the hourly
maximum wind velocity in the CK and D also reﬂected
the eﬀect of the wind breaks on decreasing wind veloc-
ity. Another regression equation y = 0.307x was gener-
ated which could be used to estimate the maximum wind
velocity in the treatment D.
Commonalities inresponses amongthespecies towind
All plant traits showed signiﬁcant diﬀerences among
species, while 10 out of 19 traits showed signiﬁcant
wind eﬀects and nine out of 19 traits signiﬁcant inter-
actions of species and wind treatment (Table1). Across
species, wind velocity signiﬁcantly aﬀected plant
height, PCA, SBD, RL, SBR(1:2), LL, total and shoot
biomass, root/shoot ratio, and E. ere were signiﬁ-
cant species by treatment eﬀects on PCA, SBD, OBR,
SBR(1:2), root/shoot ratio, SCC, I and E (Table1).
Among species, common responses mainly showed as
decreases in size-related traits and increases in wind-
resistance abilities. Plants tended to be shorter, have
smaller PCA, SBR(1:2), and lower biomass in response to
increasing wind velocity (Table2, Figs.3, 4). Height of
C. intermedia and S. ruthenica decreased with increas-
ing wind velocity (Fig.3a). PCA of A. ordosica, C. inter-
media, and S. ruthenica decreased with increasing wind
velocity (Fig.3b). OBR and SBR(1:2) of A. ordosica and
C. intermedia were markedly inﬂuenced by the wind
treatments (Table2), being lower in CK and I than in
D (Fig.3d, Additional ﬁle1: Fig. S5E). Increased wind
velocity decreased total biomass of C. intermedia,
and shoot biomass of C. intermedia and A. squarro-
sum (Table2, Fig.4). RL, SBD and I tended to increase
with increased wind velocity (Table2, Figs. 3, 5, and
Additional ﬁle1: Fig. S5). RL of A. squarrosum and S.
ruthenica increased steadily from D via CK to I while A.
ordosica only showed an increase from CK to I and C.
intermedia showed no response at all (Table2, Fig.3c).
I of A. ordosica and S. ruthenica was higher in treat-
ments CK and I than in D (Table2, Fig.5c). Stem base
diameter of A. ordosica and A. squarrosum increased in
CK and I compared to treatment D (Table2, Additional
ﬁle1: Fig. S5F).
Idiosyncratic responses amongspecies towind treatments
In contrast to the common responses to wind treatments
for size-related traits, the plants showed idiosyncratic
responses among species for some traits (Additional
ﬁle 1: Fig. S6). Responses of mechanical properties
showed large diversity. Wind decreased E in A. ordosica
and S. ruthenica. E of A. ordosica decreased markedly in
CK and I as compared to D, which resulted in an increase
of EI. In S. ruthenica, E decreased in CK. However, the
wind treatments had no signiﬁcant inﬂuence on E of A.
squarrosum and increased E of C. intermedia. Except for
A. ordosica, EI of the other three species was not signiﬁ-
cantly inﬂuenced by the wind treatments. Moreover, SCC
of A. ordosica was higher in CK than in D, while in A.
squarrosum it was lower in CK and I than in D (Fig.5).
Page 8 of 16
Zhangetal. Plant Methods (2021) 17:5
Leaf traits also responded in a species-speciﬁc manner
to the wind treatments. Leaf length of C. intermedia and
S. ruthenica and leaf area of C. intermedia and A. squar-
rosum were inﬂuenced by wind treatments (Table2). e
leaves of S. ruthenica were shorter in CK and I as com-
pared to D (Additional ﬁle1: Fig. S5A). Leaf area of C.
intermedia was markedly reduced in CK as compared to
D, and leaf area of A. squarrosum was lower in CK and I
than in D (Additional ﬁle1: Fig. S5C).
To our knowledge, this is the ﬁrst experimental study to
simultaneously compare eﬀects of both decreased and
increased wind speed on plant performance with ambient
wind speed under ﬁeld conditions. Our wind funneling
treatment has added a new experimental method to
increase wind speed under ﬁeld conditions without some
of the experimental artefacts on other environmental fac-
tors associated with treatments using wind tunnels and
fans. In particular, our ﬁeld wind-funneling baﬄes con-
tinuously increased wind speed proportionally (increased
daily ambient wind velocity by 56% and maximum wind
velocity by 114%, Fig.2) to ambient wind speed. ese
treatments were eﬀective in altering wind speed with the
intended magnitude of reduction or increase through-
out the day across the growing season (Fig.2). is new
experimental design has been able to reveal how four
morphologically diﬀerent steppe plant species, includ-
ing two shrubs and two forbs, responded in apparently
adaptive ways to both reduced and increased wind speed.
Overall, these responses showed both common patterns
among these species, especially for traits related to plant
and organ size, and idiosyncratic patterns, which were
seen mostly for traits related to shoot and leaf properties.
Correspondingly, we will ﬁrst discuss the commonalities
in trait response among the species, followed by a focus
on the traits that showed idiosyncratic responses among
species. We will also discuss the possible confound-
ing inﬂuences of the treatments on plant performance
via microclimate eﬀects on leaf boundary layer and gas
exchange, which we did not measure. Finally we will dis-
cuss the eﬀects of our wind treatments and their future
Common response patterns ofplant size related traits
In general, plants can adapt to wind stress at the whole-
plant level (i.e. besides possible eﬀects via photosynthesis
Table 1 Eects ofspecies andwind velocity treatments, andtheir interactions onplant traits
Italic type indicates signicant dierences at p < 0.05
Plant traits Species (S) Wind treatments (T) S × T
F P F P F P
Height 43.16 < 0.001 20.06 < 0.001 1.31 0.268
Projected crown area 146.12 < 0.001 32.33 < 0.001 3.09 0.012
Stem base diameter 204.97 < 0.001 10.51 < 0.001 3.00 0.013
Root length 71.48 < 0.001 10.35 < 0.001 1.59 0.167
Overall bifurcation ratio 10.52 < 0.001 2.26 0.113 3.35 0.007
Stepwise bifurcation ratio(1:2) 13.18 < 0.001 3.29 0.044 3.47 0.005
Leaf length 359.76 < 0.001 4.81 0.012 1.03 0.413
Leaf area 51.88 < 0.001 0.73 0.485 0.29 0.939
Speciﬁc leaf area 248.26 < 0.001 0.08 0.921 0.30 0.937
Biomass and allocation
Total biomass 171.96 < 0.001 9.60 < 0.001 0.35 0.908
Shoot biomass 246.47 < 0.001 10.67 < 0.001 0.17 0.846
Root biomass 19.01 < 0.001 2.63 0.080 0.77 0.599
Root/shoot ratio 61.17 < 0.001 10.82 < 0.001 7.40 < 0.001
Stem lignin content 11.71 < 0.001 0.38 0.687 0.72 0.637
Stem cellulose content 5.41 0.002 1.67 0.196 7.90 < 0.001
Lignin/cellulose ratio 11.16 < 0.001 0.7 0.500 3.11 0.010
Second moment of area (I) 16.57 < 0.001 1.51 0.231 3.79 0.003
Young’s modulus (E) 10.93 < 0.001 10.63 < 0.001 4.52 0.001
Flexural stiﬀness (EI) 14.68 < 0.001 3.09 0.054 0.69 0.660
Page 9 of 16
Zhangetal. Plant Methods (2021) 17:5
and transpiration) either by reducing the mechanical
stress through lower height and smaller crown; and/or
by increasing the resistance abilities via increasing stem
base diameter, root length, and/or changing mechani-
cal properties. Our results showed that across species,
wind velocity signiﬁcantly aﬀected plant shoot biomass
and morphology traits at the whole-plant level, particu-
larly plant height, projected crown area, stepwise bifur-
cation ratio, root length and stem diameter. Increased
wind velocity had negative eﬀects on plant height, PCA
and total biomass across all species. In the two shrub spe-
cies (A. ordosica and C. intermedia), SBR(1:2) also showed
a negative trend with the increase in wind velocity. In
contrast, there were positive trends with root length in all
species. SBR(1:2) indicates directly branching conditions
in the current year , and decreased of SBR(1:2) means
that plants reduce the numbers and densities of branches.
Together with a decrease of PCA, lower SBR(1:2) will
reduce the wind drag to whole plants. ese results are
consistent with most previous studies on single plant
species, where plant height usually decreased and stem
base diameter increased at high wind velocity [12, 14,
15], plant biomass and PCA decreased [4, 10] and root/
shoot ratio increased [9, 16]. In summary, across species,
plants will reduce the mechanical stress caused by wind
through smaller stature and increase the physical resist-
ance against wind through deeper and coarser roots.
Treatment D had stronger eﬀects on size-related traits
than treatment I, such as on PCA in A. ordosica and C.
intermedia, height of C. intermedia and S. ruthenica, and
shoot biomass of C. intermedia and A. squarrosum. How
can we explain this? Firstly, on an evolutionary time scale,
windy conditions have probably been more common than
still conditions [53–55]. e monthly mean wind speeds
during our experimental period were below 2.5m/s, i.e.
were lower than those in most other parts of the Inner
Mongolian Plateau region. Regional ecotypes may thus
be adapted to higher wind exposure. Secondly, there was
a 71% reduction and a 56% increase of daily wind velocity
and a 67% reduction and a 114% increase of daily maxi-
mum wind velocity compared to the control treatment.
is diﬀerence in treatment eﬀect may have led to diﬀer-
ent eﬀect sizes of response. By adjusting the height and
length of the artiﬁcial air baﬄes in treatment I, we should
be able to adjust the wind speed increase. Longer baﬄes
will be needed to obtain stronger wind forces to cover the
Table 2 Eects ofwind velocity treatments onplant traits ofA. ordosica, C. intermedia, A. squarrosum, andS. ruthenica
based onone-way ANOVAs
Italic type indicates signicant eects at p < 0.05
Plant traits A. ordosica C. intermedia A. squarrosum S. ruthenica
F P F P F P F P
Height 1.82 0.201 15.55 < 0.001 2.51 0.131 6.40 0.010
Projected crown area 10.97 < 0.001 10.57 0.003 1.64 0.238 25.24 < 0.001
Stem base diameter 5.50 0.016 0.59 0.567 4.11 0.038 2.13 0.155
Root length 8.57 < 0.001 0.50 0.617 5.48 0.017 3.00 0.085
Overall bifurcation ratio 6.03 0.012 4.71 0.026 1.29 0.304 16.36 < 0.001
Stepwise bifurcation ratio(1:2) 6.51 < 0.001 4.23 0.035 0.91 0.423 3.38 0.062
Leaf length 2.97 0.082 5.51 0.016 0.30 0.747 4.40 0.031
Leaf area 0.06 0.942 4.52 0.029 3.79 0.047 0.04 0.966
Speciﬁc leaf area 1.15 0.342 0.51 0.610 0.01 0.992 0.19 0.826
Biomass and allocation
Total biomass 3.04 0.078 6.94 0.007 0.93 0.415 1.85 0.192
Shoot biomass 1.22 0.323 9.22 0.002 4.45 0.030 1.99 0.171
Root biomass 1.39 0.279 2.53 0.113 0.85 0.448 0.60 0.561
Root/shoot ratio 2.50 0.121 4.52 0.049 5.06 0.021 0.10 0.907
Stem lignin content 0.59 0.566 0.01 0.993 0.14 0.870 1.51 0.252
Stem cellulose content 13.72 < 0.001 0.12 0.885 25.94 < 0.001 2.68 0.101
Lignin/cellulose ratio 4.74 0.025 0.02 0.984 4.52 0.029 0.68 0.522
Second moment of area (I) 8.80 < 0.001 2.54 0.124 1.56 0.243 6.43 0.014
Young’s modulus (E) 11.83 < 0.001 5.57 0.021 0.70 0.513 6.93 0.011
Flexural stiﬀness (EI) 3.70 0.049 0.27 0.770 1.10 0.359 1.86 0.201
Page 10 of 16
Zhangetal. Plant Methods (2021) 17:5
full range of wind velocities diﬀerent plants may experi-
ence not only in steppe but also in other ecosystems.
Idiosyncratic shoot trait response patterns amongspecies
Mechanical perturbation caused by wind in nature has
long been examined [12, 16], and is known to change the
ﬂexibility and rigidity of plant stems or petioles [9, 19,
56]. Plants may either develop ﬂexible stems for reduc-
ing the stress imposed by wind drag force; or stiﬀ stems
to resist wind [19, 57, 58]; and there may be trade-oﬀs
between these two aspects. erefore, the response pat-
terns of shoot traits, especially mechanical properties,
to wind velocity is expected to be idiosyncratic among
In this experiment, wind signiﬁcantly increased the
second moment of area (I) of A. ordosica and decreased
Young’s modulus (E) of stem of A. ordosica and S. ruthen-
ica, resulting in a signiﬁcant increase of Flexural stiﬀness
(EI); while wind did not signiﬁcantly aﬀect EI in the other
three species. is is consistent with previous studies
[59, 60], in which EI remained constant or increased. Yet,
a constant EI value does not mean that the mechanical
properties do not change, because EI depends on both E
and I. Actually, wind signiﬁcantly increased E in C. inter-
media, and I in S. ruthenica in our experiment; while
wind did not change E, I or EI in A. squarrosum. us,
both A. ordosica and S. ruthenica tended to become more
ﬂexible under strong wind stress while C. intermedia
shoots tended to increase their rigidity.
e change in EI in this experiment is likely a result of
the change in stem base diameter, change in symmetry
and/or amounts of chemical compounds such as plant
stem cellulose content, all of which may have inﬂuenced
the ﬂexibility of the stem . SBD is tightly correlated
with I, the rigidity index . In A. ordosica, the increase
of SBD in treatment I corresponded with the increase of
I, and thereby increased EI. But the increase of SBD in A.
squarrosum did not coincide with changes in mechanical
properties, which may be attributed to the cancellation
eﬀect of a reduction of cellulose content.
Fig. 3 Plant size and shape traits of four plant species under diﬀerent wind treatments. Plant height (a), projected crown area (b), root length (c)
and STEPWISE bifurcation ratio(1:2) (d). Decrease, CK (ambient) and Increase refer to wind velocity treatments. Diﬀerent lowercase letters indicate
signiﬁcant diﬀerences among treatments at p < 0.05. The error bars are plotted by means ± SE
Page 11 of 16
Zhangetal. Plant Methods (2021) 17:5
Most of the time, the eﬀects of wind on plants are mod-
iﬁed by other environmental factors. e responses to
wind are speciﬁcally modiﬁed by shading, the nutritional
status of plants, by soil water and inherent plant traits,
such as clonality [16, 19, 32, 61]. ese factors may also
enhance the idiosyncrasy of shoot trait response patterns
to wind among species. Additionally, in the Mu Us Sand-
land, Aeolian sand displacement is an important environ-
mental factor. Wind denudation and sand burial are the
two main sand mobility processes that will also modify
the eﬀects of wind on plants (see “Background”). For
example, the root to shoot ratio of A. ordosica seedlings
was found to increase while the height and stem diam-
eter decreased under wind denudation of the soil surface
[27, 62]. e E, I and EI of C. intermedia were found to
decrease with sand burial . ese traits were also
inﬂuenced by our experimental wind perturbation. us,
plant responses to sand movement may also modify plant
wind resistance traits. Although the inﬂuences of wind-
blown sand on plants were minimized in our experi-
ment, it is still a very interesting question to address in
further studies. In-depth studies are needed to focus on
the eﬀects of confounded factors related to climate, soil
properties, Aaeolian sediment and inherent plant traits
on response patterns to wind.
Indirect wind eects onspecies’ performance
Under ﬁeld conditions, besides aﬀecting plant perfor-
mance through changing their leaf traits, wind can also
inﬂuence photosynthesis and transpiration through
changing the microclimate. Diﬀerent from branches and
trunks, leaves are highly ﬂexible, and only very strong
wind gusts could cause signiﬁcant damage, such as being
torn, shredded or pulled oﬀ the branches . Under less
extreme wind condition, many previous studies on single
species found that wind reduced the number of leaves,
leaf area and leaf dry mass, with an increase of leaf thick-
ness [5, 7, 17–19, 27]. Our results showed that, across
plant species, wind signiﬁcantly aﬀected leaf length only,
but had no signiﬁcant eﬀects on leaf area or SLA. e leaf
traits measured also showed very diﬀerent response pat-
terns to wind.
Wind speed plays an important role in the micro-
climate around a leaf attached to the plant. Increased
wind velocity could result in lower leaf boundary layer
Fig. 4 Plant biomass and its allocation in four plant species under diﬀerent wind treatments. Decrease, CK (ambient) and increase refer to wind
velocity treatments. Diﬀerent uppercase and lowercase letters indicate signiﬁcant diﬀerences among treatments at p < 0.05. The error bars are
plotted by means ± SE
Page 12 of 16
Zhangetal. Plant Methods (2021) 17:5
conductance which can cool the leaves and improve gas
exchange at higher temperatures [64, 65]. At low ambient
air temperature, photosynthesis may be reduced due to
below-optimal leaf temperatures and stomatal conduct-
ance [5, 16, 18].
In contrast, at low wind speed, a thicker leaf boundary
layer may cause high leaf temperatures and inhibit pho-
tosynthesis. In treatment D, the peak wind velocity at
noon was on average less than 0.5m/s (Fig. 2b). Under
such conditions the already high air temperatures (Addi-
tional ﬁle1: Fig. S3) will be ampliﬁed by the boundary
layer potentially risking acute heat damage . e neg-
ative eﬀects of low wind treatments on leaf performance
through excessive leaf temperatures may explain why
total biomass did not increase in three species out of four
(i.e. except C. intermedia) under our low wind treatment.
Smaller or narrower leaves are more suited to withstand
high air temperature due to better convective dissipation
. is could explain why the leaf area of A. ordosica
and S. ruthenica did not increase signiﬁcantly under low
wind treatment, even though leaf area tended to increase
at low wind condition overall. At the same ambient air
temperature, pinnate leaves dissipate heat more eﬀec-
tively than simple ones . C. intermedia has pinnate
leaves, which should help it to cope with high air tem-
perature. Trichomes can also help leaves to substantial
reduce sunlight absorption, which can reduce the dam-
age due to low wind and high temperature . ere
were 10–50 trichomes/mm2 on the leaves of A. squar-
rosum and more than 50 trichomes/mm2 on the leaves
of C. intermedia . is may explain why leaf area of
C. intermedia and A. squarrosum could increase signiﬁ-
cantly under low wind velocity.
Eects ofthewind treatments andits future application
Our new ﬁeld wind-funneling design (I) showed high
potential in synchronously increasing the wind velocity
in the treatment area roughly in proportion to the ambi-
ent wind velocity. rough the four funnel entrances in
four diﬀerent directions, the wind can be accelerated in
Fig. 5 Shoot property traits linked to wind resistance in four plant species under diﬀerent wind treatments: a Stem cellulose content, b Young’s
modulus (E), c Second moment of area (I), and d Flexural stiﬀness (EI). Decrease, CK (ambient) and Increase refer to wind velocity treatments.
Diﬀerent lowercase letters indicate signiﬁcant diﬀerences among treatments at p < 0.05. The error bars are plotted by means ± SE
Page 13 of 16
Zhangetal. Plant Methods (2021) 17:5
diﬀerent directions, which ensures that with the change
of wind direction, the wind ﬂowing through the fun-
nel can be continuously accelerated in an eﬀective way.
According to formula Vout = (s+
l)Vin/s, it is possible to
adjust the wind speed increase by adjusting the length of
the artiﬁcial air baﬄes. Wind acceleration eﬀects can be
estimated according to the quadrat length and the length
of baﬄes. In another ongoing experiment of ours, we are
also attempting to use this design at community scale
to detect the response of A. ordosica community to the
increased wind speed (Additional ﬁle1: Fig. S7A). In that
experiment, 4m × 4 m quadrats were set out and four
5m long and 1.2m high baﬄes were used to increase the
wind velocity. e results showed that the wind funnel
design also had potential to be used at community scale
(Additional ﬁle1: Fig. S7B–D). e experimental wind
baﬄes (treatment I) increased daily ambient wind veloc-
ity by 47% and maximum wind velocity by 130%, while
decreasing ambient air temperature by 0.7°C on average.
e height of the baﬄe would not aﬀect wind velocity
increase in our ideal model, while increasing the baﬄe
height would make the funnel closer to the sealed sta-
tus, which should make the results for the wind increase
treatment closer to those calculated by the model. us,
we suggest that the height of the baﬄe should also be
considered when building new wind funnels for getting
better acceleration results. Additionally, according to
the ﬂuid continuity equation, the border of the exit side
of the wind funnel has the shortest width, which means
that the wind velocity peak appears at this point. Because
the quadrat was set at 1m × 1m in our experiment, i.e. a
small area, the wind velocity may not vary much within
the quadrat. In larger quadrats wind might decline
through the quadrat in the wind ﬂow direction. Such a
gradient in the wind velocity should also be considered in
Our wind shield design (D), like many other wind
reduction designs, was eﬀective in decreasing wind
velocity. However, the use of wind shields could cause
two main confounding eﬀects: light shielding and heat
trapping. e choice of using transparent plastic sheets
was based on the intention to minimize the inﬂuence
of the wind shield on the light regime. However, we
are aware that the plastic sheets used in treatment D
could alter light quality somewhat by intercepting or
reﬂecting certain wavelength bands more than oth-
ers. Especially the red: far-red ratio could have been
aﬀected, which is known to aﬀect several plant perfor-
mance parameters [67, 68]. Also, the reﬂection of the
red and far-red part of the light spectrum might have
caused a minor increase of air temperature. In future
experiments of this kind, a material should be used that
causes minor and evenly distributed interception and
reﬂection of the whole range of plant-relevant wave-
lengths and the actual interception should be meas-
ured. Our wind decreasing treatment increase the air
temperature (especially around noon) by, on average,
1.7°C; even though the rationale of this treatment was
to decrease wind velocity simultaneously with wind
velocity enhancing methods under ﬁeld conditions in
a way that had minimal inﬂuences on the other envi-
ronmental factors. is drawback may be overcome
by adjusting the fencing method. For example, we also
tried a slightly diﬀerent design with a larger fencing
area and distance between quadrat and fences (Addi-
tional ﬁle1: Fig. S7A); this design seemed to have less
inﬂuence on the air temperature (increase by 0.03°C
on average; from our unpublished data). us, we sug-
gest that, with such improvements, this wind reduction
method can still be eﬀective and economical for use in
ﬁeld wind manipulating experiments.
Finally, we did not focus on the inﬂuence of the wind
on the reproductive traits of plants, as the early harvest
(needed for accurate measurement of E) did not allow
for enough ﬂowering or fruiting. However, it would be
interesting to test whether the inﬂuence of wind could
carry over to the next generation via the phenotypes of
the seeds (e.g. through amount of reserves). Also, future
experiments could test if the wind treatments could favor
the survival and reproductive output of certain geno-
types within a population of a species, thereby potentially
also aﬀecting the performance of future generations.
us we suggest that our wind design could be the per-
fect candidate for further experiments to address this
issue by collecting the seeds from plants from diﬀerent
wind treatments and then sow them to study subsequent
Our new experimental method to continuously increase
wind speed, using wind-funneling baﬄes, enabled us to
partially validate our null hypothesis that there are com-
mon whole-plant responses to wind stress across dif-
ferent plant species; strong wind signiﬁcantly aﬀected
plant-size-related traits in our experiment. However, the
response patterns of shoot traits, especially mechanical
properties, to wind velocity were idiosyncratic among
species. Furthermore, the wind eﬀects on plant perfor-
mance through changing leaf microclimate could be
negative or positive depending on confounding factors
including plant properties and local weather. In-depth
experiments are needed to disentangle these eﬀects of
increasing wind on plant performance under diﬀerent
wind speeds achieved through diﬀerent dimensions of
wind-funneling baﬄes under ﬁeld conditions.
Page 14 of 16
Zhangetal. Plant Methods (2021) 17:5
The online version contains supplementary material available at https ://doi.
org/10.1186/s1300 7-020-00705 -2.
Additional le1: TableS1. The speciﬁcations of all the instruments used
in the experiment. Fig. S1. The distribution of wind direction of maximum
wind speed from 2005 to 2015 at the meteorological station at Ordos
ecological station. Fig. S2. Background environmental conditions during
experimental period at Ordos ecological station. (A) Rainfall and tempera-
ture pattern within 2017. (B) Monthly mean and maximum wind velocity
within the experimental period (April to October of 2017). (C) Daily mean
and maximum wind velocity within the experimental period (April to
October of 2017). (D) The distribution of wind direction of maximum wind
speed from April to October of 2017. Fig. S3. Dynamics of air tempera-
ture (A), soil temperature (B), relative humidity (C) and volumetric water
content (D) (mean ± SE) under diﬀerent wind treatment over 24 h during
the experiment period. D in the legend means decreased wind velocity
treatment, CK means ambient wind velocity treatment, and I means
increased wind velocity treatment. Fig. S4. Comparison of the wind
treatment eﬀects between the whole data set and the westerly-wind data
subset extracted from the whole data set. D means decreased wind veloc-
ity treatment, CK means ambient wind velocity treatment, and I means
increased wind velocity treatment. D-W, CK-W and I-W, respectively, are
the westerly subsets of the data for the decreased, ambient and increased
wind velocity treatments. Fig. S5. Response of various traits to diﬀerent
wind velocity treatments in four plant species: Leaf length (A), leaf width
(B), leaf area (C), SLA (D), overall bifurcation ratio (E), Stem base diameter
(F), and stem lignin content (G). Decrease means decreased wind velocity
treatment, CK means ambient wind velocity treatment, and Increase
means increased wind velocity treatment. Diﬀerent lowercase letters indi-
cate signiﬁcant diﬀerences among the three treatments at P < 0.05. The
error bars are plotted by means ± SE. Fig. S6. Results of the redundancy
analysis (RDA) for the four plant species: distribution of each treatment of
4 plant species on the RDA1 × RDA2 plane. The relationship is signiﬁcant
(p < 0.01) based on 999 permutations. The adjusted R2 is 0.61. Suﬃx D of
plant species name in the legend represents decrease wind velocity treat-
ment, CK represented ambient wind velocity treatment, and I represented
increase wind velocity treatment. The ﬁrst dimension (RDA1), which
describes 34.81% of the total variability, is positively correlated with stem
cellulose content, root/shoot ratio and I, and negatively correlated with
total biomass, shoot biomass, root biomass, plant height, leaf length, stem
base diameter, leaf area, SLA, projected crown area, stem lignin content, E,
EI, OBR and SBR(1:2). The second dimension (RDA2), which explains 16.86%
of the total variability, is positively correlated with EI, I, stem base diameter,
leaf area, OBR and SBR(1:2), while it is negatively correlated with root
length, stem cellulose content, E, projected crown area and SLA. For trait
abbreviations see the main text. Fig. S7. Application of wind manipulation
design at community scale (4 m × 4 m quadrat, A. ordosica community)
from our ongoing experiment. (A) Picture of the experiment set and detail
of treatments. The Decreased wind velocity treatment (D) was imple-
mented through a wind shield. The plastic chambers did not have a roof,
so that ambient air could mix with that in the cubicles freely. Distances
were kept between the wind shield and quadrat to avoid shading eﬀect.
There was no manipulation for the ambient treatment (CK). The Increased
wind velocity treatment (I) was implemented through experimental wind
baﬄes. We placed sheets at each side directed towards the North-East,
North-West, South-East and South-West, respectively from each quadrat
with plants, to converge and increase wind velocity. (B) Diurnal mean
wind velocity pattern (mean ± SE) over 24 h under wind treatments. (C)
Diurnal maximum wind velocity pattern (mean ± SE) over 24 h under
wind treatments. (D) Eﬀects of wind treatments. The black dashed line
with the equation (
) represents the ideal wind velocity
that could be reached by acceleration. The black dashed line with the
equation (y = x) represents the wind velocity in the treatment CK. Red
dots and line represent the relationship between the temporally matched
hourly maximum wind velocity in I and CK. Blue dots and line represent
the relationship between the temporally matched hourly maximum wind
velocity in D and CK. Regression equations and R2 are given.
Special thanks to Mr. Qiang Zhang for the great support during the
SZ, GL, XY and ZH conceived the ideas and designed methodology; SZ and
QC collected the data; SZ, GL, JHC and XY analysed the data; SZ, XY, ZH and
JHC led the writing of the manuscript. All authors contributed critically to the
drafts. All authors read and approved the ﬁnal manuscript.
This research was supported by the Strategic Priority Research Program of the
Chinese Academy of Sciences (XDA23080302), National Natural Science Foun-
dation of China (31470032), and the CAS President’s International Fellowship
Initiative (PIFI, 2018VCA0014) for JHCC.
Availability of data and materials
All data generated or analysed during this study are included in this published
article [and its additional information ﬁles].
Ethics approval and consent to participate
Consent for publication
The authors declare that they have no competing interests.
1 State Key Laboratory of Vegetation and Environmental Change, Institute
of Botany, Chinese Academy of Sciences, Beijing 100093, People’s Republic
of China. 2 University of Chinese Academy of Sciences, Beijing 100049, People’s
Republic of China. 3 Systems Ecology, Department of Ecological Science, Vrije
Universiteit, De Boelelaan 1085, 1081 HV Amsterdam, The Netherlands.
Received: 1 March 2020 Accepted: 24 December 2020
1. Cheplick GP. Responses of native plant populations on an unprotected
beach to disturbance by storm-induced overwash events. Plant Ecol.
2. He WM, Xi WM, Anten NPR. Chronic wind and plant communities in
dunes: total biomass, inter-speciﬁc competition, and species abundance.
J Arid Environ. 2013;90:69–76.
3. Mitchell SJ. Wind as a natural disturbance agent in forests: a synthesis.
4. Thomas SC, Mar tin AR, Mycroft EE. Tropical trees in a wind-exposed island
ecosystem: height-diameter allometry and size at onset of maturity. J
5. Anten NPR, Alcala-Herrera R, Schieving F, Onoda Y. Wind and mechani-
cal stimuli diﬀerentially aﬀect leaf traits in Plantago major. New Phytol.
6. de Langre E. Eﬀects of wind on plants. Annu Rev Fluid Mech.
7. Niklas KJ. Diﬀerences between Acer saccharum leaves from open and
wind-protected sites. Ann Bot. 1996;78(1):61–6.
8. Gardiner B, Berry P, Moulia B. Review: Wind impacts on plant growth,
mechanics and damage. Plant Sci. 2016;245:94–118.
9. Anten NPR, Casado-Garcia R, Nagashima H. Eﬀects of mechanical stress
and plant density on mechanical characteristics, growth, and lifetime
reproduction of tobacco plants. Am Nat. 2005;166(6):650–60.
10. Bang C, Sabo JL, Faeth SH. Reduced wind speed improves plant growth
in a desert city. PLoS ONE. 2010;5(6):8.
11. Coutand C, Julien JL, Moulia B, Mauget JC, Guitard D. Biomechanical
study of the eﬀect of a controlled bending on tomato stem elongation:
global mechanical analysis. J Exp Bot. 2000;51(352):1813–24.
Page 15 of 16
Zhangetal. Plant Methods (2021) 17:5
12. Jaﬀe MJ. Thigmomorphogenesis—response of plant-growth and devel-
opment to mechanical stimulation—with special reference to Bryonia-
dioica. Planta. 1973;114(2):143–57.
13. Coutand C. Mechanosensing and thigmomorphogenesis, a physiological
and biomechanical point of view. Plant Sci. 2010;179(3):168–82.
14. Brucher t F, Gardiner B. The eﬀect of wind exposure on the tree aerial
architecture and biomechanics of sitka spruce (Picea sitchensis, Pinaceae).
Am J Bot. 2006;93(10):1512–21.
15. Cleugh HA, Miller JM, Bohm M. Direct mechanical eﬀects of wind on
crops. Agrofor Syst. 1998;41(1):85–112.
16. Retuerto R, Woodward FI. Eﬀects of windspeed on the growth and
biomass allocation of white mustard Sinapis Alba L. Oecologia.
17. Cordero RA, Fetcher N, Voltzow J. Eﬀects of wind on the allometry of two
species of plants in an Elﬁn Cloud Forest. Biotropica. 2007;39(2):177–85.
18. Smith VC, Ennos AR. The eﬀects of air ﬂow and stem ﬂexure on the
mechanical and hydraulic properties of the stems of sunﬂowers Helian-
thus annuus L. J Exp Bot. 2003;54(383):845–9.
19. Wang YH, He WM, Dong M, Yu FH, Zhang LL, Cui QG, Chu Y. Eﬀects of
shaking on the growth and mechanical properties of Hedysarum laeve
may be independent of water regimes. Int J Plant Sci. 2008;169(4):503–8.
20. Danin A. Adaptations of Stipagrostis species to desert dunes. J Arid Envi-
21. Dech JP, Maun MA. Adventitious root production and plastic resource
allocation to biomass determine burial tolerance in woody plants from
central Canadian coastal dunes. Ann Bot. 2006;98(5):1095–105.
22. Shao Y. Physics and modelling of wind erosion. Berlin: Springer; 2008.
23. Xu L, Huber H, During HJ, Dong M, Anten NPR. Intraspeciﬁc variation of a
desert shrub species in phenotypic plasticity in response to sand burial.
New Phytol. 2013;199(4):991–1000.
24. Yu FH, Wang N, He WM, Chu Y, Dong M. Adaptation of rhizome connec-
tions in drylands: increasing tolerance of clones to wind erosion. Ann Bot.
25. Chen XH, Duan ZH. Changes in soil physical and chemical properties
during reversal of desertiﬁcation in Yanchi County of Ningxia Hui autono-
mous region, China. Environ Geol. 2009;57(5):975–85.
26. Fryrear DW, Stubbend J, McCully WG. Grass seedling response to wind
and windblown sand. Crop Sci. 1973;13(6):622–5.
27. Li SL, Werger MJA, Zuidema PA, Yu FH, Dong M. Seedlings of the
semi-shrub Artemisia ordosica are resistant to moderate wind denu-
dation and sand burial in Mu Us sandland, China. Trees-Struct Funct.
28. Zhou RL, Li YQ, Zhao HL, Drake S. Desertiﬁcation eﬀects on C and N con-
tent of sandy soils under grassland in Horqin, northern China. Geoderma.
29. Zhao WZ, Li QY, Fang HY. Eﬀects of sand burial disturbance on seedling
growth of Nitraria sphaerocarpa. Plant Soil. 2007;295(1–2):95–102.
30. Watt MS, Moore JR, McKinlay B. The inﬂuence of wind on branch charac-
teristics of Pinus radiata. Trees-Struct Funct. 2005;19(1):58–65.
31. Roden JS. Modeling the light interception and carbon gain of individual
ﬂuttering aspen (Populus tremuloides Michx) leaves. Trees-Struct Funct.
32. Henr y HAL, Thomas SC. Interactive eﬀects of lateral shade and wind on
stem allometry, biomass allocation, and mechanical stability in Abutilon
theophrasti (Malvaceae). Am J Bot. 2002;89(10):1609–15.
33. Tamasi E, Stokes A, Lasserre B, Danjon F, Berthier S, Fourcaud T, Chia-
tante D. Inﬂuence of wind loading on root system development and
architecture in oak (Quercus robur L.) seedlings. Trees-Struct Funct.
34. Xu L, Yu FH, Werger M, Dong M, Anten NPR. Interactive eﬀects of
mechanical stress, sand burial and defoliation on growth and mechanical
properties in Cynanchum komarovii. Plant Biol. 2013;15(1):126–34.
35. Gleason SM, Williams LJ, Read J, Metcalfe DJ, Baker PJ. Cyclone eﬀects on
the structure and production of a tropical upland rainforest: Implications
for life-history tradeoﬀs. Ecosystems. 2008;11(8):1277–90.
36. Meng SX, Rudnicki M, Lieﬀers VJ, Reid DEB, Silins U. Preventing crown
collisions increases the crown cover and leaf area of maturing lodgepole
pine. J Ecol. 2006;94(3):681–6.
37. Kern KA, Ewers FW, Telewski FW, Koehler L. Mechanical perturbation
aﬀects conductivity, mechanical properties and aboveground biomass of
hybrid poplars. Tree Physiol. 2005;25(10):1243–51.
38. Niklas KJ. Eﬀects of vibration on mechanical properties and biomass
allocation pattern of Capsella bursa-pastoris (Cruciferae). Ann Bot.
39. Paul-Victor C, Rowe N. Eﬀect of mechanical perturbation on the biome-
chanics, primary growth and secondary tissue development of inﬂo-
rescence stems of Arabidopsis thaliana. Ann Bot. 2011;107(2):209–18.
40. Pruyn ML, Ewers BJ, Telewski FW. Thigmomorphogenesis: changes in
the morphology and mechanical properties of two Populus hybrids in
response to mechanical perturbation. Tree Physiol. 2000;20(8):535–40.
41. Wang YH, Dong M, Yu FH, Jiang H, Yu SQ, Lin XQ, He WM. Mechanical
shaking and soil water aﬀect the growth of Psammochloa villosa in the
Mu Us Sandland. J Arid Environ. 2011;75(10):974–7.
42. Grace J. Eﬀect of wind on grasses 1. Cuticular and stomatal transpira-
tion. J Exp Bot. 1974;25(86):542–51.
43. Wu T, Zhang P, Zhang L, Wang GG, Yu M. Morphological response
of eight quercus species to simulated wind load. PLoS ONE.
44. Okin GS, Heras MM-d, Saco PM, Throop HL, Vivoni ER, Parsons AJ,
Wainwright J, Peters DPC. Connectivity in dryland landscapes: shifting
concepts of spatial interactions. Front Ecol Environ. 2015;13(1):20–7.
45. Rachal DM, Okin GS, Alexander C, Herrick JE, Peters DPC. Modifying
landscape connectivity by reducing wind driven sediment redistribu-
tion, Northern Chihuahuan Desert, USA. Aeol Res. 2015;17:129–37.
46. Editorial Board of the Flora of China of Chinese Academy of Sciences:
Flora of China. Beijing: Science Press; 1991.
47. Liu ZM, Yan QL, Baskin CC, Ma JL. Burial of canopy-stored seeds in the
annual psammophyte Agriophyllumsquarrosum Moq. (Chenopodi-
aceae) and its ecological signiﬁcance. Plant Soil. 2006;288(1–2):71–80.
48. Ma JL, Liu ZM. Spatiotemporal pattern of seed bank in the annual
psammophyte Agriophyllum squarrosum Moq. (Chenopodiaceae) on
the active sand dunes of northeastern Inner Mongolia, China. Plant
49. Lewis RJ, Lanker MD. Branching pattern of gametophytes and tetraspo-
rophytes of Ceramium codicola (Ceramiales, Rhodophyta) is related to
phase and reproductive structures. Phycologia. 2004;43(2):121–5.
50. Shah DU, Reynolds TPS, Ramage MH. The strength of plants: theory and
experimental methods to measure the mechanical properties of stems.
J Exp Bot. 2017;68(16):4497–516.
51. Ververis C, Georghiou K, Christodoulakis N, Santas P, Santas R. Fiber
dimensions, lignin and cellulose content of various plant materials and
their suitability for paper production. Ind Crop Prod. 2004;19(3):245–54.
52. Guo W, Li B, Zhang X, Wang R. Architectural plasticity and growth
responses of Hippophae rhamnoides and Caragana intermedia seed-
lings to simulated water stress. J Arid Environ. 2007;69(3):385–99.
53. Ding ZL, Derbyshire E, Yang SL, Sun JM, Liu TS. Stepwise expan-
sion of desert environment across northern China in the past 3.5
Ma and implications for monsoon evolution. Earth Planet Sci Lett.
54. Du S, Li B, Chen M, Zhang DD, Xiang R, Niu D, Wen X, Ou X. Kiloyear-scale
climate events and evolution during the last Interglacial, Mu Us Desert
China. Quatern Int. 2012;263:63–70.
55. Sun J. Origin of eolian sand mobilization during the past 2300 years in
the Mu Us Desert, China. Quatern Res. 2000;53(1):78–88.
56. Liu Y, Schieving F, Stuefer JF, Anten NPR. The eﬀects of mechanical stress
and spectral shading on the growth and allocation of ten genotypes of a
stoloniferous plant. Ann Bot. 2007;99(1):121–30.
57. Butler DW, Gleason SM, Davidson I, Onoda Y, Westoby M. Safety and
streamlining of woody shoots in wind: an empirical study across 39 spe-
cies in tropical Australia. New Phytol. 2012;193(1):137–49.
58. Puijalon S, Bouma TJ, Douady CJ, van Groenendael J, Anten NPR, Martel
E, Bornette G. Plant resistance to mechanical stress: evidence of an
avoidance-tolerance trade-oﬀ. New Phytol. 2011;191(4):1141–9.
59. Jaﬀe MJ, Telewski FW, Cooke PW. Thigmomorphogenesis: On the
mechanical properties of mechanically perturbed bean plants. Physiol
60. Telewski FW, Jaﬀe MJ. Thigmomorphogenesis: Field and laboratory
studies of Abies fraseri in response to wind or mechanical perturbation.
Physiol Plant. 1986;66(2):211–8.
61. Grace J, Pitcairn CER, Russell G, Dixon M. The eﬀects of shaking on the
growth and water relations of Festuca-arundinacea schreb. Ann Bot.
Page 16 of 16
Zhangetal. Plant Methods (2021) 17:5
fast, convenient online submission
thorough peer review by experienced researchers in your ﬁeld
rapid publication on acceptance
support for research data, including large and complex data types
gold Open Access which fosters wider collaboration and increased citations
maximum visibility for your research: over 100M website views per year
At BMC, research is always in progress.
Learn more biomedcentral.com/submissions
Ready to submit your research
Ready to submit your research
? Choose BMC and benefit from:
? Choose BMC and benefit from:
62. Li SL, Zuidema PA, Yu FH, Werger MJA, Dong M. Eﬀects of denudation
and burial on growth and reproduction of Artemisia ordosica in Mu Us
Sandland. Ecol Res. 2010;25(3):655–61.
63. Vogel S. Leaves in the lowest and highest winds: temperature, force and
shape. New Phytol. 2009;183(1):13–26.
64. Daudet FA, Le Roux X, Sinoquet H, Adam B. Wind speed and leaf bound-
ary layer conductance variation within tree crown: consequences on
leaf-to-atmosphere coupling and tree functions. Agric For Meteorol.
65. Grantz DA, Vaughn DL. Vertical proﬁles of boundary layer conductance
and wind speed in a cotton canopy measured with heated brass sur-
rogate leaves. Agric For Meteorol. 1999;97(3):187–97.
66. Liu Y, Li X, Li MM, Liu D, Zhang W. Leaf (or assimilation branch) epidermal
micromorphology of desert plant in arid and semi-arid areas of China.
Chin J Plant Ecol. 2016;40(11):1189–207.
67. Holalu SV, Finlayson SA. The ratio of red light to far red light alters Arabi-
dopsis axillary bud growth and abscisic acid signalling before stem auxin
changes. J Exp Bot. 2017;68(5):943–52.
68. Ugar te CC, Trupkin SA, Ghiglione H, Slafer G, Casal JJ. Low red/
far-red ratios delay spike and stem growth in wheat. J Exp Bot.
Springer Nature remains neutral with regard to jurisdictional claims in pub-
lished maps and institutional aﬃliations.
Terms and Conditions
Springer Nature journal content, brought to you courtesy of Springer Nature Customer Service Center GmbH (“Springer Nature”).
Springer Nature supports a reasonable amount of sharing of research papers by authors, subscribers and authorised users (“Users”), for small-
scale personal, non-commercial use provided that all copyright, trade and service marks and other proprietary notices are maintained. By
purposes, Springer Nature considers academic use (by researchers and students) to be non-commercial.
These Terms are supplementary and will apply in addition to any applicable website terms and conditions, a relevant site licence or a personal
subscription. These Terms will prevail over any conflict or ambiguity with regards to the relevant terms, a site licence or a personal subscription
(to the extent of the conflict or ambiguity only). For Creative Commons-licensed articles, the terms of the Creative Commons license used will
We collect and use personal data to provide access to the Springer Nature journal content. We may also use these personal data internally within
ResearchGate and Springer Nature and as agreed share it, in an anonymised way, for purposes of tracking, analysis and reporting. We will not
otherwise disclose your personal data outside the ResearchGate or the Springer Nature group of companies unless we have your permission as
While Users may use the Springer Nature journal content for small scale, personal non-commercial use, it is important to note that Users may
use such content for the purpose of providing other users with access on a regular or large scale basis or as a means to circumvent access
use such content where to do so would be considered a criminal or statutory offence in any jurisdiction, or gives rise to civil liability, or is
falsely or misleadingly imply or suggest endorsement, approval , sponsorship, or association unless explicitly agreed to by Springer Nature in
use bots or other automated methods to access the content or redirect messages
override any security feature or exclusionary protocol; or
share the content in order to create substitute for Springer Nature products or services or a systematic database of Springer Nature journal
In line with the restriction against commercial use, Springer Nature does not permit the creation of a product or service that creates revenue,
royalties, rent or income from our content or its inclusion as part of a paid for service or for other commercial gain. Springer Nature journal
content cannot be used for inter-library loans and librarians may not upload Springer Nature journal content on a large scale into their, or any
other, institutional repository.
content on this website and may remove it or features or functionality at our sole discretion, at any time with or without notice. Springer Nature
may revoke this licence to you at any time and remove access to any copies of the Springer Nature journal content which have been saved.
To the fullest extent permitted by law, Springer Nature makes no warranties, representations or guarantees to Users, either express or implied
with respect to the Springer nature journal content and all parties disclaim and waive any implied warranties or warranties imposed by law,
including merchantability or fitness for any particular purpose.
Please note that these rights do not automatically extend to content, data or other material published by Springer Nature that may be licensed
from third parties.
If you would like to use or distribute our Springer Nature journal content to a wider audience or on a regular basis or in any other manner not
expressly permitted by these Terms, please contact Springer Nature at