American Journal of Botany 90(8): 1253–1256. 2003.

BRIEF COMMUNICATION

FIELD MEASUREMENTS OF WIND SPEED AND

RECONFIGURATION IN ARUNDO DONAX (POACEAE)
WITH ESTIMATES OF DRAG FORCES1

OLGA SPECK2
Plant Biomechanics Group, Fakultät für Biologie, Albert-Ludwigs-Universität Freiburg, Schänzlestr. 1, D-79104 Freiburg, Germany

The giant reed (Arundo donax) is well known as a species that can withstand high wind loads without mechanical damage. To
examine wind impact, profiles of vertical wind speeds in the plant’s natural habitat (southern France) were measured at the edge and
within a stand in the main wind direction. Wind speed was recorded simultaneously at five heights. For 75 measurements of within-
canopy wind speed profiles, the attenuation coefficient was 4.4 6 0.5, a value typical for plant stands with very dense canopies. Video
recordings proved that A. donax becomes streamlined with increasing wind speed, reducing the projected surface area of leaves and
stem. The total projected surface area is a function of wind speed and can be characterized by a second-order polynomial regression
curve. For small wind velocities up to 1 m/s, the calculated drag force is proportional to the square of the wind speed. However, when
A. donax plants are subjected to higher wind speeds (1.5–10 m/s), the drag force becomes directly proportional to the wind speed.
Streamlining is a potentially important adaptation for withstanding high wind loads, especially for individual plants and plants at the
edge of stands, whereas in dense stands streamlining probably plays a minor role.

Key words: Arundo donax; biomechanics; drag force; France; Poaceae; profiles of vertical wind speeds; projected surface area.

Arundo donax L. plants grow in dense stands and are often
planted in hedge rows to serve as wind shelter because the
slender culms have a pronounced pliability and excellent wind
resistance. Wind and gravity are the two basic forces that af-
fect plants in a variety of ways (Fraser, 1962). Measurements
of air movements, such as profiles of vertical wind speeds, are
necessary to understand the biotic effects of wind (Grace,
1989). The shape of a vertical wind profile is characterized by
the wind velocity, which changes with height above ground,
surface roughness, and length of uniform surface over which
the wind has blown (Campbell, 1977). Predictions of wind
velocity within a canopy are complicated, but the wind veloc-
ity can be estimated by formulas suggested by Bussinger
(1975). Niklas and Spatz (2000) explored the effects of five
different biologically realistic wind speed profiles on safety,
wind-induced bending moments, and stresses generated in the
stems of leafless cherry trees.
Overcritical wind loads may cause severe damage in forests,
for example (cf. Gardiner and Quine, 1994; Coutts and Grace,
1995) but also, due to lodging, in crop plantings (cf. Crook
and Ennos, 2000). While a plant is subjected to wind, it re-
sponds with streamlining, especially of its branches and leaves,
a behavior that may significantly reduce the danger of me-
chanical damage caused by drag forces. Vogel (1989) has
shown that the drag on leaves varies with wind speed, an effect
caused by reconfiguration of the leaves. Daffodil flowers also
1
Manuscript received 14 November 2002; revision accepted 28 March
2003.
The author wishes to thank Prof. Dr. K. J. Niklas for fastidious measure-
ments of the projected surface areas. I am grateful to Prof. Dr. T. Speck for
his help with the field experiments and to Prof. Dr. H.-Ch. Spatz for helpful
discussions. I also thank two anonymous reviewers for helpful criticism of
the manuscript. This study was in part supported by the DaimlerChrysler AG
and Alumni Freiburg.
2
Phone: 149 (0)761-203-2803; FAX: 149 (0)761-203-2804; e-mail:
olga.speck@biologie.uni-freiburg.de.
change both shape and orientation in response to wind above
5 m/s, and both changes lead to reductions of their drag by
;30% (Etnier and Vogel, 2000).
The ultimate aim of our studies on A. donax is to analyze
how this plant avoids or reduces mechanical damage due to
wind loads. To answer this question we have studied the
plant’s response to varying wind speeds and have quantified
its oscillation and damping behavior (e.g., Spatz and Speck,
2002). Further, we examined the range and profiles of wind
speeds with which these plants have to cope and the wind-
induced streamlining of the plants. The objectives of this study
were (1) to measure wind speed profiles and (2) to analyze the
correlation between the drag force on the plant and the wind
velocity.
MATERIALS AND METHODS
Field measurements were carried out in the natural habitat of A. donax in
the Camargue (southern France).
Wind speed profile—Vertical wind speed profiles in an open area were
measured at the edge of an approximately 4-m-high A. donax stand (0.5 m
upwind in front of the stand) and at 9.0 m within this stand in the direction
of the main wind (Fig. 1). Wind speed was recorded simultaneously at five
heights on a mast. Four hot-wire anemometers were used to measure the wind
speed at 1 m, 2 m (both custom-made devices), 3.52 m, and 5 m (both Tri-
Sense Model No. 37000-00, Cole-Palmer, Niles, Illinois, USA) above ground.
One lightweight three-cup anemometer (Casella, London, UK) was used to
measure the average wind speed above the canopy (6.26 m high) during the
experimental period. Seventy-five profiles of wind speeds within the canopy
and 50 at the edge of the stand were measured. On average, a measurement
was taken every 45 s.
An equation that describes the wind speed (uZ) within the top layer of most
plant canopies (i.e., for the uppermost 36%), as well as the top and middle
layer of plant stands with uniform canopies, is given by Campbell (1977) and
Niklas (1992):
uz 5 utop e[a(z/h21)] (1)

1253
1254 AMERICAN JOURNAL
Fig. 1. Shape of wind speed profiles at five heights above ground (1 m, 2 m, 3.52 m, 5 m, and 6.26 m), and the experimental setup at the edge of a stand
of Arundo donax. Mean wind velocity and standard deviation are shown. The canopy averaged 4 m in height.
where utop is the wind speed at the top of the canopy, h the canopy height, a
the attenuation coefficient, and z the height of measurement (0 # z # h). The
attenuation coefficient ranges from 0 for very sparse canopies to 5 for very
dense canopies (Niklas and Speck, 2001).
Projected surface area and wind speed—A side view of a typical A. donax
plant (height: 3.3 m) was videotaped. The video camera (Sharp, Osaka, Japan,
Model VL-C750S) was arranged in such a way that the connecting line be-
tween camcorder and plant was parallel to the main wind direction. The wind
speed was measured with a hot-wire anemometer (Tri-Sense Model No 37000-
00, Cole-Palmer), which records wind velocity only for the main wind direc-
tion. With a second camera (Sony, Japan, Model CCD-V800E), the display
of the anemometer was videotaped. With a splitter (Panasonic, Osaka, Japan,
Digital AV Mixer WJ-MX 20), the recordings of the bending plant and the
corresponding wind speed were combined on one tape.
OF BOTANY [Vol. 90
With the aid of Occulus frame grabber software (Computer Products, Inc.),
selected video images were transferred onto the hard disk of a computer.
Projected surface areas of individual images of stems and leaves were mea-
sured electronically with the software SECTION. This program was used to
calculate the surface areas of specified regions of each image once the mag-
nification of each image is set. For each image, the program was also used
to measure the linear distance of markers fixed to the stem with respect to
ground level (cf. Niklas and Spatz, 2000).
For individual objects in high Reynolds number flow, the drag force Df(u)
is related to the square of the ambient wind velocity and the projected surface
area of stem and leaves of plants by the following equation:
Df(u) 5 0.5 CD r u2 ASL(u) (2)
where CD is the drag coefficient, r the density of air (1.29 kg/m ), and u the
3
wind speed. Assuming that CD 5 1.0 (cf. Vogel, 1981; Niklas, 1994), the
August 2003] SPECK—WIND-INDUCED
Fig. 2. The relationship between wind speed and total projected surface
area of an individual plant of Arundo donax. The second-order polynomial
regression curve has a coefficient of determination of R2 5 0.9742.
term 0.5 CD r equals 0.65. The projected surface area of stem and leaves,
ASL(u), depends on the wind speed. The drag force is proportional to the
product of ambient wind speed squared times the total projected surface area
of the plant [Df(u) 5 0.65 u2 ASL(u)].
RESULTS AND DISCUSSION
Wind speed profile—Vertical wind speed profiles in an open
area were measured at the edge of an A. donax stand (approx-
imately 4 m high) and within this stand in the direction of the
main wind. The mean wind velocity did not vary much during
the 1 h, 35 min of the recording, except for individual gusts
at different periods. Within the stand (Fig. 1), the average wind
speeds 6 SD of 75 recordings measured at four heights were
1.04 6 0.23 m/s (1 m), 0.87 6 0.16 m/s (2 m), 1.74 6 0.39
m/s (3.52 m), and 6.43 6 1.23 m/s (5 m), whereas above the
canopy (6.26 m) an average wind speed of 6.01 m/s was found
during the experimental period. At the edge of the stand, 50
measurements of wind speed averaged 2.42 6 0.87 m/s (1 m),
3.56 6 1.02 m/s (2 m), 3.22 6 0.49 m/s (3.52 m), and 4.86
6 0.66 m/s (5 m); above the canopy (6.26 m) an average wind
speed of 4.92 m/s was measured. The average wind speed was
considerably reduced within the vegetation compared to values
found at the edge of the stand. In addition, the probability of
wind peaks is considerably reduced in the stand, which is mir-
rored by the very small standard deviations of wind speeds
within the stand. Both findings are a consequence of the ‘‘self-
sheltering effect’’ caused by the dense stand growth. In both
profiles, the wind velocity abruptly increased above the closed
canopy. In the latter case, the values within the vegetation
exceeded those measured at the edge.
For 75 profiles of wind speeds, the attenuation coefficient
was calculated following Eq. 1 and yielded on average 4.4 6
0.5. This value is typical for dense stands of plants with uni-
form shapes and heights (cf. Campbell, 1977).
Projected surface area and estimated drag—The video re-
cording showed that A. donax streamlines and tends to reduce
the projected surface area of leaves and stems with increasing
wind speed. Even at relatively low wind speeds, the leaves
start to orientate leeward and the apicalmost stem part begins
to bend. At higher wind speeds, the leaves flutter and are
entirely streamlined, and the stems curvilinearly bend with an
LOADS ON ARUNDO DONAX 1255
Fig. 3. The relationship between wind speed and drag force of an indi-
vidual plant of Arundo donax. The drag force is calculated using Eq. 2 (see
text), where the projected surface area is a function of wind speed (see Fig.
2). For small wind velocities up to 1 m/s (see inset), the drag force is pro-
portional to the square of the wind velocity (R2 5 0.994), whereas a linear
proportionality results for higher wind speeds between 1.5 and 10 m/s (R2 5
0.988).
acropetally increasing curvature. At the highest wind speed
measured (10 m/s), the maximum height of the plant is re-
duced by 45%. The correlation of wind speed and total pro-
jected surface area can be characterized by a decreasing sec-
ond-order polynomial (Fig. 2), mirroring the streamlining of
the plant.
Figure 3 shows the correlation of wind speed and calculated
drag force. The drag force Df(u) was calculated using Eq. 2,
where the projected surface area ASL(u) is a function of wind
speed u. For small wind speeds (uwind , 1 m/s), the drag force
is proportional to the square of the wind velocity, whereas a
linear proportionality between drag force and wind velocity is
found for higher wind speeds (uwind: 1.5–10 m/s) (cf. Fraser,
1962; Niklas, 1992).
As a consequence of this behavior, the real drag forces that
individual A. donax plants have to withstand are smaller than
the forces predicted by Eq. 2 without considering streamlining,
especially at higher wind speeds. The difference between ex-
trapolated values using the second-order polynomial equation
(see Fig. 3, inset) and the values taking the reduction of the
projected surface area into account (Fig. 2) is 19% for uwind 5
2.5 m/s, 54% for uwind 5 5 m/s, 66% for uwind 5 7.5 m/s, and
73% for uwind 5 10 m/s. These results prove that, because of
streamlining of leaves and stem, individual A. donax plants
can considerably reduce the drag forces acting on them and
therefore withstand considerably higher wind forces without
mechanical damage.
Streamlining seems to mainly benefit plants growing in iso-
lation or at the edge of the stand. Within the stand, wind
speeds are low (,1.7 m/s) and streamlining probably plays a
minor role. However, at the edge, where wind velocities are
up to 3.6 m/s, the drag force is significantly reduced due to
streamlining. In addition, plants growing in stands interact and
may form a drag-reducing aerodynamic unit, especially at high
1256 AMERICAN JOURNAL
wind speeds, an effect well known from compound leaves (Vo-
gel, 1989). Thus, streamlining may permit plants to survive at
the edge and facilitate progressive growth of clonal stands. On
the other hand, through ‘‘self-sheltering’’ stand growth appar-
ently removes the threat of wind-damage for the great bulk of
the plants.
LITERATURE CITED
BUSSINGER, J. A. 1975. Aerodynamics of vegetated surfaces. In D. A.
DeVries and N. H. Afgan [eds.], Heat and mass transfer in the biosphere,
104–119. Wiley, New York, New York, USA.
CAMPBELL, G. S. 1977. An introduction to environmental biophysics. Spring-
er, Ulm, New York, New York, USA.
COUTTS, M. P., AND J. GRACE [EDS.]. 1995. Wind and trees. Cambridge
University Press, Cambridge, UK.
CROOK, M., AND A. R. ENNOS. 2000. A field based method of quantifying
the lodging resistance of wheat cultivars. In H.-Ch. Spatz and T. Speck
[eds.], Plant biomechanics 2000, 315–320. Thieme, Stuttgart, Germany.
ETNIER, S. E., AND S. VOGEL. 2000. Reorientation of daffodil (Narcissus:
Amaryllidaceae) flowers in wind: drag reduction and torsional flexibility.
American Journal of Botany 87: 29–32.
OF BOTANY [Vol. 90
FRASER, A. I. 1962. Wind tunnel studies of the forces acting on the crown
of small trees. Report on Forest Research 178–183.
GARDINER, B. A., AND C. P. QUINE. 1994. Wind damage to forests. Biomi-
metics 2: 139–147.
GRACE, J. 1989. Measurements of wind speed near vegetation. In R. W.
Pearcy, J. R. Ehleringer, H. A. Mooney, and P. W. Rundel [eds.], Plant
physiological ecology, 57–73. Chapman and Hall, New York, New York,
USA.
NIKLAS, K. J. 1992. Plant biomechanics: an engineering approach to plant
form and function. University of Chicago Press, Chicago, Illinois, USA.
NIKLAS, K. J. 1994. Plant allometry: the scaling of form and process. Uni-
versity of Chicago Press, Chicago, Illinois, USA.
NIKLAS, K. J., AND H.-CH. SPATZ. 2000. Wind-induced stresses in cherry
trees: evidence against the hypothesis of constant stress levels. Trees 14:
230–237.
NIKLAS, K. J., AND T. SPECK. 2001. Evolutionary trends in safety factors
against wind-induced stem failure. American Journal of Botany 88:
1266–1278.
SPATZ, H.-CH., AND O. SPECK. 2002. Oscillation frequencies of tapered plant
stems. American Journal of Botany 89: 1–11.
VOGEL, S. 1981. Life in moving fluids. Willard Grant Press, Boston, Mas-
sachusetts, USA.
VOGEL, S. 1989. Drag and reconfiguration of broad leaves in high winds.
Journal of Experimental Botany 40: 941–948.