Tree Physiology 21, 403–408
© 2001 Heron Publishing—Victoria, Canada
Enhanced transpiration in response to wind effects at the edge of a
blue gum (Eucalyptus globulus) plantation
P. J. TAYLOR,1,3 I. K. NUBERG1 and T. J. HATTON2
1
Department of Agronomy and Farming Systems, University of Adelaide, Roseworthy, SA 5371, Australia
2
CSIRO Land and Water, Private Bag, PO Wembley, WA 6014, Australia
3
Present address: Agriculture Western Australia, Manjimup Horticultural Research Institute, Locked Bag No. 7, Manjimup, WA 6258, Australia
Received June 12, 1998
Summary In Australia, tree planting has been widely promoted to alleviate dryland salinity and one proposed planting
configuration is that of strategically placed interception belts.
We conducted an experiment to determine the effect of tree position in a belt on transpiration rate. We also assessed how
much the effect of tree position can be explained by advection
and environmental conditions. Daily transpiration rates were
determined by the heat pulse velocity technique for four edge
and 12 inner trees in a 7-year-old Tasmanian blue gum (Eucalyptus globulus) plantation in South Australia. Various climatic variables were logged automatically at one edge of the
plantation. The relationship between daily sap flow and sapwood area was strongly linear for the edge trees (r 2 = 0.97), but
only moderately correlated for the inner trees (r 2 = 0.46), suggesting an edge effect. For all trees, sap flow normalized to
sapwood area (Qs) increased with potential evaporation (PE)
initially and then became independent as PE increased further.
There was a fairly close correlation between transpiration of
the edge and inner trees, implying that water availability was
partially responsible for the difference between inner and edge
trees. However, the ratio of edge tree to inner tree transpiration
differed from unity, indicating differences in canopy conductance, which were estimated by an inverse form of the Penman-Monteith equation. When canopy conductances were less
than a critical value, there was a strong linear relationship between Qs of the edge and inner trees. When canopy conductances of the edge trees were greater than the critical value, the
slope of the linear relationship was steeper, indicating greater
transpiration of the edge trees compared with the inner trees.
This was interpreted as evidence for enhancement of transpiration of the edge trees by advection of wind energy.
Keywords: advection, canopy conductance, edge effect, heat
pulse velocity technique, interception belt.
Introduction
In Australia, trees have been widely advocated as a means to
alleviate dryland salinity (Morris and Thomson 1983,
Schofield 1991, Farrington and Salama 1996). A proposed
planting configuration to control rising groundwater tables associated with dryland salinity is an interception belt, i.e., a belt
of trees strategically placed in the landscape above the break
of slope where it may intercept potential recharge water and
discharge (by transpiration) shallow groundwater (Bulman et
al. 1993). The implementation of such tree planting schemes at
the farm level, however, is likely to be undertaken only if the
amount of land taken out of production and planted with trees
is minimal and there is potential for significant financial return
from the trees. In addition, the perceived hydrological benefits
of tree planting must be significant. The success of such
schemes, therefore, is dependent not only on the choice of appropriate species but also on the efficiency of water removal
per unit land area. Optimal design of interception belts is thus
critical to effective amelioration of saline lands.
We hypothesized that, in a belt or plantation of trees, transpiration rates will be greater for trees situated at the edge because of the effect of reduced competition from fewer
immediate neighbors and a potentially increased advective effect as a result of reduced sheltering from wind. It is reasonable to assume that edge trees will exhibit faster growth than
non-edge trees because they are subject to less competition for
light and soil resources, and that faster growth will be reflected
in greater leaf and sapwood areas. Both of these parameters
are linearly correlated with transpiration (Hatton et al. 1995)
and may be used as scaling or normalizing factors in the comparison of trees of different sizes, thus neutralizing differences
in water use related to differences in tree size.
If transpiration rates are normalized against either sapwood
area or leaf area, they may still be higher for edge trees because of differences in water availability or increased advection of wind energy. The magnitude of such effects has implications for the design of interception belts. Periods of reduced
transpiration caused by drought would indicate the need for a
reduced planting density to allow greater rooting volume per
tree. A major influence of advection on tree water use would
indicate the need for widely spaced, single rows, where available wind energy could be utilized most effectively by all
trees. Conversely, if the influence of advection on transpiration is low, this would indicate the desirability of denser,
404
TAYLOR, NUBERG AND HATTON
multi-row belts, to optimize water use and simplify plantation
design and management. A field-based experiment was conducted to determine if there is an effect of tree position within
a belt or plantation on transpiration rate, how much that effect
can be explained by advection at the plantation edge, and what
environmental conditions determine the magnitude of the effect.
used to monitor a range of parameters above the tree canopy.
These were air temperature and humidity, solar radiation,
wind speed and wind direction, all of which were averaged
and recorded automatically over 15-min intervals for the duration of the experiment with a Datataker 500 (Data Electronics
Pty. Ltd., Rowville, Australia). Rainfall was recorded at
ground level.
Materials and methods
Results and discussion
Site description
Transpiration measurement
A 7-year-old Tasmanian blue gum (Eucalyptus globulus spp.
globulus Labill.) plantation at Talinga, 35 km north of
Naracoorte in the south east of South Australia (36°96′ S,
140°74′ E), was selected for the experiment. The 2-ha plantation had originally been planted by the Forestry Division of the
former Primary Industries South Australia to assess the potential of the species for pulp wood production in a rainfall zone
of less than 600 mm. Mean annual rainfall for Naracoorte is
581 mm. The plantation is 200 m long by 100 m wide with the
long axis running north–south rising slightly at the northern
end onto an east–west sand dune. The land around the plantation is reasonably flat with a wheat crop on the western and
southern sides and sheep pasture to the east. The section of the
plantation used for the experiment is on a deep sand.
Trees were planted on a 3 × 3 m grid and had reached a mean
height of 14 m with a full canopy by the start of the experiment. Some gaps were apparent where trees had not survived
but all trees sampled within the plantation had a full complement of eight neighbors. The experiment was conducted at the
start of summer, over the period December 4, 1996 to January
3, 1997, at which time it was expected that transpiration would
be at a maximum as a result of high evaporative demand and a
plentiful soil water supply after winter.
During the first measurement period from December 5 to 16,
1996, transpiration rates of 14 trees were monitored. During
the second measurement period (December 18, 1996 to January 3, 1997) only nine trees were monitored, due to malfunction of five sap flow sensors.
For the first measurement period, the relationship between
daily sap flow and sapwood area was strongly linear for the
edge trees (r 2 = 0.97) but only moderately correlated for the
inner trees (r 2 = 0.46) (Figure 1). The difference in slope of the
two lines suggested a potential edge effect. It was not possible
to demonstrate a correlation for the edge trees in the second
measurement period, because the sensors on the two trees at
the eastern edge failed.
Measurement of climatic factors and calculation of potential
evaporation
Daily potential evaporation (PE), which was calculated from
the Priestley-Taylor equation as the sum of all daytime 15-min
means for each day, is shown in Figure 2. Net radiation was
calculated as the sum of measured solar radiation (reduced to
account for an estimated albedo value of 0.2) and net
long-wave radiation (atmospheric minus terrestrial, assuming
atmospheric and ground temperatures were equal to measured
Experimental procedure
Transpiration was determined for 16 trees by the heat pulse velocity technique (Marshall 1958, Swanson and Whitfield
1981) with sap flow sensors (Models SF100 and SF300,
Greenspan Technology, Queensland, Australia) logged at
30-min intervals over the 30-day period. Pairs of sensors were
installed at four depths in the sapwood to characterize the sap
velocity profile radially, and the time taken for the temperature
differential between each pair of sensors to return to zero following a heat pulse was recorded. Heat pulse velocities were
corrected for probe wounding effects and converted to sap velocities based on volumetric wood and water contents of a
cored sapwood sample. Transpiration rates were calculated by
integrating the sap velocity profile around the bole of the tree,
assuming a circular cross section (Hatton et al. 1992). Tree parameters used were circumference at breast height, bark thickness and depth to heartwood (determined from a core sample).
A transect of trees was sampled from the western to the eastern edge, with a bias toward the edges. A 15-m tower was
erected at the western edge (prevailing wind direction) to support meteorological equipment (Monitor Systems, Australia)
Figure 1. Relationship between mean daily water use (sap flow, Q)
and sapwood area of four edge trees (open symbols) and 10 inner trees
(solid symbols) for the period December 5–16, 1996. Correlation for
all trees (not shown) is high (r 2 = 0.92) and is also high (r 2 = 0.97) for
just the edge trees.
TREE PHYSIOLOGY VOLUME 21, 2001
ENHANCED TRANSPIRATION DUE TO EDGE EFFECTS
405
Figure 2. Daily potential evaporation during the measurement periods calculated by the PriestleyTaylor equation.
air temperature). A constant of 1.26 was applied. We note that
these values of PE were not intended to represent the maximum transpiration rate for the plantation, but merely to allow
comparison of transpiration rates among days (i.e., the relative
values of PE rather than the absolute values).
During the first measurement period (December 5–16), PE
varied widely from less than 3 mm on December 6 to more
than 7 mm on four days. In contrast, the second measurement
period (December 18 to January 2) was dominated by hotter,
drier conditions with PE in excess of 7 mm on 11 of the
17 days. No rain fell in the first period and only 2 mm fell toward the end of the second period.
Relationship between transpiration and potential evaporation
Relationships between normalized sap flow, Qs (Edwards et
al. 1997) and PE for both measurement periods are shown in
Figure 3. Because the relationships were nonlinear, second-order polynomials were used to indicate lines of best fit. In both
the edge and inner trees, Qs increased with PE initially and
then became independent as PE increased further. The plots
indicate that minor but significant differences in Qs between
inner and edge trees existed at low PE and these differences increased with increasing PE until Qs became independent of
PE.
To investigate these differences, values of Qs for edge and
inner trees were plotted against each other for each day (Figure 4). Edge effects were assumed to be negligible if the ratio
of the two flows was constant and equal to unity. A constant
ratio that was not equal to unity would imply an edge effect
that would most likely result from proportional differences in
canopy conductance between the edge trees and the inner trees
(indicating different water availabilities for example).
Figure 4 indicates a reasonable correlation between edge
and inner trees in both measurement periods (r 2 = 0.81, 0.82),
suggesting a high degree of proportionality. This implies that
water availability is partially responsible for the differences.
However, the slopes of the two plots differed markedly from
unity (1.64 and 1.52), indicating possible differences in canopy conductance.
Calculating actual and critical canopy conductances
Canopy conductance was estimated by an inverse form of the
Penman-Monteith equation (Granier et al. 1990) with the
evaporation term (E) replaced by tree transpiration per unit
ground area. Because the area of ground “occupied” by each
tree is not easily quantified, we assumed that the occupied area
was proportional to the cross-sectional sapwood area (sensu
Hatton and Wu 1995, who postulated that individual trees tend
toward an equilibrium between their size and domain). For
each of the inner trees, the occupied area was calculated by
multiplying the plantation grid spacing (9 m 2) by the ratio of
actual sapwood area to mean sapwood area of all the inner
trees. The area occupied by the edge trees is more uncertain
Figure 3. Relationships between normalized (per unit sapwood area)
daily sap flow (Qs) and potential evaporation for edge trees (䊏, 䊐)
and inner trees (䉱, 䉭) for the periods December 5–12, 1996 (filled
symbols) and December 17, 1996 to January 2, 1997 (open symbols).
TREE PHYSIOLOGY ONLINE at http://heronpublishing.com
406
TAYLOR, NUBERG AND HATTON
Figure 4. Relationship between normalized (per unit sapwood area) daily
sap flow (Qs) of edge trees and inner
trees for (a) Period 1 (December 5 to
17, 1996), where the line of best fit is y
= 1.64x (r 2 = 0.82) and (b) Period 2
(December 18, 1996 to January 2,
1997), where the line of best fit is y =
1.52x (r 2 = 0.81). Error bars represent
standard errors of mean.
because the plantation grid spacing does not apply. However,
an effective grid spacing can be calculated by comparing mean
sapwood area of the edge trees with that of the inner trees and
multiplying this ratio by the actual grid spacing. The ratio of
edge tree to inner tree sapwood area was approximately 2 (see
Table 1) so the effective grid spacing for the edge trees was assumed to be 18 m 2. The area occupied by each of the edge trees
was then calculated as described for the inner trees by multiplying the effective grid spacing by the ratio of actual sapwood
area to mean sapwood area of the edge trees.
Mean daily canopy conductance, gc (mm s –1), for each tree
was then calculated (i.e., the mean of 48 × 30 min periods per
day). Note that inverting the Penman-Monteith equation may
be achieved in several ways, but where vapor density deficit is
used, it can be simplified as:
gc =
E
,
c1 − c2 E
(1)
where
sϕ N
c1 =
+ ∆ρva ,
λγga
(2)
s+γ
,
γga
(3)
and
c2 =
where ϕN = net radiation, ∆ρva = vapor density difference between the vegetation and the air at some reference point, ga =
aerodynamic or boundary layer conductance between the vegetation surface and the bulk air stream, λ = latent heat of vaporization of water, γ = psychrometric parameter (= ρcp/λ,
where ρ = air density and cp = specific heat of air at constant
pressure) and s = rate of change of saturated vapor density with
temperature. The last three terms are all physical parameters
that are temperature dependent.
Transpiration may respond to increasing wind speed
(through its effect on aerodynamic or boundary layer conduc-
tance, ga) in one of three ways, depending on the relative
values of ga and canopy conductance. A critical value of canopy conductance exists for which leaf temperature Tl is equal
to air temperature Ta. When canopy conductance, gc, is above
the critical value, Tl < Ta and transpiration increases with wind
speed because Tl increases toward Ta. The opposite happens
when gc is below this value. If canopy conductance is equal to
the critical value, then transpiration is independent of wind
speed. By rearranging the Penman-Monteith equation and setting evaporation to be independent of aerodynamic conductance, the critical value of canopy conductance can be
calculated (Thornley and Johnson 1990, Equation 14.6e):
gc ( critical) =
sϕN
.
λ∆ρva ( s + γ )
(4)
Table 1. Sapwood and basal area of sampled trees. The ratio of mean
sapwood area of edge trees to mean sapwood area of inner trees was
approximately two (2.06).
Tree ID
Basal area
(cm 2)
Sapwood area
(cm 2)
W01A
W01B
E01A
E01B
Mean of edge trees
Standard error of edge trees
397
108
357
134
249
124
280
85
259
91
179
89
W02A
W02B
W05
W09
W13
W17
E02A
E02B
E05
E09
E12
E16
Mean of inner trees
Standard error of inner trees
142
100
110
161
137
144
134
89
92
103
121
168
125
31
97
65
78
103
92
103
91
65
67
74
90
117
87
22
TREE PHYSIOLOGY VOLUME 21, 2001
ENHANCED TRANSPIRATION DUE TO EDGE EFFECTS
By calculating mean daily daytime values for critical canopy conductance and comparing these values with mean daily
daytime values of canopy conductances calculated for individual trees, it was possible to identify days when transpiration is
likely to be enhanced and days when it is likely to be suppressed by increasing wind speed. Table 2 summarizes the
critical canopy conductances (mean of 96 × 15-min periods
per day) for each day and the calculated mean canopy conductances (mean of 48 × 30-min periods per day) for each tree.
The data in Table 2 indicate that enhancement of transpiration was most likely on December 15, when the critical canopy
conductance was exceeded by all edge trees and six of ten inner trees. Conversely, it can be seen that enhancement was unlikely on six days, when none of the canopy conductances
exceeded the critical value. Taking this further, it is possible to
rank the days in order of the likelihood of critical conductance
being exceeded. If it is measurable, the enhancement of transpiration of edge trees with respect to inner trees will be most
apparent on the days when there is the highest likelihood of the
critical conductance being exceeded. On the days when there
is less likelihood of the critical conductance being exceeded,
either there should be no edge effect apparent or, in extreme
conditions, transpiration of the edge trees may be suppressed
with respect to the inner trees.
Figure 5 shows that, for the four days when canopy conductances were least likely to exceed the critical value during Period 1, a strong linear relationship existed between Qs of the
four edge trees and of the four inner trees whose conductances
never exceeded the critical value. For the days when canopy
conductances of the edge trees were more likely to be greater
than the critical value, the relationship was still linear (although less strongly so), but the slope of the line was steeper
and approached the significance level (P = 0.08), indicating
greater transpiration of the edge trees with respect to the inner
trees. This is evidence for enhancement of transpiration of the
edge trees.
Because there was only one day when canopy conductance
407
Figure 5. Relationship between normalized (per unit sapwood area)
edge and inner tree sap flow (Qs) over 4 days for subcritical (䊐) and 8
days of supercritical (䊏) canopy conductance conditions.
exceeded the critical value during Period 2, we predicted that
the transpiration of the edge trees would not be enhanced with
respect to the inner trees. Figure 6 shows that, for the 6 days
with the highest likelihood of the critical conductance being
exceeded, the relationship between edge and inner tree Qs was
good (r 2 = 0.76), as it was for the 3 days of least likelihood (r 2
= 0.94), but there was no significant difference (P > 0.2),
which is in agreement with our hypothesis.
Implications for design of interception belts
Under the conditions of this experiment, edge trees transpire at
higher rates than inner trees as the difference in water availability increases, and this difference may be further enhanced
by advection of wind energy. However, it remains to be shown
that this relationship holds under the conditions that prevail
where interception belts are likely to be established. The plantation under study was on well-drained sands in a low rainfall
Table 2. Calculated mean daily canopy conductances (mm s –1) for all sampled trees in Period 1. Values in bold type indicate days when calculated
canopy conductances exceeded the critical value, which is shown in the right-hand column.
Date
Dec 5
Dec 6
Dec 7
Dec 8
Dec 9
Dec 10
Dec 11
Dec 12
Dec 13
Dec 14
Dec 15
Dec 16
Edge trees
Inner trees
Critical
gc
W01A W01B E01A
E01B
W02B W05
W09
W13
W17
E16
E12
E09
E02A
E02B
4.6
0.9
2.8
1.9
3.5
3.0
4.7
3.1
3.9
3.2
5.4
1.8
4.8
1.1
3.0
2.2
4.1
3.0
5.0
3.2
4.5
3.7
5.2
2.1
1.7
0.3
1.0
0.7
1.5
1.1
1.9
1.1
1.5
1.1
1.9
0.6
1.8
0.4
1.3
0.9
1.7
1.4
1.8
1.2
1.4
1.4
2.0
0.8
2.5
0.7
1.9
1.5
2.6
2.0
3.0
2.0
2.4
2.2
3.2
1.4
1.6
0.3
1.4
1.0
1.8
1.5
2.1
1.3
1.6
1.3
2.2
1.5
3.1
0.8
2.3
1.6
3.0
2.4
3.3
2.1
2.8
2.4
3.5
1.7
3.5
1.0
2.5
2.0
3.2
2.6
3.8
2.6
3.3
2.7
4.1
1.8
2.0
0.5
1.6
1.2
2.0
1.6
2.4
1.6
1.9
1.8
2.7
1.1
2.6
0.8
1.9
1.4
2.4
1.8
2.5
1.7
2.5
1.9
3.2
1.4
2.5
0.6
2.0
1.5
2.4
1.9
2.7
1.8
2.1
1.9
2.6
1.2
3.9
0.9
2.6
1.8
3.5
2.8
4.1
2.7
3.4
2.9
4.8
1.7
3.6
0.8
2.3
1.8
3.1
2.5
3.9
2.6
3.6
3.0
4.5
1.6
1.1
0.2
0.8
0.6
1.1
0.8
1.3
0.8
1.0
0.8
1.5
0.6
TREE PHYSIOLOGY ONLINE at http://heronpublishing.com
4.1
1.6
4.6
2.9
2.9
4.4
4.6
4.4
3.8
2.1
2.3
2.2
408
TAYLOR, NUBERG AND HATTON
transpiration between edge and inner trees were evident,
which is in agreement with our hypothesis.
Transferring the results of this study to areas with shallow
groundwater where interception belts are likely to be used
should not be problematic. If there is no risk of soil water
stress, canopy conductance will remain high and, as a consequence, the probability of transpiration enhancement by
advection of wind energy will also be high. We conclude that
the optimal design for interception belts favors wide tree spacing to maintain high canopy conductance, increase wind energy to the entire canopy and maximize transpiration of
groundwater.
Acknowledgments
Figure 6. Relationship between edge and inner tree normalized (per
unit sapwood area) sapflow (Qs) over 6 days for subcritical (䊐) and
3 days of supercritical (䊏) canopy conductance conditions. Error bars
are large and have been omitted. Differences were not significant (P >
0.2).
environment, whereas the notional interception belt will have
access to shallow groundwater and, in some cases, considerably higher rainfall.
Under well-watered conditions where interception belts are
likely to be used, the probability of canopy conductance exceeding the critical value is much higher than under the conditions of this experiment and therefore enhancement of
transpiration in response to wind advection is also likely to be
higher. This is because canopy conductance is sharply reduced
when plants are water stressed so that the likelihood of canopy
conductance exceeding the critical value is much reduced. In
interception belts, all trees should behave as edge trees, because increasing wind energy to the entire canopy of the belt is
desirable to maximize transpiration.
Conclusions
Plots of tree water use against sapwood area indicate that sapwood area is a suitable scaler for comparing water use between
trees (r 2 = 0.92). Edge trees showed higher normalized sap
flows (i.e., greater water use per unit sapwood area) compared
with inner trees (defined as trees in all rows other than the eastern and western edge rows), particularly as PE increased.
These differences in normalized water use between the edge
trees and the inner trees can be explained largely by differences in canopy conductance, probably because edge trees had
better access to water than inner trees.
In addition, wind effects significantly enhanced transpiration of edge trees compared with inner trees by about 10% on
days when canopy conductance was higher than a critical
value. It is possible that this figure would be higher for
well-watered conditions. The original assumption that the soil
profile was well-watered appears to have been invalid. On
days of low canopy conductance, no significant differences in
We thank Trevor Pridham, property owner, and Mick Underdown
and Ian Robertson of Primary Industries South Australia for allowing
us access to the experimental site; Kerryn McEwan, CSIRO Land and
Water, Adelaide, for assistance with field work; Kevin Harrison, Paul
Harris and Tim Ellis of The University of Adelaide for engineering
support in the design and construction of the weather tower. PJT acknowledges the Land and Water Resources Research and Development Corporation, Canberra, Australia, for providing funding for this
work in the form of a postgraduate scholarship.
References
Bulman, P., B. Cohen and A. Jensen. 1993. Revegetation principles
and options. Background paper prepared for the Upper South East
Dryland Salinity and Flood Management Plan Environmental Impact Statement. South Australian Dept. Environ. Land Manage.,
Adelaide, Australia, 6 p.
Edwards, W.R.N., P. Becker and J. Èermák. 1997. A unified nomenclature for sap flow measurements. Tree Physiol. 17:65–67.
Farrington, P. and R.B. Salama. 1996. Controlling dryland salinity by
planting trees in the best hydrogeological setting. Land Degrad.
Develop. 7:183–204.
Granier, A., V. Bobay, J.H.C. Gash, J. Gelpe, B. Saugier and W.J.
Shuttleworth. 1990. Vapor flux density and transpiration rate comparisons in a stand of maritime pine (Pinus pinaster Ait.) in Les
Landes forest (France). Agric. For. Meteorol. 51:309–320.
Hatton, T.J., D. Greenslade and W.R. Dawes. 1992. Integration of
sapflow velocity in elliptical stems. Tree Physiol. 11:185–196.
Hatton, T.J., S.J. Moore and P.H. Reece. 1995. Estimating stand transpiration in a Eucalyptus populnea woodland with the heat pulse
method—measurement errors and sampling strategies. Tree
Physiol. 15:219–227.
Hatton, T.J. and H.I. Wu. 1995. Scaling theory to extrapolate individual tree water use to stand water use. Hydrol. Proc. 9:527–540.
Marshall, D.C. 1958. Measurement of sap flow in conifers by heat
transport. Plant Physiol. 33:385–396.
Morris, J.D. and L.A.J. Thomson. 1983. The role of trees in dryland
salinity control. Proc. Roy. Soc. Victoria, 95:123–131.
Schofield, N.J. 1991. Tree planting for dryland salinity control in
Australia. Proc. Conf. The Role of Trees in Sustainable Agriculture, Albury, Victoria, pp 1–19.
Swanson, R.H. and D.W.A. Whitfield. 1981. A numerical analysis of
heat pulse velocity theory and practice. J. Exp. Bot. 32:221–239.
Thornley, J.H.M. and I.R. Johnson. 1990. Plant and crop modelling: a
mathematical approach to plant and crop physiology. Clarendon
Press, Oxford, 411 p.
TREE PHYSIOLOGY VOLUME 21, 2001