1. science
2. weather

Tree Physiology 24, 1035–1044
© 2004 Heron Publishing—Victoria, Canada
Water use by fast-growing Eucalyptus urophylla plantations in
southern China
JIM MORRIS,1,2 ZHANG NINGNAN,3 YANG ZENGJIANG,3 JOHN COLLOPY1 and XU
DAPING 3
1
Forest Science Centre, Heidelberg, Victoria 3084, Australia
2
Corresponding author ([email protected])
3
Research Institute for Tropical Forestry, Chinese Academy of Forestry, Guangzhou 510520, China
Received October 6, 2003; accepted January 24, 2004; published July 1, 2004
Keywords: canopy conductance, Guangdong, heat pulse, sap
flow, soil moisture, transpiration, vapor pressure deficit.
Introduction
Eucalyptus plantations in several countries have been the subject of criticism because of their high water use and other negative environmental impacts. Examination of the evidence for
these claims has usually concluded that well-managed plantations are beneficial rather than detrimental to the environment
(Poore and Fries 1985, White et al. 1995, Casson 1997). How-
ever, studies in India (Calder et al. 1992) and in South Africa
(Dye 1996) indicate that, when water resources are limited, the
area, location and management of plantations must be carefully considered to avoid conflict with other water users.
Detailed data on the water use and water balance of plantations are essential both to evaluate their environmental impacts
and to design optimal land use strategies in catchments where
wood production is an important economic component. Recently, research in several countries, including Australia (Honeysett et al. 1996, Myers et al. 1996, Morris and Collopy
1999), Brazil (Soares and Almeida 2001), Portugal (Osorio et
al. 1998), South Africa (Scott and Smith 1997, LeMaitre et
al. 2002), India (Kallarackal and Somen 1997) and Pakistan
(Mahmood et al. 2001), has increased our understanding of the
hydrology of eucalypt and other exotic tree plantations.
Plantations of eucalypt and other exotic tree species have
been extensively established in southern China during the past
50 years (Qi 2002). The total area of eucalypt plantations in
China was around 460,000 ha in 1993 (Bai and Gan 1996),
with a more recent estimate at 1.3 million ha (FAO 2001), increasing by 100,000 ha or more per year in response to government afforestation programs and cooperative joint ventures
between provincial forestry authorities and foreign investors.
In addition to their key role in timber and fiber production,
eucalypt plantations in China have an important ecological
role in the revegetation of degraded lands (Corlett 1999, Zhou
et al. 2002).
The Leizhou Peninsula of western Guangdong province
(Figure 1) is an example of a region where eucalypt plantations have become a significant component of the regional
economy. The forest estate of the region includes more than
200,000 ha of eucalypt plantations, from which more than
60% of the wood production is used as wood chips for export
and for the domestic paper industry. It has been suggested that
high water use by eucalypt plantations may contribute to the
depletion of groundwater resources on the peninsula. Mean
annual rainfall in the area is over 1300 mm, but the topography
of the region and the deep permeable soils limit opportunities
for storage of surface water, so farmers must often extract
groundwater during the dry season to establish or maintain
Downloaded from http://treephys.oxfordjournals.org/ by guest on August 29, 2014
Summary Tree growth, water use, climate and soil water
conditions were monitored over 12 months in two 3– 4-yearold Eucalyptus urophylla S.T. Blake plantations on the Leizhou Peninsula of southern China. The Hetou plantation was
established on a sandy soil of sedimentary origin with low water storage capacity, and the Jijia plantation was established on
a clay soil formed on basalt. Sapwood area was ~50% higher at
Jijia than at Hetou because of differences in plant spacing
(1994 versus 1356 stems ha –1). Annual water use, assessed by
heat pulse measurements, was 542 mm at Hetou and 559 mm at
Jijia, with mean sap flux densities of 2772 and 1839 l m –2
day –1, respectively. Limitations to water use, imposed by climatic and soil factors, were quantified by analysis of daily canopy conductance in relation to daytime vapor pressure deficit
(VPD) and soil water content. Similar annual water use at the
two sites was a result of higher VPD and soil water availability
at Hetou compensating for the greater sapwood area at Jijia.
Potential annual water use in the absence of soil water limitation was estimated at 916 mm at Jijia and 815 mm at Hetou.
Higher water availability during the dry season and early wet
season at Hetou than at Jijia was the result of deep root systems.
The results imply that water use by plantations on soils with
high water availability and in areas of high VPD may be reduced by establishment at wider spacing. The environmental
cost of water use by plantations must be weighed against their
economic and environmental values to determine an appropriate mix of forestry, agriculture and other land uses in regions
where water resources are limited.
1036
MORRIS ET AL.
Quaternary sediments. At both locations, the plantations are
monocultures of E. urophylla planted in mid-1996. Tree spacing is 3 × 2.5 m at Hetou and 3.3 × 1.5 m at Jijia. An understory
of native shrubs developed following the removal of all vegetation during planting preparations, but is patchy and relatively
light at Hetou. Slopes at both sites are < 1%, near the top of a
broad ridge defining the catchment boundary at Jijia (70 m
above sea level) and on a low divide between streams in a
broad valley at Hetou (25 m above sea level).
Weather monitoring
crops of sugar cane and rice. Water use by plantations has been
shown to affect water table depth in other countries, but little
research has been conducted on the hydrology of eucalypt
plantations in China, and there have been no previous studies
of plantation water balance on the Leizhou Peninsula. Zhou et
al. (1995) estimated transpiration by a eucalypt plantation in
western Guangdong at 614 mm year –1 based on an empirical
function of rainfall and potential evaporation, and suggested
that eucalypts are limited by water availability as a result of
low water storage in the root zone.
As part of an investigation of hydrological impacts of plantations on the Leizhou Peninsula, we studied the water use of
plantations of Eucalyptus and other species at several locations. This paper reports daily and annual water use by stands
of Eucalyptus urophylla S.T. Blake at two contrasting sites,
and examines the influence of climate and soil water as controlling factors for plantation water use in the region.
Methods
Study areas
Two sites were selected on forest farms near Hetou (21°05′ N,
109°54′ E) and Jijia (20°54′ N, 109°52′ E) in the Nandu River
catchment of the Leizhou Peninsula (Figure 1). The topography of the peninsula is flat to undulating, with deeply weathered red soils derived from basaltic and sedimentary parent
materials. The climate is tropical, with monthly mean temperatures of around 28 °C in July and 16 °C in January. Annual
rainfall varies from 1300 mm in the south to 1800 mm in the
north of the peninsula, with high year-to-year variability (standard deviation of 350 to 400 mm). Over 80% of the rain falls
between April and September, up to half of this in typhoons
that occur two to three times per year.
The Jijia site is on a basalt-derived clay soil, whereas the
Hetou site is about 40 km north on a sandy soil formed from
λE =
ε Rn + λρDq ga
ε +1 + ( g a / g c )
(1)
where ε is the dimensionless rate of change of saturated specific humidity with temperature (estimated as 2.2), Rn is net radiation, Dq is specific saturation deficit (calculated as 0.00622
× VPD), ga is aerodynamic conductance, gc is canopy conductance (assumed to be infinite for a wet canopy), λ is the latent
heat of evaporation of water and ρ is air density. Daily net radiation was estimated from solar radiation as: Rn = 0.8R s – 90 W
m –2 and aerodynamic conductance as 0.1 times wind speed
in m s –1 (Landsberg 1986). Although both of these estimates
are approximations of the true values of Rn and ga, neither variable is of critical importance to the calculations of this study
and so the approximations were considered acceptable. Mean
daytime values of VPD and potential E were calculated by averaging half-hourly data for each day when solar radiation was
greater than zero.
Soil conditions
To monitor soil water, a set of four standing wave dielectric
soil moisture probes with a waveguide length of 60 mm (MP406, Agri-Tech Instruments, Beijing, China) were buried at
each site, at depths of 50, 150, 250 and 350 cm below the soil
surface. Output from the probes was recorded at half-hourly
intervals by the weather station data loggers. Calibration to
volumetric water content (θ) was based on soil samples col-
TREE PHYSIOLOGY VOLUME 24, 2004
Downloaded from http://treephys.oxfordjournals.org/ by guest on August 29, 2014
Figure 1. Leizhou Peninsula, western Guangdong province, showing
the Nandu River system and location of the monitored plantations
near Jijia and Hetou.
A 0.1-mm tipping-bucket rain gauge and meteorological sensors (SC1 pyranometer, PTAT temperature sensor, Vaisala humidity sensor and cup anemometer from Tain Instruments,
Box Hill, Australia) were mounted on a mast 7.5 m above
ground level in a cleared area close to the plots monitored for
water use within each of the plantations in September 1999.
Half-hourly observations of solar radiation, rainfall, air temperature, relative humidity and wind speed were recorded by a
Tain Micropower data logger. Data were downloaded at approximately monthly intervals. Daily values of solar radiation
(Rs ), rainfall, maximum and minimum temperature and mean
wind speed were calculated, and daily mean vapor pressure
deficit (VPD; kPa) was derived from the half-hourly humidity
data, with saturation vapor pressure estimated as an exponential function of temperature (Dilley 1968). Potential (wet canopy) daily evaporation (E; mm day –1) was calculated based
on the Penman-Monteith combination equation, expressed by
Landsberg and Gower (1997) as:
EUCALYPTUS PLANTATION WATER USE IN CHINA
Tree growth and sapwood area
A 40 × 40 m monitoring plot was selected in each plantation in
September 1999. Within each monitoring plot, diameters at
breast height (DBH; cm, 1.3 m above ground level) were measured on all trees at intervals of about 6 months. For a small
number of multi-stemmed trees, an equivalent diameter was
calculated as the DBH of a single stem with the same basal
area as the combined stems. A regression with diameter (H =
1.5584DBH + 0.0257DBH 2, R 2 = 0.88) was derived with
heights (H; m) of a subset of 34 trees at each plantation, measured with a Vertex hypsometer (Forestor Instruments AB,
Sweden). This estimated the heights of all trees and inferred
mean dominant height (mean of the largest 200 trees ha –1), total volume (from single-tree volume approximated as DBH 2
× H/30,000 m3 ) and mean annual volume increment over the
life of the stand. Comparison with the results of growth studies
reported by Baker et al. (2003) suggests that the simple approximation adopted for single-tree volume yields estimates
within 5% of actual stem volume.
Sapwood areas of 40 trees including those selected for water
use measurements at each monitoring plot were calculated
from measurements of DBH, bark thickness and sapwood
width from increment cores taken at 2– 4 points around the
stem. The sapwood–heartwood boundary was identified by
color difference, light transmission or with the assistance of
0.1% aqueous methyl orange stain. Second-degree polynomial
regressions of sapwood area against DBH were derived to estimate the sapwood area of all trees in each plot at the times of
diameter measurement, and hence the total plot sapwood area.
Leaf area
Canopy leaf area index (LAI) was assessed at both sites in April and August 2000, with an Accupar ceptometer (Decagon
Devices, Pullman, WA) held ~1 m above ground. The mean
below-canopy photosynthetic photon flux (PPF) was calculated over 10 s at each point in a 5 × 5 m grid within each plot,
and expressed as a fraction of PPF recorded simultaneously by
a PR1 PPF sensor (Monitor Sensors, Caboolture, Australia)
located in an open area outside the plantation. We calculated
LAI from functions provided with the ceptometer, assuming a
leaf distribution parameter of 1 (approximately spherical canopy elements) and calculating zenith angle from latitude, longitude and time. The LAI values at all grid points were averaged to derive a single plot LAI for each site. Measurements at
Jijia, but not at Hetou, unavoidably included a small but unknown contribution from understory shrubs and small trees up
to 5 m in height.
Tree water use
Water use by E. urophylla plantations at Hetou and Jijia was
monitored from September 1999 to September 2000, for a total of 583 and 544 tree-days, respectively, by the heat pulse
method described by Edwards and Warwick (1984) and modified by Olbrich (1991). A set of 18 trees were selected for water use measurement at each site, matching the diameter distributions of the monitoring plots. From these 18 trees, two were
randomly selected per site at intervals of 3 to 4 weeks for implantation with teflon thermistor probes and stainless steel
heaters for heat pulse monitoring with HeatPulser instruments
(Edwards Industries, Otaki, New Zealand) and CR10X data
loggers (Campbell Scientific, Logan, UT).
Measurements of apparent heat pulse velocity were recorded at 30-min intervals with four pairs of thermistor probes
per tree and a 1-s heat pulse duration. The data were corrected
for the effects of wound diameter, probe separation and volume fractions of water and woody matrix in the sapwood to
derive estimates of sap flux density (SFD; cm 3, cm –2 h –1). A
whole-tree SFD estimate was derived from each set of four
point estimates by calculating a second-degree polynomial regression against implantation depth below the cambium, then
integrating this function around the stem. In some trees where
probe failures reduced the number of point measurements to
less than four, a weighted mean calculation similar to that of
Hatton (1990) was adopted. Instantaneous tree water use was
derived by multiplying SFD by sapwood area, estimated from
under bark diameter and sapwood widths measured at two to
four points around the stem. Total daily water use was obtained
by integration of the half-hourly values over each day, and dividing by the sapwood area to give a mean daily SFD in units
of dm 3 m –2 day –1.
Scaling of individual tree observations to stand water use
was based on sapwood area, as described by Khanzada et al.
(1998). The daily mean SFD values recorded for individual
trees were assumed to be unbiased estimates of the mean SFD
of the whole stand. This assumption is supported by the observation that differences in SFD between trees measured simultaneously or sequentially were usually small compared with
day-to-day SFD variation within a tree. Daily plot water use
was therefore estimated as the product of plot sapwood area
(interpolated for each day from the six monthly measurements, assuming continuous year-round growth) and mean
daily SFD of the monitored trees, and annual water use was
calculated as 365 times the mean of the daily values.
Canopy conductance
For evaporation from rough forest canopies, the radiation term
of Equation 1 is often small compared with the vapor deficit
TREE PHYSIOLOGY ONLINE at http://heronpublishing.com
Downloaded from http://treephys.oxfordjournals.org/ by guest on August 29, 2014
lected at the time of installation and 6 months later from auger
holes within 1 m of the probes. The volume of stored soil water
to 4 m depth was calculated by taking the water content at each
probe as an estimate of mean water content in a 1 m stratum.
Water release characteristics and particle size distribution of
the soils were assessed from soil samples collected at both
sites in September 1999 (end of the wet season) and April
2000 (end of the dry season). Matric potential (Ψm ) of samples
from 30-cm depth intervals to 390 cm was determined by the
filter paper method (Greacen et al. 1989), and gravimetric water content of the same samples was measured. Bulk density
was determined at 20-cm intervals to 80 cm as the mean of
three samples collected with 200-cm3 fixed-volume sampling
cylinders. Plant available water was determined as the difference between field water content and water content at wilting
point (estimated as –1500 kPa), based on a power curve fitted
to the water release data for each site (Ψm = –0.05θ –10.95 at
Jijia, Ψm = – 0.05θ –5.63 at Hetou).
1037
1038
MORRIS ET AL.
term, and aerodynamic conductance is large compared with
canopy conductance (Jarvis 1981). Under these conditions,
canopy conductance (gc; m s –1) may be estimated from observations of daily water use and daytime VPD as:
gc =
γ λT
ρ C pV
(2)
Figure 2. Monthly rainfall recorded at Jijia and Hetou between October 1, 1999 and September 30, 2000.
study sites was unaffected by this brief chilling period, although frost-sensitive trees were killed in plantations 100 km
north. Measurements of wind speed were affected by shelter
from the surrounding plantations as a result of the sub-canopy
height position of the anemometers, but maximum half-hour
means reached 2.5 m s –1 at Jijia and 1.8 m s –1 at Hetou. The
dry season was windier than the wet season at both sites. Daily
Results
Weather conditions
Rainfall in the 12 months from October 1, 1999 to September
30, 2000 was 1539 mm at Jijia and 1531 mm at Hetou.
Long-term meteorological records for the Leizhou Peninsula
indicate that both 1998 and 1999 were years of below average
rainfall (73.5 and 80.8% of the 21-year mean at Zhanjiang),
whereas in 2000, the rainfall returned to the 21-year mean.
There was a clear differentiation in monthly rainfall distribution between the wet and dry seasons (Figure 2). The wettest
days were June 26, 2000 at Hetou (163 mm) and June 27, 2000
at Jijia (128 mm), and the highest rainfall intensity recorded
was 48.6 mm in 30 min. Daily total solar radiation at Jijia is
shown in Figure 3A. Radiation at Hetou followed a similar
pattern to that at Jijia, but was slightly less, averaging 94% of
the Jijia daily values. Overcast skies are common at all times
of the year in this region of the Leizhou Peninsula. Mean daytime VPD (Figure 3B) rarely exceeded 2.5 kPa, and was less
than 1.5 kPa throughout the dry season from December to
March. The VPD was 17% higher on average at Hetou than at
Jijia, as a result of temperature and relative humidity differences. Throughout the year, monthly means of daily maximum
temperature were 1 to 2 °C higher at Hetou than at Jijia, reaching a maximum of 36.8 °C at Hetou and 34.6 °C at Jijia in June.
Daily minimum temperatures rarely fell below 7 °C, except
during the week of December 21–28 when temperatures of
zero or below were recorded at both sites. Vegetation at the
Figure 3. Climate and soil conditions at Jijia and Hetou, October
1999–September 2000. (A) Daily solar radiation; (B) daily mean vapor pressure deficit during daylight hours; and (C) available soil water
(0 to 4 m depth), estimated as the difference between observed field
water content and water content at –1500 kPa.
TREE PHYSIOLOGY VOLUME 24, 2004
Downloaded from http://treephys.oxfordjournals.org/ by guest on August 29, 2014
where T is mean rate (mm s –1) of plot water use for the day, and
V (kPa) is mean daily daytime vapor pressure deficit. This formula was applied to estimate gc for the plantations at Hetou
and Jijia on each day when the vapor deficit term in the numerator of Equation 1 exceeded the radiation term by a factor of
five or more. The psychrometric constant (γ), latent heat of
evaporation of water (λ), density of air (ρ) and specific heat of
air at constant pressure (Cp ) were estimated as 66 Pa K –1,
2450 kJ kg –1, 1.2 kg m –3, and 1010 J kg –1 K –1, respectively. To
calculate T, the estimated daily stand water use was divided by
day length, calculated from latitude and date following Sellers
(1965). To test the assumption that aerodynamic conductance
exceeded canopy conductance sufficiently to make Equation 2
valid, the calculated daily values of gc were sorted into five
classes of mean daily wind speed, and plotted against V. There
was no evidence of a difference in the relationship of gc and V
between days of high and low wind speed, confirming that
aerodynamic conductance was high enough to permit use of
Equation 2 on a daily time scale.
EUCALYPTUS PLANTATION WATER USE IN CHINA
1039
potential evaporation calculated from the meteorological data
was between 0–15 mm day –1 during the dry season and 5–
25 mm day –1 during the wet season, with high short-term variation. Daily potential evaporation was on average 16% higher
at Hetou than at Jijia during both dry and wet seasons.
Table 1. Soil properties at Hetou and Jijia. Bulk density was averaged
over 20 – 80 cm depth; water content at –1500 kPa (WC) was estimated from a water release curve derived from matric potential and
volumetric water content measurements on 104 soil samples from 0 –
3.9 m depth.
Soil water content
Location
Bulk density
Water content
at saturation (%)
Estimated WC
at –1500 kPa (%)
Hetou
Jijia
1.64
1.01
38
62
16
38
Tree growth
Tree density at Hetou was approximately 70% of that at Jijia.
Plot measurements in September 1999, March–May 2000 and
August 2000 indicated that both stands grew rapidly during the
study, with an increase in basal area of 30% at Hetou and 19%
at Jijia for the study year despite declining stand density as a
result of mortality of suppressed trees (Table 2). Mean tree diameter and height were slightly greater at Hetou, but the
higher stocking at Jijia led to greater sapwood area and volume
growth. Leaf area index was relatively low at both sites, reflecting the open canopy structure; LAI increased during the
wet season, although at Jijia the measured increase was partly
a result of understory development during the period.
Polynomial regressions derived to relate sapwood area (SA;
cm2 ) to overbark diameter (DBH cm) of a sub-set of 40 trees
ranging in DBH from 4 to 16 cm were: SA = 0.359DBH2 +
Figure 4. Variation in soil particle size distribution with depth at Hetou and Jijia.
0.537DBH (R 2 = 0.96) at Hetou and SA = 0.331DBH2 +
1.298DBH (R 2 = 0.92) at Jijia. These site-specific relationships were used to calculate the stand sapwood areas in Table 2
and for scaling up single-tree water use measurements to a
stand basis, although the regression lines did not differ significantly within the DBH ranges at Jijia and Hetou.
Tree water use
Single-tree water use averaged 13.9 dm 3 day –1 at Hetou with a
maximum of 49 dm 3 day –1 compared with 8.6 dm 3 day –1 and a
maximum of 27 dm 3 day –1 at Jijia. Two small suppressed trees
at Hetou (DBH = 6.2 and 7.3 cm) were excluded from the calculation of stand water use because their sap fluxes were anomalously low compared with other trees measured during the
same period, suggesting that their canopy exposure and access
to soil water were atypical of the stand as a whole and that they
are more appropriately considered as part of the understory.
Trees of this size or smaller contributed about 10% of the sapwood area at both sites, but not all are necessarily suppressed.
Daily sap flux density (SFD) and daily stand water use recorded at the two sites are shown in Figure 5. Estimates of annual stand water use based on the daily mean values at Hetou
and Jijia (Table 3) were 542 and 559 mm, respectively, equivalent to 35.4 and 36.3% of rainfall for the period, respectively.
Wet season (April–September) water use was 59.6% of the annual total at Hetou, but only 53.9% at Jijia, where water use
was greater than at Hetou in the dry season (1.39 versus
1.25 mm day –1, averaged over 137 and 172 days, respectively)
and less than at Hetou in the wet season (1.62 versus 1.85 mm
day –1, averaged over 112 and 101 days, respectively).
Water-use efficiency, defined as total stem volume growth
per unit of water transpired, was 4.8 × 10 –3 ± 0.2 × 10 –3 m3 m –3
at Hetou and 4.2 × 10 –3 ± 0.2 × 10 –3 m3 m –3 at Jijia (P > 0.95).
Despite lower planting density and hence lower stand volume
at Hetou, volume growth at Hetou in 1999–2000 exceeded that
of the higher density Jijia plantation (stand volume increased
by 23.8 and 21.6 m3 ha –1, respectively, between September
1999 and August 2000), leading to a higher water-use efficiency at Hetou during the study period.
Despite the similar annual and shorter-term water use estimates for Hetou and Jijia in Figure 5B, SFD differed between
plantations, with an annual mean daily value about 51% higher
at Hetou than at Jijia (Table 3). Sapwood area averaged 53%
more at Jijia than at Hetou during the study year. Because water use is the product of SFD and sapwood area, the similarity
of these ratios suggests that SFD varies between stands to
TREE PHYSIOLOGY ONLINE at http://heronpublishing.com
Downloaded from http://treephys.oxfordjournals.org/ by guest on August 29, 2014
Soil texture varied with depth at both sites (Figure 4). The
coarser textured soil at Hetou had markedly higher bulk density and lower water-holding capacity than the clay-dominated
soil at Jijia (Table 1). Half-hourly data from the buried soil
moisture sensors showed extraction of soil water to at least
350 cm depth from September 1999 until the end of the dry
season in early May 2000. Rain events during that period usually had no measurable effect on soil water content below
50 cm at Hetou or below 150 cm at Jijia. Replenishment of water at 350-cm depth occurred when about 200 mm of rain fell at
both sites over four days in late June. Infiltration through the
profile was notably more rapid at Hetou than at Jijia, consistent with the greater storage capacity of the clay soil at Jijia.
Maximum available soil water in the upper 4 m of the profile
was estimated to reach 450 mm at Jijia and 200 mm at Hetou,
and was depleted at both sites to around 150 mm by the end of
the dry season (Figure 3C).
1040
MORRIS ET AL.
Table 2. Growth characteristics of 1600 m2 plots of 3–4-year-old E. urophylla plantations assessed at Hetou and Jijia in 1999–2000.
Hetou
Number of trees
Stocking (stems ha –1)
Mean diameter (cm) ± SE
Basal area (m2 ha –1)
Mean dominant height (m)
± SE (n = 32)
Mean annual volume increment
(m3 ha –1 year –1)
Sapwood area (m2 ha –1)
Leaf area index ± SE (n = 64)
Jijia
September
1999
March
2000
August
2000
September
1999
May
2000
August
2000
217
1356
8.5 ± 0.2
9.0
16.7 ± 0.2
213
1331
9.5 ± 0.3
11.0
18.0 ± 0.2
208
1300
9.9 ± 0.3
11.7
18.5 ± 0.2
319
1994
8.5 ± 0.2
12.9
16.3 ± 0.2
311
1944
9.3 ± 0.2
15.1
17.5 ± 0.2
303
1894
9.5 ± 0.2
15.3
17.8 ± 0.2
16.3
17.9
18.3
22.6
24.7
22.6
4.74
5.69
1.14 ± 0.03
6.03
1.53 ± 0.12
7.65
8.71
1.70 ± 0.08
8.77
2.98 ± 0.11
Canopy conductance
Evaporation from a forest canopy is driven by the same factors
that drive evaporation from any wet surface (cf. Equation 1).
Thus, low net radiation and VPD may limit plantation water
use, and the response to these factors will be tempered by aerodynamic and canopy conductance. We found a close relationship between observed monthly water use and potential evaporation from a wet surface calculated from Equation 1 (Figure 6).
The ratio of the vapor deficit term to the radiation term in
Equation 1 was greater than five on 76% of days at Jijia and
92% of days at Hetou, with mean values of 8.4 and 14.3, respectively. Although these values are subject to errors in the
approximation of net radiation from solar radiation, and of
aerodynamic conductance from wind speed at less than canopy height, they confirm the common observation (Jarvis
1981) that evaporation from tall rough canopies is usually
dominated by atmospheric vapor deficit. Under such conditions, Equation 2 provides a valid estimate of canopy conductance. Values of gc calculated for all days when the vapor
deficit to radiation term ratio was greater than five are plotted
against mean daytime VPD in Figure 7. The value of five, chosen as a cutoff point for data filtering, is entirely arbitrary and
is clearly affected by the approximations adopted for net radiation and aerodynamic conductance. However, a comparison of
gc data calculated for all days with filtered data sets based on
cutoff values of three to 10 for the ratio of vapor deficit to radiation terms indicated that a factor of five was effective in removing days when VPD was low enough or radiation high
enough to possibly invalidate Equation 2, while retaining a
Table 3. Sap flux density and water use by E. urophylla plantations
monitored at Hetou and Jijia between September 1999 and September
2000.
Figure 5. Sap flux density (A) and daily water use (B) of E. urophylla
at Jijia and Hetou, September 1999 –September 2000.
Hetou
Jijia
Days monitored
291
266
Mean sap flux density
(dm3 m –2 day –1) ± SE
2772 ± 66
1839 ± 50
Mean daily water use (mm) ± SE
1.49 ± 0.04
1.53 ± 0.04
Annual water use
542
559
TREE PHYSIOLOGY VOLUME 24, 2004
Downloaded from http://treephys.oxfordjournals.org/ by guest on August 29, 2014
compensate for differences in stand sapwood area, such that
variation in stand water use is minimized within a region of
generally similar environmental conditions. This would further imply that soil and climate factors may wholly determine
stand water use, independently of tree density and stand sapwood area, over the range of the plantations studied. However,
the similarity of water use at Jijia and Hetou despite their different sapwood areas might be coincidental, resulting from
differing responses to environmental conditions at the two
sites. It is therefore pertinent to examine the extent to which
canopy conductance and hence water use at the Leizhou Peninsula sites are limited by evaporative demand and soil water
availability.
EUCALYPTUS PLANTATION WATER USE IN CHINA
data set containing a broad range of daily climate conditions
for analysis.
The daily gc data in Figure 7 show the expected decline in
canopy conductance with increasing VPD, resulting from the
stomatal response to atmospheric humidity (Monteith 1995).
The envelope curves in Figure 7, fitted to define the approximate upper limit to observed gc at a given VPD, are described
by the functions gc = 0.87(VPD + 0.12) –0.86 at Jijia and gc =
0.75(VPD + 0.12) –0.86 at Hetou. These curves were derived by
assuming functions of the form gc = a (b + VPD)c with a maximum conductance of 5 cm s –1 at VPD = 0 kPa, inserting the
values of gc and VPD for two points at the upper limit of the
observed data and solving the resulting simultaneous equations for a, b and c. Schäfer et al. (2000) proposed an alternative method for fitting envelope curves to similar data sets. The
relatively small (16%) difference in VPD response for the two
sites is magnified by the higher VPD at Hetou. The tendency
for higher VPD at Hetou to lead to higher water use is there-
Figure 7. Daily canopy conductance of E. urophylla at Hetou and Jijia
plotted against mean daytime vapor pressure deficit (VPD) on days
when the ratio of the vapor pressure deficit to radiation terms in Equation 1 was greater than five. Envelope curves are fitted to define the
approximate upper limit of observations at each site.
fore partly compensated for by reduced stomatal conductance.
Over the year, the combination of higher sapwood area, higher
stomatal conductance and possibly other factors, led to a mean
gc at Jijia that was 1.34 times that at Hetou.
If it is assumed that the envelope curves of Figure 7 represent the maximum possible gc at each site under conditions
where transpiration is limited only by VPD, then we can calculate the corresponding daily potential water use of the plantations by inverting Equation 2. Figure 8 shows the resulting
curves and mean rates of water use observed on days with vapor deficit to radiation term ratios greater than five. Weekly
mean values of potential water use are compared with potential evaporation in Figure 9A. Potential water use was usually
10 to 30% of potential evaporation during the wet season, rising to 20 to 50% in the dry season at both sites, and was
slightly higher at Jijia than at Hetou all year round. The observed daily rates of water use were 60 to 100% of potential
rates during the wet season (Figure 9B), falling during December and January to less than 40%. Although limited data are
available for the late dry season, it appears that the recovery of
water use to greater than 60% of potential rates occurred during March at Hetou but not until late May at Jijia.
Potential water use for the year (calculated as 365 times the
daily mean) was 815 mm at Hetou and 916 mm at Jijia. The reduction in estimated actual annual water use below its potential value was therefore 273 mm (33%) at Hetou and 357 mm
(39%) at Jijia. The observed difference in soil water availability (Figure 3C) is the most obvious cause of the variations with
time and between sites displayed in Figure 9B. To test this hypothesis, the residuals of daily values of gc below the envelope
curves of Figure 7 were expressed as a percentage of the potential (envelope) value and compared with soil water content
and matric potential at 50-cm depth by regression. Regressions with total available water to 400 cm were also calculated.
Highly significant relationships (P < 0.0001) were found for
both sites, indicating that low values of gc compared with the
envelope value were associated with periods of low soil water
content. The nature of these relationships differed between
sites.
At Jijia, canopy conductance was most strongly correlated
with total available soil water (Figure 10A, r 2 = 0.50 for
Figure 8. Observed mean daily water use and calculated potential water use at Hetou and Jijia corresponding to the canopy conductance
observations and envelope functions in Figure 7.
TREE PHYSIOLOGY ONLINE at http://heronpublishing.com
Downloaded from http://treephys.oxfordjournals.org/ by guest on August 29, 2014
Figure 6. Monthly values of mean daily water use by E. urophylla at
Hetou and Jijia, and mean daily potential evaporation from a wet canopy. Monthly water use at Jijia in April and at Hetou in July were estimated by regression from daily observations at the other site, because
insufficient observations were recorded in these months to calculate a
reliable monthly mean.
1041
1042
MORRIS ET AL.
Figure 9. (A) Weekly means of potential daily water use by E. urophylla plantations at Hetou and Jijia as a percentage of daily potential
evaporation. (B) Weekly means of observed daily water use as a percentage of potential water use.
242 days from September to early May), whereas at Hetou,
canopy conductance was most strongly correlated with matric
potential at 50 cm, and the correlation was significantly increased by a time lag of 2 days between the matric potential
and conductance observations (r 2 = 0.22 for 242 days from
September to early May). Yin et al. (2003) also reported that
correlations between SFD and soil water content were strongest in the lower profile at Jijia but in the upper profile at
Hetou. Correlations calculated over the full period of measurement (382 days) that included the wet season of May–September 2000 were stronger at Hetou (Figure 10B, r 2 = 0.29) but
weaker at Jijia (r 2 = 0.21). These results suggest that soil water
was limiting at Hetou during both wet and dry seasons, but
only during the dry season at Jijia. The 2-day lag between canopy conductance and matric potential at 50 cm at Hetou suggests that the major part of the root system was deep in the profile; a comparison of soil moisture sensor responses to individual rain events shows that infiltrating water reached 150 cm
one day after reaching 50 cm, and 250 cm after a further
2 days, suggesting a major root presence at around 200 cm
depth. At Jijia, good correlations with both total available soil
water and soil water content at 50 cm indicate that the upper
limit of the zone of root water uptake was shallower than at
Hetou, consistent with the greater soil water storage capacity
of the clay soil.
Discussion
Automated data collection at the study sites produced a detailed data set for relating plantation water use to climate and
soil conditions in remote locations on the Leizhou Peninsula.
Although measurements of net radiation and above-canopy
wind speed would have improved the precision of daily potential evaporation estimates, the finding that vapor deficit exceeds radiation by a factor of five or more on most days would
be expected to remain true, so that the relationship between
plantation water use and climate is adequately described by the
relationship with VPD in Equation 2. The heat pulse equipment and techniques adopted for water use measurement were
a reliable and convenient means of determining SFD in the
range from about 1.5 to over 80 ml cm –2 h –1. The 1127 treedays of water use observations recorded over 12 months at the
two sites are expected to provide a satisfactory estimate of
mean daily water use and hence seasonal and annual water use.
Annual water use at the sites was similar, despite a 53%
greater sapwood area at Jijia than at Hetou. The interpretation
that SFD varies to compensate for differences in sapwood
area, such that variation in stand water use is minimized conflicts with the hypothesis of constant leaf water efficiency advanced by Hatton et al. (1998) and confirmed in terms of SFD
by Khanzada et al. (1998), Benyon et al. (1999) and other studies. Furthermore, if soil and climate determine annual evapotranspiration largely independently of vegetation, then the
denser understory and higher soil evaporation at Jijia (Lane et
al. 2003) would be expected to reduce overstory water use relative to that at Hetou, rather than only reducing SFD. We
therefore conclude that the similarity in water use between
stands on an annual scale is coincidental, arising from soil and
climate differences at the two sites that have compensated for
the difference in sapwood area of the plantations studied, and
that the observed data do not support the hypothesis of SFD
varying to maintain constant stand water use across the region.
The greater sapwood area at Jijia would be expected to increase water use by increasing the conductance of the stand as
a conduit for water from soil to atmosphere. The difference in
conductance, and hence water use, would be less than 50%,
TREE PHYSIOLOGY VOLUME 24, 2004
Downloaded from http://treephys.oxfordjournals.org/ by guest on August 29, 2014
Figure 10. (A) Residual canopy conductance (gc; %) at Jijia in relation
to total available soil water to 4 m depth (September 1999–May
2000). (B) Residual gc (%) at Hetou in relation to soil matric potential
at 50 cm depth, with a 2- day lag time (September 1999–September
2000).
EUCALYPTUS PLANTATION WATER USE IN CHINA
tations studied. This environmental cost must be weighed
against the acknowledged non-wood benefits of the plantations to agriculture including wind shelter, control of erosion
and soil loss, and contribution to atmospheric humidity (Zhou
et al. 2002). Our results provide some of the information
needed for making rational decisions about the management
of the tree growing areas of the Leizhou Peninsula and nearby
provinces of southern China.
Acknowledgments
The study was funded by the Australian Centre for International Agricultural Research (ACIAR) as part of project FST 97/77, Eucalypts
and Water. We acknowledge the support of the Forest Science Centre,
Research Institute of Tropical Forestry (RITF), China Eucalypt Research Centre (CERC) and South China Institute of Botany (SCIB) in
making staff and other resources available for the project. We are
grateful to the management and staff of Leizhou Forest Bureau, the
Tangjia, Jijia and Hetou forest farms, site watchmen and their families, and other members of the Chiniwei and Tiantou workgroups who
made it possible to carry out the monitoring reported here over a protracted period in remote locations. Finally, we thank the staff and
students of RITF, CERC and SCIB, especially Zhou Guangyi, Wu
Zhongmin, Zhou Tao, Hua Lizhong, Chen Shaoxiong, Zhang Kejian,
Zhou Guoyi, Li Zhi’an, Shen Weijun, Chu Guowei and Huang Zhihong, for assistance with data collection, establishment and maintenance of the field sites.
References
Bai, J.Y. and S.M. Gan. 1996. Eucalyptus plantations in China. In Reports Submitted to the Regional Expert Consultation on Eucalyptus, October 1993. Eds. M. Kashio and K. White. RAPA Publication 1996/44, FAO Regional Office for Asia and the Pacific, Bangkok, Thailand, pp 23–32.
Baker, T., J. Morris, M. Duncan, N.N. Zhang, Z.J. Yang, Z.H. Huang
and G.W. Chu. 2003. Tree and stand growth and biomass relationships for Eucalyptus urophylla and E. 12ABL on the Leizhou Peninsula, Guangdong province, China. In Eucalypts in Asia. Ed.
J. Turnbull. ACIAR Proc. No. 111, ACIAR, Canberra, pp 174–182.
Benyon, R.G., N.E. Marcar, D.F. Crawford and A.T. Nicholson. 1999.
Growth and water use of Eucalyptus camaldulensis and E. occidentalis on a saline discharge site near Wellington, NSW, Australia. Agric. Water Manage. 39:229–244.
Calder, I.R., R.L. Hall and P.G. Adlard. 1992. Growth and water use
of forest plantations. John Wiley and Sons, Chichester, 381 p.
Casson, A. 1997. The controversy surrounding eucalypts in social forestry programs of Asia. Economics Division Working Paper 97/1,
National Centre for Development Studies, Australian National
University, Canberra, 46 p.
Corlett, R.T. 1999. Environmental forestry in Hong Kong: 1871–
1977. For. Ecol. Manage. 16:93–105.
Dilley, A.C. 1968. On the computer calculation of vapour pressure
and specific humidity gradients from psychrometric data. J. Appl.
Meteorol. 7:717–719.
Dye, P.J. 1996. Climate, forest and streamflow relationships in South
African afforested catchments. Comm. For. Rev. 75:31–38.
Edwards, W.R.N. and N.W.M. Warwick. 1984. Transpiration from a
kiwi fruit vine as estimated by the heat pulse technique and the Penman-Monteith equation. N.Z. J. Agric. Res. 27:537–543.
FAO. 2001. Global forest resources assessment 2000: main report.
FAO Forestry Paper 140, United Nations Food and Agriculture Organization, Rome, Italy, 511 p.
TREE PHYSIOLOGY ONLINE at http://heronpublishing.com
Downloaded from http://treephys.oxfordjournals.org/ by guest on August 29, 2014
because soil, root and canopy components also influence total
conductance. Higher VPD at Hetou than at Jijia, particularly
during the wet season, would offset this difference by increasing water use, even though canopy conductance would have
been lower. In the range of VPDs observed at our sites (usually
averaging < 2 kPa during daylight hours), there was no evidence of a reduction in water use as a result of stomatal response to low humidity as observed in more arid conditions
(Monteith 1995). The net result of these opposing factors is
that the potential water use was estimated as 916 and 815 mm
year –1 for Jijia and Hetou, respectively.
The effect of limited soil water availability on water use was
greater at Jijia than at Hetou, because of a shallower root system and hence greater dependence on water in the upper soil
profile. The reduction in actual water use below its potential
value amounted to 33% at Hetou and 39% at Jijia, bringing annual water use at both sites to about 550 mm. Statistical analysis of deviations from maximum conductance in relation to
soil water availability provided evidence of a strong relationship at both sites. Direct measurements of root density variation with depth in the profile would be required to confirm the
interpretation that a deeper root zone at Hetou alleviated the
reduction in water use caused by limited soil water availability
compared with at Jijia. Water balance studies reported by Lane
et al. (2003) found evidence of dry season water uptake from
below 4 m depth at Hetou but not at Jijia. Recovery of water
use after the dry season in 2000 was later at Jijia than at Hetou,
implying a need to replenish the soil water content to some
depth before a measurable increase in water use was achieved;
lower storage capacity of the sandy soil at Hetou than at Jijia
allowed this to occur several weeks earlier, despite the greater
root system depth of trees at Hetou.
The analysis of potential water use at the study sites allows
some qualitative predictions of water use of plantations under
other conditions. If the Hetou plantation had the same soil conditions as Jijia, we may infer that its annual water use would be
less than 550 mm because of reduced water availability in the
dry season, and because the positive influence of higher VPD
would still only partly offset the lower sapwood area. On the
other hand, if the Hetou plantation was established at the same
density as Jijia, its water use would be greater than 550 mm because of the combined effects of higher VPD and higher root
zone soil water availability during the dry season than Jijia.
For plantation managers seeking to minimize water use, our
results indicate that one solution is to plant trees at lower density on sandy soils and in areas with warmer, less humid climates.
The possibility that eucalypt plantation water use may impact excessively on groundwater resources on the Leizhou
Peninsula remains unresolved. Plantations in the region are
usually established on higher ground, with agriculture in the
valleys and on lower slopes; hence in most locations there may
be limited opportunity for direct uptake of shallow groundwater. Reduction of groundwater recharge is more likely to have
an impact when plantations deplete soil water stored in the
root zone during the dry season and extract some of the infiltrating rain water during the wet season. Our results suggest
that these processes total around 550 mm year –1 in the plan-
1043
1044
MORRIS ET AL.
Myers, B.J., S. Theiveyanathan, N.D. O’Brien and W.J. Bond. 1996.
Growth and water use of Eucalyptus grandis and Pinus radiata
plantations irrigated with effluent. Tree Physiol. 16:211–219.
Olbrich, B.W. 1991 The verification of the heat pulse velocity technique for estimating sap flow in Eucalyptus grandis. Can. J. For.
Res. 21:836–841.
Osorio, J., M.L Osorio, M.M. Chaves and J.S. Pereira. 1998. Water
deficits are more important in delaying growth than in changing
patterns of carbon allocation in Eucalyptus globulus. Tree Physiol.
18:363–373.
Poore, M.E.D. and C. Fries. 1985. The ecological effects of Eucalyptus. FAO Forestry Paper No. 59, Food and Agriculture Organization, Rome, Italy, 87 p.
Qi, S.X. 2002. Eucalyptus in China. 2nd Edn. Chinese Forestry Press,
Beijing, 517 p.
Schäfer, K.V.R., R. Oren and J.D. Tenhunen, 2000. The effect of tree
height on crown level stomatal conductance. Plant Cell Environ.
23:365–375.
Scott, D.F and R.E. Smith. 1997. Preliminary empirical models to
predict reductions in total and low flows resulting from afforestation. Water S.A. 23:135–140.
Sellers, W.D. 1965. Physical climatology. University of Chicago
Press, Chicago, 272 p.
Soares, J.V. and A.C. Almeida. 2001. Modeling the water balance and
soil water fluxes in a fast growing Eucalyptus plantation in Brazil.
J. Hydrol. 253:130–147.
White, K., J. Ball and M. Kashio. 1995. Proceedings of the regional
expert consultation on Eucalyptus, October 1993. RAPA Publication 1995/6, FAO Regional Office for Asia and the Pacific, Bangkok, Thailand, 196 p.
Yin, G.C., J. Morris, G.Y. Zhou, X. Wang, Z.H. Huang and N.N.
Zhang. 2003. Relating sap flow to environment in two Eucalyptus
urophylla plantations in southeastern China. In Eucalypts in Asia.
Ed. J. Turnbull. ACIAR Proc. No. 111, ACIAR, Canberra, Australia, pp 209–216.
Zhou, G.Y., Z.Y. Yu, S.L. Peng and X.A. Cai. 1995. A comparative
study on evapotranspiration of the three ecosystems in the Xiaoliang experimental station. Acta Ecol. Sinica 15:230 –236.
Zhou, G.Y., J.D. Morris, J.H. Yan, Z.Y. Yu and S.L. Peng. 2002. Hydrological impacts of reafforestation with eucalypts and indigenous species: a case study in southern China. For. Ecol. Manage.
167:209–222.
TREE PHYSIOLOGY VOLUME 24, 2004
Downloaded from http://treephys.oxfordjournals.org/ by guest on August 29, 2014
Greacen, E.L., G.R. Walker and P.G. Cook. 1989. Procedure for the
filter paper method of measuring soil water suction. CSIRO Division of Soils, Report No. 108. CSIRO, Collingwood, Australia, 7 p.
Hatton, T.J., E.A. Catchpole and R.A. Vertessy. 1990. Integration of
sapflow velocity to estimate plant water use. Tree Physiol. 6:
201–209.
Hatton, T., P. Reece, P. Taylor and K. McEwan. 1998. Does leaf water
efficiency vary among eucalypts in water-limited environments?
Tree Physiol. 18:529–536.
Honeysett, J.L., D.A. White, D. Worledge and C.L. Beadle. 1996.
Growth and water use of Eucalyptus globulus and E. nitens in irrigated and rainfed plantations. Aust. For. 59:64–73.
Jarvis, P.G. 1981. Stomatal conductance, gaseous exchange and transpiration. In Plants and Their Atmospheric Environment. Eds.
J. Grace, E.D. Ford and P.G. Jarvis. 21st Symposium, British
Ecological Society, Blackwell Scientific Publications, Oxford, pp
175–204.
Kallarackal, J. and C.K. Somen. 1997. Water use by Eucalyptus tereticornis stands of differing density in Southern India. Tree Physiol.
17:195–203.
Khanzada, A.N., J.D. Morris, R. Ansari, P.G. Slavich and J.J. Collopy. 1998. Groundwater uptake and sustainability of Acacia and
Prosopis plantations in Southern Pakistan. Agric. Water Manage.
36:121–139.
Landsberg, J.J. 1986. Physiological ecology of forest production. Academic Press, London, U.K., 198 p.
Landsberg, J.J. and S.T. Gower. 1997. Applications of physiological
ecology to forest management. Academic Press, San Diego, 354 p.
Lane, P., J. Morris, N.N. Zhang, G.Y. Zhou and D.P. Xu. 2003. Water
balance of Eucalyptus urophylla plantations in south China. In
Eucalypts in Asia. Ed. J. Turnbull. ACIAR Proc. No. 111, ACIAR,
Canberra, Australia, pp 217–224.
Le Maitre, D.C., B.W. van Wilgen, C.M. Gelderblom, C. Bailey, R.A.
Chapman and J.A. Nel. 2002. Invasive alien trees and water resources in South Africa: case studies of the costs and benefits of
management. For. Ecol. Manage. 160:143–159.
Mahmood, K., J. Morris, J. Collopy and P. Slavich. 2001. Groundwater uptake and sustainability of farm plantations on saline sites in
Punjab province, Pakistan. Agric. Water Manage. 48:1–20.
Monteith, J.L. 1995. A reinterpretation of stomatal responses to humidity. Plant Cell Environ. 18:357–364.
Morris, J.D. and J.J. Collopy. 1999. Water use and salt accumulation
by Eucalyptus camaldulensis and Casuarina cunninghamiana on a
site with shallow saline groundwater. Agric. Water Manage. 39:
205–227.