Agricultural and Forest Meteorology 130 (2005) 1–11
www.elsevier.com/locate/agrformet
Transpiration-use efficiency of barley
Armen R. Kemanian a,*, Claudio O. Stöckle a, David R. Huggins b,1
a
Biological Systems Engineering Department, Washington State University, Pullman, WA 99164-6120, USA
b
USDA-ARS, Washington State University, Pullman, WA 99164-6421, USA
Received 2 March 2004; accepted 6 January 2005
Abstract
Transpiration-use efficiency, the ratio of biomass (Y) produced per unit of water transpired (T) by a crop, depends on crop
characteristics and on the environment in which crops develop. Transpiration-use efficiency has been described as Y/T = kc/Da,
where kc is a crop dependent constant and Da is the daytime air vapor pressure deficit. Our objectives were to determine Y/T and
kc of barley grown in Pullman, WA, and to analyze the variation in Y/T and kc of barley (Hordeum vulgare L.) and wheat (Triticum
aestivum L.) reported in the literature. Transpiration and biomass accumulation of barley crops were measured in the years 2000
and 2001. The coefficient kc was estimated as the slope of the regression between cumulative values of biomass and T/Da. It
ranged from 6.6 0.4 to 6.9 0.2 Pa. These figures are greater than 5.8 Pa obtained by applying equations developed by Tanner
and Sinclair [Tanner, C.B., Sinclair, T.R., 1983. Efficient water use in crop production: research or re-search. In: Taylor, H.M.,
et al. (Eds.), Limitations to Efficient Water Use in Crop Production. ASA, Madison, WI, pp. 1–27]. Data on kc reported in the
literature, although scarce, ranged from 3.0 to 5.9 Pa for barley, and from 2.8 to 6.7 Pa for wheat, with the lower values occurring
at low Da (<1 kPa). This variability seems to associate with the response of the internal (leaf) to external (bulk air) CO2
concentration ratio (ci/ca) to changes of the leaf-to-air vapor pressure deficit (Dl), suggesting that kc rather than a constant could
be a function of Dl. The evaluation of more field data on kc, the field validation of the response of ci/ca to Dl, and testing this
approach for different species and cultivars is needed to improve the understanding of the Y/T determination at the canopy level.
# 2005 Elsevier B.V. All rights reserved.
Keywords: Transpiration-use efficiency; Barley; Wheat; ci/ca; Air vapor pressure deficit
1. Introduction
Transpiration-use efficiency is the ratio of biomass
produced per unit of water transpired by a crop.
* Corresponding author. Tel.: +1 509 335 1578;
fax: +1 509 335 2722.
E-mail addresses: armen@wsunix.wsu.edu (A.R. Kemanian),
stockle@wsu.edu (C.O. Stöckle), dhuggins@wsu.edu
(D.R. Huggins).
1
Tel.: +1 509 335 3379; fax: +1 509 335 3842.
Bierhuizen and Slatyer (1965) proposed that, at the
leaf level, transpiration-use efficiency is represented
by:
A kl
¼ ;
E Dl
(1)
where A is the CO2 assimilation rate per unit of leaf
area, E is the rate of evaporation per unit leaf area, kl is
a constant for leaves of a given crop, and Dl is the leafto-air vapor pressure deficit. Bierhuizen and Slatyer
0168-1923/$ – see front matter # 2005 Elsevier B.V. All rights reserved.
doi:10.1016/j.agrformet.2005.01.003
2
A.R. Kemanian et al. / Agricultural and Forest Meteorology 130 (2005) 1–11
(1965) argued that kl could be scaled up to the entire
canopy (kc), since the leaf temperature appears to be
within 2–3 8C of air temperature, and as first approximation the vapor pressure deficit of the air (Da) could
be a surrogate of Dl; hence:
Y
kc
¼
;
T Da
(2)
where Y is crop biomass production, T is crop transpiration, and kc is a canopy level constant for a given
crop.
Tanner (1981) presented a rich discussion on the
scaling of leaf (A/E) to canopy (Y/T) transpiration-use
efficiency. He pointed out that kc should be a
conservative parameter (at least within a location),
considering that the ratio between the internal (leaf)
and the external (bulk air) concentration of CO2 (ci/ca)
is reasonably constant, an idea previously expressed
by Bierhuizen and Slatyer (1965). Figures for this ratio
reported by Wong et al. (1979) were around 0.7 for C3
crops and 0.3 for C4 crops. Tanner (1981) also
indicated that kc should be lower than kl for two
reasons. First, there is a loss of CO2 due to
maintenance and growth respiration. Second, if the
effectively photosynthesizing leaf area of a canopy is
less than the effectively transpiring leaf area, kc is
further decreased below kl.
Tanner and Sinclair (1983) expanded the analysis
of Tanner (1981). They developed equations to
represent biomass production and transpiration of a
canopy from which kc can be explicitly calculated.
Since we will use these equations throughout our
analysis, they are summarized in this section. Scaling
up from leaf to canopy, they concluded that biomass
production could be represented by the following
expression (their Eq. 11):
Y¼
abcca PAId
;
1:5rd
(3)
where a is the ratio of molecular weights of carbohydrates to CO2 (0.68), b is a conversion coefficient from
hexose to biomass and varies from near 0.8 for crops
with high accumulation of sugar or starch (e.g. sugarcane and potato) to near 0.45 during the seed growth of
crops with high accumulation of oil (e.g. sunflower),
c = (1 ci/ca), PAId is the sunlit plant area index, ca is
the CO2 atmospheric concentration (g m3), and rd is
the sum of boundary layer and stomatal resistance to
water vapor flux for sunlit leaves. The fundamental
assumptions to develop Eq. (3) were: (1) c is reasonably constant (Wong et al., 1979) and (2) the photosynthesis of the shaded leaves is roughly equivalent to
the maintenance respiration of the whole canopy.
Separating canopy foliage as sunlit and shaded
leaves, Tanner and Sinclair (1983) expressed transpiration as (their Eq. 12b):
T¼
reDa ½PAId Dl =Da þ ðPAI PAId Þrd =rs
;
Prd
(4)
where r is the density of the air, e is the vapor to air
molecular weight ratio, P is the atmospheric pressure,
and rs is the sum of boundary layer and stomatal
resistance to water vapor flux for shaded leaves. As
indicated by Tanner and Sinclair (1983), the term in
brackets is the effective transpiring leaf area (PAIt).
Assuming that the ratio rd/rs is approximately 0.3 for
PAI = 4, they calculated that PAIt would be around
2.2 0.2 for PAI > 3, unless the ratio Dl/Da departs
appreciably from unity. The fundamental assumption
to develop Eq. (4) was that the shaded leaves are at air
temperature.
The quotient between the equations for biomass
and transpiration is Y/T; then kc is:
kc ¼
abcCa P PAId
:
ð1:5reÞ PAIt
(5)
The merit of the Tanner and Sinclair (1983)
analysis is widely recognized; after its publication
almost every paper dealing with transpiration-use
efficiency refers to it. Several simulation models
estimate daily growth (Stockle et al., 1994; Sinclair
and Seligman, 1995) and transpiration (Keating et al.,
1999) based on kc and Eq. (2).
Sinclair (1994) argued that, since the c values that
Tanner and Sinclair (1983) used are probably in the
upper limit for C3 and C4, kc estimated with their
equations should be close to the maximum attainable.
Barley and wheat kc can be evaluated by taking
b = 0.74 (as used for corn by Tanner and Sinclair,
1983), PAId = 1.4, PAIt = 2.2, the current ca at sea level
(0.67 g m3), and the atmospheric pressure at sea
level, giving kc of 5.8 Pa g biomass g1 water (hereafter we express kc in Pa). Sinclair and Seligman
(1995) used this value in a wheat growth model.
A.R. Kemanian et al. / Agricultural and Forest Meteorology 130 (2005) 1–11
However, field-measured kc values for barley are
scarce and the available data shows great variability.
In Table 1 we summarized data of Y/T and kc of barley
and wheat reported or calculated from the literature,
assuming that they would be similar for both crops.
Different techniques were employed by different
authors to obtain kc. Tanner (1981) indicated that kc
should be calculated as the slope of the linear
regression between biomass and the daily integration
of the quotient between transpiration and daytime Da,
and that was the method employed by Condon et al.
(1993) and Marcos (2000). The other estimates in
Table 1 are the product of Y/T times a seasonal
average of daytime Da, which can give slightly
different values than the daily integration method.
Hubick and Farqhuar, 1989 presented data for several
cultivars obtained in glasshouse experiments. We
only show data for the cultivar with the highest kc
(Table 1); other cultivars had kc 20% lower. For the
water stressed treatment kc was 5.9 Pa, almost
identical to the theoretical value obtained above;
however, for the irrigated treatment kc was 4.7 Pa.
Higher Y/T in water stressed than in unstressed crops
was reported in field grown potatoes (Vos and
Groenwold, 1989), and in pearl millet (Pennisetum
glaucum [L.] R. Br.) in pot experiments (Brück et al.,
2000), suggesting that kc of water stressed crops
could be higher than that of unstressed crops. Gregory
et al. (1992) and Siddique et al. (1990) estimated kc
for barley from field experiments (Table 1). Gregory
et al. (1992) estimated kc as the product of Y/T times
Da. However, they indicated that Da varied from
around 0.51 kPa at the beginning of the season to
1.4 kPa in the last 4 weeks of growth. Siddique et al.
(1990) reported the ratio Y/T at crop maturity but did
not report Da. We estimated Da from a paper
reporting the meteorological conditions for the same
experiment (Siddique et al., 1989); it varied from
around 0.8 to 1.2 kPa. The kc reported for Gregory
et al. (1992) range from 3.16 to 3.58 Pa and the kc
calculated from Siddique et al. (1990) is 5.14 Pa. For
wheat, reported kc ranges from 2.80 Pa (Gregory
et al., 1992) to 6.7 Pa (Angus and van Herwaarden,
2001). This variability cannot be attributed entirely to
genotypic variation; for the wheat cultivar Gutha kc
ranged from 2.80 to 4.74 Pa, and for Timgalen ranged
from 3.83 to 5.18 Pa (Table 1). Although the data is
insufficient to compare barley and wheat, the range of
3
variation of kc for both species is similar. Condon
et al. (1993) found a correlation between genotypic
variation in kc and 13C discrimination in wheat,
strongly suggesting that the genotypic variation in kc
is explained by variation in c (Condon et al., 2002).
Interestingly, Angus and van Herwaarden (2001)
reported a higher kc postanthesis compared to
preanthesis. In that study, Da was also higher
postanthesis compared with preanthesis (Table 1),
suggesting that the environment could affect kc. To
the best of our knowledge, there has been no previous
attempt to explain this variability in kc.
Our objectives were to: (1) estimate kc of barley
from field measurements of biomass accumulation and
transpiration and (2) analyze the variability in barley
and wheat kc reported in the literature in terms of the
assumptions used by Tanner and Sinclair (1983) to
develop their theoretical estimate of kc.
2. Materials and methods
Measurements of crop growth and biomass
accumulation were obtained from a larger field
experiment designed to study radiation interception
and radiation-use efficiency of barley. Briefly, field
experiments were conducted in 2000 and 2001 at the
Palouse Conservation Field Station (lat 46.88N, long
117.28W, elevation 756 m), located 5 km NW of
Pullman, WA, on a Palouse silt loam (Fine-silty,
mixed, mesic Pachic Ultic Haploxerolls). Treatments
consisted of a factorial combination of two cultivars of
spring barley (Baronesse, a two-row barley and
Steptoe, a six-row barley), two seeding densities
and two seeding dates, arranged in a complete
randomized block design with four replications in
2000 and three replications in 2001. Each plot
(2.2 m 12 m) was seeded with a no-till drill rows
20 cm apart. At seeding, each crop received
157 kg ha1 of nitrogen and 51 kg ha1 of phosphorus. The sowing dates were 27 April and 6 June in
2000, and 26 April and 13 June in 2001. Target
densities were 100 and 250 plants m2. Plots were
irrigated with sprinklers. Weeds were controlled by
hand. Diseases and insect damage were prevented or
controlled with insecticides and fungicides.
In 2000, we monitored water use and crop growth
in Baronesse seeded at the highest density in two
4
A.R. Kemanian et al. / Agricultural and Forest Meteorology 130 (2005) 1–11
Table 1
Barley and wheat transpiration-use efficiency (Y/T) and kc as reported or calculated from data obtained in the literature (the average daytime
vapor pressure deficit (Da) during the period of biomass growth and transpiration measurement is also reported)
Sourcea
Site
Species and cultivar
Y/T (g kg1)
Da (kPa)
kc (Pa)
Observationsb
(1)
Pullman, WA
Barley, cv. Baronesse
4.99
3.20
4.12
4.29
1.39
1.93
1.63
1.63
6.89
6.67
6.72
7.00
2000,
2000,
2001,
2001,
cv. Steptoe
normal seeding date
late seeding date
normal seeding date
normal seeding date
(2)
Australia (glasshouse)
Barley, cv. Proctor
3.37
4.22
1.4
1.4
4.72
5.91
Irrigated, shoot + roots
Water stressed, shoot + roots
(3)c
Merredin, Australia
Barley, cv. O’Connor
Wheat, cv. Gutha
5.35
4.93
0.96
0.96
5.14
4.74
1987
(4)
East Beverley, Australia
Barley, cv. Beecher
cv. O’Connor
cv. Syrian
Wheat, cv. Gutha
5.43
5.69
5.01
4.45
0.63
0.63
0.63
0.63
3.42
3.58
3.16
2.80
1988, shoot + roots
Wheat, cv. Timgalen
5.75
4.30
5.64
3.90
4.14
3.10
4.83
4.10
3.91
3.40
0.82
1.19
0.92
1.23
1.07
1.35
0.91
1.17
0.98
1.39
4.69
5.10
5.18
4.80
4.45
4.20
4.38
4.80
3.83
4.73
1973,
1973,
1973,
1973,
1973,
1973,
1975,
1975,
1975,
1975,
Wheat, cv. Bank
7.06
8.06
7.75
9.21
0.70
0.64
0.68
0.61
5.07
5.16
5.28
5.63
1984
1985
1984
1985
(5)c
(6)c
Werribee, Australia
cv. Quarrion
D1,
D1,
D2,
D2,
D3,
D3,
D1,
D1,
D2,
D2,
preanthesis
postanthesis
preanthesis
postanthesis
preanthesis
postanthesis
preanthesis
postanthesis
preanthesis
postanthesis
(7)
Moombooldool, Australia
Wheat, cv. Gutha
cv. Quarrion
8.16
6.75
0.54
0.71
4.37
5.51
1985, preanthesis
(8)c
Toowoomba, Australia
Wheat, cv. Hartog
4.2
1.18
4.9
1993
(9)
Pucawan, Australia
Wheat, average of cv.
Comet, Janz and Kulin
7.13
5.98
4.3
3.9
0.51
0.51
1.54
1.54
3.95
3.10
6.7
6.0
Preanthesis, low N
Preanthesis, high N
Postanthesis, low N
Postanthesis, high N
(10)c
Nottinghamshire, UK
Wheat, cv. Soissons
5.66
6.17
6.63
6.66
0.60
0.63
0.60
0.63
3.68
4.01
4.31
4.32
1994
1995
1994
1995
4.59
1.13
5.90
Pooling 1998/1999 data
cv. Maris Huntsman
(11)
a
Pullman, WA
Wheat, cv. WB926R
(1) This study; (2) Hubick and Farqhuar, 1989; (3) Siddique et al. (1990), Da from Siddique et al. (1989); (4) Gregory et al. (1992); (5) Doyle and
Fischer (1979) showed data from year 1974 too, but these were excluded because of occurrence of heavy frosts; Da was estimated as 2/3 of the maximum
saturation deficit, that is approximated as the difference in vapor saturation between maximum and minimum temperature, both obtained from the
average monthly temperature reported and a daily thermal amplitude of 12 8C; (6) Connor et al. (1992) showed data for four sowing dates per year but
we excluded the last sowing date for Quarrion because it showed irregular phenological development; Da was estimated as in (5) with the reported
maximum and minimum temperatures; (8) Meinke et al. (1997); (7) Condon et al. (1993); (9) Angus and van Herwaarden (2001); (10) Foulkes et al.
(2001) they had irrigated treatments but indicated that the irrigation added uncertainty to the transpiration estimates; we only show data for non-irrigated
crops; (11) Marcos (2000).
b
Shoot + roots indicates that roots were sampled and included in the computation of biomass; to make a rough accounting of root biomass when it
was not measured, Y/T and the resulting kc were multiplied by 1.15 (sources 5 and 7) and by 1.07 (see text) for the preanthesis and the entire crop cycle
data, respectively. No apportioning of biomass to roots was made for postanthesis data.
c
Indicates that kc was estimated as the product of Y/T times Da.
A.R. Kemanian et al. / Agricultural and Forest Meteorology 130 (2005) 1–11
replications at both sowing dates. In 2001, we
monitored water use and crop growth between booting
and the beginning of grain filling (June 28 to July 13)
for Baronesse and Steptoe seeded at the highest
density in the first seeding date. The techniques
employed to estimate water use in each year differed.
In 2000, soil water content was measured from
emergence to physiological maturity with water
reflectometer probes (model CS615-L, Campbell
Scientific Inc., Logan, UT). In each plot, five probes
were installed to a depth of 1.5 m at 0.3 m interval. An
additional sensor was installed in the first layer at
angle of 208 from the soil surface to monitor water
content in the first 0.1 m of the soil profile. Each probe
was connected to a datalogger (CR10x, Campbell
Scientific Inc., Logan, UT), and the signal recorded at
midnight. Each probe was calibrated against measured
volumetric water content. Gravimetric water content
was measured four times during the growing season to
a depth of 1.8 m at 0.3 m interval. The samples were
taken at crop emergence, beginning of stem elongation, heading and harvest. Volumetric water content
was estimated as the product of the gravimetric water
content and the bulk density of the layer. Bulk density
was measured to a depth of 1.8 m at 0.3 m intervals by
taking three soil cores of 2 cm in diameter and 20 cm
long in spring.
A weather station located in the border of the
experimental area recorded hourly precipitation,
temperature, relative humidity, and wind speed.
Radiation interception was measured using one tube
solarimeter (70 cm) per plot (Marcos, 2000). After the
plants reached the two- to three-leaf stage, the
solarimeters were placed below the canopy in areas
representative of the plot. Each solarimeter was
connected to a datalogger, and the signal recorded
every 20 min. Simultaneously, solar radiation was
measured at a height of 2.5 m with a pyranometer
(LI200X, Licor Inc., Lincoln, NE, USA). The
pyranometer and the solarimeter outputs were
integrated to obtain daily solar irradiance and daily
solar radiation transmitted through the canopy, and the
values used to calculate daily fractional radiation
interception ( f i). The seasonal variation in f i was
reported in Kemanian et al. (2004). The replicates of
each treatment for f i showed very low variability (data
not shown), consistent with the observations that the
crop stand was homogenous and that the plants tillered
5
aggressively, compensating minor unevenness in the
plants distribution.
Daily evapotranspiration (ET) was calculated from
a water balance for the soil profile:
ET ¼ PP þ I DP R DS;
(6)
where, PP is precipitation, I is irrigation, DP is deep
percolation, R is runoff, and DS is the change in
storage to a depth of 1.8 m. Runoff did not occur
during the course of the experiment. Deep percolation
was only evident at the beginning of the measurement
period in the first sowing date, when the sensors at 1.2
and 1.5 m of depth showed a slight decrease in water
content, but thereafter the signal stabilized until the
crops started to uptake water from that layer towards
the end of the crop growth cycle. Daily ET was
partitioned into soil evaporation and crop transpiration
on a daily basis. Soil evaporation was taken as the
minimum of the change in storage at the first 0.1 m
soil and the product (1 f i)ET. The remaining fraction of ET was apportioned to transpiration. Aboveground biomass was estimated from samples of two
adjacent 0.5 m length rows (0.2 m2) per plot, at
intervals of 6–10 days until physiological maturity.
Samples were dried at 60 8C for 72 h and the dry
weight recorded.
In the year 2001, measurements were taken
between flowering and beginning of grain filling
(June 28 to July 13). We selected this period because it
shows the highest growth and water depletion rate of
the crop cycle, the crops fully cover the ground
( f i 0.9) minimizing soil evaporation and then errors
in the estimation of transpiration, and the probability
of occurrence of precipitation are low. No irrigation
was applied during that time interval. In each plot at
high density (three replications), two sections of about
1 m2 were flagged. The sections were visually
identical. We carefully avoided borders, areas with
uneven plants distribution or populations unusually
high or low. Aboveground biomass samples of two
adjacent rows 1 m long were taken at the beginning
and at the end of the selected period; one of the flagged
areas was sampled at each time. Samples were dried at
60 8C for 72 h and the dry weight recorded.
Concurrently, gravimetric water content was measured to a depth of 1.5 m at 0.3 m interval. An
additional sample was taken in the first 5 cm. Each
sample was a composite of two cores 30 cm long
6
A.R. Kemanian et al. / Agricultural and Forest Meteorology 130 (2005) 1–11
(5 cm in the top layer) and a diameter of 2 cm. There
was a rain of 3.3 mm during that period but there was
no runoff and we assumed deep percolation to be
negligible. Therefore, ET was approximated as the
change in storage plus the precipitation during the
period June 28 to July 13. The change in storage of the
first 5 cm was attributed to soil evaporation and the
remaining fraction of ET apportioned to transpiration.
In 2000, the transpiration efficiency (Y/T) was
estimated as slope of the linear regression between
cumulative Y against cumulative T, and the coefficient
kc was estimated as the slope of the linear regression
between Y and daily cumulative T/Da. We considered
that the estimations of T for Y > 100 g m2 ( f i of
approximately 0.35) were more accurate than the
estimations of T for Y < 100 g m2. Therefore,
although in theory the regression should pass through
the origin, we prioritize honoring the actual data and
did not set the intercept to zero. In 2001, Y/T was
calculated as the biomass gained in the period divided
by the estimated transpiration, and kc was calculated
as the product of Da times Y/T, where Da was averaged
for the period considered.
3. Results and discussion
In the year 2000, crop water uptake was restricted
to a depth of 1.5 m; no change in the water content
between 1.5 and 1.8 m was detected between heading
and maturity (data not shown). Biomass accumulation
was linearly related to both cumulative transpiration
(slopes of 4.7 0.4 and 3.0 0.6 g biomass kg1
water for the first and the second sowing date) and
cumulative transpiration normalized by Da (Fig. 1). As
observed by Tanner (1981) in potatoes, the normalization of T by Da decreased the scatter of the data; the
slopes of the first and second sowing dates were
indistinguishable giving a common kc of 6.2 0.4 Pa
(cv. Baronesse). If an accounting for roots is made
assuming that on average over the entire crop cycle
about 7% of the biomass is allocated belowground,
then kc = 6.6 0.4 Pa. The 7% figure comes from the
following calculation. Gregory et al. (1978) indicated
that roots represent about 15% of the total biomass at
anthesis and that root growth stops thereafter. In this
area, spring cereals duplicate the aboveground
biomass from beginning of anthesis to maturity.
Therefore, if root biomass is kept constant, it amounts
to about 7% of total biomass at harvest. Another
calculation that gives a similar estimate is converting
root density measurements to root biomass based on a
root length density of 240 m g1 (Gregory et al.,
1978). With a maximum root density at the surface of
4 cm cm3 (see the review table by de Willigen and
van Noordwijk, 1987, p.88) and a exponential
decrease in root density (Dwyer et al., 1998) to a
depth of 1.5 m, root biomass would represent slightly
less than 7% of the total biomass for a crop with
10 Mg ha1 of aboveground biomass. In the year
Fig. 1. Biomass as a function of cumulative transpiration (panel A) and cumulative transpiration normalized by the air vapor pressure deficit (Da)
(panel B), of spring barley cv. Baronesse grown at Pullman, WA, year 2000.
A.R. Kemanian et al. / Agricultural and Forest Meteorology 130 (2005) 1–11
2001, the average ET for the period considered was
115 7 mm, the average T was 106 7 mm, and the
average crop growth rate was 29.4 1.2 g m2 d1.
The average kc estimated was 6.9 0.2 Pa (n = 6); the
estimates for Baronesse and Steptoe, although not
statistically different, are shown separately in Table 1.
No accounting for root biomass was made in the year
2001 as root growth stops around anthesis (Gregory
et al., 1978).
Comparing the values that we obtained for barley
with the values summarized in Table 1, Y/T values are
within the range reported in the literature, while kc
values of 6.6 and 6.9 Pa are in the upper limit reported
in the literature for both barley and wheat, and so are
the Da values. Furthermore, while some kc presented in
Table 1 are similar to the 5.8 Pa obtained by applying
Eq. (5), others are noticeable lower. We graphed Y/T of
Table 1 against the daytime Da (Fig. 2) and found that,
as suggested by Bierhuizen and Slatyer (1965), Y/T
seems to be an inverse function of Da, that is well
represented by the model Y/T = 4.9Da0.59 (r2 = 0.97,
n = 38, P < 0.001). The upper envelope of the data
departs from the Y/T predicted by Eq. (2) using
kc = 5.8 Pa, particularly at Da < 1 kPa, where predicted Y/T is much higher than the observed Y/T,
Fig. 2. Transpiration-use efficiency (Y/T, g biomass kg1 H2O) as a
function of the vapor pressure deficit of the air. The points are the Y/T
reported in the literature for wheat and barley (Table 1); the dotted
line is for Y/T = 5.8/Da; the solid line is for Y/T calculated using
c = 1 (0.85 0.05Dl) (Eq. (7), see text) and (Eqs. (2), (5), (8) and
(9)); the dashed line is for Y/T calculated using
c = 1 (0.85 0.12Dl); see text for explanations. The solid line
is well approximated by Y/T = 3.8/Da3/4 and the dashed line by Y/
T = 5.2/Da3/5.
7
except the highest Y/T reported by Connor et al. (1992)
(Fig. 2). We suspect that Connor et al. (1992)
underestimated T. The information presented in
their paper allows calculating Y/T for different
phenological stages, and we found that in the third
seeding date of 1984, the cultivar Quarrion Y/T was
17 g biomass kg1 H2O in the period between floral
initiation and anthesis. This value is unusually high,
even for a C4 plant, and may be indicative of an
underestimation of transpiration and consequently an
overestimation of Y/T. There is also considerable
scattering in the data of Fig. 2, and although part of
this scattering could be genetic variation, it is probably
better explained by uncertainties in measuring
transpiration and growth. However, the upper and
lower envelopes of the data are almost parallel,
suggesting that they are truly depicting the trend of Y/T
as a function of Da (Fig. 2). The significant conclusion
here is that, to make the reported Y/T compatible with
Eq. (2), kc cannot be a constant but rather decrease
with decreasing Da, which is in fact the trend observed
in Table 1.
One of the fundamental assumptions of Bierhuizen
and Slatyer (1965) and Tanner and Sinclair (1983),
was that the ratio ci/ca, and hence c, is fairly constant.
However, there is evidence that c varies in response to
physiological and environmental factors. Among
these factors, it is well documented that ci/ca decreases
as water stress increases in field-grown potatoes (Vos
and Groenwold, 1989), wheat (Whitfield, 1990), and
sorghum (Williams et al., 2001). Indirect evidence of
that response to water deficit and the associated
decrease in stomatal conductance is the decrease in the
13
C discrimination in water stressed plants (e.g.
Hubick and Farqhuar, 1989; Condon et al., 1992).
Interestingly, there is also evidence that the ratio ci/ca
decreases with increasing Dl (Table 2), and effect
suggested in the analysis by Condon et al. (1992).
Farquhar et al. (1982) concluded that when assimilation rate is reduced by a decrease in the stomatal
conductance, ci should decrease. Then, the response of
ci/ca to Dl could be explained by a reduction in
stomatal conductance at increasing transpiration rates
(Mott and Parkhurst, 1991), caused in turn by
increasing Dl.
The magnitude of the response of ci/ca to Dl seems
to be associated to the photosynthetic metabolism.
While the extrapolated ci/ca at Dl = 0 is approximately
8
A.R. Kemanian et al. / Agricultural and Forest Meteorology 130 (2005) 1–11
Table 2
Relation between the ratio leaf-to-ambient CO2 concentration (ci/ca) and the leaf-to-air vapor pressure deficit (Dl) for several species (the Dl at
which the plants were exposed during the measurements is reported)
Sourcea
Species
Intercept
Slope (kPa1)
r2
n
Dl range (kPa)
C3 plants
(1)
Nicotiana glauca
0.81
0.051
0.63
10
0.5–3.0
(2)
Gossypium hirsutum
0.83
0.120
0.99
5
1.5–3.4
(3)
Phalaris aquatica
Oryza sativa
0.95
0.91
0.093
0.120
0.97
0.97
4
4
0.4–2.0
0.4–2.0
(4)
Phaseolus vulgaris
0.84
0.065
0.97
6
0.8–3.0
(5)
Ricinus communis
0.90
0.83
0.070
0.126
0.97
0.81
6
4
0.2–2.0
0.2–2.0
(6)
Encelia farinosa
0.89
0.027
0.85
28
1.0–4.0
(7)
Solanum tuberosum
0.93
0.070
0.55
32
0.9–3.0
Average C3
0.88
0.082
Zea mayz
Paspalum plicatulum
0.85
0.93
0.203
0.179
0.99
0.98
4
4
0.4–2.0
0.4–2.0
(6)
Pleuraphis rigida
0.80
0.101
0.88
21
1.0–4.0
(7)
Sorghum bicolor
1.00
0.246
0.79
22
1.0–2.5
Average C4
0.90
0.182
C4 plants
(3)
(1) Farquhar et al. (1980), their Fig. 4, plants grew in growth chamber, ca 330 mmol mol1, photosynthetic photon flux density
(PPFD) 480 mmol m2 s1; (2) Sharkey et al. (1982), their Fig. 2, plants grew at ca 1900 mmol mol1, measurements were at
ca 350 mmol mol1; (3) Morison and Gifford (1983), their Fig. 8, plants grew in growth chamber at Da 0.95 kPa, ca 340 mmol mol1,
1, and PPFD 670 mmol m2 s1; (4) Commstock and Ehleringer, 1993, their Fig. 2, plants grew in glasshouse with Da 2.5 kPa,
ca 350 mmol mol1, and maximum PPFD 1600 mmol m2 s1; (5) Dai et al. (1992) plants grew in growth chamber Da 1.3 kPa,
ca 345 mmol mol1, and PPFD 600 mmol m2 s1, measurements at 20 8C and at PPFD of 1000 and 1800 mmol m2 s1 shown here; (6)
Zhang and Nobel (1996), field and growth chamber study, ca 380 mmol mol1, maximum PPFD 1600 mmol m2 s1, measurements made
several times over daytime; (7) Bunce (personal communication, 2003), field study, Da varied day-to-day, measurements were on plants grown at
ca of 350 and 700 mmol mol1, the regressions were identical for both CO2 levels and were pooled, PPFD > 1500 mmol m2 s1; for potatoes,
the regression was not significant for instantaneous Da variation.
a
0.90 for both C3 and C4 plants, the slope of the
response is on average 0.08 kPa1 for C3 and
0.18 kPa1 for C4 plants (Table 2). Choudhury
(1986) calculated ci for field grown cotton based on
measured net photosynthesis and other biophysical
considerations; we estimated that the slope of ci/ca
versus Dl was 0.07 kPa1 (Choudhury, 1986, Fig. 5),
very similar to the average of 0.08 kPa1 for C3 crops
shown in Table 2. Hence, the empirical relation
between ci/ca and Dl appears robust and could be
included in an analysis of Y/T. Zhang and Nobel
(1996) developed a simple model to explain the
response of A/E to Dl, explicitly considering the linear
response of ci/ca to Dl. However, there is no report
integrating this response to the canopy level Y/T.
We calculated Y/T (and kc) using Eq. (5) but
allowing c to vary linearly with Dl:
ci
(7)
c ¼ 1 ¼ 1 ðd0 þ d1 Dl Þ;
ca
where d0 and d1 are the intercept and slope, respectively, of the regression of ci/ca versus Dl (Table 2).
Since there is no report of the response of ci/ca for
barley and wheat, we can only assume that the
response would be somewhere within the range indicated by the data gathered in Table 2. We calculated Y/
T using the average of the three steepest slopes
(d1 = 0.12 kPa1) and the average of the three less
steep slopes (d1 = 0.05 kPa1) for the C3 species
summarized in Table 2. In both cases the average d0
was 0.85. The Y/T obtained with the steepest d1 would
A.R. Kemanian et al. / Agricultural and Forest Meteorology 130 (2005) 1–11
approximate an upper boundary for C3, while the Y/T
obtained with the less steep d1 would represent a lower
boundary. With d1 = 0.12 kPa1, the value of c of 0.3
used by Tanner and Sinclair (1983) is achieved at
Dl = 1.25 kPa. Vos and Groenwold (1989) reported ci/
ca 0.77 (c = 0.23) at Da 1.4 kPa for the potato cv.
Saturna; assuming that Dl is approximated by Da,
Eq. (7) predicts c = 0.22 (d1 = 0.05 kPa1) and
c = 0.32 (d1 = 0.12 kPa1).
Two more responses to change in the vapor
pressure deficit should be considered for completeness
in applying Eq. (5): (1) the response of the stomatal
resistance and thus the ratio rd/rs, and (2) the canopy
evaporative cooling and thus the ratio Dl/Da.
Choudhury and Monteith (1986) summarized information showing that the stomatal resistance increases
with increasing Dl. Due to the exposure of sunlit
leaves, we can expect that the increase in resistance
with rising Dl be more acute in the sunlit leaves,
increasing the ratio rd/rs. Since we did not find
experimental information, we assumed that a linear
function could approximate this effect:
rd
Dl
¼ 0:2 þ 0:8 :
rs
7
(8)
This function assumes that when Dl = 0 kPa the
ratio of resistances is 0.2, and increases until a
maximum of 1 for Dl = 7 kPa. The limit of 7 kPa was
taken from Choudhury and Monteith (1986), who
assumed by extrapolation of published data that
stomata of arable crops close at Dl of approximately
7 kPa. We did not perform any calculation above
5 kPa, which is a reasonable upper limit for the Dl
experienced by field crops. Although field measurements of resistance of both sunlit and shaded leaves
should be performed to test the validity of this
approximation, the use of an intercept of 0.2 seems
to be consistent with some available measurements.
For example, this equation predicts that the ratio rd/rs
equals 0.3 at a Dl of 0.9 kPa, the ratio assumed by
Tanner and Sinclair (1983) for their computations of
kc. Kjelgaard (1993) reported the stomatal resistance
for both sunlit and shaded leaves of irrigated corn
(average PAI = 3.7, range 2.0–4.5). From this data, we
calculated that the ratio rd/rs averaged 0.4 for an
average Da of 1.7 kPa, while Eq. (8) yields 0.39.
The leaf-to-air vapor pressure deficit depends on
the canopy temperature and the air vapor pressure.
9
Although the canopy temperature depends on the
energy balance of the foliage, Idso (1982) showed that
for well-watered crops the difference between canopy
and air temperature (DT) is a linear function of Da. Even
more, Idso et al. (1987) showed that this relationship is
weakly affected by doubling ca from 330 to
660 mmol mol1. For barley Idso (1982) reported that:
DT ¼ 2:01 2:25Da :
(9)
For potato and sunflower the slopes reported by
Idso are also near 2.0 C kPa1. We used this equation
to calculate the ratio Dl/Da assuming that it gives the
sunlit leaves temperature and that the shaded leaves
are at the air temperature.
It is of interest to analyze the behavior of PAIt as a
function of Da after the introduction of these changes.
Increasing Da causes PAId to be cooler than the air and
consequently a decreasing ratio Dl/Da than under the
isothermal condition (Dl/Da = 1), and an increasing
ratio rd/rs. Hence, on the computation of PAIt the
weight of the sunlit leaf area decreases while the
weight of the shaded leaf area increases, keeping the
value reasonably bounded. For instance, for a dew
point temperature of 10 8C, air temperatures varying
from 15 to 30 8C, d1 = 0.12 kPa1, a PAI of 4 and a
PAId of 1.4, PAIt is 2.1 0.2, reasonably close to the
2.2 used by Tanner and Sinclair (1983). For this set of
conditions, c ranges from 0.22 to 0.39. Then, the
adjustment of c by Da will cause a greater impact on kc
than the adjustment of PAIt by Da.
The resulting Y/T calculated with variable kc
follows approximately the lower envelope
(d1 = 0.05 kPa1, Fig. 2, solid line) and the upper
envelope (d1 = 0.12 kPa1, Fig. 2, dashed line) of
the data gathered from the literature. The calculated kc
decrease with decreasing Da as suggested by the data
of Table 1; for instance, using d1 = 0.12 kPa1,
Eq. (5) gives kc = 3.9 Pa for Da = 0.5 kPa and
kc = 6.8 Pa for Da = 2.0 kPa. The differences in Y/T
suggested by the so-called upper and lower boundary
could represent both differences among species or
cultivars. In comparing two cultivars of potatoes, Vos
and Groenwold (1989) found that the cultivar Bintje
(high stomatal resistance) had Y/T about 14% higher
than the cultivar Saturna (low stomatal resistance).
Genetic differences in Y/T were also reported for
barley (Hubick and Farqhuar, 1989) and wheat
10
A.R. Kemanian et al. / Agricultural and Forest Meteorology 130 (2005) 1–11
(Condon et al., 1993). Therefore, although the
information presented is not conclusive, it strongly
suggests that the variability observed in Y/T could be
associated to the variation of c as a function of Da, and
that kc is not constant.
Both a constant or a variable kc imply that the water
use efficiency increases with decreasing Da, but the
advantage in biomass production at low Da compared
with high Da predicted by the variable kc is smaller
than that predicted by a constant kc of 5.8 Pa (Fig. 2).
Tanner and Sinclair (1983) warned that under low Da
the departure of the ratio Dl/Da from unity due to
canopy temperature rising above the air temperature
could cause deviation from their calculations. The
additions presented here to their equations seem to
account explicitly for that effect. Testing this approach
under high Da would be of major interest. However,
due to the season at which barley and wheat are grown,
average Da > 2 kPa are rarely obtained for a long
period of time. Consequently, we could not find data in
the literature reporting Y/T at Da > 2 kPa.
The set of equations presented can be used to
speculate on the effect of water stress on kc. The
temperature of a water stressed canopy can be 2–4 8C
above the air temperature. This would cause a rise in
both c Eq. (7) and PAIt Eq. (4). We tried several
combinations of dew point, air and canopy temperatures, and found only a marginal variation of kc. This is
in disagreement with some investigations showing that
Y/T tends to increase with increasing water stress (e.g.
Vos and Groenwold, 1989). The reason is that in
addition to the effect on c through an increase in Dl,
water stressed plants close the stomata to prevent
dehydration, a response attributed to an abscisic acidmediated root signal (Davies and Zhang, 1991), which
causes a further increase in c. This would explain an
increase in Y/T of stressed crops, and suggests that kc
of Eq. (2) should be calculated independently for
stressed and unstressed crops.
4. Conclusions
The coefficient kc of barley (cv. Baronesse)
measured at Pullman ranged from 6.6 0.4 to
6.9 0.2 Pa. Thus, in this environment and for the
cultivars selected, kc is greater than the 5.8 Pa estimate
obtained from Tanner and Sinclair (1983) Eq. (5). This
difference, as well as the variability reported in Y/T and
kc in the literature, seems to be accounted for by
allowing the ratio ci/ca to vary as a function of Dl in the
Tanner and Sinclair (1983) derivation. However, both
the quality and the quantity of available data on kc of
barley (and wheat) preclude being conclusive. The
evaluation of more field data on kc, the field validation
of the response of ci/ca to Dl, and testing this approach
for different species and cultivars is needed to improve
the understanding of the transpiration-use efficiency
determination at the canopy level.
Acknowledgment
Dr. James Bunce, from USDA-ARS Alternate
Crops and Systems, generously provided the information presented in Table 2 for potatoes and sorghum.
References
Angus, J.F., van Herwaarden, A.F., 2001. Increasing water use and
water use efficiency in dryland wheat. Agron. J. 93, 290–298.
Bierhuizen, J.F., Slatyer, R.O., 1965. Effect of atmospheric concentration of water vapour and CO2 in determining transpiration–photosynthesis relationships of cotton leaves. Agric.
Meteorol. 2, 259–270.
Brück, H., Payne, W.A., Sattelmacher, B., 2000. Effect of phosphorus and water supply on yield, transpirational water use
efficiency and carbon isotope discrimination of pearl millet.
Crop Sci. 40, 120–125.
Choudhury, B.J., 1986. An analysis of observed linear correlations
between net photosynthesis and a canopy-temperature-based
plant water stress index. Agric. For. Meteorol. 36, 323–333.
Choudhury, B.J., Monteith, J.L., 1986. Implications of stomatal
response to saturation deficit for the heat balance of vegetation.
Agric. For. Meteorol. 36, 215–225.
Commstock, J., Ehleringer, J., 1993. Stomatal response to humidity
in common bean (Phaseolus vulgaris L.): implications for
maximum transpiration rate, water-use efficiency and productivity. Aust. J. Plant Physiol. 20, 669–691.
Condon, A.G., Richards, R.A., Farquhar, G.D., 1992. The effect of
variation in soil water availability, vapor pressure deficit and
nitrogen nutrition on carbon isotope discrimination in wheat.
Aust. J. Agric. Res. 43, 935–947.
Condon, A.G., Richards, R.A., Farquhar, G.D., 1993. Relationships
between carbon isotope discrimination, water use efficiency and
transpiration efficiency for dryland wheat. Aust. J. Agric. Res.
44, 1693–1711.
Condon, A.G., Richards, R.A., Rebetzke, G.J., Farquhar, G.D.,
2002. Improving intrinsic water-use efficiency and crop yield.
Crop Sci. 42, 122–131.
A.R. Kemanian et al. / Agricultural and Forest Meteorology 130 (2005) 1–11
Connor, D.J., Theiveyanathan, S., Rimmington, G.M., 1992. Development, growth, water-use and yield of a spring and a winter
wheat in response to time of sowing. Aust. J. Agric. Res. 43,
493–519.
Dai, Z., Edwards, G.E., Ku, M.S.B., 1992. Control of photosynthesis
and stomatal conductance in Ricinus communis L. (castor bean)
by leaf to air vapor pressure deficit. Plant Physiol. 99, 1426–
1434.
Davies, W.J., Zhang, J.H., 1991. Root signals and the regulation of
growth and development of plants in drying soils. Annu. Rev.
Plant Physiol. 42, 55–76.
de Willigen, P., van Noordwijk, M., 1987. Roots, plant production
and nutrient use efficiency. Ph.D. Thesis. Agricultural University, Wageningen, 288 pp.
Doyle, A.D., Fischer, R.A., 1979. Dry matter accumulation and
water use relationships in wheat crops. Aust. J. Agric. Res. 30,
815–829.
Dwyer, L.M., Stewart, D.W., Balchin, D., 1998. Rooting characteristics of corn, soybeans and barley as a function of available
water and soil physical characteristics. Can. J. Soil Sci. 68, 121–
132.
Farquhar, G.D., O’Leary, M.H., Berry, J.A., 1982. On the relationship between carbon isotope discrimination and the intercellular
carbon dioxide concentration in leaves. Aust. J. Plant Physiol. 9,
121–137.
Farquhar, G.D., Schulze, E.-D., Küppers, M., 1980. Response to
humidity by stomata of Nicotiana glauca L. and Coryllus avellana L. are consistent with optimization of carbon dioxide uptake
with respect to water loss. Aust. J. Plant Physiol. 7, 315–327.
Foulkes, M.J., Scott, R.K., Sylvester-Bradley, R., 2001. The ability
of wheat cultivars to withstand drought in UK conditions:
resource capture. J. Agric. Sci. (Cambridge) 137, 1–16.
Gregory, P.J., McGowan, M., Biscoe, P.V., Hunter, B., 1978. Water
relations of winter wheat. I. Growth of the root system. J. Agric.
Sci. (Cambridge) 91, 91–102.
Gregory, P.J., Tennant, D., Belford, R.K., 1992. Root and shoot
growth, and water and light use efficiency of barley and wheat
crops grown on a shallow duplex soil a Mediterranean-type
environment. Aust. J. Agric. Res. 43, 555–573.
Hubick, K., Farqhuar, G., 1989. Carbon isotope discrimination and
the ratio of carbon gained to water lost in barley cultivars. Plant
Cell Environ. 12, 795–804.
Idso, S.B., 1982. Non-water-stressed baselines: a key to measuring
and interpreting plant water stress. Agric. Meteorol. 27, 59–70.
Idso, S.B., Kimball, B.A., Mauney, J.R., 1987. Atmospheric carbon
dioxide enrichment effects on cotton midday foliage temperature: implications for plant water use and crop yield. Agron. J.
79, 667–672.
Keating, B.A., Robertson, M.J., Muchow, R.C., Huth, N.I., 1999.
Modelling sugarcane production systems. I. Description and
validation of the APSIM Sugarcane module. Field Crops Res.
61, 253–271.
Kemanian, A.R., Stöckle, C.O., Huggins, D.R., 2004. Variability of
barley radiation-use efficiency. Crop Sci. 44, 1662–1672.
Kjelgaard, J.F., 1993. Comparison of models for estimating evapotranspiration in maize. M.Sc. Thesis. Washington State University, 100 pp.
11
Marcos, J., 2000. Simulation-based assessment of alternative crops
in the dryland Pacific Northwest. Ph.D. Dissertation. Washington State University, 171 pp.
Meinke, H., Hammer, G.L., van Keulen, H., Rabbinge, R., Keating,
B.A., 1997. Improving wheat simulation capabilities in Australia
from a cropping systems perspective: water and nitrogen effects
on spring wheat in a semi-arid environment. Eur. J. Agron. 7, 75–
88.
Morison, I.L., Gifford, R.M., 1983. Stomatal sensitivity to carbon
dioxide and humidity: a comparison of two C3 and two C4 grass
species. Plant Phys. 71, 789–796.
Mott, K.A., Parkhurst, D.F., 1991. Stomatal response to humidity in
air and helox. Plant Cell Environ. 14, 509–515.
Sharkey, T.D., Imai, K., Farquhar, G.D., Cowan, I.R., 1982. A direct
confirmation of the standard method of estimating intercellular
partial pressure of CO2. Plant Physiol. 69, 657–659.
Siddique, K.H.M., Beldford, R.K., Perry, M.W., Tennant, D., 1989.
Growth, development and light interception of old and modern
wheat cultivars I a Mediterranean-type environment. Aust. J.
Agric. Res. 40, 473–487.
Siddique, K.H.M., Tennant, D., Perry, M.W., Beldford, R.K., 1990.
Water use and water use efficiency of old and modern wheat
cultivars in a Mediterranean-type environment. Aust. J. Agric.
Res. 41, 431–447.
Sinclair, T.R., 1994. Limits to crop yield? In: Boote, K.J., et al.
(Eds.), Physiology and Determination of Crop Yield. Proceedings Symposium on Physiology and Determination of Crop
Yield, ASA, CSSA, SSSA, USDA-ARS, University of Florida
IFAS, Gainesville, Florida, 10–14 June 1991. ASA, Madison,
WI, pp. 509–532.
Sinclair, T.R., Seligman, N.G., 1995. Global environment change
and simulated forage quality of wheat I Nonstressed conditions.
Field Crops Res. 40, 19–27.
Stockle, C.O., Martin, S., Campbell, G.S., 1994. CropSyst, a cropping systems model: water/nitrogen budgets and crop yield.
Agric. Syst. 46, 335–359.
Tanner, C.B., 1981. Transpiration efficiency of potato. Agron. J. 73,
59–64.
Tanner, C.B., Sinclair, T.R., 1983. Efficient water use in crop
production: research or re-search. In: Taylor, H.M., et al.
(Eds.), Limitations to Efficient Water Use in Crop Production.
ASA, Madison, WI, pp. 1–27.
Vos, J., Groenwold, J., 1989. Characteristics of photosynthesis and
conductance of potato canopies and the effects of cultivar and
transient drought. Field Crops Res. 20, 237–250.
Whitfield, D.M., 1990. Canopy conductance, carbon assimilation
and water use in wheat. Agric. For. Meteorol. 53, 1–18.
Williams, D.G., Gempko, V., Fravolini, A., Leavitt, S.W., Wall,
G.W., Kimball, B.A., Pinter Jr., P.J., LaMorte, R., Ottman, M.,
2001. Carbon isotope discrimination by Sorghum bicolor under
CO2 enrichment and drought. New Phytol. 150, 285–293.
Wong, S.C., Cowan, I.R., Farquhar, G.D., 1979. Stomatal conductance correlates with photosynthetic capacity. Nature 282, 424–
426.
Zhang, H., Nobel, P.S., 1996. Dependency of ci/ca and leaf transpiration efficiency on the vapor pressure deficit. Aust. J. Plant
Physiol. 23, 561–568.