Crop Science 42:122-131 (2002)
© 2002 Crop Science Society of America
CROP PHYSIOLOGY & METABOLISM
Improving Intrinsic Water-Use Efficiency and Crop Yield
A. G. Condon*,a,
R. A. Richardsa,
G. J. Rebetzkea and
G. D. Farquharb
a CSIRO Plant Industry, GPO Box 1600, Canberra, ACT, 2601, Australia
b Environmental Biology, Australian National Univ., Canberra, ACT, 2601, Australia
* Corresponding author (tony.condon{at}pi.csiro.au)
 |
ABSTRACT
|
|---|
Greater yield per unit rainfall is one of the most important challenges in dryland agriculture. Improving intrinsic water-use efficiency (WT), the ratio of CO2 assimilation rate to transpiration rate at the stomata, may be one means of achieving this goal. Carbon isotope discrimination (
13C) is recognized as a reliable surrogate for WT and there have now been numerous studies which have examined the relationship between crop yield and WT (measured as 13C). These studies have shown the relationship between yield and WT to be highly variable. The impact on crop yield of genotypic variation in WT will depend on three factors: (i) the impact of variation in WT on crop growth rate, (ii) the impact of variation in WT on the rate of crop water use, and (iii) how growth and water use interact over the crop's duration to produce grain yield. The relative importance of these three factors will differ depending on the crop species being grown and the nature of the cropping environment. Here we consider these interactions using (i) the results of field trials with bread wheat (Triticum aestivum L.), durum wheat (T. turgidum L.), and barley (Hordeum vulgare L.) that have examined the association between yield and 13C and (ii) computer simulations with the SIMTAG wheat crop growth model. We present details of progress in breeding to improve WT and yield of wheat for Australian environments where crop growth is strongly dependent on subsoil moisture stored from out-of-season rains and assess other opportunities to improve crop yield using WT.
Abbreviations: A, instantaneous rate of CO2 assimilation • ca, CO2 concentration of the air • ci, CO2 concentration inside the leaf • 13C, carbon isotope discrimination • ET, evapotranspiration • g, stomatal conductance • T, transpiration (instantaneous rate or cumulative) • wa, water vapor concentration of the air • wi, water vapor concentration inside the leaf • WT, intrinsic water-use efficiency
|
INTRODUCTION
|
|---|
YIELDS WOULD BE GREATER in many cropping regions if more water was available for crop growth. Water is essential to plant growth because it provides the medium within which most cellular functions take place. Water also is the required unit of exchange for the acquisition of CO2 by plants. But plants differ in their capacity to regulate how much water is lost per unit carbon gained. Such differences can be referred to as differences in "intrinsic water-use efficiency", WT. The realisation that carbon isotope discrimination could provide an indirect measure of variation in WT (Farquhar et al., 1982; Farquhar and Richards, 1984) has rekindled the prospect of exploiting differences in WT to improve the yield of crop species. This review will give a brief outline of prospects, problems, and progress associated with recent attempts to improve WT of C3 crops via the use of carbon isotope analysis of plant tissue. The review will concentrate on research and breeding being conducted at CSIRO Plant Industry for wheat grown under rain-fed conditions in Australia but will attempt to draw out issues likely to be of more general significance to those working with other crop species and in other environments.
|
Intrinsic Water Use Efficiency
|
|---|
The term "intrinsic water-use efficiency" can be defined as the ratio of the instantaneous rates of CO2 assimilation (A) and transpiration (T) at the stomata. The instantaneous rates of both A and T can be described by relatively simple equations. Both A (Eq. [1]) and T (Eq. [2]) are the product of two factors: stomatal conductance (g) to either CO2 (gc) or water vapor (gw) and a concentration gradient of either CO2 (ca - ci) or water vapor (wi - wa) between the air outside the leaf and the air inside the leaf.
 | [1] |
 | [2] |
For CO2 the concentration is greater outside the leaf, while for water vapor the concentration is greater inside the leaf. Intrinsic water-use efficiency (WT), the ratio of A and T, can be closely approximated by Eq. [4] which derives simply from Eq. [3].
 | [3] |
 | [4] |
where the factor 0.6 refers to the relative diffusivities of CO2 and water vapor in air. Assuming the water vapor concentration gradient is an independent variable then Eq. [4] indicates that WT is a negative function of the ratio ci (the intercellular CO2 concentration) and ca (the atmospheric CO2 concentration).
For non-stressed plants of C3 species, the value of ci/ca is typically near 0.7 (Farquhar et al., 1989). This "operating" value of ci/ca is determined by the balance between stomatal conductance and photosynthetic capacity. Stomatal conductance influences the supply of CO2 to the leaf interior, whereas photosynthetic capacity determines the demand for CO2. Photosynthetic capacity is the amount and activity of photosynthetic machinery per unit leaf area. A lower value of ci/ca and hence improved WT can be achieved either through lower stomatal conductance or higher photosynthetic capacity or a combination of both.
Examination of Eq. [4] and Fig. 1
reveals that large improvements in WT are theoretically possible for relatively modest changes in the value of ci/ca. For example, if ci/ca were to be "lowered" from 0.7 to 0.6, then the theoretical gain in WT would be 33%, noting that WT is proportional to (1 - ci/ca). But there are likely to be trade-offs. If ci/ca is lower as a result of an increase in photosynthetic capacity (from set-point 1 to set-point 2 on Fig. 1) then there will be an increase in A per unit leaf area. However, if ci/ca is lower as a result of lower stomatal conductance (from set-point 1 to set-point 3 on Fig. 1, for example) then there will be a reduction in A.
View larger version (22K):
[in this window]
[in a new window]
|
Fig. 1. A schematic illustrating the dependence of ci on the relationship between stomatal conductance and photosynthetic capacity. In this "A/ci" plot, the two curved lines rising from near the origin represent the dependence of A on ci as determined by leaf gas-exchange measurements taken over a range of external CO2 concentrations. Variation in the initial slope of these curves reflects variation in photosynthetic capacity. The two curved lines are intersected by two straight lines originating at the ambient CO2 concentration, ca. The slope of these straight lines is the stomatal conductance to CO2, gc. The intersections indicated by numerals represent the "operating" values of ci (and of A, gc, and ci/ca) for three genotypes of a C3 species. The data shown here were derived from leaf gas-exchange measurements on three wheat varieties reported in Condon et al. (1990). (Adapted from Condon and Hall, 1997.)
|
|
Apart from a decrease in A, there is likely to be another penalty associated with a decrease in g. Unless the leaf boundary layer conductance is very large then leaf temperature and wi will increase as g decreases, and wa will decrease. These changes in wi and wa will cause an increase in the gradient driving transpiration and therefore an increase in T per unit g. The net result is that the anticipated increase in WT as g decreases will not be as great as predicted from Eq. [4]. Apart from these considerations, there are several more complications in scaling any potential improvement in WT into realized improvements in crop yield. Before expanding further on some of these issues the relationship between WT and carbon isotope discrimination will be examined.
|
Carbon Isotope Discrimination
|
|---|
Approximately 1% of the carbon in the biosphere is in the form of the stable isotope 13C. Two sorts of nomenclature appear in the literature to describe the 13C content of plant material. One is carbon isotope composition, the other is carbon isotope discrimination. Carbon isotope composition (
13C) is a measure of the 13C/12C ratio in a sample of plant relative to the value of the same ratio in an accepted international standard, the limestone Pee Dee belemnite. Thus,
where Rp is the 13C/12C ratio measured in plant material and Rs is the ratio of the standard and 
is per mil. Carbon isotope composition provides a means of relating samples of diverse origin for carbon isotope content. Samples of contemporary plant material have negative values of 13C because the 13C/12C ratio in the atmosphere is less than that in Pee Dee belemnite and because there is a net discrimination against 13C by plants during uptake and fixation of CO2 into plant dry matter.
Carbon isotope discrimination (13C) is a measure of the 13C/12C ratio in plant material relative to the value of the same ratio in the air on which plants feed (Farquhar and Richards, 1984). Thus,
 | [6] |
where Ra is the 13C/12C ratio of the atmosphere. Carbon isotope discrimination has positive values, which reflects the fact that C3 plants actively discriminate against 13C during photosynthesis.
Theory first published by Farquhar et al. (1982) and elaborated on since then indicates that the ratio of 13C/12C in dry matter of C3 plants is the result of discrimination against 13C during several processes. These processes are encapsulated in Eq. [7] and include discrimination that occurs during diffusion of CO2 through the stomata (a); discrimination by ‘Rubisco’ during the process of carboxylation of CO2 into the first products of photosynthesis (b); and some downstream fractionations associated with subsequent metabolism and (possibly) respiration (d). Thus,
 | [7] |
where a, b and d in Eq. [7] account for the fractionations described above. A useful empirical equation describing variation in 13C measured in plant dry matter is given by (Farquhar and Richards, 1984)
Equations [7] and [8] show that 13C is positively related to the ci/ca ratio. Earlier it was noted that WT should be negatively related to ci/ca (Eq. [4]). Therefore 13C and WT should also be negatively related. Since the first study on pot-grown bread wheat (Farquhar and Richards, 1984) there have been numerous studies that indicate that this negative association between 13C and WT is robust for plants of many C3 species. A short list of crop species for which this negative association has been found includes bread wheat (Condon et al., 1990; Ehdaie et al., 1991), barley (Hubick and Farquhar, 1989), peanut (Arachis spp.) (Hubick et al., 1986), common bean (Phaseolus vulgaris L.) (Ehleringer et al., 1991), cowpea (Vigna unguiculata L.) (Ismail and Hall, 1992), sunflower (Helianthus annuus L.) (Virgona et al., 1990), and chickpea (Cicer arietinum L.) (Udayakumar et al., 1998). The theory proposed by Farquhar et al. (1982) relating 13C to WT is therefore well established at both the leaf and whole-plant levels. Consequently, 13C, due to its convenience and relatively cheap cost, has become a useful indicator of differences in WT. Because plants retain much of the C they fix, another attractive feature of the measurement of 13C is that it can provide a time- and spatially-integrated estimate of relative variation in WT.
|
Issues in "Scaling-up" to Crop Yield
|
|---|
Despite its apparent utility, the measurement of 13C has several shortcomings as an indicator of crop performance. For example, measuring 13C provides no information on the magnitudes of either A or T or whether variation in 13C is being driven by variation in stomatal conductance or photosynthetic capacity. Perhaps more importantly, 13C provides no information on variation in (wi - wa) or the impact of this gradient in water-vapor concentration on the water-use efficiency of leaf gas-exchange. Both seasonal and diurnal variation in (wi - wa) may have more influence on water-use efficiency at the crop scale than variation in intrinsic water-use efficiency (Tanner and Sinclair, 1983).
The shortcomings of WT (and of 13C) are most apparent when g is the main source of variation in WT and when water supply does not impose a major limitation on crop growth. To illustrate these shortcomings, the data shown in Table 1 and Fig. 2
compare the performance of two wheat varieties grown in "field-scale" plots of 1.5 to 5 ha in two environments in eastern Australia. The two varieties were matched for height and flowering time but, compared with the variety Matong, the variety Quarrion has lower g (by about one third), lower 13C (by about 2) and higher WT (by about one third under well-watered conditions). The two environments contrasted in the extent of water limitation. Both were rain-fed but at Wagga Wagga in 1989 regular rainfall events occurred throughout the season, including well into grain filling. At Condobolin in 1990 the crops experienced an extended terminal drought with much of the water used for crop growth being stored in the soil profile from rains in the previous summer.
View this table:
[in this window]
[in a new window]
|
Table 1. A comparison of water use (as evapotranspiration, ET and transpiration, T), productivity (as above-ground biomass and grain yield) and water-use efficiency (ratios of productivity and water use) at two sites for two wheat cultivars differing in intrinsic water-use efficiency, WT, stomatal conductance, g, and carbon isotope discrimination, 13C.
|
|
View larger version (25K):
[in this window]
[in a new window]
|
Fig. 2. Patterns of soil water extraction at two locations under two wheat genotypes, Quarrion and Matong, that differ in WT because of a large difference in g.
|
|
In both environments the lower g of Quarrion was associated with slower crop growth rates and slower soil water depletion during the middle part of the growing season from about the start of stem elongation (Fig. 2). At the drier Condobolin site both varieties used close to the same amount of water. Transpiration was a little greater for Matong but this was insufficient to compensate for its relatively poor WT. The "conservative" gas-exchange of Quarrion associated with lower g and higher WT resulted in greater biomass production and grain yield at Condobolin. At Wagga Wagga the position was very different. At this site biomass production and grain yield were both less for the high WT, low g variety Quarrion. Total ET was also lower for Quarrion. Quarrion left 24 mm of water behind in the profile that was not left behind by Matong. The difference in T was even greater (58 mm). This was because soil evaporation made up a greater component of total ET for Quarrion. The slower crop growth rate of Quarrion meant that the frequently-rewetted soil surface was more exposed, resulting in greater evaporation of water directly from the soil surface.
This study emphasized the potential penalties that are likely to arise in favorable environments when higher WT is the result of lower g. But it also demonstrated the potential advantages that may arise from higher WT in drier, "stored-moisture" environments. As well, the study demonstrated that a difference in WT resulting from a difference in g does translate into a difference in crop-scale transpiration efficiency. There are sound reasons why this might not have been the case, particularly at the more favorable Wagga Wagga site. In the theory summarized by Eq. [4], the leaf-to-air vapor pressure difference (wi - wa) is assumed to be an independent variable. But this is unlikely to be the case outside "well-stirred" gas-exchange cuvettes. Leaf temperature and therefore wi will tend to be greater for genotypes with relatively low g unless the air around the leaves is very well stirred. Higher wi will drive transpiration faster per unit g than would be expected if differences in g had no effect on leaf temperature at all. The phenomenon will be compounded for an extensive canopy of a crop with lower g for which, unless stirring is very thorough, the surrounding air will be hotter and drier (i.e., wa will be less), further reducing any potential gain in WT. These problems of scaling the influence of stomata on canopy gas-exchange are associated with the impacts of unstirred "boundary layers" around both the individual leaves and the canopy as a whole (Jones, 1976; Cowan, 1988; Farquhar et al., 1988). These impacts will tend to be greater the smoother and more extensive the canopy, the higher the temperature, the lower the wind speed and at high levels of nutrition and available soil water (i.e., when boundary layer resistances are relatively large compared with crop resistances to water loss and carbon gain). Both Jones (1976) and Cowan (1988) developed mathematical descriptions of cereal canopy gas-exchange which indicated that there may be some circumstances where there would be no gain in crop transpiration efficiency despite lower g. In the study of Quarrion and Matong these "boundary layer" effects were important but not sufficient to override the difference in WT between the two genotypes. At both sites biomass/T (i.e., above-ground biomass per unit transpiration, a crop-scale measure of transpiration efficiency) was 15% greater for Quarrion than for Matong (Table 1). This difference was about half the difference in WT that might have been expected on the basis of the initial difference in 13C between the two varieties. For these well-managed sites the average water-use efficiency of grain production (grain yield/T) for Matong was 20.8 kg ha-1 mm-water-transpired-1. This value is very close to the value of 20 kg ha-1 mm-water-transpired-1 commonly used as the "potential" water-use efficiency in agronomically based studies in southern Australia (Angus and van Herwaarden, 2001). The average value for Quarrion, with low g, high WT and low 13C, was about 15% greater (24.2 kg ha-1 mm-water-transpired-1).
We took several lessons from this study. First, differences in ‘intrinsic water-use efficiency’ do translate to the field scale, although the potential gain in crop transpiration efficiency was effectively halved in these experiments and could be reduced further in even more productive environments where the effects of boundary layer may be greater. Second, differences in crop transpiration efficiency could be masked by differences in the partitioning of crop water use between evaporation from the soil surface and transpiration from plants such that differences in total crop water-use efficiency may become small or insignificant. Third, in favorable environments, the "conservative" gas-exchange associated with low g can result in considerable yield reductions as well as less water use. But, fourth, in drier "stored-moisture" environments, the conservative gas-exchange associated with low g and high WT can result in a considerable yield gain.
|
Carbon Isotope Discrimination and Yield in Wheat and Barley in Water-Limited Rainfed Environments
|
|---|
There have now been numerous field studies that have examined the association between yield and 13C in wheat and barley in rainfed environments (Condon et al, 1987; Acevedo et al, 1993; Condon and Richards, 1993; Condon et al, 1993; Lopez-Casteneda and Richards, 1994; Condon and Hall, 1997; Araus et al., 1998; Merah et al., 1999; Voltas et al., 1999). In these studies a range of bread wheat, durum wheat or barley genotypes or breeding lines have been grown that vary in 13C and usually in other characteristics such as flowering time and height. The majority of these studies have been conducted in mediterranean or similar environments where there has been strong reliance on "within-season" rainfall, but in some "stored-moisture" environments there has been a strong reliance on subsoil moisture from out-of-season rains. Most of the associations between grain yield and 13C reported from these studies have been either positive or "neutral". Even in dry environments there have been few reports of the negative associations between yield and 13C that might have been expected if WT was having a major influence on productivity. If WT was going to influence productivity then this should be reflected most directly in biomass production. For biomass too, the majority of associations reported have been either positive or neutral (Condon et al., 1987; Ehdaie et al., 1991; Condon and Richards, 1993; Condon et al., 1993; Lopez-Casteneda and Richards, 1994). Examples of the sorts of relationships that have been observed are given in Fig. 3
.
The impact of variation in 13C on the seasonal course of crop growth and water use has also been examined in some studies (Condon and Richards, 1992; Condon et al., 1993; Lopez-Casteneda and Richards, 1994). The results of these studies indicate that there are at least four factors that could account for the dominance of positive associations between grain yield and 13C observed in mediterranean and similar "within-season" rainfall environments. First, low 13C is associated with slower crop growth rate in the absence of water stress. Second, faster crop growth may not be accompanied by a correspondingly faster rate of soil water depletion. Third, in several studies a positive association has been found between harvest index and 13C. Fourth, the relationship between yield and 13C may vary depending on the plant organ used for the measurement of 13C.
The slower crop growth rate associated with low 13C in the absence of water stress is readily explained if variation in 13C is the result of variation in g, as found in the earlier comparison of Quarrion and Matong. But in cereals variation in 13C arises from variation in photosynthetic capacity as well as g (Fig. 1) (Condon et al., 1990; Morgan and LeCain, 1991). An increase in photosynthetic capacity is most often achieved in cereals by the concentration of leaf nitrogen into smaller leaves with low specific leaf area. The result is a reduction in the rate of leaf area development during early crop growth. This results in less light interception, lower canopy photosynthesis and slower dry matter production (Condon et al., 1993; Lopez-Casteneda et al., 1995). Thus, if low 13C is due to greater photosynthetic capacity per unit leaf and full light interception is maintained for only a relatively short period, the outcome may be a negative association between leaf-level A and crop biomass, and a positive association between biomass production and 13C.
A faster rate of crop growth should be accompanied by greater crop transpiration but Condon et al. (1993) and Lopez-Casteneda and Richards (1994) found that soil water depletion associated with high dry matter production at anthesis by high 13C genotypes was not as great as might have been expected. Transpiration did tend to be greater for high 13C genotypes but a substantial proportion of the extra water consumed in the production of greater biomass by these genotypes was not conserved by low 13C genotypes. There was greater evaporation from the frequently re-wetted soil surface under the slower-developing canopies of low 13C genotypes. As well, much of the extra water transpired by high 13C genotypes was consumed early in the season during the cool winter and early spring period when crop WT was high due to the low evaporative demand.
Associations between grain yield and have tended to be "more positive" than those between biomass and 13C. This is because, where biomass and grain yield have both been measured, harvest index and 13C have also tended to be positively related. In a field study on a range of wheat genotypes reported by Condon et al. (1987), the relative increase in biomass per mil 13C was 24% and for grain yield it was 35%. In a multi-site study on biomass and 13C reported by Condon and Richards (1993), relationships between grain yield and 13C were all positive except for the single site where the crop was strongly reliant on stored subsoil moisture from out-of-season rains. At this site no association between grain yield and 13C was found (Condon and Richards, 1996, unpublished data) but there was a negative association between biomass and 13C. The field data presented by Ehdaie et al. (1991), Morgan et al. (1993) and Lopez-Casteneda and Richards (1994) also indicate positive associations between harvest index and 13C.
Several factors could account for the positive associations between harvest index and 13C often observed with cereals. One may be phenology. Strong negative associations have been found between 13C and days to heading or anthesis in several of the studies conducted with cereals in mediterranean, terminal-drought environments, with low genotypes flowering many days later than high 13C genotypes (Ehdaie et al., 1991; Craufurd et al., 1991; Acevedo, 1993; Sayre et al., 1995). In terminal-drought environments late-flowering genotypes are more likely to have a lower harvest index because seed-set and grain-filling are more likely to occur under conditions of greater evaporative demand and lower available soil water. The positive association between harvest index and 13C may be further strengthened if 13C is measured on organs that expand near the time of seed-set, such as the flag-leaf, peduncle or grain, since 13C also tends to be lower when growing conditions are less favorable (Araus et al., 1997; Condon et al., 1993; Farquhar et al., 1989; Merah et al., 1999). However, in some studies where there has been little variation in flowering date (Fig. 3) (Condon et al., 1987; Morgan et al., 1993; Acevedo, 1993) or where variation in flowering date has been accounted for statistically (Sayre et al., 1995), grain yield and 13C have still been positively associated.
Another factor likely to contribute to the positive association between harvest index and 13C in cereals is the use of stored assimilate to sustain grain growth. This stored assimilate is laid down in the stems and other organs before and shortly after flowering. The dependence on stored assimilate for grain filling increases as the degree of post-anthesis water stress increases. The contribution of pre-anthesis assimilate to grain filling may vary from about 10% of final grain carbon when there is little water limitation to as much as 80% of final grain carbon under severe water stress (Palta and Fillery, 1995). It seems likely that among cereal genotypes growing in terminal-drought environments the contribution of stored assimilate to grain growth will be greater for high 13C genotypes that have produced more dry matter and used more of the soil water store at anthesis. As an example, for the wheat genotypes represented in Fig. 4
there was a strong positive association between shoot biomass production at anthesis and 13C. Both harvest index and the ratio of grain yield to post-anthesis growth were also positively correlated with 13C. Grain yield increased 16% per per mil 13C compared with a 6% increase for final biomass.
View larger version (19K):
[in this window]
[in a new window]
|
Fig. 4. Relationships between final grain yield and above-ground dry matter production for the periods from (a) sowing to anthesis and (b) anthesis to physiological maturity, for a collection of wheat genotypes grown under rain-fed conditions at Moombooldool, south-eastern Australia, in 1986. (Adapted from Condon and Hall, 1997.)
|
|
Positive associations between harvest index and 13C have also been observed in some studies in stored-moisture environments. But in these environments and in drier current-rainfall environments associations between yield and 13C have tended to be either negative or neutral. It is more likely that a faster rate of crop growth will carry a greater risk of depleting soil water in these environments because transpired water is not being replaced by frequent rainfall events. As well, soil evaporation will tend to be low and it should be similar for both fast and slow-growing canopies because the soil surface is dry for long periods. The conservative growth and water use associated with low 13C is more likely to be an advantage for cereals growing in these sorts of environments.
|
Simulating the Likely Impact of Breeding to Improve Intrinsic Water-Use Efficiency
|
|---|
The variability observed in relationships between 13C and crop yield indicates that selecting or breeding for improved WT may be desirable for some cropping environments but inappropriate for others. One way of resolving this conundrum may be to use crop growth models to simulate, for different environments, the likely impact on yield of breeding for improved WT. This has been done for wheat in eastern Australia. The SIMTAG wheat crop growth model (Stapper and Harris, 1989) was parameterized to simulate the effects of an increase in intrinsic WT of 25% (i.e., a reduction in 13C of about 1.5) but at the cost of a 10% reduction in crop growth rate in the absence of soil water stress. The simulated changes in 13C and WT were a little less than the differences in these parameters between the varieties Quarrion and Matong described earlier, and are consistent with gas-exchange data from a range of wheat genotypes published in Condon et al. (1990) and summarized in Fig. 1. The reduction in growth rate in the absence of soil water stress was required to generate variation in final biomass production under well-watered conditions similar to that published in Condon et al. (1987) and as shown here in Fig. 3 for the wet 1993 season at Condobolin and for Wagga Wagga in 1989 (Table 1). The model was run with and without these changes for two environments in eastern Australia with average annual rainfall of 500 to 600 mm. In each environment ‘spring’ wheat is typically sown just before the start of winter but the seasonal pattern of rainfall varies from summer-dominant in the sub-tropical north to winter-dominant in the temperate south. In the north, crops have a strong reliance on reserves of soil moisture stored from the summer rains. In the south there is often little reliance on soil-water reserves stored from out-of-season rainfall. In both environments year-to-year variation in rainfall can be large.
The outcome of this simulated change in WT is summarized in the left-hand histograms in Fig. 5
. The results indicate that for the southern zone ‘breeding’ to improve WT would have had little impact on average crop yield. But the northern cropping zone appears a much more promising target environment. Here, simulated yields were raised by about 11% on average, despite the impact of reduced crop growth rate in the absence of soil water stress that accompanied the modelled 25% improvement in intrinsic WT.
View larger version (28K):
[in this window]
[in a new window]
|
Fig. 5. Average simulated yield impact for two locations in eastern Australian of breeding to improve WT, early vigor and a combination of WT and early vigor in spring wheat. Simulations were done with the SIMTAG wheat crop growth model and 25–30 years of historical weather data for each location (Condon and Stapper, 1995, unpublished).
|
|
|
A Backcross Breeding Program to Improve Intrinsic Water-Use Efficiency of Wheat
|
|---|
On the basis of the outcomes of this simulation study and the results of our previous field studies of crop growth and water use in stored-moisture environments, we embarked on a backcross breeding program to improve the intrinsic water-use efficiency of the variety Hartog. This variety is grown extensively in Australia's northern wheat-growing region. Details of the breeding and selection process are given in Rebetzke et al. (2002). To test the impact of the process, 30 random low 13C and 30 random high 13C BC2 lines with similar height and phenology were grown from 1995 to 1998 in eight environments in eastern Australia varying in the extent to which crops relied on stored soil moisture. The lines also were grown in five environments in ‘mediterranean’ Western Australia, where crops grow almost entirely on within-season rainfall. The results of this comparison of low 13C and high 13C backcross lines are summarized in Fig. 6
. The figure indicates, for each environment, the relative yield advantage of the low 13C, high WT set of lines over the high 13C, low WT set. In no environment was the average yield of the low 13C set less than the average yield of the high 13C set, even at site-mean yields of five or six t/ha. In eastern Australia the relative yield advantage of the low 13C set became greater as the environment-mean yield declined and as the reliance on stored soil moisture increased. For the sites in Western Australia the range in yields was small among environments and average yields were low. Nonetheless, even in these ‘mediterranean’ environments, the relative yield advantage of the low 13C set averaged 3 to 4%.
The results of this comparison of backcross lines differing in 13C are consistent with the outcomes of the simulation exercise described above but are somewhat at odds with the results of many of the field trials with wheat and barley summarized earlier. In particular, the low 13C backcross lines were not lower-yielding than the high 13C lines at sites with high mean yields. There may be several reasons for this. For these backcross lines there was no indication that low 13C was associated with a slower crop growth rate in the absence of soil water stress. This outcome may be peculiar to the parents used here, although it may also be that the backcrossing process acts to restore genes or gene combinations from the recurrent parent that are associated with high yield under well-watered conditions and that are in some way independent of variation in WT. Seasonal rainfall distribution may also have played some role. The backcross lines were grown at sites in ‘mediterranean’ Western Australia with winter-dominant rainfall and at sites in eastern Australia which, on the basis of long-term averages, have an even distribution of rainfall throughout the year. But at most of the site/years of these trials the 3- to 4-wk period before anthesis was drier than average. At lower-yielding sites this drier pattern continued into grain-filling, promoting a strong reliance on subsoil moisture stored from earlier rains. This reliance on stored moisture may have favored the more "conservative" low 13C lines.
This backcross breeding program also has a commercial purpose. Low 13C backcross lines have undergone extensive field testing in Queensland, Australia, with several showing consistent and substantial yield advantage over the recurrent parent Hartog (Banks, unpublished data, 1996–2000). Some of these backcross lines have now completed the final phases of multi-environment testing as potential new varieties for the northern wheat-belt of Australia.
|
Other Opportunities to Improve Crop Yield by Breeding for Intrinsic Water-Use Efficiency
|
|---|
The backcross breeding program with wheat for stored-moisture environments in Australia indicates that there are circumstances where breeding for improved WT can lead to consistent gains in crop yield. Are there other circumstances? The most likely prospects are similar "stored-moisture" environments where the slower crop growth rate in the absence of soil water stress usually associated with improved WT does not limit grain yield, i.e. environments where within-season rainfall makes up a relatively small proportion of the total water available for crop growth and where stored soil moisture needs to be metered out from relatively early in the cropping season so as to sustain seed set and seed growth. Dryland winter or spring cropping environments in temperate North America and Asia and in sub-tropical regions of South Asia, South America and Africa may fall into this category.
High WT is not always associated with a slower crop growth rate. For peanut, field studies in both well-watered and water-limited environments consistently show greater biomass production to be associated with higher WT (Wright et al., 1993). For this species variation in photosynthetic capacity accounts for most of the variation in WT (Nageswara Rao et al., 1995) but importantly, high photosynthetic capacity is associated with a faster rate of leaf area growth. This is the reverse of the situation observed in many species such as cereals and cotton. It appears that for non-legume species greater N per unit leaf area is most commonly achieved through a reduction in leaf size (Bhagsari and Brown, 1986), perhaps reflecting the fact that N enters these plants from finite soil sources. Symbiotic N fixation may allow peanut and perhaps other legumes to concentrate leaf N without sacrificing leaf size and rapid leaf area growth. In contrast to peanuts, there are other legume species for which most variation in 13C appears to be related to variation in g rather than photosynthetic capacity (Ehleringer et al., 1991; Udayakumar et al., 1999). For legumes, identifying those species or genotypes in which high WT arises mainly from high photosynthetic capacity may accelerate progress in improving crop yield via WT. This could be done by combining measurements of 13C with measurements of g and/or photosynthetic capacity. Techniques are available for detecting genotypic variation in g directly and rapidly with viscous-flow porometers (Rebetzke et al., 2001) or indirectly with oxygen isotope composition of dry matter (Farquhar et al., 1994) or canopy temperature (Fischer et al., 1998). Measurements of specific leaf area (Nageswara Rao et al., 1995) or leaf chlorophyll concentration (Araus et al., 1997) may be effective means of characterizing variation in photosynthetic capacity in some species.
For crop species such as cereals for which the relationship between crop growth rate in the absence of water stress and 13C tends to be positive there may well be circumstances where it would be reasonable to select against improved WT. This might be the case for irrigated environments where the reduction in growth rate associated with high WT outweighs any water savings that might be gained. The effects of ‘scale’ discussed earlier associated with leaf and canopy boundary layers are likely to be important considerations in this sort of environment (Cowan, 1988; Farquhar et al., 1989). Selection for high 13C may be useful in irrigated rice (Fig. 7)
, wheat (Fischer et al., 1998), and other species in which variation in WT is largely driven by variation in g.
View larger version (14K):
[in this window]
[in a new window]
|
Fig. 7. The relationship between grain yield and 13C of grain for advanced breeding lines of rice grown under flood-irrigation at Yanco, south-eastern Australia, 1992 (Condon and Lewin, 1993, unpublished).
|
|
For cereals such as wheat or barley grown in within-season rainfall environments such as in the mediterranean zone, the faster growth usually associated with high 13C also is associated with higher yields in all but the driest locations (Araus et al., 1998; Merah et al., 1999; Voltas et al., 1999). The 13C of grain has been measured in many of these studies. Grain 13C is a more cryptic measurement than 13C measured on unstressed leaves. High grain 13C values may reflect faster growth rate associated with high 13C values throughout crop development, although frequently the correlation between grain and leaf 13C values has been found to be poor (Condon et al., 1992; Merah et al., 1999). Additionally, high grain 13C values may reflect greater reliance for grain filling on pre-anthesis stem reserves laid down when plants were less stressed and 13C values were much higher (Fig. 4). Mean values for grain 13C tend to be several per mil lower than mean values for leaves sampled under well-watered conditions (Condon et al., 1992; Condon et al., 1993; Merah et al., 1999). The lower mean values for grain 13C largely reflects stomatal closure in response to post-anthesis water deficit. Within an environment, high grain 13C may indicate greater access to soil moisture due to more extensive rooting (White et al., 1990) or earlier flowering or it may reflect the ability to maintain stomata more open after anthesis despite increasing soil and atmospheric water stress (Condon et al., 1993). Whatever the cause of high grain 13C, any of these characteristics would be useful in mediterranean environments.
In some circumstances it may be useful to combine high WT with other traits to obtain a yield benefit. One such trait is early vigor. The importance of early vigor for cereals growing in "within-season" rainfall environments is widely recognized and our group has initiated a breeding program to address this issue for wheat in southern Australia (Richards et al., 2002). But greater early vigor may risk early exhaustion of soil water reserves in low rainfall areas or if rainfall is less than normal in higher rainfall zones. We have conducted additional simulations using the SIMTAG wheat growth model in which both WT and early vigor have been ‘manipulated’. The effect on simulated average yields of an improvement in each trait separately and in combination is shown for two environments in Fig. 5. The procedure for conducting the simulations for modifying WT were described in an earlier section. For simulating an increase in early vigor, the model was parameterized by doubling the size of the first leaf. For cereals this can be achieved through an increase in embryo size combined with higher specific leaf area (Lopez-Castaneda et al., 1995). In practice, to achieve a doubling of early leaf area in wheats with semi-dwarf stature it is likely that GA-sensitive dwarfing genes would also need to be used (Richards et al., 2002), since they allow much better expression of traits contributing to early vigor. For simulating a combination of high WT and greater early vigor the parameterization described earlier for an increase in WT was used but first-leaf size was increased by only 85%, i.e., we assumed some restriction on our ability to maximize the expression of early growth when combined with high WT.
The simulations provided strong support for the notion of improving early vigor for southern Australia. For improved early vigor alone the simulated yield impact was almost a mirror image of that for improved WT alone, with the greatest yield gain in the southern region but little impact in the northern zone (central histograms in Fig. 5). Interestingly, the simulations indicated that combining the two traits of high WT and high early vigor would result in a synergistic effect in both cropping regions (right-hand histograms in Fig. 5). We have initiated breeding to combine these two traits so as to exploit WT in a wider range of environments in Australia. This may not be straightforward. Some proportion of early vigor in cereals can be attributed to high specific leaf area, which may contribute to higher values of 13C. Breeding for greater embryo size and the use of GA-sensitive dwarfing genes offer much greater opportunity to increase early vigor (Richards et al., 2000) with less potential impact on 13C.
|
CONCLUSION
|
|---|
Tanner and Sinclair (1983) noted that improved WT would have limited impact on crop yield for many environments unless it was associated with a faster crop growth rate. This review provides support to this conclusion, especially for those environments where water is relatively plentiful for much or all of the cropping cycle. The available data indicate that in cereals high WT is associated with conservative water use but also conservative growth in the absence of water stress. However, there are strong indications that high WT can be used to improve cereal yields in stored-moisture environments. These indications come from using a crop growth model to simulate the effects of breeding for increased WT and from the yield gains achieved in stored-moisture environments in Australia by back-crossing high WT into a widely-grown recurrent parent. Additional simulations indicate that it is likely that high WT may be effective in raising yields in other rainfed environments if it is combined with other traits such as greater early vigor. A combination of high WT and greater early vigor may not be easy to achieve in practice, but there are strong indications from our breeding program that such a combination is achievable and would be effective in many parts of Australia and other rainfed environments where cereals are grown.
For more predictable, favorable environments any benefits associated with high WT are likely to be outweighed by the cost of a relatively slow growth rate, at least for cereals. Thus selection against high WT may be an effective means of raising yield levels for irrigated production of wheat and rice, for example. For other crop species such as peanut, there are strong indications that high crop growth rate and high WT are more compatible processes and that consistent gains in crop yield may be achieved by breeding for high WT. More effort is needed to identify those other crop species or genotypes for which there is a positive association between photosynthetic capacity and crop growth rate. By whatever means it is achieved, it is clear that breaking the nexus between high WT and low crop growth rate in the absence of water stress remains the key to achieving more widespread gains in crop yield from greater WT.
|
NOTES
|
|---|
Presented at the 1999 CSSA Symposium on Water Use Efficiency, organized by Div. C-2 chair, Dr. Tom Gerik.
Received for publication September 19, 2000.
|
REFERENCES
|
|---|
- Acevedo, E. 1993. Potential of carbon isotope discrimination as a selection criterion in barley breeding. p. 399–417. In J.R. Ehleringer et al. (ed.) Stable Isotopes and Plant Carbon-Water Relations. Academic Press, San Diego, CA.
- Angus, J.F., and A.F. van Herwaarden. 2001. Increasing Water Use and Water Use Efficiency in Dryland Wheat. Agron. J. 93:290–298.[Abstract/Free Full Text]
- Araus, J.L., T. Amaro, J. Casadesus, A. Asbati, and M.M. Nachit. 1998. Relationships between ash content, carbon isotope discrimination and yield in durum wheat. Aust. J. Plant Physiol. 25:835–842.[ISI]
- Araus, J.L., J. Bort, S. Ceccarelli, and S. Grando. 1997. Relationships between leaf structure and carbon isotope discrimination in field grown barley. Plant Physiol. and Biochem. 35:553–541.
- Bhagsari, A.S., and R.H. Brown. 1986. Leaf photosynthesis and its correlation with leaf area. Crop Sci. 26:127–132.[Abstract/Free Full Text]
- Condon, A.G., G.D. Farquhar, and R.A. Richards. 1990. Genotypic variation in carbon isotope discrimination and transpiration efficiency in wheat. Leaf gas exchange and whole plant studies. Aust. J. Plant Physiol. 17:9–22.
- Condon, A.G., and A.E. Hall. 1997. Adaptation to diverse environments: genotypic variation in water-use efficiency within crop species. p. 79–116. In L.E. Jackson (ed.) Agricultural Ecology. Academic Press, San Diego, CA.
- Condon, A.G., and R.A. Richards. 1993. Exploiting genetic variation in transpiration efficiency in wheat: An agronomic view. p. 435–450. In J.R. Ehleringer et al. (ed.) Stable Isotopes and Plant Carbon-Water Relations. Academic Press, San Diego, CA.
- Condon, A.G., R.A. Richards, and G.D. Farquhar. 1987. Carbon isotope discrimination is positively correlated with grain yield and dry matter production in field-grown wheat. Crop Sci. 27:996–1001.[Abstract/Free Full Text]
- Condon, A.G., R.A. Richards, and G.D. Farquhar. 1993. Relationships between carbon isotope discrimination, water use efficiency and transpiration efficiency for dryland wheat. Aust. J. Agric. Res. 44:1693–1711.[ISI]
- Cowan, I.R. 1988. Stomatal physiology and gas exchange in the field. p. 160–172 In W.L. Steffen and O.T. Denmead (ed.) Flow and Transport in the Natural Environment: Advances and Applications. Springer-Verlag, Berlin.
- Craufurd, P.Q., R.B. Austin, E. Acevedo, and M.A. Hall. 1991. Carbon isotope discrimination and grain yield in barley. Field Crops Res. 27:301–313.
- Ehdaie, B., A.E. Hall, G.D. Farquhar, H.T. Nguyen, and J.G. Waines. 1991. Water-use efficiency and carbon isotope discrimination in wheat. Crop Sci. 31:1282–1288.[Abstract/Free Full Text]
- Ehleringer, J.R., S. Klassen, C. Clayton, D. Sherrill, M. Fuller-Holbrook, Q.A. Fu, and T.A. Cooper. 1991. Carbon isotope discrimination and transpiration efficiency in common bean. Crop Sci. 31: 1611–1615.[Abstract/Free Full Text]
- Farquhar, G.D., A.G. Condon, and J.F. Masle. 1994. Use of carbon and oxygen isotope composition and mineral ash content in breeding for improved rice production under favorable, irrigated conditions. p. 95–101. In K.G. Cassman (ed.) Breaking the yield barrier: Proceedings of a workshop on rice yield potential in favorable environments, IRRI, 29 Nov.–4 Dec. 1993. IRRI, Los Baños, Philippines.
- Farquhar, G.D., J.R. Ehleringer, and K.T. Hubick. 1989. Carbon isotope discrimination and photosynthesis. Ann. Rev. Plant Physiol. Plant Mol. Biol. 40:503–537.[ISI]
- Farquhar, G.D., K.T. Hubick, A.G. Condon, and R.A. Richards. 1988. Carbon isotope fractionation and plant water-use efficiency. p. 21–40. In P.W. Rundel et al. (ed.) Stable Isotopes in Ecological Research. Springer-Verlag, New York.
- Farquhar, G.D., M.H. O'Leary, and J.A. Berry. 1982. On the relationship between carbon isotope discrimination and the intercellular carbon dioxide concentration in leaves. Aust. J. Plant Physiol. 9:121–137.[ISI]
- Farquhar, G.D., and R.A. Richards. 1984. Isotopic composition of plant carbon correlates with water-use efficiency of wheat genotypes. Aust. J. Plant Physiol. 11:539–552.[ISI]
- Fischer, R.A., D. Rees, K.D. Sayre, Z. Lu, A.G. Condon, and A. Larque Saavedra. 1998. Wheat yield progress is associated with higher stomatal conductance, higher
TOP
ABSTRACT
INTRODUCTION
