Agronomy Journal 95:10-19 (2003)
© 2003 American Society of Agronomy
SYMPOSIUM PAPERS
Evaluating the Impact of a Trait for Increased Specific Leaf Area on Wheat Yields Using a Crop Simulation Model
Senthold Asseng*,a,
Neil C. Turnera,
Tina Botwrighta and
Anthony G. Condonb
a CSIRO Plant Industry, Private Bag no. 5, Wembley, WA 6913, Australia
b CSIRO Plant Industry, P.O. Box 1600, Canberra, ACT 2601, Australia
* Corresponding author (Senthold.Asseng{at}csiro.au)
Received for publication May 1, 2001.
 |
ABSTRACT
|
|---|
Early vigorous growth (also called early vigor) has been suggested as a characteristic to increase yields of wheat (Triticum aestivum L.) in rainfed environments. Breeders have tried to incorporate early vigor by selecting for increased specific leaf area (SLA). A crop simulation model (APSIM-Nwheat) was used to evaluate the role of SLA on potential yield. Increased early vigor was simulated by increasing the SLA trait in the model. Grain yields were simulated for the mediterranean climate of Western Australia with winter-dominant rainfall, for New South Wales with evenly distributed rainfall, and for Queensland with a subtropical climate with summer-dominant rainfall. Independent, multiseason simulations were performed with historical weather records. The largest average yield increase with greater SLA was 15% on the sandy soil in Western Australia but only with additional N inputs. In the low-rainfall region of Western Australia, with low N input, a yield increase of up to 5% was achieved on the sandy soil while yield responses were negative on the clay soil. With increased N application, the SLA trait increased yields by up to 7% on the clay soils in this region. The increased SLA trait failed to increase yields and even decreased yields on average in New South Wales and in Queensland, except in above-average rainfall seasons and only at high N application. Crop management, in particular crop nutrition, is an important factor determining yield expression of new physiological traits and needs to be considered when evaluating traits.
Abbreviations: DOY, day of year • ET, evapotranspiration • LAI, leaf area index • PAW, potential plant available soil water-holding capacity in the potential maximum rooting zone • SLA, specific leaf area
|
INTRODUCTION
|
|---|
ABOUT TWO-THIRDS of the 12 million ha sown to spring wheat every year in Australia is grown in mediterranean-type climates (Anonymous, 2000). Wheat is sown at the break of season (first rainfall after summer) in autumn (May–June). In winter (July–August), vegetative growth is slow due to low temperatures and low solar radiation. Increasing temperatures and radiation in spring (September–October) allow high growth rates. However, grain filling during this time is often affected by high temperatures and declining rainfall that restricts the growing season to a 5- to 7-mo period and results in low biomass production and grain yields (Regan et al., 1992). For this environment, early vigorous growth has been suggested as a physiological characteristic to increase wheat grain yields (Turner and Nicolas, 1987, 1998; Whan et al., 1993).
Early vigorous growth (or early vigor), as a result of enhanced leaf area produced early in the season (Rebetzke and Richards, 1999; Turner and Nicolas, 1987), has been suggested to increase growth when temperatures and vapor pressure deficit are low, thereby increasing the transpiration efficiency of the crop (Fischer, 1979). Indeed, it has been shown that water use efficiency (with water use including both soil evaporation and crop transpiration) for grain yields increased by as much as 25% due to early vigor (Siddique et al., 1990b; López-Castañeda and Richards, 1994; Turner and Nicolas, 1998). Biomass weight at the five- to six-leaf stage, one measure of early vigor, was found to be positively related to grain yield under water-limited conditions and deep sandy soils in Western Australia (Turner and Nicolas, 1987, 1998).
However, Acevedo (1987), Siddique et al. (1990a)(1990b), and Damisch and Wiberg (1991) showed equivocal relationships between early vigor and grain yield in different environments while Whan et al. (1993) reported a positive relationship for low-rainfall sites and no relationship for high-rainfall sites in Western Australia. Regan et al. (1992) showed generally lower yields with early-vigor genotypes. This large range of yield responses to early vigor highlights the complex nature of interactions between early vigor and grain yields in various growing environments.
Plant growth is initially exponential so that a small increase in growth during the vegetative stage of development will often cause a considerable increase in biomass production and water use during later stages (Loss and Siddique, 1994). While the depth and amount of water extraction was greater in early-vigor genotypes (Turner and Nicolas, 1998), the early-vigor characteristic may be detrimental to grain yield if it depletes the water in the profile by anthesis (Richards and Townley-Smith, 1987), particularly on fine-textured or clay soils with a limited depth of soil wetting (Turner, 1997). Traditionally, wheat in the mediterranean environment of southwestern Australia is deliberately grown with limited N to restrict early growth and conserve water for later reproductive growth (Anderson et al., 1991). Increasing early vigor of wheat under N-deficit conditions will accelerate N stress if N supply is not adjusted to meet the increased N demand. Low N supply in some experiments in the literature (e.g., Regan et al., 1992) may have caused the lack of benefit from early-vigor lines.
One mechanism known to increase early vigor is to increase the ratio of leaf area to dry weight, the SLA. Increased SLA for improving early vigor has been successfully backcrossed into commercial cultivars by wheat breeders (Richards et al., 1998), and these lines are currently being tested in Australia. However, because the large climatic variability affects water availability and interacts with N supply during the growing season, the impact of the SLA trait and early vigor on grain yield is not clear under the range of environmental conditions found in Australia. As traditional field evaluation of advanced lines involves multilocation testing over a number of seasons, the impact of a particular trait on yield in very variable stress environments will require many site-by-year trials and may still not cover the full range of environments experienced by the genotypes (Shorter et al., 1991).
A comprehensive crop simulation model that takes into account the dynamics of the crop–soil–weather continuum and captures the principles inherent in such a system can assist in evaluating the role of a trait such as increased SLA on yield across a range of climates, soil types, and seasons (Cooper and Hammer, 1996). In addition, such models may facilitate the use of physiological understanding in interpreting genotype x environment x management interactions (Hammer et al., 1996). Crop simulation models have been shown to be effective tools in extrapolating agronomic research findings over time and space (e.g., Asseng et al., 1998a, 2001b; Fischer et al., 1990; O'Leary and Connor, 1998) and have been used to assess physiological traits in a range of environments (Stapper and Harris, 1989; Aggarwal et al., 1996; Hammer et al., 1996; Sinclair and Muchow, 2001).
This paper demonstrates how the APSIM-Nwheat model can aid plant breeders in interpreting of the effect of increased SLA, an early-vigor trait, and its interaction with N supply on the potential yield of wheat crops growing in different environments, soil types, and seasons.
|
MATERIALS AND METHODS
|
|---|
APSIM
The Agricultural Production Systems Simulator (APSIM) (McCown et al., 1996) for wheat (APSIM-Nwheat; Keating et al., 2001) is a crop simulation model consisting of modules that incorporate aspects of soil water, N, crop residues, crop growth and development, and their interactions within a wheat crop–soil system that is driven by daily weather data. The model calculates potential yield, which is the maximum yield reached by a crop in a given environment that is not limited by pests, diseases, weeds, nor lodging but is limited by temperature, solar radiation, water, and N supply. APSIM-Nwheat has been rigorously tested against field measurements in various studies under a large range of growing conditions (Asseng et al., 2000; Probert et al., 1995; 1998; Turpin et al., 1996) and in particular, in the mediterranean climatic regions of Western Australia (Asseng et al., 1998a, 1998b, 2001a, 2001b).
Parts of APSIM-Nwheat (Keating et al., 2001) have evolved from experiences in Australia with the CERES family of crop and soil models (Ritchie et al., 1985; Jones and Kiniry, 1986) and the PERFECT model (Littleboy et al., 1992), as modified by Probert et al. (1995)( 1998). The main differences between the APSIM-Nwheat model and the CERES-Wheat model are summarized by Probert et al. (1995) and Asseng et al. (1998b). Documented model source code in hypertext format can be viewed at www.apsim-help.tag.csiro.au (verified 23 Aug. 2002).
Leaf Area Growth in APSIM-Nwheat
APSIM-Nwheat employs a modified routine (Keating et al., 2001) from CERES-Wheat (Ritchie et al., 1985) for calculating leaf area growth, which is divided into two growth stages: (i) from emergence to terminal spikelet initiation and (ii) from terminal spikelet to the end of leaf growth. During the first stage, the simulated growth in leaf area is the smaller of the potential aboveground growth of leaf area and the C supply for each day (note that root growth is created in the model as a proportion of aboveground growth). For reference, the same abbreviations as in the source code of the model are used in the following description of leaf area growth. Potential leaf area growth (plag, cm2 m-2) is calculated after Ritchie et al. (1985) as:
 | [1] |
where cumph(istage) is the cumulative phyllochron within a growth stage; optfr is a stress factor, which is the minimum of water, N, or temperature stress, and tiln is the tiller number (m-2). The daily phyllochron fraction (ti) is calculated after Ritchie et al. (1985) as:
 | [2] |
where DTT is the daily thermal time (base temperature of 0°C) and PHINT is the phyllochron interval. The SLA for a particular day is calculated as a function of the growth stage after Keating et al. (2001) (Fig. 1):
 | [3] |
where sla_new is the SLA of the new potential leaf growth (cm2 g-1), slaB is the SLA at the beginning of leaf growth at emergence (300 cm2 g-1; Keating et al., 2001), slaE is the SLA at the end of leaf growth (225 cm2 g-1; Keating et al., 2001), and istage is the growth stage in the model (1, emergence; 2, terminal spikelet; and 3, end of leaf growth). The C demand to grow leaf area (grolf, g) is then calculated after Ritchie et al. (1985) as:
 | [4] |
View larger version (18K):
[in this window]
[in a new window]
|
Fig. 1. The specific leaf area (SLA) function in APSIM-Nwheat (_______) and with the modification applied to simulate an increased SLA trait (__ __ __ __) over three growth stages (1, emergence; 2, terminal spikelet; and 3, end of leaf growth).
|
|
The leaf area growth rate (lag, cm2 m-2) is then the minimum (min) of C demand and supply after Ritchie et al. (1985):
 | [5] |
where carbo (g) is the C available for all aboveground growth on a given day. Available C for growth is calculated based on plant-intercepted solar radiation and a stress factor derived from the minimum of temperature, water, and N stress. From emergence (istage = 1) until terminal spikelet (istage = 2), stem growth in wheat is small and therefore not accounted for in the model, i.e., all available C at this stage is used for leaf growth. After terminal spikelet, stem growth is the dominant sink, and the proportion of C for stem growth increases with advancing growth stages in the model. Available C not used for stem growth is then available for leaf growth. The leaf area growth rate after terminal spikelet, after Ritchie et al. (1985) is:
 | [6] |
where ptf_grostem is the C fraction for stem growth. From the end of leaf growth until the beginning of grain filling, all available C is used for stem growth and stem reserves for later remobilization to the grain.
Simulation Experiment with an Increased Specific Leaf Area Trait
Plant breeders have characterized barley (Hordeum vulgare L.) as a model crop for early vigorous growth (López-Castañeda et al., 1996). Barley has about a 75% greater SLA than wheat shortly after emergence and declines to about a 25% greater SLA at the four-leaf stage (López-Castañeda et al., 1995). A decrease in SLA with advancing growth stages has also been shown in wheat by Rawson et al. (1987). With wheat, a twofold range in aboveground dry matter accumulation at the five- to six-leaf stage among genotypes was observed by Turner and Nicolas (1998), indicating large differences in early vigor. Early vigor in wheat and barley has been shown to be associated with a greater SLA during early stages of growth (Richards, 1996). To simulate early vigor in the model, SLA at emergence (slaB) in Eq. [3] was increased twofold (Fig. 1) to 600 cm2 g-1 (assuming a similar SLA as that of barley; López-Castañeda et al., 1995). Other possible variations of SLA caused by temperature and radiation (Hotsonyame and Hunt, 1998; Rawson et al., 1987; Rebetzke and Richards, 1999) have been ignored in the model.
Simulation experiments were performed with a range of N treatments (0–210 kg N ha-1) and soil types (sand, loam, and different clays) at key locations in the major wheat-growing areas of Australia (Table 1). Local cultivars with increased SLA at emergence were used at each location (referred to as increased SLA genotypes) and compared with unmodified cultivars (referred to as normal SLA genotypes) (Fig. 1). Three of the locations (Moora, Wongan Hills, and Merredin) were in Western Australia (mediterranean-type climate, winter-dominant rainfall) in three different rainfall zones; two locations (Barellan and Pucawan) were in New South Wales (evenly distributed rainfall) in two different rainfall zones; and one location was in Warra in Queensland with a subtropical climate (Table 1 and Fig. 2). Climate variability was taken into account by repeating the simulation experiment with >78 yr of measured rainfall records and measured and calculated temperature and radiation data from each of the locations. Growing seasons were considered independent by resetting the initial soil conditions every year to the same values (except for soil water at Warra) as explained below. Soil variability was considered for Moora, Wongan Hills, and Merredin in Western Australia by simulating the same experiments for two contrasting soil types, a deep sandy soil [low plant available water-holding capacity (PAW)] and a clay soil (high PAW) (after Asseng et al., 2001a).
View larger version (35K):
[in this window]
[in a new window]
|
Fig. 2. Average monthly mean temperature (lines) and rainfall (bars) for (a) Moora (solid bars and _______), Wongan Hills (open bars and __ __ __ __), and Merredin (cross-hatched bars and __ .. __ .. __), Western Australia, based on weather records from 1907–1996; (b) Barellan (solid bars _______) and Pucawan (open bars and __ __ __ __), New South Wales, based on weather records from 1915–1992; and (c) Warra, Queensland, based on weather records from 1906–1995.
|
|
Each simulation run commenced on 1 January [day of year (DOY) 1] and was reset each year with soil water at its lower limit on 1 January and previous crop residues of 3 to 4.5 Mg ha-1 (based on average residue amounts after Asseng et al., 1998b) to assume the same soil water and residue conditions after the previous crop. To have similar N conditions at the beginning of each growing season, all N soil profiles were reset on 20 April (DOY 110) each year with about 50 kg N ha-1 in 0- to 1-m soil depth. Because of the importance of stored soil water on sowing decisions at the Queensland location, soil water was reinitialized only every 10 yr at 74 mm of PAW in 0 to 0.6 m and at lower limit below 0.6-m depth. These values are based on local experience.
To study the effect of N supply on the expression of the SLA trait, six N treatments—0, 30, 60, 90, 150, and 210 kg N ha-1—were simulated. Up to 90 kg N ha-1, all N was applied at sowing, but at the higher N treatments, the applications were split. In the 150 kg N ha-1 treatment, 90 kg N ha-1 was applied at sowing and 60 was applied 40 d later. In the 210 kg N ha-1 treatment, 90 was applied at sowing, 60 at 40 d later, and 60 another 30 d later. All N was applied as urea; the model assumed that there were no losses by volatilization. Because about 50% of applied fertilizer on loamy soils in New South Wales has been consistently observed to be plant unavailable for unknown reasons (Van Herwaarden, 1995), twice the amount of fertilizer used in the simulations would be required in the field at Barellan and Pucawan to achieve the same N effect.
Historical daily weather records from 1907–1996 for Moora, Wongan Hills, Merredin, Barellan, Pucawan, and Warra consisted of measured daily rainfall for all years and up to 30 yr of measured maximum and minimum temperatures. Maximum and minimum temperatures that were not measured and all radiation data were generated by the Agricultural Production Systems Research Unit using a weather generator, WGEN (Richardson and Wright, 1984).
Years with no sowing opportunity, due to lack of rain during the sowing window (4 yr at Barellan, 2 yr at Pucawan, 8 yr at Warra), were excluded from the statistical analyses.
The effect of the SLA trait on grain yield was calculated relative to the simulated average yield in the normal-vigor genotypes for each N application and for a given environment–soil type.
Western Australia
Sowing was set between 5 May (DOY 125) and 31 July (DOY 212) but did not occur before 5 June (DOY 156) unless at least 25 mm of rainfall had accumulated within the previous 10 d or did not occur after 5 June until at least 10 mm of rainfall had accumulated. In addition, soil water in the 0.05- to 0.10-m layer had to exceed 50% of the extractable soil water-holding capacity to ensure successful germination conditions. The cultivar Spear (late maturing) was simulated for sowing before 5 June; otherwise, the cultivar Amery (early maturing) was simulated using the wheat cultivar coefficients in Table 2. Sowing depth was set at 0.03 m and plant density at 120 plants m-2. A deep sand and a clay soil from the central agricultural zone of southwestern Western Australia were used for the simulations [for details about soil characteristics, see Asseng et al. (2001a) and Table 1].
View this table:
[in this window]
[in a new window]
|
Table 2. Wheat cultivar coefficients for cultivars Amery (early maturing), Spear (late maturing), Janz (medium–late maturing), and Hartog (medium–late maturing).
|
|
New South Wales
Sowing was set between 5 May (DOY 125) and 31 July (DOY 212) but did not occur before 15 June (DOY 166) unless at least 20 mm of rainfall had accumulated within the previous 10 d or after 15 June until at least 15 mm of rainfall had accumulated. In addition, soil water in the 0.05- to 0.1-m layer had to exceed 50% of the extractable soil water-holding capacity to ensure successful germination conditions. The cultivar Janz (Table 2) was simulated for each sowing date. Sowing depth was set at 0.03 m and plant density at 120 plants m-2. Loamy soil characteristics, typical for this region, were used in the simulations [for details about soil characteristics, see Van Herwaarden (1995) and Table 1].
Queensland
Sowing was set between 1 May (DOY 121) and 27 August (DOY 239) but did not occur unless at least 25 mm of rainfall had accumulated within the previous 10 d and at least 60 mm of PAW. The cultivar Hartog (Table 2) was simulated for each sowing date. Sowing depth was set at 0.04 m and plant density at 100 plants m-2. Clay loamy soil characteristics, typical for this region, were used in the simulations [for details about soil characteristics, see Probert et al. (1998) and Table 1].
|
RESULTS
|
|---|
The simulated average yield in the normal-vigor genotypes varied from 1.0 to 4.8 Mg ha-1, depending on location, soil type, and N supply (Table 3). With a low N input, N supply usually restricted early leaf area development and biomass growth at all locations in Australia. Increasing SLA in the simulation increased the leaf area index (LAI) under low and high N supply compared with normal SLA, particularly early in the season as shown for a sandy soil in one year in Fig. 3. Under low N supply, despite increased LAI, the increased SLA trait did not increase biomass accumulation, and grain yield was reduced by 0.23 Mg ha-1 due to greater N stress (Fig. 3a and 3c). With higher N supply, the effect of SLA on LAI was larger, and biomass accumulation and grain yield increased by 0.22 Mg ha-1 compared with the normal genotype, as long as water supply did not limit growth (Fig. 3b and 3d).
View this table:
[in this window]
[in a new window]
|
Table 3. Simulated long-term average grain yields for the wheat genotypes with normal specific leaf area for locations in Western Australia, New South Wales, and Queensland.
|
|
View larger version (26K):
[in this window]
[in a new window]
|
Fig. 3. Simulated (a and b) daily leaf area index (LAI) and water stress and (c and d) aboveground biomass, grain yield, and N stress for (a and c) 30 kg N ha-1 and (b and d) 150 kg N ha-1 and (e) measured rainfall for Wongan Hills, Western Australia, on a sandy soil in 1968. Normal specific leaf area (SLA) is indicated by _______ and increased SLA by __ __ __ __.
|
|
When the N supply was increased from 0 to 150 kg N ha-1 in the medium-rainfall region of Western Australia, the proportion of years with higher yields due to increased SLA increased from 60 to 87% on sand and from 16 to 64% on clay soil (Fig. 4).
View larger version (21K):
[in this window]
[in a new window]
|
Fig. 4. Simulated cumulative probability distribution of the difference in grain yield between the increased specific leaf area (SLA) trait and the normal SLA trait on (a) a sandy soil and (b) a clay soil for 0 (_______), 60 (__ __ __ __), and 150 (___ . ___ . ___) kg N ha-1 treatments at Wongan Hills, Western Australia. The vertical line (...) represents zero.
|
|
The yield response to the SLA trait depended on the yield level of the normal-vigor trait, which in turn was a function of seasonal growing conditions. Data in Fig. 5a shows that the increased SLA genotype tended to yield more grain than the normal SLA genotype on clay soil in the medium-rainfall region of Western Australia, particularly in years when the yields of the normal genotype ranged between 0.3 and 2.0 Mg ha-1. In years with grain yields above 2.0 Mg ha-1 in the normal genotype, the yield response to the increased SLA trait was often negative with low N input. Higher N supply increased the chance of a positive yield response in the medium- to high-yield range (Fig. 5a). A similar yield response to the increased SLA trait, but at a lower yield level, occurred on clay soil in the low-rainfall region of Western Australia. Measured yields (Botwright et al., 2002) with early-vigor lines (increased SLA) and a vigorous cultivar grown on a similar soil type as that used in the simulation, and at the same two locations (Wongan Hills and Merredin) in 1997–1999, were within the range of simulated yield changes (Fig. 5), confirming the plausibility of the simulated direction and magnitude of yield responses to an increased SLA trait.
The effect of the increased SLA trait on maximum LAI was larger on sand than on clay soil, but for both soils, the effect was larger in the high (Moora)- and medium (Wongan Hills)-rainfall region than in the low (Merredin)-rainfall region of Western Australia (Fig. 6). Increasing N supply reduced the proportion of years with a negative effect of the increased SLA trait on maximum LAI. The number of years and the magnitude of higher yields due to increased SLA was slightly larger in the medium (Wongan Hills)- than in the high (Moora)-rainfall zone but lowest in the low (Merredin)-rainfall region due to N leaching and N limitations in the high-rainfall zone and a larger number of years with water limitations in the low-rainfall zone (Fig. 6).
View larger version (31K):
[in this window]
[in a new window]
|
Fig. 6. Simulated cumulative probability distribution of differences between the increased specific leaf area (SLA) and the normal SLA traits for (a, c, and e) maximum leaf area index (LAImax) and (b, d, and f) grain yield for (a and b) Moora, (c and d) Wongan Hills, and (e and f) Merredin, Western Australia, for 0 kg N ha-1 on a clay (_______) and a sandy soil (___ . ___ . ___) and 150 kg N ha-1 on a clay (__ __ __ __) and a sandy soil (___ .. ___ .. ___). The vertical line (...) represents zero.
|
|
The change in long-term average yield responses due to increased SLA, compared with the normal genotypes (Table 3), for the sandy and the clay soil in the high-, medium-, and low-rainfall zone of Western Australia are summarized in Fig. 7a and 7b. Yield responses to increased SLA at Barellan and Pucawan, New South Wales, and at Warra, Queensland, are shown in Figure 7c. The highest increase in average yield from incorporation of the increased SLA trait was simulated for sandy soils in the mediterranean environment of Western Australia, particularly in the medium-rainfall zone (Wongan Hills). With the increased SLA trait, average yields increased by as much as 15% at high N supply (Fig. 7a). On the sandy soil, the higher N applications increased crop transpiration and total evapotranspiration (ET) but reduced soil evaporation (Table 4). Early vigor as a result of increased SLA further increased crop transpiration and reduced soil evaporation, with a resulting small increase in total ET on this soil type (Table 4).
View larger version (32K):
[in this window]
[in a new window]
|
Fig. 7. Simulated relative change in long-term average yields with incorporation of an increased specific leaf area (SLA) trait for (a) sandy and (b) clay soil in Western Australia and (c) loamy soils in New South Wales (Barellan and Pucawan) and a clay loam in Queensland (Warra) for six N treatments. Vertical lines show the upper 25% probability of occurrence base on the long-term simulation runs as an indicator of variability. Long-term simulated average yields of the normal SLA genotypes are shown in Table 3.
|
|
View this table:
[in this window]
[in a new window]
|
Table 4. Simulated long-term average crop transpiration (mm), soil evaporation (mm), and evapotranspiration (mm) during the growing season from May to November for Wongan Hills, Western Australia; Pucawan, New South Wales; and Warra, Queensland.
|
|
On the clay soil in Western Australia, the increased SLA trait reduced yields by more than 10% at low N levels but increased yields by about 5 to 7% with medium N applications before falling again to below 5% yield increases with the highest N treatment (Fig. 7b). The higher N input increased crop transpiration while evaporation was reduced so that total ET was increased (Table 4). Incorporating the increased SLA trait did not affect crop transpiration at zero N due to poor biomass growth (e.g., Fig. 3) and/or reduced soil evaporation with a larger LAI, resulting in no change or even a small reduction of total ET (Table 4). High N input on clay led to an increase in crop transpiration, a further reduction of soil evaporation, and a small increase in total ET (Table 4).
At the two New South Wales sites and the Queensland site, which were on loamy and clay loamy soils, the increased SLA trait reduced average yields by 10 to 15% at low and medium levels of N. Medium levels of N in early-vigor crops stimulated additional N demand that the medium N supply could often not satisfy, resulting in more N stress and reduced yields. Additional N supply substantially increased the average water use before anthesis (Table 4) and, at one of the New South Wales sites, also after anthesis but not at the Queensland location. However, increased biomass growth before anthesis resulted in an increased demand for water after anthesis, which in many seasons, could not be met, resulting in a larger degree of water stress and therefore lower yields. Increased water stress after anthesis was also observed on the clay soil in Western Australia where postanthesis ET was reduced with higher N input (Table 4). Evapotranspiration after anthesis was also slightly reduced with the increased SLA trait at Pucawan and Warra, indicating increased water stress during grain filling (Table 4). With higher N inputs, the negative effect of the increased SLA trait on yield was reversed, but yield increases were only simulated with very high N input and only in a quarter of the seasons at Pucawan. Increased N application reduced the N stress induced by an increased N demand from the more vigorous growth and increased yields in some years (data not shown), but water stress was unchanged as a result of lower soil evaporation (Table 4) due to more soil cover. As a result, in New South Wales and Queensland, yields were reduced by the incorporation of the increased SLA trait in the majority of the seasons (Fig. 7c).
|
DISCUSSION
|
|---|
The yield responses to an increased SLA trait have been shown to be very variable, depending on the climatic region, seasonal rainfall, soil type, and N availability. The simulation helps to explain some of the different yield responses with increased early vigor reported in the literature (Acevedo, 1987; Damisch and Wiberg, 1991; Regan et al., 1992; Siddique et al., 1990a, 1990b; Whan et al., 1993). The results clearly show that incorporation of the increased SLA trait can be negative on the better water-holding soils (clay soils) in the mediterranean climatic region of Western Australia and particularly on the loamy soils in New South Wales and the clay loamy soil in Queensland. Furthermore, in all environments from Western Australia to Queensland, the benefits of increased SLA are not realized without additional N. The yield reductions with incorporation of early-vigor characteristics observed in Western Australia by Regan et al. (1992) were explained by poorly adapted genotypes but may also be due to insufficient N as only 50 kg N ha-1 was applied in their experiment, which had followed several years of cereal crops. Other factors have also been suggested as contributing to the phenological expression of early vigor in wheat, such as sowing date and growth environment, which mainly affect leaf growth through different daily average temperatures (Hotsonyame and Hunt, 1998; Rebetzke and Richards, 1999).
The simulation results confirm suggestions by Turner and Nicolas (1987)( 1998) and Whan et al. (1993) that early vigor might be of greatest advantage on the sandy soils in a mediterranean environment where current crop yields are often limited by poor preanthesis growth. In contrast, in subtropical environments, winter crops are often grown on stored water with little rainfall during winter. Passioura et al. (1993) argued that in these environments, rapid early leaf growth would induce the crop to transpire so much of its limited water supply before flowering that too little would be left for grain filling. Such an interaction between pre- and postanthesis water use affecting grain yield was evident at Warra, Queensland, where the simulated increased SLA trait with high N supply increased preanthesis water use and reduced postanthesis water use, resulting in reduced yields from the increased SLA trait.
Yield increases from early vigorous growth have been considered to arise from faster soil cover and reduced soil evaporation, thereby making more water available for crop transpiration and growth (López-Castañeda et al., 1996; Loss and Siddique, 1994). When an early-vigor line was compared with a normal-vigor line at Condobolin in New South Wales, Condon and Richards (1993) showed that the early-vigor line had increased crop transpiration and reduced soil evaporation, leaving total ET over the growing season essentially the same. However, on average, the long-term simulations indicated that this situation was only likely to occur with high N supply in a mediterranean-type environment. In the uniform-rainfall environment and loamy soils of New South Wales, on average, the early vigorous growth, even with high N input, increased the N stress (when high N input stimulated growth, which in turn stimulated further N demand not satisfied with the applied N amount) and did not affect or slightly reduced crop transpiration and reduced crop growth. Increased early growth also increased the transpiration efficiency due to the increased growth occurring at cooler temperatures and lower vapor pressure deficit with less water use for the same or larger biomass growth and yield (López-Castañeda et al., 1996).
The water-holding capacity of the soil and rainfall amounts and distribution affect the water availability for crop transpiration and soil evaporation. On the sandy soils, yields of genotypes with normal vigor were often limited due to poor leaf area development and biomass accumulation before anthesis (Asseng et al., 2001b). However, on clay soils, the crops grew more vigorously before anthesis, but grain yields were restricted by the limited water available from the shallow wetting depth on this soil (Asseng et al., 2001b). Therefore, the increased SLA trait was more effective in increasing grain yields on the low water-holding sands than on the high water-holding clay soils. Richards and Townley-Smith (1987) and Turner (1997) argued that higher water use before anthesis through increased early growth could be detrimental to grain yield if it depletes the water in the profile by anthesis and does not make more water available by deeper rooting. In studies on a deep sand, Turner and Nicolas (1998) showed that in a mediterranean-type climate, early vigor increased water use before anthesis due to deeper extraction of water by the roots, resulting in more water available to the plants and higher yields. In contrast, on clay soils where rooting depth may be restricted by water penetration, particularly in low-rainfall regions (Asseng et al., 2001b), early vigor may lead to lower yields from greater soil water depletion before anthesis.
Field studies with increased SLA and early-vigor genotypes have been limited to a few seasons and locations and have never been assessed for soil effects and, in particular, N supply. This simulation study allowed a comprehensive assessment of the increased SLA trait and early vigor in wheat that included all of these factors. Although the simplicity of the model and the simulated increased SLA trait (changed slaB) might oversimplify the real weather–crop–soil interactions, this simplicity is also an advantage as it provides an understanding of the major driving forces for yield development in a very variable environment. Nitrogen availability, PAW, and rainfall distribution were key factors in determining whether a crop with the increased SLA trait yielded better than a crop with normal SLA. The study indicates that field testing of increased SLA lines needs to take these factors into account if the incorporation of increased SLA on grain yield is to be comprehensively evaluated. Other traits, e.g., earliness, faster early root growth, and increased transpiration efficiency, might be associated with early-vigor characteristics (Regan et al., 1992; R.A. Richards, personal communication, 1998), making it even more difficult for plant breeders to analyze the impact of SLA alone. Such traits and their interaction with different growing conditions can be analyzed separately and in any combination using a crop simulation model. However, because of the simplistic nature of the model, to be useful, these results need to be interpreted together with plant breeders and in relation to experimental data.
|
CONCLUSIONS
|
|---|
Using a modeling approach highlighted how a simulation analysis can assist plant breeders in quantifying the effect of a physiological trait on yield in very variable growing environments. The long-term simulations showed that increased SLA as a single trait is most effective on the sandy, low water-holding soils in a mediterranean environment but only when the N supply is increased to satisfy the higher N demand of the more vigorous crop. Yield increases are possible in this environment on better water-holding clay soils but with lower probability and only with increased N supply. The increased SLA trait reduced grain yields on loamy soils in the uniform-rainfall environment of New South Wales and on clay loamy soils in the subtropical environment of southern Queensland.
|
ACKNOWLEDGMENTS
|
|---|
We would like to thank B.A. Keating and N. Huth for APSIM support; V. Campisi and W. Vance for technical support, H. Meinke for supplying the weather data; and A. Weiss, S. Abbo, and M. Poole for valuable comments on the paper.
|
NOTES
|
|---|
T. Botwright, current address: Int. Rice Res. Inst., DAPO Box 7777, Manila, Philippines.
|
REFERENCES
|
|---|
- Acevedo, E. 1987. Assessing crop and plant attributes for cereal improvement in water-limited Mediterranean environments. p. 303–320. In J.P. Srivastava et al. (ed.) Drought tolerance in winter cereals. Proc. Int. Workshop, Capri, Italy. John Wiley & Sons, Chichester, UK.
- Aggarwal, P.K., M.J. Kropff, R.B. Matthews, and C.G. McLaren. 1996. Using simulation models to design new plant types and to analyse genotype by environment interactions in rice. p. 403–418. In M. Cooper and G.L. Hammer (ed.) Plant adaptation and crop improvement. CAB Int., Wallingford, UK.
- Anderson, W.K., M. Seymour, and M.F. D'Antuono. 1991. Evidence for differences between cultivars in responsiveness of wheat to applied nitrogen. Aust. J. Agric. Res. 42:363–377.
- Anonymous. 2000. Australian crop report no. 115. Australian Bureau of Agric. and Resour. Econ., Canberra, ACT, Australia.
- Asseng, S., F.X. Dunin, I.R.P. Fillery, D. Tennant, and B.A. Keating. 2001a. Potential deep drainage under wheat crops in a Mediterranean climate: I. Temporal and spatial variability. Aust. J. Agric. Res. 52:45–56.
- Asseng, S., I.R.P. Fillery, G.C. Anderson, P.J. Dolling, F.X. Dunin, and B.A. Keating. 1998a. Use of the APSIM wheat model to predict yield, drainage, and NO3- leaching for a deep sand. Aust. J. Agric. Res. 49:363–377.
- Asseng, S., B.A. Keating, I.R.P. Fillery, P.J. Gregory, J.W. Bowden, N.C. Turner, J.A. Palta, and D.G. Abrecht. 1998b. Performance of the APSIM-wheat model in Western Australia. Field Crops Res. 57:163–179.
- Asseng, S., N.C. Turner, and B.A. Keating. 2001b. Predicting water and nitrogen use efficiency of wheat in a Mediterranean climate. Plant Soil 233:127–143.
- Asseng, S., H. Van Keulen, and W. Stol. 2000. Performance and application of the APSIM Nwheat model in the Netherlands. Eur. J. Agron. 12:37–54.
- Botwright, T.L., A.G. Condon, G.J. Rebetzke, and R.A. Richards. 2002. Field evaluation of early vigour for genetic improvement of grain yield in wheat. Aust. J. Agric. Res. (in press).
- 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.
- Cooper, M., and G.L. Hammer. 1996. Synthesis of strategies for crop improvement. p. 591–623. In M. Cooper and G.L. Hammer (ed.) Plant adaptation and crop improvement. CAB Int., Wallingford, UK.
- Damisch, W., and A. Wiberg. 1991. Biomass yield—a topical issue in modern wheat breeding programmes. Plant Breed. 107:11–17.
- Fischer, R.A. 1979. Growth and water limitation to dryland wheat yield in Australia: A physiological framework. J. Aust. Inst. Agric. Sci. 45:83–94.
- Fischer, R.A., J.S. Armstrong, and M. Stapper. 1990. Simulation of soil water storage and sowing day probabilities with fallow and no-fallow in southern New South Wales: I. Model and long term mean effects. Agric. Syst. 33:215–240.
- Hammer, G.L., D.G. Butler, R.C. Muchow, and H. Meinke. 1996. Integrating physiological understanding and plant breeding via crop modelling and optimization. p. 419–441. In M. Cooper and G.L. Hammer (ed.) Plant adaptation and crop improvement. CAB Int., Wallingford, UK.
- Hotsonyame, G.K., and L.A. Hunt. 1998. Seeding date, photoperiod and nitrogen effects on specific leaf area of field-grown wheat. Can. J. Plant Sci. 78:51–61.
- Jones, C.A., and J.R. Kiniry (ed.). 1986. CERES-Maize: A simulation model of maize growth and development. 1st ed. Texas A&M Univ. Press, College Station, TX.
- Keating, B.A., H. Meinke, M.E. Probert, N.I. Huth, and I. Hills. 2001. NWheat: Documentation and performance of a wheat module for APSIM. Trop. Agric. Tech. Memo 9. CSIRO, Indooroopilly, QLD, Australia.
- Littleboy, M., D.M. Silburn, D.M. Freebairn, D.R. Woodruff, G.L. Hammer, and J.K. Leslie. 1992. Impact of soil erosion on production in cropping systems: I. Development and validation of a simulation model. Aust. J. Soil Res. 30:757–774.
- López-Castañeda, C., and R.A. Richards. 1994. Variation in temperate cereals in rainfed environments: III. Water use and water-use efficiency. Field Crops Res. 39:85–98.
- López-Castañeda, C., R.A. Richards, and G.D. Farquhar. 1995. Variation in early vigor between wheat and barley. Crop Sci. 35:472–479.[Abstract/Free Full Text]
- López-Castañeda, C., R.A. Richards, G.D. Farquhar, and R.E. Williamson. 1996. Seed and seedling characteristics contributing to variation in early vigor among temperate cereals. Crop Sci. 36:1257–1266.[Abstract/Free Full Text]
- Loss, S.P., and K.H.M. Siddique. 1994. Morphological and physiological traits associated with wheat yield increases in Mediterranean environments. Adv. Agron. 52:229–276.
- McCown, R.L., G.L. Hammer, J.N.G. Hargreaves, D.P. Holzworth, and D.M. Freebairn. 1996. APSIM: A novel software system for model development, model testing and simulation in agricultural systems research. Agric. Syst. 50:255–271.[Web of Science]
- O'Leary, G.J., and D.J. Connor. 1998. A simulation study of wheat crop response to water supply, nitrogen nutrition, stubble retention, and tillage. Aust. J. Agric. Res. 49:11–19.
- Passioura, J.B., A.G. Condon, and R.A. Richards. 1993. Water deficits, the development of leaf area and crop productivity. p. 253–263. In J.A.C. Smith and H. Griffiths (ed.) Water deficits: Plant responses from cell to community. Bios Sci. Publ., Oxford, UK.
- Probert, M.E., J.P. Dimes, B.A. Keating, R.C. Dalal, and W.M. Strong. 1998. APSIM's water and nitrogen modules and simulation of the dynamics of water and nitrogen in fallow systems. Agric. Syst. 56:1–28.
- Probert, M.E., B.A. Keating, J.P. Thompson, and W.J. Parton. 1995. Modelling water, nitrogen, and crop yield for a long-term fallow management experiment. Aust. J. Exp. Agric. 35:941–950.
- Rawson, H.M., P.A. Gardner, and M.J. Long. 1987. Sources of variation in specific leaf area in wheat grown at high temperature. Aust. J. Plant Physiol. 14:287–298.
- Rebetzke, G.J., and R.A. Richards. 1999. Genetic improvement of early vigour in wheat. Aust. J. Agric. Res. 50:291–301.
- Regan, K.L., K.H.M. Siddique, N.C. Turner, and B.R. Whan. 1992. Potential for increasing early vigour and total biomass in spring wheat: II. Characteristics associated with early vigour. Aust. J. Agric. Res. 43:541–553.
- Richards, R.A. 1996. Defining selection criteria to improve yield under drought. Plant Growth Regul. 20:157–166.
- Richards, R., G. Rebetzke, and T. Condon. 1998. Taking the cap off wheat yields. Australian Grain. June–July 1998, p. 19–20.
- Richards, R.A., and T.F. Townley-Smith. 1987. Variation in leaf area development and its effect on water use, yield and harvest index of droughted wheat. Aust. J. Agric. Res. 38:983–992.
- Richardson, C.W., and D.A. Wright. 1984. WGEN: A model for generating daily weather variables. Natl. Tech. Inf. Serv., Springfield, VA.
- Ritchie, J.T., D.C. Godwin, and S. Otter-Nacke (ed.). 1985. CERES-Wheat. A simulation model of wheat growth and development. Texas A&M Univ. Press, College Station, TX.
- Shorter, R., R.J. Lawn, and G.L. Hammer. 1991. Improving genotypic adaptation in crops—a role for breeders, physiologists and modellers. Exp. Agric. 27:155–175.
- Siddique, K.H.M., R.K. Belford, and D. Tennant. 1990a. Root:shoot ratios of old and modern, tall and semi-dwarf wheats in a Mediterranean environment. Plant Soil 121:89–98.
- Siddique, K.H.M., D. Tennant, M.W. Perry, and R.K. Belford. 1990b. 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., and R.C. Muchow. 2001. System analysis of plant traits to increase grain yield on limited water supplies. Agron. J. 93:263–270.[Abstract/Free Full Text]
- Stapper, M., and H.C. Harris. 1989. Assessing the productivity of wheat genotypes in a mediterranean climate, using a crop-simulation model. Field Crops Res. 20:129–152.
- Turner, N.C. 1997. Further progress in crop water relations. Adv. Agron. 58:293–338.
- Turner, N.C., and M.E. Nicolas. 1987. Drought resistance of wheat for light-textured soils in a Mediterranean climate. p. 203–216. In J.P. Srivastava et al. (ed.) Drought tolerance in winter cereals. CCSHAU, Hisar, and MMB, New Delhi.
- Turner, N.C., and M.E. Nicolas. 1998. Early vigour: A yield-positive characteristic for wheat in drought-prone Mediterranean-type environments. p. 47–62. In R.K. Behl et al. (ed.) Crop improvement for stress tolerance. CCSHAU, Hisar, and MMB, New Delhi.
- Turpin, J.E., N. Huth, B.A. Keating, and J.P. Thompson. 1996. Computer simulation of the effects of cropping rotations and fallow management on solute movement. p. 558–561. In Proc. Australian Agron. Conf., 8th, Toowoomba, QLD, Australia. 30 Jan.–2 Feb. 1996. Australian Soc. of Agron., Toowoomba, QLD, Australia.
- Van Herwaarden, A.F. 1995. Carbon, nitrogen and water dynamics in dryland wheat, with particular reference to haying off. Ph.D. thesis. Australian Natl. Univ., Canberra, Australia.
- Whan, B.R., G.P. Carlton, K.H.M. Siddique, K.L. Regan, N.C. Turner, and W.K. Anderson. 1993. Integration of breeding and physiology: Lessons from a water-limited environment. p. 607–614. In D.R. Buxton et al. (ed.) International crop science I. Proc. Int. Crop Sci. Congr., Ames, IA. 14–22 July 1992. CSSA, Madison, WI.
This article has been cited by other articles:
|
|
|
|
 
N. C. Turner
Agronomic options for improving rainfall-use efficiency of crops in dryland farming systems
J. Exp. Bot.,
November 1, 2004;
55(407):
2413 - 2425.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
A. G. Condon, R. A. Richards, G. J. Rebetzke, and G. D. Farquhar
Breeding for high water-use efficiency
J. Exp. Bot.,
November 1, 2004;
55(407):
2447 - 2460.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
S. Chapman, M. Cooper, D. Podlich, and G. Hammer
Evaluating Plant Breeding Strategies by Simulating Gene Action and Dryland Environment Effects
Agron. J.,
January 1, 2003;
95(1):
99 - 113.
[Abstract]
[Full Text]
[PDF]
|
|
|
TOP
ABSTRACT
INTRODUCTION
) and 90 (
) kg N ha-1 on clay soil at (a) Wongan Hills and (b) Merredin, Western Australia and measured yield response to early-vigor lines (BC2F2–derived lines of cultivar Amery crossed with a high-vigor donor line) for 1997–1999 with low (•) and high (
) N input and the high-vigor cultivar Westonia with high N input (
) (