| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
a USDA-ARS, Agron. Physiol. and Genet. Lab., Univ. of Florida, P.O. Box 110965, Gainesville, FL 32611-0965
b CSIRO, Cunningham Lab., Division of Tropical Crops and Pastures, 306 Carmody Rd., St. Lucia, Brisbane, QLD 4067, Australia
Corresponding author (trsincl{at}gnv.ifas.ufl.edu)
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Abbreviations: AMAX, area of the largest leaf on the plant • DHI, rate of harvest index increase • FTSW, fraction of transpirable soil water • HI, harvest index • RUE, radiation use efficiency • TU, thermal units
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
There are, of course, several approaches to investigating plant traits for the purpose of increasing yields under water-limited conditions. One experimental approach is to examine the changes in plant traits that may have occurred in cycles of selection leading to increased yields under water deficit conditions. The difficulty in this approach is that there are likely to be a number of interacting and confounding factors. For example, Bolanos et al. (1993) found that few plant traits were correlated with the increase in yield under water deficit conditions observed through eight cycles of selection in a tropical maize population. Of the many traits studied, including osmotic adjustment, only those factors reflecting a reduction in the days to flower, which apparently allowed an escape of drought stress in this environment, were associated with the yield increase.
In the approach used here, a system analysis is presented that focuses on specific plant traits that have been hypothesized as approaches leading to higher yields under water-limited conditions. The influence of individual plant traits are examined in a system analysis using a mechanistic crop model to simulate crop growth under a range of environments. The outputs from these simulations offer quantitative predictions of anticipated yields in response to changing any individual trait of the plant. In addition, the simulation of crop growth over a number of years offers a perspective on the year-to-year variability that is not sampled in conventional experimental approaches.
System analyses of crop yield under water-limited conditions are not new. Previous analyses, however, examined the potential response in crop growth to alterations in only a few traits. Jordan et al. (1983) found that deeper rooting would clearly increase crop yield while earlier maturity and osmotic adjustment had little or no benefit. Similarly, Jones and Zur (1984) showed that an increased soil volume occupied by roots was the most effective adaptive mechanism for increasing growth during simulations of a 10-d drying cycle. Neither of these studies considered the possibility that water deficits may develop that result in the premature senescence of the crop. A simulation study by Sinclair (1994) on the influence of rooting depth on maize yield did include, however, the possibility of crop termination due to severe soil drying. For three locations in the USA, these simulations confirmed the advantage of deeper rooting across a number of years. Other simulation analyses of plant traits for water deficit conditions include those of Muchow et al. (1991) and Hammer et al. (1996).
In contrast to the few plant traits that have been subjected to systems analyses, a large number of traits have been speculated as potentially contributing to increased crop productivity under water-limited conditions. Ludlow and Muchow (1990) discussed 16 such traits, and they recommended that eight of the traits were suited for intermittent stress conditions in modern agriculture. Their top three recommendations in order of priority were to match plant phenology to water supply, osmotic adjustment, and rooting depth. Because their priority list is inconsistent with the results of previous simulation analyses and because they listed a number of plant traits that were not previously considered in a simulation analysis, it is useful to reconsider the possible contribution of various plant traits to increasing yield under water-limited conditions.
Consequently, this analysis was undertaken to examine in a maize model the putative advantage of a number of plant traits in a water deficit environment. The approach was to modify individual traits that were appropriate for maize so that they reflected the behavior of those traits in sorghum. Sorghum is generally felt to be more suited to dry climate conditions than maize. The traits that were modified included the depth of soil water extraction, individual leaf size, the rate of leaf appearance, CO2 assimilation activity, early stomata closure, delayed stomata closure, and grain growth rate and duration. In all cases, the crop was assumed to be vulnerable to crop termination due to severe water deficits.
One trait that was not considered for adjustment was the transpiration efficiency coefficient used to calculate the daily crop transpiration rate based on the daily crop mass accumulation (Muchow and Sinclair, 1991). The transpiration efficiency coefficient for crop gas exchange is well defined and is relatively stable over a range of conditions (Tanner and Sinclair, 1983). Certainly, if this coefficient is less in a particular cultivar than might be expected for that species, then transpiration efficiency and likely yield can be increased by genetically increasing this coefficient toward the expected value. In this analysis, it will be assumed that the transpiration efficiency coefficient of the crop is stable at 9 Pa, which is near the maximum value for C4 species. Previously, this value for the transpiration efficiency coefficient was assumed to be 9 Pa for both maize and sorghum (Muchow and Sinclair, 1991; Sinclair et al., 1997). However, maintaining a constant transpiration efficiency for crop gas exchange does not preclude substantial changes in the overall ratio of crop yield to total water loss because of the many factors influencing transpirational water loss, soil evaporation, and the formation of grain yield.
A number of years were simulated because the weather scenario influencing the development of water deficits during the season can influence crop response. However, this analysis is illustrative because it was done for only a single location, and altered environmental conditions could well influence the sensitivity of crop growth and yield to any particular trait. It is anticipated that this example will illustrate the quantitative merits of specific plant traits for increasing yields under water deficit conditions.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
For many of the plant traits examined, the value of a variable describing the behavior of a trait in maize was simply changed to a value representative of sorghum. Each trait was singly changed in the maize model, and the resulting yields were compared with the nominal maize yields. A final set of simulations was done in which all of the putative traits were included for increased yields under water deficit conditions.
Below are the specific traits that were examined for improving crop yield, and consequently, the ratio of grain yield to evapotranspiration.
Depth of Soil Water Extraction
Experimental evidence from sorghum has shown that the water extraction depth of a drought tolerant cultivar was at least 40 cm deeper than a cultivar that lacked tolerance (Salih et al., 1999). This trait was found in previous simulation studies to be particularly important in making more water available to the crop if water is stored at depth in the soil. In these simulations, the potential store of soil water is calculated as the depth of soil water extraction multiplied by 0.13, which is a reasonable estimate of the volumetric fraction of plant available soil water for many soils (Ratliff et al., 1983). The original maize simulations of Muchow and Sinclair (1991) assumed a store of plant available soil water of 135 mm, or a water extraction depth of approximately 100 cm, based on experimental observations. For the analysis presented here, the nominal rooting depth was assumed to be 80 cm (104 mm water). While this may be slightly less than would be expected of maize, it is not much less than that observed for maize by Carberry et al. (1989). The use of a minimal value for the extraction depth of maize allows for an increased occurrence in the development of stressed conditions in the simulations to emphasize the potential for trait changes to increase yields. Two sets of simulations were undertaken with an increased depth of soil water extraction: 100 and 120 cm (130 and 156 mm water). The results of Birch et al. (1990) showed in a side-by-side comparison that the depth of water extraction by sorghum was greater than that of maize.
Leaf Size
Salih et al. (1999) found that the leaf area produced by a drought tolerant cultivar of sorghum was only about 45% of that produced by a cultivar that lacked tolerance. In these simulations, the leaf area was decreased by changing the single variable (AMAX) that defined the area of the largest leaf on the plant. All of the other leaves on the plant are scaled based on the value of AMAX. Changes in AMAX simulated a change in the leaf area development and the overall potential in the crop leaf area index. In the maize model of Muchow and Sinclair (1991), AMAX was set equal to 750 cm2. A possibility for increasing yield on limited water is to decrease the leaf area so that soil water is conserved to avoid stress conditions, especially late in the season. The influence of this trait on yield was tested by changing AMAX to 500 cm2, which is representative of the smaller leaves usually produced by sorghum (Muchow, 1988a).
Rate of Leaf Appearance
Similar to the above approach, the rate at which leaves appear will influence the rate at which the canopy leaf area develops. An important consequence of changing only the rate of leaf appearance is that slowing this rate lengthens the duration of the vegetative development phase. The rate of maize leaf appearance (Muchow and Carberry, 1989) was changed to a function describing the slower appearance rate of sorghum leaves (Muchow and Carberry, 1990).
Carbon Dioxide Assimilation
Because transpiration is intimately linked to CO2 assimilation by the constancy in the transpiration efficiency coefficient, any change in CO2 assimilation results in a proportional change in water loss. In the model, CO2 assimilation is defined by the radiation use efficiency (RUE), which for maize, was equal to 1.6 g MJ-1 total solar radiation (Muchow and Sinclair, 1991). Decreasing RUE results in decreased crop growth, and consequently, a decreased transpiration rate so that the development of a possible soil water deficit is delayed or avoided. The importance of this trait was tested by changing the RUE to 1.25 g MJ-1, which is more representative of sorghum (Muchow and Davis, 1988). These assumed values for maize and sorghum are fully consistent with values reported by Kiniry et al. (1989) based on a summary of published results up to that point.
Early Stomata Closure
Radiation use efficiency is down-regulated in the model as soil water deficits develop, which are an expression of stomata closure. This down regulation occurs as a function of the fraction of transpirable soil water (FTSW). The value of the FTSW ranges from 1.0 at field capacity to 0.00 when all of the water available for transpiration has been extracted (the value of the FTSW can be <0.00 as a result of continued soil evaporation even though the transpiration rate is zero). Transpiration as a function of the FTSW has been described in a number of species, and a relatively stable exponential function has been obtained across a range of conditions (Sinclair et al., 1998). Down regulation does not usually occur until the FTSW has decreased to a value of about 0.30 to 0.35. For example, the response of down regulation in maize does not result in an RUE decrease to a value that is 90% of the well-watered crop until the FTSW has decreased to a value of 0.29 (Eq. [1] of Muchow and Sinclair, 1991). To conserve soil water, down regulation might be induced at an earlier stage in soil drying. While there appears to be only small variation among crop species in the FTSW at which the down regulation is triggered, Ray and Sinclair (1997) did identify a cultivar of maize in which down regulation was triggered at a significantly higher FTSW than the other cultivars that were tested. Even though maize and sorghum have an equivalent stomata response to soil drying, the possibility of inducing early stomata closure was simulated. Early stomata closure was induced so that RUE was decreased to 90% of the well-watered value when the FTSW decreased to only 0.44.
Delayed Stomata Closure
An alternative strategy to the conservative response, in which down regulation is induced earlier in soil drying, is to sustain plant growth by delaying stomata down regulation. In this case, the crop continues to extract water at a high rate until the soil becomes very dry, i.e., there is a very low FTSW, in anticipation that there will be rain or irrigation before severe water deficits develop. This strategy, therefore, avoids the initial negative consequences of soil drying by continuing plant growth. This approach is one likely expression of osmotic adjustment in that the decrease in stomata closure is delayed (Ludlow and Muchow, 1990). While no difference between maize and sorghum for delayed stomata closure has been reported, this response was simulated in the model so that RUE did not decrease to 90% of the well-watered value until the FTSW reached 0.11.
Grain Growth Rate and Duration
A long period of grain growth results in soil water loss that could subject the crop to severe water deficits. Therefore, yield might be enhanced in water-limited situations if the rate of grain growth is increased and the duration of grain growth is shortened. For example, the data of Ehdaie (1995) showed that there was a negative association between yield and the length of the period from anthesis to maturity among eight wheat (Triticum aestivum L.) cultivars grown under water-deficit conditions. In the model, the daily rate of grain growth is calculated based on a constant linear increase in the harvest index (DHI). A low DHI value results in a low estimate of the rate of grain growth. The DHI for maize is approximately 0.0150 while it is approximately 0.0185 for sorghum (Muchow, 1988b). Coupled with the difference in the DHI is a variation in the duration of grain growth, which is expressed in thermal units (TU). For maize, the length of the grain duration period was defined as 1150 TU (base temperature = 0°C), and for sorghum, the duration was defined as 800 TU (base temperature = 0°C).
Combined Traits
As a final test of the yielding capability under water-limited conditions of crops with sorghum traits in contrast to maize traits, leaf size, rate of leaf appearance, CO2 assimilation, and grain growth rate and duration were combined to convert the maize crop to one with sorghum-like traits. In this case, the adjustment in the depth of soil water extraction for the sorghum-like trait was increased from 80 to 100 cm.
A key feature of all of these tests was the inclusion of a threshold soil water content at which the crop was assumed to be terminated. Therefore, crop survival of severe water deficit conditions was not assured in these analyses, and large decreases in yield could result from the termination of crop growth. In these analyses, the soil water content at which crop growth was terminated was set at a FTSW of -0.02 or less when the HI of the crop was >0.2. The HI restriction was included to allow a minimum production of crop yield even under the most extreme water deficits. The fatal FTSW of -0.02 was slightly less than the value of 0.00 that we previously assumed for maize (Muchow and Sinclair, 1991). Nevertheless, with an assumed soil depth of 80 cm, this threshold resulted in early crop termination in about half of the years in the tested environment.
Simulated Environment
The results of a system analysis, obviously, depend on the environment for which the simulations are done. For this analysis, weather data for Columbia, MO were used because of the high interannual variability in precipitation of this location during the growing season. The mean 20-yr annual rainfall during the growing season for 1969 to 1988 was 411 mm, but the range in growing-season rainfall was 169 to 772 mm.
It was assumed that the soil water profile was fully charged at the beginning of each growing season. At crop emergence, the depth of soil water extraction was taken to be 30 cm. Because it was assumed that the volumetric faction of plant available soil water was 0.13, the initial store of available soil water was 39 mm. The depth of soil water extraction was increased by 3.35 cm d-1 (Robertson et al., 1993) until the maximum depth was achieved.
The sowing date in each year was held at 23 April. Crop emergence was simulated to occur at 87 TU following sowing. The plant population was held constant throughout the simulations at 7 plants m-2. Also, the number of leaves per plant was maintained at 18.3 for all simulations.
For each year, the yield for a simulation with an altered trait was compared with the simulated yield for the nominal maize crop. The difference between the yields was calculated for each year and plotted against the nominal yield for that year. Such a presentation highlights the range of yield changes simulated when a trait is altered from the nominal case.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
|
|
|
Leaf Size
While a smaller leaf size resulted in a substantial yield advantage in 1982 by avoiding a fatal FTSW, this was a unique case among the 20 yr (Fig. 4). No other year offered a weather scenario that resulted in an increased yield over the yield achieved by the nominal maize simulations. The decreased light interception as a consequence of the smaller leaves resulted in decreased yields in other years so that the mean yield for the 20 yr was 39 g m-2 less than the nominal simulations. Therefore, the smaller leaf size is not advantageous for increasing yields in this water-limited environment.
|
|
|
|
Grain Growth and Duration
An increase in the DHI allowed for a shortening of the length of the grain growth period by increasing the rate of grain growth such that the crop might avoid the severe stress that could develop late in the season. This approach failed, however, to eliminate the effects of water deficit on yield (Fig. 8). In several years, this alteration resulted in substantial yield decreases because the threshold HI of 0.2, at which stress-induced termination was reached, occurred at an earlier date. In all years, yield was decreased to some extent by the shortening of the crop growing duration so that less crop mass was accumulated. With a mean yield of only 368 g m-2, this plant trait resulted in the lowest mean yield of any trait tested.
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Similar to previous analyses, these simulations demonstrated the advantage of increasing the depth at which crops extract soil water (Table 1). Yields in all years of the 20 yr of simulations were benefitted by an increased depth of water extraction (Fig. 3). It seems clear that efforts to increase the soil water extraction depth, particularly if a crop is currently limited to a depth as shallow as 80 cm, are likely to be rewarded with increasing crop yields.
Traits that decreased the rate of soil water extraction did not, in general, offer much benefit in yield increase (Table 1). Decreased leaf size and RUE resulted in little or no yield gain. These traits offered a benefit only in one year (1982) where the slowed soil water extraction rate allowed the lethal water stress to be avoided. The results from a study of six cultivars of pigeonpea [Cajanus cajan (L.) Huth] with differing RUE under irrigated conditions similarly showed no significant correlation of grain yield under water-limited conditions with irrigated RUE (Nam et al., 1998).
On the other hand, decreasing the water extraction rate by slowing the rate of leaf development resulted in yield increases both in those years vulnerable to lethal soil water deficits and in many other years with more adequate water. However, yield was decreased in 5 yr as a result of slowed leaf development, showing a variable response in yield to changes in this trait. The simulated yield decreases are consistent with the results of Grumet et al. (1987) from field experiments with barley (Hordeum vulgare L.) in which a comparison of isopopulations with differing rates of leaf development resulted in lower grain yields for the lines with slow leaf development.
Altered stomata closure relative to the soil water content offered few benefits (Table 1). While early stomata closure allowed the avoidance of the lethal soil water content in the simulations for 1982, this trait had little influence on yield in most years (Fig. 7a). The recommendation for osmotic adjustment to delay stomata closure with soil drying proved generally to be deleterious (Fig. 7b). Any putative benefit of osmotic adjustment would likely have to influence crop survival during periods of exposure to potentially lethal periods of soil water deficit.
The simulations that brought together the various sorghum-like traits in the maize model illustrated the benefit of these traits for dryland conditions. In all cases where the yield level calculated with the maize model was <550 g m-2, crop yield was increased by adopting the sorghum-like traits. These results highlighted those traits of sorghum that make it the preferred crop for water-limited environments. In contrast, in those years when maize yields were simulated to be >550 g m-2, crop yields were decreased by incorporating sorghum-like traits. These simulation results are consistent with the experimental comparison of maize and sorghum yields reported by Muchow (1989) for water-limited conditions in Katherine, Australia. In that study, maize yields were superior to sorghum yields when maize yields were at least 600 g m-2. Sorghum yields were superior when maize yielded <600 g m-2. Consequently, sorghum and sorghum-like traits appear appropriate only for very water-limited environments as a means to achieve greater yields, and thus, a greater yield for the fixed amount of available water.
Overall, these simulations are encouraging in showing that crop traits can be adjusted to change yield under water-limited conditions. In particular, sorghum-like traits were generally advantageous in the simulated environments under low-yielding conditions (<550 g m-2), and maize-like traits were generally advantageous under high-yielding conditions (>550 g m-2). Further, the yield modifications directly resulted in shifts of the ratio between grain yield and evapotranspiration (Table 1), which is directly relevant to the concerns of growers for achieving the most efficient use of water.
Received for publication January 28, 2000.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  P. M. Neumann Coping Mechanisms for Crop Plants in Drought-prone Environments Ann. Bot., May 1, 2008; 101(7): 901 - 907. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L. Fan, R. Linker, S. Gepstein, E. Tanimoto, R. Yamamoto, and P. M. Neumann Progressive Inhibition by Water Deficit of Cell Wall Extensibility and Growth along the Elongation Zone of Maize Roots Is Related to Increased Lignin Metabolism and Progressive Stelar Accumulation of Wall Phenolics Plant Physiology, February 1, 2006; 140(2): 603 - 612. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L. C. Purcell, T. R. Sinclair, and R. W. McNew Drought Avoidance Assessment for Summer Annual Crops Using Long-Term Weather Data Agron. J., November 1, 2003; 95(6): 1566 - 1576. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. E. Clay, S. A. Clay, J. Jackson, K. Dalsted, C. Reese, Z. Liu, D. D. Malo, and C. G. Carlson Carbon-13 Discrimination Can be Used to Evaluate Soybean Yield Variability Agron. J., March 1, 2003; 95(2): 430 - 435. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Asseng, N. C. Turner, T. Botwright, and A. G. Condon Evaluating the Impact of a Trait for Increased Specific Leaf Area on Wheat Yields Using a Crop Simulation Model Agron. J., January 1, 2003; 95(1): 10 - 19. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| The SCI Journals | Crop Science | Vadose Zone Journal | |||
| Journal of Natural Resources and Life Sciences Education |
Soil Science Society of America Journal | ||||
| Journal of Plant Registrations | Journal of Environmental Quality |
The Plant Genome | |||
