Crop Science Journal of Natural Resources and Life Sciences Education
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via ISI Web of Science (13)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Abbate, P. E.
Right arrow Articles by Suero, E. E.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Abbate, P. E.
Right arrow Articles by Suero, E. E.
Agricola
Right arrow Articles by Abbate, P. E.
Right arrow Articles by Suero, E. E.
Related Collections
Right arrow Water Use
Right arrow Crop Growth and Development
Right arrow Wheat
Published in Crop Sci. 44:474-483 (2004).
© 2004 Crop Science Society of America
677 S. Segoe Rd., Madison, WI 53711 USA

CROP PHYSIOLOGY & METABOLISM

Climatic and Water Availability Effects on Water-Use Efficiency in Wheat

P. E. Abbate*,a, J. L. Dardanellib, M. G. Cantareroc, M. Maturanoc, R. J. M. Melchiorid and E. E. Sueroa

a Unidad Integrada Balcarce, Estación Experimental Agropecuaria (EEA) Balcarce of INTA (Instituto Nacional de Tecnología Agropecuaria) and Facultad de Ciencias Agrarias of Univ. Nacional de Mar del Plata, CC 276 (7620), Balcarce, Bs.As., Argentina
b EEA Manfredi, INTA, Ruta 9, Km 636 (5988), Manfredi, Córdoba, Argentina
c EEA Pergamino, INTA, CC 31 (2700), Pergamino, Bs.As., Argentina
d EEA Paraná, INTA, CC 128 (3100), Paraná, Entre Ríos, Argentina

* Corresponding author (menegot{at}mdp.edu.ar).


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 REFERENCES
 
In Argentina, wheat (Triticum aestivum L.) is cropped over a wide range of climatic conditions. Considerable variability in the ratio of dry weight produced per unit of transpired water, usually referred to as water-use efficiency (WUE), is expected as variation in climatic factors affects photosynthesis and transpiration in different ways. Also, previous studies have shown that water supply limitations may affect WUE in wheat. The objective of this study was to quantify the effects of climatic environment and water availability on WUE in wheat crops. Six experiments were conducted at different locations of the Argentine wheat belt and crop dry weight and water use were measured in periods when water use was dominated by transpiration. Three of the experiments included both irrigated and rainfed treatments. Mean daily values of (i) pan evaporation, (ii) relative humidity, (iii) potential water use, and (iv) vapor pressure deficit, were used to find a general relationship that explained effects of the climatic environment on WUE. For experiments with high water availability, daytime vapor pressure deficit was better related to WUE than the other climatic factors. WUE was greater for experiments with water limitation, probably because stomatal closure to restrict transpiration rate occurred around midday when vapor pressure deficit was highest. As a consequence, relative dry weight under water limitation was not linearly related to relative water use as proposed in previous studies. A quadratic relationship that better represented this response was derived.

Abbreviations: DW, dry weight • GAI, green area index • GAR, green area ratio • H, high water availability treatments (irrigated) • L, low water availability treatments (rainfed) • LAI, leaf area index • PAN, pan evaporation • PAR, photosynthetically active radiation • PEN, Penman-FAO potential water use • RH, percentage relative humidity • SE, standard error of mean • SEE, standard error of estimate • VPD, vapor pressure deficit • WU, water use • WUE, water-use efficiency


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 REFERENCES
 
WATER SUPPLY is often the most critical factor limiting crop growth and yield in rainfed areas and the most expensive input of irrigated crops. Therefore, crop production usually requires maximizing yields on limited available water resources. In absence of nutrient limitations, the most critical period for yield determination in wheat (Fischer, 1979; Abbate et al., 1997) generally takes place while water use (WU) is dominated by transpiration. One of the key components of crop production is to achieve the greatest ratio of dry weight to transpiration, usually known as water-use efficiency (WUE). Water-use efficiency is strongly influenced by weather conditions affecting transpiration and assimilation by leaves, plants, and crop differently (de Wit, 1958; Fischer and Turner, 1979; Fischer, 1980; Tanner and Sinclair, 1983). The effect of climate on WUE based on transpiration, is to be assessed both for modeling purposes and to adopt management practices that allow WUE to be maximized.

In Argentina, wheat is mainly cropped from 32 to 39°S and 57 to 63°W (Hall et al., 1992). Considerable variability in climatic conditions and WUE is expected throughout this zone. Several authors have proposed the use of environmental indexes to explain changes in WUE across environments. Some of the most frequently used indexes were mean pan evaporation (de Wit, 1958), mean relative humidity (Arkley, 1963), mean vapor pressure deficit (VPD) (Bierhuizen and Slatyer, 1965), and mean potential water use (Doorenbos and Kassam, 1979). Vapor pressure deficit is a widely used and concise index that explains changes in the ratio yield/water use (Angus and van Herwaarden, 2001). The effect of VPD on WUE has been reported for corn (Zea mays L.), sorghum [Sorghum bicolor (L.) Moench], potato (Solanum tuberosum L.), alfalfa (Medicago sativa L.), soybean [Glycine max (L.) Merr.] (Tanner and Sinclair, 1983), and barley (Hordeum vulgare L.) (Monteith, 1986). Comparisons of the abilities of different meteorological indexes to explain environmental effects on WUE (under well watered or water deficit conditions) are scarce. To our knowledge, no relationship between WUE and VPD has been reported for wheat.

On the other hand, several reports have shown that water-supply limitations create other additional source of WUE variation besides the climatic conditions. When WUE in wheat under water limitation was compared with well-watered checks, the results showed that WUE either increased (Barraclough et al., 1989; Fischer, 1980; Mohamed and Abdel Monem, 1994; Zhang et al., 1998), did not change (van den Boogaard et al., 1996b), or decreased (El Hafid et al., 1998). The aim of this study was to quantify, under field conditions across the Argentine wheat belt (i) the ability of different meteorological indexes to explain environmental effects on WUE in well-watered crops and (ii) to assess the effects of water supply limitations on WUE.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS AND DISCUSSION
 REFERENCES
 
Six experiments (Table 1) were conducted in Argentina: two at Córdoba (31°30' S, 64°00' W, altitude 360 m), one at Paraná (31°47' S, 60°29' W, altitude 79 m), one at Pergamino (33°56' S, 60°33' W, altitude 65 m), and two at Balcarce (37°45' S, 58°18' W, altitude 130 m). The soil at Córdoba was a silty loam Enthic Haplustoll (USDA Soil Taxonomy) with 23 g kg–1 organic matter in the top 20 cm of soil; at Pergamino, it was a silty loam Typic Argiudol with 30 g kg–1 organic matter; and at Paraná, it was a silty loam Aquic Argiudol with 33 g kg–1 organic matter. These soils had no physical restrictions. The soil at Balcarce was a loam Typic Argiudol with 55 g kg–1 organic matter in the top 25 cm, and a petrocalcic layer between 70- and 80-cm depth, which restricted root penetration.


View this table:
[in this window]
[in a new window]
  		Table 1. Details of the six field experiments in Argentina.

 
Experiment Management
Sowing dates varied from June to August (Table 1) and seeding rate was 250 to 350 seeds m–2, with the lowest seeding rates for June sowings and the highest for sowings in August. Experiments were conducted with cv. PROINTA Oasis and PROINTA Federal (Paraná site only), both commercial Argentine awned, semidwarf, spring, bread-wheat cultivars. The plots were 5.5 to 7.0 m long and seven rows wide in experiments without rainfed treatments, and 7 to 15 m long and 21 to 25 rows in experiments with rainfed treatments. Rows were 0.20 m apart. The experiments were surrounded by a crop border at least seven rows wide. Lodging was prevented by nets, and pests and diseases were adequately controlled. In irrigated treatments, the soil water content was kept above 50% of maximum soil available water (water content retained between 33 and 1520 kPa) up to the depth in which soil water content was measured (1 or 2 m), by irrigating with a drip system. The crops received 11.1 to 20.0 g N m–2 divided in two applications, and 1.2 to 2.8 g P m–2 broadcasted and incorporated before seeding. These amounts of fertilizer were sufficient to ensure no nutrient limitations. Weather records (Table 2) were obtained at meteorological stations located 0.5 to 2 km from each experimental site.


View this table:
[in this window]
[in a new window]
  		Table 2. Main meteorological variables for each of the six field experiments, during the period with green area index ≥ 5 in the high water availability (irrigated) treatments and dry weight sampling frequency.

 
Water Use
Volumetric soil water content was measured every 7 to 10 d in each plot. Measurements were taken at 20-cm-depth intervals by field calibrated neutron probe, or gravimetrically (Paraná). Depth of measurement were 1 m in irrigated plots and 2 m in rainfed ones at Paraná and Pergamino, and 2 and 3 m, respectively, at Córdoba. Balcarce measurements were taken only to the depth of the petrocalcic layer. Soil water content in the top 20 cm was determined gravimetrically in all experiments. Gravimetric water content was converted in volumetric content using the bulk density of each layer and then accumulated across depths to calculate the water stored within the soil profile. Cumulative WU (mm) for a given period, was determined by accumulating the water balance between successive soil moisture measurements by the following equation:

[1]
where PE (mm) is water supplied to soil by effective precipitation, I (mm) is the irrigation, and {Delta}S (mm) is the change in the stored water within the soil profile. The PE was estimated by measuring the soil water content after each precipitation event. When this procedure was not possible to follow, daily PE was calculated from daily precipitation (P, mm), by an equation proposed by Dardanelli et al. (1992) for daily values higher than 15 mm and for soils with surface characteristics similar to those in these experiments:

[2]
Irrigation application efficiency was assumed to be 100%. Soil water content at the deepest measured layer changed ≤10% through the experimental period; thus, water loss by deep drainage was assumed to be negligible.

Crop Measurements
To measure crop dry weight (DW), plants were cut at ground level from the four to five innermost rows in each plot at 7- to 14-d intervals, except at Paraná (28 d, Table 2). Quadrat length was 50 cm before anthesis and 80 cm afterwards. A distance of at least 35 cm was left between successive quadrats. Subsamples of 40 to 50 shoots were dried at 65°C to constant weight to determine DW.

Green area ratio (GAR, cm2 g–1) was calculated as the quotient of green area and DW of a subsample of 25 shoots. The green area of the subsample was determined by measuring the surface (one side) of all green organs (i.e., leaves, stems, and spikes when present) with an area meter LI-3000 (LI-COR, Lincoln, NE). Green area index (GAI) was then calculated as the product of plot DW (g m–2) and GAR. In Exp. PA95, GAR was not measured, but was estimated on the basis of a relationship obtained from irrigated treatments in Balcarce and Córdoba, and nonshaded treatments from Abbate et al. (1997) for cv. PROINTA Oasis as follows:

[3]
where x (0.4–1.2) is time after sowing expressed as a fraction of the days between sowing and anthesis.

Water-Use Efficiency
Water-use efficiency (g m–2 mm–1) was obtained as the ratio between DW and WU changes during the period in which the measured GAI in the irrigated treatments were ≥5 (Table 3), to minimize the contribution of soil evaporation (≤10%) to WU. This GAI value was greater than the average threshold for leaf area index (LAI) reported from previous experiments (Table 4). The threshold value in those studies varied between 3.7 and 6.0, except when the surface soil was dry (2.7). Abbate et al. (1998) reported that stem and spike projected area increased from 6% of GAI 20 d before anthesis to 14% of GAI 7 d after anthesis. Thus, the values of GAI in this work are 6 to 14% greater than LAI (including green leaf only), and 4 to 8% [6 or 14% x ({pi}/2 – 1)] smaller than expected if LAI included half the surface of stems and spikes considered as cylinders, as usually calculated in wheat.


View this table:
[in this window]
[in a new window]
  		Table 3. Days from sowing to anthesis; beginning, end and duration of the period with green area index (GAI) ≥ 5 (in irrigated treatments); mean GAI, dry weight (DW) and water use (WU) changes, and water-use efficiency (WUE) during that period; for high (H, irrigated) and low (L, rainfed) water availability treatments, in each of the six experiments.

 

View this table:
[in this window]
[in a new window]
  		Table 4. Leaf area index (LAI) thresholds for soil evaporation ≤ 10% of total water use, reported or calculated for well-watered crops under different soil surface moisture conditions.

 
To match DW with WU dates of measurement, WU (measured more frequently than DW) was linearly interpolated over time between successive soil water measurements when calculating WUE. In the rainfed treatments, WUE was estimated for the same periods that the irrigated treatments had GAI ≥ 5, although the rainfed treatments did not achieve such values of GAI (Table 3).

Data Analysis
For the selected period (GAI ≥ 5 in irrigated treatments), the effect of meteorological conditions on WUE and DW, was examined using four common meteorological indexes, by applying the general expressions proposed by de Wit (1958):

[4]
and as WUE = DW/WU, then

[5]
where {alpha} and ß are empirical constants, and X is the meteorological index expressed as the average daily value during the selected period. The indexes used were as follows.

  1. Mean daily pan evaporation (PAN, mm d–1), as suggested by de Wit (1958).
  2. Mean daily relative air dryness (RD, %; Arkley, 1963), which was calculated as:

    [6]
    where RHmax and RHmin are maximum and minimum daily atmospheric relative humidity (%), and {theta} is a weighing parameter (0 ≤ {theta} ≤ 1).

  3. Mean daily potential WU (PEN, mm d–1), which was calculated as potential evapotranspiration (without crop coefficient correction) by the Penman-FAO method (Doorenbos and Kassam, 1979), via the algorithm computed by the CERES 3.5 program (Hoogenboom et al., 1994) with an albedo value of 0.23.
  4. Mean daily vapor pressure deficit (kPa), as suggested by Tanner and Sinclair (1983). The value of VPD was calculated daily as the difference between the saturated vapor pressure (ea) and the actual vapor pressure (ed), using daily maximum and minimum temperature (Tmax, Tmin) and daily RHmax and RHmin, following the procedure of Allen et al. (1998):

    [7]


[8]

[9]

[10]

[11]

[12]
where Ti is Tmax or Tmin, and {theta} a weighing parameter similar to that defined for RD.

The use of {theta} in Eq. [6] and [8] entails that the higher its value, the higher the value of the meteorological index. As a first approximation, {theta} was set at a value of 0.5 for RD and VPD calculations. This means that calculated values of RD and ea represent the mean daily value.

For fitting Eq. [4], WUE was calculated over the entire period selected, but Eq. [5] was fitted using all the available accumulated DW and WU data during the period.

Statistical Analyses
The experimental design and number of replications are shown for each experiment in Table 1. Within each experiment, the mean treatment values were compared using the least significant difference method when analysis of variance revealed significant differences of means (Steel and Torrie, 1980). Regression analyses were based on mean values of each treatment. The parameters of nonlinear equations (Eq. [4] and [5]) were fitted iteratively by minimizing the standard error of estimation (SEE) between the estimated (yi) and measured (yi) dependent variable:

[13]
where n was the number of data points, p the number of parameters estimated and n – p the degrees of freedom. The goodness of fit for the different meteorological indexes was compared by an F test on the SEE at the P ≤ 0.05 significance level.


   RESULTS AND DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
REFERENCES
 
Water-Use Efficiency under High Water Availability
The experiments were conducted under a wide variation in environmental conditions. Thus, for high water availability treatments, WUEH (the subindex denoting high water availability) for the period with GAI ≥ 5 had a variation greater than 100% (range 2.9–7.0 g m–2 mm–1, Table 3). The highest WUE values were obtained in Balcarce (Exp. BA97) and the lowest, in Córdoba late sowing (Exp. CA98).

Table 5 shows the SEE of WUEH when different meteorological indexes were fitted using Eq. [4] for the well-watered treatments. If the exponent ßH is fixed at zero, {alpha}H takes the value of the average WUEH among all the experiments, and the SEE of WUEH (1.29 g m–2 mm–1, 27%) equates to SE of the mean WUEH. Setting ßH = –1 decreased the SEE of WUEH to 10 to 12%, depending on the meteorological index. The lowest SEE was obtained with VPD, although it was not significantly different from SEE associated with the PEN and RD indexes. However, the fitted line obtained with ßH = –1, overestimated the observed WUEH values at low VPD and underestimated them at high VPD (Fig. 1a , line a). When the ßH was also fitted (ßH = –0.73), no biases were observed (Fig. 1a, line b), and the SEE was significantly reduced (Table 5). In this case the SEE of WUEH using VPD was 93 and 80% lower than for ßH = 0 and ßH = –1, respectively. To our knowledge, no previous studies have used VPD as a meteorological index for WUE predictions in wheat. Pan evaporation has been reported as a meteorological index for normalizing differences in WUE (de Wit, 1958; Fischer, 1979). From daily PAN data with the quadratic equation proposed by Fischer (1979), the SEE obtained (0.59 g m–2 mm–1) was higher than that obtained by PAN in Eq. [4].


View this table:
[in this window]
[in a new window]
  		Table 5. Standard errors of estimation (SEE) for water-use efficiency (WUE) fitted to four meteorological indexes using Eq. [4], for high water availability (H, irrigated) crops during the period with green area index ≥ 5. Values ( ±standard error) of {alpha}H and ßH, are given when Eq. [4] was fitted with ßH = –1 or unconstrained; for the latter, the probability of fitted ßH being equal to –1 is given.

 


View larger version (21K):
[in this window]
[in a new window]
  		Fig. 1. WUE vs. VPD showing fit of Eq. [4], for crops with high (H, irrigated) or low (L, rainfed) water availability, during the period with green area index ≥ 5 in H crops. (a) VPD computed with {theta} = 0.50; line a: {alpha}H fitted with ßH = –1; line b: {alpha}H and ßH both fitted. (b) VPD computed with {theta} = 0.72 (adjusted for H crops); line a: {alpha}H fitted with ßH = –1; line b: {alpha}L fitted with ßL = –1, data from Exp. PE98 was excluded (see text). See Materials and Methods for explanation of the {theta} parameter, and Table 5 and Table 6 for fitting details.

 

View this table:
[in this window]
[in a new window]
  		Table 6. Standard errors of estimation (SEE) for water-use efficiency (WUE) using Eq. [4], fitted {alpha}H and ßH with {theta} = 0.75 or fitted {alpha}H and {theta} with ßH = –1, for high water availability (H, irrigated) crops during the period with green area index ≥ 5. Values (±standard error) of fitted parameters and the probability of fitted ßH being equal to –1 are given.

 
It is perhaps surprising that the ßH value for VPD was significantly different from –1 (Table 5), but de Wit (1958), using PAN, suggested that the ßH value might vary between 0 (Netherlands) and –1 (U.S. Great Plains). Howell (1990), in examining DW to transpiration association, pointed out that the relationships between WUE and daily VPD–1 may not be constant across different environments. However, Tanner and Sinclair (1983) reported that ßH = –1 was a reliable exponent for WUEH estimations using daytime VPD. The discrepancy may have originated in the procedure used for VPD calculations. Since the daytime values of transpiration and assimilation should be the most important in explaining WUE variation, daytime VPD is likely to be a better meteorological index than daily mean VPD (Howell, 1990). These considerations also hold when RD was used as meteorological index, but not when PEN or PAN were used, because daily values of these indexes reflected mostly daytime conditions. Daytime VPD can be calculated by increasing {theta}, as the values of {theta} > 0.5 give more weight to Tmax and RHmin, which corresponds to daytime values. But it is not clear how best to define the value for {theta}. Tanner and Sinclair (1983) used {theta} ranging between 0.66 and 0.75 for maize growing in different environments. In our study, we refitted Eq. [4] for VPD and RD indexes using {theta} = 0.75. The ßH parameters obtained did not differ significantly from –1 in either case (Table 6). However, {theta} = 0.75 may not be the value that best represented daytime conditions. By setting ßH = –1, {theta} could be optimized and resultant {theta} values were close to 0.75 (Table 6). The SEE obtained using the VPD index was lower than that for RD. Using VPD with optimized {theta} (Fig. 1b, line a), we confirmed the validity of ßH = –1 and the constancy of the ratio WUEH/VPD throughout environments as proposed by Tanner and Sinclair (1983). Although the SEE for the fit with VPD using optimized {theta} (4%, Table 6) was not lower than that obtained using daily VPD ({theta} = 0.50, Table 5), both were small and acceptable.

The extreme WUEH values found in our study (Table 3) were close to those calculated under high water availability from previous studies during the period of full canopy cover under similar VPD (computed with {theta} = 0.72) conditions. The lowest WUEH value (2.9 g m–2 mm–1) in our study was similar to the 3.3 g m–2 mm–1 found in southern high plains of Texas (USA) (Howell et al., 1995), during a period when mean VPD was 1.8 kPa. The highest value in our study (7.0 g m–2 mm–1) was close to the 8.0 g m–2 mm–1 found in Rothamsted (U.K.) (Barraclough et al., 1989) when mean VPD was 0.66 kPa. The relationship found for high water availability conditions (Fig. 1b, line a), predicted these values well with estimates of 3.1 and 8.4 g m–2 mm–1 for the Howell et al. (1995) and Barraclough et al. (1989) data, respectively (data not shown).

While DW under high water availability was not closely related to WU (Fig. 2a) , it could be well predicted when WU was corrected for VPD using the findings for WUEH (fitted {alpha}H with {theta} = 0.72 and ßH = –1) applied to the Eq. 5 (Fig. 2b), which did not improve when an independent term was included. Improvements in DW estimation with WU/VPD rather than WU as the independent variable were also reported by Tanner (1981) for potato and Monteith (1986) for barley, independent of water availability.



View larger version (26K):
[in this window]
[in a new window]
  		Fig. 2. Dependence of cumulative DW on cumulative WU and cumulative WU VPD–1, for crops with high (H, irrigated) or low (L, rainfed) water availability, during the period with green area index ≥ 5 in H crops. (a) WU VPDß with ß = 0. (b) fit of Eq. [5] with ß = –1, VPD computed with {theta} = 0.72, {alpha} fitted; line a: H crops; line b: L crops, data from Exp. PE98 were excluded (see text). See Materials and Methods for explanation of the {theta} parameter. SEE, standard error of estimations of DW (as percentage of mean dry weight in brackets).

 
Water-Use Efficiency and Water Availability
Comparisons between water regimes within each experiment revealed higher WUE under rainfed conditions, which was significant in Exp. CA98 and BA98 (Table 7, mean 37% increase), but not in Exp. PE98 (8% increase). The mean percentage WU reductions were 44% in Exp. CA98 and BA98, but only 16% in Exp. PE98. Thus, we could assume that in Exp. PE98, drought was minor during the period analyzed. In spite of the observed increase in WUE induced by water deficit, crop DW at low water availability was always equal to or lower than at high water availability (Table 3). The calculated increase in WUE caused by water deficits might also be underestimated because more soil evaporation might have occurred in low water availability treatments as GAI ≥ 5 was not achieved (Table 3) and 13 to 34 mm rainfall occurred (17–36% of WU) during the period used for calculation (Table 2).


View this table:
[in this window]
[in a new window]
  		Table 7. Values reported or calculated for water-use efficiency (WUE) in wheat under high (H) and low (L) water availability, and the associated percentage WUE increase and percentage water use (WU) reduction.

 
With only three available data points it was not appropriate to fit {alpha}L and ßL parameters (the subindex denoting low water availability) at the same time. We tried to fit {alpha}L, assuming ßL = ßH, and to fit ßL assuming {alpha}L = {alpha}H, by calculating VPD with {theta} = 0.72. The SEE for WUEL was significantly lower in the first case (0.18 g m–2 mm–1 = 3%, Fig. 1b, line b) than in the second (2.14 g m–2 mm–1 = 36%). Thus, water deficits seemed to modify the ratio WUE/VPDß expressed by the {alpha} parameter, rather than the exponent ß. The difference in WUE (Fig. 1b) and {alpha} coefficient (Fig. 2b) across water regimes might be overestimated if soil evaporation was underestimated, particularly under high water availability. Assuming zero soil evaporation under low water availability, for {alpha}H to equal {alpha}L soil evaporation must have been 43% [(8.0/5.6 – 1)x100] of WU under high water regime, or at worst 28% so that the difference between {alpha}L and {alpha}H be statistically nonsignificant. Such thresholds of soil evaporation are much greater than expected in crops with GAI ≥ 5.

Assuming similar ß values in situations with and without water deficit, it follows from Eq. [4] that

[14]
so if WUEH and WUEL were evaluated at the same level of VPD, the relative increase in WUE under water limited conditions would be the same across a wide range of VPD. This increment was found to be reasonably similar among our experiments (excluding PE98, Table 7), ranging from 34 to 42%, when mean WU reduction was 44% and VPD (calculated with {theta} = 0.72) ranged from 0.9 to 2.0 kPa. Consequently, the highest absolute increment in WUE would occur at the lowest VPD.

De Wit (1958) recognized that WUE for plants experiencing severe water limitation might be higher than WUE for plants with adequate soil moisture. However, he had limited information at his disposal at that time. Several papers have reported an increase in WUE for crops of wheat exposed to water limitations (Table 7). Zhang et al. (1998) found that WUE, calculated over the entire crop cycle after removing soil evaporation based on estimates from a physically based model, increased by about 20% over four growing seasons. They attributed the higher WUE under stress to enhanced stomatal conductance under irrigation that favored elevated transpiration more than elevated photosynthesis. Recalculated values of WUE from Mohamed and Abdel Monem (1994), for the period with low soil evaporation from stem elongation to milk dough, showed a 56% increase associated with water limitation between environments. Reprocessed data from Barraclough et al. (1989), for the period with LAI > 4, showed a WUE increment of 31% associated with water limitation between environments. Similar trends were reported by Fardad and Pessarakli (1995) for a semiarid environment. Water-use efficiency increased 20 and 37% for low and severe water deficits, respectively. In a pot study under controlled conditions, van den Boogaard et al. (1996a) demonstrated that plant WUE increased progressively and proportionally to transpiration reductions until transpiration was reduced by 70%. These results demonstrated that the increase in WUE under water deficit occurred over a wide range of VPD and drought levels, as noted in this study.

On the other hand, van den Boogaard et al. (1996b) and El Hafid et al. (1998) concluded that drought did not increase WUE. However, they calculated WUE for the whole growing period, and did not discriminate between evaporation and transpiration (El Hafid et al., 1998), or estimated soil evaporation as the intercept on the x axis (WU) of the DW to WU linear relationship. It is known that this crude method of estimating soil-evaporation losses during the growth period may underestimate soil evaporation (Angus and van Herwaarden, 2001). Although there may be some contradictory results for the effect of water deficit on WUE, when the contribution of soil evaporation was removed with a reliable procedure, the WUE based on transpiration alone was consistently increased over a wide range of drought intensity and climatic environments, which is in agreement with our findings in the present study. The increase in WUE based on transpiration could be explained to some extent by stomatal control on transpiration. Sinclair et al. (1984) suggested that stomatal sensitivity for preventing high transpiration rates could be important for improving WUE, in particular, midday closure of stomata during periods of high VPD. Results of carbon isotope discrimination studies (Farquhar and Richards, 1984; Ehdaie et al., 1991) are compatible with this idea.

Crop Growth and Water Use
Hanks (1974) considered that if WU was evaluated with either low or high water availability in the same environment, and WU represents transpiration, then Eq. [5] condenses to

[15]
Further, he assumed that water availability did not modify the value of {alpha}, so

[16]

Equation [16] is widely known and has been adopted in several models (Howell, 1990). However, one immediate consequence of our finding of water deficit increasing WUE, is that the ratio {alpha}L/{alpha}H would be ≥1, if ßL = ßH. As a first approximation, it can be suggested that water availability affecting WU induces a relative increase in {alpha} (and in WUE) of a similar magnitude to the relative fall in WU. This approach is based on the fact that, when WU is not affected by water deficit (i.e., 1 – WUL/WUH = 0%), we expect neither an increase in relative WUE (i.e., WUEL/WUEH – 1 = 0%) nor in {alpha}L/{alpha}H – 1 = 0% (cf. Eq. [14]); and according to our data, when 1 – WUL/WUH was 44% {alpha}L/{alpha}H – 1 was 43% (calculated for Fig. 1b), thus

[17]
and then

[18]

Substituting Eq. [18] in Eq. [15] we obtain

[19]
This function indicates that when relative WU declines from 100 to 50%, relative DW is reduced less than relative WU (slope < 1), and the opposite occurs when relative WU declines from 50 to 0% (slope > 1), but relative WUE (i.e., the ratio between relative DW and relative WU) always increases according to

[20]

In Fig. 3 , data are compared both from this study and from others previously reported (Table 7), with the functions proposed by Hanks (1974) (Eq. [16]), with our deduced Eq. [19] and with a quadratic line of regression. The latter was the relationship producing the least SEE for relative DW (4.8%), which did not improve when an independent term was included. The SEE of Eq. [19] (5.8%, Fig. 3, line b) was not statistically greater than in the quadratic regression (Fig. 3, line c), while the SEE produced by Hanks' (1974) function was clearly higher (18.6%).



View larger version (31K):
[in this window]
[in a new window]
  		Fig. 3. Different functional relationships between relative DW and relative WU when observations under low water availability are expressed relative to observations under high water availability (see Table 7 for data). Line a: Hanks (1974) relationship (Eq. [16]); line b: Deduced Eq. 19; and line c: Quadratic regression. SEE, standard error of estimations of relative DW.

 
The quadratic form derived as Eq. [19] is based on Eq. [17], which assumes an inverse linear association between relative WUE and relative WU. This may be questioned. It is evident in Table 7 that the relative increase in WUE did not always result in an equal relative fall in WU. The linear regression between these variables was not strong (r2 = 0.56, df = 8), the intercept was equal to zero, but the slope was different to the expected value of one. This could be due to the high error inherent in the quantification of relative WUE, as it is the result of at least four measurements per replication, and the error may be even higher if data were to come from different sources as in Fig. 3. Another assumption underlying Eq. 19 is that ßL = ßH. If ßL!=ßH, the value of {alpha}L/{alpha}H might depend not only on the relative WU but also on VPD or some other climatic variable. The available information does not allow this question to be elucidated. Despite this concern, the approach suggested improved estimations of relative DW variations associated with changes in relative WU.

In conclusion, when WU was dominated by transpiration (>90%) under conditions of high water availability, VPD was more closely associated with WUE than with PAN, RD, or PEN. The value of the ß coefficient (Eq. [4]) was compatible with –1 when VPD was calculated with {theta} = 0.72, which is consistent with previous theories (de Wit, 1958; Fischer and Turner, 1979; Tanner and Sinclair, 1983). Water deficiency increased WUE, probably because stomatal control prevented high transpiration rates by means of midday closure during periods of high VPD. Low water availability apparently increased the value of the {alpha} parameter without affecting ß. Thus, in absolute terms, the increase in WUE was higher with low rather than with high VPD, but in relative terms the increase was independent of VPD. As a consequence, DW under low- relative to high-water-availability environments did not relate linearly with relative WU, as proposed in previous studies (Hanks, 1974). This response was better represented by a quadratic relationship of the form derived in this study.


   ACKNOWLEDGMENTS
 
We would like to thank R.A. Fischer and V.O. Sadras for their valuable comments and suggestions, and A.C. Menegotto for her help with the manuscript preparation.

Received for publication December 22, 2001.


   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 REFERENCES
 




This article has been cited by other articles:


Home page
 Agron. J.Home page
X. Zhang, D. Pei, S. Chen, H. Sun, and Y. Yang
Performance of Double-Cropped Winter Wheat-Summer Maize under Minimum Irrigation in the North China Plain
Agron. J., October 31, 2006; 98(6): 1620 - 1626.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via ISI Web of Science (13)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Abbate, P. E.
Right arrow Articles by Suero, E. E.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Abbate, P. E.
Right arrow Articles by Suero, E. E.
Agricola
Right arrow Articles by Abbate, P. E.
Right arrow Articles by Suero, E. E.
Related Collections
Right arrow Water Use
Right arrow Crop Growth and Development
Right arrow Wheat

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
The SCI Journals Agronomy Journal 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