Comparative forage yield, water use, and water use
efficiency of alfalfa, crested wheatgrass and spring wheat
in a semiarid climate in southern Saskatchewan
Paul G. Jefferson and Herb W. Cutforth
Semiarid Prairie Agricultural Research Centre, Agriculture and AgriFood Canada, Box 1030, Swift Current,
Saskatchewan, Canada S9H 3X2 (e-mail: JeffersonP@agr.gc.ca). Received 14 July 2004, accepted 27 April 2005.
Jefferson, P. G. and Cutforth, H. W. 2005. Comparative forage yield, water use, and water use efficiency of alfalfa, crested wheatgrass
and spring wheat in a semiarid climate in southern Saskatchewan. Can. J. Plant Sci. 85: 877–888. Crested wheatgrass
(Agropyron cristatum L. Gaertn.) and alfalfa (Medicago sativa L.) are introduced forage species used for hay and grazing by cattle across
western Canada. These species are well adapted to the semiarid region but their long-term responses to water stress have not been previously
compared. Two alfalfa cultivars with contrasting root morphology (tap-rooted vs. creeping-rooted) and two crested wheatgrass
(CWG) cultivars with different ploidy level (diploid vs. tetraploid) were compared with continuously cropped spring wheat (Triticum
aestivum L.) for 6 yr at a semiarid location in western Canada. Soil water depletion, forage yield, water use efficiency, leaf water potential,
osmotic potential and turgor were compared. There were no consistent differences between cultivars within alfalfa or CWG for variables
measured. However, these two species exhibit different water stress response strategies. Leaf water potential of CWG was lower
during midday stress period than that of alfalfa or wheat. Alfalfa apparently had greater capacity to osmotically adjust to avoid midday
water stress and maintain higher turgor. Soil water use patterns changed as the stands aged. In the initial years of the trial, forage crops
used soil water from upper layers of the profile. In later years, soil water was depleted down to 3 m by alfalfa and to 2 m by crested wheatgrass.
Alfalfa was able to deplete soil water to lower concentrations than crested wheatgrass or wheat. Soil water depletion by wheat during
the non-active growth season (after harvest to fall freeze-up) was much less than for CWG or alfalfa as expected for annual vs.
perennial crops. As a result, more soil water was available to wheat during its active growth period. In the last 3 yr, the three species
depleted all available soil water. Forage yield responses also changed over time. In the initial 3 yr, crested wheatgrass yielded as much
as or more than alfalfa. For the last 3 yr of the experiment, alfalfa yielded more forage than crested wheatgrass. Forage crops deplete
much more soil water during periods of aboveground growth dormancy than wheat. Water use efficiency of crested wheatgrass declined
with stand age compared with fertilized continuous spring wheat. Alfalfa exhibited deep soil water extraction and apparent osmotic
adjustment in response to water stress while CWG exhibited tolerance of low water potential during stress.
Key words: forage yield, soil water, water potential, water use, water use efficiency, drought
Jefferson, P. G. et Cutforth, H. W. 2005. Comparaison du rendement fourrager, de l’utilisation de l’eau et de l’efficacité de
l’utilisation de l’eau chez la luzerne, l’agropyre à crête et le blé de printemps dans une région semi-aride du sud de la
Saskatchewan. Can. J. Plant Sci. 85: 877–888. L’agropyre à crête (Agropyron cristatum L. Gartn.) et la luzerne (Medicago sativa
L.) sont des espèces fourragères introduites employées pour la production de foin et la paissance des bovins partout dans l’ouest
du Canada. Ces espèces se sont bien acclimatées aux conditions semi-arides de la région, mais jamais encore n’a-t-on comparé
leur réaction à long terme au stress hydrique. Les auteurs ont comparé deux cultivars de luzerne aux racines de morphologie différente
(racines pivotantes ou traçantes) et deux cultivars d’agropyre à crête de ploïdie différente (diploïde et tétraploïde) à la
monoculture du blé de printemps (Triticum aestivum L.) pendant six ans, dans une région semi-aride de l’Ouest canadien. Les
paramètres comparés étaient l’extraction de l’eau du sol, le rendement fourrager, l’efficacité de l’utilisation de l’eau, le potentiel
hydrique foliaire, le potentiel osmotique et la turgescence. Les auteurs n’ont relevé aucun écart cohérent entre les cultivars de
luzerne et d’agropyre pour les variables examinées. Les deux espèces recourent toutefois à des stratégies différentes pour composer
avec le stress hydrique. L’agropyre a un potentiel hydrique foliaire plus faible que celui de la luzerne ou du blé pendant la
période de stress du milieu de la journée. Apparemment, la luzerne utilise mieux l’osmose pour éviter le stress hydrique du milieu
de la journée et rester turgescente. L’utilisation de l’eau du sol évolue avec l’âge du peuplement. Les premières années de l’expérience,
les cultures fourragères utilisaient l’eau présente dans les couches supérieures du profil. Vers la fin du projet, la luzerne
avait épuisé l’eau du sol jusqu’à 3 m de profondeur contre 2 m pour l’agropyre à crête. La luzerne extrait plus d’eau du sol que
l’agropyre à crête ou le blé. Le blé extrait beaucoup moins d’eau du sol pendant sa période de croissance non active (de la récolte
au gel en automne) que l’agropyre à crête ou la luzerne ainsi qu’on pourrait s’y attendre avec des cultures annuelles et vivaces. En
conséquence, le blé dispose d’une plus grande quantité d’eau pendant sa période de croissance active. Au cours des trois dernières
années, les trois espèces ont épuisé toute l’eau présente dans le sol. Le rendement fourrager a lui aussi évolué dans le temps. Au
cours des trois premières années, l’agropyre à crête avait un rendement identique ou supérieur à celui de la luzerne. Lors des trois
dernières cependant, le rendement de la luzerne dépassait celui de l’agropyre. Les cultures fourragères extraient beaucoup plus
d’eau du sol que le blé pendant la période de dormance des organes aériens. L’efficacité avec laquelle l’agropyre utilise l’eau
diminue avec l’âge du peuplement comparativement au blé de printemps cultivé continuellement avec fertilisation. La luzerne
extrait l’eau du sol en profondeur et procède apparemment à un ajustement osmotique quand l’eau vient à manquer alors que l’agropyre
à crête tolère la réduction du potentiel hydrique en période de stress.
Mots clés: Rendement fourrager, eau du sol, potentiel hydrique, utilisation de l’eau, efficacité de l’utilisation de l’eau, sécheresse
877
Abbreviations: CWG, crested wheatgrass; DM, dry matter; WUE, water use efficiency; ?pd, predawn leaf water potential;
?md, midday leaf water potential; ?pd, predawn turgor; ?md, midday turgor; ?pd, predawn osmotic potential; ?md, midday
osmotic potential
878 CANADIAN JOURNAL OF PLANT SCIENCE
Crested wheatgrass (CWG) (Agropyron cristatum L.
Gaertn.) has been seeded across the prairie region of western
Canada since the 1930s (Gray 1996) and used for hay
and spring grazing by beef cattle. It establishes quickly,
excludes weedy competitors, and complements native
rangeland by providing sustainable forage production during
the spring and fall (Knowles and Kilcher 1983). Forage
production is correlated to precipitation but this relationship
changes as the CWG stands age over the first 3 yr (White
1985). Thirty-five-year-old stands of CWG were stable
communities with 90% of plant biomass produced by CWG
and only 10% by native plants (Looman and Heinrichs
1973). These authors concluded that CWG is a long-term
replacement for native range for sustainable forage production.
In a recent report, Campbell et al. (2000) reported that
CWG in a long-term crop rotation experiment exhibited a
decline in forage production after 5 yr compared with continuous
spring wheat rotation. These results suggested that
CWG forage production may not be as sustainable over time
as Looman and Heinrichs (1973) had concluded.
Diploid CWG (A. cristatum) (2N = 2??= 14) is commonly
grown in western Canada while tetraploid CWG (A.
desertorum) (2N = 4?= 28) is more commonly grown in the
United States of America because the tetraploids are more
drought tolerant (Bruynooghe 1996). Kirk CWG is a natural
tetraploid cultivar of A. cristatum that exhibits improved
forage production in western Canada over diploid cultivars
(Knowles 1990). However, its adaptation to water stress was
not known.
Forage yield and quality of CWG can be improved by
seeding with alfalfa (Medicago sativa L.). Alfalfa roots
extract soil moisture below 1.5 MPa tension, which is normally
assumed to be the lower limit of available water to
crop plants (Cutforth et al. 1991). This drying of the soil
profile was greater with older alfalfa stands. Jefferson and
Cutforth (1997) reported that alfalfa yield was less dependent
on precipitation in the first 2 yr after establishment,
likely because the plants were exploiting stored soil water.
They suggested that soil water depletion should be measured
in dryland alfalfa stands to account for alfalfa’s apparent
drought tolerance. There has been confusion about the
impact of alfalfa root morphology on its capabilities for soil
water extraction. For example, Henry et al. (1987) describe
alfalfa as exhibiting a tap root that can reduce excess soil
water that contributes to salinization of prairie soils.
However, in a table of recommended forage species for
seeding on saline soils, these same authors recommend
Rambler alfalfa, a creeping-rooted cultivar. Creeping-rooted
cultivars were developed for persistence, particularly under
grazing (Heinrichs 1963), and we were not aware of any
research on the impact this trait might have on soil water
use.
A summary of reported alfalfa water use efficiency
(WUE) from eight locations in the USA indicated that it
produces 15.2 ± 2.1 kg ha–1 mm–1 (Sheaffer et al. 1988). In
California, irrigated alfalfa WUE was reported to be 23.2 kg
ha–1 mm–1 (Grimes et al. 1992). Dryland alfalfa WUE has
not been reported for semiarid conditions in western
Canada. Several water use and long-term crop production
studies at Swift Current, Saskatchewan have studied spring
wheat (Triticum aestivum L.) (Campbell et al. 1987, 1988).
We used this common annual crop for comparison to perennial
forage crops and to provide a benchmark that can be
readily compared with previous results.
The objective of this work was to compare spring wheat,
two cultivars of CWG, and two alfalfa cultivars for forage
production, water use, water use efficiency, and plant water
status in a long-term experiment at Swift Current,
Saskatchewan.
MATERIALS AND METHODS
An experiment was conducted at Swift Current,
Saskatchewan (50°16?N 107°44?W, elev. 825 m) on a
Brown Chernozem Swinton loam soil (Agriculture Canada
Expert Committee on Soil Survey 1987) from 1993 to 1998.
Two CWG cultivars, Kirk (Knowles 1990) and Parkway
(Alderson and Sharp 1994), two alfalfa cultivars, Beaver
(Bolton et al. 1963) and Rangelander (Heinrichs et al. 1979)
were compared with continuous spring wheat. These two
alfalfa cultivars remain popular with local producers despite
their age. They represent contrasting root morphology as
Beaver exhibits a tap-root architecture while Rangelander is
predominantly (80%) creeping-rooted (Heinrichs et al.
1979). The experimental design was a randomized complete
block of five treatments with nine replications. The replications
were grouped three per range (oriented east–west) on
three ranges stacked south to north. The individual plots
were 12 seeded rows spaced 0.3 m apart for plot dimensions
of 3.6 ??36 m. CWG and alfalfa were seeded on 1992 May
25 at 6 and 4 kg ha–1, respectively. Annual weeds were controlled
by clipping during 1992 and no data were recorded
during the establishment of perennial forage plots.
The site was irrigated during the spring of 1993 to raise
the soil water concentration to field capacity to a depth of 3
m prior to seeding wheat. A total of 175 mm of irrigation
water was applied during a 7 d period. Irrigation ensured
that the soil water concentration was as uniform as possible
in order to compare water depletion by depth as suggested
for alfalfa by Jefferson and Cutforth (1997). A 21-mo fallow
period at this location can store an additional 96 mm of soil
water in dryland rotations (Campbell et al. 1987) while irrigated
alfalfa typically requires about 400 mm of irrigation
water at this location (Pohjakas et al. 1967). So we concluded
that the irrigation water provided only in 1993 to permit
root exploration of the soil profile did not invalidate the
characterization of this trial as dryland.
This irrigation resulted in a delay of spring growth by the
alfalfa and CWG and a delay in seeding the spring wheat in
1993 (Table 1). Spring wheat was no-till seeded at recommended
rates (Table 1) on the same plots year after year and
fertilized with N at 80 kg ha–1 and P2O5 at 31 kg ha–1, while
alfalfa and CWG were not fertilized because this is the typical
agronomic practice of forage producers in this region. The
wheat cultivar was Leader in 1993 and 1994 but this was
changed to Lancer from 1995 to 1998. Lancer has greater
genetic resistance to wheat stem sawfly (Cephus cinctus), a
common problem on continuous wheat crops in this region.
Weeds were controlled in the spring wheat with post-emerJEFFERSON
AND CUTFORTH — WATER USE EFFICIENCY OF FORAGE CROPS IN A SEMIARID LOCATION 879
gence herbicide application. In 1992, aluminum access tubes
were installed to 3-m depth in the centre of each plot.
Volumetric soil water was determined by neutron probe at 10,
30, 50, 70, 90, 110, 130, 150, 170, 190, 210, 230, 250 and 270
cm depths at spring greening, forage harvest, wheat harvest,
and freeze-up in the fall to track soil water use by the plants.
Available soil water (field capacity minus permanent wilting
point) is 268 mm in this soil. Non-growing season water use
was summed from harvest date to fall freeze-up. Daily rainfall
and Class-A pan evaporation were recorded at an automated
weather station located 200 m from the experiment.
Water use (rainfall plus soil water depletion) was summed
from spring green-up to fall freeze-up for all three crops.
Growing season water use was summed from seeding to grain
harvest date for wheat or from spring green-up to forage harvest
date for alfalfa and CWG.
Two leaves were sampled on plants adjacent to neutron
access tubes for replications 3, 4, and 5 at several dates each
season for determination of plant water status from 1994
to1998 (Table 1). The youngest, fully expanded leaf of
CWG or wheat or the youngest, fully expanded trifoliate
leaf of alfalfa were excised with a razor blade and placed in
a plastic bag to minimize water loss (Turner and Long
1980). The leaf was then placed in a pressure chamber apparatus
(PMS Instrument Co., Corvallis, OR) with the cut end
exposed. The chamber was pressurized with N2 gas until
water extruded from xylem vessels when viewed with a 10?
binocular microscope mounted above the chamber. Leaf
water potential (?) was determined at pre-dawn (?pd) and
midday (?md) periods, which represent minimum and maximum
diurnal water stress. At the same time, leaf tissue of
the same age and orientation was excised and placed in 5-
mL plastic syringes with a small amount of glass wool in the
tip end. These syringes were placed in a freezer to rupture
the cell membranes of the leaves. Within 3 mo, these samples
were thawed and the plant material manually pressurized
with the plunger to exude sap (cell solubles and
apoplastic water) on to filter paper disks. The osmotic
potential (?) was determined with a Vapour Pressure
Osmometer (Model 5500XR, Wescor, Logan, UT) without
correction for apoplastic water dilution of cell solubles.
Turgor ( ?) was determined from the components of water
potential (?) within plant tissues (Fitter and Hay 1983):
??= ??– ??(1)
where ??is osmotic potential and ?is turgor. The water potential
component for gravity was negligible (<0.01 MPa) (Fitter
and Hay 1983). Negative turgor values were obtained by calculation
of ??but this reflected an underestimation of ??due to
dilution of cell sap by apoplastic (cell wall) water (Campbell
et al. 1979). We report the negative ?values but recognize that
negative turgor does not occur in nature. Water potential,
osmotic potential and turgor readings from both predawn and
midday samples were analyzed by date within each year.
Forage yield of CWG was determined in late June or
early July of each year (Table 1) by harvesting 0.6 ??36 m
area of each plot with a flail plot harvester and recording the
fresh forage mass. A 300-g subsample was weighed, dried
in a forced-air oven at 60°C for 48 h and reweighed to determine
dry matter (DM) concentration. Forage DM yield was
calculated for each plot. Regrowth in 1997 was sufficient for
a second harvest for alfalfa (Table 1) but not for CWG.
Biomass yield (grain, chaff, and straw) of wheat was determined
at the time of grain yield harvest. In 1994 and 1997
wheat biomass yield data were lost so biomass yield of
wheat was derived by calculation based on grain yield and
harvest index for each plot.
Water use efficiency was derived by dividing forage DM
yield by water use (Eq. 2) from spring green-up to fall
freeze-up, similar to the season-long approach of Scheaffer
et al. (1988). No attempt was made to separate soil water
depletion into evaporation and transpiration components.
WUE (kg ha–1 mm–1) = forage DM yield (kg ha–1) (2)
_ water use (mm)
We considered the Campbell et al. (1988) approach to calculate
soil water use by wheat from seeding to harvest plus growing
season precipitation. This would have created a
Table 1. Dates of seeding and harvest for continuous wheat, dates of spring growth initiation, fall freeze-up, and harvest on forage crops, dates of
water potential sampling, and total growing season (April to August) precipitation for 6 yr
1993 1994 1995 1996 1997 1998
Wheat seeding date Jun. 09 Apr. 29 May 09 May 23 May 14 May 13
Wheat harvest date Sep. 08 Aug. 22 Aug. 22 Aug. 29 Aug. 21 Sep. 03
Forage spring green date May 28 Apr. 11 Apr. 12 Apr. 17 Apr. 30 Apr. 08
Forage harvest date Aug. 09 Jun. 28 Jul. 06 Jun. 27 Jul. 11 Jul. 08
21 Aug.
Fall freeze-up datez Oct. 19 Oct. 19 Oct. 16 Oct. 16 Oct. 27 Sep.03x
Water potential sampling dates None May 18 May 18 May 29 Jun. 02 May 26
Jun. 01 Jun. 01 Jun. 12 Jun. 17 Jun. 04
Jun. 15 Jun. 14 Jun. 26 Jul. 01 Jun. 23
Jun. 30 Jun. 28 Jul. 10 Jul. 15 Jul. 07
Jul. 13 Jul. 12 Jul. 24 Jul. 28 Jul. 21
Jul. 27 Jul. 26 Aug. 08 Aug. 11 Aug. 11
Aug. 10 Aug. 22
Aug. 24
zLast date of soil water monitoring after a –5°C frost.
xTermination date of experiment.
880 CANADIAN JOURNAL OF PLANT SCIENCE
confounding factor in our calculations, however, because the
spring wheat would have a different period of water use than
CWG or alfalfa. Our approach to calculating WUE includes
soil water use by the perennial crops during periods of aboveground
growth dormancy, such as during late summer and in
the autumn, because this water use is important for maintaining
crown and root tissues and carbon reserves for regrowth
initiation. Restricting WUE calculation to water used only during
the period of active aboveground growth would over-estimate
WUE of perennial crops. We concluded that the water
use from green-up to fall freeze-up would be the best period
for comparison between annual and perennial crops.
Analysis of variance based on a two-factor [replication (n
= 9), crop (n = 5)] model was calculated for each year or
sampling date with JMP software (SAS Institute, Inc.1995).
Analysis of variance in JMP is based on the General Linear
Model of SAS and least square means were generated.
Contrasts were calculated to compare: forage crops vs.
spring wheat, CWG vs. alfalfa, Rangelander vs. Beaver
alfalfa, and Kirk vs. Parkway CWG.
RESULTS AND DISCUSSION
Water Use
From green-up to freeze-up, differences between species in
water use (WU) occurred during the first 3 yr of the study
(Table 2). There were no differences in WU between cultivars
within alfalfa or CWG (data not shown). Therefore,
amounts and patterns of WU for a given species were determined
by averaging across cultivars within species. Alfalfa
used more water than CWG in 1994 and 1995 and more
water than wheat in 1993 and 1995. There were no differences
between species in WU from 1996 to 1998, although
alfalfa tended to use more water than CWG or wheat. CWG
used more water than wheat only in 1993. Because the time
periods from green-up and freeze-up were the same for both
perennial species, water use per day (WU d–1) followed the
same significance and trends as WU.
Because the duration of the growing season (growth period
over which machine-harvestable dry matter was produced)
was 19 or more days longer for wheat compared with alfalfa
and CWG, wheat used more water during the growing season
than alfalfa and CWG (Table 2). Except for 1993 and 1996,
alfalfa used more water during the growing season than
CWG. The opposite occurred for WU d–1. During 1993 and
1994 WU d–1 was higher for alfalfa and CWG compared with
wheat. Thereafter, WU d–1 was highest for wheat. Alfalfa and
CWG were not fertilized, but wheat was fertilized. During
1993 and 1994, fertility and water were sufficient to support
good growth of all three species. Thus, despite the much
shorter growing season, WU for alfalfa and CWG was only
slightly less than and WU d–1 slightly greater than for wheat.
The growth of alfalfa and CWG decreased substantially compared
with wheat in the last 3 yr, resulting in wheat using
more water and having higher daily water use compared with
alfalfa and CWG. This decline in growth may be related to the
decline in natural soil fertility as the stands aged.
The duration of the non-growing season was much longer
for alfalfa and CWG than for wheat (Table 2). Partly
because of the longer duration, WU during the non-growing
season was much greater for alfalfa and CWG compared
with wheat. The differences in non-growing season WU
were so large that, despite the longer durations, WU d–1 was
larger for alfalfa and CWG compared with wheat. In general,
there were no differences in WU and WU d–1 between
alfalfa and CWG.
The WU of irrigated alfalfa at this site was 637 mm
(Pohjakas et al. 1967), including 400 mm of irrigation water.
Irrigated spring wheat used 526 mm of soil water and irrigation
over the whole season (Pohjakas et al. 1967). In comparison,
dryland alfalfa in our experiment used an average
of 414 mm and wheat used an average of 387 mm of soil
water. Our data suggest that perennial forage crops will use
re-charge soil water as effectively as an annual crop such as
wheat and can also exploit soil water stored in deeper soil
layers than wheat.
Precipitation received from fall freeze-up to seeding for
wheat or to spring green-up for CWG and alfalfa varied
among years (Table 3), reflecting the different growth periods
for these species. Over-winter precipitation was lowest for the
1998 season and was highest for 1996 (wheat) or 1997 (CWG
and alfalfa). Wheat benefitted from precipitation received
from harvest to freeze-up, which contributed to soil water
recharge. This was accounted for in water use efficiency calculations
as the additional soil water would be measured in the
spring of the next year. For alfalfa and grass, a large proportion
of the forage harvest to freeze-up precipitation would be
transpired as the plants are still living, and this transpired water
would, therefore, be lost from the soil profile and unavailable
for biomass production the following year. However, photosynthesis
during the aboveground growth dormancy period
likely contributed to root energy reserves that are essential to
regrowth in the next spring. Approximately 25% more rainfall
occurred during the growing season for wheat (seeding to harvest)
than for alfalfa or grass (green-up to harvest).
Evaporation indicated that alfalfa and CWG experienced
water stress in 1993 and moist conditions in 1996 (Table 3).
Wheat experienced water stress in 1998.
Soil Water Depletion
From 1993 to 1995, alfalfa depleted more water than CWG
and wheat from lower depths (Fig. 1). For example, alfalfa
depleted more water than wheat from the surface to 1.3-m
layer in 1993, and more water than wheat and CWG from the
1.3-m to 2.1-m layer in 1994, and from the 1.7-m to 2.3-m
layer in 1995. From 1996 to 1998, in general, alfalfa depleted
more water than CWG and wheat from below 1.7 m.
Generally, throughout the 6 yr of the study, alfalfa tended to
deplete more water than CWG and wheat from below 1.5 m.
Only in 1993 did CWG deplete more water than wheat to 1.3
m, whereas the reverse occurred in 1994. Thereafter, CWG
and wheat depleted similar amounts of water to 1.3 m. In most
years CWG and wheat depleted similar amounts of water from
the 1.3-m to 2.1-m layer except in 1994 when CWG depleted
more water from 1.5- and 1.7-m depths than wheat.
The species-dependent water use patterns were a reflection
of the rooting depth differences between species. Our
soil water depletion data suggested that the maximum rootJEFFERSON
AND CUTFORTH — WATER USE EFFICIENCY OF FORAGE CROPS IN A SEMIARID LOCATION 881
ing depth for wheat was about 1.5 m, about 2.3 m for CWG,
and >2.7 m (probably at least 3 m) for alfalfa. Campbell et
al. (1988) also reported water use by spring wheat to 1.5 m
depth. In contrast to our results, Bittman (1985) reported
soil water depletion by CWG ranged from 0.8 to 1.2 m depth
at Melfort, Saskatchewan.
Freeze-up Soil Water Content
The differences between species in freeze-up soil water content
with depth reflects the species differences in soil water
use patterns (Fig. 2). As the years progressed, the thickness
of the soil layer where alfalfa had a lower soil water content
at freeze-up compared with CWG increased and moved
Table 2. Duration (d) and water use (WU) and water use per day (WU d–1) by alfalfa, crested wheatgrass (CWG) and spring wheat from green-up
to freeze-up, and for the growing season and for the non-growing season when no harvestable dry matter was produced
Green-up to freeze-up Growing season Non-growing season
Days WU WU d–1 Days WU WU d–1 Days WU WU d–1
Year Crop (mm) (mm) (mm)
1993 Alfalfa 145 450 3.1 73 311 4.3 72 139 1.9
CWG 145 433 3.0 73 298 4.1 72 135 1.8
Wheat 145 389 2.7 92 346 3.8 53 42 1.0
SEz 22** 0.2** 23** 0.3** 20** 0.3**
1994 Alfalfa 192 453 2.4 79 306 3.9 113 148 1.3
CWG 192 424 2.2 79 290 3.7 113 134 1.2
Wheat 192 440 2.3 116 366 3.2 76 74 0.9
SE 19* 0.1* 14** 0.2** 14** 0.1**
1995 Alfalfa 183 467 2.6 86 149 1.7 97 318 3.3
CWG 183 406 2.2 86 126 1.6 97 280 2.9
Wheat 183 405 2.2 106 308 3.0 77 97 1.2
SE 18** 0.1** 13** 0.1** 12** 0.2**
1996 Alfalfa 184 432 2.3 72 163 2.3 112 269 2.4
CWG 184 413 2.2 72 141 2.0 112 273 2.5
Wheat 184 416 2.3 99 362 3.7 85 54 0.6
SE NS NS 23** 0.3** 19** 0.2**
1997 Alfalfa 181 408 2.3 72 234 3.3 109 174 1.6
CWG 181 385 2.1 72 199 2.8 109 186 1.7
Wheat 181 405 2.2 99 299 3.0 81 107 1.2
SE NS NS 19** 0.2** 20** 0.2**
1998 Alfalfa 121 274 2.3 64 123 1.9 57 151 2.7
CWG 121 269 2.2 64 105 1.6 57 164 2.9
Wheat 121 269 2.2 114 257 2.3 7 12 1.2
SE NS NS 14** 0.2** 14* 0.4*
zSE is the standard error.
*, ** significantly different at P < 0.05, and P < 0.01, respectively; NS, non-significant at P = 0.05.
Table 3. Precipitation and evaporation (Class A pan) totals for the time periods indicated. In 1993, the study started on May 28, shortly after the soil
profile had been filled to field capacity to 3 m by irrigation. The study ended on 8 September 1998, the day spring wheat was harvested. Therefore,
freeze-up’ in 1998 was Sep. 08
Wheat Alfalfa/crested wheatgrass
Freeze-up seeding harvest to Freeze-up green-up 1st cut 1st/2nd
Year to seeding to harvest freeze-up to green-up to 1st cut to 2nd cut cut to freeze-up
Precipitation (mm)
1993 – 314 60 – 173 – 210
1994 104 173 62 94 146 – 99
1995 99 261 104 74 176 – 213
1996 196 139 112 117 158 – 168
1997 174 199 45 172 148 53 45
1998 57 200 – 43 170 – 44
Evaporation (mm)
1993 – 617 97 – 710 – 92
1994 NA 662 251 NA 400 – 728
1995 38 742 208 NA 452 – 536
1996 99 719 155 NA 289 – 696
1997 106 766 315 NA 540 333 315
1998 110 959 – NA 494 – 574
NA not available due to low temperature.
882 CANADIAN JOURNAL OF PLANT SCIENCE
downward. For example, in 1993, alfalfa had a lower soil
water content than CWG from 1.1 m to 1.5 m, but the thickness
of this layer increased to 1.1 m to 2.1 m in 1994, and to
0.3 m to 2.3 m in 1995. As the upper portion of the soil profile
dried, alfalfa used more water at deeper depths so that
by 1997 and 1998, alfalfa had the lowest soil water content
at the deeper depths compared with CWG and wheat.
Even though for any given year CWG did not deplete significantly
more water than wheat from below 1.3 m (Fig. 1),
CWG tended to deplete more water than wheat from the 1.3-m
to 2.1-m layer. Therefore, the freeze-up soil water content from
1.3 m to 2.1 m was significantly lower for CWG compared
with wheat from the beginning to the end of the study (Fig. 2).
By 1997, all three species had used similar amounts of
water from the surface to 0.9-m depth so there were no differences
between species in soil water to 0.9 m (Fig. 2).
There were differences in soil water content below 0.9 m
with alfalfa having the lowest soil water content below 0.9
m, and with CWG having a lower soil water content than
wheat from 1.3 m to 2.1 m.
Forage Yield
Forage yield comparisons among the crops changed with
time. In 1993, CWG yielded more than alfalfa, which produced
more than spring wheat (Table 4). In 1994, alfalfa
yielded more than CWG while spring wheat was intermediate.
From 1995 to 1998, spring wheat biomass yield was
greater than alfalfa forage yield, except in 1997 when they
were similar. During 1994 to 1998, CWG was the lowestyielding
species. Looman and Heinrichs (1973) reported
Fig. 1. Annual soil water depletion curves to 2.8 m for alfalfa, crested wheatgrass and spring wheat from 1993 to 1998. LSD bars (P < 0.05)
at a given depth are for comparing means among species.
JEFFERSON AND CUTFORTH — WATER USE EFFICIENCY OF FORAGE CROPS IN A SEMIARID LOCATION 883
that CWG forage yield stabilized at a low level over time,
but we did not observe this trend in our results. There were
no differences between Kirk and Parkway CWG or between
Rangelander and Beaver alfalfa for forage yield.
Forage yield at this location is correlated to growing season
precipitation but newly seeded forage stands may be
less dependent on precipitation while depleting soil water
from deeper layers (Jefferson and Cutforth 1997). Forage
producers in this region generally recognize the higher productivity
of newly seeded stands and adjust their management
by preferentially haying younger stands and deferring
haying or grazing of older stands during “wet” growing seasons.
WUE
As previously observed for forage yield, WUE differences
changed with time (Table 5). In 1993, CWG exhibited the
highest WUE, while spring wheat had the lowest. The WUE
values for CWG in 1993 were similar to those reported for
this species in 1984 at Melfort, Saskatchewan (Bittman
1985). In 1994, alfalfa exhibited the highest WUE while
CWG had the lowest. In 1995, there were no crop species
differences in WUE, but from 1996 to 1998, alfalfa had the
highest and CWG had the lowest WUE. There was a significant
negative correlation between time (age of stand in
years) and WUE in CWG (r = –0.89, n = 6, P = 0.05). White
(1985) reported a curvilinear response in CWG between age
Fig. 2. Fall freeze-up soil profile water content to 2.8 m for alfalfa, crested wheatgrass, and spring wheat from 1993 to 1998. LSD bars (P
< 0.05) at a given depth are for comparing means among species.
884 CANADIAN JOURNAL OF PLANT SCIENCE
of stand and WUE and argued that CWG yield and WUE
stabilized 3 yr after establishment. We did not observe this
response as WUE appeared to decline linearly with age of
stand in this study. White (1985) used only seasonal precipitation
and did not measure soil water depletion in his calculation
of WUE. So his values may differ from ours
because we accounted for soil water depletion. The alfalfa
WUE values in 1997 (Table 5) were similar to those reported
by Sheaffer et al. (1988), while values from 1994 to 1996
were similar to those reported by Grimes et al. (1992). As
with forage yield and WU, there were no cultivar differences
for WUE between Kirk and Parkway CWG or
between Beaver and Rangelander alfalfa (Table 5). The
creeping-rooted trait in Rangelander alfalfa was selected to
impart persistence and tolerance to grazing (Heinrichs 1963)
and did not exhibit improved soil water uptake characteristics
or WUE compared with tap-rooted Beaver alfalfa.
Tetraploid and diploid Russian wildrye [Psathyrostachys
juncea (Fisch.) Nevski] had similar WUE at a site in North
Dakota (Frank and Berdahl 1999). Our results suggest that
the reported yield advantage of tetraploid A. desertorum
cultivars over diploid A. cristatum (Bruynooghe 1996) cultivars
in semiarid environments may be due to traits other
than ploidy level.
The WUE of continuous spring wheat in this study was
similar to previous reports by Campbell et al. (1987). Those
authors reported considerable variation in WUE over 18 yr
of fertilized continuous spring wheat rotation, so the yearto-
year variation in our 6-yr study was not surprising. The
most interesting result of our study was alfalfa’s superior
WUE compared with CWG or continuous wheat, particularly
in the last 3 yr (1996 to 1998). Alfalfa is frequently grown
under irrigation in North America and is known to respond
to supplementary water application (Sheaffer et al. 1988).
These results clearly show that dryland alfalfa has WUE
superior to either CWG or continuous wheat. This observation
leads us to speculate that biological N fixation in alfalfa
may contribute to improved WUE by reducing N
limitation to growth under water-limiting conditions.
However, CWG experienced both water- and N-limiting
conditions. While we did not examine the impact of fertility
on WUE of CWG, we speculate that additional fertility
would improve it. Producers in this region argue that large
annual variation in precipitation makes fertilization of
perennial forages uneconomical. If long-range weather predictions
could be made more accurately in the future, then
this agronomic practice should be re-evaluated.
Water Potential
Midday water potential (?md) over the growing season
exhibited different responses related to the weather and
potential water stress. For example, ?md declined during the
Table 4. Forage yield of crested wheatgrass and alfalfa and biomass of continuous spring wheat grown on dryland for 6 yr at Swift Current,
Saskatchewan
1993 1994 1995 1996 1997 1998 6-yr mean
Crop Cultivar (Mg ha–1)
CWG Kirk 8.71 3.49 2.42 1.94 2.92 0.77 3.38
Parkway 9.47 4.27 2.37 1.82 2.50 0.56 3.50
Alfalfa Beaver 8.08 6.33 3.13 3.79 5.43 2.01 4.80
Rangelander 8.29 6.20 3.23 4.01 5.89 1.94 4.93
Spring wheat 7.14 5.50 6.48 5.03 4.87 2.37 4.93
Contrast pobabilities
Wheat vs. CWG & Alfalfa <0.01 NS <0.01 <0.01 0.03 <0.01 <0.01
CWG vs. Alfalfa 0.02 <0.01 <0.01 <0.01 <0.01 <0.01 <0.01
Kirk vs. Parkway NS 0.07 NS NS NS 0.08 NS
Beaver vs. Rangelander NS NS NS NS NS NS NS
SE 1.12 0.90 0.67 0.43 0.83 0.25 0.35
Table 5. Water use efficiency (WUE) of crested wheatgrass, alfalfa, and continuous spring grown for 6 yr at Swift Current, Saskatchewan
1993 1994 1995 1996 1997 1998 6 year mean
Crop Cultivar kg ha–1 mm–1
CWG Kirk 28 12 20 11 8 5 14
Parkway 32 14 19 11 6 2 14
Alfalfa Beaver 26 22 22 23 15 16 21
Rangelander 26 20 22 24 16 18 21
Spring wheat 20 15 23 11 14 8 15
Contrast probabilities
Wheat vs. CWG & Alfalfa <0.01 0.09 NS <0.01 <0.01 0.06 <0.01
CWG vs. Alfalfa 0.01 <0.01 0.08 <0.01 <0.01 <0.01 <0.01
Kirk vs. Parkway 0.08 NS NS NS NS 0.08 NS
Beaver vs. Rangelander NS NS NS NS NS NS NS
SE 4 4 4 4 2 3 2
JEFFERSON AND CUTFORTH — WATER USE EFFICIENCY OF FORAGE CROPS IN A SEMIARID LOCATION 885
summer season in 1994 but it declined much less in 1996
(Fig. 3). This was consistent with evaporation potential differences
between these two seasons (Table 3).
There were no consistent differences between Rangelander
and Beaver alfalfa cultivars or between Kirk and
Parkway CWG cultivars. The contrasts for CWG or for
alfalfa cultivars were significant only twice for ??in each
species. Therefore, we present species means of all sampling
dates within each year to illustrate species comparisons.
Mean predawn water potential (?pd) over all dates within a
year did not differ among the crops except in 1998 (Fig. 4A).
Fig. 3. Mid-day leaf water potential for three crop species in each of 5 yr at Swift Current, Saskatchewan. Vertical bars indicate LSD value
among means on sampling dates where crops were significantly different.
886 CANADIAN JOURNAL OF PLANT SCIENCE
In 1998, alfalfa and CWG had lower ?pd values than wheat.
The ?pd values from 1994 to 1997 suggest that all three
species began each sampling day with similar water status.
The ?md differed among the crops in each year (Fig. 4B).
In 1994 and 1995, CWG exhibited the lowest ?md values
and wheat the highest. In 1994, CWG leaves experienced
–3.0 MPa water potential, a value that indicated severe
water stress. In 1996, both CWG and wheat had lower water
potentials than alfalfa. In 1997 and 1998, CWG had the lowest
?md values while alfalfa and wheat were similar.
Overall, CWG experienced the lowest water potentials or
the highest degree of water stress among the three species.
This suggests that CWG exhibits a different strategy to cope
with water stress than alfalfa or wheat. Alfalfa and wheat
maintain high water potential in their tissues in order to
avoid stress (Fitter and Hay 1983) but CWG tolerates very
low ?md within its leaf tissues during the mid-day period of
water stress (Fig. 4B) and then recovers to water potential
values similar to alfalfa and wheat at night (Fig. 4A). These
?md values for CWG are similar to drought-stressed values
reported by Bittman and Simpson (1989). They found that
CWG ?md values were lower than those of Altai wildrye
[Leymus angustus (Trin.) Pilger] or smooth bromegrass
(Bromus inermis Leyss.) at a site in northeastern
Saskatchewan. They also reported that CWG exhibited leaf
rolling under water stress (Bittman and Simpson 1989) as a
strategy to reduce transpirational water loss from adaxial
leaf stomata and radiation interception by reducing leaf area.
We did not rate leaf rolling in our study but it was observed
within the CWG plots.
Predawn osmotic potential (?pd) differed among the
crops in every year (Fig. 5A). Wheat consistently had the
highest (least negative) ?pd values in every year. Alfalfa
and CWG had similar ?pd values in every year except 1997
when alfalfa was lower than CWG. Midday osmotic potentials
(?md) varied among crops in every year except 1995
(Fig. 5B). As with ?pd observations, ?md was higher for
wheat in every year. CWG and alfalfa had similar ?md in
most years.
Predawn turgor ( ?pd) differed among the crops in 1995,
1996 and 1998 (Fig. 6A). In those 3 yr, either alfalfa or
CWG had the highest ?pd while wheat had the lowest in 2 of
3 yr. Midday turgor ( ?md) varied among the crops in all 5 yr
(Fig. 6B). CWG had the lowest ?md in 4 out of 5 yr, while
alfalfa had the highest ?md in 1994, 1996, 1997 and 1998.
High ?md would permit continued carbon assimilation, cell
extension, and growth during the period of water stress.
Alfalfa maintains high turgor during low water stress periods
at predawn, and during high water stress periods during
the day. This is a very different strategy compared with
CWG, which allows very low ?md and ?md values to occur
in its tissues. CWG does not appear to adjust to water stress
through the active accumulation of osmotic solutes. Rather
through an elastic cell wall, CWG tolerates low ?md and
exhibits other water-conserving strategies such as leaf
rolling (Bittman and Simpson 1989). Active osmotic adjustment
occurs in water-stressed wheat leaves (Morgan 1984).
Measure of active osmotic adjustment is done at full turgor
or by extrapolation of ??vs. relative water content curve to
its intercept point. We did not measure either relative water
content or ??at full turgor, so we cannot conclusively state
that CWG has a more elastic cell wall than wheat. However,
it is the most logical explanation of the very low ?md
observed for this species.
As ?md declines, plants that can osmotically adjust to
water stress show a declining ?md value, so the slope of the
regression of ?md against ?md can give an indication of
osmotic adjustment (S. Angadi pers. comm.). We found a
higher slope (0.61 ± 0.05 MPa MPa–1) for alfalfa than for
CWG (0.40 ± 0.08 ) or wheat (0.39 ± 0.09 ) (P < 0.01, n =
32, 28, 20 respectively, regressions not shown). This suggests
that alfalfa is osmotically adjusting to water stress
more than wheat or CWG.
Alfalfa exhibits high levels of abscisic acid (ABA), a
plant hormone that regulates stomatal aperture and conductance
of water vapour from leaf surfaces (Chen and Chen
1988). ABA control of transpirational water loss may allow
alfalfa to tolerate very low ?md observed in our study.
Alfalfa and CWG were identified as drought-tolerant forage
crops during the 1930s and 1940s. Our results confirm
that they respond to water stress with a different combination
of strategies to avoid (alfalfa) or tolerate (CWG) leaf
tissue water stress.
CONCLUSIONS
Alfalfa cultivars with very different root architecture did not
differ in soil water depletion, forage yield, water use efficiency,
or water potential. Both alfalfa cultivars extracted
soil water to a depth of 2.7 m and this trait contributes to
alfalfa’s drought tolerance.
Two CWG cultivars from Agropyron cristatum genetic
material but with contrasting ploidy (diploid vs. tetraploid)
did not differ in the traits we measured.
Alfalfa and CWG used soil water during periods with no
machine-harvestable aboveground growth, usually during
July, August, and September. Perennial forage crops in
semiarid environments maintain root and crown tissues during
periods of growth dormancy and this maintenance costs
soil water. The proportion of soil water used during growth
dormancy can be equivalent to soil water used for forage
production.
Dryland alfalfa had 30% higher WUE than CWG or
wheat. These results illustrate the essential role of alfalfa for
productive forage hay and pasture systems in the semiarid
region of western Canada.
Alfalfa and CWG exhibited different responses to water
stress. CWG exhibited very low water potentials during
midday water stress, which is indicative of tissue tolerance
to low water potential and very low turgor. Alfalfa exhibited
apparent osmotic adjustment to maintain leaf turgor during
midday water stress.
ACKNOWLEDGMENTS
The authors thank Doug Judiesch, Rod Ljunggren, Russ
Muri, Marquita Duncan, and many summer students for
technical assistance, Darcy Schott for preparation of graphics,
Jodiene Cooke for the preparation of the tables, and Drs.
JEFFERSON AND CUTFORTH — WATER USE EFFICIENCY OF FORAGE CROPS IN A SEMIARID LOCATION 887
Fig. 5. Leaf osmotic potential at predawn (?pd) (A) and midday (?md) (B) sampling times averaged over all sampling dates within each year
for 1994 to 1998 for three crop species. Vertical bars indicate SE value among means.
Fig. 4. Leaf water potential at predawn (?pd) (A) and midday (?md) (B) sampling times averaged over all sampling dates within each year
for 1994 to 1998 for three crop species. Vertical bars indicate SE value among means.
Fig. 6. Leaf turgor at predawn ( ?pd) (A) and midday ( ?md) (B) sampling times averaged over all sampling dates within each year for 1994
to 1998 for three crop species. Vertical bars indicate SE value among means.
888 CANADIAN JOURNAL OF PLANT SCIENCE
Brian McConkey and Sangu Angadi for pre-submission
review of the manuscript. The helpful suggestions from two
anonymous reviewers are also acknowledged.
Agriculture Canada Expert Committee on Soil Survey. 1987.
The Canadian system of soil classification. 2nd ed. Agriculture
Canada, Ottawa, ON. Publication 1646. 164 pp.
Alderson, J. and Sharp, W.C. 1994. Grass varieties in the United
States. USDA-ARS. Handbook No. 170 Washington, DC. 296 pp.
Bittman, S. 1985. Physiological and agronomic responses to
drought of three forage grasses: crested wheatgrass, smooth
bromegrass, and Altai wildrye. Ph.D. thesis, University of
Saskatchewan, Saskatoon, SK. 244 pp.
Bittman, S. and Simpson, G. M. 1989. Drought effect on leaf
conductance and leaf rolling in forage grasses. Crop Sci. 29:
338–344.
Bolton, J. L., Peake, R. W. and Downey, R. K. 1963. Beaver
alfalfa. Can. J. Plant Sci. 43: 615–617.
Bruynooghe, J. D. 1996. Forage production and performance of
beef yearlings grazing diploid and tetraploid crested wheatgrasses.
M.Sc. thesis, University of Saskatchewan, Saskatoon, SK. 127 pp.
Campbell, G. S, Papendick, R. I., Rabie, E. And Shayo-Ngowi,
A. J. 1979. A comparison of osmotic potential, elastic modulus,
and apoplastic water in leaves of dryland winter wheat. Agron. J.
71: 31–36.
Campbell, C. A., Zentner, R. P. and Johnson, P. J. 1988. Effect
of crop rotation and fertilization on the quantitative relationship
between spring wheat yield and moisture use in southwestern
Saskatchewan. Can. J. Soil Sci. 68: 1–16.
Campbell, C. A., Zentner, R. P., Selles, F., Biederbeck, V. O.,
McConkey, B. G., Blomert, B. and Jefferson, P. G. 2000.
Quantifying short-term effects of crop rotations on soil organic
carbon in southwestern Saskatchewan. Can. J. Soil Sci. 80:
193–202.
Campbell, C. A., Zentner, R. P. and Steppuhn, H. 1987. Effect
of crop rotations and fertilizers on moisture conserved and moisture
use by spring wheat in southwestern Saskatchewan. Can. J.
Soil Sci. 67: 457–472.
Chen, T. H. H. and Chen, F. S. C. 1988. Relations between photoperiod,
temperature, abscisic acid, and fall dormancy in alfalfa
(Medicago sativa). Can. J. Bot. 66: 2491–2498.
Cutforth, H. W., Jefferson, P. G. and Campbell, C. A. 1991.
Lower limit of available water for three plant species grown on a
medium textured soil in southwestern Saskatchewan. Can. J. Soil
Sci. 7l: 247–252.
Fitter, A. H. and Hay, R. K. M. 1983. Environmental physiology
of plants. Academic Press, London, UK. 355 pp.
Frank, A. B. and Berdahl, J. D. 1999. Soil water use by diploid
and tetraploid Russian wildrye. Crop Sci. 39: 1101–1106.
Gray, J. H. 1996. Men against the desert. Western Producer
Prairie Books, Saskatoon, SK.
Grimes, D. W., Wiley, P. L. and Sheesley, W. R. 1992. Alfalfa
yield and plant water relations with variable irrigation. Crop Sci.
32: 1381–1387.
Heinrichs, D. H. 1963. Creeping alfalfas. Adv. Agron. 15:
317–337.
Heinrichs, D. H., Lawrence, T. and McElgunn, J. D. 1979.
Rangelander alfalfa. Can. J. Plant Sci. 59: 491–492.
Henry, L., Harron, B. and Flaten, D. 1987. The nature and management
of salt-affected land in Saskatchewan. Saskatchewan
Agriculture, Agdex 518.
Jefferson, P. G. and Cutforth, H. W. 1997. Sward age and
weather effects on alfalfa yield at semi-arid location in southwestern
Saskatchewan. Can. J. Plant Sci. 77: 595–599.
Knowles, R. P. 1990. Registration of Kirk crested wheatgrass.
Crop Sci. 30: 749.
Knowles, R. P. and Kilcher, M. R. 1983. Crested wheatgrass.
Agriculture Canada, Research Branch Contribution 1983–18E. 18 pp.
Looman, J. and Heinrichs, D. H. 1973. Stability of crested
wheatgrass pastures under long-term pasture use. Can. J. Plant Sci.
53: 501–506.
Morgan, J. M. 1984. Osmoregulation and water stress in higher
plants. Ann. Rev. Plant. Physiol. 35: 299–319.
Pelton, W. L. and Korven, H. C. 1969. Evapotranspiration estimates
in a semiarid climate. Can. Agric. Eng. 11: 50–53.
Pohjakas, K., Read, D. W. L. and Korven, H. C. 1967.
Consumptive use of water by crops at Swift Current,
Saskatchewan. Can. J. Soil Sci. 47: 131–138.
SAS Institute, Inc. 1995. JMP statistics and graphics guide.
Version 3.2 SAS Institute, Inc., Cary, NC. 593 p.
Sheaffer, C. C., Tanner, C. B. and Kirkham, M. B. 1988.
Alfalfa water relations and irrigation. Pages 373–409 in A. A.
Hanson, D. K. Barnes, and R. R. Hill, Jr., eds. Alfalfa and alfalfa
improvement. No. 29 in Agronomy Series, ASA, Madison, WI.
Turner, N. C. and Long, M. J. 1980. Errors arising from rapid
water loss in the measurement of leaf water potential by the pressure
chamber technique. Aust. J. Plant Physiol. 7: 527–537.
White, L. M. 1985. Stand age, precipitation, and temperature
effects on forage yield. J. Range Manage. 38: 39–43.