Bạn đang xem bản rút gọn của tài liệu. Xem và tải ngay bản đầy đủ của tài liệu tại đây (15.52 MB, 301 trang )
J. S. B O E R
United States have become scarce as municipalities and environmental needs
compete for the same water. Salt-laden water usually is not an alternative because evaporation removes the water and not the salt, degrading the soil. As a
consequence, new irrigation is becoming less possible than in the past and there
is increasing interest in improving the efficiency of water use in irrigation and
determining whether plants can yield well under water deficient conditions (Boyer, 1982).
A number of methods exist for improving the efficiency of water use and have
been summarized by Taylor et al. (1983) and by Stewart and Nielsen (1990). The
methods can be classified in three broad categories: (1) increasing the efficiency
of water delivery and the timing of water application, (2) increasing the efficiency of water use by the plants, and (3) increasing the drought tolerance of the
plants. The first method is practiced most because it depends on engineering and
minimally on the crop. Transporting water with low evaporative loss, preventing
runoff, storing water in catchments, delivering water only to the root zone, and
timing irrigation to the needs of the plant have been successful in improving
productivity per unit of water delivered to the farm. There are estimates that just
by improving irrigation timing, the amount of applied water can be decreased by
half in some crops while maintaining high levels of production (e.g., Bordovsky
et al., 1974). The second and third methods depend on understanding the biology
of the crop and whether it can be manipulated to achieve the same productivity
with less water. The state of knowledge in this area is the focus of this review.
11. WATER USE EFFICIENCY
Water use efficiency (WUE) usually is defined as the total dry matter produced
by plants per unit of water used,
where D is the mass of dry matter produced (usually aboveground) and W is the
mass of water used (usually including direct evaporation from the soil). For a
field experiment, D and W would be expressed on the basis of land area. For a
single plant, D and W would be measured in the same plant and expressed on the
basis of the whole plant. Sometimes, the D is the economically valuable part of
the crop (for example the grain, tuber, or fruit) and WUE refers to yield. One
may also consider the water use efficiency of a single leaf, and so on. The higher
the production per unit of water use, the higher the efficiency.
There is extensive evidence that WUE varies among species in the same
environment and among climates for the same crop (Briggs and Shantz, 1914; de
ADVANCES IN DROUGHT TOLERANCE IN PLANTS
Wit, 1958; Hanks in Taylor er al., 1983; Tanner and Sinclair in Taylor et al.,
1983). Taking advantage of the species and climate effects can help manage
limited water supplies in agriculture. For example, alfalfa (Medicago sativa L.)
has a lower water use efficiency than maize (Zea mays L.) when grown in nearby
sites in the same year (Hanks in Taylor et al., 1983; Miller, 1938). Thus, simply
by changing crops, water consumption can be reduced with little sacrifice in dry
matter production. Relocating production to a new climate with lower evapotranspiration is another possible approach. For economic reasons, however,
these options are not often employed and probably will not be until the cost of
water rises to a level that forces change. What then are the prospects for improving water use efficiency within a species, or protecting against yield loss in a
particular climate when irrigation is not possible?
The dry mass of plants consists mostly of the C and 0 atoms fixed photosynthetically from CO,. These elements are much heavier than the H atoms that
also are prevalent in the dry mass and that originate in the water photolyzed
during the photosynthetic process. As a consequence, D of Eq. ( I ) represents
mostly the net C0,-fixing activity of the plant. Before fixation, the CO, must
diffuse into the leaf and dissolve in the wet surface of the cells where it becomes
available to be fixed. The wet surfaces are exposed to the atmosphere inside the
leaf and transpiration is inevitable. As a result, the photosynthetic cells dehydrate
to varying degrees. Water absorbed from the soil replenishes the water lost by the
cells, but the water potential of the cells must be low enough to maintain absorption. The stomata and waxy cuticle of the epidermis control the transpiration rate
and thus the amount of water needing to be absorbed, and lower water potentials
cause stomata1 closure. This regulation of transpiration and absorption affects the
balance between net CO, gain and water loss and in turn the WUE. Depending on
the leaf anatomy and physiology, the dry matter produced per unit of water used
can vary widely.
In addition to these factors, water use also is affected by physical factors. CO,
enters by diffusing down a concentration gradient to the leaf interior, and the
water vapor in the intercellular spaces inside the leaf likewise diffuses in the
opposite direction. The lower the external humidity, the faster transpiration will
be when all other factors are constant. Leaf temperature plays an important role
by affecting the vapor pressure of water in the leaf. The higher the leaf temperature, the higher the vapor pressure and the more rapid the transpiration. Water
use will differ among sites and seasons for these reasons and the water use
efficiency in Eq. ( I ) thus reflects a complex of plant and environmental factors.
Briggs and Shantz (1914) conducted an extensive survey of the water use
J. S. BOER
efficiency of crops, and they expressed it as the water requirement, that is, the
amount of water transpired to produce a unit Df aboveground dry matter, which is
the reciprocal of the water use efficiency. They grew the plants in large containers
of soil and made measurements for the entire growing season. This had the
advantage that a large number of crops could be compared in a uniform climate
during a single season. In their experiments, the transpiration ratio of maize,
sorghum (Sorghum bicolor L.), and millet (Panicurn rniliaceum L. and Sefaria
italica (L.) Pal. = Chaetochloa) was less than for the other crops and, although
Briggs and his co-workers could not have known at the time, the three crops are
C , species possessing a special anatomy and biochemistry that allows CO, to be
concentrated around the site of fixation. This resulted in more photosynthesis per
unit of water transpired and accounted for the lower water requirement.
After the experiments of Briggs and his co-workers, various investigators
measured water use efficiency under field conditions where all the adaptations of
the crop could express themselves (de Wit, 1958; Hanks in Taylor et a l . , 1983).
Typically, the experiments involved season-long exposure to differing amounts of
irrigation. Figure 1 shows examples for Logan, Utah, where wheat (Triticum
aestivum L.) and maize were grown with varying amounts of irrigation in 1975
Wafer Use Efficiency
EVAPOTRANSPIRATION (xi O6 Kg H2O.ha -’)
Figure 1 Production of aboveground shoot dry matter at various levels of water use in several
crops near Logan, Utah. The years in which the crops were grown are given in the symbol key. Water
use was controlled by irrigation that held conditions essentially constant for the growing season.
Water use is the combined evaporation from the soil and transpiration from the plants. A positive
evapotranspiration intercept indicates the amount of water obtained from soil stores. The slope of the
linear relation is the water use eficiency, which was 4.49 g of dry mass per kilogram of H,O for
maize, 2.50 for wheat, 2.36 for alfalfa, and 2.1 I for barley. Maize is a C, plant and the others are C,.
Adapted from Hanks (in Taylor et al., 1983).
ADVANCES IN DROUGHT TOLERANCE IN PLANTS
and alfalfa and barley were grown in other years (Hanks in Taylor et a / ., 1983).
There is a linear relationship between water use and dry matter production. The
linearity is mostly caused by the diffusion link between photosynthesis and
transpiration because the visible radiation input is almost completely absorbed by
crops after the canopy closes and, in a given climate, the input tends to be
partitioned in a constant proportion between energy for transpiration and energy
for photosynthesis. The slope of the linear relationship is the water use efficiency
[Eq. ( I ) ] , and the straight line indicates that the water use efficiency does not
vary as the availability of water varies. However, it differs among species, as is
apparent in Fig. 1 for the C, maize and C, wheat, alfalfa, and barley. These
experiments confirm the differences noted by Briggs and his co-workers and
further indicate that water use efficiency does not differ under varying availabilities of soil water in this type of experiment. However, it differs among species,
climates, and from year to year (Briggs and Shantz, 1914; Brown and Simmons,
1979; Garrity et al., 1982; Hanks in Taylor et al., 1983; Kawamitsu el al., 1987;
Pandey et a / . , 1984a,b; Robichaux and Pearcy, 1984; Tanner and Sinclair in
Taylor et al., 1983), and there is a possibility that it will vary with different
mineral nutrient availabilities, plant spacing, and other cropping practices.
In this respect, it is important to note that while differences between C, and C,
species are apparent, similar tests have not been made in species exhibiting
Crassulacean acid metabolism. Pineapple (Ananas cornoms (L.) Merr.) has this
form of metabolism and it concentrates CO, by temporarily fixing the gas in
organic acids at night and releasing it the next day for photosynthesis. During
release, the stomata are closed and water is conserved. This allows CAM plants
to achieve even higher water saving than C, plants, and limited estimates of
water use efficiency are about 20 g of aboveground dry mass per kilogram of
water for pineapple (Joshi et a/., 1965; Neales er a / ., I968), compared to 3 to 5
for C4 plants, and 2 to 3 for C3 plants.
An alternate approach to the usual description of water use efficiency is to
normalize water use for evaporative demand (de Wit, 1958; Tanner and Sinclair
in Taylor el al., 1983) and dry mass for the potential productivity of the crop
(Hanks in Taylor et a / . , 1983). Thus, modified expressions of WUE have been
used, such as
where the fractional dry mass is DID,,,, and is expressed relative to the maximum dry mass produced with optimum water D,,.
The fractional water use
WI W,,,,, is likewise expressed relative to the maximum evapotranspiration W,,,
that would occur with optimum water. This normalization procedure has the
advantage that for a water use of, say, half the maximum evapotranspiration, half
J. S. BOYER
EVAPOTRANSPIRATION(x106 Kg H2O.ha -’)
Figure 2 Seed yield at various levels of water use by three sorghum genotypes in West Central
Nebraska. The experiment was conducted under conditions similar to those of Fig. 1. The water use
efficiency for seed yield is the slope of the line and was 1.8 g of dry mass per kilogram of H,O in
RS626. 1.9 in NC+55X, and 1.2 in NB505. The water use efficiency for total shoot dry mass was
3.3 in RS626, 3.2 in NC+55X, and 2.0 in NB505. Adapted from Garrity er al. (1982).
the maximum dry mass would be predicted. This can simplify the job of predicting the impact of water shortages but it requires a knowledge of the maximum
dry matter yield and evapotranspiration of the crop for the year, which will vary.
For practical purposes, the maximum yield and water use usually are not
known and normalization may not be done easily, so the absolute expression in
Eq. (1) is preferable. Moreover, farm income for an irrigated crop is generally
based on the absolute dry mass or economic yield rather than normalized yields,
and expense is based on the absolute amounts of water used. There needs to be a
high absolute production of dry mass to justify pumping large amounts of water
and there should be a high production of marketable yield. For example, Fig. 2
shows that the water use efficiency differed for production of grain dry mass in
sorghum genotypes RS626 and NB505 (Garrity el al., 1982). Normalizing according to Eq. (2) would not distinguish which genotype gives the highest grain
production, but Eq. (1) would detect the difference.
OF HARVEST INDEX
The fraction of the crop that is economically valuable, termed the harvest
index, is part of the total dry mass and thus part of WUE. There has been a
ADVANCES IN DROUGHT TOLERANCE IN PLANTS
general increase in yields of modern crops with little change in the total aboveground biomass, and according to Gifford (1986) the increase is attributable to an
increase in the harvest index. The increase has come without much change in the
amount of water used, and the result has been a natural improvement in WUE for
yield (Richards et al., 1993). There is a maximum to which the harvest index can
be increased and the maximum probably is being approached in many modern
crops. Therefore, the maintenance of the harvest index is of critical importance
when water is in short supply.
In Fig. 2, the harvest index was nearly constant among the various treatments.
Instead, the differences in WUE were attributable mostly to differences in the
total aboveground dry matter (Garrity et a l . , 1982). The experiments involved
season-long steady exposure to limited water, and the crops acclimated by growing smaller shoots, with flowering and grain fill adjusting in proportion. In this
steady environment, the acclimation allowed the harvest index to be maintained.
Under many field conditions, however, plants encounter variable water deficits
that do not allow the acclimation possible in long-term experiments, and the
harvest index can decrease. This effect can be extreme, and water deficits can
give a harvest index as low as zero (see Boyle et a l . , 1991b, for an example).
Therefore, considerable opportunity exists for maintaining the harvest index in
the face of variable environments, which we will explore more fully later.
The most accurate means of measuring water use efficiency is to monitor the
evapotranspiration and harvest the crop for biomass measurements at the end of
the season. The WUE can be determined for the total biomass or any part of the
biomass. However, these are labor intensive and costly measurements. Less
expensive methods have been sought, and one has been to measure directly the
COz and H,O exchange of individual leaves (Bierhuizen and Slatyer, 1965;
Brown and Simmons, 1979; Robichaux and Pearcy, 1984). Because the CO,
molecule contributes most of the dry mass, the gas exchange efficiency can be
defined as the ratio of the mass of CO, gained to the mass of H,O lost. Martin
and Thorstenson (1988) compared the gas exchange efficiency with the actual
water use efficiency for the whole growing season in tomato (Lycopersicon
esculentum Mill. ), its wild relative Lycopersicon pentzeflii (Cor.) D’Arcy, and
hybrids between them. The relationship was poor because of additional factors
affecting dry mass accumulation but not gas exchange. For example, the mass of
the plant is determined not only by photosynthesis but also by respiratory losses
at night and partitioning to nonphotosynthetic organs such as roots. It is altered
by temperature and the molecular composition of the dry mass. Gas exchange for
short times during the day does not detect these additional factors. Therefore,
J. S. BOYER
while the gas exchange efficiency gives valuable insight into the physiologic and
metabolic controls that might operate during photosynthesis and transpiration,
the method is being used less frequently than in the past.
Another method is based on the relative abundance of natural isotopes in plant
tissue. Although most of the CO, in the atmosphere is T O , , a small amount is
T O 2 . Because the I2CO2 is lighter, it diffuses more rapidly than 13C02. Also,
ribulose- 1,5-bisphosphate carboxylase fixes the ' T O 2more rapidly than I3CO2.
Consequently, the cells accumulate relatively more 12C than 13C, and the unused
C diffuses out according to the extent of stomata1 opening. This outward diffusion is correlated with transpiration. Because the inward diffusion and use of
I2CO2correlates with photosynthesis and dry mass but the outward diffusion of
13C0, correlates with transpiration, the relative uptake of I2C and I3C correlates
with the water use efficiency. Generally, higher water use efficiency correlates
with lower tissue 12C relative to 13C for wheat, peanut (Arachis hypogaea L.),
barley (Hordeum vulgare L.), and other crops as shown in Fig. 3. Therefore, the
measurements detect differences in WUE among individuals within a species and
they only require the ratio of the isotopes in tissue samples to be compared to a
standard (Bowman et al., 1989; Brugnoli et al. 1988; Condon et al., 1987, 1990;
Farquhar and Richards, 1984; Hubick and Farquhar, 1989; Hubick et al., 1986).
The ratio technique makes it possible to survey a large number of plants at
moderate cost. Differences integrate the conditions over which the plants were
Figure 3 Water use efficiency and carbon isotope discrimination compared in various genotypes
of (A) wheat, (B) barley, (C) peanut, and (D) wheatgrass. Adapted from Farquhar and Richards
(1984), Hubick ef al. (1988). Johnson et at. (1990). Hubick and Farquhar (19891, and 'hrner (1993).
ADVANCES IN DROUGHT TOLERANCE IN PLANTS
grown. Analyzing the entire shoot indicates the water use efficiency for the time
required to grow the shoot whereas analyzing only leaf starch indicates the water
use efficiency during the time necessary to accumulate the starch. One may
integrate over long or short times which avoids one of the problems of the gas
Martin and Thorstenson (1988) used this technique to show that differences in
water use efficiency were present between the domestic tomato species and L .
pennellii and their hybrids. Differences in water use efficiency were detectable in
isotope ratio data between the parents and the hybrids particularly when water
was optimally available. The domesticated parent had the lowest efficiency, the
wild parent the highest efficiency, and the hybrids showed intermediate behavior.
Because the species could be crossed, it was possible to correlate the differences
in water use efficiency with restriction fragment maps of the tomato DNA (Martin et af.,1989). Three loci were found to be predictors of the variation in water
use efficiency in field grown tomato. This landmark effort indicates that water
use efficiencies are determined by relatively few genetic loci and implies not only
that agriculturally relevant differences exist but that they might be genetically
manipulated in a simple fashion.
The success of the method suggests that differences in water use efficiency
exist in individual species and might be usefully incorporated into breeding
programs, although this is still in its infancy (Bowman et al., 1989; Brugnoli et
al., 1988; Condon et a l . , 1987, 1990; Hubick and Farquhar, 1989; Hubick e t a l . ,
1986). Genetic variation clearly exists but in crop canopies the variation becomes
less clear. Substantial water limitation usually gives a negative relationship between discrimination and WUE as shown in Fig. 3, but under relatively favorable
conditions, the relationship tends to become less negative or even positive. In the
field, this can obscure relationships developed from pot experiments. For example, Condon and Richards (1993) showed that two wheat genotypes differing by
40-50% in leaf diffusive conductance only differed by 15% in canopy transpiration efficiency. In the field, this difference was not reflected in improved WUE for
the crop because soil evaporation differed in opposition, canceling the 15%
effect. There was a slower development of the canopy in one genotype than in the
other, and this was responsible for the canceling effect (Condon and Richards,
1993). Therefore, differences in WUE may need to be combined with other crop
traits to be realized as water savings.
Opportunities to improve WUE generally involve many genes and many interactions. Perhaps this is not surprising in view of the massive changes in plant
form and anatomy that have occurred as plants colonized the land. The develop-
J. S. BOYER
ment of roots, cuticle, stomata, vascular systems, and seeds all occurred after
plants left their original water environment and, in evolutionary time, these
developments are relatively recent. As a consequence, single genes that substantially change the WUE of plants are difficult to find. With the increased ability to
transform plants genetically in recent years, it would be desirable to apply the
tools of molecular biology to the improvement of WUE. However, the tools work
best when single genes are manipulated. The problem can be seen in C, species
where many genes code for enzymes and leaf anatomy that differ from those in
C, species. While C, photosynthesis confers a clear improvement in WUE, the
genetic complexity of the trait makes it difficult to incorporate in C, plants.
However, there are traits of more limited scope that probably have less complex
genetics and might benefit from the application of molecular biology. Some of
these are considered below as ways to improve the drought tolerance of the plant.
111. DROUGHT TOLERANCE
Plants showing improved growth with limited water are considered to tolerate
drought regardless of how the improvement occurs or whether the water use
efficiency is affected. Some species can avoid drought by maturing rapidly before
the onset of dry conditions or by reproducing only after rain. Examples of these
drought avoiders are ephemerals such as California poppy (Eschscholtzia californica (Cham.)) that can complete their life cycle in a few weeks, or tree crops
such as coffee (Cofea arabica (L.)) and cacao (Theobroma cacao (L.)) that
flower and fruit after drought followed by rain (Alvim, 1960, 1985). Others can
postpone dehydration by growing deep roots or sealing themselves tightly against
transpiration or accumulating large stores of water in fleshy tissues. Examples of
dehydration postponers are upland rice (Oryza sariva L.) with deep roots compared to paddy rice (Chang et al., 1974), or agave (Agave deserti (Engelm.)) or
saguaro cactus (Carnegiea gigantea (Engelm) Britt. and Rose) with thick cuticle
or fleshy tissue. Still other species allow dehydration of the tissues and simply
tolerate it by continuing to grow when dehydrated or by surviving severe desiccation. Certain intertidal algae such as Fucus vesiculosus (L.) or lower vascular
plants such as Selaginella lepidophylla (Hook. & Grev.) can carry out photosynthesis at very low water contents and tolerate desiccation to the air dry state
without losing viability. The seeds of most angiosperms also can tolerate severe
These effects are generally distinct from the factors controlling water use
efficiency. Drought avoiders depend on the timing of development which is under
internal control. They tend to reproduce themselves after a minimal accumulation of dry matter and their success ensures that they are represented in the next
ADVANCES IN DROUGHT TOLERANCE IN PLANTS
generation. Dehydration postponers having deep roots may have a water use
efficiency identical to that of other species but will accumulate more dry weight
because of their ability to gain access to a larger amount of water than shallow
rooted types. In effect the slope of the water use efficiency relation in Fig. 4 may
be the same but the deep rooted species work farther out on the curve. Their
adaptations are mostly structural and take time to build, requiring the expenditure
of photosynthetic products. Finally, dehydration tolerators may have the same
water use efficiency as dehydration-sensitive species when water is available but
the tolerators can grow at tissue hydration levels that the other species cannot.
Of the three forms of drought tolerance, dehydration tolerance is most intriguing because it often requires only slight repartitioning of dry mass. An example is
osmotic adjustment (Morgan, 1984; Munns, 1988) which occurs because dry
mass normally used to synthesize new cells instead accumulates in the cells as
solute (Meyer and Boyer, 1972; 1981) or is deposited in fewer or smaller cells
(Fraser et al., 1990; Sharp et a l . , 1990). Only a brief decrease in biosynthesis of
tissue is necessary to accomplish this (Meyer and Boyer, 1981), but the increased
concentration of solutes can markedly increase the ability of the cells to extract
water from the soil. The increased solute is present only under dry conditions. In
other words, there is little cost to the plant when water is scarce and no cost when
water is plentiful.
A. APPROACHESTO IMPROVING
From these examples it can be seen that water use efficiency is important, but
crop improvement under conditions of limited water involves more than water
EVAPOTRANSPIRATION (x106 Kg HpO.ha -')
Figure 4 Effect of increasing the amount of water available to a crop without changing the water
use efticiency. Production moves from A to B. An example might be increasing rooting depth.
J. S. BOYER
use efficiency. Basing improvement solely on water use efficiency is tempting
because breeding programs could select solely for high productivity when water
is plentiful. The idea is that, for a given climate, water use efficiency will be
highest when dry matter production is highest and the linear relationship of dry
matter to water use (Figs. 1 and 2) would allow the high productivity to carry
over to drought conditions. However, it is clear that many opportunities will be
missed if superior selections are based only on this concept of water use efficiency. Characters such as osmotic adjustment are called into play only during a
water deficit. Roots may penetrate deeper soil layers or leaves may persist better
during a water deficit in some genotypes than in others, and so on. Without plant
selection under water deficient conditions, these beneficial traits will be missed.
Figure 5 Yields of continuous maize crops at Morrow Plots at University of Illinois. The first
crop was planted in native prairie. Soil nutrients and organic matter were added to some of the plots
beginning around 1905. Hybrids were introduced to all plots in 1937. Yields are 5-year averages.
Genotypes are those popular at the time.