Coping Mechanisms for Crop Plants in Drought-prone Environments

Abstract

Background

Drought is a major limitation to plant productivity. Various options are available for increasing water availability and sustaining growth of crop plants in drought-prone environments.

Scope

After a general introduction to the problems of water availability, this review focuses on a critical evaluation of recent progress in unravelling mechanisms for modifying plant growth responses to drought.

Conclusions

Investigations of key regulatory mechanisms integrating plant growth responses to water deficits at the whole-organism, cellular and genomic levels continue to provide novel and exiting research findings. For example, recent reports contradict the widespread conception that root-derived abscisic acid is necessarily involved in signalling for stomatal and shoot-growth responses to soil water deficits. The findings bring into question the theoretical basis for alternate-side root-irrigation techniques. Similarly, recent reports indicate that increased ABA production or increased aquaporin expression did not lead to improved drought resistance. Other reports have concerned key genes and proteins involved in regulation of flowering (FT), vegetative growth (DELLA), leaf senescence (IPT) and desiccation tolerance (LEA). Introgression of such genes, with suitable promoters, can greatly impact on whole-plant responses to drought. Further developments could facilitate the introduction by breeders of new crop varieties with growth physiologies tailored to improved field performance under drought. Parallel efforts to encourage the introduction of supplementary irrigation with water made available by improved conservation measures and by sea- or brackish-water desalination, will probably provide comprehensive solutions to coping with drought-prone environments.

Growth, drought resistance, flowering, senescence, root signals, ABA, FT, DELLA, IPT, LEA, desalination, conservation

INTRODUCTION

Sub-optimal availability of water for unrestricted plant growth and transpiration, i.e. drought, is a major limitation to agricultural production (Boyer, 1982; Delmer, 2005). Water can become limiting for agricultural (and natural) plant communities as a result of inadequate rainfall, excessive levels of salts in the soil solution or the increasing diversion of limited fresh-water resources to competing urban and industrial uses. Future water availability may also be affected by ongoing changes in global climate.

The first plant-stress symptom induced by drought is often a rapid inhibition of shoot and, to a lesser extent, root growth. This is closely followed by partial or complete stomatal closure with associated reductions in transpiration and CO₂ uptake for photosynthesis. If not relieved, drought then leads to interrupted reproductive development, premature leaf senescence, wilting, desiccation and death (Hsaio, 1973; Schulze, 1986).

Drought situations can be classified as either terminal or intermittent. During terminal drought, the availability of soil water decreases progressively and this leads to premature plant death. Intermittent drought is the result of finite periods of inadequate irrigation occurring at one or more intervals during the growing season and is not necessarily lethal. In the poorer countries of the world, the ability to survive longer and maintain even limited plant development under intermittent or terminal drought conditions may allow for the production of subsistence yields that make a contribution to the economy of the local population. Such subsistence yields are of no economic value to farmers in more developed countries. The ideal goal for all farmers is to grow plants that can survive and maintain relatively high yields despite drought conditions. After a general introduction to the problems of water availability and potential solutions, this review considers recent advances in selected areas of basic plant research, which might facilitate the eventual introduction by breeders of plant varieties that can show improved performance in drought-prone environments.

Options for improving crop performance in drought-prone environments

One ideal approach for avoiding the drought symptoms induced by inadequate rainfall is to utilize water reserves to provide supplementary crop irrigation. If sufficient water reserves are not available, their availability can be increased by the appropriate introduction of water conservation, water engineering and agronomic technologies (Tal, 2006). For example: Another less expensive approach is associated with the observation that farmers in semi-arid regions may be routinely applying more irrigation water to their crops than the amount needed to ensure maximum harvestable yields. The potential advantages of regulated-deficit irrigation are of interest in this context. During deficit irrigation, carefully effected reductions in the amount and timing of supplementary irrigation may save water without reducing the amount or quality of the harvested yield (Chaves et al., 2007; Fereres and Soriano, 2007; Tambussi et al., 2007).

1. construction of river dams and reservoirs, rainwater catchments and non-leaky water distribution networks;

2. recycling of urban waste water for use in irrigation;

3. introduction of water-saving irrigation technologies and preferential cultivation of crops with relatively low water requirements;

4. construction of units for desalination of sea-water or brackish-water and inclusion of desalinated water sources in irrigation schemes in order to reduce potential build-up of salinity in artificially irrigated soils.

In addition, the performance of plants under drought conditions might be improved through selection, breeding or genetic-engineering approaches. This requires clear and achievable goals. Condon et al. (2004) listed some potential goals for breeders aiming to increase water-use efficiency from rain-fed or irrigated grain crops: Appropriate modifications to such goals would be required for perennial horticultural crops such as vines or fruit trees, where residual soil water after harvest may play an essential role in determining the quantity and quality of successive harvests.

1. move more of the available water through the crop while minimizing (a) evaporation from the soil surface, (b) drainage beyond the root zone, and (c) water left behind in the root zone after harvest;

2. acquire more photosynthate in exchange for each unit of water transpired during CO₂ fixation by the crop;

3. partition more of the acquired photosynthate into harvestable product.

Recent advances in plant genomics have led to the identification of a vast number of potentially beneficial water-stress-related genes, plus technologies for gene over-expression or silencing. Moreover, these can be introduced into transgenic plants under the control of appropriate promoters and are transmitted to subsequent generations (Delmer, 2005; Ma and Bohnert, 2007). The characterization of key plant physiological mechanisms that restrain plant performance under drought, together with the associated regulatory genes, could therefore facilitate the development by breeders of improved crop varieties showing increased water-use efficiency and drought resistance. The next section critically reviews selected aspects of recent whole-plant and genomic research that could be relevant to such developments. The interested reader is also referred to some excellent recent reviews of related topics in plant water-stress research by Sperry et al. (2002), Chaves et al. (2003), Bray (2004), Vinocor and Altman (2005), Humphreys et al. (2006), Maggio et al. (2006), and Shinozaki and Yamaguchi-Shinozaki (2007).

PLANT GROWTH

Ultimately, the optimization of plant performance and hence crop sustainability under variable environmental stress conditions will be dependent on the degree to which plant vegetative and reproductive growth patterns can be regulated. Plant growth is the result of daughter-cell production by meristematic cell divisions and subsequent massive expansion of the young cells. Cell expansion is in turn dependent on biophysical changes, which include a regulated loosening of primary cell walls and subsequent yielding to the hydrostatic (turgor) pressure generated by solute and water uptake into the cells (Neumann, 1995; Cosgrove, 1997). Cell growth rate and its inhibition by water deficit are regulated by a complex, multigenic series of metabolic processes and are not simply a function of water availability for turgor maintenance (e.g. Bassani et al., 2004; Fan and Neumann, 2004; Achard et al., 2006; Fan et al., 2006; Ma and Bohnert, 2007; Smith and Stitt, 2007). Ongoing research into key mechanisms regulating plant growth should make directed alteration of plant growth responses to drought via genetic manipulations an increasingly realistic goal.

Reproductive growth

An instructive example of altered growth patterns providing an adaptation to terminal drought environments is seen in some desert plants. These have evolved a much shortened growth cycle in which limited vegetative growth, flowering and seed set are completed within a few weeks of the end of the rainy season, i.e. before the onset of the hot, dry summer period. Such an adaptation would clearly be undesirable in crop plants, which ideally develop relatively large yields over long growth periods. However, a moderate shortening of the vegetative growth period and associated acceleration of the onset of flowering might be a desirable goal for crops grown without supplemental irrigation in terminal-drought environments. It is therefore encouraging that understanding of the processes regulating the onset of flowering has recently increased.

For example, the long search for the identity of ‘florigen’, a photoperiod-regulated factor that acts to induce flowering, may now be nearing an end. Thus, a leaf-produced, phloem-transported protein (FT) that can induce floral transition in shoot apical meristems has been genetically characterized in diverse crop species (Lifschitz et al., 2006; Corbesier et al., 2007; Lin et al., 2007; Tamaki et al., 2007). In contrast, Razem et al. (2006) characterized a protein (FCA) that acts as an abscisic acid receptor and then interacts with RNA metabolism to delay the onset of flowering in Arabidopsis. This suggests a possible mechanistic link between the increases in ABA levels induced by water deficits and delayed onset of flowering. Conceivably, the manipulation of the activity of such proteins could help to accelerate the onset of flowering so as to minimize the loss of reproductive yields associated with premature drought-induced desiccation of still-vegetative plants. The introduction by breeders of new varieties with shortened growing season might be useful to growers, for example when climate predictions indicate a drought year.

Root growth

Ongoing root growth is essential for effective plant retrieval of untapped water and mineral nutrient reserves in the soil. Paradoxically, root growth is inhibited by water deficits. Root growth inhibition involves spatially variable reductions in growth rate, outward proton pumping, wall extensibility and gene expression (Wu et al., 1996; Fan and Neumann, 2004; Fan et al., 2006; Ma and Bohnert, 2007). For example, the inhibition by water deficits of growth and wall extensibility in the more basal regions of the maize (Zea mays) root elongation zone is associated with down-regulation of wall acidification, up-regulation of CCR gene transcription, lignin deposition and phenolic cross-linking in the cell walls (Fan et al., 2006). These authors suggested that the basally localized inhibition of growth could promote the survival of the more apical meristem region by increasing the relative availability of water, minerals and sugars no longer required for growth. Increased meristem survival would serve to maintain the important ability to rapidly renew root growth after renewal of irrigation.

However, for plants in deep soils, the inhibition of root growth can have negative implications. As the upper soil layers dry down, the remaining water reserves may be increasingly situated in deeper soil layers. In this situation, the onset of more severe water stress could be delayed and survival enhanced by an improved ability to maintain root growth towards the water in the deeper soil layers. Such effects have been predicted by computer simulations and recently quantified in field experiments (Sinclair and Muchow, 2001; Padilla and Pugnaire, 2007). Thus, the selection, breeding or genetic engineering of varieties that better maintain root growth rates during drought situations could be a rewarding approach for crops on deep soil profiles.

Interestingly, stress-induced abscisic acid (ABA) accumulation, in addition to its well-characterized effects on stomatal closure, appeared to promote root growth in maize seedlings at low water potentials (Saab et al., 1990). In addition, exogenous ABA increased root hydraulic conductivity in soybean (Glycine max, Glinka, 1980). Conversely, a deficiency in endogenous ABA was shown to decrease root hydraulic conductivity in tomato (Solanum lycopersicum) plants (Tal and Nevo, 1973). Root hydraulic conductivity determines the ease with which water can be taken up and transported within the plant (LoGullo et al., 1998; Lu and Neumann, 1999). Thus, increased endogenous ABA levels produced by over-expression of ABA synthesis genes could conceivably result in increased root hydraulic conductivity, decreased stomatal transpiration and hence more drought-resistant plants. Thompson et al. (2007) investigated this approach by generating transgenic tomato lines that over-express a gene encoding 9-cis-epoxycarotenoid dioxygenase, an enzyme that catalyses a rate-limiting step in ABA biosynthesis.

In greenhouse trials, higher endogenous ABA levels in the transgenic plants were accompanied by partial stomatal closure and improved water-use efficiency under well-watered conditions. On the whole, growth of wild-type and transgenic plants under well-watered conditions were similar. Soil water was therefore conserved by the transgenic plants. However, the responses to water deficit were disappointingly similar for the wild-type and transgenic plants. Moreover, time to seedling establishment and partitioning of biomass to reproductive growth were both reduced in the transgenic plants. Thus, the path to enhanced ABA genotypes showing advantageous phenotypes in the field will most likely require additional developments.

Root hydraulic conductivity tends to decrease under water deficit situations (LoGullo et al., 1998; Lu and Neumann, 1999). In many cases, the symplastic pathway of radial water transport through the roots, a contributor to total root hydraulic conductivity, appears to be influenced by the number and activity of aquaporin water channels in the plasma membranes of the root cells (see reviews by Javot and Maurel, 2002; Tyerman et al., 2002; Maurel, 2007). Transgenic plants that over-express aquaporin genes might therefore show increased hydraulic conductivity, i.e. plant resistance to uptake of available water would decrease and growth potential might then be better maintained. However, increased root hydraulic conductivity might also lead to reduced soil–root hydraulic contact or facilitate unwanted water efflux if the water potential in the drying soil were to drop below that in the root. Moreover, Kaldenhoff et al. (1998) found that antisense-induced reductions in aquaporin expression in tobacco resulted in compensatory increases in root mass, so that the overall water-supply capacity of the root system was maintained. This suggests that aquaporin over-expression might cause compensatory reductions in root growth.

In all events, recent investigations into the effects of constitutive over-expression of aquaporin genes have given little indication that it is beneficial to drought resistance. For example, Aharon et al. (2003) found that over-expression of the plasma membrane aquaporin (PIP1b) in Arabidopsis improved the vigour of well-watered plants but had adverse effects on plants under drought or salt stress. Similarly, Katsuhara et al. (2003) found that over-expression of a barley (Hordeum vulgare) aquaporin gene (PIP2;1) increased root hydraulic conductivity in transgenic rice (Oryza sativa) plants but reduced root and shoot growth under salt stress. Finally, Jang et al. (2007) examined the effects of drought stress on transgenic Arabidopsis and tobacco (Nicotiana tabacum) plants that constitutively over-expressed Arabidopsis PIP1;4 or PIP2;5. No significant differences in growth rates and water transport were found between the transgenic and wild-type plants under favourable growth conditions. However, under drought stress, the transgenic plants over-expressing PIP1;4 or PIP2;5 showed retarded germination and growth as compared with wild-type plants. The use of drought-induced promoters rather than constitutive promoters might perhaps lead to more encouraging findings.

Shoot growth

A primary response to water deficit is the inhibition of shoot growth. This response can benefit drought survival by progressively limiting the leaf area available for evaporative loss of limited soil water reserves. In addition, the photosynthetic production of sugars is limited under water deficit and root uptake of mineral ions is likely to be similarly impaired: the inhibition of leaf growth may then allow diversion of essential solutes from growth requirements to stress-related house-keeping functions, such as osmotic adjustment that improves cell water retention and turgor maintenance. Shoot growth inhibition in response to water deficits may therefore extend the period of soil water availability and plant survival and can be considered as an adaptive response (Chapin, 1991; Neumann, 1995; Aachard et al., 2006).

However, stress adaptations that increase the chances of plant survival under stressful conditions by inhibiting growth will also decrease plant size and hence limit yield potential. The inhibition of shoot growth could certainly be a counter-productive response in the case of crop plants exposed to episodes of moderate water deficit. In such cases, plant survival might not be threatened but stress-induced reductions in shoot growth would still limit yield potential. Development of new varieties with an ability to maintain near-normal shoot growth rates during intermittent episodes of moderate water or salinity stress might therefore be advantageous for crop production under such conditions (cf. Neumann, 1997).

Interestingly, Achard et al. (2006) found that nuclear-localized DELLA proteins are involved in a general mechanism for integrating plant growth inhibitory responses. They were able to show that seedlings with quadruple knock-out mutations of the appropriate DELLA genes produced more than twice as much dry matter as wild-type plants after 14 d in a stressful environment (root media containing 100 mm NaCl). More recently, Djakovic-Petrovic et al. (2007) have shown that DELLA proteins are also involved in growth regulation by environmental changes in mature plant canopies. The goal of using molecular manipulations to reduce the degree of plant growth inhibition under moderate water deficits may therefore become more feasible as the factors associated with DELLA gene expression are further clarified.

Drought-induced leaf senescence is not directly related to shoot vegetative growth but it may be beneficial for transferring nutrients from leaves to growing seeds in terminal-drought scenarios (Yang and Zhang, 2006; Foulkes et al., 2007). In contrast, premature induction of leaf senescence processes in vegetative plants exposed to intermittent episodes of drought will progressively limit photosynthetic capacity and lead to leaf death. In an exciting recent development, Rivero et al. (2007) have linked a drought-activated promoter (SARK, senescence-associated receptor protein kinase) to an IPT gene that promotes cytokinin synthesis. Cytokinins are known to inhibit leaf senescence. The constructs were introduced into tobacco plants. In pot trials, two third-generation lines of the resultant transgenic plants were better able to survive and effect growth recoveries after intermittent water deficit episodes than the wild-type plants. Moreover, shoot development in well-watered transgenic plants did not appear to be adversely affected. It remains to be seen whether similarly encouraging findings can be demonstrated under field conditions.

ROOT-TO-SHOOT SIGNALS REGULATING STOMATAL APERTURE AND SHOOT GROWTH

The effects of water deficits on root growth, hydraulic architecture, stomatal aperture and shoot growth will ultimately determine crop yield potentials. The mechanisms by which growth and other responses of roots and shoots to soil water deficits are integrated at the whole-organism level therefore continue to generate considerable research interest. A particularly attractive hypothesis in this context is that roots can sense water deficits in the rhizosphere and can then induce appropriate adaptive responses by transmitting chemical, hydraulic or electrical signals to the shoot. Research into the identities of stress-induced signals from the roots, and their effects on stomatal closure and/or shoot growth inhibition, can suggest new approaches to regulating whole-plant responses to drought. For example, the plant hormone abscisic acid, which can inhibit stomatal opening and possibly growth in water-stressed leaves, is present in the xylem sap that is transported from the roots to the shoots of transpiring plants. Moreover, sap levels of ABA can increase in response to root water deficits. Thus, the xylem transport of root-derived ABA signals (possibly modulated by xylem pH and nitrate levels) has been considered a primary effecter of shoot changes induced by root water deficits (reviewed by Davies and Zhang, 1991; Wilkinson and Davies, 2002; Davies et al., 2005).

The ABA signaling hypothesis has provided a theoretical basis for modified water-deficit irrigation techniques in which different sides of the root system are subjected to alternate drying and irrigation. This is thought to enhance ABA signalling from the dryer roots to the shoots without inducing excessive limitations to soil and plant water availability. Increased root delivery of ABA to the leaves is then expected to partially reduce stomatal opening and increase water-use efficiency by inhibiting water transport to a slightly greater degree than CO₂ transport (Kang et al., 1998; Davies et al., 2002).

However, an increasing body of evidence now indicates that, in many cases, hydraulic signals are likely to be primarily involved in root-to-shoot signalling of drought, rather than root-derived ABA. For example, rapid decreases in leaf cell wall extensibility and leaf growth in maize seedlings were induced by root water deficits (root media containing non penetrating PEG 6000 at –0·5 MPa water potential); the growth inhibitory responses to root water deficit occurred even when the root tissues were killed, i.e. when no hormonal or electrical signals could be transmitted to the leaves (Chazen and Neumann, 1994). It therefore seemed that hydraulic signals from the stressed roots, in the form of increased xylem tension, were involved in inducing the leaf growth responses; albeit, levels of ABA in the elongation zones of the expanding maize leaves were increased by water stress (Chazen et al., 1995). Subsequently, Neumann et al. (1997) studied the effects of the same water-stress regime on ABA accumulation in the leaf elongation zones of maize and wheat (Triticum aestivum) seedlings with live or killed roots. The degree of stress-induced ABA accumulation in the leaves of maize seedlings was reduced if the roots were killed. Some of the leaf ABA increase could therefore have been root derived. However, similar stress-induced increases in ABA levels in the leaves of wheat seedlings were assayed in seedlings with live or killed roots; the source of the stress-induced increases in ABA accumulation therefore appeared to be the affected leaves themselves. Conceivably, root-derived hydraulic signals initiated increased ABA synthesis in the leaves. It has long been known that ABA can be independently produced by water-stressed leaves that are detached from the roots (Wright and Hiron, 1969).

Further evidence against the hypothesized involvement of root-derived ABA signals in leaf growth inhibition by water deficits was provided by Voisin et al. (2006). In a carefully controlled set of water-stress experiments they assayed leaf growth and ABA levels in six sense or antisense transformants for ABA synthesis. In addition, they assayed four lines showing natural variability in ABA levels. Leaf elongation was assayed at night to avoid the complicating effects of potential differences in stomatal opening. Moreover, different genotypes were grown in the same pots to ensure identical soil water status. Importantly, they found little or no correlation between xylem ABA levels and the reductions in leaf elongation rates induced in maize plants subjected to gradual water stress. Voisin et al. also investigated the hypothesis that increased rates of ethylene production under water deficit might interact with ABA in regulating leaf elongation (cf. Sharp, 2002). However, lines that showed elevated rates of ethylene production under water deficit showed the same inhibition of leaf elongation as that shown by lines with eight-times lower ethylene production rates.

Evidence against the involvement of root-derived ABA signals in the stomatal closure induced by water stress was reported by Holbrook et al. (2002). They grafted roots or shoots of ABA-deficient mutants onto wild-type tomato plants and measured leaf responses to soil water deficits. Their findings indicated that stomatal closure in response to soil drying was not dependent on ABA production by the roots. They suggested that another chemical signal from the roots might initiate the increased ABA production observed in the leaves and in turn induce stomatal closure. Small peptide signal molecules with hormone-like activity might be potential candidates for this role. They are known to induce a wide range of local or systemic responses in plants (Ryan et al., 2002; Matsubayashi and Sakagami, 2006). Moreover, Neumann (2007) found that peptide signal activity was present in the xylem sap of tomato plants and that its level was increased by salinity stress. Although the precise identity of the signal peptide(s) and in vivo site(s) of action remain to be determined, this finding confirms that signal compounds other than ABA may be involved in chemical communication between roots and shoots.

Additional evidence against the involvement of root-derived ABA signals in stress-induced stomatal closure was recently obtained by use of a sensitive and non-destructive ABA imaging technique. This revealed progressive increases in leaf ABA accumulation in Arabidopsis thaliana at 2, 4, 6 and 10 h after root exposure to water deficit (Christmann et al., 2005). However, increased root accumulation of ABA was only observed after 10 h of water deficit. The leaf increases therefore did not appear to be root derived. Moreover, Christmann et al. (2007) generated reciprocal grafts between wild-type scion and root stock from wild-type or ABA-deficient Arabidopsis mutants. Grafted plants containing a wild-type scion on root stock derived from either wild-type or ABA-deficient plants closed their stomata similarly in response to root water deficit. Thus, there appeared to be no requirement for root-derived ABA. These results confirmed and extended the findings of Holbrook et al. (2002) with grafted tomato plants. Christmann et al. (2007) also showed that the stomatal closure induced in Arabidopsis by root-applied water deficits could be relieved by supplying water directly to the leaves. Bogoslavsky and Neumann (1998) found that the leaf growth inhibition induced in maize seedlings by root-applied water deficit could also be relieved, at least temporarily, by injecting water into the leaf elongation zone. Thus, in these cases, hydraulic restraints appeared to be the primary limiting factor for both stomatal closure and leaf growth inhibition.

Taken together, the findings reviewed in this section indicate that the proposed role of root-produced ABA in signalling rhizosphere water status to the shoots requires further evaluation. The theoretical justification for split-root irrigation technologies, or other efforts to manipulate root-ABA delivery so as to increase water-use efficiency and drought resistance, do not appear to be adequately established.

This review has focused on critical evaluations of potential modifications to root and shoot growth responses to drought that might facilitate future improvements in plant performance. The beneficial effects reported for enhancement of DELLA gene expression and for drought-activated promoters of cytokinin synthesis provide reasons to be cautiously optimistic about future developments. Moreover, research into, and possible manipulation of, many additional features of plant metabolism that are not directly related to growth, such as (1) improving the antioxidant status of water- or salt-stressed tissues (Shalata and Neumann, 2001; Huang et al., 2005; Tokunaga et al., 2005; Eltayeb et al., 2007), (2) minimizing the onset of drought-induced xylem embolisms and acceleration of the refilling process (Sperry et al., 2002; Bucci et al., 2003; Salleo et al., 2004; Stiller et al., 2005) or (3) introgression of LEA (late embryogenesis abundant) genes associated with dessication resistance (Ingram and Bartels, 1996; Xiao et al., 2007) may also lead to future improvements in plant performance under drought. The report by Xiao et al. (2007) is particularly encouraging as it demonstrates increased drought resistance in transgenic rice under field conditions.

CONCLUSIONS

The era of rapid progress in plant genomics, proteomics, transcriptomics, metabolomics, organomics and systems' biology is clearly upon us and we are probably on the threshold of breakthroughs in our ability to understand and manipulate plant physiological responses to water deficit. Ongoing investigations into the key regulatory mechanisms integrating plant growth responses to water deficits therefore continue to provide a fascinating and important topic of basic plant research. Such research may facilitate the breeding of new plant varieties that are tailored to give improved performance under specific drought scenarios. However, several caveats need to be considered. Thus, most of the basic work related to plant drought responses involves relatively rapid imposition of water stress in short-term pot or hydroponic experiments under controlled environments, using model plants at specific stages of development. Application of such findings to the optimization of specific crop plant responses to drought episodes of variable timing, duration and intensity, as encountered together with other stress factors under field conditions, will be challenging. For example, actual field performance may be limited by unforeseen developmental penalties induced by interactions between drought and other environmental stresses, or by the availability and accuracy of rainfall and temperature predictions to help growers choose the most appropriate varieties for planting, Parallel efforts to encourage the introduction of supplementary crop irrigation, with water made available by improved conservation measures and by sea- or brackish-water desalination, will most likely provide comprehensive solutions to coping with crop production in drought-prone environments.