Module 4: Water Use & Drought Tolerance

Wheat is grown on more hectares than any other crop, much of it in water-limited Mediterranean and semi-arid continental environments where rainfall rather than light or temperature sets the yield ceiling. This module treats the water economy of the wheat plant from the guard-cell level — the ABA signalling cascade that closes stomata in minutes — up through xylem hydraulics, root architecture, and whole-canopy energy balance, and closes with the Passioura (1977) transpiration-efficiency decomposition that remains the dominant framework for dryland breeding.

1. The Water Cost of Carbon

Every mole of CO&sub2; entering the mesophyll also carries with it the inevitable exit of ~200 moles of water vapour through the same stomatal aperture. At leaf-level, transpiration is an unavoidable side-effect of CO&sub2; diffusion through an aperture that must bridge internal water-saturated air (RH ~ 99%) and usually-dry atmospheric air (RH ≈ 40–60%). The leaf-to-air vapour pressure deficit (VPD) drives water out ~1.6× faster than CO&sub2; moves in, so the concentration gradients of the two gases are coupled but not identical.

\[\mathrm{WUE} \;=\; \frac{A}{E} \;=\; \frac{g_s (C_a - C_i)/1.6}{g_s \cdot D} \;=\; \frac{C_a - C_i}{1.6\,D}\]

VPD D (in kPa) scales nonlinearly with air temperature via the Clausius–Clapeyron relation: a heat wave raises D and punishes WUE faster than temperature alone would suggest. In the Indo-Gangetic plains the pre-monsoon heat spike that coincides with wheat grain filling can double VPD compared to the vegetative phase, making terminal drought and terminal heat biophysically almost inseparable.

The 200:1 Water Price

For a wheat canopy producing 6 t DM ha⁻¹ the season’s transpiration is roughly 300–400 mm, equivalent to 3000–4000 m³ ha⁻¹ of water vapour lost to the atmosphere. To appreciate the scale: the 700 Mt of wheat grown globally each year (FAO 2023) transpires ~700 km³ of water, more than the volume of Lake Erie. A 10% improvement in whole-plant transpiration efficiency (TE) would therefore save enough water to supply the annual domestic needs of several large countries.

2. Stomatal Regulation: the Guard Cell

Each wheat stoma is a pair of kidney-shaped guard cells whose coordinated turgor change opens or closes the aperture between them. Open stomata reflect a high guard-cell osmotic pressure Π ≈ 1 MPa, generated by inward loading of K⁺ ions (50–400 mM) plus counter-anions (Cl⁻, malate²⁻) plus sucrose. The turgor exerted against the epidermal counter-pressure mechanically opens the pore.

\[\Pi \;=\; i R T \sum_j c_j \quad\text{(van 't Hoff)}\]

\[P_{\text{turgor}} \;=\; \Pi_{\text{guard}} - \Pi_{\text{apoplast}}\]

Inward K⁺ flux is driven by a plasma-membrane H⁺-ATPase whose proton pumping creates a −180 mV membrane potential; inward K⁺-rectifier channels (KAT1-like in wheat) then conduct K⁺ down the electrochemical gradient. Blue-light activates the H⁺-ATPase via a phototropin + BLUS1 + PP1 cascade, giving the characteristic dawn-responsive stomatal opening.

Osmoticum Composition

Fully-open wheat guard cells contain ~400 mM K⁺, ~100 mM Cl⁻, ~50 mM malate²⁻, and 50–100 mM sucrose. Sucrose accumulates via photosynthesis and/or import and replaces K⁺-malate as the dominant osmoticum late in the day, a switch that fine-tunes aperture against the declining solar-driven transpiration demand. Under prolonged drought the K⁺/sucrose balance shifts toward sucrose, reflecting a metabolic-cost-minimising strategy (Lawson 2017).

3. ABA Signalling & the Schroeder Cascade

The stress hormone abscisic acid (ABA) couples soil-water status to guard-cell aperture in minutes. The core signalling module was dissected by the Schroeder laboratory through the 1990s and early 2000s and formalised in Schroeder (2001); its structural basis was solved in a flurry of 2009–2012 publications (Park 2009, Ma 2009, Cutler 2010).

\[\text{ABA} \xrightarrow{\text{PYR/PYL/RCAR}} \text{PP2C}\!\downarrow \xrightarrow{\text{SnRK2/OST1}\uparrow} \text{SLAC1}\uparrow, \text{KAT1}\downarrow\]

The sequence: ABA binds the soluble START-domain receptor PYR/PYL/RCAR, which then clamps onto clade-A PP2C phosphatases (ABI1, ABI2, HAB1), inhibiting their activity. Freed from phosphatase-mediated dephosphorylation, the SnRK2 kinase OST1 (open-stomata-1, also SnRK2.6) autophosphorylates at its activation loop and targets two substrates in the guard-cell membrane: the slow anion channel SLAC1 (phosphorylation at S120) and the inward K⁺ channel KAT1 (phospho-inhibition).

Activated SLAC1 conducts Cl⁻ and malate²⁻ outward, depolarising the membrane from −180 mV to −40 mV. The depolarisation opens the outward K⁺ rectifier GORK, and K⁺ floods out with its counter-anions. The loss of osmolyte collapses guard-cell turgor, and the aperture closes within 5–30 minutes of ABA perception.

Ca²⁺ and ROS Amplification

ABA also triggers NADPH-oxidase RBOHF production of apoplastic H&sub2;O&sub2;, which activates plasma-membrane Ca²⁺ channels; the resulting [Ca²⁺]ᵢ spikes reinforce SLAC1 activity via Ca²⁺-dependent CPK6/CPK23 kinases. The ROS–Ca²⁺ arm makes the ABA response digital: once threshold is crossed, closure commits even if ABA is withdrawn.

Recovery and Re-opening

When soil water is restored, ABA is catabolised by CYP707A monooxygenases to 8′-OH-ABA and phaseic acid, PP2Cs dephosphorylate OST1, SLAC1 activity falls, and the stoma gradually reopens over ~30–90 minutes. The asymmetric kinetics (fast close, slow open) biases the plant toward conservative water use.

Schroeder ABA cascade schematic

ABA -> PYR/PYL/RCAR -> PP2C off -> OST1 on -> SLAC1 -> K+ efflux -> closureABA(root-sourced)PYR/PYL/RCARreceptorPP2C (ABI1/2)inhibitedOST1 (SnRK2.6)active kinaseSLAC1 anionCl-, malate outDepolarisation-180 to -40 mVGORK K+ outoutward rectifierTurgor lossclosure < 30 minSchroeder 2001; Park 2009; Ma 2009; Cutler 2010The cascade is evolutionarily conserved across seed plants; wheat orthologs are known

Simulation 1: Guard-Cell ABA → Closure ODE

Couples the Schroeder ABA→PYR→PP2C→OST1→SLAC1→GORK cascade to a first-order osmolyte decay, converts osmolyte to turgor via van’t Hoff, and shows closure kinetics over several ABA doses. A Hill–Langmuir receptor occupancy feeds the effective closure rate constant; trapezoidal integration quantifies the water saved versus a non-ABA control.

Python
script.py143 lines

Click Run to execute the Python code

Code will be executed with Python 3 on the server

4. Osmotic Adjustment: Compatible Solutes

When soil water potential falls below −0.5 MPa, stomatal closure alone cannot sustain cell function; cells must actively lower their internal water potential by accumulating osmotically-active solutes. This is called osmotic adjustment (OA). The accumulated solutes must be “compatible” — osmotically effective but metabolically inert with respect to enzyme kinetics, membrane structure, and protein folding. The three canonical compatible solutes in wheat are:

  • Proline: synthesised from glutamate via P5CS / P5CR; accumulates 10–100-fold under drought; stabilises protein hydration shell and scavenges radicals.
  • Glycine betaine: methylated derivative of glycine; stabilises PSII D1 and membrane lipid phase state; wheat produces it in moderate amounts (2–5 µmol g⁻¹ FW).
  • Sugars (sucrose, fructans): wheat stems store large fructan reserves that mobilise during grain filling; sucrose protects membranes during freezing-induced dehydration.

\[\Psi_w \;=\; \Psi_p + \Psi_\pi + \Psi_m\]

\[\text{OA} \;=\; \Pi_{\text{full turgor, stressed}} - \Pi_{\text{full turgor, control}}\]

A wheat cultivar that accumulates 0.4 MPa of osmotic adjustment can sustain turgor (and therefore photosynthesis) at leaf water potentials as low as −2.0 MPa, where an unadjusted genotype would already have reached turgor loss point. The Morgan 1983 screening paradigm — measuring Π at zero turgor after dehydration and rehydration cycles — remains the standard OA phenotyping assay.

Turgor Loss Point (TLP)

The turgor loss point is the leaf water potential at which turgor P = 0, i.e., Ψ₩€ = Ψₙ. Pressure–volume (P–V) curves give TLP directly. Wheat TLP ranges from −1.5 MPa in drought-sensitive genotypes to −2.8 MPa in well-adjusted durum wheat. TLP is a strong cross-species proxy for drought tolerance (Bartlett 2012, Ecology Letters).

5. Root Hydraulics & Xylem Transport

Water flows from bulk soil through the rhizosphere, across the root cortex, through endodermal symplast or Casparian-belt apoplast, into the xylem, up the stem, and out the leaf. In steady state the flow is driven by a tension gradient set by transpiration and resisted by a series of hydraulic conductances kᵢ. Hagen–Poiseuille gives the conductance of a cylindrical xylem vessel:

\[K_{\text{HP}} \;=\; \frac{\pi r^4}{8 \eta}\]

\[J_v \;=\; -K \frac{d\Psi}{dx}\]

The r⁴ dependence means a few wide vessels conduct more water than many narrow ones — but wide vessels are also more susceptible to drought-induced cavitation. Wheat xylem vessel diameters are 30–80 µm, narrow by comparison with diffuse-porous angiosperm trees (100–300 µm) but wide enough to meet high transpiration demand in a fast-growing canopy. The leaf-to-root water potential drop in actively transpiring wheat is typically 0.5–1.5 MPa.

Aquaporin Regulation

Water crosses membranes mostly through plasma-membrane intrinsic proteins (PIP1, PIP2 families), a subclass of major intrinsic proteins (aquaporins). PIP gating is post-translationally regulated by cytosolic pH, Ca²⁺, phosphorylation, and heteromerisation; entire segments of root hydraulic conductance can switch on or off in minutes. Wheat TaPIP2;2 and TaPIP1;1 show diurnal expression oscillation matching transpiration demand, and drought induces rapid closure of PIP channels in the root cortex (Cochard 2007, Maurel 2008).

Cavitation & Embolism

Under sufficiently negative xylem pressure (−2 to −4 MPa in wheat), air is sucked through bordered-pit membranes from an already embolised neighbouring vessel, nucleating a gas embolus that fills the vessel and disables it. The hydraulic vulnerability curve Pₜ₀ (xylem pressure causing 50% loss of conductance) is a species-level trait: wheat Pₜ₀ ≈ −3.5 MPa, intermediate between drought-sensitive mesophytes and desert perennials. Embolus refilling in monocots is aided by root pressure at night; catastrophic cavitation at the whole-plant level coincides with terminal drought death.

6. Avoidance, Tolerance, Escape

Levitt (1972) classified drought-response strategies into three mutually-exclusive categories, a framework still central to breeding:

  • Drought escape — short life cycle, early flowering, so that grain filling completes before the terminal drought. Early-maturing wheat ideotypes, Mexican CIMMYT photoperiod-insensitive cultivars.
  • Drought avoidance — maintain high tissue water status despite soil drying, via deep roots, thick cuticle, reduced leaf area, and reflective waxes. Australian dryland wheats classically fit this mode.
  • Drought tolerance — operate biochemistry at low tissue water potential via osmotic adjustment, HSPs, compatible solutes, and anti-oxidant enzymes. Desert crops (tef, pearl millet) and some durum landraces.

Trade-offs are common: drought escape costs yield potential in wet years; avoidance via early stomatal closure caps photosynthesis; tolerance via solute accumulation costs carbon. Breeders target the dominant stress envelope of the target environment and accept the compromise.

Isohydric vs Anisohydric

Tardieu & Simonneau (1998) distinguished two stomatal-behaviour types.Isohydric genotypes maintain a constant midday leaf water potential by aggressively closing stomata as VPD or soil water deficit rises — protective but conservative. Anisohydric genotypes let Ψₙ track the environment, keeping stomata open longer and transpiring more, at the risk of cavitation. Wheat cultivars span the spectrum; durum wheats lean anisohydric (better under deep soil water access), bread wheats more isohydric.

Stay-Green Phenotype

Delayed senescence — the “stay-green” trait — preserves flag-leaf photosynthetic capacity during grain filling and extends the effective grain-filling duration by days to weeks. In wheat, stay-green QTLs on chromosomes 2B, 2D, and 7A have been cloned (Christopher 2014). Stay-green benefits yield in moderate drought but may compromise water use efficiency by keeping canopy transpiration running when roots are depleted; its ideal expression depends on soil water holding capacity.

7. Passioura’s Transpiration-Efficiency Framework

Passioura (1977) proposed the simplest possible multiplicative decomposition of dryland yield:

\[Y \;=\; T \times \mathrm{TE} \times \mathrm{HI}\]

Here T is seasonal transpiration (mm water), TE is transpiration efficiency (kg DM per mm water, ≈ 55 kg DM ha⁻¹ mm⁻¹ for wheat), and HI is harvest index. Passioura’s insight was that under a fixed rainfall envelope, agronomic management should maximise T (minimise evaporation E and deep drainage D that reduce available water) while breeding should push TE upward via genotypic differences in WUE-related traits.

\[R \;=\; T + E + D \quad\text{(season water balance, no irrigation)}\]

French & Schultz (1984) empirically fitted an envelope from Australian rainfed wheat data: Y ≈ 0.020 (R − 110) t ha⁻¹, i.e. ~20 kg grain per mm rainfall above a soil-evaporation threshold of ~110 mm. Global median WUE for rainfed wheat sits at 2–3 kg grain mm⁻¹ rainfall. The best-irrigated high-technology wheat systems in the Yaqui Valley (Mexico) or Toowoomba (Australia) reach 3.5–4 kg mm⁻¹.

Carbon Isotope Discrimination

Rubisco discriminates against ¹³CO&sub2;, so grain with more negative δ¹³C integrates a history of relatively high Cᵢ/Cₐ, i.e., high stomatal conductance and low WUE. Condon (2004) showed that carbon-isotope discrimination Δ¹³C selection in wheat shifts WUE in the expected direction and has been used operationally in Australian breeding since the 1990s. Δ¹³C in grain is a surrogate for season-integrated WUE.

Monteith’s RUE under Drought

Monteith’s RUE decomposition (Module 3) interfaces with Passioura’s via: RUE→TE through stomatal limitation of A. Drought reduces RUE roughly linearly from ~1.35 g MJ⁻¹ well-watered to ~0.7 g MJ⁻¹ under moderate drought, mostly through stomatal closure and only secondarily through biochemical decline. The Monteith–Passioura crossover identifies the water-limited-yield regime where the relevant capacity constraint is water rather than light.

8. Penman–Monteith Reference ET

The Penman–Monteith equation (Monteith 1965 from Penman 1948) combines an energy balance term and an aerodynamic term to give reference evapotranspiration ET₀, the water loss of a hypothetical short-grass canopy under given weather:

\[\lambda E \;=\; \frac{\Delta (R_n - G) + \rho c_p \, \mathrm{VPD}/r_a}{\Delta + \gamma\,(1 + r_s/r_a)}\]

Here Δ is the slope of the saturation vapour pressure–temperature curve, Rₙ is net radiation, G is soil heat flux, ρ and cₚ are air density and heat capacity, VPD is vapour pressure deficit, rₐ is aerodynamic resistance, and rₛ is surface (canopy + stomatal) resistance. γ is the psychrometric constant. The FAO-56 implementation (Allen 1998) packages Penman–Monteith with standard reference conditions (rₛ = 70 s m⁻¹) and a crop coefficient Kₐ to scale ET₀ to actual crop ETₐ.

Wheat Crop Coefficient

For wheat Kₐ rises from ~0.3 at emergence, to 1.10–1.20 at peak canopy (booting through mid-grain filling), falling to ~0.4 at maturity. Full-season wheat ETₐ in continental climates is typically 450–550 mm; under Mediterranean winter-spring wheat ~350–450 mm; under Indian irrigated wheat ~400 mm.

\[\mathrm{ET}_c \;=\; K_c \cdot \mathrm{ET}_0\]

Operational irrigation scheduling uses Penman–Monteith daily ET₀ × Kₐ as the gross irrigation requirement, adjusted by soil-water-balance accounting. In remote-sensing-based phenotyping, ETₐ is estimated from thermal imagery and used to compute CWSI (crop water stress index) at field scale.

9. Canopy Temperature Depression

An actively transpiring wheat canopy is cooler than the ambient air by 3–6 °C during midday — the canopy temperature depression (CTD). CTD reflects the evaporative cooling of transpiration minus convective heating by the surrounding air. Araus et al. (2008) argued that CTD is one of the few proxies for whole-plant transpiration (and hence for root water uptake capacity) measurable at the plot scale.

\[\mathrm{CTD} \;=\; T_{\text{canopy}} - T_{\text{air}}\]

Under well-watered conditions cultivars with higher CTD (cooler canopies) consistently yield better — a proxy for vigorous transpiration and active photosynthesis. Under severe drought the relationship can invert: cultivars that remain cool via profligate water use may exhaust soil moisture before grain filling, yielding poorly. Reynolds and colleagues at CIMMYT have used CTD via infrared thermometry as a routine selection tool since the 1990s, and airborne / UAV thermal imaging has scaled it to breeding-program throughput.

Crop Water Stress Index (CWSI)

Idso (1981) normalised CTD against a water-sufficient baseline and a non-transpiring reference to give CWSI = (T⁰ − Tₗₑ&#209b;)/(Tࡁ − Tₗₑ&#209b;), ranging 0 (well-watered) to 1 (fully stressed). CWSI bands from thermal imagery underpin modern precision irrigation.

10. Root Architecture Ideotypes

Tuberosa (2012) argued that root architecture is the final frontier of crop improvement: invisible below ground, heritable but historically neglected by breeders, and now accessible via phenotyping platforms like minirhizotrons, X-ray CT, and shovelomics. Wheat root systems are fibrous monocot systems with seminal roots (emerging from the embryo at germination) and nodal roots (emerging from tillers later in development).

Deep vs Shallow Ideotypes

Under Mediterranean-climate drought, crops sow into stored soil moisture and rely on deep roots to track receding water. A deep root ideotype — narrow root angle, vigorous deep penetration — out-yields broad-rooted controls under terminal drought. Wasson (2012) and others have developed breeding populations selected for depth at depth >1 m. Conversely, shallow soils (vertisols with hardpans, Indo-Gangetic plains) reward wide-spreading, lateral-rooted ideotypes that scavenge the topsoil.

DREB and TaDRO1 Genes

In rice, the DRO1 (deeper rooting 1) gene controls root gravitropism and its introduction into shallow-rooted genotypes lowered root distribution by 30 cm (Uga 2013). Wheat orthologs TaDRO1-A, TaDRO1-B have been characterised and are targets for deep-root ideotype breeding (Richard 2018). DREB transcription factors reprogram root morphology and drought-responsive gene networks; TaDREB2 overexpression improves tolerance but at a yield cost in unstressed environments.

Root Cortical Aerenchyma

Root cortical aerenchyma — gas-filled cavities replacing cortex parenchyma — reduces the metabolic cost of maintaining deep roots (Lynch 2013). Wheat produces aerenchyma less readily than maize but genotypic variation exists and is under selection in Penn State’s “steep, cheap and deep” breeding programs.

Simulation 2: Passioura Yield × Rainfall Gradient

Implements the Passioura Y = T×TE×HI framework across a 150–750 mm rainfall gradient, overlays the French–Schultz empirical envelope, decomposes rainfall into transpiration / soil evaporation / deep drainage components, computes a seasonal Penman–Monteith reference ET time-course with a wheat Kₐ profile, and illustrates canopy-temperature depression under well-watered vs drought regimes.

Python
script.py159 lines

Click Run to execute the Python code

Code will be executed with Python 3 on the server

11. Breeding for Water-Limited Environments

CIMMYT’s Wheat Physiological Breeding program (Reynolds, Pask, 1990s–present) assembles high-throughput phenotyping platforms — infrared canopy thermometry, NDVI, chlorophyll content (SPAD), root mass, grain δ¹³C, green-area duration — into an index selected under managed drought and heat environments in Cd. Obregón, Sonora and Tlaltizapán. The resulting CIMMYT elite bread wheat lines dominate the spring wheat mega-environment globally.

Dryland vs Irrigated Breeding

Breeding populations targeted at dryland environments share little germplasm with irrigated-system populations: the relevant trait profiles diverge. Dryland wheats emphasise deep roots, narrow leaf canopy (reducing vegetative water use), osmotic adjustment, tight stomatal control, and short grain-filling duration. Irrigated wheats emphasise maximum LAI, high-RUE photosynthesis, long grain-filling duration, lodging resistance (semi-dwarf Rht alleles), and disease resistance stacks. The Green Revolution shift of the 1960s was largely an irrigated-wheat revolution; dryland breeding progressed more slowly and remains a major gap.

HB4 & Engineered Drought Tolerance

The HB4 transgene from sunflower (Helianthus annuus HAHB-4 homeodomain-leucine zipper) was transferred into wheat by Chan and colleagues (Argentina) and commercialised as HB4 wheat (Trigo Bioceres). In multi-environment trials HB4 wheat yielded ~20% more than non-transgenic controls under moderate drought. HB4 modulates ABA signalling and delays senescence; regulatory approval was granted in Argentina, Australia, Brazil, Colombia, New Zealand, and the United States between 2020 and 2023, making it the first commercial GM drought-tolerant wheat.

12. Synthesis

Wheat drought physiology is best understood as a set of nested control loops operating across timescales:

  • Seconds–minutes: guard-cell membrane biophysics; SLAC1 opening after OST1 activation; K⁺ efflux through GORK.
  • Minutes–hours: ABA perception, osmolyte loss, aquaporin gating; whole-leaf stomatal closure.
  • Hours–days: osmotic adjustment; solute accumulation; root growth downward.
  • Days–weeks: canopy development, root architecture, phenological escape.
  • Seasons–decades: Passioura decomposition, breeding for TE, CIMMYT physiological-breeding pipelines, HB4-style biotechnology.

The biophysical treatment connects these scales: van’t Hoff osmotic pressure in the guard cell, Poiseuille flow in xylem, Penman–Monteith at canopy, and Passioura at the field. The module that follows (M5) picks up the other dominant limitation — nitrogen — with which water interacts: deep roots access deep soil nitrogen, and drought affects the mineralisation rates of soil organic matter.

Key References

• Passioura, J. B. (1977). “Grain yield, harvest index and water use of wheat.” J. Aust. Inst. Agric. Sci., 43, 117–120.

• Schroeder, J. I., Allen, G. J., Hugouvieux, V., Kwak, J. M. & Waner, D. (2001). “Guard cell signal transduction.” Annu. Rev. Plant Physiol., 52, 627–658.

• Park, S. Y. et al. (2009). “Abscisic acid inhibits type 2C protein phosphatases via the PYR/PYL family of START proteins.” Science, 324, 1068–1071.

• Ma, Y. et al. (2009). “Regulators of PP2C phosphatase activity function as abscisic acid sensors.” Science, 324, 1064–1068.

• Cutler, S. R., Rodriguez, P. L., Finkelstein, R. R. & Abrams, S. R. (2010). “Abscisic acid: emergence of a core signaling network.” Annu. Rev. Plant Biol., 61, 651–679.

• French, R. J. & Schultz, J. E. (1984). “Water use efficiency of wheat in a Mediterranean-type environment. I.” Aust. J. Agric. Res., 35, 743–764.

• Monteith, J. L. (1977). “Climate and the efficiency of crop production in Britain.” Phil. Trans. R. Soc. B, 281, 277–294.

• Araus, J. L., Slafer, G. A., Royo, C. & Serret, M. D. (2008). “Breeding for yield potential and stress adaptation in cereals.” Crit. Rev. Plant Sci., 27, 377–412.

• Condon, A. G., Richards, R. A., Rebetzke, G. J. & Farquhar, G. D. (2004). “Breeding for high water-use efficiency.” J. Exp. Bot., 55, 2447–2460.

• Tuberosa, R. (2012). “Phenotyping for drought tolerance of crops in the genomics era.” Front. Physiol., 3, 347.

• Tardieu, F. & Simonneau, T. (1998). “Variability among species of stomatal control under fluctuating soil water status and evaporative demand.” J. Exp. Bot., 49, 419–432.

• Reynolds, M. P. et al. (2015). “Achieving yield gains in wheat.” Plant Cell Environ., 38, 1821–1842.

• Bartlett, M. K., Scoffoni, C. & Sack, L. (2012). “The determinants of leaf turgor loss point.” Ecol. Lett., 15, 393–405.

• Maurel, C., Verdoucq, L., Luu, D. T. & Santoni, V. (2008). “Plant aquaporins.” Annu. Rev. Plant Biol., 59, 595–624.

• Levitt, J. (1972). Responses of Plants to Environmental Stresses. Academic Press.

• Allen, R. G. et al. (1998). “Crop evapotranspiration: guidelines for computing crop water requirements.” FAO Irrig. Drainage Paper 56.