15.3.2 Crop water use and irrigation

Printer-friendly version

(a)  Evaporation and evapotranspiration


Figure 15.13 Peanuts under flood irrigation at Narrabri, New South Wales. Water is distributed to each row via a regulated 'hyroflume' to reduce seepage loss from an earthen supply channel. (Photograph courtesy P.E. Kriedemann)

Seasonal evaporation from plant communities almost always exceeds precipitation, so that supplementary water is needed to achieve full expression of a crop’s genetic capacity for yield (Figure 15.13). In applying such water to best effect, questions arise as to how crop water loss via evapotranspiration is best estimated, and whether there are soil or plant criteria that can be used in conjunction with meteorological data to schedule irrigation.

A combination of factors drives evaporation from a free water surface, namely solar radiation, wind speed, turbulence and humidity (generally expressed as atmospheric vapour pressure deficit, or VPD). Crop plants are subject to the same combination of driving variables. Their evapotranspiration (ET) includes soil evaporation + plant transpiration. For present purposes, ET can be regarded as synonymous with crop water use, and is commonly expressed as mm d–1.

Actual ET of a sparse crop in dry ground will fall short of potential ET due to soil and plant resistances to moisture loss (dry soil surface plus hydraulic and stomatal resistances of plants plus boundary-layer resistance of the crop community). By contrast, a dense and well-watered crop can lose water somewhat faster than estimated values for potential ET based on incident radiation plus the drying power of an air mass due to an abundance of evaporating leaf surfaces which exceed projected ground cover (high leaf area index (LAI); Section 6.4). In either event, irrigation is intended to forestall moisture stress during a production season by matching actual ET with supplementary water. Derivation of ET then becomes a critical issue, and can be inferred from (1) direct measurement of nearby evaporation, (2) application of a combination equation that provides an estimate of actual ET based on weather variables and crop attributes, or (3) application of micro-meteorological methods such as the Bowen ratio and eddy covariance (discussed later).

The simplest estimate of actual ET for a given crop (ETcrop) relies on direct measurement of evaporation from a class A pan (Epan), and application of an empirical pan coefficient (Kpc) appropriate to that crop, where:


Table 15.3

This empirical pan coefficient, Kpc, is derived by comparing measurements of Epan with either direct measurements of water loss from crops in weighing lysimeters, or changes in soil moisture content. Kpc generally ranges between 0.2 and 1.3, but does vary according to species, canopy development, soil conditions, topography and atmospheric moisture (Table 15.3). Moreover, as soil moisture is extracted to meet eva-potranspiration, stomatal responses to vapour pressure deficit are accentuated, so that Kpc is strongly influenced by plant-extractable soil moisture. As demonstrated by Eastham and Rose (1988) with pasture in an agroforestry experiment, Kpc decreased linearly from about 0.75 to 0.25 as soil moisture content in their lysimeters decreased from 0.4 to 0.1 (change in volumetric water content relative to water content of saturated soil).

Despite these limitations, local calibration and careful measurement of Epan produce useful approximations of ETcrop for irrigation management.

An alternative approach to estimation of ETcrop is to rely on physical principles underlying evaporative processes, rather than on empirical observations of class A pan evaporation. These physical principles are applied via formulae which combine effects of radiation and atmospheric conditions that drive evaporative demand with canopy attributes that restrict transpirational loss. Direct radiant energy from sunlight plus energy associated with turbulence and atmospheric humid-ity are involved. Penman (1948) published the first version of what became known as combination methods, formulae that provide an estimate of evaporation, using meteorological data that combine radiant energy with turbulent transfer components. A modern version of the Penman equation (Equation 15.7) combines a water vapour and sensible heat transfer equation with a physiologically derived canopy resistance term (Monteith 1965). Using well-established physical principles, the energy exchange associated with evapo-transpiration from a plant community ET) can be derived as follows:


where λ = (2.442 MJ kg–1);
ET = the quantity of water evaporated (mm);
s = the slope of the saturation water vapour pressure versus temperature curve (kPa K–1);
Rn = net radiation (W m–2);
cp = specific heat capacity of air at constant pressure (J kg K–1);
δe = ambient water vapour pressure (kPa);
rc = canopy resistance (s m–1);
a = total aerodynamic resistance (s m–1);
γ = psychrometric constant;
ρ = air density.

Under conditions of neutral atmospheric stability the aerodynamic resistance (ra) can be estimated by:


where z = height of measurement, d = displacement height, a measure of the difference between the actual height of the canopy (h) and a point within the canopy where wind speed (u) tends to zero, zo = is the roughness length of the canopy, k = von Karman’s constant (= 0.41).

zo and d can be estimated empirically for each site, and for forests, zo is approximately 0.076h and d = 0.78h. For row crops zo = 0.123h and d = 0.67h (Burman and Pochop 1994).

Calculating ET for a forest and for a crop, the following conditions will apply:

Constants: air density = 1.22 kg m–3, specific heat capac-ity of air = 1.012 kJ kg–1 K–1, psychrometric constant (γ) = 0.0662 kPa K–1 (not strictly a constant since the slope of the relationship between temperature and saturated water vapour pressure increases with temperature, but change in the value of g between 15ºC and 30ºC is small), latent heat of vaporisation (λ) = 2.442 MJ kg–1.

Meteorological variables: Rn (energy available to the canopy) = 15 MJ d–1, mean daily Tair = 15ºC, mean daily VPD = 0.98 kPa.


Using Equation 15.7, with h = 0.4 m, u = 1.5 m s–1, then ra = 50.2 s m–1, and a typical value for rc = 50 s m–1. Thus, daily ET:



Units are converted to MJ d–1 (86400s in a day) and converted to a depth equivalence of water evaporated by dividing by the latent leat of vaporisation, λ (2.442 MJ kg–1).


The same calculation can be performed for a forest com-munity, under similar meteorological conditions. Due to their height and rough canopy, the aerodynamic resistance of forests (ra) is approximately an order of magnitude lower than that for crops or grassland. A typical value for forest ra is around 5s m–1.

In this particular example, assume forest height is 30 m, measurement height (z) is 35 m, wind speed is 3 m s–1; then from Equation 15.8 and estimates of zo and d, ra = 5.2 s m–1.

Off-setting the low aerodynamic resistance of forests is the larger stomatal control that forests exert over water loss, with typical values of rc for forest communities being approxi-mately 150 s m–1 (Shuttleworth 1989). From Equation 15.7, for a forest community, with the same meteorological conditions as above, daily forest ET is approximately 4 mm d–1, that is, approximately 30% less than that of crops.

(b)  Potential, reference and actual evaporation

In publications on estimation of evaporation, the terms potential evaporation (or evapotranspiration) and reference evaporation are used. The term potential evaporation was originally used to convey the concept of evaporation that could occur from plant surfaces without any limitation of water availability or restriction in the vapour transfer pathway. However, because of canopy architecture, leaf characteristics and stomatal behaviour, different species evaporate water at different rates, even if all other factors are the same. To alleviate this ambiguity, the term reference or reference crop evapo-ration was introduced (Jensen et al. 1970). The idea was to specify and calibrate the evaporation equations against a crop with particular characteristics. Two crops, lucerne and grass, are generally used as reference crops. On any given day lucerne will have a higher evaporation loss than grass.

To estimate the actual rate of evaporation (Ea) from a crop or other vegetation type, the daily reference crop evaporation value (Er) is modified by an empirical crop coefficient (Kc):


Crop coefficients are derived by measuring Ea, usually with precision weighing lysimeters, and finding the ratio Ea/Er for each growth stage. They vary with crop species, stage of phasic development (mostly enhanced by leaf area) and soil water availability. Different values of Kc are needed if Er uses lucerne or grass as its reference crop. Typically values range from 0.2 (bare soil) to >1.0 (dense canopies).

Plant influences and the Penman–Monteith equation


Figure 15.14 Crop coefficient (Kc) is strongly dependent on canopy growth up to a leaf area index (LAI) of around 5 or 6, and approaches an asymptote at higher values. In this case, Kc is the ratio of daily evaporation by a wheat crop (Ea) that was measured directly with a precision weighing lysimeter and an estimate of evapotranspiration by a reference crop (Et) derived from meteorological data (i.e. Kc = Ea/Er) (Based on original data from W.S. Meyer)

Estimating Er (most commonly on a daily basis) takes only limited account of the effects of differences between species, state of canopy development (Figure 15.14) or degree of hydration (droughted versus non-droughted). These effects are adjusted through the Kc term so that actual crop evaporation can be estimated. This is sometimes referred to as the ‘two-step’ evaporation estimation. To describe more completely evaporation from vegetation in a one-step process Monteith (1965) modified the Penman method. The Penman–Monteith equation includes conductances to vapour transfer terms ascribed to both plants (conceptually, the effect of aggregated stomatal aperture) and the air layers immediately surrounding leaves in the vegetation canopy (commonly called the boundary layer). These two terms, the bulk stomatal or canopy conductance (gc) and the aero-dynamic conductance (ga), explicitly incorporate the vegetation interaction with evaporation. However, neither of these terms can be measured directly; gc is estimated as a function of individual leaf stomatal conductance and canopy leaf area index (gc = gs/LAI) while ga is estimated using formulae developed from fluid mixing theory.

The complexity involved in measurements of gc and ga has hindered the widespread use of the Penman–Monteith equation. As an interim measure until better methods for determining these variables are found, an FAO working group (Smith 1991) has proposed an approximate form of the equation which can be used as a replacement for the Penman equation. This equation is:


where Δ = slope of the saturation vapour pressure–temperature curve at mean daily temperature (kPa °C–1);

γ = psychrometric constant (kPa °C–1);

Rn = net radiant energy (MJ m–2 d–1);

G = ground heat flux (positive when direction of flux is into the ground) (MJ m–2 d–1);

U = Wind run (commonly measured with a cup anemometer) expressed as km d–1;

eo = mean daily saturation vapour pressure at mean dry bulb temperature (kPa);

ed = actual mean daily vapour pressure at mean dew point temperature (kPa);

Tm = mean leaf temperature (°C).

Standardisation and associated constants have been derived largely from measurements in humid regions. In semi-arid regions of southeastern Australia this formulation may produce reference crop evaporation estimates of up to 30% less than those currently calculated (Meyer 1995). Kc depends in part on stomatal conductance (gs) and, as shown earlier (Section 15.2), gs decreases in response to low humidity and thus restricts transpiration. Values for Kc will then vary ac-cording to vapour pressure deficit, hence the reduction in estimates of reference crop evaporation in southeastern Australia.

(c)  Direct measurement of water use by plants


Weighing lysimeters consist of a large block of soil (from 1 m3 to 400 m3 of soil) encased in a thin metal sleeve and base, sitting on a very large precision balance. The balance is located beneath the block of soil, sometimes several metres below the surrounding ground surface so that the top of the block of soil within the metal sleeve is aligned with the surrounding soil surface. Care is required to maintain an undisturbed soil structure within the sleeve and vegetation growing on the block of soil should be as representative of the surrounding vegetation as possible. This is easy to achieve with crops because seed can be sown in the lysimeter just as easily as in the surrounding soil. Measurements on native communities usually involve re-establishment of vegetation.

High-quality lysimeters can detect losses of approximately 0.01 mm of water, with time resolution of an hour or less. Because the vegetation in the lysimeter is surrounded by vegetation of the same composition and structure it is exposed to a realistic microclimate. Consequently estimates of water use using lysimeters give a better measure of rates of water use in the field than those determined using isolated plants.

Micrometeorological methods

Bowen ratio  Evapotranspiration involves exchange of both sensible heat (H) and latent heat (or latent energy, LE) (Section 14.5). Sensible in this context simply means ‘felt’ by evaporating surfaces and sensed with instruments. LE exchange amounts to 2.45 MJ per kilogram of water evaporated. The Bowen ratio (B) can then be summarised as:


Put another way, the Bowen ratio is simply the ratio of sen-sible heat exchange to latent heat exchange, and assuming the conductivity coefficients for sensible heat and water vapour are the same, then Equation 15.13 can be approximated by:


where ΔT = T1T2 and Δe = e1e2. Temperature (T) and water vapour pressure (e) are measured at heights 1 and 2, typically 1–2 m apart, above a canopy (1 m apart in Figure 15.15). a is a proportioning factor. If the other components of the energy balance are measured it is possible to calculate evaporation as:


Figure 15.15 Evapotranspiration from a lupin crop an Wongan Hills, Western Australia, is being measured in real time with a Bowen ratio apparatus which incorporates a heat plate at ground level (not shown), a pair of net radiometers and a pair of reversing psychrometers spaced at 1m, (lower unit is about 30 cm above the lupin crop). A wind vane (rectangular tail, upper right side of photograph) ensures detectors are pointing up wind. (Photograph by W. van Aken, and supplied by F.X. Dunin, CSIRO Division of Plant Industry, Rural Research Laboratories, Floreat Park, Western Australia)

Equipment requirements are relatively small (Figure 15.15) and include a net radiometer (to measure Rn), a soil heat flux plate (to measure G) and a pair of matched precision psych-ro-meters (to measure wet and dry bulb temperatures at two heights). Under non-advective conditions the Bowen ratio estimates can closely match those obtained from lysimeters.

Eddy correlation

Air turbulence arises as an air mass moves across a rough surface such as a crop or forest canopy, and the air mixing that results is fundamental to continued evaporation from that surface. If the rate of evaporation from a canopy is 10 mm per day, this represents 10 kg of water leaving every 1 m2 of that surface. In the absence of any turbulence or wind, the air close to the canopy would rapidly become saturated and evaporation would cease. Continued evaporation depends upon continuous removal of air close to the evaporating surface and replacement with drier air from well above the canopy.

Turbulence produces random fluctuations in wind speed and wind direction at any point above but close to the canopy. Eddy correlation (or eddy covariance) involves high-frequency measurements of atmospheric humidity and vertical wind speed above the transpiring canopy. Frequency is typically 10 times per second (10 Hertz), while location depends upon canopy architecture. Variables are sensed at about 0.5 m for low vegetation such as a mown lawn compared with 10 m above the canopy of a tall, aerodynamically rough surface such as a forest.

Eddies of wind that have a net downward velocity will have a lower water vapour pressure than eddies that have a net upward velocity because wind moving up from the canopy contains water lost as transpiration from the canopy. By integrating over a suitable time period (typically 20 min), and subtracting the downward flux of water from the upward flux of water, the net loss of water from the canopy can be calculated. This technique has been successfully applied to Amazon forests (Grace et al. 1995) and coniferous forests of Europe (Jarvis and McNaughton 1986).

(d)  Soils and plant-available water

Soil comprises silica and clay particles in a porous matrix. Depending on packing, and therefore the density of the matrix, there will be a variable void space. Soil with a bulk density of 1.3 (mass per unit volume) has about a 50% void volume. Water can move into this void volume, largely dis-placing any gas it contains. If water fills the void volume, the soil is then saturated.

Soil water held between field capacity and wilting point (Section 15.1) defined as being plant-available on physical grounds (Ψsoil is less negative than Ψroot) is not necessarily extractable by roots because of spatial separation between soil storage sites and absorbing surfaces on roots. Moreover, not all of the water in the plant-available range is equally available to plants. With irrigated crops growing in deep, well-fertilised, soft soils leaf extension begins to decline once 70–80% of nominally available water has been evaporated.

During soil drying, plants exhibit a number of responses. Probably the most important but least well studied is root growth. Drying in upper soil layers is accompanied by increasing growth of roots into deeper, wetter layers. Often this increased growth and the water gained is enough to overcome the first detectable signs of impending water deficit. However, if transpirational demand cannot be satisfied by root extension, shoot growth is soon affected.

Roots generally grow in voids which present the least resistance to extension growth. In structured soils, roots are clumped in the cracks and larger pores around soil peds. By and large, roots do not spread in a homogeneous fashion through soil (Section 3.1). Moreover, water within peds moves only slowly to outer surfaces to make contact with roots. Consequently, much of this water is not available to plants at a rate that will keep them from experiencing water deficit. Due to these physical and biological limitations, crop plants commonly experience drought stress even though sub-soil water content seems adequate for continued growth. In these cases, roots cannot grow either rapidly enough or permeate the soil densely enough to extract the apparently available water.

As evaporation continues and available water decreases, the question of when to irrigate arises. Provided the onset of drying is reasonably gradual and soil root-zone conditions are not overly restrictive, 70% of plant-available water can be used before extension growth is affected. Further drying (>80% of plant-available water consumed) will cause a decline in shoot extension and leaf gas exchange. To ensure full production, a general conservative recommendation is to irrigate once a soil had dried to about 50% of plant-available water.