Tải bản đầy đủ - 0 (trang)
VI. Water Uptake by Roots

VI. Water Uptake by Roots

Tải bản đầy đủ - 0trang

THE INFLUENCE OF SOIL STRUCTURE



145



and radial resistances for intact and field-grown roots. Axial resistance R, is

the fractional resistance to flow in the xylem. It is often calculated from the

Poiseuille-Hagen equation:

Q



= rtr4/8q AL = Al///R,



(13)



where Q is the flux (in cubic meters per second), r is the radius of the xylem in

meters, q is the viscosity of xylem sap (in meter. seconds), AL is the length of

the root (in meters), and A+ is the decrease in hydraulic potential in the

direction of flow. Axial resistance is therefore likely to be greatest in narrow

xylem vessels, such as those of graminaceous roots, and least in widediameter xylem vessels, such as those of tap-rooted species, where secondary

xylem tissue greatly increases total xylem volume. Axial resistance can also

increase with the age of the root and the depth of the roots in the soil

(Ponsana, 1975). The increase in secondary tissue xylem of tap-rooted

(dicotyledoneous) species may partially offset this, however. Taylor and

Klepper (1975) found cotton had negligible xylem resistance, whereas Willatt

and Taylor (1978) found substantial axial resistance in soybeans as they

matured. Passioura (1983) has pointed out that either adverse or beneficial

effects may accrue where large axial resistance occurs in the seminal roots of

wheat, depending on the water regime in the soil and the climatic environment. Cornish (1981) has demonstrated that deep sowing of cereals can

produce a similar increase in axial resistance as occurs with the seminal roots

through lengthening of the coleoptile internode.

Table V summarizes some published data on the range and maximum rates

of specific root water uptake U , for different species (in cubic meters per

meter per day). The data are either taken from field experiments or from

rhizotron environments where the areal portions of the plant were exposed to

the ambient atmosphere.

Maximum rates of uptake of 10-l’ m’/s are common for cereals, but grain

legumes seem to have higher uptake rates when grown in comparable or the

same environments. When we recall that soil-water diffusivities are seldom

less than

m2/s even when the soil is at low matric potential, a perirhizal

resistance will occur only if a continuous water film does not exist at the

root-soil interface or if only a fraction of the root is taking up water in that

soil volume. The increase in total resistance commonly reported as a plant

grows may reflect, in part, the increasing proportion of the root system which

is suberized and nonabsorbing (despite the fact that La also increases with

plant age). Such compensatory (or sometimes synergistic) effects of wholeplant growth may explain some of the conflicting evidence in the literature on

soil-plant resistances. Similarly, plant resistances which are calculated from

averaged or random measurements of leaf water potential taken during the

mid-day period do not represent the full non-steady-state conditions of the



146



ANN P. HAMBLIN



Table V

Specific Water Uptake Rates for Various Species in Temperate and Subtropical

Environments’

Species



Reference



Temperate grasses

Oats

Winter wheat

Spring wheat



Welbank et a/. (1974)

Ehlers eta/. (1981a)

Gregory et al. (1978a)

Greacen and Hignett (1976)

H a m b h and Perry (unpubl.)

Allmaras et al. (1975)

Taylor and Klepper (1973)

Willatt and Taylor (1978)

Willatt and Olsson (1982)

Allmaras et a/. (1975)

Hamblin and Perry (unpubl.)

Hamblin and Perry (unpubl.)



Maize

Soybean



Field pea

Lupin

Cotton



BarYosefandLambert(1981)

Taylor and Klepper (1975)



Water uptake rates



(I, are



Water uptake rate

t x 10-l’ (average)



8 x 10-12-1x lo-”

2 x 10-13-2x 10-12

1 x 10-”-2 x lo-’’

4 x 10-l3-lx lo-’’

3 x 10-12-2.4

x lo-’’

1 x 10-”-2.4 x lo-”

2.8 x 10-”-2.2 x lo-’’

2 x 10-”-9 x lo-’’

2 x 10-”-3 x lo-’’

1 x 10-”-2 x 10-l’

3 x 10-“-2 x

5 x 10-12-5x lo-”

<10-14-8 x lo-’’



measured in m3/m s, or m’/s.



plant over longer time intervals, such as a week’s drying cycle. Jones et al.

(1981) found plant resistances calculated on a daily basis were three to five

times as great as those calculated from hourly readings which assumed quasisteady-state conditions.

In a homogeneous, wet soil into which young plant roots are rapidly

elongating, the major resistance to flow is in the plant (Reicosky and Ritchie,

1976); much of it is radial resistance. The flux of water is greatest in the

central cell layers, where the cylinder volume is smallest, hence resistance

could increase progressively inward if potential gradients remained constant

across the root. As the plant grows into undisturbed subsoil, the spatial

pattern of water uptake becomes more variable and some parts of the root

system may be at very different soil-water potentials, even at the same depth

in the soil, as we recall from Fig. 12b. In many environments increased plant

growth is accompanied by higher levels of transpiration than can be supplied

without substantial soil drying. Soil shrinkage may cause perirhizal resistance

as canopy leaf area approaches its maximum or in the ripening stages. Tinker

(1976) has best outlined the possible development of soil-root interface

resistance in relation to soil pore expansion, water vapor transport, mucigel

production, and root hair development. Faiz and Weatherley (1978) produced evidence for the development of perirhizal resistance (Rpe) but in a

controlled environment. The calculated drop in R,, was two or three times

that of the pararhizal (R,J resistance, but the rate of onset of stress was more



THE INFLUENCE OF SOIL STRUCTURE



147



rapid than could occur in the field. In the field the number of parameters

which can be adequately measured in even the water balance equation, let

alone in a root uptake equation, must leave us in doubt as to the absolute

validity of many field-obtained resistance values. van Bavel et al. (1968a,b)

discussed the problem of variability of field measurements of soil hydraulic

properties and root water uptake with reality. They considered that flux

calculations by “instantaneous” methods break down when a vigorous daily

demand is superimposed on a slowly adjusting hydraulic system, and that

accurate values may be impossible to obtain in these circumstances.

At this juncture we should reexamine the proposition that the flux of water

into the root sink is proportional to the root length density L,. Field data do

not always bear this out, and indeed there appear to be two commonly

observed situations which are somewhat contradictory. In some situations,

greatest uptake rates occur deep in the profile, with a relatively small

proportion of the total root length responsible for much of the uptake. In

other instances, more water is extracted from shallow layers with high root

densities and deeper soil water is not always completely extracted, even when

there is a high potential gradient. Specific root uptake rates (U,) and

integrated root-water uptake (r(z):m3/m3 s) increase sequentially down the

profile with time in the first case, reaching maximum values during or just

after periods of greatest leaf area and evaporative demand. Gregory et al.

(1978b), for example, found greatest r(z) values at depths greater than 1 m

with 3 % of the total root length of winter wheat being responsible for 20 % of

the water transpired. Ehlers et al. (1981a) found U , to be always highest near

the rooting front in the region of young oat roots, with maximum r(z) values

at 0.6 to 0.8 m in a total 1.0-m root zone. Similarly, Allmaras et al. (1975)

recorded the highest v(z) values at the lowest L, values for both soybean and

corn (Zea mays), despite a difference of 0.4m in the total rooting depth

between the two species and very much higher U , values for soybean than for

corn. Rickman et al. (1978) reported a very similar extraction pattern for

winter wheat in which the maximum r(z) values occurred after peak leaf area

index (LAI), at 1.2 to 1.4m in the profile. A common feature of these

examples is that the soils were generally loams or silts, with coarser-textured

subsoils which had high K(8) values. Some had perched water tables deep in

the subsoil. In these circumstances U , is more closely associated with the soil

water status than with L,.

In the contrasting situation, roots known to be present in the subsoil fail to

extract all the available water and U , is greatest in the regions of high L,

(Jordan and Miller, 1980; Willatt and Olsson, 1982; Hurd, 1974; Walter and

Barley, 1974). The same range of species is involved, sometimes even the same

cultivars. The authors themselves frequently consider L, to be too low to

extract the water in the context of the evaporative demand and the hydraulic



148



ANN P. HAMBLIN

Table VI



Maximum Depth of Spring Wheat Root Penetration (zmax)and Depth of Seasonal Wetting Front

(WF,) for Four Soil Types at Tammin, Western Australia"

1969

Soil type



1. Uniform loamy sand >4-m deep

2. Sandy loam, 14% clay

increasing to 23 % at 1.5 m

3. Grey cracking clay: calcrete at

30-45 cm

4. Sand over grey clay (10-25 cm

sand)

Growing season rainfall (mm)

a



Zmax



WFs



1.40

1.58b



1.40

0.90



1.69

1.73b



>2.0

1.40



1.65

1.68b



~2.0

1.30



0.26



0.25



0.31



0.30



0.28



0.30



0.61



0.60



0.73



0.75



0.70



0.70



126



223



165



From Tennant (1976).



* Permanent groundwater



available below that year's wetting front.



conductivities of the soils. These studies have in common a fine-textured soil,

often with marked ped formation in the subsoil. These are situations where

nonuniform root distributions are, I suggest, responsible for the low values of

~ ( z reported.

)

Nonuniform distribution may also be responsible for high

(calculated, or assumed) perirhizal and axial resistances, as Ehlers et al.

(1981b) noted. Studies in fine-textured soils have often been carried out in

drought-affected environments where the soil profile was dry prior to the

growing season. The root elongation rate then depends on the rate of wetting

front advance and may be restricted by high soil strength both downward

and laterally into massive peds or blocks. Table VI shows the extreme

dependence of maximum root depth on soil type (and soil wetness) in a

Mediterranean environment.

McCoy et al. (1984) modeled the effect of varying L, and root diameter in a

water uptake simulation which could further explain the differences found in

the two groups of results discussed above. They used soil data sets from a

Pachappa sand, an Indio silt loam, and a Chino clay, which incidentally

showed wide variation in the functional form of the D(0) curves. The

simulation predicted that the lower value of D at the time of maximum flux,

the greater was the difference between Rperand Rpar.Low root densities of

less than 1 x lo4 m-' (corresponding to pararhizal cylindrical distances of

less than 20 mm) were used. A reduction in pararhizal distance, caused by

increasing L,, creates a reduced flux at the root surface because demand

generates a flux proportional to the total soil volume. Hence less negative

Rpervalues occur as L, increases, with most of the reduced matric potential in



THE INFLUENCE OF SOIL STRUCTURE



149



the first millimeter of soil adjacent to the root. Interestingly the simulation

also suggested that high L,s damp down fluctuations in Rperand that matric

potential-dependent changes in D(0) play a large role in water flux only at

very low values of D or for very sparse root systems. This value of D occurred

at much lower matric potentials in the sand than in the silt loam or the clay.

m2/s at

The values of D in the sand were very large (e.g., D = 2 x

I/Js = - 1.0 MPa), whereas those of the silt, loam, and clay were closer to

other published values (e.g., D = 5 x lo-’ m-’/s at I/Js = - 1.0 MPa).

Eavis and Taylor (1979) approached the same topic experimentally,

specifically to answer the question, “What is the effect of increasing the plant

root system on uptake rates and transpiration during a drying period,” using

Eq. (1 1) as their model. Their experimental method preserved environmental

fidelity while allowing variation in root length to leaf area ratios. Their results

appear to show that, when all roots are surrounded by soil of the same water

content, the total root length is not important in controlling transpiration

rate. The mean soil water content and leaf area were the two most significant

( P > 0.001) parameters tested in relation to transpiration rate. Increasing the

root to shoot ratio did not increase transpiration rate or leaf water potential.

Several interpretations of the data are offered by the authors, and perhaps

their impartiality in ascribing any particular part of the total resistance

responsible for differences in uptake may be viewed as yet another failure to

resolve a thorny problem. Instead, I take a more pragmatic view. We know

that relatively simple soil physics models may often provide satisfactory

working predictions of water use and crop yield, provided the crop and

climatic environment are precisely mimicked in the model. We find that

specific soil and climatic environments affect root and shoot responses in

ways which are predictable, when we have sufficently detailed and accurate

information about those environments. I think that models built on the

concept of an “ideal” plant or soil are inappropriate, if not wrong, in most

real field situations. We need to work toward models which are consistent

among plant, soil, and climatic parameters and are biologically corrrect for

the crop type ( C , or C,, monocot or dicot, annual or perennial) being

considered.



VII. SPECULATION: ARE WE MEASURING AND

AVERAGING AT CONSISTENT SCALES?

Crop production is concerned, above all else, with weights per unit area.

The “area” of commercial farming operations is seldom smaller than 0.1 ha

or larger than 100 ha. Within these areal limits farmers, agronomists, and

plant breeders have strived toward uniformity of crop response, selecting



150



A N N P. HAMBLIN



germplasm which has a very low variance for each phasic stage, and mixing

and manipulating the soil surface to remove microtopographic irregularities

which would interfere with this uniform plant development. Concern for the

influence of soil variability on crop yield prompted literally thousands of

uniformity trials by breeders and agronomists some decades ago which were

the basis to well-established criteria for trial design and size (Cochran and

Cox, 1957). Uniformity trials established that spatial heterogeneity of crop

response was not distributed randomly: contiguous plots were more likely to

have similar yields than those further apart. Now, although plot size is

constrained on the one hand by the limitations imposed by row-spacing

requirements and edge effects and on the other by field size and the number of

treatments, within a range of, say, 1 to 240 m2,CVs for yields from uniformity

trials over a wide range of crops and soil types are seldom greater than 15 %

and usually around 10 % (Frey and Baten, 1953; Hallauer, 1964; Elliott et al.,

1952). This is remarkably low when we recall the very large CVs mentioned

for soil hydraulic properties in Section IV,B,l. However, the crop is as much a

product of its aerial environment as its soil environment, and relatively little

spatial heterogeneity occurs in the atmosphere at a field scale of 100 ha by

1 m in height, with a uniform crop canopy. In addition, it may be that there is

a periodicity in the autocorrelation interval of soil properties. In other words,

samples which are correlated at small spatial scales may be uncorrelated at

intermediary scales but are again correlated at larger scales. Finally, it must

be remembered that the crop is integrating the environmental effects through

time. Temporal variations observed at hourly or daily scales may be

smoothed out over a season just as averaging procedures themselves may

obliterate critical extreme values which set the limits to growth processes.

Gardner and Gardner (1983) drew attention to this type of variation of

response from one scale to another when they compared the near perfect

linearity which exists between evapotranspiration (water use) and dry matter

production measured on uniform experimental plots with the parabolic

relationship which exists for the same parameters measured on a large field

scale. The departure from linearity was ascribed to nonuniformity of available soil water. In the same paper they demonstrated that nonuniformity of

water distribution lowers yields in crop ecosystems which are relatively well

supplied with water, but where precipitation is less than about 25% of the

optimum, nonuniformity of soil-water distribution may enhance the yield

averaged over large areas, allowing some parts of the crop to achieve nearer

their potential at the expense of others.

It may seem faint hearted to close on so intriguing an enigma. I hope that

the moral to be gained from it, and from other unexpected results which have

been mentioned in this paper, provide us with the stimulus to anticipate

nonuniformity as the norm, not the exception, and to adapt our thinking and



Tài liệu bạn tìm kiếm đã sẵn sàng tải về

VI. Water Uptake by Roots

Tải bản đầy đủ ngay(0 tr)

×