Tải bản đầy đủ - 0 (trang)
V. Patterns of Root Growth

V. Patterns of Root Growth

Tải bản đầy đủ - 0trang

128



ANN P. HAMBLIN



depth is often only 1.5 m. Nevertheless, the proliferation of roots in nonrestricted environments, such as well-fertilized topsoils, can be prodigious.

Graminaceous species may commonly have total root lengths of 2-6 x lo4 m

perm2 of ground area (Nye and Tinker, 1977), more than half of which

occurs in the surface soil. Are such profuse root concentrations advantageous

to the crop’s water and nutrient uptake, or do they represent a wasteful

squandering of assimilates similar to the production of nonfertile tillers in

cereals? The answer may well depend on the uniformity of water supply to

the crop during its life cycle. In some drought-affected environments, where

crops grow on stored water, it has been demonstrated that restriction of

either the roots or the shoots can give a more equitable water balance

between vegetative and reproductive growth in wheat (Passioura, 1972;

Islam and Sedgley, 1981). However, the root-shoot ratios of many droughted

crops are larger than those of well-watered crops (Schultz, 1974); and

~



Root Length LVrnrW3x lo4



0



1



2



3



4



5



6



7



0.4



Q



1.2



:::r

1.6

1.2



Spring Wheat

(Australia)



Oats

(Germany)



Frc. 7. Root-length densities (I+, m/m’) at anthesis for winter wheat grown in a sandy loam

over gravely clay [redrawn from Gregory et al. (1978a)], spring wheat in a loamy sand [redrawn

from Hamblin et al. (1982)], and oats in a loess [redrawn from Ehlers et al. (1981a)l.



THE INFLUENCE OF SOIL STRUCTURE



129



drought tolerance in some species has been correlated with those cultivars

which have longer, fine-branched root systems (OToole and Soemartono,

1981 ; Fischer and Turner, 1978). Nevertheless, the possibility of genetic

selection of genotypes with root characteristics which will make more

efficient use of water in the soil seems a daunting task in view of the enormous

plasticity displayed by many root systems over a range of environments and

the similarities in overall root morphologies between species, let alone

cultivars.

Figure 7 shows the distribution of root-length density (Lv;m/m3) for

winter wheat grown in a sandy loam over gravely clay in southern England, a

spring wheat grown in a loamy sand in Western Australia, and oats grown in

a podzolic loess in Germany. Both the spring wheat and the oat crop were

grown on sites which contained a traffic pan and were part of experiments

comparing tillage systems. Roots from the spring and winter wheats were

extracted by coring and washing techniques and grid counted after removal

of dead material. The oat roots were counted directly from cut faces of the soil

profile and a conversion factor was applied to transform root numbers to

lengths. Yet, for all of this, there are probably more similarities than

differences betpeen distribution, and in the case of the two wheats many of

the parameters are very similar. This may be consoling to theoretical

modelers of water extraction by roots, but how much scope does this allow

for detection of specific or varietal difference in root parameters designed to

improve crop water use?



A. GENOTYPIC

VARIATION

Criteria for optimal root configuration have been more thoroughly identified for nutrient than for water uptake. Admirable treatments of the nutrient

aspect have been given by Barley (1970) and Nye and Tinker (1977). This

may be due to the difficulty of identifying which parts of the root system are

taking up water. Some authors have computed that less than a tenth of the

total root length need be involved; Passioura (1980) calculated active root

length was 30 % in young wheat plants and Weatherley (1975) found less than

1.O % were necessary to satisfy the uptake equations he used. Moreover, most

measurements of field-grown roots do not consider the fraction of the root

system which is suberized and relatively impervious to water. Histochemical

staining techniques are available which may help to discriminate these

portions [e.g., Ward et al. (1978)], but more development is needed for

unambiguous detection of viable tissue from field environments. From field

evidence of root growth and water uptake, which will be discussed in detail in

Se=tion VI, there is some evidence to suggest that water is most rapidly taken



130



A N N P. HAMBLIN



up from younger roots in the 10-50-cm zone behind the root tips. This could

often represent less than 50% of the root length. However, Taylor and

Klepper (1975) could find no difference in root water uptake rates in cotton

when different parts of the soil-plant system were compared at equivalent

potentials-a point not always considered in other studies.

Identification and selection of specific root morphological or anatomical

characters for improving water-use efficency are thus hampered by lack of

knowledge on the functioning of the root system in realistic field situations.

Differences in root characters of cultivars have been demonstrated, but many

parameters measured on whole-root systems [as, for example, the work of

OBrien (1979) on spring wheats] are correlated with shoot development and

are profoundly influenced by the test environment. There is good evidence for

cereals (Gregory et al., 1978a; Klepper et al., 1984) to show that tiller

development may be synchronized with root axis production, and the

number of roots on a tiller may be predicted from the number of leaves on the

same tiller. Selection of specific features with a realized heritability, such as

root-hair length in clovers (Caradus, 1979) or xylem diameter in wheat

(Richards and Passioura, 198l), for well-defined environmental stress situations is clearly a more realistic goal.

Root characters which are adaptations to environments having substantially too little or too much water have been increasingly studied in recent

years. When water deficiency is a regular part of the crop’s environment,

either during vegetative growth or as an inevitable, increasing, drought stress

during maturation, “aridopassive” species maintain their transpiration rate

as near to the potential maximum for as long as possible, usually until less

than one-third of the available soil water is left (Fischer and Turner, 1978).

These authors suggest that adjustments to achieve this may include rapid

root growth into wetter soil zones (Ritchie, 1973) or increased root density

( L J . An ability for rapid lateral root initiation and for deep root elongation

would thus be advantageous. The depth-of-rooting argument is most convincingly demonstrated in soils of low water storage capacity [high K ( 0 )

values], in which water redistribution may well occur faster than root

extension, while rainfall continues. Species which continue to grow new

subsoil roots after flowering, or have deeper root systems anyway, may have

up to an additional 0.5-1.5 m of soil water storage to tap, which may be

equivalent to 100-200 mm of additional water during high evaporation

periods. Hamblin and Hamblin (1985) compared the maximum rooting

depths of forage and grain legumes in a Mediterranean climate on three

sandy soils in the same latitude which differed markedly in growing-season

rainfall. They found no significant differences between maximum root depth

within species between sites, but highly significant differences between species



THE INFLUENCE OF SOIL STRUCTURE



131



at each site. Lupin roots grew to 2 m, whereas pea and clover roots only grew

to an average depth of 0.7 m. Differences in rooting depth have also been

found between varieties. Kaspar et al. (1978) concluded from a review of

soybean data that new root extension continued to occur in soybeans as late

as the seed-fill stage, and that this frequently benefited seed yield in drier than

average seasons, despite the additional root sink for assimilates. They found

differences of 0.3 to 0.4 m in maximum root depth between seven commercial

cultivars over the flowering to pod-fill stages.

With the advent of newer high-yielding wheat varieties, which incorporate

dwarfing genes, concern has been expressed that root systems of these

varieties might also be shortened, with undesirable consequences when they

are grown in water- or nutrient-stressed environments. An early comparison

of European semidwarf and taller-stemmed winter wheat varieties was

carried out by Lupton et al. (1974). They concluded that there was no

significant difference in root density ( L J ,total root length per unit area (La),

or dry root weight at any stage between the semidwarf and tall varieties.

Cholick et al. (1977) made a similar comparison with varieties selected for

American rain-fed conditions and drew similar conclusions for a wet year.

However, in a drier-than-average year there was a significant difference in

soil-water depletion curves in favor of a semidwarf variety, though in this

instance all the varieties grew roots to depths of 3.0 m. Unfortunately,

detailed leaf- or root-water data were not given to allow a more complete

interpretation of reasons for this difference.

Recent studies in upland rice have highlighted some of the problems in the

correct identification and selection of “deep-rooting’’ characteristics. Jones et

al. (1979) reported that semidwarf upland rice varieties appeared less drought

resistant when compared with older, taller upland varieties when grown in

acid, coarse-textured Brazilian Latosols. The lowland rice parents from

which the semidwarf varieties had been selected had significantly lower L,

values, especially in the subsoils, when compared with the upland varieties.

However, the Brazilian subsoils were all very much more acid than the Asian

soils from which the lowland strains originated, and sensitivity to aluminium

toxicity was a confounding feature. Subsequent work by Mambani and La1

(1983a,b) has confirmed the relationship of drought resistance in upland rice

to subsoil root density, using less acid environments. Even then, the question

remains as to why some cultivars should exhibit more pronounced downward root elongation than others when their heights do not differ significantly. Downward extension is markedly affected by soil mechanical

impedance, the soil water status, and the turgor pressure in the plant.

Genotypic differences in the rate of root extension in compacted soils may

provide the answer to this question.



132



ANN P. HAMBLIN



B. ENVIRONMENTAL

INFLUENCES

1. Mechanical Impedance



Mechanical impedance is probably the most ubiquitous of physical

constraints to the unhindered growth of crop roots through soil. Soil

compaction restricts root growth, particularly in weakly pedal, silty, and

sandy soils, as a result of tillage and traffic (Barnes et al., 1971). Additionally,

many soil types contain indurated horizons such as fragipans, which are

naturally resistant to root penetration.

The pressure applied by growing root tips to the soil is approximately

radial and was shown by Greacen et al. (1968) to be most conveniently

measured by a cone-tipped cylindrical rod or probe. The root elongation rate

has an exponential function with penetrometer resistance in a given soil. As

bulk density decreases in any one soil, a suite of functional curves develop

further and further from the intercept (Eavis, 1972). The rate of root

elongation also decreases with increasing penetrometer resistance as soil

temperature and aeration decrease (Greacen et al., 1968).

Soil mechanical strength is dependent upon water content (or water

potential). The sheer strength of soil (z) with normal loading (a,) depends

upon the cohesive (coulombic) forces (C) and the angle of internal friction



($1:

T =



C



+ 6, tan



(10)



Coulombic forces are dependent on absorbed water layers and surface area;

therefore C increases with clay content and clay surface area and decreases at

high values of 8. The angle of internal friction depends on the angularity of

particles and their packing arrangement, as well as on 8. At high values of 8,

primary particles are lubricated and slide past each other. Some soils develop

secondary cementation as they dry, through the evaporation of soluble salts

and the dehydration of hydrous metal oxides, which form films around

primary and aggregated particles. Plinthite and fragipan layers, which are

soft when wet and very hard when dry, are examples of this.

Roots cannot normally grow into rigid pores narrower than their own

diameters (Wiersum, 1958). When such pores are encountered roots may be

able to exert sufficient pressure to expand the pores; otherwise they will be

deflected. The root tip is then buckled, often with expansion of the vertical

cross-sectional diameter (Camp and Lund, 1964) and a proliferation of root

hairs behind the tip. Considerable vertical pressure can be exerted by root

tips. Pfeffer’s (1893) experimental values of 10 MPa have been verified by

many more recent workers. Radial pressures are normally only half or less the

axial pressure, but they act over a much larger surface area of soil. Root-



THE INFLUENCE OF SOIL STRUCTURE



133



created pores, which are conspicuous in many otherwise massive clay subsoil

peds, have been formed by such pressures. Nevertheless, Russell and Goss

(1974) and Abdalla et al. (1969) measured reductions of 20% in the root

elongation rate at applied pressures as low as 5 kPa, and 80 % reductions at

50 kPa, the response curve being exponential. These are much greater

reductions than would be obtained with equivalent osmotic pressures in the

soil solution. However, their experimental system, of a flexible-membrane

pressure cell filled with ballatoni beads, may have given rise to larger

pressures at the root tip (through arching of the bead bed) than measured at

the external wall of the membrane. The problem of accurate measurement of

the forces encountered by roots and the relative strength measured by metal

probes (penetrometers) has been extensively treated by Barley and Graecen

(1967) and Graecen et al. (1968). Critical values at which root growth is

inhibited range from 1 to > 4 MPa penetrometer resistance, depending on

the soil's mechanical composition, the pore water pressure (total potential),

and the plant species. Ehlers et al. (1983) also pointed out that soil structural

differences down the profile can lead to differences in critical penetrometer

values. They found that a value of 3.6 MPa inhibited oat root growth in the

tilled Ap horizon, but that the critical value rose to 5 MPa in the untilled

subsoil because a continuous pore system had developed in these horizons,

from old root channels and worm holes, which roots could penetrate but

which could not be sensed by the penetrometer.

The effects of mechanical impedance on root and shoot growth differ under

field conditions from those observed in controlled environments. Most

laboratory experiments must perforce be constrained by scale considerations

to observations on the seedling or the young vegetative stages of plants, yet in

the case of crop-water relations the influence of the size of the root system is

most apparent when leaf areas are greater than 1 m2/m2 and the soil water

uptake demand exceeds 2 or 3 mm/day. In addition, controlled environments

cannot normally mimic thermal or energy gradients in the soil, plant canopy,

and air-canopy interface. Thus one or a few plants growing in a container

may have a very different distribution of pressure potentials, a different water

flux regime, and soil temperatures with larger diurnal fluctuations than in the

field. This may produce quite abberant growth rates and very different

root-shoot ratios from those of field-grown crops. Finally, the root tip and

meristematic region are the site of hormone production and maximum

nutrient uptake; complex feedback interactions occur with the shoots when

the environment of the root apices is varied, the outcome of which is not yet

physiologically predictable.

Greacen (1977) emphasized the distinction between roots growing through

sands and fine-structured soils, where pores are of smaller dimensions than

those of the roots, and root growth through coarse-structured, fine-textured



134



ANN P. HAMBLIN

5.0

4.5

16



4.0



14



3.5



-



-



12



3.0



-



5



10



2.5



E



a,



.ul



5m

C



2

;j



m

+

m

m



8



2.0



g



6



1.5



4



1.o



2



0.5



7



n

5



em



0



I



1



1



I



I



I



0.1



0.2



0.3



0.4



0.5



0.6



2



0.7



Fractional Soil Volume Intercepted

by Root Cylinder



FIG.8. The interaction of aggregate size (circles) and aggregate strength (diamonds) on the

o),

fractional soil volume intercepted by the cylinder (root plus root hair cylinder) for wheat (0

and peas (a+).(Constructed from data in Dexter, 1978.)



soils, where roots tend to be restricted to regions of lowest resistance (major

crack planes), despite the reduction which this imposes on lateral branching.

Dexter (1978) developed a model to predict the probability of a root either

entering or being deflected by an aggregate, depending on its strength and

size. The disposition of aggregates was defined in terms of the scales

encountered in a tilled topsoil. Although Dexter tested his model for nutrient

rather than water uptake, the results are relevant in the present context. With

the soil water potential held constant, the degree of root branching was

determined by aggregate size. The optimum ped structure depended on the

strength of aggregates and the plant species. Figure 8 is constructed from

Dexter’s data and shows the effects of aggregate size and strength on the

computed fractional soil volume which is intercepted by the root-plus-roothair cylinders for some crop species. Because increasing soil strength removed

the effect of aggregate size, Dexter hypothesized that greater root branching

occurred since the beds of aggregates were composed of progressively smaller

size. His model is stochastic and is based upon the statistical probability of

each root having either a penetrating or deflecting angle of incidence to each

aggregate as it approaches it, in a statistically generated arrangement of

solids and voids.



THE INFLUENCE OF SOIL STRUCTURE



135



The effect of. ped orientation on root penetration was considered by

Gardner and Danielson (1964). They found fewer root apices penetrated a

wax slab set in the soil the further the slab was tilted from the horizontal.

Whiteley and Dexter (1983) have extended this work to assess the influence of

ped surface incidence-angle and critical ped strength on the rate of root

elongation in several crop species growing in cracking clays. They found that

where roots were forced to grow in cracks at very different angles to the

preferred geotropic direction, extension rates were reduced to a half or a third

of the unrestricted rate, even when cracks were wide, soil water potentials

high, and soil strength less than 2 MPa. However, when cracks were oriented

at 45" to 90" from the horizontal, elongation rates were consistently higher,

irrespective of crack width, than for roots growing through soil peds. Even at

the high water potentials used (- 1 to 20 kPa) ped strength was a significant

barrier to penetration. This approach breaks new ground in quantifying the

interactions between soil structure and root growth. However, it still relies

heavily on relationships established in controlled environments with seedlings.

Many field studies have been made on the effect of traffic pans on rootgrowth parameters and the effect of subsoiling on crop yields. Loosening of

compacted layers often increases some aspect of root and (or) shoot growth

significantly, but the effect on yield is more variable. Russell (1956), for

example, found 53 % of subsoiled clay sites in England gave yield increases,

6 % gave negative yield responses, and 3 1 % showed no difference. On sandy

soils the figures were 32,26, and 42 %, respectively. In the drier environment

of Southwest Western Australia, on the other hand, Jarvis (1983) found

positive responses to ripping of traffic pans on sandy soils in 88 % of cases,

but in only 10% of clay-loam sites. Such simple yield data do not provide

sufficient clues as to why the soil loosening was or was not effective. Subsoils

may have very different nutrient or pH reactions from topsoils, and subsoiling may invert soil layers, diluting nutrient accumulation at the surface.

Waterlogging is a problem frequently related to mechanical impedance in

clay soils, and drainage may be improved by subsoiling, whereas substantial

loosening of sandy soil may cause seeding machinery to sink into soft surfaces

and sow crops too deeply. There are enough pointers in the literature,

however, to show that yield reduction in the presence of compacted soil layers

is associated with either or both water and aeration stress (Unger, 1979:

M.A.F.F., 1975). Fine sands and loamy sands, particularly those composed of

bimodally distributed sand sizes (Bodmin and Constantin, 1965; Coughlan et

al., 1978) are both readily compacted and also have low water and nutrient

capacities. For example, in the Mississippi River basin extensive coarsetextured alluvial soils occur, which develop severely restrictive traffic pans

with bulk densities of 1.75-2.0 ton/m3. Restriction to penetration frequently



136



A N N P. HAMBLIN



inhibits cotton and soybean roots to a total soil depth of only 0.25 m of soil in

which there may be, at best, only 30 mm available water-barely enough to

supply a leafy crop for a week (Camel, 1980). In other cases impedance is not

so severe, but still reduces growth significantly. Rowse and Stone (1981)

found deep loosening of a sandy clay loam increased the rate of root

elongation and gave significantly greater root lengths in the layers beneath

the loosened top one-half meter of the profile compared with the same depths

in shallow-ploughed treatments. More water was extracted from the deeper

parts of the profile and less from the top 0.3 m in the loosened treatments.

The authors argued that the pattern of root distribution in the loosened soil

led to lower total plant resistance to water flow and hence higher uptake and

transpiration rates, but the data were circumstantial.

In the experiments previously referred to in Fig. 7, where root densities

were compared, the amounts of water taken up from the traffic pan layers of

the loessial German soil by oats and the Australian loamy sand by wheat

were smaller than in the adjacent horizons. Yet, in the case of the oats, the

root density in the 0.2-0.3-m zone was not significantly different between two

tillage treatments, despite significant differences in bulk density, unsaturated

hydraulic conductivity, and penetrometer resistance. Ehlers et al. (1981b)

noted, however, that roots from this layer were much distorted and thickened

and suggested they might exert a far higher radial resistance to water uptake.

Other physiological alterations would have accompanied such anatomical

differences; however, they were not considered in their discussion.

2. Interactions between Mechanical impedance, Water Status, and Aeration

The interaction of soil strength, bulk density, and water content and their

relations to root impedance have been treated in many texts, but nowhere

better than in the review by Barley and Greacen (1967). The relationship

between penetrometer resistance, water potential, and soil pore space is not

so well defined in many field soils as in the idealized description in Section

V,B,l. Soils with high stone or gravel contents give spuriously high values for

penetrometer resistance, while vertically or areally heterogeneous soils have

high CVs, which may simply reflect differences in water content between

adjacent layers or peds. Normal averaging procedures of penetrometer values

obtained from such soils mask small-scale variations in soil strength and

water status which markedly influence root growth distribution. McIntyre

and Tanner (1959) and Hewitt and Dexter (1983) have reported positively

skewed penetrometer resistance distributions. Figure 9 shows variations

which can occur in the depth of penetration into a columnar-structured xeric

Alfisol (red-brown earth) taken at 0.1 m increments in a rectangular grid after

50mm of rain had fallen on a previously dry soil. There is a difference of



THE INFLUENCE OF SOIL STRUCTURE



137



crn



1 7 < 6 c m

18-24cm



m 6 - 1 2 c m

24-30cm



u 1 2 - 1 8 c m

-30-36crn



FIG.9. Variations in depth of penetration into a newly wetted, previously dry red-brown

earth (Rhodoxeralf) across a 1.8 x 0.6 rn grid. (From Hamblin, 1984.)



0.3 m in the depth to wetting-front, at which point the penetrometer’s

resistance exceeded 4.5 MPa (Hamblin, 1984).

When roots grow against materials which offer mechanical resistance, the

area of contact between roots and solid-plus-liquid phases increases. The

proportion of root surface exposed to gaseous oxygen is thus reduced. At the

same time, there is an increase (albeit small in proportion to the total

respiration rate) in the oxygen demand (Greenwood, 1968). In situations

where soils are already very wet, this reduction in air-filled pore space may be

critical to the oxygen status of the root, particularly if the root compresses the

soil as it passes through. I suggest that this may occur frequently in wet clay

soils, which have small median pore sizes.

Because the diffusivity of 0, in air is some lo4 times greater than in water,

the existence of a few continuous gas-filled pores to the soil surface may

suffice to provide sufficient aeration when root and soil respiration rates are

low (as in winter conditions in Europe, when temperatures are low and plants

are relatively small). However, the critical distance for O2 diffusion is seldom

the length of that continuous pathway to the soil surface; rather, it is the

distance through a water-filled soil volume from a root surface to the closest

aerobic pore cylinder. Greenwood (1968) has discussed the influence of pore

size on aeration, an&pointed out that for equivalent volumes of air-filled pore

space, oxygen diffusion into anaerobic peds occurs more rapidly the higher

the proportion of small pores, since the average diffusion path length across

the gas-liquid interface is less. He computed that a soil system composed of

m would require

gas channels where the radius of the channels rc = 5 x

a gas space ratio (x) of 0.2 to maintain an oxygen uptake rate of 1.3 x

lo-* s-l, whereas a soil of rc = 1 x

m would only need an x ratio of

0.05 to maintain the same rate.

In partially saturated soils (say, I,$,

= - 1 kPa) the actual supply of oxygen

to the roots is not well described by the bulk air-filled pore space (x). Nor is



138



ANN P. HAMBLIN



C

rn



Critical Value for Root Growth

- 1- -1 -1 -1 -1 -1 -1 -1 -1

I

1

1

1

1

1



W



>



6



0



4



8



12



16



20



24



28



% Pores > 30 urn Diameter



FIG.10. Oxygen diffusion rate for tropical Alfisols (squares) and Ultisols (circles) at 10 kPa

(0.) and 50 kPa (00).

(Constructed from data presented by Pla, 1978.)



the average oxygen concentration of the liquid phase very much more helpful.

Primary roots may be occupying larger gas-filled pores, while laterals with

smaller diameters are growing in pores which remain filled with water and

drain orders of magnituide more slowly. Hence the much quoted value of

x = 10 % necessary for unimpeded root growth is an oversimplification and

probably derives from the coincidence of the zero value for the diffusion ratio

DID, (where D is the self-diffusion rate of 0, in the porous medium and Do

the unimpeded rate in air) with 10% air porosity in many soils tested in the

laboratory (Currie, 1962). Data from Pla (1978) [quoted by La1 (1980)l

shows the relationship between macroporosity, bulk density, and 0, diffusion rates in compacted tropical Alfisols and Ultisols (Fig. 10). Some soils

were compacted to greater than 1.75 ton/m3 and yet still apparently retained

adequate 0, diffusion rates for root growth (see Fig. lo), since oxygen

diffusion rates of less than 3 x

pg/m2 s are taken as the limiting value

for restricting root growth. However, Armstrong (1980) has demonstrated

that 0, flux is a more accurate measure of root 0, demand than is

concentration. Currie (1984) attempted to relate the D/D,ratio to x in a silty

clay loam, wetted and compacted to different porosities, but found no general

relationship even for a single soil. It is probable that no such relationship can

be found given the variations in pore shape, tortuosity, degree of swelling,

and hysteresis which may occur as soils wet and drain.

Nevertheless, it is instructive to observe the general shape of the curves

which enabled Currie to identify the transition between inter- and intraaggregate gas diffusion (Fig. 11). As samples wetted and became swollen, the

relationship between x and D/Do followed an exponential form (DID,= axb),

but as the soil drained and the inter-aggregate pores became gas filled, the

relationship was fitted by a fourth-order polynomial of x. Thus in a sequence



Tài liệu bạn tìm kiếm đã sẵn sàng tải về

V. Patterns of Root Growth

Tải bản đầy đủ ngay(0 tr)

×