Tensile strength of soil cores in relation to aggregate strength, soil fragmentation and pore characteristics

Tensile strength of soil cores in relation to aggregate strength,soil fragmentation and pore characteristics

Lars J. Munkholma,*, Per Schjønninga, Bev D. Kayb

aDepartment of Crop Physiology and Soil Science, Danish Institute of Agricultural Sciences, Research Centre Foulum,

P.O. Box 50, DK-8830 Tjele, DenmarkbDepartment of Land Resource Science, University of Guelph, Guelph, Ont., Canada N1G 2W1

Abstract

Tensile failure of soil is desirable in tillage. Soil tensile strength estimates most often are obtained from compression tests of

dry aggregates. As tillage is performed under moist conditions, it would be highly relevant to measure tensile strength at high

water contents.

Plough-layer soil was sampled in a compacted soil (PAC) and in a non-compacted reference soil (REF). Tensile strength

was measured in a new direct tension test using undisturbed soil cores (4.45 cm in diameter and 5.00 cm in height) adjusted to

either �50 or �100 hPa matric potential. The air-filled pore space, ea, was determined from water retention measurements. Air

permeability, Ka, was determined at �30, �100 and �300 hPa matric potentials and from these measurements an index of

pore organization ðPO ¼ Ka=eaÞ was calculated. Soil behaviour in the field was evaluated at approximately �300 hPa matric

potential by measuring soil strength using the torsional shear box method and soil fragmentation using a simple soil drop test.

The direct tensile strength results showed that the PAC soil had significantly higher tensile strength than the REF soil (e.g. 3.2 and

2.0 kPa, respectively at �100 hPa matric potential). This finding was in accordance with the aggregate tensile strength results and

also agreed well with soil fragmentation in the field (i.e. geometric mean diameter (GMD) equal to 38.7 and 14.2 mm, respectively,

for PAC and REF samples dropped from 75 cm height). The tensile strengths of the soil cores were close to the predicted values

determined from the aggregate tensile strength results. The energy input in the soil drop test (i.e. approximately 8.9 J kg�1 dry soil)

was low in comparison with the energy input in tillage but high compared with the specific rupture energy of soil aggregates (e.g.

0.4 and 0.3 J kg�1, respectively for PAC and REF aggregates adjusted to �100 hPa matric potential). The relatively poor frag-

mentation in the soil drop test indicated that a substantial amount of the energy input was stored as volumetric strain energy and/or

lost to processes such as plastic deformation. The tensile strength of soil cores was negatively correlated to the macroporosity of

the soil, whereas the ease of soil fragmentation was positively correlated to PO. # 2002 Elsevier Science B.V. All rights reserved.

Keywords: Tensile strength; Direct tension test; Soil fragmentation; Soil strength; Soil pore characteristics; Pore organization

1. Introduction

In seedbed preparation, tensile failure is the desired

mode of soil failure in which the soil breaks down in to

smaller fragments without disturbing soil microstruc-

ture. Methods of measuring tensile strength in a

compression test on single aggregate (e.g. Dexter

and Kroesbergen, 1985) or on soil cores (e.g. the

Brazilian method (Kirkham et al., 1959)) are well

known. However, it is difficult to measure tensile

strength in a compression test at soil water contents

similar to those in the field at tillage (i.e. close to field

Soil & Tillage Research 64 (2002) 125–135

* Corresponding author. Tel.: þ45-899-91768;

fax: þ45-899-91719.

E-mail address: [email protected] (L.J. Munkholm).

0167-1987/02/$ – see front matter # 2002 Elsevier Science B.V. All rights reserved.

PII: S 0 1 6 7 - 1 9 8 7 ( 0 1 ) 0 0 2 5 0 - 1

water capacity for sandy and loamy Danish soils).

Plastic deformation will to some extent occur in wet

soil in compression. At low confining stress the mode

of failure is expected to shift from pure tensile in dry

state to shearing and compression in wet state (Hatibu

and Hettiaratchi, 1993). Therefore, aggregate tensile

strength has up till now almost exclusively been deter-

mined on air-dry or even oven-dry aggregates. Extra-

polating these results to soil behaviour in the field may

be problematic. This is particularly a problem when

seedbed preparation is carried out at water contents

around field capacity, which usually is the case in

humid climates like that in Denmark. Methods to mea-

suresoil tensilestrengthinadirect tension testhavebeen

introduced by a number of authors (Gill, 1959; Farrell

et al., 1967; Nearing et al., 1988; Junge et al., 2000).

However, tensile strength was measured on remoulded

soil packed into cylinders or moulds in all cases.

Ease of soil fragmentation has been assessed by the

drop-shatter test suggested by Marshall and Quirk

(1950) and refined by Hadas and Wolf (1984). These

authors applied the test to relatively dry soil. Recently,

Schjønning et al. (2001) have applied the drop-shatter

test to field moist soil. Although a relation between

tensile strength and the ease of soil fragmentation is

implicitly assumed, only a few researchers have

related strength measurements to soil fragmentation

in the field (e.g. Macks et al., 1996).

Soil pore characteristics are expected to affect the

soil mechanical behaviour, especially when tensile

failure occurs. Tensile failure occurs due to the pro-

pagation of cracks in the stressed sample (e.g. Hallett

et al., 1995a). An applied stress is concentrated at

crack tips and the crack propagates if the stress

exceeds the strength in the crack tip, which may lead

to catastrophic failure of the sample. The stress con-

centration increases with increased length and narrow-

ness of the cracks tips. Therefore, stress concentration

is expected to increase with increased pore continuity

and to decrease with increased pore tortuousity. For a

moist soil, stress concentration will take place in the

air-filled cracks and pores. The water-filled pores will

show no stress concentration because the load is

uniformly borne by the pore water (Snyder and Miller,

1989). Experimental results have confirmed the strong

influence of air-filled cracks and pores on soil frag-

mentation and tensile strength. Hallett et al. (1995b)

found that dry natural soil blocks fragmented mainly

along pre-existing crack surfaces. Guerif (1990) found

a strong negative correlation between macroporosity

and tensile strength of dry soil.

The primary objective of this study was to evaluate

a method to measure tensile strength in a direct tension

test on undisturbed, field-sampled soil cores at matric

potentials around field capacity. An additional objec-

tive was to investigate the relation between tensile

strength or rupture energy of cores and aggregates on

one hand and soil fragmentation and soil pore char-

acteristics on the other hand.

2. Materials and methods

2.1. Soil type and field trial

The soils used in this study were obtained from a

field experiment conducted on a sandy loam located at

Rugballegard Experimental Station, Denmark. Some

basic soil characteristics are presented in Table 1. A

compacted soil, labelled (PAC) was compared with a

non-compacted reference soil, labelled (REF). The

experiment was established in 1997 and the treatments

were applied to plots in a randomized block design

with three blocks every year. In each experimental

year (1997–1999), the soil was mouldboard ploughed

to 20 cm depth at or above a water content of field

capacity. Immediately after ploughing, the PAC treat-

ment was applied by creating adjacent wheel tracks

using a 6–8 t tractor (inflation pressure 125 kPa).

Secondary tillage and sowing was carried out later in

the spring using a rotary harrow and a traditional drill

at moisture contents around field capacity. Small grain

cereals were grown in all experimental years.

Table 1

Basic soil characteristics for the soil under investigation as averaged

for both tillage treatments (no significant difference)

Soil parameter

Organic matter (g 100 g�1) 3.1

Clay, <2 mm (g 100 g�1) 13

Silt, 2–20 mm (g 100 g�1) 14

Fine sand, 20–200 mm (g 100 g�1) 39

Coarse sand, 200–2000 mm (g 100 g�1) 31

pH (CaCl2) 5.9

CECa (mmolc kg�1) 111

a Cation exchange capacity (Kalra and Maynard, 1991).

126 L.J. Munkholm et al. / Soil & Tillage Research 64 (2002) 125–135

2.2. Sampling and soil preparation

Soil sampling was performed in May 1999 shortly

after plant emergence at water content slightly lower

than field water capacity (Table 2). Soil was sampled

from the 7 to 12 cm depth in order to collect only

plough-layer soil that had not been directly disturbed

by the seedbed preparation and sowing operations.

Within each plot, samples were taken in three ran-

domly located 0.5 m2 areas.

Two-piece steel cylinders were constructed. Each

cylinder had an inner and outer diameter of 4.45 and

5.09 cm, respectively, and a height of 2.5 cm. To

secure sampling of undisturbed soil, the edge of the

bottom cylinder was bevelled outwards. A strong PVC

tape was used to hold the cylinders together during

sampling and storage. Undisturbed soil cores were

taken using the unified two-piece cylinders described

above. These were driven horizontally into the walls of

carefully created soil pits by means of a hammer. The

two-piece cylinders were held in position by a special

flange ensuring a horizontal sideward movement into

the soil wall. After careful removal of the soil-filled

cylinder from the bulk soil, the end surfaces were

trimmed with a knife. The cylinders were protected

from physical disruption and evaporation during trans-

port and later storage by means of a plastic lid. Thirty

soil cores were retrieved in each plot. The soil cores

were stored at 2 8C before adjusting their matric

potential and testing for tensile strength. The samples

were saturated on sandboxes and adjusted to matric

potentials of �50, �100 or �300 hPa.

Additional soil cores (6.1 cm diameter, 3.4 cm

height, 100 cm3 volume) were collected at the 7–

12 cm depth. These cores were collected using metal

cylinders forced into the soil by means of a hammer.

The cylinders were held in position by a special flange

ensuring a vertical downward movement into the soil.

After careful removal of the soil-filled cylinder from

the bulk soil, the end surfaces were trimmed with a

knife and mounted with plastic caps. Two replicate

samples were collected at each of the three sampling

areas in all plots (i.e. 18 replicates per treatment). The

minimally disturbed soil cores were stored at 2 8Cuntil analysis for pore characteristics could take place.

2.3. Direct tension test

The direct tension test was performed in an auto-

matically operated mechanical press (Fig. 1). The

lower half of the two-piece cylinder was fixed in a

specially designed rigid frame by the use of three

screws horizontally driven against the cylinder wall. A

plastic cap was put on the upper half of the two-piece

cylinder and fixed similarly. The top cap was con-

nected to the pressure transducer by a steel bar of

adjustable length. Immediately before testing, the tape

that held the two-piece cylinder together was pulled

off as gently as possible. The cylinders were pulled

apart with a longitudinal strain rate of 2 mm min�1.

The force was measured by a pressure transducer

ð0�100 � 0:03 NÞ and recorded automatically on a

computer. After the tensile strength measurement, the

top cylinder with soil was weighed in order to account

for the force needed to counterbalance the weight of

the sample.

Problems with the soil cores sliding in the rings

were encountered in the tests. Therefore, an approxi-

mately 3 mm deep furrow was made with a knife at the

boundary to the metal ring. This was performed both

at the bottom and at the top of the two-piece cylinder.

Silicone glue was squeezed into the furrow to improve

the contact between the soil and the metal ring. This

worked out quite well for soil tested at �50 and

�100 hPa, but was not applicable for soil drained to

�300 hPa.

Table 2

Water retention, density and porosity of the investigated soils

(7–12 cm depth)

Treatment

PAC REF

Water content (sampling) (m3 100 m�3) 30.0 aa 29.2 a

Water content (�100 hPa) (m3 100 m�3) 34.3 b 32.7 a

Water content (�300 hPa) (m3 100 m�3) 30.5 b 28.1 a

Particle density (Mg m�3)b 2.613 a 2.609 a

Bulk density (Mg m�3) 1.57 b 1.41 a

Porosity (m3 100 m�3) 39.7 a 46.2 b

Density, 2–4 mm aggregates (Mg m�3)c 1.83 b 1.76 a

Density, 4–8 mm aggregates (Mg m�3)c 1.71 b 1.63 a

Density, 8–16 mm aggregates (Mg m�3)c 1.67 b 1.61 a

a Figures followed by the same letter are not significantly

different at the P < 0:05 level.b Determined from the pycnometer method (Blake and Hartge,

1986).c Data from Munkholm and Kay (2002).

L.J. Munkholm et al. / Soil & Tillage Research 64 (2002) 125–135 127

As the soil cores were kept inside the cylinders

during the tension test, some soil–metal friction could

not be avoided, i.e. the cores did not always break at

the interface between the two cylinders. The cores that

broke at the very top or bottom were rejected. To be

able to take account of the soil–metal friction, the

height of the broken halves of non-rejected cores was

measured after the test. The deviation from a break

exactly at the interface between the cylinders (labelled

breakpoint deviation) was calculated as the mean

height of soil measured above the lower half cylinder

plus the mean height of soil above the upper half

cylinder. The height of the broken soil core above the

steel cylinder was measured at three points for both

halves. A number of 30 soil cores were tested for each

combination of treatment and matric potential.

The tensile strength of the soil cores (Ycore) was

calculated as

Ycore ¼Fmax � Flift

Acore

(1)

where Fmax (N) is the maximum longitudinal force

measured, Flift (N) the force needed counterbalance

the weight of the sample and Acore the sectional area of

the soil core.

2.4. Pore size distribution

The soil cores were capillary wetted to saturation in

sandboxes and then drained progressively to matric

potentials of �4, �10, �16, �30, �50, �100, �160,

�300 and �1000 hPa. Finally, they were oven-dried

at 105 8C for 24 h. The weight of each sample was

recorded at each matric potential and after oven

drying. Soil porosity was estimated from bulk soil

density and particle density (Table 2). The volumetric

water content at each potential was calculated from

the weight loss on oven drying. Air-filled porosity,

ea, at a given matric potential was calculated as the

difference between total porosity and the water

retained at the specific potential. An estimate of water

retention at the �1.5 MPa, W�1.5 MPa, matric poten-

tial (permanent wilting point) was calculated from

the pedo-transfer function obtained by Hansen

(1976).

2.5. Air permeability

Air permeability, Ka, was measured by the steady-

state method recently described by Iversen et al.

(2001) on cores that had been equilibrated at the

�30, �100 and �300 hPa matric potentials. The flow

of air through the sample was controlled by a pressure

regulator and measured by a precision flow meter.

The flow was recorded at a pressure difference of

5 hPa as measured by a water manometer. Prior to

permeability measurements, gentle pressing the soil

at the very edge of the metal ring minimized the risk

of air leaking along the ring (Ball and Schjønning,

2002). An empirical index of pore continuity/pore

organization (PO, mm2) was calculated (Groenevelt

et al., 1984):

PO ¼ Ka

ea

(2)

High values of PO express a high capacity of a given

air-filled pore volume as to conduct air, i.e., a high

continuity or a low tortuousity.

Fig. 1. The experimental set-up of the direct tension test on soil

cores: (A) adjustable steel bar connected to the pressure transducer;

(B) plastic cap attached to the upper half of the two-piece cylinder;

(C) two-piece cylinder enclosing the undisturbed soil core; (D)

rigid frame to which the lower half of the two-piece cylinder is

fastened.

128 L.J. Munkholm et al. / Soil & Tillage Research 64 (2002) 125–135

2.6. Soil fragmentation

The ease of soil fragmentation was determined

in the field just after plant emergence in May 1999

using the soil drop test described by Schjønning et al.

(2001). The soil drop test is a modification of the drop-

shatter fragmentation method described by Hadas

and Wolf (1984). An undisturbed cubic soil sample

ð7:0 cm � 8:0 cm � 11:5 cmÞ was collected from a

depth of 7–14 cm with a specially designed metal

shovel (Schjønning et al., 2001). The field moist soil

cube was then dropped from a height of 75 cm into a

metal box and the aggregate size distribution deter-

mined. The field moist soil was passed through a nest

of sieves with openings of 32, 16, 8, 4 and 2 mm and

each size fraction was recorded by weighing the mass

of soil. Non-dropped cubic samples were also sieved

and the size distribution recorded. The fragmentation

was expressed as thegeometric mean diameter for drop-

ped (GMDD, mm) and non-dropped samples (GMDN,

mm). The relative change in GMD from non-dropped

to dropped DGMD/GMDN was also calculated. Nine

replicate measurements were carried out for dropped

as well as non-dropped samples (i.e. three per plot).

2.7. Shear strength

In situ shear strength, t, was determined at the time

of sampling when the matric potential was close to

�300 hPa. A 10.0 cm diameter torsional shear box

operated at five normal stresses, P, ranging from 7.3 to

32.3 kPa was placed on the soil surface after removal

of the 0–7 cm soil layer and forced to a depth of

about 12 cm below the original surface (Payne and

Fountaine, 1952). The soil surrounding the metal box

was carefully removed before shearing. For all soils,

one series of determinations (five normal stresses) was

performed at two of the three sampling areas within

each plot, i.e. six determinations for each combination

of normal load and traffic treatment.

2.8. Statistical analysis

All parameters were analysed for normal distribu-

tion and transformed to yield normality if necessary.

Averages were calculated for each sampling area and

used in the F-tests, taking the inter-sampling area

variation as the residual error.

3. Results and discussion

3.1. Tensile strength of soil cores and aggregates

The suggested method was applicable at high

matric potentials (�50 and �100 hPa), while it failed

to work at �300 hPa. To be able to measure tensile

strength at lower matric potentials the approach of

Farrell et al. (1967) may be applicable. They used epoxy

resin, Araldite, to seal the ends of the soil cores to end

plates.

The heavily trafficked soil had significantly higher

tensilestrengthatbothmatricpotentials tested(Table3).

The difference in strength between the matric poten-

tials tested was rather small and the measured strengths

were also small. According to Farrell et al. (1967),

tensile strength increases only slightly with decreasing

potential at matric potentials around �100 hPa. More-

over, it was more difficult to measure tensile strength at

�100 hPa than at �50 hPa. At the lower matric poten-

tial, more problems with soil sliding in the metal cylin-

ders occurred to a greater extent and the deviation from

a break at the cylinder interface (breakpoint deviation)

increased. Although the PAC treated soil had a rather

largedeviationfromabreakat the interfaceat�100 hPa

matric potential, the effect of breakpoint deviation was

not significant.

The range of tensile strengths measured in this

study (2.0–3.2 kPa) agrees with results reported by

Nearing et al. (1988). They measured tensile strength

on repacked soil cores (3.88 cm in diameter) at matric

potentials comparable to those applied in this study.

On the other hand, Farrell et al. (1967) measured

tensile strength between 10 and 15 kPa on repacked

soil cores (3.8 cm in diameter) at matric potentials

between �100 and �500 hPa. The latter study used

soil with approximately the same textural composi-

tion as the soil used in this investigation, however,

remoulded and packed to a bulk density as high as

1.7 Mg m�3.

The results from the direct tensile strength mea-

surements agree well with the aggregate tensile

strengths determined by Munkholm and Kay (2002)

for the same treatments (Table 3). In both cases, the

PAC soil showed the largest tensile strength. The

tensile strengths of the soil cores were markedly lower

than those of the soil aggregates measured at the same

matric potential (�100 hPa, Table 3). In general,

L.J. Munkholm et al. / Soil & Tillage Research 64 (2002) 125–135 129

tensile strength has been found to scale linearly with

sample size on a log–log scale (Utomo and Dexter,

1981). As illustrated in Fig. 2, the predicted tensile

strength (from regression line) of soil units with a

cross-sectional area of 1555 mm2 is close to the tensile

strength measured on soil cores with the same cross-

sectional area. This finding thus indicates an excellent

correlation between the results obtained by the direct

and the indirect tension tests.

3.2. Soil shear strength and fragmentation

The data obtained in the laboratory correspond well

with the observed soil behaviour in the field. The PAC

soil had significantly higher shear strength at all

normal loads (Fig. 3). The treatments differed signifi-

cantly in apparent soil cohesion (i.e. intercept with the

y-axis, Fig. 3), but not in internal friction (slope of

lines). It is remarkable that both treatments showed

approximately 10 times higher apparent soil cohesion

than tensile strength of soil cores even though the

measurements were performed at fairly similar matric

potentials. The torsional shear test operated at low

normal loads (<32 kPa) allows the soil to fail along

‘‘natural’’ weak planes, i.e. as a result of crack pro-

pagation, the same as for the tension test. However,

torsion loading is less effective than tension loading to

induce crack propagation (Anderson, 1995) and inter-

nal friction may to some extent have been included in

the estimate of apparent soil cohesion. Inhomogeneity

Table 3

Tensile strength, Y, measured on undisturbed soil cores in a direct tension test at �50 and �100 hPa (geometric mean values), and deviation

from a break exact at the interface (breakpoint deviation) between the cylinders (geometric mean values)a

Matric potential (hPa) Treatment

PAC REF

Tensile strength (soil cores) (kPa) �50 3.1 bb 2.1 a

Tensile strength (soil cores) (kPa) �100 3.2 b 2.0 a

Breakpoint deviation (mm) �50 3.0 a 3.3 a

Breakpoint deviation (mm) �100 6.4 b 3.9 a

Tensile strength (2–4 mm aggregates) (kPa)c �100 16.9 a 15.5 a

Tensile strength (4–8 mm aggregates) (kPa)c �100 11.7 b 9.6 a

Tensile strength (8–16 mm aggregates) (kPa)c �100 7.1 a 5.9 a

Rupture energy (2–4 mm aggregates) (J kg�1)c �100 1.7 a 1.8 a

Rupture energy (4–8 mm aggregates) (J kg�1)c �100 0.7 a 0.8 a

Rupture energy (8–16 mm aggregates) (J kg�1) �100 0.4 a 0.3 a

Soil fragmentation, non-dropped (GMDN, mm) Sampling 44.1 b 30.2 a

Soil fragmentation, dropped (GMDD, mm) Sampling 38.7 b 14.2 a

Soil fragmentation, DGMD/GMDN (%)d Sampling 13 a 53 b

a Tensile strength, Y, measured on single aggregates in an indirect tension test (geometric mean values) and specific rupture energy

(geometric mean values). Soil drop test results—GMD for dropped and non-dropped samples (field test at sampling).b Figures with the same letter are not significantly different at P ¼ 0:05 level.c Data from Munkholm and Kay (2002).d DGMD ¼ GMDN � GMDD.

Fig. 2. The relationship between tensile strength, Y (kPa) and

sample cross-sectional area, A (mm2) on a log–log scale. Lines

indicate linear regression for log aggregate tensile strength vs. log

aggregate cross-sectional area (i.e., direct tension test results not

included). All measurements were performed at �100 hPa matric

potential: (—) PAC, (- - -) REF; bars indicate �standard error of

the mean.

130 L.J. Munkholm et al. / Soil & Tillage Research 64 (2002) 125–135

in crack and pore characteristics may also explain

some of the difference as the soil was forced to fail

along the horizontal and the vertical plane, respec-

tively, for the torsional shear and tension tests.

The PAC soil fragmented poorly when performing

the soil drop test (Table 3 and Fig. 4). Both with and

without dropping, the PAC soil fragmented into large

clods, whereas the REF soil fragmented into signifi-

cantly smaller aggregates. The relative change in

GMD from non-dropped to dropped (DGMD/GMDN)

may be considered as an empirical index of ease of

fragmentation and showed a strong treatment effect

(i.e. DGMD=GMDN ¼ 53 and 13% for REF and PAC,

respectively).

The extent of correlation between the tensile

strength of soil cores and GMDD or DGMD/GMDN

varied with matric potential and compaction treatment.

Within each treatment no significant correlation was

found between DGMD/GMDN and tensile strength of

the soil cores (data not shown). For the REF soil, a

significant positive correlation ðR2 ¼ 0:65��Þ was

found between GMDD and tensile strength of soil

cores measured at �50 hPa matric potential but not

at �100 hPa. Such a correlation could not be found

for the PAC treated soil, which exhibited little frag-

mentation in the soil drop test (data not shown).

Obviously, the energy input in the soil drop test was

too low to induce substantial fragmentation of the

PAC treated soil. The energy input in the soil drop test

is approximately 7.4 J kg�1 field moist soil, which

corresponds approximately to 8.9 J kg�1 dry soil when

assuming a wet and dry bulk density of 1.7 and

1.4 g cm�3, respectively. This level is low in compar-

ison with the estimated energy input in different tillage

operations. For instance, Patterson et al. (1980) esti-

mated that a secondary tine cultivation on clay loam

required an energy input of 24 MJ ha�1 at the imple-

ment connection. When assuming a dry bulk density

of 1.4 g cm�3 and a tillage depth of 5 cm this corre-

sponds approximately to an energy input of 34 J kg�1

dry soil. However, the energy input in the soil drop test

supposedly is realistic as a substantial part of the

Fig. 3. Peak shear strength, t, related to normal load, P, as

determined by the torsional shear box method in the field at field

capacity: (—) PAC, (- - -) REF; bars indicate �standard error of the

mean.

Fig. 4. Aggregate size distribution for the soils when subjected to the soil drop test at approximately �300 hPa in the field: (—) PAC, (- - -)

REF; bars indicate þ1 standard error of the mean.

L.J. Munkholm et al. / Soil & Tillage Research 64 (2002) 125–135 131

energy input in tillage is lost in friction, and transfer

and re-orientation of the structural units (Hadas and

Wolf, 1983).

Based on the rupture energy of the aggregates,

much higher soil fragmentation would have been

expected in the soil drop test, i.e. the rupture energy

was <2 J kg�1 dry soil for all aggregate sizes at

�100 hPa (Table 3). The relatively low fragmentation

implies that a substantial amount of the energy input in

the soil drop test also did not dissipate into soil

fragmentation. It may have been stored as volumetric

strain energy (Chancellor et al., 1969) and/or lost as

fragment rebound and heat evolution related to plas-

tic deformation. It is also noteworthy that there was

no difference between the treatments in the specific

rupture energy of soil aggregates. This was evident at

all potentials in the range �100 hPa to �166 MPa

(Munkholm and Kay, 2002). The results imply that

caution must be taken in predicting soil fragmenta-

tion in tillage from measurements of rupture energy of

single aggregates.

3.3. Soil pore characteristics

The matric potential at sampling and when perform-

ing field measurement was close to �300 hPa, as indi-

cated by the water retention results in Table 2. The PAC

soil contained significantly more water at all the matric

potentials. The compaction treatment resulted in lower

total porosity (i.e. 40 and 46 m3 100 m�3 for PAC and

REF, respectively) (Table2).Likewise thedensityof the

aggregates was also highest for the PAC treatment (e.g.

1.71 and 1.63 Mg m�3 for 4–8 mm aggregates from the

PAC and REF, respectively). The compaction treatment

caused a decrease in macroporosity ðpores > 30 mmÞ(6 and 13 m3 100 m�3 pores > 30 mmforPACandREF,

respectively) (Fig. 5). The compaction treatment also

significantly reduced the volume fraction of medium

sized pores (10–30 mm) (Fig. 5). The influence of soil

compaction on the large pores agrees with other studies

(Rasmussen, 1985; McAfee et al., 1989).

The compaction treatment also caused a significant

reduction in the ability of the soil to transport gas by

convection (i.e., a lower air permeability, Ka) (Table 4).

Moreover, the significantly lower index of pore orga-

nization (PO) found for the PAC soil at �30 and

�100 hPa indicate a relatively poorer continuity

of the macro-pores in the PAC treated soil. At

�300 hPa there was no significant difference between

the soils in pore continuity. The negative effect of soil

traffic on gas transport characteristics and PO found in

this study corresponds with the results by Ball et al.

(1988) and the results from simulated traffic in the

laboratory by O’Sullivan et al. (1999).

3.4. Influence of pore characteristics on tensile

strength and soil fragmentation

The tensile strength of the soil cores was signifi-

cantly negatively related to macroporosity. At a matric

potential of �50 hPa, the tensile strength of soil cores

Fig. 5. Fractions of pores of selected size classes as derived from

water retention measurements, assuming the approximate relation

d ¼ �3000=Cm, where d is the tube equivalent pore diameter in

mm and Cm the matric potential in hPa. Fractions labelled with

different letters (a/b) are significantly different ðP ¼ 0:05Þ.

Table 4

Air-filled pore space, ea, air permeability, Ka (geometric meanvalues)

and PO ðPO ¼ Ka=eaÞ (geometric mean values) for soil sampled

vertically at the 7–12 cm depth

�30 hPa �100 hPa �300 hPa

PAC REF PAC REF PAC REF

ea (m3 100 m3) 0.03 aa 0.08 b 0.06 a 0.13 b 0.09 a 0.18 b

Ka (mm2) 0.28 a 1.18 b 0.68 a 2.82 b 2.05 a 6.91 ab

PO (mm2) 7.1 a 21.2 b 16.0 a 33.2 b 28.6 a 45.1 ab

a Values in the same row and at the same matric potential with

the same letter are not significantly different at the P < 0:05 level.b Significant interaction between treatment and block. When

testing against treatment � block variation no significant difference

between treatments.

132 L.J. Munkholm et al. / Soil & Tillage Research 64 (2002) 125–135

showed a linear decrease with increasing volume

fraction of >60 mm pores ðR2 ¼ 0:54���Þ (Fig. 6, left).

At �100 hPa, the tensile strength decreased linearly

with volume fraction of >30 mm pores (Fig. 6, right).

The observed significant influence of macroporosity

on tensile strength is in accordance with studies by

Guerif (1990) and Hallett et al. (1995b). The correla-

tion between tensile strength and pore characteristics

were not improved by including the index of PO in the

statistical model (data not shown). In itself, there was

no significant correlation between PO and tensile

strength within the treatments (data not shown). The

poor correlation may be due to a large small-scale

variability of air permeability, Ka (e.g. Koszinski et al.,

1995; Iversen et al., 2001). Therefore, the PO in the

failure zone in the cores used for tensile strength

measurements may be distinctly different than in those

taken for measurement of pore characteristics.

Interestingly, the soil fragmentation data showed a

stronger correlation to PO than to total or air-filled

porosity. Only a weak and insignificant relation was

found between soil fragmentation and total or air-

filled porosity, although GMDD tended to decrease

with total porosity and air-filled porosity, ea, at �50

and �100 hPa. In contrast, the GMDD decreased

linearly with log PO at �100 hPa for the REF treat-

ment ðR2 ¼ 0:76��Þ (Fig. 7), whereas DGMD/GMDN

increased significantly with log PO at �100 hPa when

including all data ðR2 ¼ 0:78��Þ (data not shown). The

results imply that soil fragmentation depends on the

geometrical characteristics of the air-filled pores and

cracks, which is in accordance with brittle fracture

mechanics (Hallett et al., 1995a). A high PO indicates

that the air-filled pore space to a large extent consisted

of long straight pores and cracks, i.e. pores and cracks

where crack propagation most likely will start.

4. Conclusions

Tensile strength of moist soil can be measured using

direct tension methods without making assumptions

about the mode of failure. The suggested method

worked well at high matric potentials (�50 and

Fig. 6. Tensile strength of soil cores measured at �50 and �100 hPa as related to the volume fraction of pores >60 and >30 mm, respectively.

The symbols represent a mean value for each sampling area. Lines represent linear regressions.

Fig. 7. GMD of dropped samples, GMDD, as related to the pore

continuity, PO of air-filled pores at �100 hPa matric potential.

L.J. Munkholm et al. / Soil & Tillage Research 64 (2002) 125–135 133

�100 hPa), while it failed to work at �300 hPa. Modi-

fications of the method may make it possible to

measure tensile strength at lower matric potentials.

The heavily trafficked soil had a significantly higher

tensile strength than the REF soil in the direct tension

test. The results corresponded well with the predicted

values determined from indirect measurements of

aggregate tensile strength.

The tensile strength results were in accordance with

soil behaviour in the field as estimated by shear

strength and soil fragmentation techniques.

The energy input in the soil drop test was low in

comparison with the energy input in tillage but high

compared with the specific rupture energy of soil

aggregates. For the latter, no difference was found

between the treatments. The relatively low fragmenta-

tion of the soils in the soil drop test indicated that a

considerable amount of the energy input was stored as

volumetric strain energy and/or lost to processes such

as plastic deformation. Furthermore, the findings in

this study imply that caution must be taken in pre-

dicting soil fragmentation in tillage solely from mea-

surements of rupture energy of single aggregates.

The tensile strength of the soil cores was negatively

correlated to the macroporosity of the soil, whereas the

ease of soil fragmentation was positively correlated to

PO.

Acknowledgements

The technical assistance of Bodil B. Christensen

and Michael Koppelgaard is gratefully acknowledged.

This work was financed by the Danish Environmental

Research Programme and was performed in the con-

textof theDanishResearchCenterforOrganicFarming.

References

Anderson, T.L., 1995. Fracture Mechanics—Fundamentals and

Applications, 2nd Edition. CRC Press, Boca Raton, FL, 688 pp.

Ball, B.C., Schjønning, P., 2002. Air permeability. In: Dane, J.H.,

Topp, G.C. (Eds.), Methods of Soil Analysis, Part 1, 3rd

Edition. Agronomy Monograph. ASA and SSSA, Madison, WI,

in press.

Ball, B.C., O’Sullivan, M.F., Hunter, R., 1988. Gas diffusion, fluid

flow and derived pore continuity indices in relation to vehicle

traffic and tillage. J. Soil Sci. 39, 327–339.

Blake, G.R., Hartge, K.H., 1986. Particle density. In: Klute, A.

(Ed.), Methods of Soil Analysis. Part 1. Physical and

Mineralogical Methods. American Society of Agronomy,

Madison, WI, pp. 377–382.

Chancellor, W.J., Vomocil, J.A., Aref, K.S., 1969. Energy

disposition in compression of three agricultural soils. Trans.

ASAE 12, 524–528.

Dexter, A.R., Kroesbergen, B., 1985. Methodology for determina-

tion of tensile strength of soil aggregates. J. Agric. Eng. Res.

31, 139–147.

Farrell, D.A., Greacen, E.L., Larson, W.E., 1967. The effect of

water content on axial strain in a loam soil under tension and

compression. Soil Sci. Soc. Am. Proc. 31, 445–450.

Gill, W.R., 1959. The effects of drying on mechanical strength of

Lloyd clay. Soil Sci. Soc. Am. Proc. 3, 255–257.

Groenevelt, P.H., Kay, B.D., Grant, C.D., 1984. Physical assess-

ment of a soil with respect to rooting potential. Geoderma 34,

101–114.

Guerif, J., 1990. Factors influencing compaction-induced increases

in soil strength. Soil Till. Res. 16, 167–178.

Hadas, A., Wolf, D., 1983. Energy efficiency in tilling dry clod-

forming soils. Soil Till. Res. 3, 47–59.

Hadas, A., Wolf, D., 1984. Refinement and re-evaluation of the

drop-shatter soil fragmentation method. Soil Till. Res. 4, 237–

249.

Hallett, P.D., Dexter, A.R., Seville, J.P.K., 1995a. The application

of fracture mechanics to crack propagation in dry soil. Eur. J.

Soil Sci. 46, 591–599.

Hallett, P.D., Dexter, A.R., Seville, J.P.K., 1995b. Identification of

pre-existing cracks on soil fracture surfaces using dye. Soil Till.

Res. 33, 163–184.

Hansen, L., 1976. Jordtyper ved statens forsøgsstationer. Tidsskr.

Planteavl. 80, 742–758.

Hatibu, N., Hettiaratchi, D.R.P., 1993. The transition from ductile

flow to brittle failure in unsaturated soils. J. Agric. Eng. Res.

54, 319–328.

Iversen, B.V., Schjønning, P., Poulsen, T.G., Moldrup, P., 2001. In

situ, on-site and laboratory measurements of soil air perme-

ability: boundary conditions and measurement scale. Soil Sci.

166, 97–106.

Junge, T., Grasle, W., Gisela, B., Horn, R., 2000. Effect of pore

water pressure on tensile strength. J. Plant Nutr. Soil Sci. 163,

21–26.

Kalra, Y.P., Maynard, D.G., 1991. Methods Manual for Forest Soil

and Plant Analysis. Information Report No. R-x-319. Forestry

Canada, Northwest Region, Northern Forestry Centre, Canada.

Kirkham, D., De Boodt, M.F., De Leenheer, L., 1959. Modulus of

rupture determination on undisturbed soil core samples. Soil

Sci. 87, 141–144.

Koszinski, S., Wendroth, O., Lehfeldt, J., 1995. Field scale

heterogeneity of soil structural properties in a moraine land-

scape of northeastern Germany. Int. Agrophys. 9, 201–210.

Macks, S.P., Murphy, B.W., Cresswell, H.P., Koen, T.B., 1996. Soil

friability in relation to management history and suitability for

direct drilling. Aust. J. Soil Res. 34, 343–360.

Marshall, T.J., Quirk, J.P., 1950. Stability of structural aggregates

of dry soil. Aust. J. Soil Res. 1, 266–275.

134 L.J. Munkholm et al. / Soil & Tillage Research 64 (2002) 125–135

McAfee, M., Lindstrom, J., Johansson, W., 1989. Effects of pre-

sowing compaction on soil physical properties, soil atmosphere

and growth of oats on a clay soil. J. Soil Sci. 40, 707–717.

Munkholm, L.J., Kay, B.D., 2002. Effect of water regime on

aggregate tensile strength, rupture energy and friability. Soil

Sci. Soc. Am. J., in press.

Nearing, M.A., West, L.T., Bradford, J.M., 1988. Consolidation of

an unsaturated illitic clay soil. Soil Sci. Soc. Am. J. 52, 929–934.

O’Sullivan, M.F., Robertson, E.A.G., Henshall, J.K., 1999. Shear

effects on gas transport in soil. Soil Till. Res. 50, 73–83.

Patterson, D.E., Chamen, W.C.T., Richardson, C.D., 1980. Long-

term experiments with tillage systems to improve the economy

of cultivation for cereals. J. Agric. Eng. Res. 25, 1–35.

Payne, P.C.J., Fountaine, E.R., 1952. A field method of measuring

the shear strength of soils. J. Soil Sci. 3, 136–144.

Rasmussen, K.J., 1985. Jordpakning ved forskellig belastning (Soil

compaction with different surface pressures). Tidsskr. Plan-

teavl. 89, 31–45.

Schjønning, P., Elmholt, S.E., Munkholm, L.J., Debosz, K., 2001.

Soil quality aspects of humid sandy loams as influenced by

organic and conventional long-term management. Agric.

Ecosyst. Environ., in press.

Snyder, V.A., Miller, R.D., 1989. Soil deformation and fracture

under tensile forces. In: Larson, W.E., Blake, G.R., Allmaras,

R.R., Voorhees, W.B., Gupta, S.C. (Eds.), Mechanics and

Related Processes in Structured Agricultural Soils. NATO ASI

Series E, Vol. 172. Kluwer Academic Publishers, Dordrecht,

Netherlands, pp. 23–35.

Utomo, W.H., Dexter, A.R., 1981. Soil friability. J. Soil Sci. 32,

203–213.

L.J. Munkholm et al. / Soil & Tillage Research 64 (2002) 125–135 135