Soil and Environmental Analysis: Physical Methods - Chapter 10 docx

This equation sumes that the soil is homogeneous and isotropic, that the frictional resistancebetween the penetrometer shaft and the soil is negligible, that the cone angle ofthe penetro

Penetrometer Techniques in Relation

to Soil Compaction and Root Growth

A Glyn Bengough

Scottish Crop Research Institute, Dundee, Scotland

Donald J Campbell and Michael F O’Sullivan

Scottish Agricultural College, Edinburgh, Scotland

I INTRODUCTION

Soil hardness is the resistance of the soil to deformation, be it by a plant root, theblade of a plow, or the tip of a penetrometer Hard soils are a major problem inagriculture worldwide; they restrict root growth and seedling emergence, increasethe energy costs of tillage, and impose restrictions on the soil management re-gimes that can be used

Penetrometers are used commonly to measure soil strength If a standardprobe and testing procedure is used, penetrometers give an empirical measure ofsoil strength that enables comparisons between different soils A penetrometerconsists typically of a cylindrical shaft with a conical tip at one end, and a devicefor measuring force at the other (Fig 1) Penetration resistance is the force re-quired to push the cone into the soil divided by the cross-sectional area of its base(i.e., a pressure) The American Association of Agricultural Engineers specified

a standard penetrometer design that gives a measurement called the cone index(ASAE, 1969) This standard has been adopted widely, but many nonstandardpenetrometers are in use Nonstandard penetrometers and testing procedures aremore appropriate for some applications, as long as comparisons are made usingthe same procedure The principles behind the testing procedure must be under-stood so that the results can be interpreted sensibly

In this chapter we describe the theory behind the measurement of tion resistance, and how penetration resistance is related to other soil properties

penetra-We then consider the practical aspects of penetrometer measurements, includingthe design of the apparatus, the availability of equipment, the measurement pro-cedure, and the interpretation of data In the final section we discuss how to applythe technique to studies of trafficability, tillage, compaction, and root growth.

II THEORY

A Soil Penetration by Cones

Penetration resistance can, in principle, be estimated from the bulk mechanicalproperties of the soil Farrell and Greacen (1966) developed a model of soil pene-tration in which penetration resistance consisted of two components: the pressurerequired to expand a cavity in the soil, and the frictional resistance to the probe

Penetrometer resistance, Q, is given by Eq 1 (Farrell and Greacen, 1966),

includ-ing the effects of adhesion (Bengough, 1992):

Q ⫽ s(1 ⫹ cot a tan d) ⫹ c cot aa (1)wheres is the stress normal to the cone surface, a is the cone semiangle, d is the

angle of soil–metal friction, and cais the soil–metal adhesion This equation sumes that the soil is homogeneous and isotropic, that the frictional resistancebetween the penetrometer shaft and the soil is negligible, that the cone angle ofthe penetrometer is sufficiently small so that no soil-body accumulates in front ofthe cone, and that the stress is distributed uniformly on the cone surface

as-The normal stress, s, was equated with the pressure required to expand

a cylindrical or spherical cavity in the soil Expansion of the cavity occurred

Fig 1 Schematic diagram of a penetrometer showing cone, shaft, and force transducer

through compression of the soil surrounding the probe Two distinct zones wereidentified: a zone of compression with plastic failure surrounding the probe, with

a zone of elastic compression immediately outside it (Farrell and Greacen, 1966).Calculatings required measurements of many soil mechanical properties Thevalue ofs was predicted for three soils at different bulk densities and matric po-tentials For cylindrical soil deformation,s was only 0.25 – 0.45 of that for spheri-cal deformation Greacen et al (1968) suggested that roots and penetrometerswith narrow cone angles cause cylindrical soil deformation, while penetrometerswith larger cone angles cause spherical deformation

The detailed measurements and calculations required to predicts show that

it is much easier to measure penetration resistance than to predict it One of themajor findings of this work was the large contribution of friction to penetrationresistance Friction on a 5⬚ semiangle probe accounts for more than 80% of thetotal penetration resistance (Eq 1) This has been tested using a penetrometer with

a rotating tip (Bengough et al., 1991, 1997) Rotation of the penetrometer tipdecreased the resultant component of friction directed along the penetrometershaft The measured penetration resistance agreed closely with the predicted resis-tance in a range of soils

When the cone angle exceeds 90⬚ ⫺ f, where f is the angle of internalfriction of the soil, a cone of soil builds up on the probe tip (Koolen and Kuipers,1983) This body of soil moves with the probe, so that friction occurs between thesoil body and the surrounding soil, instead of between the metal and soil surfaces.Equation 1 can therefore be applied only to probes with relatively narrow coneangles Penetrometer design, testing procedure, and the effects on penetration re-sistance are considered in Sec III

B Effects of Soil Properties on Penetration Resistance

Penetration resistance depends on soil type —the distribution of particle sizes andshapes, the clay mineralogy, the amorphous oxide content, the organic matter con-tent, and the chemistry of the soil solution (Gerard, 1965; Byrd and Cassel, 1980;Stitt et al., 1982; Horn, 1984) Within a given soil type, the penetration resistancedepends on the bulk density, water content, and structure of the soil Penetrationresistance can be affected by the pretreatment of the soil prior to testing Hencethe penetration of samples that have been dried, sieved, rewetted, and remoldedwill probably be very different from the penetration resistance of the soil in thefield The purpose of the experiment must therefore be considered carefully beforethe soil is sampled or penetration resistance is measured

Penetration resistance decreases with increasing soil water content, and itincreases with increasing bulk density Gravimetric water content is a useful mea-sure of water status, as matric potential and volumetric water content may change

as soil is compressed during penetration (Koolen and Kuipers, 1983) Matric

potential, however, is the mechanistic link to effective stress and hence to soilstrength, via the surface tension of water-films holding the soil particles together(Marshall et al., 1996) Water content has little effect on cone resistance in loosesoil, but its effect increases with bulk density The influence of bulk density oncone resistance is greater in dry than in wet soil Different functions have beenproposed to describe these relations (Perumpral, 1983) For a given soil, the sim-plest suitable function is

Q ⫽ k ⫹ k u ⫹ k r ⫹ k ru1 2 m 3 4 m (2)whereumis gravimetric water content,r is dry bulk density, and k1 k4areempirical constants (Ehlers et al., 1983) This relation is applicable widely and isillustrated in Fig 2, using values of the constants for a loess soil In some soils,however, the changes in cone resistance with bulk density and water content arenot linear: cone resistance changes most rapidly at high bulk densities and lowwater contents The linear model (Eq 2) may still be appropriate if the ranges ofbulk density and water content are small or soil variability is high, but other mod-els may be valid more generally (Perumpral, 1983)

Fig 2 Variation of penetrometer resistance with water content at different bulk densities.(Based on data from Ehlers et al., 1983.)

The relation between soil strength (in this case measured as penetrationresistance) and matric potential is known as the soil strength characteristic Themain problem in deriving and applying such empirical relations is that soilstrength changes with time, even if bulk density and water content remain constant(Davies, 1985) Soil management practices affect soil structure, changing the con-stants in these empirical relations.

At constant water content and bulk density, cone resistance tends to increasewith decreasing particle size (Ball and O’Sullivan, 1982; Horn, 1984) Thus a claywill have a larger penetration resistance for a given gravimetric water content than

a sand This is due to the greater effective stress associated with the lower matricpotential in the finer textured soil In general, the decrease in organic matter as-sociated with the intensive cultivation or deforestation of soils is associated with

an increase in the gradient of the soil strength characteristic (Mullins et al., 1987)

III PENETROMETER DESIGN

Details of a selection of commercially available penetrometers are given inTable 1 Penetrometers can be classified broadly as ‘‘needle’’ type if they have

a diameter smaller than about 5 mm Most needle penetrometers are used for ing of soils in the laboratory, though some have been used in the field Penetrom-eters that are used in the field often have a diameter greater than 10 mm Manypenetrometers have also been designed for specific purposes Needle penetrometermeasurements can be made in the laboratory by attaching a suitable probe to theforce transducer of a loading frame designed for material testing In the followingsections, the effects of penetrometer design and testing procedure on penetrationresistance measurements are considered

test-A Cone Angle and Surface Properties

Penetrometer tips are generally cones, although flat-ended cylinders (Groenevelt

et al., 1984) and shapes resembling the tips of plant roots (Eavis, 1967) have beenused The shape of the tip determines both the mode of soil deformation and theamount of frictional resistance on the tip Penetrometer resistance is a minimum

at a cone angle of 30⬚ (Fig 3; Gill, 1968; Voorhees et al., 1975; Koolen and drager, 1984) Increased cone resistance is associated at small cone angles withthe increased component of soil–metal friction and, at large cone angles, with soilcompaction in front of the cone (Gill, 1968; Mulqueen et al., 1977).Figure 3,which was derived from measurements made in 67 agricultural fields (Koolen andVaandrager, 1984) shows the relationship between cone resistance and cone anglefor a fixed cone base area Soil tends to be displaced laterally at small cone angles,whereas the direction of displacement becomes more vertical with increasing coneangles (Gill, 1968; Tollner and Verma, 1984) Lateral soil displacement relatesmore closely to the mechanics of root growth than does the more axial displace-

Vaan-ment produced by probes with larger cone angles (Greacen et al., 1968) versely, the load-bearing characteristics of the soil are more closely related to theresistance encountered by larger cone angles Penetrometers that are availablecommercially are generally fitted with 30⬚ or 60⬚ cones, but these can be easilyinterchanged.

Con-The surface roughness of the cone is not an important factor in penetrometerdesign, as abrasion by soil particles quickly removes any minor irregularities Lu-brication of the cone decreases penetration resistance by decreasing soil– conefriction and the movement of soil in the axial direction (Gill, 1968; Tollner andVerma, 1984) Use of such a lubricated penetrometer is of questionable advantage,

as the mechanics of penetration of a lubricated cone is poorly understood, and thelubricating technology may be difficult to standardize

Table 1 Suppliers of Some Penetrometers, Force Transducers, and Load Frames

Available Commercially

Approximatecost (US$)ELE Inter-

national Ltd

In the UK:

Eastman Way, HemelHempstead, Hertfordshire,HP2 7HB

In the USA:

86 Albrecht Drive,P.O Box 8004, Lake Bluff,Illinois 60044-8004

Field penetrometer withdata logger, hand-held

pene-1000

Ametek Mansfield & Green Division,

8600 Somerset Drive,Largo, Fl 34643, USA

Wide range of loadingframes and force trans-ducers Agents also inUK

Pioden

Con-trols Ltd

Graham Bell House, RoperClose, Roper Road, Canter-bury, Kent CT2 7EP, UK

Force transducers suitableranges for needlepenetrometers

Force transducers suitableranges for needlepenetrometers

From about 225

Inclusion in this list does not constitute any recommendation of the product.

B Cone Base Diameter

In general, the diameter of needle penetrometers is important and must be takeninto account when comparing results from different instruments Diameter is lessimportant when comparing field penetrometers

The diameter of the cone bases range from large field penetrometers(⬎10 mm) (Ehlers et al., 1983) to small needle penetrometers (⬍0.2 mm)(Groenevelt et al., 1984) Although cone resistance is expressed as a force per unitbase area, it tends to increase with decreasing base area (Freitag, 1968) For fieldpenetrometers, the standard of the American Society of Agricultural Engineers(ASAE, 1969) allows cone base areas of 320 mm2and 130 mm2, both with a 30⬚cone angle A 3% decrease in diameter is allowed for cone wear In Europe, cones

of 100 mm2base area are common, but cones with base areas of up to 500 mm2have been used

Even in homogeneous soil, penetration resistance can depend on probe ameter as soil particles of finite size must be displaced Diameter dependence is

di-Fig 3 Variation of penetrometer resistance with cone angle for a fixed cone base area

(From Koolen and Vaandrager, 1984 Reproduced with permission from the Journal of Agricultural Engineering Research.)

most noticeable for very small probes, which may have to displace particles ofcomparable size The effect of probe diameter on penetration resistance depends

on the soil type, water content, and structure (Whiteley and Dexter, 1981) Inremolded soil cores with textures ranging from clay to sand, resistance to a 1 mmprobe was typically 45 –55% greater than to a 2 mm diameter probe (Whiteleyand Dexter, 1981) Other studies found no significant effect of diameter among 1,

2, and 3 mm diameter probes in remolded sandy loam (Barley et al., 1965), tween 3.8 and 5.1 mm probes in undisturbed cores (Bradford, 1980), and between

be-1 and 2 mm probes in both undisturbed clods and remolded soils (Whiteley andDexter, 1981) There is need for a comprehensive study over a wide range ofpenetrometer diameters and soil textures

In soils with well-developed structural units, the mechanism of penetrationmay differ between cones of different sizes A cone with a small diameter, relative

to the size of structural units, may penetrate aggregates or planes of weaknessbetween aggregates, whereas a large cone will tend to deform aggregates (Jamie-son et al., 1988)

C Shaft Diameter

The surface area of a penetrometer shaft is directly proportional to its diameter,whereas the force on the penetrometer tip is proportional to the square of the tipdiameter Thus shaft friction is relatively more important for smaller probes, andthis has been confirmed by experiment (Barley et al., 1965) To decrease soil–metal shaft friction, a relieved shaft (i.e., a shaft with a diameter 20% smaller thanthe probe tip) is used commonly

Shaft friction can significantly increase the resistance even to a standardASAE penetrometer, especially in wet clay (Freitag, 1968; Mulqueen et al., 1977).Freitag (1968) found that increasing the shaft diameter from 9.5 mm to 15.9 mm(the ASAE standard) increased the resistance threefold at 0.3 m depth on a stan-dard 20.3 mm diameter cone Similarly, Reece and Peca (1981) used a shaft 8 mm

in diameter to eliminate the clay –shaft friction on the standard 20.3 mm ter cone

diame-IV PENETROMETER INSERTION AND MEASUREMENT

A Force Measurement

The commonest and most easily interpreted penetrometer results are from suring the resistance to a probe driven into soil at a constant speed Other designsmeasure the magnitude or the rate of probe penetration under different constantloads (van Wijk, 1980) In this chapter only penetrometers designed to be used at

mea-a constmea-ant rmea-ate mea-are considered

1 Laboratory Needle Penetrometers

To obtain a constant rate of penetration in the laboratory, it is necessary either todrive the probe downward into the soil with some sort of motor (Barley et al.,1965) or to raise the soil sample on a moving platform toward a stationary probe(Eavis, 1967) The movable crosshead of a strength testing machine has a conve-nient drive capable of a wide range of speeds, and can accept force transducers tomeasure the force resisting penetration (Fig 4; Callebaut et al., 1985; Bengough

et al., 1991) Proving rings, strain gauges, and electronic balances have all beenused to measure the force resisting penetration (Barley et al., 1965; Eavis, 1967;

Fig 4 Needle penetrometer attached to a force transducer on a loading frame

Misra et al., 1986a) The advantage of an electronic balance or force transducer isthat the output can be logged using the analog-to-digital converter of a datalogger

or personal computer Proving rings that are too flexible can result in small voidsgoing undetected, as the proving ring expands when unloaded

2 Field Penetrometers

A field penetrometer may be mounted on a rack to allow easy and precise location(Soane, 1973; Billot, 1982) This facilitates measurements on a regular, closelyspaced grid Hand-held penetrometers are more portable, are cheaper, and can beused in inaccessible field sites (Fig 5)

Automatic logging of force is very advantageous, as it is difficult for theoperator to record measurements at predefined depths Analog recording using a

Fig 5 Field penetrometer with data storage unit

chart recorder records even rapid changes with depth However, the graphical put must then be digitized for statistical analysis, which can be laborious.Digital recording has the disadvantage that maxima and minima may be not

out-be identified This loss of information can out-be important when depth incrementsare large, especially if cone resistance changes abruptly with depth or if the depth

of a cultivation pan varies between penetrations Averaging data at predetermineddepths can disguise such features

B Rate of Penetration

1 Laboratory Needle Penetrometers

Needle penetrometers are used most commonly to estimate the penetration tance of the soil to roots Roots elongate typically at a rate of 1 mm/h or less,which is an inconveniently slow rate at which to conduct penetrometer tests Mostneedle penetrometer measurements are performed at rates of penetration betweenone and three orders of magnitude faster than root growth rates (Whiteley et al.,1981) Eavis (1967) found no effect of rate of penetration on the penetrometerresistance of a silty clay loam at rates between 5 and 0.1 mm/min At slower rates

resis-of penetration, however, the resistance decreased, but only by 13% at a penetrationrate 20 times slower A small decrease in the penetrometer resistance of sandyloam and clay was noted at rates below 0.02 mm/min (Voorhees et al., 1975) Insaturated clay, penetrometer resistance increases with penetration rate because wa-ter must be displaced as the probe compresses the soil (Barley et al., 1965) Insuch a saturated system, the penetration resistance depends on the saturated hy-draulic conductivity in the soil surrounding the probe Penetrometer resistance isrelatively weakly dependent on penetration rate in unsaturated sandy soils at typi-cal rates of testing Given the large difference in penetration rate between rootsand penetrometers, it is still an important factor that must be evaluated if estimat-ing the penetration resistance to roots

2 Field Penetrometers

Increasing penetration speed increases cone resistance in fine-textured soils(Freitag, 1968), in which strength depends on strain rate (Yong et al., 1972) Inmost soils, however, cone resistance is relatively insensitive to penetration ratewithin the range expected from operators of manual penetrometers aiming for theASAE standard rate of 30 mm/min (Carter, 1967; van Wijk and Beuving, 1978;Anderson et al., 1980) The constant penetration rate possible with mechanicallydriven penetrometers is not a significant advantage Exceptions are saturated clay(Turnage, 1973) and soils with a strong layer overlying a weak layer The largeforce required to penetrate the strong layer may cause an excessive penetrationrate in the underlying layer

C Variability

Penetration resistance readings can be very variable, even when penetrations aremade close together (O’Sullivan et al., 1987) The coefficient of variation is typi-cally between 20 and 50%, though it may be more than 70% near the soil surface(Voorhees et al., 1978; Cassel and Nelson, 1979; Gerrard, 1982; Kogure et al.,1985) Small cones give more variable results than large cones (Bradford, 1980).The resistance readings may have a skewed distribution, so that a logarithmic(McIntyre and Tanner, 1959; Cassel and Nelson, 1979) or square root (Mitchell

et al., 1979) transformation is necessary to normalize the data Data at individualdepths may be normally distributed (Cassel and Nelson, 1979; Gerrard, 1982;O’Sullivan and Ball, 1982), but a logarithmic transformation may be necessary ifdepth is included as a factor in analyzing results

The number of measurements, N, required can be predicted using the

relation assumes that the data is normally distributed and is illustrated inFig 6

for values of CV that represent the normal range encountered A fourfold increase

in the number of replicates is required to double the expected degree of precision.The ASAE recommends seven measurements, giving a 95% confidence intervalbetween about 15 and 38% of the mean This is a very large error compared withthe maximum 5% error they allow for cone wear, though such wear is a source ofsystematic error (ASAE, 1969)

Our estimates of the number of penetrations required assume that all surements are independent O’Sullivan et al (1987) found that measurements mademore than about 1 m apart were independent, but Moolman and Van Huyssteen(1989) found evidence of spatial dependence that extended to about 9 m

mea-The penetrometer is ideal for investigating the uniformity of a site becausethe measurements can be made cheaply, quickly, and easily Furthermore, cone re-sistance is related to many other soil properties Hartge et al (1985) used the pene-trometer to identify areas within a field experiment for more detailed investigation.Schrey (1991) showed that cone resistance data could be used to identify areas ofshallow or compact soil or plow pans

D Problems in Use

1 Laboratory Needle Penetrometers

Most penetrometers designed for small cones are unsuitable for field use ford, 1980) Large field penetrometers have been used successfully in root growth

(Brad-studies (Ehlers et al., 1983; Barraclough and Weir, 1988; Jamieson et al., 1988),but these are very different from growing roots, in terms of diameter and penetra-tion rate.

Care must be taken, when sampling soils for needle-penetrometer ments, that the soil is compressed as little as possible during coring Soil is com-pacted if cores are sampled too close together, or if soil is trampled by the field-worker Such compaction increases the penetrometer resistance

measure-Lateral confinement of the soil core may increase penetrometer resistance

if the core diameter is less than about 20 times that of the probe (Greacen et al.,1969) Tensile failure of the core may occur if the core is unconfined laterally,decreasing the penetrometer resistance as the core cracks Penetrometer resistancemay also be affected if more than one penetration is performed on each core —cracks of tensile failure may form between the penetration holes (Greacen et al.,1969) though, under other circumstances, penetration resistance could be in-creased by compaction around the neighboring penetration hole

Stones cause rapid increases in penetration resistance that can damage sitive force transducers Overload cutoffs should be included, if possible, to pro-

sen-Fig 6 Variation of the 95% confidence interval about the mean with the number of coneresistance observations, for two coefficients of variation