Xray Computed Tomography

European Journal of Soil Science, June 2013, 64, 279–297 doi: 10.1111/ejss.
12028
Review
Applications of X-ray computed tomography

for examining biophysical interactions and structural
development in soil systems: a review
J . R . H e l l i w e l la , C . J . S t u r r o c ka , K . M . G r a y l i n ga , S . R . T r a c ya , R . J . F l a v e lb , I . M . Y o u n gb ,
W . R . W h a l l e yc & S . J . M o o n e ya
a
Division of Agricultural and Environmental Sciences, University of Nottingham, Sutton Bonington Campus, Leicestershire LE12 5RD,
UK, b School of Environmental and Rural Sciences, University of New England, Armidale, New South Wales 2351, Australia, and
c Sustainable Soils and Grassland Systems, Rothamsted Research, West Common, Harpenden AL5 2JQ, UK
Summary
Soil systems are characterized by the spatial and temporal distribution of organic and mineral particles,
water and air within a soil profile. Investigations into the complex interactions between soil constituents
have greatly benefited from the advent of non-invasive techniques for structural analysis. In this paper we
present a review of the application of one such technique, X-ray computed tomography (CT), for studies of
undisturbed soil systems, focusing on research during the last 10 years in particular. The ability to undertake
three-dimensional imaging has provided valuable insights regarding the quantitative assessment of soil features,
in a way previously unachievable because of the opaque nature of soil. A dynamic approach to the evaluation
of soil pore networks, hydro-physical characteristics and soil faunal behaviour has seen numerous scanning
methodologies employed and a diverse range of image analysis protocols used. This has shed light on functional
processes across multiple scales whilst also bringing its own challenges. In particular, much work has been
carried out to link a soil’s porous architecture with hydraulic function, although new technical improvements
allowing the characterization of organic matter and the influence of soil biota on structural development are
showing great promise. Here we summarize the development of X-ray CT in soil science, highlight the major
issues relating to its use, outline some of the applications for overcoming these challenges and describe the
potential of future technological advances for non-invasive soil characterization through integration with other
complementary techniques.
Introduction ordered in a complex and dynamic hierarchical manner, advances

in understanding of soil behaviour will only come with in-
The physical structure of soil constituents is crucial to its
depth spatial evaluations of its constituent configurations in three
functioning, with quantification of internal soil architecture being
dimensions (3D) and quantification of the underlying processes
key to understanding the processes governing its development.
that cause changes over time.
The diverse and dynamic array of mineral particles, air- and water-
filled pores and organic material over a range of scales make Over the last 30 years, X-ray computed tomography (CT) has
studies of soil a challenging task, and it is for this reason it is often provided a non-destructive means for visualizing and quantifying
described as ‘the most complex biomaterial on the planet’ (Young soils in 3D, allowing researchers to gain fundamental insights
& Crawford, 2004). As a result, a detailed understanding regarding regarding soil function. X-ray CT has been applied in many
the interactions between soil structure formation, stabilization different studies concerning the spatial distribution of soil
and degradation processes still remains incomplete. As soil is properties (Heeraman et al ., 1997; Young et al ., 2001; Pierret
et al ., 2002; Rogasik et al ., 2003; Nunan et al ., 2006), with
Correspondence: S. J. Mooney. E-mail: [email protected] applications evolving to cover the extensive characterization of
Received 23 April 2012; revised version accepted 17 December 2012 pore-space geometry and fractures within a soil sample with
© 2013 The Authors

Journal compilation © 2013 British Society of Soil Science 279
280 J. R. Helliwell et al.
respect to different variables such as bulk density (Petrovic for studies of the physical soil environment. This work catalysed
et al ., 1982; Anderson et al ., 1990), layer detection (Macedo future research interests and subsequent pioneering work by
et al ., 1998; Lipiec & Hatano, 2003), spatial correlation and Hainsworth & Aylmore (1983) investigated root-related hydrology
tortuosity (Perret et al ., 1999), porosity (Grevers et al ., 1989; in relation to water absorption properties at a spatial resolution
Heijs et al ., 1995), pore network structures (Al-Raoush & Willson, of 2 × 2 × 5 mm. The comparison of results obtained from a
2005), permeability (Ketcham & Carlson, 2001; Mooney, 2002), commercially-produced medical X-ray system with a conventional
volumetric water content (Crestana et al ., 1985; Hopmans et al ., gamma ray scanner indicated, for the first time, the suitability
1992), solute breakthrough (Clausnitzer & Hopmans, 2000), of X-ray CT for soil-plant studies, with a clear application for
ice-lens formation (Torrance et al ., 2008), seedbed preparation resolving spatial changes in soil water content with time in the
(Atkinson et al ., 2007) and calculated fractal properties (Zeng presence of a single plant root. Shortly afterwards, Crestana et al .
et al ., 1996; Gantzer & Anderson, 2002; De Gryze et al ., 2006; (1985) demonstrated the linear dependence of X-ray attenuation
Papadopoulos et al ., 2008). and soil water content, enabling them to obtain ‘reproducible
Recent advances in tomographic imaging are now providing and quantitative information on 3D space and time of water in
researchers with the tools for further understanding of dynamic soil’. Flannery et al . (1987) explored the advantages of using
pore-scale processes in soil systems. By visualizing samples in synchrotron-based X-ray microtomography as a means of spatially
3D and crucially relating results to complementary laboratory resolving elemental maps and exploring dynamic flow profiles, at
measurements, this will help soil scientists to unravel the processes a resolution of 2.8 μm. Such pioneering studies demand imaging
of pore formation and development with time. The aim of this across multiple scales and clearly demonstrated the validity of
review is to focus on work (in the last 10 years in particular) X-ray CT for studies of undisturbed soil systems.
that has used X-ray CT to visualize and quantify soil physical
characteristics, and to illustrate the further technological advances
Basic principles
that will potentially revolutionize spatio-temporal studies of the
inner space of soil systems. Here we provide only a brief general description of the principles
and processes of X-ray CT. For a more complete technical
description of the theory behind X-ray CT we refer the reader to
Overview of X-ray computed tomography (CT)
the comprehensive accounts provided by Clausnitzer & Hopmans
X-ray CT has demonstrated potential for examining the undis- (2000), Ketcham & Carlson (2001), Wildenschild et al . (2002),
turbed internal structure of soils. It is less affected by paramagnetic Taina et al . (2008), Pires et al . (2010) and Mooney et al . (2012).
elements in soil than other imaging techniques, such as magnetic X-ray CT is a non-invasive technique that can be used to
resonance related approaches (including nuclear magnetic reso- visualize the interior of objects in 2D and 3D based on the
nance (NMR) and magnetic resonance imaging (MRI)) (Macedo principle of attenuation of an electromagnetic wave. A typical scan
et al ., 1998). Previous limitations of image resolution and a poor involves the collection of a series of radiograph images of a sample
ability to distinguish between different phases of media are rapidly acquired at incremental angular positions, normally over 360◦ .
being overcome with further technological development. It is now Generally, X-ray CT scanners consist of three common parts: an
possible to detect sub-micrometre changes in soil aggregation, X-ray source (either synchrotron light or a conventional X-ray
root structures and pore space (both air and water filled), allow- tube), a sample manipulation stage and a detector. X-rays emitted
ing for the direct measurement of soil structural changes both from the source pass through the sample and are progressively
spatially and temporally through soil development cycles (Lang- attenuated by absorption and scattering as the object itself becomes
maack et al ., 2002; Pierret et al ., 2003; Feeney et al ., 2006; Otten a secondary source of X-rays and electrons through atomic
et al ., 2010; Crawford et al ., 2011). interactions (Mooney et al ., 2012). The characteristic of a material
to either absorb or scatter a photon is called the attenuation
coefficient. Attenuation coefficients are related to the density of
History
the absorbing material, electron density of the voxel of interest
Initially developed as a tool for diagnostic medicine because of its and incident X-ray energy, but are predominantly controlled by
capacity to produce non-destructive, high-contrast cross-sectional four dominant processes: (i) photoelectric absorption, (ii) Rayleigh
images quickly, the earliest X-ray CT scans comprised only five scattering, (iii) Compton scattering and (iv) pair production.
image slices taking 35 minutes to acquire, and were optimized However, it should be noted that pair production occurs at X-
for soft tissue and bone visualization (Hounsfield, 1973). Within ray energies (1.022 MeV) far greater than conventional X-ray CT
a decade of its advent, CT had been adapted and extended systems. The reader is referred to Duliu (1999), Carlson et al .
for use in a wide variety of academic disciplines, including (1999), Ketcham & Carlson (2001) and Seibert & Boone (2005)
soil science. Petrovic et al . (1982) were the first to apply X- for detailed descriptions of the interaction of photons with matter.
ray CT in a geological study of soil bulk density, where the Integration of the linear X-ray attenuation coefficient values
linear relationship observed between bulk density and the X-ray from the multitude of radiographic images form the basis of
attenuation coefficient demonstrated the technological potential tomographic reconstruction (Taina et al ., 2008). It is based
© 2013 The Authors

Journal compilation © 2013 British Society of Soil Science, European Journal of Soil Science, 64, 279–297
X-ray computed tomography for soil physical properties 281
frequently on mathematical filtered back-projection algorithms, and the acquisition of low-noise data with few artefacts make
through which cross-sectional 2D image slices are generated from synchrotron light tomography an excellent way of visualizing soil
radiograph projection images (extensively covered in reviews by structure, as demonstrated by Lehmann et al . (2006) who pre-
Wildenschild et al . (2002) and Stock (2008)). Each non-invasive dicted fluid phase distributions in sand. However, restricted access
tomographic ‘slice’ consists of discrete units known as voxels (3D to synchrotron beamlines have limited fully replicated studies of
pixels), the size of which reflects the spatial resolution of the scan. soils, whilst the required sample size is also small. For further
details concerning synchrotron X-ray CT applications, we refer
the reader to Spanne et al . (1994).
Typical X-ray CT systems
X-ray CT systems can generally be grouped into three categories,
Image segmentation
namely industrial, medical and synchrotron (Figure 1). Recently,
Vaz et al . (2011) discussed the current trend in types of scanners Discrete material phases within X-ray CT data can be differenti-
used in soil science. Although much of the early pioneering work ated by image segmentation techniques, which classify individual
was performed on medical or custom-designed systems, there has voxels with a common mean grey scale value. The acquired
recently been a significant increase in the use of industrial X-ray images from soil-based X-ray CT scans frequently have three main
CT systems as scanner costs have reduced. phases: mineral grains, pore space (both water and air-filled) and
Industrial and medical scanners produce a spectrum of electron organic materials such as roots. The grey values in the images are
frequencies, yielding a polychromatic beam. Because of the ordered according to attenuation density, so that mineral grains as
nature of the requirements of X-ray CT systems for medical the densest are represented by bright voxels (highly attenuating),
research with optimized tissue differentiation and minimized and air-filled pore spaces, the least dense, have darker voxels (low
patient exposure times, a limited dose of low-energy X-rays attenuation). Two approaches have generally been used to seg-
(typically <125 kV) is coupled with a large (mm scale), high- ment materials of interest in soils: ‘global thresholding’ based on
efficiency detector (Ketcham & Carlson, 2001). The scale of an estimate from the histogram of the image, or ‘localized adap-
observation in medical research is typically metres, resulting tive or dynamic thresholding’, where a different threshold value
in typical scanning resolutions of millimetre scales. Scanner is applied to different regions in an image, and thus accounts for
methodologies for medical systems are the same in principle the spatial phase dependency (Iassonov et al ., 2009; Wang et al .,
as industrial scanners but the source and detector rotate around 2011).
a static object (Figure 1b), whereas in industrial scanners the
objects typically rotate in a cone beam of polychromatic X-rays
Limitations to X-ray CT
(Figure 1a). As the entirety of the object passes through the
beam with each angular rotation, a reduced focal spot size allows As with all analytical equipment X-ray CT has limitations, which
smaller objects to be moved nearer to the X-ray source and into at present limit its capacity to characterize certain soil properties.
a narrower section of the X-ray beam. This permits increased Whilst X-ray CT is capable of resolving features at the microscale,
resolution through improved detector utilization, to image smaller complications in phase isolation can be problematic if scanning
sub-sections of the object in a single view (Ketcham & Carlson, artefacts are present within the data. These artificial structures,
2001). Hence industrial scanners can span a range of scales such as rings and streaks, can obscure details of interest or cause
from metre-wide scales of observation for large objects such as changes in the grey value of a material depending on spatial
geological rock forms to large (≤1 μm) resolutions when resolving location within the sample. This is known as beam hardening,
soil pores in a single micro-aggregate, for example. In medical X- which results in an increase in mean X-ray beam energy as it
ray CT, the resolution and image quality is somewhat limited, passes through a sample. The increase results from the greater
so from a soil perspective, if research objectives are to resolve attenuation of lower-energy photons compared with higher-energy
individual soil phases and classify individual microscopic pore- photons, giving a beam with reduced overall intensity but greater
characteristics, an industrial research X-ray or synchrotron system mean energy. The effect, which is typically shown at the outer
is preferable (Kinney & Nichols, 1992). edges of an object appearing brighter (larger grey value) than
Synchrotron scanners use electromagnetic radiation emitted by at the centre, can be reduced by filtering the incident X-ray
high-speed electrons spiralling in a magnetic field of a particle beam to remove lower-energy photons using thin sheets of
accelerator (Figure 1c). Synchrotron radiation emission occurs as aluminium, copper or tin. The majority of CT reconstruction
the electrons decelerate during beam focusing by large electro- software packages now offer software-based reduction of common
magnets. As such, synchrotron scanners produce high intensity X-ray CT data artefacts.
(number of photons per second), highly collimated (source conver- Studying the organic fraction of soil can be difficult because of
gence reducing image blur) polychromatic radiation, which can be a poor ability to distinguish between media phases as, depending
split with a monochromator into different wavelengths, giving the on the soil type, the edges of organic and soil (mineral) material
ability to tune the radiation over a wide range (Kinney & Nichols, can have similar attenuation densities (Figure 2). This can cause
1992). The advantages of rapid scan speed (typically <15 minutes) problems in characterizing organic constituents or segmenting root
© 2013 The Authors

Figure 1 Examples of tomography arrangements for different CT systems: (a) industrial scanners, (b) medical systems and (c) synchrotron systems.
Modified from Wildenschild et al . (2002) with permission from Elsevier.
networks satisfactorily (Mooney et al ., 2012), as the identification 1997; Kaestner et al ., 2006; Flavel et al ., 2012). Therefore the
of single objects through image thresholding may be limited by minimum object size detectable accurately is often viewed as twice
individual voxels containing multiple materials. The attenuation (Rogasik et al ., 2003) or three times (Wildenschild et al ., 2002)
of materials within single voxels is proportionally averaged, the scanning resolution.
potentially leading to ‘partial volume effects’ as voxels are To date there has been no widely adopted method for the
misclassified (Clausnitzer & Hopmans, 2000; Ketcham & Carlson, segmentation of roots, organic material, air and water from soils.
2001). Such effects have been a particular problem for root The use of global thresholding and local adaptive thresholding
studies (Kaestner et al ., 2006; Lontoc-Roy et al ., 2006; Perret (Pierret et al ., 1999a; Capowiez et al ., 2003; Gregory et al ., 2003;
et al ., 2007; Flavel et al ., 2012), although enhancing image Lontoc-Roy et al ., 2006; Perret et al ., 2007; Flavel et al ., 2012)
contrast between the organic components and supporting media techniques is susceptible to ‘partial volume effects’, sensitive to
through filter applications and appropriate thresholding can help image quality and, in the case of Flavel et al . (2012), represented
to minimize this problem. Despite some debate in the literature, substantial investment of time. Furthermore, thresholding can
the magnitude of this effect is intensified as the resolution is ‘over-segment’ images because of the presence of water in
increased below the pore size of the soil used (Heeraman et al ., both roots and soil pores, and is consequently often followed
© 2013 The Authors

when performed with different image analysis protocols. The need

for a more standardized approach to X-ray CT data evaluation is
critical if we are to be confident of soil quantification and suitable
comparisons are to be made between individual studies.
Using X-ray computed tomography in soil studies

The classification and separation of soil phases
Destructive techniques and observations in 2D, including thin
sections and electron microscopy, have been used traditionally
to interpret soil structure (Young et al ., 2001). X-ray CT has
been used in many studies to describe the spatial nature of soil
constituents in undisturbed systems, as summarized in Table 1.
However, technological advances in system design have now seen
recent studies using voxel sizes of <15 μm (Feeney et al ., 2006;
Wang et al ., 2012), allowing, for the first time, interpretations
of fractions finer than sand to be made. Further developments
Figure 2 Histograms showing the three phases frequently present in X-
in image resolution and scanning refinements should enable the
ray CT scans. Note the cross-over in greyscale values between the phases,
detection of discrete clay fractions, highlighting the influence of
making segmentation a challenging task.
microscale processes on soil development such as weathering,
nutrient liberation and carbon storage.
by a ‘cleaning’ connectivity or geometric routine to remove Interpretations of the size, distribution and orientation of
misclassified structures and to reconnect continuous compositions. soil fractions are somewhat dependent upon the ability to
However this can result in a loss of ‘fine’ sample features discriminate between isolated soil phases. The attenuation of
such as roots (Kaestner et al ., 2006). Although the majority of X-rays is generally greater at smaller energy levels, and is a
segmentation methods are still based on subjective principles, principle that can be used to discriminate between minerals of
various manual and automated methods have been used to reduce different compositions but of a similar mass (Taina et al ., 2008).
bias. For example, Heeraman et al . (1997) used a manual per However, because of difficulties in phase isolation, we still lack
voxel classification method to segment bush bean roots from a a standardized approach to quantifying soil particles, air-filled
fine sand substrate, and more recently Mairhofer et al . (2012) pore space and water volumes. Another approach, when studying
applied object tracking approaches to segment roots from soil minerals with similar attenuation coefficients but different electron
images automatically. Ketcham (2005) and Sleutel et al . (2008) compositions, is to combine scans of multiple energy levels
used automated Blob3D and Morpho+ software to investigate soil (Ketcham, 2005). In their study on the effects of soil structure on
mineralogy and organic matter, respectively. Peth et al . (2010) functional processes, Rogasik et al . (1999) used a medical scanner
used automatic kriging algorithms in 3dma_rock software to with energy levels of 80 and 120 kV to calculate water, air and
investigate soil fracturing under mechanical loading. phase composition distributions in two silt loam subsoils. At a
At present the diversity of scanning methodologies and image resolution of 0.25 × 0.25 × 1 mm, the weighted mean of derived
segmentation protocols adopted by the research community using soil constituent data agreed with the phase composition determined
X-ray CT is hindering the comparability between different by weighing, indicating that multiple energy X-ray CT may play
intra-disciplinary studies. The majority of algorithms used for a key role in phase separation, although further work in this area
volume reconstructions are currently bespoke and system specific. has been surprisingly sparse.
Furthermore, at the coarse resolutions (hundreds of micrometers)
typically used until recently, the majority of micropore space is
Soil mineralogy
not visualized. This results in poor interpretation of observed
shifts in soil physical properties using X-ray CT, because fine Considerable attention has been given to mineral classification
changes have been misclassified or overlooked entirely. The need by X-ray CT in petrological studies (Norrish & Hutton, 1969;
for rigorous, objective methodologies for image analysis protocols Denison & Carlson, 1997; Carlson et al ., 1999; Kalukin
is urgent, as was recently highlighted by Baveye et al . (2010) in a et al ., 2000; Van Geet et al ., 2000; Ketcham, 2005; Stock,
‘round-robin’ study involving 13 international groups. Variations 2008). However, as the adaptations of mineral classifications in
in porosity estimates of between 13 and 73% for the same 2D geological applications have become more refined, their relevance
X-ray CT image demonstrates the potential for large operator bias to applications in soil mineralogy studies has also become
during the selection of simple thresholding techniques, yielding realized. Orsi et al . (1994) were the first to apply medical X-
dramatically varied and unrepresentative results. Marcelino et al . ray CT to analyse offshore sediment cores non-destructively,
(2007) stressed the dangers in comparing soil porosity values with subsequent investigations able to characterize the sediment
© 2013 The Authors

Table 1 Selected key research papers in which X-ray computed tomography (CT) has been used to characterize soil properties
Reference Measured properties System Scan parameters Voxel size / mm Soil type
Petrovic et al . (1982) Soil bulk density Medical X-ray CT (American 125 kV, 30 mA 0.25 × 0.25 × 2 Sandy loam
Science and Engineering)
© 2013 The Authors

Hainsworth & Aylmore (1983) Volumetric water content Medical X-ray CT (EMI 1007) 120 kV 1.5 × 1.5 × 10 –
Crestana et al . (1985) Volumetric water content Medical X-ray CT (GE 8800) 120 kV – Sandy loam
Grevers et al . (1989) Macroporosity comparison Medical X-ray CT (GE CT 140 kV, 140 mA 0.75 × 0.75 × 1 –
with thin sections 9800)

Anderson et al . (1990) Macropore geometry Medical X-ray CT (Siemens 125 kV, 410 mA 0.7 × 0.7 × 2 Silt loam
Somatom DRH)
Anderson et al . (1992) Solute transport through Medical X-ray CT (Siemens 125 kV, 410 mA 0.5 × 0.5 × 2 Silt loam
aggregates Somatom DRH)
Heijs et al . (1995) Water infiltration and flow X-ray CT 0.27 × 0.27 × 1 Clay soil
patterns
Heeraman et al . (1997) Plant roots Industrial X-ray CT (SMS 420 kV, 3 mA 0.16 × 0.16 × 0.2 Loamy sand
Model 101 B+ CITA)
Perret et al . (1999) Macropore network Medical X-ray CT (ADVENT 120 kV, 50 mA 0.195 × 0.195 × 2 –
geometry and pore HD-200)
connectivity
Pierret et al . (1999b) Tree rooting spatial Medical X-ray CT (GE CT 140 kV, 140 mA 0.4 × .0.4 × 2 Sandy soil
distributions Pace)
Rogasik et al . (1999) Gravimetric water content, Medical X-ray CT (Siemens 80 kV, 190 mA and 120 kV, 0.25 × 0.25 × 1 Silt loam
bulk density, soil phases Somatom Plus) 85 mA
Kalukin et al . (2000) Comparison of mineralogy Industrial desktop X-ray CT 50 kV, 70 kV, 90 kV, 110 kV 0.022 × 0.022 Calcite, gypsum and
with thin sections (Skyscan 1072) and 130 kV barite
Adderley et al . (2001) Pore space, organic material, Industrial high-resolution X-ray 100 kV and 0.5 mA 0.03 × 0.03 × 0.03 Ash midden deposits
mineralogy CT (BIR ACTIS)
Baveye et al . (2002) Bulk density, volumetric Medical X-ray CT (Siemens 80 kV and 120 kV 0.25 × 0.25 × 1 Silt loam
water content, porosity Somatom Plus)
Gantzer & Anderson (2002) Bulk density and macropore Medical X-ray CT (GE 120 kV, 100 mA 0.19 × 0.19 × 1 Silt loam
characteristics Genesis-Zeus)High-resolution 180 kV, 0.133 mA 0.15 × 0.15 × 0.1
X-ray CT (Bio-imaging
research)
Mooney (2002) Macroporosity and water Medical X-ray CT (Picker 120 kV, 100 mA 0.46 × 0.46 × 0.46 Loamy sand, sandy loam,
flow pathways PQ6000) clay loam
Pierret et al . (2002) Macropore characterization Medical X-ray CT (GE CT 140 kV, 130 mA 0.4 × 0.4 × 2 Loamy clay
Pace)
Anderson et al . (2003) Soil porosity and solute Medical X-ray CT (Siemens 2×2×8 Silt loam
breakthrough Somatom DRH)
Gregory et al . (2003) Root system architecture Industrial high-resolution X-ray 50 kV, 0.5 mA 0.1 × 0.1 × 0.1 Clay
CT (Custom built – Jenneson
et al . (2003))
Table 1 Continued.
Reference Measured properties System Scan parameters Voxel size / mm Soil type
Rogasik et al . (2003) Bulk density Medical X-ray CT (Siemens 120 kV and 165 mA 0.25 × 0.25 × 1 Loamy, silty
Somatom Plus)
Al-Raoush & Willson (2005) Algorithm-derived pore Synchrotron high-resolution 33.22 kV 0.0115 × 0.0115 × 0.0115 Sand
network structures, X-ray CT 0.023 × 0.023 × 0.023
porosity and specific
surface area
Ketcham et al . (2005) Mineralogy Industrial high-resolution X-ray 420 kV, 4.7 mA 0.25 × 0.25 × 0.5 Schist sample
CT (Bio-Imaging Research)
Kaestner et al . (2006) Root-system architecture Medical micro-CT (Scanco) 50 kV, 0.114 mA 0.036 × 0.036 × 0.036 Sand
Lontoc-Roy et al . (2006) Root -system architecture X-ray CT 130 kV, 100 mA 0.12 × 0.12 × 0.1
Perret et al . (2007) Root -system architecture X-ray CT 130 kV, 150 mA 0.275 × 0.275 × 0.1 Sand and loam
Sleutel et al . (2008) Microstructure and organic Industrial high-resolution X-ray 60 kV and 120 kV Artificial sand-OM mix
matter CT
Torrance et al . (2008) Structure of frozen soils Industrial micro-CT (GE MS8) 120 kV, 0.19 mA 0.043 × 0.043 × 0.043 Silt loam
during ice lense formation
Hargreaves et al . (2009) Root system architecture Industrial high-resolution X-ray 50 kV, 0.5 mA 0.1 × 0.1 × 0.121 Sandy loam
CT (Custom built – Jenneson
et al . (2003)
Quinton et al . (2009) Pore size, shape and Medical X-ray Micro CT(GE 100 kV, 80 mA 0.0167 × 0.0167 × 0.0167 Peat
distribution. Organic MS8X-130)
matter
Elliot et al . (2010) Pore networks for saturated Industrial X-ray Micro CT(GE 100 kV, 80 mA 0.02 × 0.02 × 0.02 Clay loam, loamy sand
hydraulic conductivity MS8X-130)
modelling
Peth et al . (2010) Localized soil deformation Industrial X-ray 240 kV and 150 kV 0.04 × 0.04 × 0.04 Silt
under shrink-swell cycles microtomography (GE 0.0253 × 0.0253 × 0.0253
and confined compression Phoenix V|tome|x 240 and
GE Phoenix nanotom)
Flavel et al . (2012) Root length Industrial X-ray 100 kV, 0.270 mA 0.68 × 0.68 × 0.68 Sand clay mix
microtomography (GE
Phoenix V|tome|x 240)
Mairhofer et al . (2012) Root-system architecture Industrial X-ray 80–130 kV, 0.09–0.18 mA 0.044 × 0.044 × 0.044 Sand, loamy sand and
microtomography (GE 0.023 × 0.023 × 0.023 clay loam
Phoenix nanotom)
Schmidt et al . (2012) Root:soil contact Industrial X-ray Micro CT 145 kV, 0.14 mA 0.0349 × 0.0349 × 0.0349 Sandy loam and
(Metris X-Tek HMX) vermiculite
Tracy et al . (2012b) Pore and root architecture Industrial X-ray 100 kV, 100 μA 0.0174 × 0.0174 × 0.0174 Sandy loam
microtomography
(SkyScan-1076)
X-ray computed tomography for soil physical properties
Gaps represent unreported data in the literature.
© 2013 The Authors

285
morphology successfully (Orsi & Anderson, 1995). Later, Géraud

et al . (2003) used a medical system to map spatially the 3D
distributions of quartz, feldspar and micas on the basis of their
radiological densities. By using modified Hounsfield units (an
adapted scale of greyscale values, based on the transformation of
the linear attenuation coefficient into one where the radio-density
of water at standard temperature and pressure equals zero), Géraud
et al . (2003) compared the radiological densities of the main
mineral phases with measured values on macroscopic crystals,
confirming mineral presence in the sample. Kalukin et al . (2000)
successfully used principal components analysis (PCA) of images
from X-ray scans to enhance contrast between individual minerals
more clearly than was possible in single-energy X-ray CT scans.
Unfortunately, because of the nature of PCA (with no calibration
of the data to an external standard and no prior knowledge of
the expected transmission curve) no mineral constituents can be
identified specifically, which is a major limitation in using this
approach to map soil mineralogy. However, PCA undoubtedly Figure 3 Three-dimensional representation of a peat sample at −40-cm
has the potential to direct the researcher to specific regions of soil water pressure. Lighter areas represent the soil and liquid phases,
interest in the sample that may be missed in raw single-energy highlighting the difficulties in segmenting organic constituents from the
images. Furthermore, by enhancing object contrast within samples, water phase. Taken from Quinton et al . (2009) with permission from
the accuracy of segmenting multiple materials and feature Elsevier.
identification and statistical measurement may be improved.
the sample with lead (II) nitrate in order to visualize the structure
of Sphagnum stems and branches. Recently, Elyeznasni et al .
Organic constituents
(2012) detected coarse-sized organic matter concentrations in X-
Organic compounds in soil, such as plant roots or decomposing ray CT images, and recovered part of the discontinuous pattern of
residues, are far more difficult to detect from X-ray CT images organic matter fragments in the macroporosity. However, it is still
than soil pores. Less research has thus resulted, with the majority widely acknowledged that as a result of difficulties in accurately
focused on the identification of roots (Tollner et al ., 1994; Lontoc- discriminating between certain soil phases, quantification of
Roy et al ., 2006; Mooney et al ., 2007; Pierret et al ., 2007; Tracy organic matter, and in particular fine-sized assemblages, using
et al ., 2012b). One issue in single threshold X-ray CT images X-ray CT remains a challenging prospect.
is that significant portions of organic phases are misclassified as
part of the mineral phase, and pore space can be misclassified
Quantification of physical soil properties
as being organic material. Sleutel et al . (2008) overcame this
by assigning ‘uncertainty ranges’ to thresholded images of a Characterization of soil porosity and structure. Many ecologi-
sand-organic matter mix, using scan data from four different X- cal functions in soils are influenced by structure, through impacts
ray CT systems, which used different X-ray sources of different on the spatio-temporal distributions of soil particles, water and
energy spectra, detectors of differing sensitivity and beam filters air within soil profiles playing a pivotal role in water movement
of different thicknesses. Through this approach they were able and nutrient displacement. The strong contrast in X-ray attenuation
to discriminate between organic, mineral and pore phases, and between soil pores and related solid material has seen the develop-
concluded that medium energy systems were optimal for phase ment of various methods for the quantification of temporal changes
segmentation because of an increased signal:noise ratio compared in soil pore networks under different environmental stresses (Pires
with low energy scans. et al ., 2007; Peth et al ., 2010). Information on the 3D arrangement
X-ray CT has also been used to study highly organic soils of the soil pore system can be derived from measurements of the
such as peats (Quinton et al ., 2009). A significant challenge actual morphology of soil pores, rather than inferring the soil pore
exists in distinguishing between water and organic matter, as system from flow dynamics and idealized pore diameter measure-
the latter is very porous and so attenuates X-rays in a manner ments (Grevers et al ., 1989). To quantify properties such as pore
that is very similar to water when it is saturated (Figure 3). connectivity and continuous pore networks within the soil matrix
Consequently, at small soil water potentials, the volumetric linear and to use them adequately to model pore networks, the visualiza-
attenuation coefficient was considered negligible because of partial tion must be in 3D (Mooney, 2002). However, because of issues
volume effects. Likewise, Kettridge & Binley (2011) found of scale still associated with X-ray CT imaging, our understanding
similar difficulties in interpreting quantitatively the structural of how soils operate at the microscale has been limited, with most
characteristics of the organic constituents of peat, and so flushed research focused on soil macroporosity (Nunan et al ., 2006). For
© 2013 The Authors

example, Perret et al . (1999) characterized pore tortuosity (the effect of soil compaction on 3D macropore geometry was recently
ratio of total pore length to that of the shortest possible path), characterized by Kim et al . (2010) in undisturbed field cores.
numerical density and hydraulic radius in undisturbed cores at a By using X-ray CT in addition to recording decreases in overall
coarse resolution of 0.195 × 0.195 × 2 mm, whilst Rogasik et al . porosity and bulk density, Kim et al . (2010) were able to reveal the
(2003) assessed macropore size, length and connectivity of pores total macro- and mesopore numbers associated with an increase
under different agricultural management practices at the spatial in surface compaction. Furthermore, they were able to assess
scale of 0.25 × 0.25 × 1 mm. the degree of correlation between additional CT-measured pore
Many studies have used X-ray CT to quantify various properties characteristics, including largest pore area, pore circularity and
of macropore network morphology, including pore network fractal dimensions.
structure (Baveye et al ., 2002; Aravena et al ., 2011), pore Soil management plays a key role in altering soil stability. The
diameter (Anderson et al ., 1992; Zeng et al ., 1996), pore loss of vegetation cover, mismanagement of water applications
circularity (Gantzer & Anderson, 2002) and crack formation (Peth and the pollution of soil systems can all influence soil degradation,
et al ., 2010). Used in this way, X-ray CT provides data not and hence be detrimental to the physical soil environment.
obtainable by other means. Baveye et al . (2002) helped inform Quantification of the impacts of different agricultural tillage
future research by revealing the dependence of macroscopic practices on the mechanical stability of soils and the resulting
soil properties such as bulk density, volumetric water content seedbeds they create has been achieved for intact soil samples
and air content on sampling volume, shape and positioning. By from conventional chisel-disc plough and no-tillage systems
comparing soil characteristics from different types and size of by Gantzer & Anderson (2002), with the chisel-disc system
sampling volume, they were able to show how properties in small resulting in double the number of macropores and a 62 and
volumes can exhibit erratic fluctuations in measurements, which 94% increase in macropore perimeter and circularity, respectively.
can be stabilized as sampling volume increases. Likewise, a recent However, Gantzer & Anderson (2002) acknowledged that whilst
important study by Peth et al . (2010) at a resolution of 25.3 μm measurements such as these appear to be useful, crucially they are
revealed a very heterogeneous microscale soil environment limited in their capacity to describe the behaviour of soil volumes
when examining soil structure, localized soil deformation and as a whole. It is 10 years since this work and one might suggest
local strain analysis over very small spatial scales (<1 mm). critically this thought still holds true, with many studies still failing
Our understanding of these mechanisms at the microscale to fully use the wealth and richness of the datasets that can be
level is crucial to further understanding of the processes obtained by X-ray CT. Future research that uses X-ray CT needs
behind soil stabilization and structural development and support to focus on demonstrating the actual benefit of the 3D datasets
improvements in conceptual model development of geometric over traditional measurements, as opposed to simply comparing
pore networks. Delerue et al . (2003) successfully developed a results of the same property obtained by different means.
pore network directly from soil images by integrating pore size Atkinson et al . (2009) used X-ray CT to describe the
and connectivity parameters, which enabled calculation of the temporal evolution of a seedbed at the mesoscale, from pre-
equivalent hydraulic conductivity from a 3D image of any porous through to post-cultivation. The work demonstrated that discing
soil. Likewise, Al-Raoush & Willson (2005) used skeletonization and rolling treatments can create porous seedbeds, which can
algorithms (thinning operations, which systematically remove lead to reduced crop establishment because of poor seed-soil
voxels from an object until a minimal but topologically identical contact, especially on fine-textured soils. Papadopoulos et al .
structure is produced) to extract pore-bodies, pore-throats and size (2009) compared stable and unstable aggregate fractions from
distributions in physically realistic pore network structures, which organically and conventionally managed soils to evaluate the
enabled the discrimination of active and inactive pores and the role of management on aggregate structure and stability. X-
characterization of redundant pore throats, demonstrating a clear ray μCT-derived data indicated that intra-aggregate porosity was
advantage of X-ray CT over other, invasive techniques. independent of soil stability, because of the organic matter held
inside stable microaggregates which resisted their breakdown by
Soil compaction, bulk density and stability. Soil compaction slaking.
has been described as the most serious environmental problem Changes in soil structure that occur during wetting and
caused by conventional agriculture (McGarry, 2001), and is drying events have also been quantified by using X-ray CT,
one component of the land-degradation ‘syndrome’ that has allowing previously unachievable comparisons of alterations
several influences on many aspects of general soil management, in the same sample to be made using classical techniques.
including soil erosion (Batey, 2009). Traditional characteristics A sophisticated study by Peth et al . (2010) quantified the
used to determine the effect of soil compaction such as dry effect of hydraulic stress on the dynamic rearrangement of
bulk density and total soil porosity (Lipiec & Hatano, 2003) are solid particles at a resolution of 38.4 μm, and provides a
derived from simple but crude water-holding capacity estimates, good example of the potential of X-ray CT to address spe-
which generally assume uniformity in structure. However non- cific biophysical questions that can not be addressed by other
destructive imaging allows for both the true heterogeneity and means. By exerting capillary pressure at a matric potential
finer details of the porous architecture to be quantified. The of −0.4 kPa to saturate soils prior to air-drying, Peth et al .
© 2013 The Authors

or earthworm burrows must either be removed or filled with a

fixative and thin sectioned for effective measurement (Capowiez
et al ., 1998; Gregory, 2006). These measures are generally very
labour intensive, preclude any further investigations using the
same sample which is a major limitation, only generate single
2D sample slices which is a very laborious process, and can alter
finer structural features during acquisition. Conversely, X-ray
CT is non-destructive and therefore permits repeat probing and
measurement of the same sample over time, allowing in situ mea-
surements from thousands of sections of biotic characteristics as
they respond to changing conditions. Crucially, the spatial nature
of CT scans allows precise 3D measurement of soil features in
situ, a major advantage that is absent from many other techniques.
For example, root architectural and pore space structural char-
acteristics, such as branching or liquid contact angles, tortuosity
and fractal dimensions, can be precisely measured under differing
environmental stresses (Gregory et al ., 2003; Lontoc-Roy et al .,
2006; Crawford et al ., 2011), without the need to interpolate from
2D measurements. However, it should be noted that although
the scan times have been shortened in recent years, the image
analysis is still labour intensive and time consuming.
Figure 4 Greyscale tomographs of the top and central regions of
Avdat loess samples subjected to a shrink-swell regime with different
initial bulk densities (loose = 1.00 g cm−3 , intermediate = 1.2 g cm−3 , Plant development. The development of X-ray CT technology
dense = 1.33 g cm−3 ) in dry and wet conditions (hom = homogenized air has improved to the point of permitting resolutions relevant to
dry, sat = saturated, dry = air dry). Taken from Peth et al . (2010) with root systems, with practical computational times. The relative
permission from Elsevier. abundance of particularly high-resolution industrial X-ray CT
systems has also overcome the previously cited access issues
(2010) were able to induce shrinkage and cause localized (Mooney, 2002), allowing more complex, time series-based or
crack formation (Figure 4). Upon wetting, water menisci at replicated trials. Until recently, much of the literature has been
inter-particle contacts entered pores with sufficient pressure to dominated by short-time studies validating the use of X-ray
cause a rearrangement of pore space, and resulted in a dispersion CT methodology and segmentation algorithms against established
of microaggregates which could be quantified. X-ray CT enabled root system measures. This trend reflects the complexity of
Peth et al . (2010) to observe reorganization of the pore space by the segmentation issues as well as the inability to source a
using the shift in the frequency distribution of greyscale values. standardized, easily accessible solution (Mairhofer et al ., 2012).
Scatter in greyscale values from X-ray CT data during soil Despite the wide range in equipment and processing methods,
saturation indicated aggregate structural reorganization to a more Gregory et al . (2003), Perret et al . (2007) and Flavel et al . (2012)
heterogeneous state. By masking connected pore volumes with observed similar systematic under-estimates of X-ray CT root
an isosurface and colour coding based on size, Peth et al . (2010) length measures of 10, 10 and 8%, respectively, when compared
were able to map soil porosities spatially under different hydraulic with destructive methods over a wide range of root lengths.
loadings, and provide a framework that can be used subsequently Heeraman et al . (1997) observed a 35% over-estimation of root
for 3D spatial analysis of soil shrinkage. Additionally, the appli- volume and a 48% over-estimation of root length density by X-
cation of optical flow-based methods to model strain fields on the ray CT when compared with destructive methods. However, this
soils under different extents of compression showed that most of is thought to be a direct result of voxel misclassification in a
the soil deformation occurred via plastic flow at the crack borders. sample with a large root length density. There appeared to be
greater variation in root diameter (in the studies that measured it)
than in root length estimates, with a range of 4–36% recorded by
Characterization of biotic soil characteristics
Gregory et al . (2003) and a similar range by Perret et al . (2007).
Imaging and distinguishing between similarly low attenuating Despite the increase in variation, the average root diameters were
substances inside a dense heterogeneous substance is challenging, consistent between destructive and X-ray CT techniques (Gregory
but X-ray CT has considerable potential to measure biological et al ., 2003). Other authors did not report destructive root
features such as root systems inside an otherwise opaque and system measurements but the images presented demonstrate clear
complex medium (Gregory et al ., 2003; Young & Crawford, relationships between destructively taken samples and 3D images
2004). Most alternative methods for measuring biotic features (Kaestner et al ., 2006; Mairhofer et al ., 2012; Tracy et al ., 2012b).
within soil are destructive, and characteristics such as root systems It is expected that the programmes and architectural measurement
© 2013 The Authors

systems developed for gel-based plant root phenotyping systems the investigation of water processes in the rhizosphere. Further
(such as RootReader3D) may soon be applied to X-ray CT experimental testing of advanced techniques such as root:soil
tomographs, allowing for complex root system architectural traits contact determination and water flow is required across a range
and time sequences to be quantified (Iyer-Pascuzzi et al ., 2010; of soil textures as well as the mainly coarser textured substrates
Clark et al ., 2011). The measurement of root length density and currently used, which tend to yield better quality images because
the spatio-temporal calculation of root extension rates in situ of a relatively small water content.
represent significant improvement in the ability to understand
and observe root system development and responses to varied Soil microorganisms. The currently available X-ray CT technol-
environmental stimuli (Pierret et al ., 1999a; Tracy et al ., 2012b). ogy is not capable of resolving microorganisms in soil. There
This opens up the possibility to explore soil environment-genotype are many reasons for this limitation, not least of which is the
interactions and specific root traits of plants using X-ray CT, highly attenuating nature of soil, combined with the limited atten-
offering several exciting opportunities for targeted crop research uation of microbes that inhabit thin water films in the soil (Young,
(Gregory et al ., 2009). For a more comprehensive review of the 1998; O’Donnell et al ., 2007). O’Donnell et al . (2007) described
application of X-ray CT to root visualization in soil, see Mooney the possibility of visualizing heavy-metal accumulating (labelled)
et al . (2012). microbes, provided that the metal concentration is sufficient to
achieve adequate attenuation. However, this has yet to be demon-
Water supply to roots. There has been a substantial volume of strated. In an unrelated investigation into fungal attack of struc-
work using ‘low-resolution’ (voxel size of around 2 mm) X-ray tural timbers, Van den Bulcke et al . (2009) used high resolution
CT to quantify water content spatially in the rhizosphere and (0.7 μm voxel size), low power (50 keV, 40 μA) and very small
bulk soil (Hainsworth & Aylmore, 1983; Hamza & Aylmore, samples (1 mm3 ) to visualize fungal hyphae but with limited suc-
1992; Hamza et al ., 2001). Despite the linear calibration regularly cess, as most fungal species produced insufficient contrast to be
achieved between attenuation and water content, single energy able to discriminate them.
X-ray CT is not able to discriminate readily between changes However, despite this inability to visualize microbial com-
in bulk density and water content, and thus relies on the large munities directly, X-ray CT provides an opportunity to observe
assumption of uniform bulk density (Aylmore, 1993; Gregory their habitat and the structural dynamics therein. For example,
& Hinsinger, 1999; Moradi et al ., 2011). The suitability of Thieme et al . (2003) used X-ray microscopy to measure the spa-
X-ray CT for detection of edaphic structures and their spatial tial arrangement of bacteria in the 3D colloidal flocculates of clay
arrangement makes it ideal for use in combination with other particles, oxides and organic substances at sub-200 nm resolution.
techniques that have greater sensitivity to water content such as Thus X-ray CT used in combination with other techniques such
MRI, neutron tomography or micro-tensiometers (Macfall et al ., as thin sectioning and molecular biology techniques rather than
1991; Koebernick, 2011; Moradi et al ., 2011). Hamza et al . (2001) as a stand-alone tool can provide very useful information about
observed changes in root diameters with transpiration rates and, habitat structure, resource flow, microbial populations (size and
similarly, Carminati et al . (2009) observed the development of structure) and spatial distributions (Young & Crawford, 2004;
a ‘gap’ between lupin roots and the soil after transpiration rates Crawford et al ., 2011). Several researchers have suggested that
declined, indicating that the gap was a result, and not the cause, the microstructure of soil is self-organizing under the influence
of water limitation to the plant. Spatial information about the of microbial activity, resource flow, aggregation and rhizosphere
microstructure of the soil in the rhizosphere obtained from high- effects (Young & Crawford, 2004; Feeney et al ., 2006; O’Donnell
resolution X-ray CT is also useful for the modelling of fluid et al ., 2007; Crawford et al ., 2011). This insight has only been
flow and in understanding the impact of root axial expansion made possible with the use of high-resolution X-ray CT for the
on water movement. Aravena et al . (2011) recently provided measurement of fine pore structure at resolutions relevant to mois-
evidence that under small bulk densities, the compaction of soil ture storage. The new perspective on the nature of pore space at
in the rhizosphere will improve soil:root contact and increase the this high resolution has also challenged aspects of accepted larger-
unsaturated hydraulic conductivity between aggregates; this has a scale ecosystem models (Feeney et al ., 2006), such as hierarchical
net positive effect on the availability and transport of water to the aggregation and carbon capture. Future non-invasive decoupling
root. Schmidt et al . (2012) reported a new method for quantifying of the role that specific microbial communities play in modify-
root:soil contact from 3D X-ray CT images, by segmenting the ing the spatial architecture of soils may eventually allude to the
root and growth media separately and dilating the growth medium function of specific microbial populations at the field scale.
volume by 1 pixel to create a region of overlap that corresponds
to the area of contact of the root with the growth media. This Insects and earthworms. The use of X-ray CT in earthworm
approach has great potential for unravelling the precise responses studies has been well documented (Taina et al ., 2008), and will
of plant roots and soils under water stress, but has only been be mentioned only briefly here. The advantages in using X-ray CT
tested on loamy sand and vermiculite substrates. Despite the for the quantitative and qualitative characterization of invertebrate
current limitations of X-ray CT for the visualization of soil water burrows in soil are much the same as those for root systems,
distribution, this technique can provide useful information for allowing observation of the burrows through time and with the
© 2013 The Authors

concurrent measurement of soil physical properties. In the case multiple scanning of the same sample to investigate formation and
of measuring burrows, X-ray CT overcomes the problems of subsequent changes with time. Whilst there are moves towards the
more labour-intensive methods, such as using plaster of Paris, adoption of this approach (Tracy et al . 2012a), there still remains
polyurethane or lead to create moulds (Capowiez et al ., 1998). a research gap in that there has yet to be any work that has
X-ray CT studies of earthworm activity have been conducted as sought to quantify the impact that repeated scanning of the same
either time-sequence measurements in pots (Joschko et al ., 1993), sample might have on constituents such as microbial community
or as contrasting intact cores taken from the field (Capowiez composition, root growth or even soil structure.
et al ., 1998; Langmaack et al ., 1999). Langmaack et al . (1999,
2002), Schrader et al . (2007) and Jegou et al . (2002) explored the
Sample size and resolution
effects of soil compaction on earthworm burrowing, and Capowiez
et al . (1998) explored seasonal variation in burrowing behaviour. High-resolution X-ray CT studies have often been limited to
In addition, many other workers have quantified species-specific small sample sizes to achieve the required resolution. Synchrotron
burrowing behaviour (Taina et al ., 2008). However, there appear tomography continues to provide excellent resolution (achieving
to be three sources of error in the use of this technique for burrow resolutions of <1 μm) and contrast and particularly high flux
quantification: (i) the false identification of pores such as root (brilliance), allowing fast acquisition times. For example, the
channels as burrows (Joschko et al ., 1993), (ii) burrows that Advanced Photon Source, Illinois, as used by Wang et al .
earthworms have filled with casts are not identified (Capowiez (2001), achieved voxel resolutions 1 μm–100 nm in a few minutes
et al ., 1998) and (iii) deformation of the burrows because of per scan, with high performance computers to reconstruct in
localised wetting and drying cycles for example. Schrader et al . a comparable time-frame. However, limitations in the range of
(2007) demonstrated the unique potential of X-ray CT in this area source energies that can be generated (commonly less than 50 keV)
by measuring the concentric compaction from axial expansion continue to restrict the sample size, particularly of high density
around earthworm burrows and calculating the potential volume materials such as soils (Wildenschild et al ., 2002).
of soil that could be affected by a single Lumbricus terrestris L. For industrial X-ray CT equipment (with cone beam sources),
earthworm, but conceded that the calculation was dependent on the spatial resolution is largely the function of the focal spot
experimental conditions and that the density measurements were size, performance of the detector and the magnification of the
possibly confounded by moisture content. Likewise, Langmaack system (distance between sample and detector). Some current
et al . (2002) demonstrated the positive influence that L. terrestris commercial systems claim feature recognition of down to 50 nm
earthworms have on soil physical properties in relation to the for small samples using low energies. More common are voxel
rehabilitation of degraded soil structure caused by tillage-induced resolutions as small as 0.7–1 μm, with sufficient source power
compaction. By scanning undisturbed soil monoliths collected in for soil analysis (Sun et al ., 2012). In medical and some
the field, Langmaack et al . (2002) were able to visualize rapidly industrial scanners there is often a fixed distance between source
the pore volume, pore length, tortuosity and total connectivity of and detector, in which case the limitation becomes the static
burrow systems under both conventional and conservation tillage relationship between the size of the sample and the resolution
treatments, and concluded that intrinsic soil processes had a bigger achievable. This is dependent on the distance between source
impact on soil rehabilitation than tillage operations. and detector, the size of the detector and the density of the
X-ray CT has also been used to map the nesting architecture pixels in the detector panel. In other words there is a ‘trade
of other burrowing invertebrate species such as blue banded off’ between sample size and resolution (high resolution is
bees (Greco et al ., 2006). Greco et al . (2006) observed 3D nest achieved when small samples are placed closer to the source,
architectures for the first time, and highlighted the potential for allowing greater magnification; Dhondt et al ., 2010). This is of
monitoring fungal parasitism load in addition to following the particular significance where large samples are required for field-
lifecycle and development of individual lava and pupa. representative sampling, but high resolution is required for the
scientifically relevant detail of water storage pores or fine root
distributions for example (Young et al ., 2001). It is of particular
Future technological expectations from X-ray systems
concern for biotic studies (such as earthworm burrows or root
Carlson (2006) reflected that the developments in X-ray CT have systems), where time-sequence scans are taken of ‘microcosms’
resulted historically from technological breakthroughs and that in which biotic factors are rapidly constrained and can distort
this trend is likely to continue with improved resolution, faster realistic results. This is principally why many fine root plant
scan times and acquisition efficiencies, and an associated greater studies have been limited to the development of young root
requirement for computational power. Following a review of the systems. However, the development of larger, denser detector
industrial X-ray CT market, Sun et al . (2012) concluded recently arrays of greater sensitivity allows far larger samples to be scanned
that although the demand is growing rapidly, the development at greater resolution.
of the technology is in its infancy. Our review supports this Traditionally, it has also been considered to be beneficial to
conclusion, as the majority of previous studies have failed to ensure that the scanned object fits inside the field of view, but
exploit a major benefit of non-invasive imaging; that is, the the successful extensive application of ‘region-of-interest’ local
© 2013 The Authors

tomography in other disciplines such as geology (Stock, 2008) the material), where better phase contrast can be achieved between
demonstrates potential benefit, which should be considered and water and air (Wildenschild et al ., 2002).
further exploited for soil research. In region of interest tomography
the field of view is smaller than the sample size, so reconstruction
Characterization of soil chemical properties
is focused on a sub-sample. This allows greater spatial resolutions,
providing further insights into spatial geometries at resolutions not The linear attenuation coefficient calculated by X-ray CT is
currently accessible from stand-alone systems, but at the expense relative to the average atomic number of the voxel. Single
of complete sample imaging and, usually, reduced image quality. energy X-ray CT is limited by the additive effects of photo-
Carminati et al . (2009) used local tomography effectively to electric absorption and Compton scattering. So it follows that if
visualize pore formation around drying Lupin roots. By zooming substances within a sample area are sufficiently homogeneous,
on specific regions of the sample, they were able to increase it should be possible to visualize their distribution with X-ray
image resolution from 320 μm for the entire sample to 90 μm for CT and create a 3D elemental map. Recently, Godel et al .
isolated regions of the tap and lateral roots, allowing previously (2010) used high-resolution X-ray CT with backscattered SEM
unachievable soil and root shrinkage to be observed at coarser thin sections to describe the spatial arrangement of platinum
resolutions. group elements and formation processes in ore cores from
South Africa. Godel et al . (2010) were able to assign chemical
compositions to typical attenuation values because of the relative
Image contrast
homogeneity of substances within the sample, allowing deposits to
Image contrast is a function of the X-ray source energy, detector be visualized in 3D. Hapca et al . (2011) have undertaken similar
characteristics and the range of electron densities in the sample. work in soil, developing a protocol to automatically register
The relative energy of the X-ray source in relation to the average 2D elemental maps generated by scanning electron microscope-
density of the sample determines how many X-rays penetrate energy dispersive X-ray spectroscopy (SEM-EDX) thin sections,
the sample. Where source energies are too large, the sample and then mapping them onto X-ray CT images. Using co-kriging,
becomes transparent to the X-rays; conversely, when they are a 3D spatial distribution of elemental composition (for elements
too small there are insufficient X-rays passing through the object C, O, K, Ca and Si) was calculated from their attenuation values.
for acceptable counting statistics (Wildenschild et al ., 2002). The Verification of the method indicated that the technique is limited
information that can be acquired in a set amount of time is to substances that are very spatially pure or in great abundance,
therefore relative to the detector type and specifications. Detector with less application for more diffuse (smaller concentration)
technology is continually improving, with regular increases in and heterogeneously distributed elements such as nitrogen or
dynamic range, sensitivity and acquisition rates (which are phosphorus. Nonetheless, this technique reveals excellent potential
inextricably linked) (Sun et al ., 2012). for this application and was useful for demonstrating the spatial
A common way to improve the contrast between substances chemical heterogeneity in soil.
of similar attenuation is to increase the electron density of one
substance artificially by using doping (tracer) agents. The addition
High-throughput X-ray CT systems
of tracer solutions such as potassium or sodium iodide and
calcium bromide to the liquid phases in soil has frequently been The advent of high-flux X-ray sources and the increased efficiency
employed in solute transport experiments, and has allowed the of detectors have seen dramatic decreases in scan times in
evaluation of flow morphology through soils because of their high recent years. Most modern X-ray CT systems are now able to
contrasting X-ray attenuations (Hopmans et al ., 1992; Clausnitzer scan samples within about an hour, albeit with variable image
& Hopmans, 2000; Anderson et al ., 2003). Artificial doping is quality. However, the use of constant rotation tomography or
a common practice in studies of interfacial dynamics seeking to ‘fast scanning’ by Flavel et al . (2012) and Tracy et al . (2012a)
improve the contrast of water-air or water-oil phases (Culligan has seen scanning times of about 4 and 20 minutes respectively,
et al ., 2004; Carlson, 2006; Prodanović et al ., 2006; Kaestner both providing images of sufficiently good quality with low
et al ., 2008; Wildenschild et al ., 2011). Tracers change the noise to differentiate root growth from surrounding soil phases.
density, surface tension and contact angle between the liquid and Three years ago Gregory et al . (2009) described one of the
solid phase (dependent on the concentration used) (Kaestner et al ., main challenges of X-ray CT adoption in soil science as the
2008). However, with improved detector contrast and dynamic limitation of typically analysing <10 samples in 1 day, reducing
range, the need for such measures for examining soil water the potential for replicated trials. However, recent work has
may now be less necessary. Mooney (2002) and more recently demonstrated the potential for high-throughput, highly replicated
Tippkötter et al . (2009) were able to achieve good contrast trials in modern commercially available small-footprint systems.
between soil water and air-filled pore space without doping For example, Tracy et al . (2012a) were able to scan 12 tomato
agents. This is especially the case when scanning energies are plants daily over 10 days of growth, with three replicates of the
above 100 keV and where attenuation is dominated by Compton four different treatments investigated. This short temporal range
scattering (which is mainly dependent on the electron density of allowed the assessment of rooting characteristics with techniques
© 2013 The Authors

such as convex hull, centroid and enclosing circle analysis sensitivity of the technique to hydrogen molecules. However, it
throughout early-stage plant development. Likewise, Yang et al . still remains unsuited to high throughput studies (Nagel et al .,
(2011) recently reported a fully automated high-throughput X- 2009; Van Dusschoten, 2011) and is limited by ferrous soil
ray CT system, in which an X-ray CT scanner connected to constituents. Nagel et al . (2009) used MRI in combination with
an industrial conveyor can automatically deliver pot-grown rice positron emission tomography (PET) to identify the allocation of
11
plants to be scanned, and subsequent automated analysis. The C isotope-labelled photosynthate in root systems under different
complete automated procedure is reported to take 25 s per plant temperature regimes. In this way the combination of structural
(nine for plant transport, nine for image acquisition, one for tomographic techniques can be complemented by the functional
manipulator stage to rise, one for the manipulator stage to fall and identification through PET scanning (Garbout et al ., 2012).
five for image processing and analysis), resulting in the possibility
of analysing 4320 pots per continuous 24-hour working day. This
Application of phase contrast systems
compares with a reported image analysis time of about 200 s per
plant by manual counting, which also involves human disturbance X-ray CT calculates the average attenuation value for every
of the plant. Whilst the analysis of above-ground plant samples given voxel and as a result, the boundary between similarly
(in this case tiller number) is far simpler than soil environments attenuating substances (phases) can be difficult or impossible to
because there are fewer material phases present, the transfer of delineate. Phase contrast techniques have been developed using
similar technologies to soil science will undoubtedly have huge both synchrotron radiation as well as laboratory-based sources
benefits for high-throughput research. to use the average attenuation values as well as refractive index
variation between phases to produce images with greatly increased
contrast at phase boundaries (Moran et al ., 2000; Lewis et al .,
Combining X-ray CT with other complementary techniques
2005). Moran et al . (2000) proposed a method for the use of
The high resolution detection of structural features using X-ray 2D phase contrast imaging of roots from sections of intact cores,
CT can be complemented by techniques that are either (i) able allowing clearer identification of root section characteristics (of
to distinguish features with greater resolution or low density length or diameter) and observation of the root-soil interface.
detail, such as microbial community distribution, or (ii) of smaller Phase contrast has also been widely applied and developed in
resolution but are able to provide functional information, such as medical fields to image complex organs, such as lungs and
water content or photosynthate accumulation. The combination their microstructure, that are very weakly attenuating with large
of 2D elemental or chemical information with X-ray CT has the numbers of phase boundaries (Kitchen et al ., 2005; Lewis et al .,
potential to allow 3D visualization of soil chemical properties, 2005). It is therefore likely that suitable phase contrast could be
with implications for understanding the spatial heterogeneity of achieved in soils to improve the delineation between air and water
soils (Jacobson et al ., 2007; Hapca et al ., 2011). In a similar phases as well as boundaries between roots and water, which
fashion, it has been suggested that sequential biological thin have been a particular challenge for root segmentation methods.
sectioning and geostatistical methods using 3D information from However, the dramatically increased scan times (>10 hours) have
X-ray CT would complement our understanding of the distribution hindered the uptake of laboratory-based systems, along with the
of microbial communities in the soil (Young et al ., 2001, 2008; increased cost of scanners (often twice that of typical industrial
O’Donnell et al ., 2007). X-ray CT systems).
Neutron radiography and tomography are widely recognized as
complementary techniques to X-ray CT, as they have different
Conclusions
elemental sensitivities (Mees et al ., 2003; Vontobel et al ., 2006;
Moradi et al ., 2011). Whilst X-ray radiation is predominantly Many traditional and current analytical methods for exploring a
attenuated by interactions with electrons and therefore more soil’s biophysical structure and subsequent behaviour do not fully
strongly attenuated by high atomic number elements, neutron account for the heterogeneity and variability of field soils, limiting
radiation is attenuated by atomic nuclei and is thus more strongly the successful integration of data into soil models. As with many
attenuated by elements with a low atomic number (Kaestner et al ., emerging technologies, a significant volume of previous research
2008). Several neutron radiography and tomography studies have has focused on exploring the potential applications of X-ray CT in
been conducted to explore water uptake dynamics around root soil science. As a result, one might fairly consider that there has
systems (Carminati et al ., 2010; Moradi et al ., 2011). Kaestner been an over-emphasis on collecting data that might have been
et al . (2008) demonstrate how neutron imaging, which has a ascertained quicker and with reduced cost by more traditional
smaller spatial resolution but does have the ability to calculate approaches. However, as shown in this paper, the most recent
water content at high temporal resolution, can be combined with research is now revealing the full benefit of a technique that
X-ray tomography, with high spatial 3D pore geometry, and be enables us to truly visualize and measure the undisturbed 3D
used to refine fluid flow and distribution models. inner space of soil at highly relevant resolutions for key soil
MacFall et al . (1991) used MRI to observe heterogeneous water functions. Furthermore, with super fast scan times (within the
uptake of Loblolly pine seedlings in 2D, exploiting the greater order of 5–10 minutes), we are almost able to visualize dynamic
© 2013 The Authors

soil behaviour in high-throughput environments, enabling us to Aylmore, L.A.G. 1993. Use of computer-assisted tomography in studying
quantify multiscale uncertainty and heterogeneity. Improvements water movement around plant roots. Advances in Agronomy, 39, 1–54.
in X-ray CT detectors now allow for easier discrimination Batey, T. 2009. Soil compaction and soil management – a review. Soil
between soil phases, including the quantification of root systems Use and Management, 25, 335–345.
embedded within soil, enabling the visualization of the emerging Baveye, P., Rogasik, H., Wendroth, O., Onasch, I. & Crawford, J.W.
2002. Effect of sampling volume on the measurement of soil physical
rhizosphere. However, there still remain certain technological and
properties: simulation with X-ray tomography data. Measurement
user constraints that are hindering progress. Whilst the cost of
Science and Technology, 13, 775–784.
X-ray CT systems has dramatically decreased and accessibility
Baveye, P.C., Laba, M., Otten, W., Bouckaert, L., Dello Sterpaio, P.,
issues have been reduced, the uptake of the technology for soils Goswami, R.R., et al . 2010. Observer-dependent variability of the
research has been slow. Previous research has made huge progress thresholding step in the quantitative analysis of soil images and X-ray
in visualizing pore channels and soil constituents, but often this microtomography data. Geoderma, 157, 51–63.
has not been focused on trying to unravel how soil features Capowiez, Y., Pierret, A., Daniel, O., Monestiez, P. & Kretzschmar, A.
form or change with time. Likewise, there have been very few 1998. 3D Skeleton reconstructions of natural earthworm burrow systems
appropriately replicated studies to date, which has often led to using CAT scan images of soil cores. Biology and Fertility of Soils, 27,
X-ray CT being perceived as an exclusive, novel and a largely 51–59.
inaccessible research tool. Whilst this may have been true 10 years Capowiez, Y., Pierret, A. & Moran, C. 2003. Characterisation of the
ago, X-ray CT should no longer be considered as a novel approach three-dimensional structure of earthworm burrow systems using image
but, instead, an accepted and tested means of characterizing soil analysis and mathematical morphology. Biology and Fertility of Soils,
systems. 38, 301–310.
Carlson, W.D. 2006. Three-dimensional imaging of earth and planetary
materials. Earth and Planetary Science Letters, 249, 133–147.
Acknowledgements Carlson, W., Denison, C. & Ketcham, R. 1999. High-resolution X-
ray computed tomography as a tool for visualization and quantitative
Funding for this research was provided by the Lawes Agricultural
analysis of igneous textures in three dimensions. Visual Geosciences, 4,
Trust, the University of Nottingham and the University of New
1–14.
England. Carminati, A., Vetterlein, D., Weller, U., Vogel, H.-J. & Oswald, S.E.
2009. When roots lose contact. Vadose Zone Journal , 8, 805–809.
References Carminati, A., Moradi, A.B., Vetterlein, D., Vontobel, P., Lehmann,
E., Weller, U., et al . 2010. Dynamics of soil water content in the
Adderley, W.P., Simpson, I.A. & MacLeod, G.W. 2001. Testing high-
rhizosphere. Plant and Soil , 332, 163–176.
resolution X-ray computed tomography for the micromorphological
Clark, R.T., MacCurdy, R.B., Jung, J.K., Shaff, J.E., McCouch, S.R.,
analyses of archaeological soils and sediments. Archaeological Prospec-
Aneshansley, D.J., et al . 2011. Three-dimensional root phenotyping
tion, 8, 107–112.
with a novel imaging and software platform. Plant Physiology, 156,
Al-Raoush, R.I. & Willson, C.S. 2005. Extraction of physically realistic
455–465.
pore network properties from three-dimensional synchrotron X-ray
Clausnitzer, V. & Hopmans, J.W. 2000. Pore-scale measurements of solute
microtomography images of unconsolidated porous media systems.
breakthrough using microfocus X-ray computed tomography. Water
Journal of Hydrology, 300, 44–64.
Resource Research, 36, 2067–2079.
Anderson, S.H., Peyton, R.L. & Gantzer, C.J. 1990. Evaluation of con-
Crawford, J.W., Deacon, L., Grinev, D., Harris, J.A., Ritz, K., Singh,
structed and natural soil macropores using X-ray computed tomography.
B.K., et al . 2011. Microbial diversity affects self-organization of the
Geoderma, 46, 13–29.
Anderson, S.H., Peyton, R.L., Wigger, J.W. & Gantzer, C.J. 1992. soil–microbe system with consequences for function. Journal of the
Influence of aggregate size on solute transport as measured using Royal Society, Interface Online only at present. doi: 10.1098/rsif.
computed tomography. Geoderma, 53, 387–398. 2011.0679.
Anderson, S.H., Wang, H., Petyton, R.L. & Gantzer, C.J. 2003. Estimation Crestana, S., Mascarenhas, S. & Pozzi-Mucelli, R.S. 1985. Static and
of porosity and hydraulic conductivity from X-ray CT-measured solute dynamic 3 dimensional studies of water in soil using computed
breakthrough. In: Applications of X-ray Computed Tomography in the tomographic scanning. Soil Science, 140, 326–332.
Geosciences (eds F. Mees, R. Swennen, M. Van Geet & P. Jacobs), pp. Culligan, K.A., Wildenschild, D., Christensen, B.S.B., Gray, W.G., Rivers,
135–149. Geological Society Special Publication, London. M.L. & Tompson, A.F.B. 2004. Interfacial area measurements for
Aravena, J.E., Berli, M., Ghezzehei, T.A. & Tyler, S.W. 2011. Effects of unsaturated flow through a porous medium. Water Resource Research,
root-induced compaction on rhizosphere hydraulic properties - X-ray 40, W12413.
microtomography imaging and numerical simulations. Environmental De Gryze, S., Jassogne, L., Six, J., Bossuyt, H., Wevers, M. & Merckx,
Science and Technology, 45, 425–431. R. 2006. Pore structure changes during decomposition of fresh residue:
Atkinson, B.S., Sparkes, D.L. & Mooney, S.J. 2007. Using selected soil X-ray tomography analyses. Geoderma, 134, 82–96.
physical properties of seedbeds to predict crop establishment. Soil and Delerue, J.F., Perrier, E., Timmerman, A. & Swennen, R. 2003. 3D Soil
Tillage Research, 97, 218–228. image characterization applied to hydraulic properties computation. In:
Atkinson, B.S., Sparkes, D.L. & Mooney, S.J. 2009. Effect of seedbed Applications of X-ray Computed Tomography in the Geosciences (eds F.
cultivation and soil macrostructure on the establishment of winter wheat Mees, R. Swennen, M. Van Geet & P. Jacobs), pp. 167–176. Geological
(triticum aestivum). Soil and Tillage Research, 103, 291–301. Society Special Publication, London.
© 2013 The Authors

Denison, C. & Carlson, W.D. 1997. Three-dimensional quantitative Hainsworth, J. & Aylmore, L. 1983. The use of computer assisted
textural analysis of metamorphic rocks using high-resolution computed tomography to determine spatial distribution of soil water content.
X-ray tomography: part II. Application to natural samples. Journal of Australian Journal of Soil Research, 21, 435–443.
Metamorphic Geology, 15, 45–57. Hamza, M. & Aylmore, L. 1992. Soil solute concentration and water
Dhondt, S., Vanhaeren, H., Van Loo, D., Cnudde, V. & Inzé, D. 2010. Plant uptake by single lupin and radish plant roots. Plant and Soil , 145,
structure visualization by high-resolution X-ray computed tomography. 187–196.
Trends in Plant Science, 15, 419–422. Hamza, M.A., Anderson, S.H. & Aylmore, L.A.G. 2001. Studies of soil
Duliu, O.G. 1999. Computer axial tomography in geosciences: an water drawdowns by single radish roots at decreasing soil water content
overview. Earth-Science Reviews, 48, 265–281. using computer-assisted tomography. Soil Research, 39, 1387–1396.
Elliot, T.R., Reynolds, W.D. & Heck, R.J. 2010. Use of existing pore Hapca, S.M., Wang, Z.X., Otten, W., Wilson, C. & Baveye, P.C. 2011.
models and X-ray computed tomography to predict saturated soil Automated statistical method to align 2D chemical maps with 3D
hydraulic conductivity. Geoderma, 156, 133–142. X-ray computed micro-tomographic images of soils. Geoderma, 164,
Elyeznasni, N., Sellami, F., Pot, V., Benoit, P., Vieuble-Gonod, L., Young, 146–154.
I., et al . 2012. Exploration of soil micromorphology to identify coarse- Hargreaves, C., Gregory, P. & Bengough, A. 2009. Measuring root traits in
sized OM assemblages in X-ray CT images of undisturbed cultivated barley (hordeum vulgare ssp. Vulgare and ssp. Spontaneum) seedlings
soil cores. Geoderma, 179, 38–45. using gel chambers, soil sacs and X-ray microtomography. Plant and
Feeney, D.S., Crawford, J.W., Daniell, T., Hallett, P.D., Nunan, N., Ritz, Soil , 316, 285–297.
K., et al . 2006. Three-dimensional microorganization of the soil-root- Heeraman, D.A., Hopmans, J.W. & Clausnitzer, V. 1997. Three dimen-
microbe system. Microbial Ecology, 52, 151–158. sional imaging of plant roots in situ with X-ray computed tomography.
Flannery, B.P., Deckman, H.W., Roberge, W.G. & D’Amico, K.L. 1987. Plant and Soil , 189, 167–179.
Three-dimensional X-ray microtomography. Science, 237, 1439–1444. Heijs, A.W.J., Delange, J., Schoute, J.F.T. & Bouma, J. 1995. Computed-
Flavel, R.J., Guppy, C.N., Tighe, M., Watt, M., McNeill, A. & Young, tomography as a tool for nondestructive analysis of flow patterns in
I.M. 2012. Non-destructive quantification of cereal roots in soil using macroporous clay soils. Geoderma, 64, 183–196.
high-resolution X-ray tomography. Journal of Experimental Botany, 63, Hopmans, J.W., Vogel, T. & Koblik, P.D. 1992. X-ray tomography of soil
2503–2511. water distribution in one-step outflow experiments. Soil Science Society
Gantzer, C.J. & Anderson, S.H. 2002. Computed tomographic measure- of America Journal , 56, 355–362.
ment of macroporosity in chisel-disk and no-tillage seedbeds. Soil and Hounsfield, G.N. 1973. Computerized transverse axial scanning (tomog-
Tillage Research, 64, 101–111. raphy): part 1. Description of system. British Journal of Radiology, 46,
Garbout, A., Munkholm, L.J., Hansen, S.B., Petersen, B.M., Munk, O.L. 1016–1022.
& Pajor, R. 2012. The use of PET/CT scanning technique for 3D Iassonov, P., Gebrenegus, T. & Tuller, M. 2009. Segmentation of X-
visualization and quantification of real-time soil/plant interactions. Plant ray computed tomography images of porous materials: a crucial step
and Soil , 352, 113–127. for characterization and quantitative analysis of pore structures. Water
Géraud, Y., Surma, F. & Mazerolle, F. 2003. Porosity and fluid flow Resource Research, 45, W09415.
characterization of granite by capillary wetting using X-ray computed Iyer-Pascuzzi, A.S., Symonova, O., Mileyko, Y., Hao, Y., Belcher, H.,
tomography. Geological Society, London, Special Publications, 215, Harer, J., et al . 2010. Imaging and analysis platform for automatic
95–105. phenotyping and trait ranking of plant root systems. Plant Physiology,
Godel, B., Barnes, S.J., Barnes, S.-J. & Maier, W.D. 2010. Platinum 152, 1148–1157.
ore in three dimensions: insights from high-resolution X-ray computed Jacobson, A.R., Dousset, S., Andreux, F. & Baveye, P.C. 2007. Electron
tomography. Geology, 38, 1127–1130. microprobe and synchrotron X-ray fluorescence mapping of the
Greco, M., Bell, M., Spooner-Hart, R. & Holford, P. 2006. X-ray heterogeneous distribution of copper in high-copper vineyard soils.
computerized tomography as a new method for monitoring amegilla Environmental Science and Technology, 41, 6343–6349.
holmesi nest structures, nesting behaviour, and adult female activity. Jegou, D., Brunotte, J., Rogasik, H., Capowiez, Y., Diestel, H., Schrader,
Entomologia Experimentalis et Applicata, 120, 71–76. S., et al . 2002. Impact of soil compaction on earthworm burrow systems
Gregory, P.J. 2006. Plant Roots: Growth, Activity and Interaction with using X-ray computed tomography: preliminary study. European
Soils. Blackwell Publishing, Oxford. Journal of Soil Biology, 38, 329–336.
Gregory, P.J. & Hinsinger, P. 1999. New approaches to studying chemical Jenneson, P.M., Gilboy, W.B., Morton, E.J. & Gregory, P.J. 2003. An
and physical changes in the rhizosphere: an overview. Plant and Soil , X-ray micro-tomography system optimised for the low-dose study of
211, 1–9. living organisms. Applied Radiation and Isotopes, 58, 177–181.
Gregory, P.J., Hutchison, D.J., Read, D.B., Jenneson, P.M., Gilboy, Joschko, M., Müller, P.C., Kotzke, K., Döhring, W. & Larink, O. 1993.
W.B. & Morton, E.J. 2003. Non-invasive imaging of roots with high Earthworm burrow system development assessed by means of X-ray
resolution X-ray micro-tomography. Plant and Soil , 255, 351–359. computed tomography. Geoderma, 56, 209–221.
Gregory, P.J., Bengough, A.G., Grinev, D., Schmidt, S., Thomas, W.T.B., Kaestner, A., Schneebeli, M. & Graf, F. 2006. Visualizing three-
Wojciechowski, T., et al . 2009. Root phenomics of crops: opportunities dimensional root networks using computed tomography. Geoderma,
and challenges. Functional Plant Biology, 36, 922–929. 136, 459–469.
Grevers, M.C.J., Jong, E.D. & St. Arnaud, R.J. 1989. The characterization Kaestner, A., Lehmann, E. & Stampanoni, M. 2008. Imaging and image
of soil macroporosity with CT scanning. Canadian Journal of Soil processing in porous media research. Advances in Water Resources, 31,
Science, 69, 629–637. 1174–1187.
© 2013 The Authors

Kalukin, A.R., Van Geet, M. & Swennen, R. 2000. Principal components McGarry, D. 2001. Tillage and soil compaction. In: 1st World Congress on
analysis of multienergy X-ray computed tomography of mineral Conservation Agriculture (eds L. Garcia-Torres, J. Benites, A. Martinez-
samples. IEEE Transactions on Nuclear Science, 47, 1729–1736. Vilela & A. Holgado-Cabrera), pp. 307–316. Springer, Madrid.
Ketcham, R.A. 2005. Three-dimensional grain fabric measurements using Mees, F., Swennen, R., Van Geet, M. & Jacobs, P. 2003. Applications
high-resolution X-ray computed tomography. Journal of Structural of X-ray Computed tomography in the Geosciences. Geological Society
Geology, 27, 1217–1228. Special Publications, London.
Ketcham, R.A. & Carlson, W.D. 2001. Acquisition, optimization and Mooney, S.J. 2002. Three-dimensional visualization and quantification of
interpretation of X-ray computed tomographic imagery: applications to soil macroporosity and water flow patterns using computed tomography.
the geosciences. Computers and Geosciences, 27, 381–400. Soil Use and Management, 18, 142–151.
Kettridge, N. & Binley, A. 2011. Characterization of peat structure using Mooney, S.J., Morris, C., Craigon, J. & Berry, P. 2007. Quantification
X-ray computed tomography and its control on the ebullition of biogenic of soil structural changes induced by cereal anchorage failure: image
gas bubbles. Journal of Geophysical Research, 116, G01024. analysis of thin sections. Journal of Plant Nutrition and Soil Science,
Kim, H., Anderson, S.H., Motavalli, P.P. & Gantzer, C.J. 2010. 170, 363–372.
Compaction effects on soil macropore geometry and related parameters Mooney, S., Pridmore, T., Helliwell, J. & Bennett, M. 2012. Developing
for an arable field. Geoderma, 160, 244–251. X-ray computed tomography to non-invasively image 3-D root systems
Kinney, J.H. & Nichols, M.C. 1992. X-Ray tomographic microscopy architecture in soil. Plant and Soil , 352, 1–22.
(XTM) using synchotron radiation. Annual Review of Materials Science, Moradi, A.B., Carminati, A., Vetterlein, D., Vontobel, P., Lehmann,
22, 121–152. E., Weller, U., et al . 2011. Three-dimensional visualization and
Kitchen, M.J., Lewis, R.A., Yagi, N., Uesugi, K., Paganin, D., Hooper, quantification of water content in the rhizosphere. New Phytologist,
S.B., et al . 2005. Phase contrast X-ray imaging of mice and rabbit lungs: 192, 653–663.
a comparative study. British Journal of Radiology, 78, 1018–1027.
Moran, C., Pierret, A. & Stevenson, A. 2000. X-ray absorption and phase
Koebernick, N. 2011. In-situ observation of root-growth by X-ray
contrast imaging to study the interplay between plant roots and soil
tomography – relationship between root and water distribution during
structure. Plant and Soil , 223, 101–117.
a drying cycle. In: Rhizosphere, 3 International Conference (ed. H.
Nagel, K.A., Kastenholz, B., Jahnke, S., Van Dusschoten, D., Aach,
Lambers). Perth.
T., Muhlich, M., et al . 2009. Temperature responses of roots: impact
Langmaack, M., Schrader, S., Rapp-Bernhardt, U. & Kotzke, K. 1999.
on growth, root system architecture and implications for phenotyping.
Quantitative analysis of earthworm burrow systems with respect to
Functional Plant Biology, 36, 947–959.
biological soil-structure regeneration after soil compaction. Biology and
Norrish, K. & Hutton, J.T. 1969. An accurate X-ray spectrographic method
Fertility of Soils, 28, 219–229.
for the analysis of a wide range of geological samples. Geochimica et
Langmaack, M., Schrader, S., Rapp-Bernhardt, U. & Kotzke, K. 2002.
Cosmochimica Acta, 33, 431–453.
Soil structure rehabilitation of arable soil degraded by compaction.
Nunan, N., Ritz, K., Rivers, M., Feeney, D.S. & Young, I.M. 2006.
Geoderma, 105, 141–152.
Investigating microbial micro-habitat structure using X-ray computed
Lehmann, P., Wyss, P., Flisch, A., Lehmann, E., Vontobel, P., Krafczyk,
tomography. Geoderma, 133, 398–407.
M., et al . 2006. Tomographical imaging and mathematical description
O’Donnell, A.G., Young, I.M., Rushton, S.P., Shirley, M.D. & Crawford,
of porous media used for the prediction of fluid distribution. Vadose
J.W. 2007. Visualization, modelling and prediction in soil microbiology.
Zone Journal , 5, 80–97.
Nature Reviews Microbiology, 5, 689–699.
Lewis, R.A., Yagi, N., Kitchen, M.J., Morgan, M.J., Paganin, D., Siu,
K.K.W., et al . 2005. Dynamic imaging of the lungs using X-ray phase Orsi, T.H. & Anderson, A.L. 1995. X-ray computed tomography
contrast. Physics in Medicine and Biology, 50, 5031–5040. of macroscale variability in sediment physical properties, offshore
Lipiec, J. & Hatano, R. 2003. Quantification of compaction effects on soil Louisiana. In: Transactions of the Forty-Fifth Annual Convention of the
physical properties and crop growth. Geoderma, 116, 107–136. Gulf Coast Association of Geological Societies (eds C.J. John & M.R.
Lontoc-Roy, M., Dutilleul, P., Prasher, S.O., Han, L., Brouillet, T. & Byrnes), pp. 475–480. Gulf Coast Association of Geological Societies,
Smith, D.L. 2006. Advances in the acquisition and analysis of CT scan New Orleans.
data to isolate a crop root system from the soil medium and quantify Orsi, T.H., Edwards, C.M. & Anderson, A.L. 1994. X-ray computed
root system complexity in 3-D space. Geoderma, 137, 231–241. tomography; a nondestructive method for quantitative analysis of
Macedo, A., Crestana, S. & Vaz, C.M.P. 1998. X-ray microtomography sediment cores. Journal of Sedimentary Research, 64, 690–693.
to investigate thin layers of soil clod. Soil and Tillage Research, 49, Otten, W., Grinev, D., Baveye, P.C., Wang, Z., Hapca, S.M. & Wilson,
249–253. C. 2010. Integrating physical and chemical techniques to characterise
Macfall, J.S., Johnson, G.A. & Kramer, P.J. 1991. Comparative water soil microsites. In: 19th World Congress of Soil Science: Soil Solutions
uptake by roots of different ages in seedlings of loblolly pine (pinus for a Changing World (eds R.J. Gilkes & N. Prakongkep), pp. 60–63.
taeda L.). New Phytologist, 119, 551–560. International Union of Soil Sciences, Brisbane.
Mairhofer, S., Zappala, S., Tracy, S.R., Sturrock, C., Bennett, M., Mooney, Papadopoulos, A., Bird, N.R.A., Mooney, S.J. & Whitmore, A.P. 2008.
S.J., et al . 2012. RooTrak: automated recovery of three-dimensional Fractal analysis of pore roughness in images of soil using the slit island
plant root architecture in soil from X-ray microcomputed tomography method. Vadose Zone Journal , 7, 456–460.
images using visual tracking. Plant Physiology, 158, 561–569. Papadopoulos, A., Bird, N.R.A., Whitmore, A.P. & Mooney, S.J. 2009.
Marcelino, V., Cnudde, V., Vansteelandt, S. & Carò, F. 2007. An evalua- Investigating the effects of organic and conventional management on
tion of 2D-image analysis techniques for measuring soil microporosity. soil aggregate stability using X-ray computed tomography. European
European Journal of Soil Science, 58, 133–140. Journal of Soil Science, 60, 360–368.
© 2013 The Authors

Perret, J., Prasher, S.O., Kantzas, A. & Langford, C. 1999. Three- Sleutel, S., Cnudde, V., Masschaele, B., Vlassenbroek, J., Dierick, M., Van
dimensional quantification of macropore networks in undisturbed soil Hoorebeke, L., et al . 2008. Comparison of different nano- and micro-
cores. Soil Science Society of America Journal , 63, 1530–1543. focus X-ray computed tomography set-ups for the visualization of the
Perret, J.S., Al-Belushi, M.E. & Deadman, M. 2007. Non-destructive soil microstructure and soil organic matter. Computers and Geosciences,
visualization and quantification of roots using computed tomography. 34, 931–938.
Soil Biology and Biochemistry, 39, 391–399. Spanne, P., Jones, K.W., Prunty, L. & Anderson, S.H. 1994. Potential
Peth, S., Nellesen, J., Fischer, G. & Horn, R. 2010. Non-invasive 3D applications of synchrotron computed microtomography to soil science.
analysis of local soil deformation under mechanical and hydraulic In: Tomography of Soil-Water-Root Processes, Special Publication No
stresses by mu CT and digital image correlation. Soil and Tillage 36 (eds S.H. Anderson & J.W. Hopmans) pp. 43–57. Soil Science
Research, 111, 3–18. Society of America, Madison.
Petrovic, A.M., Siebert, J.E. & Rieke, P.E. 1982. Soil bulk density analysis Stock, S.R. 2008. Recent advances in X-ray microtomography applied to
in three dimensions by computed tomographic scanning. Soil Science
materials. International Materials Reviews, 53, 129–181.
Society of America Journal , 46, 445–450.
Sun, W., Brown, S. & Leach, R. 2012. An Overview of Industrial X-ray
Pierret, A., Moran, C.J. & Pankhurst, C.E. 1999a. Differentiation of soil
Computed Tomography. National Physical Laboratory Report No 32,
properties related to the spatial association of wheat roots and soil
National Physical Laboratory, Teddington.
macropores. Plant and Soil , 211, 51–58.
Taina, I.A., Heck, R.J. & Elliot, T.R. 2008. Application of X-ray computed
Pierret, A., Capowiez, Y., Moran, C.J. & Kretzschmar, A. 1999b. X-
tomography to soil science: a literature review. Canadian Journal of Soil
ray computed tomography to quantify tree rooting spatial distributions.
Science, 88, 1–19.
Geoderma, 90, 307–326.
Pierret, A., Capowiez, Y., Belzunces, L. & Moran, C.J. 2002. 3D Thieme, J., Schneider, G. & Knöchel, C. 2003. X-ray tomography of
reconstruction and quantification of macropores using X-ray computed a microhabitat of bacteria and other soil colloids with sub-100 nm
tomography and image analysis. Geoderma, 106, 247–271. resolution. Micron, 34, 339–344.
Pierret, A., Kirby, M. & Moran, C. 2003. Simultaneous X-ray imaging of Tippkötter, R., Eickhorst, T., Taubner, H., Gredner, B. & Rademaker,
plant root growth and water uptake in thin-slab systems. Plant and Soil , G. 2009. Detection of soil water in macropores of undisturbed soil
255, 361–373. using microfocus X-ray tube computerized tomography (μCT). Soil and
Pierret, A., Doussan, C., Capowiez, Y., Bastardie, F. & Pagès, L. 2007. Tillage Research, 105, 12–20.
Root functional architecture: a framework for modeling the interplay Tollner, E.W., Rasmseur, E.L. & Murphy, C. 1994. Techniques and
between roots and soil. Vadose Zone Journal , 6, 269–281. approaches for documenting plant root development with X-ray
Pires, L.F., Bacchi, O.O.S. & Reichardt, K. 2007. Assessment of computed tomography. In: Tomography of Soil-Water-Root Processes,
soil structure repair due to wetting and drying cycles through 2D Special Publication No 36 (eds S.H. Anderson & J.W. Hopmans) pp.
tomographic image analysis. Soil and Tillage Research, 94, 537–545. 115–133. Soil Science Society of America, Madison.
Pires, L.F., Borges, J.A.R., Bacchi, O.O.S. & Reichardt, K. 2010. Twenty- Torrance, J.K., Elliot, T., Martin, R. & Heck, R.J. 2008. X-ray computed
five years of computed tomography in soil physics: a literature review tomography of frozen soil. Cold Regions Science and Technology, 53,
of the Brazilian contribution. Soil and Tillage Research, 110, 197–210. 75–82.
Prodanović, M., Lindquist, W.B. & Seright, R.S. 2006. Porous structure Tracy, S.R., Black, C.R., Roberts, J.A., Sturrock, C., Mairhofer, S.,
and fluid partitioning in polyethylene cores from 3D X-ray micro- Craigon, J., et al . 2012a. Quantifying the impact of soil compaction
tomographic imaging. Journal of Colloid and Interface Science, 298, on root system architecture in tomato (solanum lycopersicum) by X-ray
282–297. micro-computed tomography. Annals of Botany, 110, 511–519.
Quinton, W.L., Elliot, T., Price, J.S., Rezanezhad, F. & Heck, R. Tracy, S.R., Black, C., Roberts, J., McNeill, A., Davidson, R., Tester,
2009. Measuring physical and hydraulic properties of peat from X-ray M., et al . 2012b. Quantifying the effect of soil compaction on three
tomography. Geoderma, 153, 269–277.
varieties of wheat (triticum aestivum L.) using X-ray micro computed
Rogasik, H., Crawford, J.W., Wendroth, O., Young, I.M., Joschko, M.
tomography (CT). Plant and Soil , 353, 195–208.
& Ritz, K. 1999. Discrimination of soil phases by dual energy X-ray
Van Den Bulcke, J., Boone, M., van Acker, J. & van Hoorebeke, L. 2009.
tomography. Soil Science Society of America Journal , 63, 741–751.
Three-dimensional X-ray imaging and analysis of fungi on and in wood.
Rogasik, H., Onasch, I., Brunotte, J., Jegou, D. & Wendroth, O. 2003.
Microscopy and Microanalysis, 15, 395–402.
Assessment of soil structure using X-ray computed tomography. In:
Van Dusschoten, D. 2011. Quantitative 3D root imaging in natural-like soil
Applications of X-ray Computed Tomography in the Geosciences (eds F.
by MRI. In: Rhizosphere 3 International Conference (ed. H. Lambers).
Mees, R. Swennen, M. Van Geet & P. Jacobs), pp. 151–165. Geological
Society Special Publication, London. Perth.
Schmidt, S., Bengough, A.G., Gregory, P.J., Grinev, D.V. & Otten, W. Van Geet, M., Swennen, R. & Wevers, M. 2000. Quantitative analysis
2012. Estimating root–soil contact from 3D X-ray microtomographs. of reservoir rocks by microfocus X-ray computerised tomography.
European Journal of Soil Science, 63, 776–786. Sedimentary Geology, 132, 25–36.
Schrader, S., Rogasik, H., Onasch, I. & Jegou, D. 2007. Assessment of soil Vaz, C.M.P., De Maria, I.C., Lasso, P.O. & Tuller, M. 2011. Evaluation
structural differentiation around earthworm burrows by means of X-ray of an advanced benchtop micro-computed tomography system for
computed tomography and scanning electron microscopy. Geoderma, quantifying porosities and pore-size distributions of two Brazilian
137, 378–387. oxisols. Soil Science Society of America Journal , 75, 10.
Seibert, J.A. & Boone, J.M. 2005. X-ray imaging physics for nuclear Vontobel, P., Lehmann, E.H., Hassanein, R. & Frei, G. 2006. Neutron
medicine technologists. Part 2: X-ray interactions and image formation. tomography: method and applications. Physica B: Condensed Matter,
Journal of Nuclear Medicine Technology, 33, 3–18. 385, 475–480.
© 2013 The Authors

Wang, W., Kravchenko, A.N., Smucker, A.J.M. & Rivers, M.L. 2011. Yang, W.N., Xu, X.C., Duan, L.F., Luo, Q.M., Chen, S.B., Zeng, S.Q.,
Comparison of image segmentation methods in simulated 2D and 3D et al . 2011. High-throughput measurement of rice tillers using a
microtomographic images of soil aggregates. Geoderma, 162, 231–241. conveyor equipped with X-ray computed tomography. The Review of
Wang, W., Kravchenko, A.N., Smucker, A.J.M., Liang, W. & Rivers, Scientific Instruments, 82, 025102–025107.
M.L. 2012. Intra-aggregate pore characteristics: X-ray computed Young, I.M. 1998. Biophysical interactions at the root-soil interface: a
microtomography analysis. Soil Science Society of America Journal , 76, review. Journal of Agricultural Science, 130, 1–7.
1159–1171. Young, I.M. & Crawford, J.W. 2004. Interactions and self-organization in
Wang, Y., De Carlo, F., Mancini, D.C., McNulty, I., Tieman, B., the soil-microbe complex. Science, 304, 1634–1637.
Bresnahan, J., et al . 2001. A high-throughput X-ray microtomography Young, I.M., Crawford, J.W. & Rappoldt, C. 2001. New methods and
system at the advanced photon source. The Review of Scientific models for characterising structural heterogeneity of soil. Soil and
Instruments, 72, 2062–2068. Tillage Research, 61, 33–45.
Wildenschild, D., Hopmans, J.W., Vaz, C.M.P., Rivers, M.L., Rikard, D. Young, I.M., Crawford, J.W., Nunan, N., Otten, W. & Spiers, A. 2008.
& Christensen, B.S.B. 2002. Using X-ray computed tomography in Microbial distribution in soils: physics and scaling. In: Advances
hydrology: systems, resolutions, and limitations. Journal of Hydrology, in Agronomy, Volume 100 (ed. D.L. Sparks) pp. 81–121. Elsevier
267, 285–297. Academic Press Inc, San Diego, CA.
Wildenschild, D., Armstrong, R.T., Herring, A.L., Young, I.M. & William Zeng, Y., Gantzer, C.J., Payton, R.L. & Anderson, S.H. 1996. Fractal
Carey, J. 2011. Exploring capillary trapping efficiency as a function dimension and lacunarity of bulk density determined with X-ray
of interfacial tension, viscosity, and flow rate. Energy Procedia, 4, computed tomography. Soil Science Society of America Journal , 60,
4945–4952. 1718–1724.
© 2013 The Authors


Xray Computed Tomography

Uploaded by

Copyright:

Available Formats

Xray Computed Tomography

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Xray Computed Tomography

Uploaded by

Copyright:

Available Formats

European Journal of Soil Science, June 2013, 64, 279–297 doi: 10.1111/ejss.

Applications of X-ray computed tomography

Introduction ordered in a complex and dynamic hierarchical manner, advances

© 2013 The Authors

© 2013 The Authors

© 2013 The Authors

© 2013 The Authors

when performed with different image analysis protocols. The need

Using X-ray computed tomography in soil studies

© 2013 The Authors

© 2013 The Authors

with thin sections 9800)

Gaps represent unreported data in the literature.

© 2013 The Authors

morphology successfully (Orsi & Anderson, 1995). Later, Géraud

© 2013 The Authors

© 2013 The Authors

or earthworm burrows must either be removed or filled with a

© 2013 The Authors

© 2013 The Authors

© 2013 The Authors

© 2013 The Authors

© 2013 The Authors

© 2013 The Authors

© 2013 The Authors

© 2013 The Authors

© 2013 The Authors

© 2013 The Authors

You might also like