Soinne et al. (2023). Effects of organic carbon and clay contents on structure‐related properties of arable soils with high clay content

|

Riikka Keskinen

|

Mika Tähtikarhu

|

Jukka Kuva

Creative Commons Attribution

|

Jari Hyväluoma

2,4

1

Natural Resources Institute Finland

Luke, Helsinki, Finland

2

Natural Resources Institute Finland

Luke, Jokioinen, Finland

3

Geological Survey of Finland GTK,

Espoo, Finland

4

Häme University of Applied Sciences

HAMK, Mustiala, Finland

Correspondence

Helena Soinne, Natural Resources
Institute Finland Luke, Latokartanonkaari
9, FI-00790 Helsinki, Finland.
Email:

helena.soinne@luke.fi

Funding information

Academy of Finland via the RAMI
infrastructure project, Grant/Award
Number: 293109; CSC-IT Center for
Science through project, Grant/Award
Number: 2005253; Maa- ja Vesitekniikan
Tuki Ry; Niemi-säätiö; Salaojituksen
Tukisäätiö sr; Uusimaa Centre for
Economic Development, Transport and
the Environment through Rural
Development Programme for Mainland
Finland 2014-2020

Abstract

Understanding of factors governing soil structural features is necessary for

managing key processes affecting crop productivity and environmental impacts

of agriculture, for example, soil water balance, aeration, and root penetration.

Organic matter is known to act as a major binding agent in soil aggregation

and thus constitutes a central pillar in soil structure formation. However,

knowledge of the structural role of organic matter or carbon (OC) in soils

highly rich in clay-sized particles (<0.002 mm) is limited. In this study, the

effects of clay and OC contents on aggregate stability, water holding capacity,

near-saturated hydraulic conductivity, total porosity, and pore size distribution

were assessed in cultivated fields with high clay content located in private crop

production farms in southern Finland. Significant positive correlations were

found between OC content and proportion of water stable aggregates and spe-

cific pore sizes from the range of 30

μ

m up to 1 mm diameter determined by

image analysis. Porosities on a smaller size range derived from water retention

measurements likewise showed a positive correlation with OC in <0.2

μ

m sizes.

On the range of 0.2

–

1

μ

m, a negative relationship was observed, which induced

a negative effect of OC on soil plant available water reserves. In line with the

positive correlation between OC and larger soil pores, free water, representing

the amount of water that can be drained by gravity, exhibited a positive relation-

ship with OC suggesting that OC content can enhance aeration of soils with

high clay content. Compared to OC, clay content tended to have an adverse

effect on soil structural properties. Clay correlated negatively with pores larger

than 30

μ

m, free water content, and extrapolated field saturated hydraulic

conductivity. Further, our imaging results showed how saturated hydraulic

conductivity was controlled by pore morphology, and there was a power law

relationship between the conductivity and critical pore diameter.

K

/

d

2

c

in

agreement with the percolation theory. Overall, the structural impacts and

Received: 3 May 2023

Revised: 27 September 2023

Accepted: 1 October 2023

DOI: 10.1111/ejss.13424

This is an open access article under the terms of the

License, which permits use, distribution and reproduction in any medium, provided

the original work is properly cited.

European Journal of Soil Science

published by John Wiley & Sons Ltd on behalf of British Society of Soil Science.

Eur J Soil Sci.

2023;74:e13424.

wileyonlinelibrary.com/journal/ejss

1 of 16

https://doi.org/10.1111/ejss.13424

hydrological implications of OC and clay in heavy clay soils vary by pore size

ranges and their emergent practical impacts are thus not straightforward.

K E Y W O R D S

aggregate stability, hydraulic conductivity, plant available water, pore size distribution, pore
volume, texture, water retention, X-ray tomography

1

|

I N T R O D U C T I O N

Structural features exert an overarching effect on soil
properties through impacts on water balance, aeration,
root penetration, soil biota, and dynamics of nutrient
cycling (Dexter,

1988

; Rabot et al.,

). Considering soil

functioning, the sizes, spacings, continuity and tortuosity
of the voids, that is, the pore space intervening the solid
phase, is decisive (Pagliai & Vignozzi,

2002

). Moreover,

different-sized pores have different implications regard-
ing soil functions. The smallest of pores (<0.5

μ

m) medi-

ate interparticle forces and serve as nutrient storages,
while pores in the size range of 0.5

–

50

μ

m hold water

against gravity but in plant available state, whereas the
pores with diameters greater than 50

μ

m transmit air and

in wet conditions drain water (Greenland,

1977

). The

pores larger than 300

μ

m contribute to the fast non-

equilibrium water and solute transport and have thus
impacts on the drainage and transport of substances
(Jarvis,

). Optimal soil water balance, encompassing

minimal surface runoff, efficient infiltration and trans-
mission, adequate water storage and sufficient drainage,
necessitates a continuous pore system with variation in
pore sizes (Nimmo,

2013

In clay soils rich in mineral particles of less than

0.002 mm in diameter, flocculation of the primary parti-
cles into aggregates is essential for constructing a func-
tional porous matrix of otherwise a massive, dense
structure (Horn et al.,

1994

; Qadir & Schubert,

2002

).

Individual soil particles are bound into aggregates by
electrostatic forces through ionic bridging, by organo-
mineral associations, and by cementation via hydroxides,
carbonates, and disordered aluminosilicates (Bronick &
Lal,

2005

; Tisdall & Oades,

1982

). Excluding oxide-

dominated soils, organic matter constitutes the major
binding agent as a part of clay

–

polyvalent cation

–

organic

matter

complexes

(Oades

&

Waters,

1991

;

Six

et al.,

2004

).

Considering soil water dynamics, water stability of

macroaggregates is crucial since slaking into microaggre-
gates or dispersion into primary particles may cause
blocking of pores transmitting and storing water
(Le Bissonnais & Arrouays,

1997

). Soil organic matter

contents and organic inputs are known to exhibit a

positive relationship with aggregate stability, which
derives from enhanced cohesion within aggregates
and/or increased hydrophobicity (Abiven et al.,

Chaney & Swift,

; Chenu et al.,

). The relation-

ship is dependent on the clay content of the soil such that
the higher the clay content, the higher the amount of soil
organic matter or soil organic carbon (OC) needed for
achieving an equal effect (Feller & Beare,

1997

; Johannes

et al.,

). Consequently, the ratio of OC to clay has

been suggested as a useful indicator in soil structural
assessments (Johannes et al.,

; Soinne et al.,

).

In addition to natural factors modifying soil aggrega-

tion and porosity (e.g. texture, mineralogy, weathering,
wet-dry or freeze

–

thaw cycles and vegetation), the stabil-

ity of aggregates and, therefore, the ability of soil to retain
its structure under stress conditions (Amézketa,

1999

) are

affected by management practices including tillage, lim-
ing, and fertilization (Bronick & Lal,

2005

). For soil

aggregates, mechanical stress induced by the tillage
equipment is known to be deleterious and aggregates in
tilled soils are typically less stable than their counterparts
in no-till soils (e.g. Abdollahi et al.,

2014

; Nunes

et al.,

). Furthermore, fertilization, liming, soil condi-

tioning or irrigation induced changes in the composition
and/or concentration of electrolytes in the soil solution
can either improve or impair soil structure (Abu-Sharar
et al.,

1987

; Ding et al.,

2019

; Dontsova & Norton,

2001

).

The role of OC in supporting soil functioning is well

established, but recently more detailed information has

Highlights

•

Knowledge on the role of organic carbon in the
structure of soils highly rich in clay is limited

•

The structural impacts of carbon and clay con-
tent were opposite and varied by pore size
range

•

Hydraulic conductivity was controlled by criti-
cal pore diameter both decreasing with increas-
ing clay

•

Organic carbon enhanced soil aeration but was
not beneficial for plant available water in clay
soils

2 of 16

SOINNE

ET AL

.

13652389, 2023, 5, Downloaded from https://bsssjournals.onlinelibrary.wiley.com/doi/10.1111/ejss.13424 by Hayley Peter-Contesse , Wiley Online Library on [18/10/2026]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License

been published related to the effects of OC on water hold-
ing capacity in soils with varying texture (Minasny &
McBratney,

). However, the effect of OC on soil pore

structure and pore size distribution has rarely been stud-
ied in heavy clay soils, where formation of secondary
structure with pores of different sizes is controlling the
flow of water and substances (e.g. Jarvis,

; Turunen

et al.,

2013

). Moreover, the structural properties are often

studied with indirect methods (such as water retention
curves), whereas X-ray microtomography could provide
more direct information on the morphology of the struc-
tural features and their changes. In addition to the pore
size distribution, other features of pore structure such as
pore connectivity can practically set key controls on the
flow of water and substances in a structured soil
(e.g. Koestel et al.,

Therefore, the aim of this study was to assess the

effects of clay and OC contents of cultivated clay soils on
key structure-related properties: aggregate stability, water
holding capacity, near-saturated hydraulic conductivity,
total porosity, and pore size distribution. Emphasis was
given on soils with high clay content and real-farm condi-
tions, of which little information is available from previ-
ous studies. We hypothesized that increase in OC affects
the observed structural attributes even in the high range
of OC contents (2.2%

–

8.6% C in 0

–

10 cm surface layer)

found in these high-clay soils.

2

|

M A T E R I A L S A N D M E T H O D S

2.1

|

Study sites, soil sampling, and

analyses

The studied fields were confined to 19 cooperating pri-
vate crop production farms located in the Uusimaa
region in southern Finland. The area is dominated by
post-glacial clays formed subaquatically during the vari-
ous stages of the Baltic Sea following the Weichselian
deglaciation (Koljonen,

1992

). Mineralogically, the clay-

sized fraction of these soils typically contains a consider-
able proportion of fine ground rock-forming minerals (K-
feldspar, plagioclase, and quartz) in addition to actual
clay minerals (chlorite, vermiculite, smectite, and kaolin-
ite) (Keskinen et al.,

In 2020, soil samples from two field sites from each

farm were collected for measuring basic soil characteris-
tics and aggregate stability. Only limited information was
available on the management history of the sampled
fields. Of the 38 fields, 12 were currently under no-till
management and three fields were regularly ploughed in
the autumn. In majority of the fields (23), the prevailing
tillage method was reduced tillage, where the plant

residues are mixed in the topmost 10 cm. All farms were
crop production farms indicating infrequent or non-
existing manure applications. Soil samples were taken
within a GPS-positioned (10 cm accuracy) equilateral tri-
angle with a 5 m side width. For textural analysis, one
composite sample per site was compiled by bulking
ca. 10 subsamples taken from the depth of 0

–

20 cm below

the surface. The samples were air-dried and ground to
pass through a 2-mm sieve and the proportions of sand,
silt, and clay were determined by the pipette method of
Elonen (

1971

). Furthermore, soil pH and electrical con-

ductivity (EC) were measured from soil

–

water suspen-

sions (1:2.5). For total bulk density (BD) and OC analysis,
three individual soil cores of 4.7 cm diameter were col-
lected per site in 10 cm segments to 20 cm depth. Each
soil core segment of known volume was dried at

+

37

C

and weighed. Samples were further sieved (2 mm) and
analysed for total C by dry combustion (Leco TruMac
CN). In these acidic soils, the total C can be taken to be
equal to OC (Nelson & Sommers,

1996

).

Samples for analysis of aggregate stability were com-

posed of three 200 cm

3

cylinders of soil taken from the

5-cm surface layer and gently emptied by hand to form
one composite sample per site. Aggregate stability was
assessed with the wet sieving method (Heikkinen et al.,

2019

) using an Eijkelkamp 08.13 equipped with eight

0.25 mm mesh size sieves. Four grams of air-dried
2

–

5 mm aggregates were weighed for the analysis and

sieves containing the samples were immersed in water
(100 mL) for 15 min prior to a 3-min up-and-down run
with the wet sieving apparatus. Material remaining on
the sieve after wet sieving was air-dried and the share of
water stable aggregates (WSA%) was calculated by divid-
ing the weight of the material on the sieves after subtract-
ing the mass of >0.2 mm particle size fractions by the
total weight of the material. The suspension containing
soil material disintegrated during the wet sieving was
transferred into a 100 mL container and left to settle for
21 h. After settling, samples for turbidity measurement
were taken with a pipette from the surface of the suspen-
sion. Turbidity (nephelometric turbidity units, NTU),
indicating the risk for colloid dispersion, was determined
using a HACH 2100AN IS turbidimeter (Hach Company).
Wet sieving was carried out on three replicates per one
composite sample.

From the total set of 38 sites, ten sites were selected

for measurement of soil water holding capacity (2.2.) and
a further subset of 5 was picked for the more laborous
analysis of tension infiltrometry (2.3.) and X-ray tomogra-
phy (2.4.) (Figure

, Table

). The selection aimed at vari-

ation in OC content. No significant correlations between
clay content and OC content were found in the sample
sets. The sites selected for water holding capacity

SOINNE

ET AL

.

3 of 16

13652389, 2023, 5, Downloaded from https://bsssjournals.onlinelibrary.wiley.com/doi/10.1111/ejss.13424 by Hayley Peter-Contesse , Wiley Online Library on [18/10/2026]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License

analysis, X-ray tomography, and near-saturated hydraulic
conductivity measurements included fields with no-till
and fields where reduced tillage was practiced. To reduce
the impact of these management practices, the samples
were taken below the mixed layer (below the depth
of 10 cm).

2.2

|

Soil water holding capacity

For analysis of soil water holding capacity, six undis-
turbed soil samples per site were collected in 200 cm

3

cyl-

inders at a depth of ca. 10

–

15 cm. The soil moisture

characteristic curves were determined using 100, 300,
and 500 kPa ceramic plates in pressure extractors
(Soilmoisture Equipment Corp., Goleta, CA). Before the
measurements, samples were saturated for 14 days.
The soil moisture characteristic curves were determined

as drying curves by weighing the samples after equilibra-
tion in the following suctions: 0.1, 0.2, 1.0, 10, 31,
100, and 310 kPa. The equilibration times varied between
3 and 21 days depending on the suction. The highest suc-
tion (wilting point 1600 kPa) was produced osmotically
by vapour pressure equilibrium with saturated ammo-
nium oxalate solution. Both vertical and horizontal
shrinkages of samples were measured at each suction
using a Vernier calliper and a feeler gauge, respectively.
The moisture content of soil samples was calculated as a
quotient of the water volume in the sample by the vol-
ume of the shrunken sample. In the following, plant
available water (PAW) refers to the amount of water
between suctions 10 and 1600 kPa, readily plant available
water (RPAW) between 10 and 310 kPa and free water
(FW) is the amount of water below suction 10 kPa.
Throughout the article, we use the sign convention where
suction is defined as negative of the matric potential. To

F I G U R E 1

Soil organic

carbon content (OC, %) in
(a) 0

–

10 cm and (b) 10

–

20 cm

soil layers plotted against clay

content (%). All samples were
analysed for aggregate stability,

darker grey in diamonds and

triangles indicates samples
included in water holding

capacity analysis and triangles

indicate samples included in
X-ray tomography and

near-saturated hydraulic

conductivity measurements.

T A B L E 1

Characteristics of the

soils (pH; electric conductivity, EC;
texture; organic carbon content, OC;

bulk density, BD) included in the water

holding capacity analysis, and X-ray
tomography and near-saturated

hydraulic conductivity measurements

(Inf. XR)

Soil

pH

EC

Clay

Silt

Sand

OC

BD

0

–

20 cm

0

–

20 cm

0

–

10 cm

10

–

20 cm

10

–

20 cm

Inf.

mS m

-1

%

g cm

3

XR

1

6.5

21

74

14

12

4.7

3.8

1.4

2

6.3

16

38

42

20

2.3

2.2

1.6

3

6.0

12

69

14

17

2.6

2.5

1.3

x

4

6.2

9

55

32

13

3.6

2.9

1.5

x

5

6.4

10

53

29

18

3.3

2.7

1.5

6

5.8

8

62

31

7

8.6

7.6

1.1

x

7

5.9

8

54

34

12

4.3

3.8

1.3

x

8

5.4

7

77

16

7

6.5

6.3

1.2

9

6.2

17

65

24

11

4.5

1.3

x

10

6.3

13

74

12

14

3.6

1.4

Note

: The bulk densities (BD) from 10 to 20 cm layer correspond to the depth of structural samples.

4 of 16

SOINNE

ET AL

.

13652389, 2023, 5, Downloaded from https://bsssjournals.onlinelibrary.wiley.com/doi/10.1111/ejss.13424 by Hayley Peter-Contesse , Wiley Online Library on [18/10/2026]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License

derive pore size distribution from water retention curves,
the suction values can be converted to equivalent pore
diameters d by using the Young-Laplace equation
d

=

4

γ

/p, where

γ

is the surface tension of water and p is

the suction.

2.3

|

Tension infiltrometry

Near-saturated hydraulic conductivities were measured
on site with tension infiltrometers manufactured by Soil
Measurement Systems (Huntington Beach, CA) similar
to that described by Ankeny et al. (

1988

). The used

devices consist of a porous disk with a diameter of 20 cm,
a separate water reservoir, and bubble towers. At each
measurement point, the vegetation was cut, and the soil
surface was carefully levelled and smoothed. To ensure
good contact between the disk and the soil surface, a thin
layer of moist fine sand was added to the soil surface in
an area equal to the infiltration disk. Supply pressure
heads of

4,

2, and

1 cm were used in an ascending

sequence. Measurement at each supply pressure head
was continued until a steady-state infiltration was
reached. The hydraulic conductivity was calculated with
the method described by Ankeny et al. (

1991

) based on

Wooding's

theory

of

unconfined

three-dimensional

infiltration (Wooding,

1968

). Field saturated hydraulic

conductivity K

fs

was calculated by extrapolating the near-

saturated hydraulic conductivity values to zero supply
pressure (see, e.g. Jarvis & Messing,

1995

). At each mea-

surement site four measurements were done.

2.4

|

X-ray tomography and flow

simulations

For imaging, intact soil cores were sampled in alumin-
ium cylinders (inner diameter 46 mm and height
40 mm). From each sampling location, five samples were
collected.

The

sampling

depth

was

approximately

10

–

15 cm below the soil surface. The collected soil

samples were wrapped with plastic film to preserve their
natural soil moisture and stored in

+

5

C until imaging.

The 3D X-ray microtomography imaging was con-

ducted at Geological Survey of Finland (GTK) with a GE
Phoenix v

j

tome

j

x s device (General Electric) using a

240 kV microfocus tube. An accelerating voltage of
100 kV and a tube current of 350

μ

A were used, which

resulted in a total power 35.0 W. No beam filter was used
in the scans. The number of projections for each 3D
image was 2500 during a full rotation of 360

. Exposure

time was set to 250 ms. At each angle, the detector waited
for one exposure time and then took an average over

three exposures. The 3D images had a voxel size of
30.0

μ

m and the dimensions were 2024

2024

voxels. Each soil sample was imaged with the same set-
tings. The resulting radiographs were converted to gray-
scale image stacks (16-bit) with the phoenix datos

j

x

software.

The greyscale images were converted to binary form

for quantitative analyses of pore space properties, which
was done in several steps. The greyscale values of images
were calibrated, and the images converted to 8-bit
images using the greyscale values for the aluminium cyl-
inder

and

air

as

reference

densities

(cf.

Larsbo

et al.,

2014

; Turunen et al.,

2019

). Denoising of the grey-

scale images was done using the variance-weighted mean
filter with a radius of 5 voxels (Gonzalez & Woods,

;

Turpeinen,

2015

). Images were then segmented to pores

and solid voxels using Otsu's segmentation method
(Otsu,

1979

) and the resulting binary image was further

filtered with a majority filter (radius 1 voxel). Characteri-
zation of pore structure was performed using the result-
ing binary samples. A region of interest (ROI) was
selected for subsequent analyses such that possible dis-
turbances caused by soil sampling were excluded from
the analyses. The diameter and height of the cylinder-
shaped ROI were 40 and 20 mm, respectively, such that
sample volume close to cylinder walls as well as top and
bottom of the samples were excluded.

Binarized images were analysed for porosity, pore size

distribution, and critical pore diameter in the vertical
direction. Porosity is the fraction of pore volume in the
ROI and was obtained as the quotient between the num-
ber of pore voxels and the total number of voxels in the
ROI, that is, porosity is a dimensionless quantity ranging
between 0 and 1. Pore size distribution was calculated by
successive application of a morphological opening opera-
tion (Horgan,

1998

) with increasing diameter of the struc-

turing element. The critical pore diameter (D

c

) is defined

as the diameter of the largest sphere that can pass
through the pore space in vertical direction (e.g. Koestel
et al.,

). The determination of D

c

was based on the

distance transform of the pore spaces, that is, labelling
each pore voxel with distance to the closest pore wall.
This was calculated using the weighted Chamfer dis-
tances d

3,4,5,7

with a scaling procedure as described by

Svensson and Borgefors (

2002

). Chamfer distances

approximate

the

Euclidean

distances

with

integer

weights, and they can be calculated at lower computa-
tional cost than exact Euclidean distances.

Transport properties of the imaged soil samples were

determined by simulating a fluid flow through the
imaged pore system. The pore-scale steady-state flow field
for saturated pressure-driven flow was determined with
the lattice Boltzmann method, which operates on a

SOINNE

ET AL

.

5 of 16

13652389, 2023, 5, Downloaded from https://bsssjournals.onlinelibrary.wiley.com/doi/10.1111/ejss.13424 by Hayley Peter-Contesse , Wiley Online Library on [18/10/2026]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License

uniform cubic lattice, whereby the segmented image
obtained by X-ray tomography was directly used as the
computational grid in order to describe the complex
pore geometry of soil. We used the D3Q19 lattice Boltz-
mann model with the two-relation time

–

time collision

operator (Ginzburg et al.,

) and the bounce-back

boundary conditions to enforce no-slip flow at the pore
walls. The flow through the sample was driven by using
a forcing term in the vertical direction. The material
intrinsic permeability of the soil (k) was determined
from the obtained steady-state flow field using Darcy's
law which describes creeping flow through porous
media as

q

z

¼

k

μ

∂

p

∂

z

,

where

q

z

is the volumetric flux,

μ

the dynamic viscosity,

and

∂

p

/

∂

z

the pressure gradient in the vertical (flow)

direction. Saturated hydraulic conductivity K

sim

for water

is then obtained from permeability as K

sim

=

k

ρ

g

/

μ

,

where

ρ

is the density of water and

g

the gravitational

acceleration.

2.5

|

Statistical analysis

Some structure-related parameters were only measured
for the smaller set of soils. As these parameters were not
normally distributed, statistical correlations between
them and the OC and clay content of soil were deter-
mined using the Spearman rank correlation coefficient

ρ

.

In these analyses, each measurement was treated as an
individual paired observation regardless of them being
from same or different field site. The observed trends are
generalizable due to the sampling setup containing
uncontrollable variation in soil management. However,
quantitativeness is not achieved due to unfilled underly-
ing assumptions for regression analysis. A significance
level of

α

=

0.05 was used in all analyses. These statistical

analyses were performed with Python using the Stats
module in the SciPy library (Virtanen et al.,

) and in

the R environment (version 4.2.1).

3

|

R E S U L T S

3.1

|

Aggregate stability and BD

The proportion of WSA in the 0

–

5 cm surface soil layer

varied from 43% to 91% and correlated positively with OC
and clay contents (Figure

, Table

). Turbidity values

(NTU) correlated negatively with OC and positively with

clay/OC ratio indicating lower clay-sized particle disper-
sion in soils with a higher OC content (Table

, Figure

BD correlated negatively with OC content (

ρ

=

0.77,

p

< 0.001) but no correlation with clay content was

found.

3.2

|

Imaging

Three-dimensional visualizations of imaged soil samples
from five sampling sites are shown in Figure

. Overall,

visually the increasing OC content appeared to change
the pore structure towards relatively closely located,
connected, and tortuous pore space. Spearman's rank
correlations of structure-related quantities determined
by image analysis with soil carbon and clay content
are shown in Table

. In Figure

4

, image analysis

results for all replicate samples are shown as functions
of OC and clay contents. Notice that all imaging
results considered refer to macroporosity visible in
X-ray tomography images with the used imaging reso-
lution (30

μ

m).

All structure-related quantities (i.e. porosity, median

pore size, critical pore diameter) determined by image
analysis had a negative correlation with clay content
(Table

). The porosity quantified by X-ray tomography

ranged between 0.01 and 0.15 and showed an ascending
trend with respect to OC content. Median pore sizes did
not show a clear dependence on OC. Generally, OC con-
tent seemed to have less significant effects on structural
properties than clay content. Clay/OC ratio, in turn, had
the same statistically significant, but negative, correla-
tions as OC.

Considering the pore size distributions, pore sizes

ranged up to 5 mm, but typically the largest pores were
ca. 2 mm except for a couple of samples. Pore size distri-
butions for imaged soils are given in Table

(

The

correlation

coefficients

of

selected pore size classes determined by image analysis
are shown in Table

4

. A positive correlation between OC

content and porosity was observed for pore sizes up to
1 mm. Clay content, in turn, showed a negative correla-
tion in all pore size classes except for the pore size range
from 200 to 600

μ

m. All pore size classes had a negative

correlation with the Clay/OC ratio.

While there was a lot of variation between samples in

the above-mentioned structural properties, there was
even more variation for critical pore diameter which ran-
ged from 30 to 1140

μ

m. The simulated hydraulic con-

ductivity also had remarkable variation from 1 to
1700 cm h

1

. The simulated hydraulic conductivities

were found to relate to the critical pore diameter with a
power law dependence

6 of 16

SOINNE

ET AL

.

13652389, 2023, 5, Downloaded from https://bsssjournals.onlinelibrary.wiley.com/doi/10.1111/ejss.13424 by Hayley Peter-Contesse , Wiley Online Library on [18/10/2026]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License

K

sim

/

d

β

c

with exponent

β

=

2.0 (Figure

5

).

3.3

|

Soil hydraulic properties

For near-saturated hydraulic conductivity measurements,
three pressure heads were used and the determined
values ranged from 0.2 to 2.1 mm h

1

at a

4 cm pres-

sure head, from 0.60 to 9.5 mm h

1

at a

3 cm pressure

head,

and from 2.7 to 31 mm h

1

at a

1 cm

pressure head. OC content did not have a clear impact on
hydraulic conductivity at any of the used pressure heads
(Table

5

). Increasing clay content, on the other hand,

showed a positive correlation with hydraulic conductivity
at a pressure head of

4 cm. Field saturated hydraulic

conductivities were determined by extrapolation from
near-saturated hydraulic conductivities and these results
showed a negative trend with increasing clay content,
while OC content did not have a noticeable effect on this
property.

PAW, RPAW, and FW were derived from soil water

retention measurements. Regarding dependence on OC
content, PAW showed a declining trend whereas the
opposite was found for FW. For the clay content, in turn,
only statistically significant correlation was observed for
FW, which declined with increasing clay content
(Figure

6

).

Porosities derived from water retention curves corre-

lated positively both with OC and clay contents (Table

6

For pore size class <0.2

μ

m, porosities correlated posi-

tively both for OC and clay contents. Porosities in the

T A B L E 2

Spearman's rank correlation coefficients for water

stable aggregates (WSA, %) and turbidity (NTU) indicating the
amount of dispersible clay-sized particles.

OC

Clay

Clay/OC

ρ

p

ρ

p

ρ

p

WSA, %

0.40

0.013

0.71

0.000

0.24

0.138

NTU

0.43

0.007

0.03

0.841

0.33

0.043

Note

: Statistically significant Spearman rank correlations are bolded.

F I G U R E 2

Share of water

stable aggregates (WSA, %)
plotted against (a) soil organic

carbon (OC, %) and (b) clay

content (Clay, %), and an
indicator of the risk for colloid

dispersion, turbidity (NTU g

1

),

plotted against (c) OC and
(d) Clay/OC-ratio. Darker grey

in diamonds and triangles

indicates samples included in
water holding capacity analysis

and triangles indicate samples

included in X-ray tomography
and near-saturated hydraulic

conductivity measurements. The

arrows indicate the statistically
significant Spearman rank

correlations and act as a guide

for the eye.

SOINNE

ET AL

.

7 of 16

13652389, 2023, 5, Downloaded from https://bsssjournals.onlinelibrary.wiley.com/doi/10.1111/ejss.13424 by Hayley Peter-Contesse , Wiley Online Library on [18/10/2026]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License

pore size class of 300

–

1200

μ

m also correlated positively

with OC content while negative correlation was found
for pore size class 0.2

–

1

μ

m. In respect of clay content,

porosities in pore size classes with pore size >30

μ

m cor-

related negatively with clay content. Pore size distribu-
tions derived from water retention curves are given in
Table

S2

(

supplementary material

4

|

D I S C U S S I O N

Properties of arable clay-rich soils in Southern Finland
were studied with the initial hypothesis of discernible
effects of OC contents on structure-related properties.
The hypothesis was confirmed regarding positive rela-
tionships observed between OC and aggregate stability
and soil total porosity, and a negative relationship
between OC and particle dispersion. However, a decisive
effect of clay content was apparent in most of the attri-
butes. The study focused on soils with high clay content
in which the clay contents ranged between 28% and 79%
in the whole dataset (aggregate stability) and centred

around the higher end in the subsets of specified analyses
(38%

–

77% clay in WHC, and 54%

–

69% in hydraulic con-

ductivity and image analyses), while maintaining a wide
range in OC (ca. 2%

–

7.6%). Moreover, the focus was on

the practical environment (real farm conditions) at fields
which were located scattered within the Uusimaa region
and encompassed variation in management. This pre-
vented rigorous standardizations, consequently increas-
ing variability but on the other hand also generalizability.
Thus, the detected trends demonstrate how OC and clay
can affect the key structural features in different environ-
mental conditions. The results of water holding capacity,
tension infiltrometry, and X-ray tomography measure-
ments illustrate the impacts of clay and OC on soil pores
and hydraulic conductivity below the managed surface
layer of no-till and reduced tilled soils.

The aggregate stability results of clay soils with

diverse

cultivation

histories

agreed

with

1984

; Haynes,

2000

) and

showed the positive effect of OC on aggregate stability.
Also, as the clay-sized particles are important compo-
nents

in

formation

of

stable

aggregate

structure

F I G U R E 3

Example 3D

visualizations of X-ray visible
soil macropore system for the

five sampling locations. Each

visualization represents a
cubical subvolume with side

length of 2 cm. OC content is

increasing from (a) to (e), and
the OC contents and porosities

were following: (a) 2.5% 0.019;

(b) 2.9% and 0.030; (c) 3.8% and
0.046: (d) 4.5% and 0.022;

(e) 7.7% and 0.11.

T A B L E 3

Spearman's rank

correlation coefficients for image
analysis results.

Porosity

Median pore size

D

c

K

sim

ρ

p

ρ

p

ρ

p

ρ

p

OC

0.463

0.020

0.926

0.151

0.470

0.165

0.431

Clay

0.510

0.009

0.514

0.009

0.549

0.005

0.604

0.001

Clay/OC

0.667

0.000

0.106

0.614

0.311

0.131

0.345

0.091

Note

: Statistically significant differences are bolded. D

c

denotes the critical pore diameter and K

sim

the

simulated hydraulic conductivity.

8 of 16

SOINNE

ET AL

.

13652389, 2023, 5, Downloaded from https://bsssjournals.onlinelibrary.wiley.com/doi/10.1111/ejss.13424 by Hayley Peter-Contesse , Wiley Online Library on [18/10/2026]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License

(Tisdall & Oades,

1982

), we measured the highest aggre-

gate stabilities in heavy clay soils with clay content
higher than 60%. The lower dispersibility of clay-sized

particles in soils with higher OC content is in accordance
with the findings of Watts and Dexter (

1997

) and our

results also indicate the risk of clay dispersion to increase

F I G U R E 4

Image analysis results for all replicate samples plotted against soil OC (a, d, and g), and clay (b, e, and h) contents and clay/

OC ratio (c, f, and i). The arrows indicate the statistically significant Spearman rank correlations (all data points treated as individual

observations) and act as a guide for the eye.

T A B L E 4

Spearman's rank correlation coefficients for pore size distributions derived from X-ray tomography.

<200

μ

m

200

–

400

μ

m

400

–

600

μ

m

600

–

1000

μ

m

>1000

μ

m

ρ

p

ρ

p

ρ

p

ρ

p

ρ

p

OC

0.584

0.002

0.459

0.021

0.533

0.006

0.404

0.045

0.220

0.291

Clay

0.522

0.007

0.373

0.067

0.377

0.064

0.522

0.007

0.635

0.001

Clay/OC

0.788

0.000

0.647

0.000

0.675

0.000

0.612

0.001

0.428

0.033

Note

: Statistically significant differences are bolded.

SOINNE

ET AL

.

9 of 16

13652389, 2023, 5, Downloaded from https://bsssjournals.onlinelibrary.wiley.com/doi/10.1111/ejss.13424 by Hayley Peter-Contesse , Wiley Online Library on [18/10/2026]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License

with increasing clay/OC ratio. Clay/OC ratio has been
suggested as an indicator of soil structural stability with
values lower than a threshold value of 10 indicating good
soil structure (Dexter et al.,

; Prout et al.,

). The

stability of aggregates largely governs the ability of a soil
to

retain

its

structure

under

stress

conditions

(Amézketa,

1999

), and reduced stability resulting in dis-

integration of aggregates may lead to reduced permeabil-
ity and increased surface runoff (Rabot et al.,

).

Soil total porosity and pore size distribution were

examined with X-ray tomography and water retention
measurements that probe partially different pore size
ranges. Imaging was used to describe macroporosity, and
it showed direct information of the pore system. Water
retention measurements, in turn, showed information on
the wider pore size distribution range, but the interpreta-
tion of the pore sizes is indirect as it is based on the capil-
lary bundle model (Young-Laplace equation). Thus,
results obtained with these two methods complement
each other.

Clay content correlated positively with total porosity

derived from water retention results, but in contrast, total
porosity derived with imaging (pores

≥

30

μ

m) was nega-

tively related to clay content. These ostensibly contrasting
results reflect the differences in detectable pore sizes
between the methods. Porosity derived with imaging

correlated negatively with clay content in all size classes,
although for pore sizes 200

–

600

μ

m, no statistically sig-

nificant correlation was observed. This result was sup-
ported by our water retention measurements (in large
pores

≥

30

μ

m) and by the findings of Naveed et al. (

2013

)

who reported decreasing macroporosity (pore size
>1 mm) with increasing clay content.

According to imaging, total porosity as well as poros-

ity in the studied pore size classes showed a significant
positive correlation with OC content from 30

μ

m to

1 mm. A positive correlation between total porosity and
OC content was also found in water retention results as
well as in pore size class 300

–

1200

μ

m, which overlaps

with the pore sizes observed with imaging. Our work is
mostly consistent with previous research (Schlüter
et al.,

2011

; Xu et al.,

; Zaffar & Sheng-Gao,

2015

;

Zhou et al.,

) reporting results of the positive influ-

ence of OC on the soil porosity at length scales typically
visible in X-ray tomography images. The results of Larsbo
et al. (

), however, indicated that the positive relation-

ship between large pores (200

–

1200

μ

m) and OC weakens

or disappears in high-OM soils where the clay/OC ratio is
lower than 10. According to Dexter et al. (

), at a

clay/OC ratio < 10, clay particles are complexed with soil
organic matter and the dispersibility of clay is reduced.
Higher OC contents (clay/OC ratio < 10) would thus not
further enhance aggregation in soil and, therefore, would
not increase the pore space between the aggregates. Min-
asny and McBratney (

) suggested that the impact of

organic matter especially on large pores originates from
the formation of macroaggregates. In our sample set
included in the X-ray tomography and water retention
measurements, only one soil had a clay/OC ratio lower
than 10 and our results showed a significant negative cor-
relation between macropores and clay/OC ratio. How-
ever, imaged macroporosity and measured FW did not
level off when the clay/OC ratio decreased towards 10.

The FW representing the volume of pores >30

μ

m

was calculated based on water retention measurements,
and similarly as the macropore regime, correlated posi-
tively with OC content. As suggested by Minasny and
McBratney (

), the larger share of stable aggregates in

high OC soils likely contributed to the high FW. The
higher FW values suggest that OC content can increase
the amount of water that can be drained by gravity from
clay soils leading to increased volume of air-filled large
pores above the drainage depth. The increased air space
can reduce the probability of water logging and conse-
quently saturation-excess surface runoff generation.
Increased FW indicates a higher macroporosity which
could have implications regarding preferential flow gen-
eration and transport of substances (e.g. Jarvis,

However, OC content had no significant impact on the

F I G U R E 5

Simulated hydraulic conductivity for water flow

through imaged soil samples as a function of critical pore diameter.
Both values for each dot in the figure have been determined for the

same soil sample. Line is a linear fit (R

2

=

0.788) to the data points

and has a slope of 2.02, whereby on linear scale simulated
hydraulic conductivity follows as a function of critical pore

diameter a power law with exponent equal to the slope.

10 of 16

SOINNE

ET AL

.

13652389, 2023, 5, Downloaded from https://bsssjournals.onlinelibrary.wiley.com/doi/10.1111/ejss.13424 by Hayley Peter-Contesse , Wiley Online Library on [18/10/2026]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License

saturated hydraulic conductivity determined by simula-
tions or field measurements. This means that the drain-
age speed may not be affected by changes in the OC
content. Instead, a negative relationship between clay
content

and

saturated

hydraulic

conductivity

was

revealed by both image-based flow simulations and field
measurements with tension infiltrometers. This relation-
ship was explained by the negative correlations between
clay content and porosity in macropore regime and criti-
cal pore diameter.

The simulated and measured hydraulic conductivities

had partially large differences such that the largest simu-
lated values were 1 or 2 orders of magnitude greater than
the measured one, while the smallest values were of the
same magnitude. This can be understood from the small
sample size used in imaging. A similar discrepancy
between simulated and measured hydraulic conductivi-
ties was detected by Schlüter et al. (

), who also rea-

soned that some assumptions in numerical simulations,
such as complete saturation and perfect wettability, are
not entirely realistic. Our simulations showed that
hydraulic conductivity was related to the critical pore
diameter, that is, the diameter of the largest sphere that
can pass through the pore system in the vertical direc-
tion. This demonstrated the marked role of pore mor-
phology in the formation of soil hydraulic properties.
For samples with limited height, it becomes more prob-
able to find a continuous large vertical pore through
the sample. Such a flow pathway would not convey
water as efficiently in a sample where the path passes
through a smaller bottleneck in the pore system. Perco-
lation theory by Katz and Thompson (

1986

) links the

saturated hydraulic conductivity to critical pore radius
via a power law with exponent 2. This model has been
found to describe the hydraulic conductivity of soils

(e.g. Ghanbarian et al.,

; Koestel et al.,

). Our

simulated data followed a power law with exponent of
2 very accurately. Schlüter et al. (

) also reported

that percolation theory and critical pore diameter pre-
dicted the saturated hydraulic conductivity well in soil
with no-till management while for conventionally tilled
soil the predictive power was clearly weaker. The soil
samples used in our simulations were taken from fields
under no-till or reduced tillage and in the latter case
from the soil layer that is not affected by reduced tillage
operations. Our results showed that the relationship
was clear in a range of real farm conditions, and thus
underlines the generality of the relationship. In com-
parison with reported correlations between critical pore
diameter

and

measured

hydraulic

conductivities

(Koestel et al.,

), our result based on simulated

hydraulic conductivity is strong, which is likely due to
the incomplete wetting and air entrapment always pre-
sent in experiments (cf. Schlüter et al.,

). The finite

size of the imaged samples can be expected to cause
exaggerated hydraulic conductivity values from simula-
tions based on the images especially as hydraulic con-
ductivity increases faster than linearly with increasing
critical pore diameter. In any case, simulated saturated
hydraulic conductivities showed a similar negative cor-
relation with clay content and a lack of correlation with
C content as the unconfined measurements with ten-
sion infiltrometers.

Below the wilting point, the water is mainly held by

adsorptive forces on surfaces of solid matter and the large
specific surface area of clay-sized particles explains the
positive correlation between clay and porosity for pores
<0.2

μ

m detected here and reported in previous studies

(Fukumasu et al.,

; Torres et al.,

). Similarly, the

specific surface area of SOM is large and, therefore,

T A B L E 5

Spearman's rank

correlation coefficients for

measurements of soil hydraulic
properties.

K (

4 cm)

K (

2 cm)

K (

1 cm)

K

fs

ρ

p

ρ

p

ρ

p

ρ

p

OC

0.061

0.797

0.233

0.323

0.245

0.297

0.258

0.273

Clay

0.601

0.005

0.368

0.111

0.172

0.469

0.466

0.038

Clay/OC

0.159

0.502

0.031

0.898

0.221

0.350

0.349

0.131

PAW

RPAW

FW

ρ

p

ρ

p

ρ

p

OC

0.463

0.000

0.203

0.120

0.307

0.017

Clay

0.060

0.649

0.244

0.060

0.383

0.003

Clay/OC

0.677

0.000

0.114

0.385

0.742

0.000

Note

: Statistically significant differences are bolded. K(h) denotes near-saturated hydraulic conductivity at

supply pressure head h (

4,

2,

1 cm) and K

fs

is the field saturated hydraulic conductivity. PAW, RPAW,

and FW are plant available water, readily plant available water, and free water, respectively.

SOINNE

ET AL

.

11 of 16

13652389, 2023, 5, Downloaded from https://bsssjournals.onlinelibrary.wiley.com/doi/10.1111/ejss.13424 by Hayley Peter-Contesse , Wiley Online Library on [18/10/2026]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License

contributes to water retention below the wilting point
(Rawls et al.,

2003

). The water retention measurements

done to unravel the effects of OC on fine-sized pores did
show a positive correlation between OC and water held
in <0.2

μ

m pores.

In contrast to <0.2

μ

m pores, we found a negative

correlation

for

pore

size

class

0.2

–

1

μ

m

with

OC. Hydrophobicity of organic matter could affect the
water retention properties in high OC soils. According to
Rasa et al. (

), subcritical water repellence is present

in boreal clay soils and their results indicate that soil
higher in organic matter exhibits more severe water

repellence. In addition, complexation of organic mole-
cules with mineral matter and microaggregate formation
creates organic matter-filled spaces between the mineral
particles (McCarthy et al.,

). McCarthy et al. (

)

reported a sharp drop in the share of organic matter-filled
spaces with a diameter larger than 1

μ

m, indicating

that organic matter would fill spaces comparable to
<1

μ

m pores and, therefore, have a decreasing effect

especially on the pore size class 0.2

–

1

μ

m. Possibilities for

the formation of microaggregates and filling of intra-
microaggregate pores with OC are higher in high-OC
soils supporting the negative correlation between OC and

F I G U R E 6

Soil hydraulic properties (plant availabe water, PAW; readily plant available water, RPAW; free water, FW) plotted against

soil OC (a, d, and g) and clay (b, e, and h) contents, and clay/OC ratio (c, f, and i). The arrows indicate the statistically significant Spearman

rank correlations (all data points treated as individual observations) and act as a guide for the eye.

12 of 16

SOINNE

ET AL

.

13652389, 2023, 5, Downloaded from https://bsssjournals.onlinelibrary.wiley.com/doi/10.1111/ejss.13424 by Hayley Peter-Contesse , Wiley Online Library on [18/10/2026]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License

0.2

–

1

μ

m pores. However, Fukumasu et al. (

) who

also used water retention measurements to determine
soil pore size distributions, found the pore diameter
0.2

–

5

μ

m to have a positive correlation with OC. The con-

trasting results may partly originate from different pore
size classes reported but also from different textures in
the studied soils. It should be noted that, apart from the
study by Larsbo et al. (

) with clay contents ranging

from 63.7% to 70.5%, all the above-mentioned studies
were conducted in fields with clearly lower clay content
than those in this study.

The PAW and RPAW results stem from the water

retention measurements which explains the negative cor-
relation between OC content and PAW. However, this
negative influence results solely from pore size class of
0.2

–

1

μ

m, whereby a similar negative correlation is not

observed for RPAW, which is derived from water stored
in larger pores. According to the meta-analysis by Min-
asny and McBratney (

), increase in soil organic mat-

ter has only a small effect on PAW and the effect is
almost

negligible

for clay

soils. A similar

weak

correlation between PAW and OC was also reported by
Libohova et al. (

), who also reported that for clay

soils, mineralogy plays an important role such that OC is
more clearly correlated with water retention in kaolinitic
than smectitic clays. For the heavy clay soils studied here,
site-specific mineralogy was not determined, but differ-
ences

between

sites

are

likely

small

(Keskinen

et al.,

). The negative correlation between OC and

PAW arose from the low PAW values determined for the
soils with highest OC content (>6%). These soils could
hold a clearly lower amount of PAW determined as a dif-
ference between suctions 10 and 1600 kPa than the soils
with lower OC content. However, recent articles have
reported that plants grown in clay soils do not wilt at the
permanent wilting point of 1500 kPa but at higher

suctions and under lower water contents when com-
pared to those grown in coarser soils (Torres et al.,

;

Wiecheteck et al.,

). Considering the reported dif-

ferences between the soil matric potential at which the
physiological wilting of different plant species occurs,
and the permanent wilting point measured as a water
content at 1500 kPa, the practical significance of OC on
PAW and especially the decreasing effect on volume of
0.2

–

1

μ

m pores needs further research especially in

high-clay soils.

For many of the structural parameters studied here,

increase in OC and clay content appear to have an oppo-
site effect. Therefore, for clay-rich soils, a higher amount
of OC is required to achieve comparable levels in many
structural parameters. In several earlier studies, clay/OC
ratio has been recognized as a viable parameter to quan-
tify soil aggregate stability (Soinne et al.,

) or a soil

quality indicator to assess appropriate level of soil organic
matter (Johannes et al.,

; Prout et al.,

), which

emphasize the importance of maintaining a sufficient
level of OC from the point of view of crop production.
Our results showed the importance of clay and OC for
aggregate stability also in heavy clay soils.

In this work, we investigated the influences of OC

and clay content on the soil structure on regional scale at
private farms. Therefore, our results likely contain
uncontrollable sources of variation due to possibly differ-
ent management histories or weather conditions. Mineral
type has also been found to influence the soil hydraulic
properties (Lehmann et al.,

; Libohova et al.,