R E S E A R CH AR T I C L E
Soil organic carbon fractions and storage potential in
Finnish arable soils
Anna-Reetta Salonen1,2 | Ron de Goede1 | Rachel Creamer1 |
Jussi Heinonsalo3,4 | Helena Soinne5
1Soil Biology Group, Department of
Environmental Sciences, Wageningen
University & Research, Wageningen,
Netherlands
2Environmental Soil Science, Department
of Agricultural Sciences, University of
Helsinki, Helsinki, Finland
3Department of Forest Sciences,
University of Helsinki, Helsinki, Finland
4Institute for Atmospheric and Earth
System Research/Forest Sciences, Faculty
of Agriculture and Forestry, University of
Helsinki, Helsinki, Finland
5Natural Resources Institute Finland
(LUKE), Helsinki, Finland
Correspondence
Anna-Reetta Salonen, Soil Biology Group,
Department of Environmental Sciences,
Wageningen University & Research,
PO Box 47, 6700AA Wageningen,
Netherlands.
Email: anna-reetta.salonen@wur.nl
Abstract
Understanding the factors affecting the total amount and distribution of soil
organic carbon (OC) across different functional carbon pools is important to
better define the future management of soil OC stocks. The interactions
between soil management practices, local physicochemical soil properties and
climate are essential for determining the OC content of the soil. Nevertheless,
how these factors affect the total amount of OC and its distribution across car-
bon pools, i.e., more labile particulate (POC) and more stable mineral-
associated (MAOC) organic carbon, is only partly known. In this study, we
assessed topsoil (0–20 cm) samples from 93 arable farms in the southern half
of Finland to determine the total amount of OC, and its distribution in MAOC
and POC, along with relevant soil properties (amount of clay and silt, alumin-
ium and iron oxides and pH), climate (precipitation and temperature) and fer-
tilization (mineral versus organic). The fertilization did not affect the total soil
carbon content (12–58 g OC kg
1 soil). The share of OC in the MAOC fraction
(on average 86% of total OC) was relatively stable across the large range of OC
contents and clay contents (2%–68%). We assessed the highest feasible MAOC
of the soils with boundary line analyses and their OC saturation state with
Hassink's equation (Hassink, 1997). Only soils with the lowest clay content
(<10% clay) were assumed to be carbon-saturated, suggesting that most of the
studied soils have a capacity to accrue more MAOC. Simple linear regression
showed that clay, aluminium and iron oxides explained 9%, 21% and 22% of
the variation in MAOC, respectively. Multiple regression analyses including
the amount of clay, clay+silt, aluminium and iron oxides, pH, type of fertiliza-
tion, precipitation and temperature as explanatory variables explained
33%–53% of the variation in OC and MAOC. In all soils, aluminium oxides
were important explanatory variable for MAOC, whereas Fe oxides were signifi-
cant only in soils with higher clay content (>30%). In soils with a low clay content
(<30%), pH had added value in explaining MAOC. Altogether, it seems that vari-
ous climatic, edaphic and soil management-related factors are context-dependently
Received: 15 April 2024 Revised: 24 May 2024 Accepted: 14 June 2024
DOI: 10.1111/ejss.13527
This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided
the original work is properly cited.
© 2024 The Author(s). European Journal of Soil Science published by John Wiley & Sons Ltd on behalf of British Society of Soil Science.
Eur J Soil Sci. 2024;75:e13527. wileyonlinelibrary.com/journal/ejss 1 of 17
https://doi.org/10.1111/ejss.13527
controlling OC and that soil textural information alone is not necessarily an
adequate predictor to assess the MAOC saturation state of the soil.
KEYWORD S
aluminium oxides, boundary line analyses, carbon saturation, clay, fertilizing, Hassink's
equation, iron oxides, mineral-associated organic carbon, particulate organic carbon, soil
texture
1 | INTRODUCTION
Soil organic matter (SOM) is recognized as a key for a
multitude of soil functions (Hoffland et al., 2020;
Wiesmeier et al., 2019), one of which, climate regulation,
a potential sink for atmospheric carbon dioxide (CO2;
Paustian et al., 2016) is receiving a lot of attention. Agri-
cultural management has resulted in the depleted soil
organic carbon (OC) globally (Paustian et al., 2016;
Sanderman et al., 2017). There is now a strong focus on
carbon farming practices which promote the incorpora-
tion of organic matter into soils as part of agricultural
management to maintain or increase soil carbon stocks.
However, to date, our mechanistic understanding of the
soil processes relating to the soil OC dynamics and stabi-
lization is still inadequate and results in variable out-
comes in relation to carbon farming practices (Angst
et al., 2023).
SOM is a complex mixture of organic molecules in
different phases of decomposition (Schmidt et al., 2011)
and hence presents in the soil in a multitude of forms
(Simpson & Simpson, 2012). To determine the forms of
SOM in the soil, it is often fractionated into mineral-
associated (MAOM) and particulate (POM) organic mat-
ter (Cambardella & Elliott, 1992; Lavallee et al., 2020;
Tiessen & Stewart, 1983). In mineral soils, the majority
of OC is often present as mineral-associated organic
carbon (MAOC; Kögel-Knabner et al., 2008, Georgiou
et al., 2022) which is regarded as a relatively stable and
long-lasting form of soil OC since its interactions with
soil minerals guards it against further decomposition
(Lavallee et al., 2020). In contrast, particulate organic car-
bon (POC) has only limited interaction with soil minerals
and is hence considered more labile than MAOC, but soil
aggregation or soil physicochemical status may still slow
the rate of its decomposition (Lavallee et al., 2020).
Around half of MAOC occurring in arable soils is pre-
sumed to be of microbial origin and half directly of plant
origin (Angst et al., 2021) whereas POC comprises rela-
tively undecomposed plant material (Lavallee et al., 2020).
MAOC is considered important from a climate miti-
gation perspective, as it potentially decomposes slower
than POC and can form a relatively stable OC storage
(Lavallee et al., 2020). The more labile POC is used as an
energy source for soil biota (Yu et al., 2022) which in turn
also becomes a source for MAOC production (Angst
et al., 2023). Approximately 20% of the OC content of ara-
ble soil is considered to be POC (Guillaume et al., 2022;
Mayer et al., 2022; Sokol et al., 2022); however, the POC
pool is considered to be dynamic which means its pool
size can readily change as the result of soil management
interventions. One of the soil management practices to
which total OC, POC and MAOC may respond on a rela-
tively short timescale is the type of fertilization applied
(Angst et al., 2023). The application of organic fertilizers
introduces OC in the form of POC into the soil, but
depending on climatic and edaphic factors, most of the
OC may decompose relatively soon after application
(Maillard & Angers, 2014; Mayer et al., 2022), meaning
that maintaining a certain level of soil OC stocks
would require a continuous input of organic materials
(Powlson & Neal, 2021). Decomposition of the added
organic materials results in the respiration of some of
the OC in the form of carbon dioxide (CO2) and there-
fore a loss of OC from the soil. The rate of this loss is
dependent on the temperature, moisture and quality of
the organic materials (Berthelin et al., 2022; Powlson &
Neal, 2021). However, this process also results in the
assimilation of soil organic carbon into the microbial
biomass, and this biologically processed OC contributes
to a relatively higher proportion of OC in the MAOC
fraction (Angst et al., 2021).
Highlights
• Organic fertilization practices did not increase
OC compared to mineral fertilization.
• A high proportion of OC in MAOC (86 ± 3%)
in soils that represented a large range of clay
contents (2%–68%).
• The majority of studied soils have a capacity to
accrue more MAOC.
• Detecting universally applicable predictors for
soil OC is challenging.
2 of 17 SALONEN ET AL.
 13652389, 2024, 4, D
ow
nloaded from
 https://bsssjournals.onlinelibrary.wiley.com/doi/10.1111/ejss.13527 by Duodecim Medical Publications Ltd, W
iley Online Library on [10/07/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on W
iley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
The long-term stabilization of OC in the form of
MAOC is proposed to be strongly dependent on soil
properties such as mineral particles and various metal
oxides, and soil pH (Rasmussen et al., 2018; Wiesmeier
et al., 2019). The amount of clay-sized (<2 μm) and silt
(<20 μm or <60 μm) particles has traditionally been
considered an adequate predictor of the soil OC storing
capacity (Dexter et al., 2008; Feng et al., 2013;
Wiesmeier et al., 2019) as the proportion of these parti-
cles provides a reference of mineral surface area avail-
able for MAOC binding. Furthermore, clay is a
paramount factor in soil aggregation which may hinder
further OC decomposition as aggregates often form
around POC particles, making it physically inaccessible
for decomposer organisms (Besnard et al., 1996; Totsche
et al., 2018). However, despite the acknowledged impor-
tance of the clay-sized soil minerals for OC storage, it
remains unclear what the limitations are for the maxi-
mum binding capacity for MAOC on these mineral sur-
faces, and some studies find no upper limit for the
operationally defined MAOC (Begill et al., 2023;
Matus, 2021; Salonen et al., 2023; Urbanski et al., 2023).
Finnish soils are relatively young as they were formed
by Weichselian glaciation around 10,000 years ago
(Koljonen, 1992). During the glaciation, soils and their
clay and silt-size minerals were formed from rock by abra-
sion of the ice. The felsic parent material (Yli-Halla &
Mokma, 2002) from which these soils are formed is
strongly durable and resistant to in-situ weathering, lead-
ing to a relatively large proportion (more than 50%;
Keskinen et al., 2022) of the minerals in the clay and silt-
sized fractions of Finnish soils being only moderately
weathered. This means that they have a relatively low sur-
face activity in comparison to soils that are further weath-
ered and hence containing a larger share of 1:1 or 2:1 type
of clay minerals that are important in stabilizing soil OC
(Georgiou et al., 2022; Six et al., 2002; Six et al., 2024).
As soil mineralogy is spatially variable, so are the asso-
ciated surface properties such as charge density and sur-
face area (Mitchell & Soga, 2005) and soils with
corresponding clay+silt content may have differing types
of mineral surfaces. Therefore, the surface activity and
MAOC storage capacity may be different despite a similar
clay+silt content (Feng et al., 2013; Georgiou et al., 2022).
Different metrics applying the amount of specific-sized soil
minerals (i.e., clay and/or silt) or the type of the dominat-
ing clay mineral (i.e., 1:1 or 2:1) can be used to assess the
OC saturation state of soil and to estimate OC accrual
potential of soil (e.g., Dexter et al., 2008; Feng et al., 2013;
Georgiou et al., 2022; Hassink, 1997; Schjønning
et al., 2012; Six et al., 2024). Estimating the soil OC satura-
tion state can provide information on the OC sequestra-
tion potential of agricultural soils, which can be
potentially useful from the climate mitigation perspective
since soils can act as a carbon sink (Keenan &
Williams, 2018; Paustian et al., 2016).
In addition to clay and silt, oxides of aluminium
(Al) and iron (Fe) can act efficiently as OC binding
agents (Hall & Thompson, 2022; Mendez et al., 2020;
Wiesmeier et al., 2019; Wiseman & Püttmann, 2006).
Interaction of OC with Al and Fe can include OC
adsorption onto oxide surfaces, strong inner-sphere bond
formation between OC and Al and Fe oxides, and co-
precipitation of OC and oxides (Gerke, 1992; Schneider
et al., 2010; Tamrat et al., 2019). Their oxalate-extractable
species (Al oxides and Fe oxides) have been shown to be
better predictors for soil OC than soil texture (i.e., clay or
clay+silt content) in soils of humid cold climates with
relatively low pH (Fukumasu et al., 2021; Rasmussen
et al., 2018; Salonen et al., 2023; Wiesmeier et al., 2019).
Such soils are typical of boreal conditions in Finland
(Metzger et al., 2012) where relatively cold mean annual
temperatures hinder evapotranspiration, leading to prev-
alent moist conditions that can hamper SOM decomposi-
tion (Wiesmeier et al., 2019) and that may lead to
relatively low soil pH which may further retard microbial
decomposition of soil carbon (Keiluweit et al., 2016;
Slessarev et al., 2016).
Our objectives were: (i) to assess the amount of OC
and its distribution across MAOC and POC carbon frac-
tions in arable soils receiving organic or mineral fertiliza-
tion; (ii) to estimate the OC saturation state of the
studied soils; (iii) to assess which soil properties (clay,
silt, Al and Fe oxides, and pH) explain the different pro-
portions of the carbon fractions; and (iv) to determine
whether climatic factors, that is, precipitation and tem-
perature, control the total amount of OC and its dis-
tribution into MAOC and POC. We analysed topsoil
(0–20 cm) samples from 93 farms with varying soil tex-
tures (clay content 2%–68%) located across the southern
half of Finland (60N–65N) for the total amount of OC
and its distribution between MAOC and POC. We used a
boundary line analysis (Feng et al., 2013; Georgiou
et al., 2022; Six et al., 2024) to assess the highest feasible
MAOC values for these soils under the current agricul-
tural management. Further, the MAOC saturation state
of the soil was assessed with Hassink's equation
(Hassink, 1997). We hypothesized that: (1) there will be
higher levels of total OC, MAOC and POC in organically
fertilized soils due to higher inputs of organic matter;
(2) as cultivation often depletes soil OC stocks, the stud-
ied soils will not be OC-saturated; (3) Al and Fe oxides
will be better predictors of MAOC than clay content as
they act as the preferential binding sites; and (4) soil OC
concentration increases with increasing precipitation and
decreases with increasing temperature.
SALONEN ET AL. 3 of 17
 13652389, 2024, 4, D
ow
nloaded from
 https://bsssjournals.onlinelibrary.wiley.com/doi/10.1111/ejss.13527 by Duodecim Medical Publications Ltd, W
iley Online Library on [10/07/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on W
iley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
2 | MATERIALS AND METHODS
2.1 | Soil sampling and basic soil
characterization
Soil samples were collected during the years 2018 and
2019 right after establishing the Finnish Carbon Action
project (Mattila et al., 2022) focussing on the effects of a
range of soil management practices that are estimated to
have the potential to increase soil OC levels. At the start,
participating farmers selected a field within their farm for
the project and reported its soil management history
for the 5 years preceding the start and baseline soil sam-
pling (i.e., for the years 2014–2018 or 2015–2019, depend-
ing on the year that the farm entered the project).
Studied soils were in arable use and there were no per-
manent grasslands included. Most of the farms had
cereals (n = 78) and grasses (n = 52) included in the five-
year crop rotation before the soil sampling. Based on the
soil management history from 5 years before the soil
sampling, the studied farms were divided into two
groups: farms applying mainly mineral fertilizers and
farms applying mainly organic fertilization. The quality
and amounts of applied fertilizing (in both, farms that
applied mineral and organic fertilizing) varied from farm
to farm, considering that the applied maximum amounts
of nutrients were in line with the national regulations.
Individual rates of the amounts of nutrients applied with
fertilizing were not available. Despite the nature of the
data (i.e., only a 5-year period prior to sampling had been
considered and lack of the quantities of OC applied with
the organic fertilizing), we decided to divide the farms into
minerally and organically fertilized farms to assess
whether effects of fertilization could be seen in the total
OC, or its distribution between the MAOM and POM frac-
tions. The farms were located across the southern half of
Finland (60N–65N; Figure 1). Long-term mean annual
precipitation (MAP) in the studied farms was 636 mm,
with a median of 631 mm (range 548–727 mm year 
1).
Long-term mean annual temperature (MAT) was 5.1C,
with a median of 5.2C (range of 3.3–6.4C). MAP and
MAT are 19-year averages before the soil sampling (years
1998–2017; Finnish Meteorology Institute. Data from the
weather stations closest to each farm).
FIGURE 1 Locations of the studied farms in Finland and mean annual temperature (1A, MAT,C) and precipitation (1B, MAP, mm).
MAT and MAP are 19-year averages.
4 of 17 SALONEN ET AL.
 13652389, 2024, 4, D
ow
nloaded from
 https://bsssjournals.onlinelibrary.wiley.com/doi/10.1111/ejss.13527 by Duodecim Medical Publications Ltd, W
iley Online Library on [10/07/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on W
iley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
The samples analysed in this study were taken from
three geo-referenced sampling points per field as part of
the baseline sampling. Baseline sampling took place in
the autumn of the start year, after the last harvest, before
the carbon farming practices were commenced. The top-
soil (0–20 cm) was sampled by the participating farmers
who collected 10 soil core (Ø 14 mm) samples on the cir-
cumference of a 10-m radius circle. The 30 samples were
bulked to form one composite sample which was sent in
a plastic container by a courier service to the University
of Helsinki (Helsinki, Finland), all samples arrived
within 3 days of sampling. Soils were sieved (mesh size
10 mm) and part of the samples were stored at 
20C
prior pending further analyses, and the other part was
analysed for elemental C and N (LECO, Michigan, USA).
2.2 | MAOM and POM fractionation, Al
and Fe oxide extraction, and soil textural
analyses
Soils were fractionated by size into MAOM (<53 μm) and
POM (>53 μm) according to Cotrufo et al. (2019). In
short, frozen soils (stored at 
20C prior to the analyses)
were melted at +4C, sieved (mesh size 2 mm) and air-
dried. Then the soil samples were dispersed by shaking in
5 g l
1 sodium hexametaphosphate and glass beads for
18 h at 120 rpm (Ika LabortechnikKS 501 digital). After
the dispersion, the soil was rinsed on a sieve (mesh size
53 μm). The soil that passed through the sieve was col-
lected as MAOM (<53 μm), and the soil on top of the
sieve was collected as POM (>53 μm). Fractions were
dried to a constant weight in a forced-air oven (60C). To
confirm reproducibility of the soil dispersion, two inter-
nal reference soils (clay- and coarse-textured soil) were
included in each set of fractionated soils. Samples
were analysed for OC (MAOC and POC) by dry combus-
tion (LecoCHN 628, St. Joseph, Michigan, USA). OC
recovery was on average 97% and ranged from 100 to
84%. All the measured C was taken to represent OC as
due to the inherently low pH in Finnish soils, carbonate
minerals are absent (Nelson & Sommers, 1982).
Poorly crystalline Al and Fe oxides were extracted
with acid ammonium oxalate extraction according to
Niskanen (1989; 0.05 M oxalate, pH 3.3). Prior to the
analysis, soil was stored at 
20C, then defrosted to
+4C, sieved (mesh size 2 mm) and air-dried. We used
5 g of air-dried soil in the extraction. Al and Fe were
measured with inductively coupled plasma optical
emission spectrometry (ICP–OES Thermo Scientific
iCAP 6300 MFC DUO).
Soil texture (size classes: clay <2 μm; fine silt
2–20 μm; coarse silt 20–60 μm; fine sand 60–200 μm; and
coarse sand >200 μm) was analysed with the pipette
method (Elonen, 1971). Following the protocol, soils
were pre-treated with H2O2 for SOM removal and then
acidified with 2 M HCl and dispersed with 0.05 M
Na4P2O7.
2.3 | Estimations of soil OC saturation
and OC accrual potential
We applied a modified boundary line analysis to estimate
the maximum capacity for MAOC stabilization (as in
Feng et al., 2013, Georgiou et al., 2022, Six et al., 2024).
In the boundary line analyses, the soils having the high-
est MAOC contents were used in analysing a feasible
maximum for OC accrual in the studied climate and
under the current arable land management. For the anal-
ysis, data were sorted by mass proportions of clay+silt
(<60 μm) particles in bulk soil (g kg
1 soil) and then sep-
arated into groups with intervals of 100 g of the clay+silt
fraction kg
1 soil. Next, the upper tenth percentile of the
MAOC content of clay+silt particles in each group was
identified and then the upper tenth percentile of the
MAOC contents of clay+silt particles and the corre-
sponding mass proportions from each group were com-
bined and used in a regression analysis where the
intercept was forced through zero. We then used the
slope from the boundary line analyses (Figure 2) to esti-
mate the highest feasible MAOC content (g OC kg
1 clay
+silt) for the studied soils as follows (Table 2): Maximum
FIGURE 2 Prediction of the maximal MAOC content of clay
+silt (<60 μm). Regression is based on boundary line analyses with
the data of the upper tenth percentile of MAOC contents included
within clay+silt 100 g kg
1 soil intervals (marked with asterisks).
SALONEN ET AL. 5 of 17
 13652389, 2024, 4, D
ow
nloaded from
 https://bsssjournals.onlinelibrary.wiley.com/doi/10.1111/ejss.13527 by Duodecim Medical Publications Ltd, W
iley Online Library on [10/07/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on W
iley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
MAOC = clay+silt content (%)  maximum MAOC for
the studied soils (55 g C kg
1 silt+clay). We also esti-
mated the maximum MAOC for the studied soils assum-
ing 2:1 or 1:1 clay mineral dominated (Six et al., 2024) as
follows: Maximum MAOC = clay+silt content (%) 
maximum MAOC for 1:1 (46 g C kg
1 silt+clay) or 2:1
(82 g C kg
1 silt+clay).
Soil texture was used to estimate the MAOC accrual
capacity for each soil (Table 2). We calculated the capac-
ity of fine-sized particles to stabilize MAOC using the
Hassink's (1997) equation: MAOC (g kg
1) at OC
saturation = 4.09 + 0.370  (clay+silt%). For clay and
silt, we used two size cut-offs: clay+silt <20 μm, as in
Hassink, 1997, and clay+silt <60 μm. Larger clay+silt
(<60 μm) was included as it is suggested that minerals up
to this size have relevant specific surface area for MAOC
binding (Six et al., 2024). These accrual capacity estima-
tions were compared with measured MAOC to assess the
potential for further MAOC accrual of the soils. Further,
we divided soils into saturated and non-saturated accord-
ing to Hassink's equation (with clay+silt size cut-off
<20 μm) to test the effect of fertilizing on the distribution
of soil OC into MAOC and POC (Table 3).
2.4 | Data analyses
Out of the 105 farms participating in the Carbon Action
project, soil samples from 97 were available for our ana-
lyses. Among these 97 farms, there were four soils that
were classified as highly organic (OC >7%, OM 12%–20%;
Paasonen-Kivekäs et al., 2009) according to the Finnish
soil classification system. As these highly organic soils
had considerably higher OC content than the rest of the
studied soils, we excluded them from the statistical ana-
lyses, leaving 93 soils for a part of the statistical analyses.
For three out of the remaining 93 farms, details on
the fertilization history were not available (not reported
by farmers) and hence these three farms were also excluded
from the statistical analyses where fertilization (mineral vs
organic) was assessed. The remaining 90 farms were divided
into two groups based on their fertilization history (mineral,
n = 46 and organic, n = 44). Two-way ANOVA with a
Tukey HSD test (package ‘stats’) was conducted to evaluate
statistical differences between OC and nitrogen (N) concen-
trations and OC to N ratios of total soil, MAOM and POM,
OC contents of MAOM and POM, soil pH, and clay+silt
content (%; <60 μm) in mineral and organically fertilized
soils grouped by clay content (0–100, 100–200, 200–300 g
clay kg
1soil etc.). Division to the clay content groups was
chosen as some studies have reported increasing OC con-
tent with increasing clay content (Arrouays et al., 2006;
Hassink, 1997; Kaiser & Guggenberger, 2003).
Further, to assess whether fertilization (mineral
vs. organic) would yield different outcomes depending
on the soil OC saturation state, we divided soils into
two groups: MAOC-saturated and non-saturated. We
then employed a t-test (using the ‘rstatix’ package;
Kassambara, 2022) to evaluate differences in OC and N
contents.
We performed a multiple linear regression to identify
predictors for soil OC and MAOC. The selection of the
best explanatory variables was performed on an AIC-driven
model selection approach (‘dredge’ function from the pack-
age MuMIn Barton, 2023). As explanatory variables, we
used textural class (clay <2 μm or clay+silt <60 μm)
(g kg
1 soil; see details below), amount of Al and Fe oxides
(g kg
1 soil), pH, cumulative annual rainfall (mm; yearly
average from the 19-year time period 1998–2017), mean
annual temperature (C; yearly average from 19-year time
period 1998–2017) and fertilization (mineral vs. organic fer-
tilizers). We divided soils into two groups according to their
clay content: (1) <30% of clay and (2) >30% of clay. The
30% cut-off was chosen as according to the Finnish soil clas-
sification system clay soil has >30% of clay (Paasonen-
Kivekäs et al., 2009). The variance inflation factor for the
covariates of the models was checked using the ‘vif’ func-
tion of the car package (Fox & Weisberg, 2019) and this was
found to be <2, meaning that no co-linearity of the covari-
ates of the best-fit models was observed. We inspected the
distribution of the residuals of the final models visually to
check that model assumptions were not violated.
All statistical analyses were performed with R
(R Core Team, 2023, version 4.3.1), and all the plots were
created using R packages ‘ggplot2’ and ‘ggpubr’
(Kassambara, 2020; Wickham et al., 2016). We regarded
p-values of 0.05 or less as statistically significant.
3 | RESULTS
3.1 | MAOC and POC across clay
contents
Soil OC and N contents (g kg
1 soil) tended to increase
with increasing clay content, with the exception of the
clay content interval of 200–300 g kg
1 which had less
OC than the previous clay intervals, and the clay interval
of 300–400 g kg
1 that had the highest measured OC con-
tent (Table 1). N contents were significantly higher in
soils having >400 g kg
1 clay than in soils that contained
less clay. OC content (C%) of the <53 μm soil fraction
(i.e., MAOM) was highest in the lowest clay content
interval (0–100 g kg
1, p-value <0.05). In contrast, the
OC content (C%) of the >53 μm soil fraction (i.e., POM)
increased with increasing clay content.
6 of 17 SALONEN ET AL.
 13652389, 2024, 4, D
ow
nloaded from
 https://bsssjournals.onlinelibrary.wiley.com/doi/10.1111/ejss.13527 by Duodecim Medical Publications Ltd, W
iley Online Library on [10/07/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on W
iley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
The proportion of the total OC in the MAOM fraction
varied between 83% and 89% (Table S1). The total
amount of MAOC expressed per g OC kg
1 total soil was
lowest in the clay content interval of 200–300 g kg
1 and
tended to be different in the soils with differing clay con-
tents whereas the total amount of POC expressed per g
OC kg
1 total soil did not differ between the clay classes.
The bulk soil C to N ratio in the different clay classes was
15 on average (Table 1), and a significantly higher ratio
(C to N ratio = 18) was observed in the 0–100 g kg
1 clay
content interval compared to the soils with a higher
clay content (C to N ratio = 13–15). Such a decrease in C
to N ratio with increasing soil clay content was found
also for MAOM (decrease from C to N ratio 18 to 12) but
not for POM, where the C to N ratio increased from 18 to
23. No statistically significant differences were observed
in soil pH between the different clay content intervals.
There were no statistically significant differences
observed in the carbon- and nitrogen-associated parame-
ters across the clay content intervals for either mineral or
organically fertilized sites (Table S2).
3.2 | The OC saturation status of the
soils under current agricultural
management
With the boundary line analyses, the estimated feasible
practical maximum MAOC in the studied climate and
under the current agricultural management was 55 g
MAOC kg
1 clay+silt (Figure 2).
The estimated feasible highest possible MAOC con-
tents based on the boundary line analyses were higher
than the estimates assuming 1:1 clay mineral dominated
soils (46 g OC kg
1 clay+silt; Six et al., 2024) and lower
than when estimated for 2:1 clay mineral dominated soils
(82 g OC kg
1 clay+silt). Only the soils in the lowest clay
content interval (0–100 g kg
1) reached their MAOC sat-
uration according to the boundary line analyses as they
contained 118% of the estimated feasible MAOC whereas
the other clay content intervals contained 45%–66% of
the estimated feasible MAOC.
To critically assess the results of the boundary line
analysis, we applied Hassink's equation (Hassink, 1997)
for verification. When the MAOC saturation was esti-
mated based on Hassink's equation (Hassink, 1997)
including a clay+silt fraction sized <20 μm, the MAOC
accrual capacities were exceeded in the low clay soils
(0–100 and 100–200 g clay kg
1 soil; Table 2). However,
when including also the larger silt fraction (clay+silt
<60 μm), only soils in the lowest clay content interval
(0–100 g clay kg
1 soil) were estimated to be carbon-
saturated, and soils with a clay content >100 g kg
1 wereT
A
B
L
E
1
A
ve
ra
ge
(±
SE
)
or
ga
n
ic
ca
rb
on
an
d
n
it
ro
ge
n
co
n
ce
n
tr
at
io
n
(g
kg


1
so
il)
of
to
ta
ls
oi
l,
th
e
C
to
N
ra
ti
o
(C
:N
)
of
th
e
to
ta
ls
oi
l,
M
A
O
M
(<
53
μm
so
il)
an
d
PO
M
(>
53
μm
so
il)
,t
h
e
O
C
co
n
ce
n
tr
at
io
n
(g
kg

1
so
il)
an
d
co
n
te
n
ts
(%
)
of
M
A
O
C
an
d
PO
C
in
th
e
<
53
an
d
>
53
μm
so
il
m
as
s
fr
ac
ti
on
,r
es
pe
ct
iv
el
y,
an
d
so
il
pH
in
di
ff
er
en
t
cl
ay
cl
as
se
s
by
10
0
g
kg


1
in
te
rv
al
s.
C
la
y
(g
k
g

1 )
n
T
ot
al
O
C
(g
k
g

1 )
T
ot
al
N
(g
k
g

1 )
So
il
C
:N
M
A
O
M
(<
53
u
m
so
il
)
C
-%
M
A
O
C
(g
k


1
so
il
)
M
A
O
M
C
:N
P
O
M
(>
53
u
m
so
il
)
C
-%
P
O
C
(g
k
g

1
so
il
)
P
O
M
C
:N
p
H
0–
10
0
14
24
.3
±
3.
1a
b
1.
90
±
0.
16
a
18
±
0.
9a
7.
1
±
0.
9a
20
.2
±
2.
7a
b
18
±
1.
0a
0.
7
±
0.
1a
4.
1
±
0.
6
18
±
1.
1a
6.
43
±
0.
19
10
0–
20
0
12
29
.0
±
3.
6a
b
1.
90
±
0.
16
a
15
±
0.
8b
4.
0
±
0.
6b
25
.3
±
3.
2a
b
14
±
0.
8b
1.
8
±
0.
4a
3.
7
±
0.
5
20
±
1.
6a
b
6.
38
±
0.
15
20
0–
30
0
12
22
.6
±
1.
4a
1.
69
±
0.
09
a
13
±
0.
5b
2.
6
±
0.
3b
18
.7
±
1.
3a
13
±
0.
4b
2.
5
±
0.
5a
b
3.
9
±
0.
8
21
±
1.
4a
b
6.
33
±
0.
12
30
0–
40
0
23
33
.9
±
2.
3b
2.
24
±
0.
11
ab
15
±
0.
4b
3.
4
±
0.
2b
29
.3
±
2.
1b
14
±
0.
4b
5.
4
±
0.
9b
4.
6
±
0.
4
22
±
0.
7a
b
6.
38
±
0.
09
40
0–
50
0
8
31
.1
±
3.
4a
b
2.
35
±
0.
33
b
14
±
0.
6b
3.
3
±
0.
4b
27
.8
±
3.
3a
b
13
±
0.
6b
3.
5
±
0.
8a
b
3.
4
±
0.
3
22
±
2.
0a
b
6.
24
±
0.
13
50
0–
60
0
12
30
.1
±
1.
2a
b
2.
20
±
0.
06
b
14
±
0.
4b
3.
0
±
0.
1b
26
.2
±
1.
2a
b
13
±
0.
4b
3.
9
±
0.
9a
b
3.
9
±
0.
5
22
±
0.
5a
b
6.
35
±
0.
10
60
0–
70
0
12
32
.6
±
2.
7a
b
2.
45
±
0.
2b
13
±
0.
4b
3.
1
±
0.
3b
28
.0
±
2.
4a
b
12
±
0.
3b
5.
8
±
0.
6b
4.
5
±
0.
7
23
±
1.
1b
6.
40
±
0.
11
N
ot
e:
L
et
te
rs
de
n
ot
e
st
at
is
ti
ca
ld
if
fe
re
n
ce
s
(T
uk
ey
's
H
SD
,s
ig
n
if
ic
an
ce
le
ve
l<
0.
05
).
C
la
y
an
d
si
lt
co
n
te
n
ts
w
er
e
n
ot
te
st
ed
fo
r
st
at
is
ti
ca
ld
if
fe
re
n
ce
s,
n
=
n
u
m
be
r
of
fa
rm
s
sa
m
pl
ed
(i
n
to
ta
ln
=
93
).
T
h
e
h
or
iz
on
ta
ll
in
e
be
tw
ee
n
th
e
cl
ay
cl
as
se
s
of
20
0–
30
0
g
kg


1
an
d
30
0–
40
0
g
kg

1
in
di
ca
te
s
th
e
cu
t-
of
f
be
tw
ee
n
lo
w
cl
ay
(<
30
%
)
an
d
h
ig
h
cl
ay
(>
30
%
)
so
il
s.
SALONEN ET AL. 7 of 17
 13652389, 2024, 4, D
ow
nloaded from
 https://bsssjournals.onlinelibrary.wiley.com/doi/10.1111/ejss.13527 by Duodecim Medical Publications Ltd, W
iley Online Library on [10/07/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on W
iley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
T
A
B
L
E
2
C
la
y
(%
;<
2
μm
)
an
d
cl
ay
+
si
lt
co
n
te
n
t
(%
;<
20
μm
an
d
<
60
μm
)
of
th
e
so
ils
,e
st
im
at
ed
m
ax
im
um
M
A
O
C
(g
O
C
kg


1
so
il)
ac
co
rd
in
g
to
H
as
si
n
k'
s
eq
ua
ti
on
(w
it
h
cl
ay
+
si
lt
si
ze
cu
t-
of
fs
<
20
μm
an
d
<
60
μm
),
an
d
m
ax
im
um
O
C
ac
cr
u
al
es
ti
m
at
es
by
bo
un
da
ry
lin
e
an
al
ys
es
(g
O
C
kg


1
cl
ay
+
si
lt
<
60
μm
)
fo
r
1:
1
an
d
2:
1
cl
ay
m
in
er
al
do
m
in
at
ed
so
ils
,a
n
d
es
ti
m
at
ed
fe
as
ib
le
M
A
O
C
(g
M
A
O
C
kg


1
cl
ay
+
si
lt
)
in
th
e
st
ud
ie
d
so
ils
u
n
de
r
cu
rr
en
t
ag
ri
cu
lt
ur
al
m
an
ag
em
en
t
in
di
ff
er
en
t
cl
ay
cl
as
se
s
by
10
0
g
kg

1
in
te
rv
al
s.
H
as
si
n
k
’e
q
u
at
io
n
E
st
im
at
ed
m
ax
im
u
m
O
C
co
n
te
n
t
w
it
h
bo
u
n
d
ar
y
li
n
e
an
al
ys
es
C
la
y+
si
lt
<
20
μm
C
la
y+
si
lt
<
60
μm
C
la
y
(g
k
g

1 )
n
C
la
y
(<
2
μm
)
%
C
la
y
+
si
lt
(<
20
μm
)
%
C
la
y
+
si
lt
(<
60
μm
)
%
T
ot
al
O
C
(g
k
g

1
so
il
)
M
A
O
C
(g
k

1
so
il
)
E
st
im
at
ed
m
ax
im
u
m
(g
O
C
k
g

1
so
il
)
M
A
O
C
fr
om
es
ti
m
at
ed
m
ax
im
u
m
(%
)
E
st
im
at
ed
m
ax
im
u
m
(g
O
C
k
g

1
so
il
)
M
A
O
C
fr
om
es
ti
m
at
ed
m
ax
im
u
m
(%
)
D
om
in
at
in
g
cl
ay
m
in
er
al
E
st
im
at
ed
fe
as
ib
le
M
A
O
C
in
ow
n
so
il
s
C
u
rr
en
t
M
A
O
C
fr
om
th
e
es
ti
m
at
ed
fe
as
ib
le
(%
)
1:
1
2:
1
(g
O
C
k
g

1
cl
ay
+
si
lt
)
0–
10
0
14
5
±
0.
6
15
±
1.
7
31
±
4.
5
24
.3
±
3.
1a
b
20
.2
±
2.
7a
b
9.
5
±
0.
6
21
3
±
27
.4
a
15
.6
±
1.
7
13
0
±
20
.4
a
14
26
17
11
8
10
0–
20
0
12
15
±
0.
8
43
±
4.
2
70
±
5.
4
29
.0
±
3.
6a
b
25
.3
±
3.
2a
b
20
.1
±
1.
6
12
6
±
21
.6
b
30
.2
±
2.
0
84
±
12
.9
b
32
59
39
66
20
0–
30
0
12
24
±
0.
6
64
±
5.
1
76
±
5.
2
22
.6
±
1.
4a
18
.7
±
1.
3a
27
.6
±
1.
9
68
±
8.
2b
32
.1
±
1.
9
58
±
6.
7b
35
64
42
45
30
0–
40
0
23
36
±
0.
5
75
±
2.
4
87
±
2.
2
33
.9
±
2.
3b
29
.3
±
2.
1b
32
.0
±
0.
9
92
±
5.
8b
36
.2
±
0.
8
81
±
4.
9b
40
73
48
61
40
0–
50
0
8
46
±
1.
0
77
±
3.
9
85
±
3.
5
31
.1
±
3.
4a
b
27
.8
±
3.
3a
b
32
.4
±
1.
4
86
±
10
.4
b
35
.5
±
1.
3
78
±
9.
1b
39
71
47
59
50
0–
60
0
12
55
±
0.
8
80
±
2.
6
85
±
2.
8
30
.1
±
1.
2a
b
26
.2
±
1.
2a
b
33
.6
±
1.
0
78
±
3.
6b
35
.7
±
1.
0
73
±
2.
8b
39
71
47
56
60
0–
70
0
12
62
±
0.
6
85
±
1.
1
92
±
0.
9
32
.6
±
2.
7a
b
28
.0
±
2.
4a
b
35
.7
±
0.
4
78
±
7.
2b
38
.0
±
0.
3
74
±
6.
2b
42
77
51
55
N
ot
e:
L
et
te
rs
in
M
A
O
C
fr
om
es
ti
m
at
ed
m
ax
im
um
w
it
h
H
as
si
n
k'
s
eq
ua
ti
on
de
n
ot
e
st
at
is
ti
ca
ld
if
fe
re
n
ce
s
(T
uk
ey
's
H
SD
,s
ig
n
if
ic
an
ce
le
ve
l<
0.
05
),
n
=
n
um
be
r
of
fa
rm
s
sa
m
pl
ed
.O
th
er
va
ri
ab
le
s
w
er
e
n
ot
te
st
ed
fo
r
st
at
is
ti
ca
l
di
ff
er
en
ce
s.
T
h
e
h
or
iz
on
ta
ll
in
e
be
tw
ee
n
th
e
cl
ay
cl
as
se
s
of
20
0–
30
0
g
kg

1
an
d
30
0–
40
0
g
kg

1
in
di
ca
te
s
th
e
cu
t-
of
f
be
tw
ee
n
lo
w
cl
ay
(<
30
%
)
an
d
h
ig
h
cl
ay
(>
30
%
)
so
il
s.
8 of 17 SALONEN ET AL.
 13652389, 2024, 4, D
ow
nloaded from
 https://bsssjournals.onlinelibrary.wiley.com/doi/10.1111/ejss.13527 by Duodecim Medical Publications Ltd, W
iley Online Library on [10/07/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on W
iley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
found to be below their estimated saturation level. In the
>200 g clay kg
1 soils, the estimated maximum MAOC
saturation capacity was not observed.
When we divided the soils into carbon-saturated and
non-saturated soils according to Hassink's equation (with
clay+silt size cut-off <20 μm; Table 3), we observed
no fertilization (mineral vs. organic) related differences
(t-test) in carbon- and nitrogen-associated soil parameters
within the saturated or non-saturated soils.
3.3 | Soil properties, climate and
fertilization regime as explanatory
variables of total OC and its distribution
into MAOC and POC
When observing all the studied soils together (clay con-
tent 2%–68%), clay content strongly correlated with the
concentration of Fe oxides. While Al oxides did not sig-
nificantly correlate with clay content, there was a signifi-
cant correlation between Al and Fe oxides concentration
(Table 4). Likewise, in the clay <30% soils, clay and Fe
oxides correlated strongly whereas in clay >30% soils,
Fe oxides and clay did not correlate but Al oxides and
clay, as well as Fe oxides and Al oxides correlated.
For total OC and MAOC, simple linear regression
analyses for all soils together showed positive relation-
ships with clay, Al and Fe (p-values <0.05; Figure 3).
There was a significant positive relationship between soil
clay content and total OC and MAOC (p = 0.005 for
both), which accounted for 9% of the variation in the car-
bon data. Al and Fe oxides accounted for more of the var-
iation, describing around 21%–23% of the variation in
total OC and MAOC. Neither clay nor Al and Fe oxides
explained variation in POC.
Multiple regression models explained 33%–50% of the
variation in total OC and 36%–53% of the variation in
MAOC (Table 5). Yearly annual temperature (C) was
not chosen as an explanatory variable in any of the
models, nor was the clay content (g kg
1 soil), whereas
clay+silt content (g kg
1 soil) was a significant explana-
tory variable in the total OC and MAOC models (p-value
<0.001; Table 5, Table S3). Al oxide was the only explana-
tory parameter that contributed (p-value <0.05) to the
explanation of variance in all models, whereas Fe oxides
were not included in the models for soils with low
(<30%) clay content for both total OC and MAOC.
Instead, pH was important in explaining total OC and
MAOC in the clay <30% soils.
Generally, best-fit models explaining total OC and
MAOC per clay content class (< or >30% clay) were very
similar. The only deviation between the models was that
mean annual precipitation (MAP) was not a significant T
A
B
L
E
3
So
il
or
ga
n
ic
ca
rb
on
pr
op
er
ti
es
an
d
pH
of
so
ils
ar
e
di
vi
de
d
ac
co
rd
in
g
to
th
e
M
A
O
C
sa
tu
ra
ti
on
le
ve
l(
sa
tu
ra
te
d
an
d
n
on
-s
at
ur
at
ed
,r
es
pe
ct
iv
el
y,
ac
co
rd
in
g
to
H
as
si
n
k'
s
eq
ua
ti
on
w
it
h
cl
ay
+
si
lt
si
ze
cu
t-
of
f
<
20
μm
)
an
d
fe
rt
il
iz
er
ap
pl
ic
at
io
n
.
M
A
O
C
sa
tu
ra
ti
on
st
at
e
of
th
e
so
il
F
er
ti
li
za
ti
on
n
T
ot
al
O
C
(g
k
g

1 )
T
ot
al
N
(g
k
g

1 )
So
il
C
N
M
A
O
C
(g
k
g

1
so
il
)
M
A
O
M
(<
53
μm
so
il
)
C
-%
M
A
O
M
C
N
P
O
C
(g
k
g

1
so
il
)
P
O
M
(>
53
μm
so
il
)
C
-%
P
O
M
C
N
p
H
C
la
y
%
C
la
y+
si
lt
%
A
l-
ox
(g
k
g

1
so
il
)
F
e-
ox
(g
k
g

1
so
il
)
M
ea
n
H
as
si
n
k
sa
tu
ra
ti
on
(%
)
Sa
tu
ra
te
d
M
in
er
al
16
32
±
4
2.
0
±
0.
2
16
±
0.
6
28
±
3.
2
4.
9
±
0.
6
16
±
0.
6
3.
7
±
0.
5
2.
6
±
0.
8
18
±
1.
2
6.
4
±
0.
1
25
±
5
65
±
7
1.
9
±
0.
2
4.
1
±
0.
6
14
5
O
rg
an
ic
19
35
±
3
2.
2
±
0.
2
17
±
0.
7
30
±
2.
9
6.
1
±
0.
7
16
±
0.
8
4.
4
±
0.
3
3.
4
±
1.
1
21
±
1.
1
6.
2
±
0.
1
22
±
4
56
±
7
2.
5
±
0.
2
4.
0
±
0.
6
17
9
N
on
-s
at
ur
at
ed
M
in
er
al
30
27
±
1
2.
0
±
0.
1
14
±
0.
2
23
±
0.
9
2.
8
±
0.
1
13
±
0.
2
4.
3
±
0.
4
3.
7
±
0.
5
22
±
0.
7
6.
5
±
0.
1
43
±
3
84
±
2
1.
9
±
0.
1
4.
7
±
0.
3
73
O
rg
an
ic
25
26
±
1
1.
9
±
0.
1
14
±
0.
3
22
±
1.
2
2.
5
±
0.
1
13
±
0.
3
3.
7
±
0.
4
4.
3
±
0.
5
22
±
0.
8
6.
4
±
0.
1
40
±
3
88
±
2
1.
8
±
0.
1
4.
5
±
0.
2
70
N
ot
e:
W
it
h
in
sa
tu
ra
te
d
an
d
n
on
-s
at
ur
at
ed
so
ils
,n
o
st
at
is
ti
ca
lly
si
gn
if
ic
an
t
di
ff
er
en
ce
s
w
er
e
ob
se
rv
ed
be
tw
ee
n
th
e
so
il
s
w
it
h
m
in
er
al
an
d
or
ga
n
ic
fe
rt
il
iz
at
io
n
(n
=
n
u
m
be
r
of
fa
rm
s
sa
m
pl
ed
,t
ot
al
n
u
m
be
r
of
fa
rm
s
=
90
).
SALONEN ET AL. 9 of 17
 13652389, 2024, 4, D
ow
nloaded from
 https://bsssjournals.onlinelibrary.wiley.com/doi/10.1111/ejss.13527 by Duodecim Medical Publications Ltd, W
iley Online Library on [10/07/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on W
iley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
explanatory variable for MAOM in clay <30% soils,
whereas it was significant in all the other models.
4 | DISCUSSION
4.1 | OC, MAOC and POC in the soils
Contrary to our expectation, we did not find that the
implementation of organic fertilization practices resulted
in a higher total OC content of soils, when comparing to
mineral fertilization (Table S2). OC in minerally and
organically fertilized soils was on average 29.9 and 28.8 g
OC kg
1 kg soil, respectively. This result differs from
studies that showed increase in OC at least in the longer
term after changing from mineral to organic fertilization
(Alvarez, 2021; Just et al., 2023; Maillard & Angers, 2014;
Poulton et al., 2018).
There are several potential explanations for the simi-
larity of soil OC contents in our soils that differed in fer-
tilization regime. As plants are important in bringing OC
to the soil via root exudates and crop residues (Jackson
et al., 2017), it is possible that as long as there is sufficient
amount of nutrients to maintain a specific level of pri-
mary productivity, the type of fertilization (mineral
vs. organic) in some cases is not a decisive factor for the
soil OC storage (Gocke et al., 2023; Gregorich et al., 1996;
Ladha et al., 2011). However, it is likely that longer-term
data on fertilization, including the specific amounts of
OC applied with it, would have been needed to ascertain
the effects of mineral vs. organic fertilization on the OC
in our study. It is also possible that in combination with
the aforementioned limitations in our data, we did not
detect higher OC stocks in organically fertilized soils due
to relatively high baseline levels of OC in the Finnish
arable soils (Heikkinen et al., 2013). In this case, the pre-
existing OC pool may mask OC additions by organic fer-
tilization as it is difficult to discern small gradual changes
in soil OC levels (e.g., Schrumpf et al., 2011). Further-
more, one possibility is that part of the OC from the
organic fertilizers is located below the ploughed layer
(i.e., below 20 cm soil depth we studied here) as plough-
ing and leaching may translocate OC from the organic
fertilizer into deeper soil layers (e.g., Kätterer et al., 2014;
Salonen et al., 2023; Skadell et al., 2023).
Similarly, the type of fertilization had no effect on the
OC contents of the MAOM and POM fractions
(Table S1). This is in line with what has been previously
reported for the Finnish arable topsoil (0–20 cm) in field
trials comparing mineral vs organic systems after
35 years (Mayer et al., 2022) and after 24 years (Salonen
et al., 2023). This may further indicate that comparable
MAOC and POC levels can be achieved with both organic
and mineral fertilization, as long as in mineral farming
systems, there are enough nutrients to support the plant
growth providing OC inputs to the soil.
Across the studied soils, the proportion of OC in
MAOC was relatively constant, (83%–89% of the total
OC) and this result agrees with what has been deter-
mined before for the Finnish arable clay soil (83%–87% of
OC in MAOC; Salonen et al., 2023), for temperate arable
soils (86%; Begill et al., 2023) and in a global meta-
analysis (83%, Matus, 2021), indicating that soils of the
agroecosystems across differing environments and cli-
mates are having corresponding proportions of the OC in
the MAOC and POC fractions.
4.2 | Testing OC saturation concepts in
the studied soils
There was a strong positive linear relationship between
total OC and MAOC across the range of OC contents
(12 to 58 g total OC kg
1 soil, Figure S1). A similar rela-
tionship was found by Begill et al. (2023). In contrast,
some studies have shown that after a specific OC content,
the relationship between total OC and MAOC dimin-
ishes. For example, Cotrufo et al. (2019) observed this
levelling off after a total OC content of 50 g kg
1 in a set
of European soils, which was not the case in the current
study despite using a similar physical MAOM/POM frac-
tionation method. Our study was restricted to a set of
TABLE 4 Pearson correlation coefficients (r) for the soil clay content (%), aluminium (Al-ox, g kg
1) and iron (Fe-ox, g kg
1) oxides in
all studied soils, and in soils with clay <30% or clay >30%.
All soils Clay <30% Clay >30%
Clay Al-ox Fe-ox Clay Al-ox Fe-ox Clay Al-ox Fe-ox
Clay 1 Clay 1 Clay 1
Al-ox 0.16 1 Al-ox -0.28 1 Al-ox 0.30 1
Fe-ox 0.63 0.23 1 Fe-ox 0.69 0.016 1 Fe-ox 0.16 0.31 1
Note: Significant correlations bolded (p < 0.05).
10 of 17 SALONEN ET AL.
 13652389, 2024, 4, D
ow
nloaded from
 https://bsssjournals.onlinelibrary.wiley.com/doi/10.1111/ejss.13527 by Duodecim Medical Publications Ltd, W
iley Online Library on [10/07/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on W
iley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
mineral arable soils from the boreal climate zone with a
very specific pedogenic history (i.e., formed by abrasion
by ice, Koljonen, 1992). Therefore, our results possibly
better reflect soil OC saturation status in the specific case
of boreal agricultural soils.
We estimated the feasible maximum amount of
MAOC under the prevailing agricultural management for
our soils with boundary line analyses and found it to be
55 g MAOC kg
1 clay+silt (Figure 2, Table 2). This esti-
mate sits between the proposed maxima (Six et al., 2024)
FIGURE 3 Simple linear regressions for total OC, MAOC and POC (g kg
1 soil) with clay, Al and Fe oxides as explanatory variables in
fields (n = 90) with mineral (blue, n = 46) or organic (orange, n = 44) fertilization. In case of a significant relationship (p < 0.05), a
regression line was added to illustrate the relationship.
SALONEN ET AL. 11 of 17
 13652389, 2024, 4, D
ow
nloaded from
 https://bsssjournals.onlinelibrary.wiley.com/doi/10.1111/ejss.13527 by Duodecim Medical Publications Ltd, W
iley Online Library on [10/07/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on W
iley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
for 1:1 clay-mineral-dominated soils (46 g MAOC kg
1
clay+silt) and 2:1 clay-mineral-dominated soils (82 g
MAOC kg
1 clay+silt). The result of the boundary line
analyses is consistent with what could be expected based
on background knowledge of the mineralogy of Finnish
soils since they are formed by ice abrasion and are rela-
tively young (Koljonen, 1992) and non-weathered
(Keskinen et al., 2022) and thus are assumed to have
lower surface activity than 2:1 clay mineral dominated
soils. According to the boundary line analyses, studied
soils had 45%–118% of the estimated feasible MAOC
(g OC kg
1 clay+silt) under the current arable land
management, and only the lowest clay content interval
(0–100 g kg
1) was above the estimated maximum (118%
of the estimated feasible MAOC). However, it is worth
emphasizing that we do not know whether the soils
selected in this study and for the boundary line analyses
are truly representative of a maximum achievable OC
content with the current agricultural management as we
do not have time series measurements of the soils that
confirm a steady state of OC in terms of a specific level of
OC inputs. Nonetheless, since our dataset is relatively
large and contains a diverse selection of farms from the
southern half of Finland with different farming histories,
we can assume that the results reveal a reasonable esti-
mate of the maximum MAOC of <60 μm particle size
fraction under the current agricultural management.
We also assessed soil carbon saturation with Has-
sink's equation (Hassink, 1997; Table 2) with two particle
size cut-offs for clay+silt: <20 μm (as in Hassink, 1997)
and <60 μm (as theoretically, minerals up to this size
have relevant specific surface area for MAOC binding;
Six et al., 2024). With this approach, all soils with the clay
content >200 g kg
1 (72% of the studied soils) had OC
saturation deficit while soils with the lowest studied clay
content (0–100 g kg
1) were found to be above their theo-
retical maximum (i.e., >100%) MAOC saturation.
As some of the soils, with a clay content <100 g kg
1,
were found to be oversaturated with MAOC (i.e., having
more than 100% of the maximum MAOC proposed by
the Hassink's equation), it is possible that the used
approaches are not well suited for the soils in question.
These saturation estimation approaches oversimplify the
binding capacity associated with MAOC as they only take
into account soil texture, whereas there is a multitude of
other factors (such as climate, vegetation, soil chemistry,
Al and Fe oxides) affecting soil OC stabilization soils
(e.g., Guillaume et al., 2022; Rasmussen et al., 2018; Six
et al., 2002; Soinne et al., 2024).
Furthermore, observed oversaturation could also indi-
cate methodological error in the size fractionation of
MAOM (<53 μm soil) / POM (>53 μm soil). Similarities
of the C to N ratios of MAOM and POM in the low clay
soils (<100 g clay kg
1 soil, C to N ratio = 18, Table 1)
may imply POM contamination, for example, fine-sized
POM passing through the sieve to the MAOM fraction
during the MAOM/POM fractionation, which would
increase the OC content of the MAOM. Due to the rela-
tively high clay content of most of our soils, we needed to
use glass beads to assist full dispersion of the soils, and
the glass bead approach may have led to some crumbling
of the larger POM fragments into smaller, MAOM-sized
fragments (<53 μm).
Yet, it is possible that the observed similarity of the C
to N ratios of the MAOM and POM in the <100 g clay
kg
1 soil is not a sign of contamination but instead indi-
cates a chemical resemblance of MAOC and POC in the
low clay soils. Some studies have found MAOM and POM
to be chemically resembling (e.g., Chang et al., 2024; Yu
et al., 2022), and Yu et al. (2022) reported C to N ratios
TABLE 5 Best-fit regression models from AIC-driven model selection explaining total OC and MAOC.
OC fractiona Soils Regression model Adj. R 2
1 Total OC Clay <30% 0.03 clay + 5.06 Alox 
 0.86 pH 
 0.10 MAP 
 4.26 fertilization + 129.9 0.33
2 Clay >30% 
0.02 clay + 6.56 Alox+ 1.57 Feox 
 0.10 MAP + 83.2 0.50
3 All soils 4.91 Alox+ 1.91 Feox 
 3.91 pH 
 0.08 MAP + 89.8 0.44
4 All soils 0.008 clay & silt+ 5.26 Alox+ 1.26 Feox 
 3.78 pH 
 0.09 MAP + 87.5 0.46
5 MAOM-C Clay <30% 3.56 Alox 
 0.97 pH 
 0.06 MAP 
 3.94 fertilization + 116.4 0.36
6 Clay >30% 
0.019 clay + 6.34 Alox+ 1.39 Feox 
0.09 MAP + 76.4 0.53
7 All soils 4.65 Alox+ 1.69 Feox 
 4.48 pH 
0.08 MAP + 86.8 0.47
8 All soils 0.008 clay& silt+ 5.01 Alox+ 1.01 Feox 
 4.34 pH 
0.08 MAP + 84.4 0.49
Note: Models were tested for all soils together (n = 90) and for soils with a clay content <30% (n = 35) and >30% (n = 55), separately. Explanatory variables
were clay content (g kg
1), clay plus silt content (g kg
1), Al and Fe oxides (g kg
1), cumulative long-term annual precipitation (MAP; mm/year) and yearly
mean temperature (MAT; oC; 19-year average), pH and fertilization (mineral or organic, coded 0 and 1, respectively). Significant (p < 0.05) explanatory
variables bolded. More model details in Table S3.
aIn the models 1–3 and 5–7, clay as an explanatory variable. In the models 4 and 8, clay + silt as an explanatory variable.
12 of 17 SALONEN ET AL.
 13652389, 2024, 4, D
ow
nloaded from
 https://bsssjournals.onlinelibrary.wiley.com/doi/10.1111/ejss.13527 by Duodecim Medical Publications Ltd, W
iley Online Library on [10/07/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on W
iley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
corresponding to the ones in here in MAOM and POM
fractionated with the same method (size cut-off 53 μm).
Also, it has been shown that in coarse soils, POM-sized
MAOM (i.e., mineral-bound MAOC >53 μm) can have
higher C to N ratio than smaller-sized MAOM (<53 μm;
Samson et al., 2020, Sokol et al., 2022). Hence, we cannot
rule out the possibility that the resembling C to N ratios of
the MAOM and POM in the studied low clay soils are a
true finding.
4.3 | Soil properties, climate parameters
and OC, MAOC and POC
As expected, we found that Al and Fe oxides were more
strongly correlated with OC and MAOC than clay
(Figure 3). However, as simple linear regressions could
explain only 9%–23% of the variation in OC and its frac-
tions, neither clay, Al and Fe were well suited as single
predictors to explain OC and its fractions in the studied
soils. As soil OC and clay content have previously been
reported to be connected (e.g., Arrouays et al., 2006;
Kaiser & Guggenberger, 2003; Wiesmeier et al., 2019), it
was somewhat surprising that the soil clay content was
not associated strongly with OC (clay explained only 9% of
OC and MAOC variation; Figure 3). Some previous studies
have also stated that clay content is not the only or even
an important determinant of OC (Prout et al., 2021;
Rasmussen et al., 2018; Salonen et al., 2023). One possible
explanation for this in the current study relates to the rela-
tively low primary productivity of soils with a high clay
content. As very clayey soils in Finland tend to have lower
yield levels than coarser soils, there is possibly a lower
plant-derived OC input in the soil (Soinne et al., 2021)
which could lead to lower SOM levels compared to soils
with a higher primary productivity.
While Al and Fe oxides had a stronger correlation
with OC and MAOC than the soil clay content, also they
both remained poor single predictors for the amount of
soil carbon. Other studies from similar soils have found
larger correlations between Al oxides and OC. Rasmus-
sen et al. (2018) found Al and Fe oxides to be most impor-
tant predictors of OC in humid climates. In our data, Al
oxides explained 21% of the variation in OC (clay content
2%–68%), whereas in a Swedish field experiment, they
explained 48% in soils with a clay content of 17%–42%
(Fukumasu et al., 2021). However, our soils estimated to
be MAOC-saturated had relatively more Al and Fe
oxides per gram of clay or clay+silt than non-saturated-
soils (Table 3), indicating a close relationship between
oxides and MAOC (Hall & Thompson, 2022; Rasmussen
et al., 2018). In the multiple regression models (Table 5),
Al oxides were particularly important as an explanatory
variable in all the models, whereas Fe oxides were sig-
nificant for OC and MAOC only in soils with high clay
content (>30%). Soil pH correlated negatively with OC
and MAOC in soils with low clay (<30%). This is consis-
tent with previous research, as low soil pH has been
found to correlate with slower OC decomposition
(Keiluweit et al., 2016; Slessarev et al., 2016) and OC
accumulation (Lugato et al., 2021). On the other hand,
low pH could lead to lower primary productivity and
lower OC input via roots (Young et al., 2021), possibly
lowering soil OC.
Consistent with the simple linear regressions, clay
content was not a significant predictor in any of the mul-
tiple regression models. However, clay+silt was a signifi-
cant predictor of OC and MAOC when all the studied
soils were included in the analysis. It is possible that this
is a consequence of higher primary productivity and
plant-derived OC inputs in coarser soils (Soinne
et al., 2021), as lower porosity of the more clayey soils
may limit plant OC inputs.
In a global meta-analysis, Jobbagy and Jackson (2000)
reported that soil total OC content increased with precipi-
tation and decreased with temperature. Likewise, Hansen
et al. (2024) found that MAP was connected to higher
MAOC levels on a global scale. Opposite to these findings
and our expectations, OC decreased with increasing pre-
cipitation, and clay and temperature had no relationship
with OC or MAOC in our multiple regression analyses. It
is possible that the range of mean temperature in our
data (3.3–6.4C) is not wide enough to make a noticeable
difference in the soil OC content as for the large part of
year, the studied soils are far from the optimal tempera-
ture for the microbial growth (20–25C for Boreal agricul-
tural soil; Pietikäinen et al., 2005) which halts OC
decomposition and transformation processes.
Our finding that the mean annual precipitation
(range 548–725 mm year
1) was negatively correlated
with OC and MAOC was unexpected as greater MAP has
been previously connected to larger OC as prevalent
moist conditions can cause anoxic conditions which hin-
der SOM decomposition (Jobbagy & Jackson, 2000;
Wiesmeier et al., 2019). On the other hand, prevalent rain
may restrain OC inputs into the soil as it may limit plant
growth and consequently curb OC inputs to the soil due
to waterlogging and following anoxic conditions
(Drew, 1997). As another example, temperature affects
yields, and higher temperature could lead to higher OC
inputs to the soils via roots. In addition, temperature
affects microbial activity and higher temperature could
lead to a faster decomposition of SOM. These examples
illustrate that many of the factors affecting OC are inter-
connected. Therefore, reliably disentangling effects that
they individually have on OC is challenging. In addition,
SALONEN ET AL. 13 of 17
 13652389, 2024, 4, D
ow
nloaded from
 https://bsssjournals.onlinelibrary.wiley.com/doi/10.1111/ejss.13527 by Duodecim Medical Publications Ltd, W
iley Online Library on [10/07/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on W
iley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
climate-related gradients can be further confounding
assessing the effects of the individual factors.
All in all, the best-fit models explained 33%–50% of
the variation in total OC and 36%–53% of the variation in
MAOC. The relatively low model fit could indicate that
other factors explaining soil OC that were not included in
this study (such as plant productivity, topography, long-
term land use and fertilizing history) could affect
OC/MAOC levels in our soils. In particular, data on soil
productivity/ yield levels could improve the models as
most of our soils were found to have low carbon satura-
tion, with the implication that mineral adsorption sur-
faces were not a limiting factor.
5 | CONCLUSIONS
We fractionated soil from 93 arable fields with a physical
MAOM (<53 μm) / POM (>53 μm) fractionation method.
The total amount of OC (12–58 g kg
1 soil) was not
affected by the type of fertilization (mineral vs organic)
and the share of OC in MAOC (86 ± 3%) was found to be
relatively stable and also not affected by fertilizer type.
The largest total OC and MAOC contents were measured
in high clay soils, but considering the general assumption
of the significant role of clay in defining the protective
capacity for MAOC, we observed surprisingly high
MAOC contents even in the low clay soils. The applied
soil OC saturation assessment with Hassink's equation
revealed that the lowest clay content soils were above
their predicted maximum (i.e., >100%) MAOC saturation,
and as expected, soils with higher clay content have fur-
ther OC accrual potential. Similarly, based on the bound-
ary analysis, the low clay soils had the highest carbon
saturation. As the low clay soils seemed to be oversatu-
rated with OC/MAOC, it is possible that the applied OC
saturation estimation approaches misjudge the
OC accrual capacity of these soils because they rely
merely on soil textural information. Also, as clay was not
an important predictor for soil OC in the multiple regres-
sion models, there are indications that in the studied, rel-
atively moderately weathered soils, clay content is not a
decisive factor for the amount of stabilized OC. In our
soils, single soil properties (clay, Al and Fe oxides) were
not well-suited as predictors for total OC or MAOC, but
multiple regression models revealed that soil parameters
(in particular Al and Fe oxides and pH) and mean annual
precipitation combined could explain more of the varia-
tion in soil carbon. Multiple regression models at their
best explained 53% of the variation of MAOC, indicating
that there are other context-dependent factors (such as
long-term land use history) contributing to soil OC-
related processes that deserve attention in further studies.
AUTHOR CONTRIBUTIONS
Anna-Reetta Salonen: Conceptualization; investigation;
writing – original draft; formal analysis; data curation.
Ron de Goede: Conceptualization; writing – original
draft; writing – review and editing; formal analysis.
Rachel Creamer: Conceptualization; writing – review
and editing. Jussi Heinonsalo: Conceptualization;
writing – original draft; writing – review and editing;
funding acquisition. Helena Soinne: Conceptualization;
writing – original draft; writing – review and editing.
ACKNOWLEDGEMENTS
We would like to thank all the farmers participating in
the Carbon Action project and Jenni Jääskeläinen, from
BSAG for helping with the soil samples and data. We
also thank Eija Hagelberg and Tuomas Mattila, and
everyone at BSAG for their work that enables the Car-
bon Action project, as well as all the Carbon Action
farmers for their participation in the project. We thank
Julius Vira and Liisa Kulmala from the Finnish Meteo-
rological Institute for providing the weather data. We
are grateful to Oona Uhlgren, Aku Pakarinen, Miia
Collander, Janni Ryynälä and Reija Heinonen
(University of Helsinki) for skillful technical assistance
with the laboratory analyses and Kari Koppelmäki for
providing some helpful guidance. We sincerely thank
Mirjam Breure, Carmen Vazquez Martin and the C
group: Guusje Koorneef, Margot Lahens, Karen Moran
Rivera, Jonas Schepens, Shevani Murray, Sophie Plan-
chenaut and Gabriel Moinet (Wageningen University &
Research) for providing valuable comments relating to
this work.
FUNDING INFORMATION
This study was funded by the Strategic Research Council
(SRC) at the Academy of Finland as part of the project
‘Multi-benefit solutions to climate-smart agriculture’
(MULTA, grant numbers 352435 (Natural Resources Insti-
tute Finland) and 352436 (University of Helsinki)), by
Maa-ja vesitekniikan Tuki ry (grant number 4268), and
Drainage Foundation sr (grant number H-9-2020-8-12).
CONFLICT OF INTEREST STATEMENT
The authors have no conflicts of interest to declare.
DATA AVAILABILITY STATEMENT
The data that support the findings of this study are avail-
able from the corresponding author upon reasonable
request.
ORCID
Anna-Reetta Salonen https://orcid.org/0000-0002-3189-
5105
14 of 17 SALONEN ET AL.
 13652389, 2024, 4, D
ow
nloaded from
 https://bsssjournals.onlinelibrary.wiley.com/doi/10.1111/ejss.13527 by Duodecim Medical Publications Ltd, W
iley Online Library on [10/07/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on W
iley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Ron de Goede https://orcid.org/0000-0002-9786-5187
Rachel Creamer https://orcid.org/0000-0003-3617-1357
Helena Soinne https://orcid.org/0000-0002-7965-6496
REFERENCES
Alvarez, R. (2021). Organic farming does not increase soil organic
carbon compared to conventional farming if there is no carbon
transfer from other agroecosystems. A meta-analysis. Soil
Research, 60(3), 211–223.
Angst, G., Mueller, K. E., Castellano, M. J., Vogel, C.,
Wiesmeier, M., & Mueller, C. W. (2023). Unlocking complex
soil systems as carbon sinks: Multi-pool management as the
key. Nature Communications, 14(1), 2967.
Angst, G., Mueller, K. E., Nierop, K. G., & Simpson, M. J. (2021).
Plant-or microbial-derived? A review on the molecular compo-
sition of stabilized soil organic matter. Soil Biology and Bio-
chemistry, 156, 108189.
Arrouays, D., Saby, N., Walter, C., Lemercier, B., & Schvartz, C.
(2006). Relationships between particle-size distribution and
organic carbon in French arable topsoils. Soil Use and Manage-
ment, 22(1), 48–51.
Barton, K. (2023). MuMIn: Multi-Model Inference. R package
version 1.47.5. https://CRAN.R-project.org/package=MuMIn
Begill, N., Don, A., & Poeplau, C. (2023). No detectable upper limit
of mineral-associated organic carbon in temperate agricultural
soils. Global Change Biology., 29, 4662–4669.
Berthelin, J., Laba, M., Lemaire, G., Powlson, D., Tessier, D.,
Wander, M., & Baveye, P. C. (2022). Soil carbon sequestration
for climate change mitigation: Mineralization kinetics of
organic inputs as an overlooked limitation. European Journal
of Soil Science, 73(1), e13221.
Besnard, E., Chenu, C., Balesdent, J., Puget, P., & Arrouays, D.
(1996). Fate of particulate organic matter in soil aggregates dur-
ing cultivation. European Journal of Soil Science, 47(4), 495–503.
Cambardella, C. A., & Elliott, E. T. (1992). Particulate soil organic-
matter changes across a grassland cultivation sequence. Soil
Science Society of America Journal, 56(3), 777–783.
Chang, Y., Sokol, N. W., van Groenigen, K. J., Bradford, M. A.,
Ji, D., Crowther, T. W., & Ding, F. (2024). A stoichiometric
approach to estimate sources of mineral-associated soil organic
matter. Global Change Biology, 30(1), e17092.
Cotrufo, M. F., Ranalli, M. G., Haddix, M. L., Six, J., & Lugato, E.
(2019). Soil carbon storage informed by particulate and
mineral-associated organic matter. Nature Geoscience, 12(12),
989–994.
Dexter, A. R., Richard, G., Arrouays, D., Czyz, E. A., Jolivet, C., &
Duval, O. (2008). Complexed organic matter controls soil physi-
cal properties. Geoderma, 144(3–4), 620–627.
Drew, M. C. (1997). Oxygen deficiency and root metabolism: Injury
and acclimation under hypoxia and anoxia. Annual Review of
Plant Biology, 48(1), 223–250.
Elonen, P. (1971). Particle-size analysis of soils. Acta Agr.Fenn., 122,
1–122.
Feng, W., Plante, A. F., & Six, J. (2013). Improving estimates of
maximal organic carbon stabilization by fine soil particles. Bio-
geochemistry, 112, 81–93.
Fox, J., & Weisberg, S. (2019). An R companion to applied regression
(Third ed.). Sage.
Fukumasu, J., Poeplau, C., Coucheney, E., Jarvis, N., Klöffel, T.,
Koestel, J., & Larsbo, M. (2021). Oxalate-extractable aluminum
alongside carbon inputs may be a major determinant for
organic carbon content in agricultural topsoils in humid conti-
nental climate. Geoderma, 402, 115345.
Georgiou, K., Jackson, R. B., Vindusˇkova, O., Abramoff, R. Z.,
Ahlström, A., Feng, W., & Torn, M. S. (2022). Global stocks and
capacity of mineral-associated soil organic carbon. Nature Com-
munications, 13(1), 3797.
Gerke, J. (1992). Phosphate, aluminium and iron in the soil solution
of three different soils in relation to varying concentrations of
citric acid. Zeitschrift für Pflanzenernährung Und Bodenkunde,
155(4), 339–343.
Gocke, M. I., Guigue, J., Bauke, S. L., Barkusky, D., Baumecker, M.,
Berns, A. E., & Amelung, W. (2023). Interactive effects of agricul-
tural management on soil organic carbon accrual: A synthesis of
long-term field experiments in Germany. Geoderma, 438, 116616.
Gregorich, E. G., Liang, B. C., Ellert, B. H., & Drury, C. F. (1996).
Fertilization effects on soil organic matter turnover and corn
residue C storage. Soil Science Society of America Journal, 60(2),
472–476.
Guillaume, T., Makowski, D., Libohova, Z., Bragazza, L.,
Sallaku, F., & Sinaj, S. (2022). Soil organic carbon saturation in
cropland-grassland systems: Storage potential and soil quality.
Geoderma, 406, 115529.
Hall, S. J., & Thompson, A. (2022). What do relationships between
extractable metals and soil organic carbon concentrations
mean? Soil Science Society of America Journal, 86(2), 195–208.
Hansen, P. M., Even, R., King, A. E., Lavallee, J., Schipanski, M., &
Cotrufo, M. F. (2024). Distinct, direct and climate-mediated envi-
ronmental controls on global particulate and mineral-associated
organic carbon storage. Global Change Biology, 30(1), e17080.
Hassink, J. (1997). The capacity of soils to preserve organic C and N
by their association with clay and silt particles. Plant and Soil,
191, 77–87.
Heikkinen, J., Ketoja, E., Nuutinen, V., & Regina, K. (2013). Declin-
ing trend of carbon in Finnish cropland soils in 1974–2009.
Global Change Biology, 19(5), 1456–1469.
Hoffland, E., Kuyper, T. W., Comans, R. N., & Creamer, R. E.
(2020). Eco-functionality of organic matter in soils. Plant and
Soil, 455, 1–22.
Jackson, R. B., Lajtha, K., Crow, S. E., Hugelius, G.,
Kramer, M. G., & Piñeiro, G. (2017). The ecology of soil carbon:
Pools, vulnerabilities, and biotic and abiotic controls. Annual
Review of Ecology, Evolution, and Systematics, 48, 419–445.
Jobbagy, E. G., & Jackson, R. B. (2000). The vertical distribution of
soil organic carbon and its relation to climate and vegetation.
Ecological Applications, 10(2), 423–436.
Just, C., Armbruster, M., Barkusky, D., Baumecker, M.,
Diepolder, M., Döring, T. F., & Wiesmeier, M. (2023). Soil
organic carbon sequestration in agricultural long-term field
experiments as derived from particulate and mineral-associated
organic matter. Geoderma, 434, 116472.
Kaiser, K., & Guggenberger, G. (2003). Mineral surfaces and soil
organic matter. European Journal of Soil Science, 54(2),
219–236.
Kassambara, A. (2020). Ggpubr: ‘ggplot2’ based publication ready
plots. R package version 0.4.0.999. https://rpkgs.datanovia.
com/ggpubr/
SALONEN ET AL. 15 of 17
 13652389, 2024, 4, D
ow
nloaded from
 https://bsssjournals.onlinelibrary.wiley.com/doi/10.1111/ejss.13527 by Duodecim Medical Publications Ltd, W
iley Online Library on [10/07/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on W
iley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Kassambara, A. (2022). Rstatix: pipe-friendly framework for basic
statistical tests. 2021.
Kätterer, T., Börjesson, G., & Kirchmann, H. (2014). Changes in
organic carbon in topsoil and subsoil and microbial community
composition caused by repeated additions of organic amend-
ments and N fertilisation in a long-term field experiment in
Sweden. Agriculture, Ecosystems & Environment, 189, 110–118.
Keenan, T. F., & Williams, C. A. (2018). The terrestrial carbon sink.
Annual Review of Environment and Resources, 43, 219–243.
Keiluweit, M., Nico, P. S., Kleber, M., & Fendorf, S. (2016). Are oxy-
gen limitations under recognized regulators of organic carbon
turnover in upland soils? Biogeochemistry, 127, 157–171.
Keskinen, R., Hillier, S., Liski, E., Nuutinen, V., Nyambura, M., &
Tiljander, M. (2022). Mineral composition and its relations to
readily available element concentrations in cultivated soils of
Finland. Acta Agriculturae Scandinavica, section B—Soil. Plant
Science, 72(1), 751–760.
Kögel-Knabner, I., Ekschmitt, K., Flessa, H., Guggenberger, G.,
Matzner, E., Marschner, B., & von Lützow, M. (2008). An inte-
grative approach of organic matter stabilization in temperate
soils: Linking chemistry, physics, and biology. Journal of Plant
Nutrition and Soil Science, 171(1), 5–13.
Koljonen, T. (Ed.). (1992). The geochemical atlas of Finland part 2:
Till (p. 218). Geological Survey of Finland.
Ladha, J. K., Reddy, C. K., Padre, A. T., & van Kessel, C. (2011). Role
of nitrogen fertilization in sustaining organic matter in cultivated
soils. Journal of Environmental Quality, 40(6), 1756–1766.
Lavallee, J. M., Soong, J. L., & Cotrufo, M. F. (2020). Conceptualiz-
ing soil organic matter into particulate and mineral-associated
forms to address global change in the 21st century. Global
Change Biology, 26(1), 261–273.
Lugato, E., Lavallee, J. M., Haddix, M. L., Panagos, P., &
Cotrufo, M. F. (2021). Different climate sensitivity of particulate
and mineral-associated soil organic matter. Nature Geoscience,
14(5), 295–300.
Maillard, E., & Angers, D. A. (2014). Animal manure application
and soil organic carbon stocks: A meta-analysis. Global Change
Biology, 20(2), 666–679.
Mattila, T. J., Hagelberg, E., Söderlund, S., & Joona, J. (2022). How
farmers approach soil carbon sequestration? Lessons learned
from 105 carbon-farming plans. Soil and Tillage Research, 215,
105204.
Matus, F. J. (2021). Fine silt and clay content is the main factor
defining maximal C and N accumulations in soils: A meta-anal-
ysis. Scientific Reports, 11(1), 6438.
Mayer, M., Krause, H. M., Fliessbach, A., Mäder, P., & Steffens, M.
(2022). Fertilizer quality and labile soil organic matter fractions
are vital for organic carbon sequestration in temperate arable soils
within a long-term trial in Switzerland. Geoderma, 426, 116080.
Mendez, J. C., Hiemstra, T., & Koopmans, G. F. (2020). Assessing
the reactive surface area of soils and the association of soil
organic carbon with natural oxide nanoparticles using ferrihy-
drite as proxy. Environmental Science & Technology, 54(19),
11990–12000.
Metzger, M. J., Shkaruba, A. D., Jongman, R. H. G., &
Bunce, R. G. H. (2012). Descriptions of the European environ-
mental zones and strata No. 2281. Alterra.
Mitchell, J. K., & Soga, K. (2005). Fundamentals of soil behavior
(Vol. 3). John Wiley & Sons.
Nelson, D. W., & Sommers, L. E. (1982). Total carbon, organic car-
bon, and organic matter. In Methods of soil analysis: Part 2
chemical and microbiological properties (Vol. 9, pp. 539–579).
Niskanen, R. (1989). Extractable aluminium, iron and manganese
in mineral soils: III comparison of extraction methods. Agricul-
tural and Food Science, 61(2), 89–97.
Paasonen-Kivekäs, M., Peltomaa, R., Vakkilainen, P., &
Äijö, H. (2009). Maan vesi-ja ravinnetalous. Ojitus, kastelu
ja ympäristö. Salaojayhdistys ry. Gummerus Kirjapaino Oy,
Jyväskylä.
Paustian, K., Lehmann, J., Ogle, S., Reay, D., Robertson, G. P., &
Smith, P. (2016). Climate-smart soils. Nature, 532(7597),
49–57.
Pietikäinen, J., Pettersson, M., & Bååth, E. (2005). Comparison of
temperature effects on soil respiration and bacterial and fungal
growth rates. FEMS Microbiology Ecology, 52(1), 49–58.
Poulton, P., Johnston, J., Macdonald, A., White, R., & Powlson, D.
(2018). Major limitations to achieving “4 per 1000” increases in
soil organic carbon stock in temperate regions: Evidence from
long-term experiments at Rothamsted research, United
Kingdom. Global Change Biology, 24(6), 2563–2584.
Powlson, D. S., & Neal, A. L. (2021). Influence of organic matter on
soil properties: By how much can organic carbon be increased
in arable soils and can changes be measured? In Proceedings of
the international Fertiliser society (Vol. 862, pp. 1–32). Interna-
tional Fertiliser Society (IFS).
Prout, J. M., Shepherd, K. D., McGrath, S. P., Kirk, G. J., &
Haefele, S. M. (2021). What is a good level of soil organic mat-
ter? An index based on organic carbon to clay ratio. European
Journal of Soil Science, 72(6), 2493–2503.
R Core Team. (2023). R: A language and environment for statistical
computing. R Foundation for Statistical Computing. https://
www.R-project.org/
Rasmussen, C., Heckman, K., Wieder, W. R., Keiluweit, M.,
Lawrence, C. R., Berhe, A. A., & Wagai, R. (2018). Beyond clay:
Towards an improved set of variables for predicting soil organic
matter content. Biogeochemistry, 137, 297–306.
Salonen, A. R., Soinne, H., Creamer, R., Lemola, R., Ruoho, N.,
Uhlgren, O., & Heinonsalo, J. (2023). Assessing the effect of
arable management practices on carbon storage and fractions
after 24 years in boreal conditions of Finland. Geoderma
Regional, 34, e00678.
Samson, M. E., Chantigny, M. H., Vanasse, A., Menasseri-
Aubry, S., & Angers, D. A. (2020). Coarse mineral-associated
organic matter is a pivotal fraction for SOM formation and is
sensitive to the quality of organic inputs. Soil Biology and Bio-
chemistry, 149, 107935.
Sanderman, J., Hengl, T., & Fiske, G. J. (2017). Soil carbon debt of
12,000 years of human land use. Proceedings of the National
Academy of Sciences, 114(36), 9575–9580.
Schjønning, P., de Jonge, L. W., Munkholm, L. J., Moldrup, P.,
Christensen, B. T., & Olesen, J. E. (2012). Clay dispersibility
and soil friability—testing the soil clay-to-carbon saturation
concept. Vadose Zone Journal, 11(1), vzj2011-0067. https://doi.
org/10.2136/vzj2011.0067
Schmidt, M. W., Torn, M. S., Abiven, S., Dittmar, T.,
Guggenberger, G., Janssens, I. A., & Trumbore, S. E. (2011).
Persistence of soil organic matter as an ecosystem property.
Nature, 478(7367), 49–56.
16 of 17 SALONEN ET AL.
 13652389, 2024, 4, D
ow
nloaded from
 https://bsssjournals.onlinelibrary.wiley.com/doi/10.1111/ejss.13527 by Duodecim Medical Publications Ltd, W
iley Online Library on [10/07/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on W
iley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Schneider, M. P. W., Scheel, T., Mikutta, R., Van Hees, P.,
Kaiser, K., & Kalbitz, K. (2010). Sorptive stabilization of organic
matter by amorphous Al hydroxide. Geochimica et Cosmochi-
mica Acta, 74(5), 1606–1619.
Schrumpf, M., Schulze, E. D., Kaiser, K., & Schumacher, J. (2011).
How accurately can soil organic carbon stocks and stock changes
be quantified by soil inventories? Biogeosciences, 8(5), 1193–1212.
Simpson, M. J., & Simpson, A. J. (2012). The chemical ecology of
soil organic matter molecular constituents. Journal of Chemical
Ecology, 38, 768–784.
Six, J., Conant, R. T., Paul, E. A., & Paustian, K. (2002). Stabiliza-
tion mechanisms of soil organic matter: Implications for
C-saturation of soils. Plant and Soil, 241, 155–176.
Six, J., Doetterl, S., Laub, M., Müller, C. R., & Van de Broek, M.
(2024). The six rights of how and when to test for soil C satura-
tion. The Soil, 10(1), 275–279.
Skadell, L. E., Schneider, F., Gocke, M. I., Guigue, J., Amelung, W.,
Bauke, S. L., & Don, A. (2023). Twenty percent of agricultural
management effects on organic carbon stocks occur in
subsoils–results of ten long-term experiments. Agriculture, Eco-
systems & Environment, 356, 108619.
Slessarev, E. W., Lin, Y., Bingham, N. L., Johnson, J. E., Dai, Y.,
Schimel, J. P., & Chadwick, O. A. (2016). Water balance creates
a threshold in soil pH at the global scale. Nature, 540(7634),
567–569.
Soinne, H., Hyyrynen, M., Jokubė, M., Keskinen, R., Hyväluoma, J.,
Pihlainen, S., & Heikkinen, J. (2024). High organic carbon con-
tent constricts the potential for stable organic carbon accrual in
mineral agricultural soils in Finland. Journal of Environmental
Management, 352, 119945.
Soinne, H., Keskinen, R., Räty, M., Kanerva, S., Turtola, E.,
Kaseva, J., & Salo, T. (2021). Soil organic carbon and clay con-
tent as deciding factors for net nitrogen mineralization and
cereal yields in boreal mineral soils. European Journal of Soil
Science, 72(4), 1497–1512.
Sokol, N. W., Whalen, E. D., Jilling, A., Kallenbach, C., Pett-
Ridge, J., & Georgiou, K. (2022). Global distribution, formation
and fate of mineral-associated soil organic matter under a
changing climate: A trait-based perspective. Functional Ecology,
36(6), 1411–1429.
Tamrat, W. Z., Rose, J., Grauby, O., Doelsch, E., Levard, C.,
Chaurand, P., & Basile-Doelsch, I. (2019). Soil organo-mineral
associations formed by co-precipitation of Fe, Si and Al in pres-
ence of organic ligands. Geochimica et Cosmochimica Acta, 260,
15–28.
Tiessen, H. J. W. B., & Stewart, J. W. B. (1983). Particle-size frac-
tions and their use in studies of soil organic matter:
II. Cultivation effects on organic matter composition in size
fractions. Soil Science Society of America Journal, 47(3),
509–514.
Totsche, K. U., Amelung, W., Gerzabek, M. H., Guggenberger, G.,
Klumpp, E., Knief, C., Lehndorff, E., Mikutta, R., Peth, S.,
Prechtel, A., & Ray, N. (2018). Microaggregates in soils. Journal
of Plant Nutrition and Soil Science, 181(1), 104–136.
Urbanski, L., Kalbitz, K., Rethemeyer, J., Schad, P., & Kögel-
Knabner, I. (2023). Unexpected high alkyl carbon contents in
organic matter-rich sandy agricultural soils of northwest Cen-
tral Europe. Geoderma, 439, 116695.
Wickham, H., Chang, W., & Wickham, M. H. (2016). Package
‘ggplot2’. Create elegant data visualisations using the grammar
of graphics. Version, 2(1), 1-189.
Wiesmeier, M., Urbanski, L., Hobley, E., Lang, B., von Lützow, M.,
Marin-Spiotta, E., & Kögel-Knabner, I. (2019). Soil organic car-
bon storage as a key function of soils-a review of drivers and
indicators at various scales. Geoderma, 333, 149–162.
Wiseman, C. L. S., & Püttmann, W. (2006). Interactions between
mineral phases in the preservation of soil organic matter. Geo-
derma, 134(1–2), 109–118.
Yli-Halla, M., & Mokma, D. L. (2002). Problems encountered when
classifying the soils of Finland.
Young, M. D., Ros, G. H., & de Vries, W. (2021). Impacts of agro-
nomic measures on crop, soil, and environmental indicators: A
review and synthesis of meta-analysis. Agriculture, Ecosystems &
Environment, 319, 107551.
Yu, W., Huang, W., Weintraub-Leff, S. R., & Hall, S. J. (2022).
Where and why do particulate organic matter (POM) and
mineral-associated organic matter (MAOM) differ among
diverse soils? Soil Biology and Biochemistry, 172, 108756.
SUPPORTING INFORMATION
Additional supporting information can be found online
in the Supporting Information section at the end of this
article.
How to cite this article: Salonen, A.-R.,
de Goede, R., Creamer, R., Heinonsalo, J., &
Soinne, H. (2024). Soil organic carbon fractions
and storage potential in Finnish arable soils.
European Journal of Soil Science, 75(4), e13527.
https://doi.org/10.1111/ejss.13527
SALONEN ET AL. 17 of 17
 13652389, 2024, 4, D
ow
nloaded from
 https://bsssjournals.onlinelibrary.wiley.com/doi/10.1111/ejss.13527 by Duodecim Medical Publications Ltd, W
iley Online Library on [10/07/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on W
iley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License