Contribution of roots and amendments to

Agriculture, Ecosystems and Environment 200 (2015) 79–87

Contents lists available at ScienceDirect

Agriculture, Ecosystems and Environment
journal homepage: www.elsevier.com/locate/agee

Contribution of roots and amendments to soil carbon accumulation
within the soil profile in a long-term field experiment in Sweden
Lorenzo Menichetti a, *, Alf Ekblad b , Thomas Kätterer c
a

Swedish University of Agricultural Sciences, Department of Soil and Environment, P.O. Box 7014, Uppsala 75007, Sweden
School of Science and Technology, Örebro University, Örebro 70182, Sweden
c
Swedish University of Agricultural Sciences, Department of Ecology, P.O. Box 7044, Uppsala 75007, Sweden
b

A R T I C L E I N F O

A B S T R A C T

Article history:
Received 5 December 2013
Received in revised form 30 October 2014
Accepted 3 November 2014
Available online xxx

The contribution of different C inputs to organic carbon accumulation within the soil profile in the Ultuna
long-term continuous soil organic matter experiment, established in 1956, was determined. Until 1999,
C3-crops were grown at the site, but since then maize (C4) has been the only crop. The effect of a total of
10 different inorganic nitrogen and organic amendment treatments (4 Mg C ha 1 yr 1) on SOC in topsoil
and subsoil after 53 years was evaluated and the contribution from maize roots to SOC after 10 years of
cultivation was estimated.
Soil organic carbon (SOC) and d13C signature were measured down to 50 cm depth. The C content in the
topsoil (0–20 cm depth) was 1.5% at the start of the experiment. After 53 years of treatments, the average
topsoil C content varied between 0.9 and 3.8% of soil dry weight, with the open fallow having the lowest
and the peat amended the highest value. Nitrogen seemed to promote C accumulation in the topsoil
treatment effects were smaller below 20 cm depth and only two of the amendments (peat and sewage
sludge) significantly affected SOC content down to 35 cm depth. Despite this, penetrometer measurements showed significant treatment differences of compaction below 41 cm depth, and although we
could not explain these differences this presented some evidence of an initial treatment-induced subsoil

differentiation. Ten years of maize growth affected the d13C of SOC down to 22.5 cm depth, where it varied
between 25.16 and 26.33(m), and an isotopic mass balance calculation suggested that maize C
accounted for 4–8% of total SOC in the topsoil. Until less than 2500 years ago the site was a post-glacial sea
floor and the 14C data suggest that marine sediment C still dominates the SOC in deeper soil layers.
Overall, the results suggest that 53 years of treatments has caused dramatic changes on the stored C in the
topsoil in several of the treatments, while the changes in the subsoil is much less dramatic and a small C
accumulation in the upper subsoil was found in two of the treatments.
The contribution from roots to SOC accumulation was generally equal to or greater than the
contribution from amendments. The retention coefficient of root-derived C in the topsoil was on average
0.30  0.09, which is higher than usually reported in the literature for plant residues but confirms
previous findings for the same experiment using another approach. This strengthens the conclusion that
root-derived SOC contributed more to SOC than above-ground crop residues.
ã 2014 Elsevier B.V. All rights reserved.

Keywords:

d13C
Root humification
SOC
Subsoil

Topsoil

1. Introduction
The global carbon (C) sink is expected to increase with the
increase in primary productivity driven by higher temperatures
and CO2 concentrations (Kirschbaum, 2000). However, a simultaneous increase in emissions due to land use change or an increase
in soil respiration after an increase in temperature may counteract
almost all this positive effect (Eglin et al., 2010). More than one-

* Corresponding author. Tel.: +46 7 68549268; fax: +46 18673156.
E-mail address: Lorenzo.Menichetti@slu.se (L. Menichetti).
http://dx.doi.org/10.1016/j.agee.2014.11.003
0167-8809/ ã 2014 Elsevier B.V. All rights reserved.

third (37%) of global land is used in agriculture and 10% of global
land is under annual crops (FAO Statistical Database, 2013), and
therefore exposed to quick changes in management. On agricultural land, the range of possible interventions for climate change
mitigation is constrained by the global requirements for food
production (Powell and Lenton, 2012), but C sequestration in
agricultural soils can be increased through changes in management practices (Lal, 2004; Kätterer et al., 2013; Stockmann et al.,

2013).
Most previous studies on C sequestration have focused on the
topsoil (e.g., Lorenz and Lal, 2005). Although this is understandable, since the concentrations and turnover of SOC are usually

80

L. Menichetti et al. / Agriculture, Ecosystems and Environment 200 (2015) 79–87

higher in topsoils than in subsoils, globally subsoils store more
than half of total SOC (Jobbagy and Jackson, 2000) and are
therefore potentially important for C sequestration strategies. A
few studies have been conducted (e.g., Paul et al., 1997; Jobbagy
and Jackson, 2000; Bird and Kracht, 2003; Jenkinson and Coleman,
2008; Jenkinson et al., 2008; Kirchmann et al., 2013), but there is
still relatively large uncertainty regarding C in deeper soil layers
(Lorenz and Lal, 2005). Areas of uncertainty include the response
to changes in management of a large part of the C stored in soil
(Poeplau et al., 2011) and its stabilization mechanisms (Chabbi
et al., 2009), making precise quantitative predictions difficult. It is
therefore important to obtain accurate information on the

reactivity of different C pools to management changes and on
how different C inputs contribute to the formation of SOC over the
whole soil profile. In particular, the contribution of root-derived
material can be particularly relevant for SOC accumulation because
of the possible associated protection mechanisms (Rasse et al.,
2005), and might so far be underestimated in the literature
(Kätterer et al., 2011).
During recent decades, long-term experiments have been
shown to produce information on the formation of SOC stocks
from different organic C inputs (Kätterer et al., 2011) and on how
the older SOC decays (Barré et al., 2010). Therefore in this study we
used data from the Ultuna long-term field experiment, established
in Sweden in 1956 and managed since then by additions of a fixed
amount of C in the form of different amendments, in combination
with or without nitrogen (N) fertilization.
The aim of the study was to examine the following questions:
(1) How have N fertilization and addition of organic amendments
affected SOC accumulation in topsoil and subsoil in the Ultuna
long-term field experiment? and (2) How have C inputs from roots
contributed to SOC formation and accumulation?

To determine how management practices affect soil C
accumulation or release in agricultural topsoil and subsoil, we
analyzed soil cores from 0 to 50 cm depth in 10 different
experimental treatments. This allowed us to evaluate the effect
of different C inputs on SOC stocks on the whole soil profile since
the start of the experiment in 1956.
To consider the inputs from roots, we based our investigation on
the fact that the crops cultivated in the experiment shifted from C3
to C4 crops in the year 2000. Several authors have tried to
determine the decay of SOC of different ages by exploiting the
natural difference in 13C content of plants with C3 and C4
photosynthetic cycles (Wynn and Bird, 2007; Blagodatskaya
et al., 2011). The analysis of d13C signatures down the whole soil
profile allowed us to quantify the contribution to SOC from maize
rhizodeposition to a depth of 50 cm over 10 years in the different
experimental treatments. We then determined the retention
coefficient of the newly formed material for each experimental
treatment, investigating the influence of rhizodeposition on SOC
accumulation and decay at different depths.

2. Material and methods
2.1. Study site and treatments
The long-term field experiment is located in Ultuna, close to
Uppsala (59.82  N, 17.65  E), in a Dfb climate (warm summer
hemiboreal) according to the Köppen classification (Peel et al.,
2007), with mean annual precipitation of 570 mm and mean
annual air temperature of +5.4  C. The topsoil (0–20 cm) is a clay
loam with 36.5% clay, 41% silt (0.002–0.06 mm) and 22.5% sand
(0.06–2 mm) and is classified as a Eutric Cambisol (IUSS Working
Group, 2007). The parent material consists of post-glacial sediments and illite is the main clay mineral (Gerzabek et al., 1997). In
1956, the topsoil had an organic C content of 1.5%, an N content of
0.17% and a pH of 6.6. The site has been in agricultural use for at
least 300 years. Since the start of the experiment, the plots have
been cultivated manually.
From 1956 to 1999, annual C3 crops such as oats, spring barley,
sugar beet, oilseed rape, turnip rape and white mustard were
cultivated. Prior to 2000, cultivated plants had an average d13C
signature of 28.0  0.1m (Menichetti et al., 2013). In 2000, C3
crops were replaced with forage maize, a plant with a C4
photosynthetic cycle and an average d13C signature of 12.3  0.1

m (Menichetti et al., 2013).
The experimental design consisted of 15 treatments with four
replicate plots in a randomized block design. Each plot is 2 m  2 m,
separated by 40 cm high steel frames extending to a depth of
30 cm. Approximately the same amount of C (4 Mg ha 1) is added
in 10 of the treatments in autumn every second year as different
organic amendments (Table 1). Inorganic N fertilizer is added
annually during spring at a rate of 80 kg N ha 1 yr 1 in the
N-fertilized treatments. The experiment also includes a treatment
that receives neither N fertilizer nor organic amendments and a
bare fallow treatment that is kept free from vegetation by regular
weeding. All plots are fertilized annually with 22 kg P and 35–38 kg
K ha 1. Above-ground biomass is harvested by cutting the crop
close to the soil surface, and both grain and above-ground yields
are recorded each year. A sample archive, managed together with
the experiment, stores samples from topsoil, plant materials and
amendments taken every second year since 1983 and intermittently between 1956 and 1982. From the 15 treatments, the subset
of 10 selected for this study covered the whole range of SOC quality
in the experiment (Table 1).
2.2. Sampling and analysis

Soil sampling was carried out after crop harvest in September
2009. Samples were taken with an auger at increasing depth
intervals of: 0–15, 15–17.5, 17.5–20, 20–22.5, 22.5–25, 25–27.5,
27.5–30, 30–35, 35–40 and 40–50 cm. These sampling depths have
been chosen to increase the resolution in the intervals with more

Table 1
Topsoil (0–20 cm) characteristics (means with standard errors; n = 4) of treatments A–O in the Ultuna long-term experiment studied here.
Treatment
A
B
C
F
G
H
J
M
N
O
a

Crop Fertilisera
(80 kg N ha
No
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes

No
No
Yes
No
Yes
No

No
Yes
Yes
No

Fertilised with Ca(NO3)2.

1

yr

1

)

Amendment
(4 Mg C ha 1)
No
No
No
Straw
Straw
Green manure
Farmyard manure
Peat
Sawdust
Sewage sludge

d13C 1999
(m)
26.1  0.0
26.7  0.1
26.5  0.0
26.9  0.2
27.0  0.1
26.9  0.1
27.1  0.2
26.3  0.1
26.1  0.1
26.1  0.0

d13C 2009
(m)
25.2  0.4
25.6  0.1
25.5  0.1
25.9  0.1
26.3  0.2
25.8  0.2
26.3  0.2
25.8  0.1
25.4  0.0
25.4  0.1

C 2009(%)

N 2009(%)

0.97  0.02 0.11  0.00
1.12  0.02 0.12  0.01
1.37  0.03 0.13  0.01
1.49  0.01 0.15  0.00
1.92  0.04 0.17  0.02
1.54  0.02 0.18  0.00
2.21  0.05 0.20  0.01
3.78  0.07 0.23  0.01
2.2  0.05 0.16  0.01
2.66  0.02 0.25  0.02

C/N
2009

d15N 2009 Maize yields
(Mg ha 1)
(m)

Bulk density
2009

8.2
8.3
10.0
10.0
10.6
8.9
9.5
15.7
11.9
9.6

8.1  0.3
8.0  0.2
7.6  0.3
7.7  0.1
6.8  0.6
7.4  0.2
8.3  0.4
5.0  0.3
5.5  1.2
7.6  0.1

1.43  0.06
1.40  0.02
1.28  0.04
1.38  0.07
1.21  0.01
1.34  0.02
1.24  0.05
1.05  0.04
1.23  0.08
1.02  0.06

NA
3.3  0.2
7.1  0.4
4.1  0.2
8.3  0.5
6.3  0.2
8.2  0.6
10.1  0.6
7.7  0.2
9.7  0.1

L. Menichetti et al. / Agriculture, Ecosystems and Environment 200 (2015) 79–87

expected variation. Since the soil is ploughed to 20 cm, we
expected that variation would be the highest in the transition zone
between top- and subsoil, between 15 and 30 cm depth. Due to
limitations in the amount of subsoil that can be taken from the
small plots in this experiment, only one core (diameter 2 cm) per
plot was taken below 20 cm depth. This was extracted within
40 cm from two borders of each plot. The topsoil (0–20 cm) was
sampled according to standard procedures at five locations within
each plot. The samples were brought to a cool room within 2 h of
sampling and stored at 5  C for a period that varied between two
weeks and one month until further treatment. From each depth
interval, a sample of 10–15 g was taken by aggregating five
subsamples (2–3 g each) taken at approximately 1-cm intervals.
The bulk samples were then dried at 105  C for 12 h, homogenized
and milled to a powder with an agate mortar and pestle, and
subsequently analyzed.
Topsoil C and N were analyzed using an elemental analyzer
(Leco CN-2000, St. Joseph, Michigan) and d13C signature was
determined with an elemental analyzer (model EuroEA3024;
Eurovector, Milan, Italy) coupled online to a continuous flow
Isoprime isotope-ratio mass spectrometer (GV Instruments;
Manchester, UK) at Örebro Isotope Laboratory. Soil samples from
1956 and 1999 and samples of the amendments used from 1975 to
2009 were taken from the historical archive and also analyzed for
total C and d13C. The resulting d13C values were expressed in parts
per thousand (m) relative to the international standard of Vienna
Pee Dee Belemnite (V-PDB), where d13C = 1000  (Rsample Rstan13
C–12C. The d13C value of the
dard)/Rstandardm and R is the ratio of
amendments was measured by averaging direct measurements of
samples from 1975, 1979, 1989, 1993, 1995 and 2005. The resulting
d13C values for farmyard manure, green manure and sewage sludge
were 27.9  1.2, 27.4  0.9 and 25.7  0.1m, respectively. The
d13C value for peat ( 25.6  0.4m) was taken from Gerzabek et al.
(1997).
The physical conditions in the soil profile were characterized by
measuring penetration resistance in September 2009 down to
45 cm depth with an Eijkelkamp penetrologger (model P1.52,
Eijkelkamp, Arnhem, Netherlands). Each penetration profile
represented the average of four plots, with 7 repeated measurements within each plot for a total of 28 measurements for each
penetration profile.
Soil cores from a previous sampling performed in 1997 in the
unamended, fertilized treatment by Bergkvist et al. (2003) were
analyzed for 14C content. The analyses were performed after
bulking the four replicates into two separate samples. The results
of 14C analyses were expressed as per cent of modern carbon
(pMC), where pMC = Asn/Aon (1/8267  (Y 1950))% and Asn represents the particle count per minutes of the sample, Aon
represents the particle count per minutes of the international
standard and Y refers to the year of sampling. The value 8267 refers
to the average life of 14C relative to this particular convention
(Stuiver and Polach, 1977). Conventional radiocarbon age was
calculated according to the formula CRA = 8033 ln(Asn/Aon) and
expressed in years before present (BP). The value 8033 refers to the
average life of 14C relative to this particular convention (Stuiver and
Polach, 1977).
2.3. Data treatment and analysis
Data analysis was performed within the R statistical environment (R Development Core Team, 2011). Based on the penetrometer data, three depth zones were identified: (a) the topsoil (0–
20 cm), where tillage and fertilization takes place, (b) the interface
between topsoil and subsoil (20–25 cm) just below ploughing
depth and (c) the upper subsoil (25–50 cm).

81

Treatment effects were tested with ANOVA independently for
each depth interval and also aggregated (by weighted average) by
depth zones, followed then by a Tukey’s Honestly Significant
Difference (HSD) test. Correlations were expressed as Pearson’s
correlation coefficient.
The topsoil in the straw treatment (F) has been roughly in C
equilibrium since the start of the experiment, having gained only a
0.03% C or 0.26 g C cm 3 over the last 56 years. We therefore
considered it as a reference and assumed the difference between
the SOC profile in the straw treatment and that in the other
treatments to represent the C gains or losses due to the different
treatments.
All the past profile data, taken sometimes during the whole
history of the experiment at different depth intervals for different
analyses, were reconciled by linear interpolation with a vertical
resolution of 1 cm in order to have compatible measurements for
subsequent calculations. The bulk density profiles for 2009 were
estimated by linear interpolation between the topsoil bulk density
in 2009 (Kätterer et al., 2011) and the bulk density values down to
100 cm depth taken in 1956. Changes in bulk density in the subsoil
since the start of the experiment were assumed to be negligible.
The calculations of maize contributions to SOC were performed
according to the concept of equivalent soil depth (Kätterer et al.,
2011), in order to account for the relative change in soil density and
the consequent change in the tilled part of the soil. This change was
assumed to influence the preferential rooting zone so application
of the equivalent soil depth, as calculated for the same experiment
by Kätterer et al. (2011), was therefore particularly relevant in this
study. From the relative contribution of maize to SOC, we then
estimated the amount of material added by maize cultivation over
10 years and its retention coefficient, which is defined as the
amount of C from maize-derived inputs that is retained as SOC.
Since all above-ground crop residues are removed from the field in
our experiment, the retention coefficient refers to carbon derived
from below-ground biomass which was estimated from aboveground yields using linear allometric functions (Bolinder et al.,
2007).
The contribution from maize to total SOC was estimated
according to Bayesian principles. We utilized a Monte Carlo
Markov Chain (MCMC) search algorithm running the model in R
through the JAGS sampler (Plummer, 2003), which utilizes a
formal log-likelihood cost function. We chose for the target value a
prior distribution generated as uniform distribution in a range
between 0 and 0.3 for the maize proportion, and for the measured
values a normal distribution centred on the mean of the value and
with standard deviation corresponding to the measured error of
d13C signatures. The model was calibrated based on four chains of
350,000 runs each. The application of Bayesian principles allowed
us to calibrate the desired value propagating the known
uncertainty involved in the calculation, and to estimate a
probability distribution for each value. This technique becomes
particularly useful when working with small isotope intervals for
which errors need to be quantified precisely to determine
significant differences, since it gives much more detailed
information on the probability distribution of the results and
permits more rigorous error propagation compared with deterministic techniques (Parnell et al., 2010).
A principal component analysis (PCA) has been run (Venables
and Ripley, 2002) on our dataset in combination with data from
Börjesson et al. (2011), in order to assess the possible correlations
between SOC mineralization and microbial ecology.
2.3.1. Estimation of maize root contribution to SOC
The contributions of C3 and C4 material to total SOC were
calculated according to the following mass-balance equation:

82

L. Menichetti et al. / Agriculture, Ecosystems and Environment 200 (2015) 79–87

Table 2
Soil carbon stocks, contributions from maize and maize retention coefficients in the topsoil (considering equivalent soil depths according to Kätterer et al., 2011) with
standard errors of the mean (n = 4). For treatment codes A–O see Table 1.
C stocks, Mg ha
23.58  4.59
27.2  8.27
32.47  6.47
38.68  5.34
48.55  12.82
41.56  8.34
56.12  8.38
86.98  17.26
45.35  10.29
74.72  11.7

A
B
C
F
G
H
J
M
N
O

pC4 ¼

1

d13 C2009
d13 Cmaize

C4(%)

Maize stocks, Mg ha

NA
0.07  0.01
0.08  0.02
0.07  0.03
0.05  0.02
0.07  0.03
0.05  0.03
0.04  0.02
0.05  0.02
0.05  0.02

NA
1.88  0.57
2.45  0.49
2.61  0.36
2.41  0.64
2.94  0.59
2.91  0.44
3.66  0.73
2.25  0.51
3.84  0.60

d13 C1999
d13 C1999

(1)

where pC4 represents the proportion from C4 to C, d13C2009 the

measured d13C signature in 2009, d13C1999 the signature of SOC in
1999 and d13Cmaize the average isotopic signature of maize. The
d13C signature change due to addition of amendments was
considered by calculating the annual d13C change in SOC due to
amendments for each treatment from 1956 to 1999, and
extrapolating the linear trend to the period 2000–2009 to estimate
the theoretical d13C signature of SOC in 2009 that would have
resulted if only C3 plants had been grown. This value was then used
to calculate the proportion due to maize in order to account for the
effect of amendments.
3. Results
3.1. SOC, N and

14

C in topsoil and subsoil

**The highest C stocks (Table 3) were found in the peat (M) and
sludge (O) treatments, which contained 87.0 and 74.7 Mg C ha 1,
respectively. This is more than three times the total C of the bare
fallow (A), which amounted to 23.6 Mg C ha 1 and about twice that
of the straw (F) treatment, which amounted to 38.7 Mg C ha 1.
Nitrogen fertilization positively influenced SOC accumulation in
the topsoil, an effect which was particularly clear when comparing
the treatments that differed only in N fertilization. The N+ straw
treatment (G) had more SOC than the straw (F) treatment and the
+N treatment (C) had more SOC than the control (B). But the
accumulation of C4-derived SOC did not show any statistically
significant difference between the +N and the +straw treatment
(Table 2).
Significant increases between 25 and 35 cm were generally
recorded in the richer amended treatments (peat and sludge),
which displayed an increase relative to the straw treatment of on
average 0.27 g C cm 3 (and ranged between 0.12 and 0.31 g C cm 3),
while the C content of the non-amended treatments (A, B and C)

1

Maize retention coefficient
NA
0.41
0.25
0.46
0.21
0.34
0.26
0.26
0.21
0.29

was not significantly different from the straw treatment at this
depth. Treatment differences in C content were significant down to
35 cm depth when tested by comparing values aggregated by layer
with an ANOVA (Table A1). Considering C variation in the subsoil
(depth below 25 cm) as a linear function of depth, treatments M
(peat) and O (sewage sludge) showed significantly (C.I. 95%) higher
C contents below this depth than in the straw treatment.
Interestingly, when tested with ANOVA the penetrometer values
showed significant treatment differences only above 13 cm and
below 41 cm (Fig. 2).
To calculate the SOC changes induced by the treatments over
the whole soil profile, we considered the straw treatment, which is
roughly in C equilibrium, as zero (Table 3). Relative to the straw
treatment, all the amendment treatments gained SOC (between
0.26 and 2.27 g C cm 3 in the 0–15 cm range), while the nonamended treatments lost SOC (between 0.42 and 0.85 g C cm 3 in
the 0–15 cm range).
The N content of the soil profiles followed the SOC content. The
d15N signature in the topsoil varied according to the different
fertilization regimes, but with no apparent relationship with total
SOC and probably depending mainly on the signature of the
fertilizer and amendments added. Treatment differences were
significant in the topsoil but no significant differences were
detected in the subsoil.
The analysis of 14C content within the profile of the unamended
fertilized plots (treatment C) in 1997 revealed a steep decline in
modern carbon throughout the profile, which followed a nonlinear
decrease from around 100% in the topsoil to around 40% at 70 cm
depth (Fig. 3). The average SOC age between 40 and 60 cm depth
was 1759  296 years BP, while below 60 cm depth it was
7747  100 years BP.
3.2. Contribution from maize to SOC
Significant effects of treatments on soil organic d13C signature
were found to 22.5 cm depth according to analysis of variance
(Fig. 4). When comparing the d13C signature of the topsoil (after
aggregating the data by weighted average for the three depth

Table 3
Differences (g cm3) in soil organic carbon stocks relative to treatment F. For treatment codes A–O see Table 1.
Treatment

Soil depth (cm)
0–15

A
B
C
G
H
J
M
B
O

0.85  0.20
0.39  0.13
0.42  0.08
0.26  0.05
0.28  0.03
0.79  0.10
2.27  0.10
0.50  0.09
0.97  0.15

15–17.5
0.50  0.07
0.32  0.10
0.12  0.01
0.2  0.04
0.18  0.02
0.58  0.13
1.74  0.33
0.00  0.00
0.77  0.16

17.5–20
0.36  0.06
0.27  0.09
0.16  0.01
0.17  0.03
0.21  0.01
0.52  0.16
1.16  0.41
0.06  0.01
0.58  0.08

20–22.5
0.15  0.02
0.10  0.03
0.17  0.02
0.10  0.02
0.07  0.01
0.42  0.12
0.47  0.09
0.02  0.00
0.40  0.06

22.5–25
0.15  0.01
0.09  0.02
0.11  0.01
0.05  0.01
0.05  0.01
0.19  0.05
0.12  0.01
0.01  0.00
0.31  0.06

25–27.5
0.02  0.00.
0.02  0.00
0.00  0.00
0.13  0.03
0.12  0.02
0.16  0.03
0.17  0.01
0.03  0.00
0.31  0.05

27.5–30
0.05  0.01
0.02  0.00
0.05  0.01
0.32  0.09
0.05  0.00
0.22  0.05
0.28  0.02
0.12  0.02
0.21  0.02

30–35
0.00  0.00
0.04  0.01
0.10  0.02
0.15  0.04
0.14  0.03
0.26  0.07
0.26  0.03
0.05  0.01
0.13  0.02

35–40
0.03  0.01
0.05  0.01
0.21  0.06
0.01  0.01
0.30  0.10
0.15  0.03
0.32  0.05
0.02  0.01
0.10  0.02

40–50
0.05  0.02
0.18  0.03
0.22  0.08
0.11  0.06
0.32  0.12
0.08  0.02
0.39  0.04
0.06  0.02
0.12  0.03

L. Menichetti et al. / Agriculture, Ecosystems and Environment 200 (2015) 79–87

Fig. 1. Soil C content with depth in six of the 10 selected treatments. Shaded areas
represent standard error of the mean (n = 4). Values above the horizontal dashed
line show significant treatment differences (p < 0.1). For treatment codes A–O see
Table 1.

83

Fig. 4. d13C signatures along the soil profile. Shaded areas represent standard error
of the mean (n = 4). Values above the horizontal dashed lines show significant
treatment differences (p < 0.1). For treatment codes A–O see Table 1.

Fig. 2. Penetration force along the soil profile. For treatment codes A–O see Table 1.

Fig. 5. Relationship between maize-derived SOC and above-ground yield in each
treatment. For treatment codes A–O see Table 1.

Fig. 3. Percentage of modern carbon (pMC) along the soil profile in treatment C (no
organic amendment, fertilised with calcium nitrate) in the Ultuna long-term
continuous soil organic matter experiment.

zones), a Tukey’s HSD test found significant differences only
between amended and non-amended treatments (Table A2).
The contribution of 10 years of maize cultivation to the d13C
signature of SOC in the topsoil, calculated according to Eq. (1),
varied between 4 and 8% of total SOC. The non-amended
treatments displayed less uncertain values, probably due to the
fact that there were no amendments contributing to the C stock. In
treatments with high C accumulation (M, N, O and J, see Table 1),
the relative contribution of maize was between 4 and 5% of total
SOC, although the absolute C stocks of maize origin were in general
higher than in treatments with lower C accumulation ranging from
2.41 Mg ha 1 in the N+ straw treatment to 3.88 Mg ha 1 in the
sludge treatment (Table 2). A similar interaction of amendments

84

L. Menichetti et al. / Agriculture, Ecosystems and Environment 200 (2015) 79–87

Fig. 6. Probability density of the contribution of maize to SOC in the topsoil in the
different treatments during 10 years, 1999–2009. Each probability density is
calculated from four Monte Carlo Markov chains of 250,000 runs and expresses the
probability distribution of the calibrated value. For treatment codes see Table 1.

with C4–C has been already noticed in the gas fluxes of the same
experiment (Menichetti et al., 2013). Particularly interesting is the
comparison between the straw and N+ straw treatments, where
the maize contribution to SOC was 2.61  0.36 and 2.41  0.64 Mg
ha 1, respectively. This contrasts to the almost doubled yield in the
N+ straw compared to straw treatment.
There was a statistically significant positive relationship
(R2 = 0.54; p < 0.02) between above-ground biomass yield and
the amount of maize-derived SOC (Fig. 5), indicating the
importance of roots from the last 10 years to the formation of
soil C stocks. Most treatments showed a ratio between maizederived soil C stocks and maize above-ground dry matter yield of
between 0.29 and 0.45, while the straw (F) treatment showed a
ratio of 0.62. The average retention coefficient, which is based on
the ratio between maize C stocks and root inputs (with the latter
linearly derived from yield), was 0.30  0.09 (where the error
represents the standard deviation for all treatments taken
together, and has been calculated from the probability distributions showed in Fig. 6). The highest retention coefficient for the
maize roots was found in the straw (F) and control (B) treatments
and the lowest in the N+ straw (G) and sawdust (N) treatments.
4. Discussion
The 53 years of various treatments have caused significant
changes in the SOC content down to 35 cm depth but not below this
level (Fig. 1). A corresponding influence of fertilization treatments
on the upper (in this case 30–40 cm) subsoil was reported recently
by Kirchmann et al. (2013) in two other long-term fertilisation
experiments in Sweden over a similar time frame. In that case the
fertilization treatment affected in a more evident way other soil
properties and produced two markedly different soil profile
descriptions.
The analysis of 14C content over the profile of the nitrogenfertilised treatment revealed a distribution of modern carbon
(Fig. 3) compatible with the root distribution function used by
Kätterer et al. (2011). This particular treatment only shows root
inputs since the start of the experiment, and therefore roots have a
great influence on SOC formation. According to the distribution of
14
C, the influence of C inputs over the soil profile decreases as an
exponential-like function of depth. Similarly shaped 14C profiles
were found for agricultural soils cropped with maize and wheat at
Rotthalmünster and Halle in Germany (Rethemeyer et al., 2005).

The soil at the Ultuna site is formed from post-glacial marine
sediments and is presently situated at 14 m above sea level. Based
on estimated shore displacements at Gamla Uppsala (Eriksson,
1999), a nearby site, the Ultuna site emerged from the sea less than
2500 years ago. Consequently, the average ages of the deep soil
layers suggests that the soil C in these is still dominated by preterrestrial C from marine sediments.
The penetrometer revealed some significant differences between treatments in the compaction of deeper horizons (below
41 cm depth), suggesting influences of the treatments on structure
in the subsoil. These variations could be due to direct or indirect
crop activity, for example to drier conditions induced by higher
evapotranspiration. The organic material eluviated from the upper
layers might for example also differ in nature between the
treatments, thus, producing different effects on the aggregates in
the deeper layers. Several other explanations might be equally
possible and a more detailed investigation would be needed for
obtaining a precise answer.
We calculated the SOC balance relative to the straw treatment
(Table 3) as the topsoil in this treatment is roughly in C equilibrium
(inputs = outputs) and is therefore assumed to be in steady state.
The resulting SOC balance profile suggested again a major
influence of the treatments on the topsoil.
The effect of maize roots and rhizodeposition was not yet
detectable below 22.5 cm depth after 10 years (Fig. 4). Considering
instead the time elapsed since the start of the experiment, the
richer treatments (M and O) were the only treatments to show a
significant SOC increase also in the subsoil (Fig. 1) which can have
been due to translocation induced by soil fauna, dissolved organic
carbon or accumulation of rhizodeposits. Although differences in
SOC between quite extreme treatments, such as bare fallow and
compost-amended soil, were shown to be significant to 40 cm
depth already after 13 years in a Swedish experiment (Kätterer
et al., 2014), the accumulation of SOC in the subsoil, at least in
agricultural soils in Nordic environments, can be considered to be
relatively slow relative to the rate of changes in the topsoil.
When calculating the contribution of maize to SOC, the isotopic
effect of atmospheric CO2 enrichment (Francey et al., 1999) is
contained, as error, in the average maize d13C signature over the
period of maize cultivation. The isotopic enrichment due to
fractionation mentioned by Balesdent and Mariotti (1996) can be
estimated from the bare fallow treatment as 0.005  0.001m yr 1 if
such an effect is approximated with a linear relationship between
d13C signature and time (Schweizer et al., 1999). In such a relatively
short period as 10 years, a linear approximation should not be
distinguishable from a more detailed exponential approximation
such as the Rayleigh equation (Balesdent and Mariotti, 1996;
Menichetti et al., 2014). Over 10 years it can therefore be regarded
as quite small, given also the difference between the two end
members (C3 and C4 signatures). In any case, this enrichment is
contained in the correction applied to each single treatment, which
considers the changes in d13C signature recorded in each treatment
from 1956 to 1999. This correction also takes care of the effect of
the different amendments on the d13C signature.
The average retention coefficient calculated for maize material
over all treatments (0.30  0.09) was compatible with that
previously calculated by Kätterer et al. (2011). Those authors
utilized SOC mass balance and found an average retention
coefficient for root-derived C during 53 years of 0.27  0.09.
Although these two coefficients were calculated on two different
periods using different approaches, they are very similar.
In general, the retention coefficient of root material seems
higher than previously calculated (e.g., Plénet et al., 1993 and
Bolinder et al., 1999). Our results are in line with the hypotheses
proposed by Rasse et al. (2005), who suggested that besides

L. Menichetti et al. / Agriculture, Ecosystems and Environment 200 (2015) 79–87

chemical protection, root-derived C is also more protected than
other forms of C inputs due to physicochemical and physical
interactions. In particular, the C contained in root hairs, associated
mycorrhizae and fine roots is considered to enter the soil directly at
the scale of physically protected C inside aggregates or fine pores
(Rasse et al., 2005; Mendez-Millan et al., 2010). Root growth also
actively structures the soil and it is able to promote soil
aggregation and coating formation around the roots, which could
contribute to subsequent SOC stabilization. The retention coefficients of roots were generally also higher than those of nonhumified amendments such as straw, sawdust or green manure, as
found previously by Kätterer et al. (2011), suggesting that roots
contribute more than these amendments to SOC formation. This
finding is in line with the above-mentioned hypotheses, as C
originating from amendments would lack the physical protection
that root C has.
Calculating a specific retention coefficient for each treatment
allowed us to discuss eventual differences between the different
fertilisation regimes and to explore more in details the limitations
of a uniform allometric function for every treatment. Some
treatments (unfertilized, straw and green manure) which produced relatively low yields (Table 1) also gave a slightly higher
retention coefficient for maize roots than treatments with higher
yields (Table 2). This observation could indicate some kind of stress
factor, possibly related to nutrient deficiencies within root tissues
or even in the soil that might have influenced the decomposability
of the root material during the first decade (Freschet et al., 2012). In
a more long-term perspective, however, nitrogen limitation may
rather lead to lower microbial substrate use efficiency and lower C
retention (Poeplau et al., 2015). This is supported by the effect of N
fertilization on straw C retention observed in the Ultuna
experiment. During the period 1956–1999 with only C3-crops,
yields were in average 91% higher in N-fertilized (treatment C and
G) compared to unfertilized treatments irrespective of whether
straw was added or not (B and F). Thus, it is reasonable to assume
that the root-derived C contribution to SOC was very similar in the
treatment pairs C/B and G/F. Nevertheless, total C stocks in G were
about 10 Mg higher than in F whereas those in C were only 5 Mg
higher than in B (Table 2). Consequently, the retention of straw C
was higher in G than in F. In absolute amounts, straw addition
resulted in 16 Mg more SOC in G compared with C but only in 11 Mg
more in F than in B.
Factors generating a non-linear response of yield to C
accumulation caused by nutrient deficiencies seemed present,
but only in extreme cases. The logarithmic relationship between
maize-derived C and yield resulted only in a marginally higher
coefficient of determination (R2 = 0.54) than a linear relationship
(R2 = 0.52) (Fig. 5). This supports the validity of the linear
relationship between roots and yields proposed by Bolinder
et al. (2007) in case of low yields, although such coefficients of
determinations might be considered perfectly acceptable or really
low depending on the purpose of the study. Assuming an uniform
root:shoot value seems a viable approach in case of lack of data, but
one must bear in mind the limitations associated with such an
approach.
We also observed correlations between retention coefficient
and certain phospholipid fatty acid (PLFA) classes previously

85

measured in the same site by Börjesson et al. (2011). Two PLFA
classes indicating Gram-negative bacteria (16:1v7 and 16:1v9),
which are considered to be less efficient decomposers than Grampositive bacteria (Zogg and Zak, 1997), were found to be negatively
correlated with the retention coefficient (r = 0.2 and r = 0.4,
respectively). Class 18:2, which is commonly associated with
fungal cells (Bååth and Anderson, 2003), was instead positively
correlated with the retention coefficient (r = 0.6). A multivariate
analysis (PCA) pointed out similar correlations. It is difficult to
draw precise conclusions from these relationships, since the two
datasets are coming from two different years although from the
same experimental field. But also these results suggest a direct
relationship of microbial ecology and SOC mineralization and its
interaction with agricultural practices.
5. Conclusions
Application of organic amendments and N fertilisation in the
Ultuna long-term experimental plots substantially affected SOC in
the topsoil, giving rise to a 4 times range in carbon stock. Below
this depth a significant accumulation of C down to 35 cm depth
was only found in recalcitrant or processed amendments such as
peat and sewage sludge, the two treatments which also had the
largest accumulation in the topsoil. In contrast, the physical
conditions in deeper soil layers were affected in several of the
treatments. Maize roots significantly contributed to SOC in the
topsoil after 10 years of cultivation, and contributed approximately 4–8% of the newly formed SOC through rhizodeposition.
However, SOC accumulation in most of the subsoil was not
detectable after 10 years. The SOC below 40 cm depth showed an
average age >1500 years, suggesting again that SOC in deeper soil
layers is influenced quite slowly by changes in the topsoil.
However, since organic matter stored in the subsoil may
decompose slower than in the topsoil, C accumulation in the
subsoil induced for example by deep-rooting varieties of crops
remains a promising strategy for C sequestration.
The contribution of roots to SOC was similar to that reported
recently using another methodology. Roots contributed relatively
more to SOC than the same mass of carbon derived from aboveground plant material. These results suggest that root contributions to SOC are often underestimated and that root-derived C
should be considered one of the most effective inputs for C
sequestration in soil.
Acknowledgements
We are grateful to former colleagues, especially Olle Gunnarsson, Hans Nömmik and Jan Persson for starting and keeping this
experiment running despite several threats of closure. Pär Hillström has managed the experiment over the last 25 years. Financial
support for keeping the long-term experiment and for the present
study was provided by the Faculty of Natural Resources and
Agricultural Sciences at SLU.
Appendix A
See Tables A1 and A2

86
Table A1
Carbon concentrations (%) in the experimental treatments at different depths with standard error (n = 4). Superscript letters represent the groups identified by a Tukey’s HSD test (a = 0.1). NA = not available. For treatment codes A–O
see Table 1.
1–15

15–17.5

17.5–20

20–22.5

22.5–25

25–27.5

27.5–30

30–35

35–40

40–50

50–60

0.89  0.11e
1.02  0.11de
1.30  0.15cde
1.63  0.04cd
1.97  0.17c
1.75  0.12cd
2.31  NAbc
3.79  0.15a
2.13  0.18c
3.01  0.14b

0.88  0.10e
1.06  0.12de
1.31  0.12cde
1.6  0.03bcde
1.89  0.24bcd
1.79  0.07bcd
2.02  0.18bcd
4.06  0.15a
2.19  0.04bc
2.63  0.44b

0.90  0.05d
0.91  0.13d
1.28  0.07cd
1.33  0.12cd
1.57  0.18bcd
1.43  0.09cd
1.87  0.18bc
3.30  0.35a
1.42  0.05cd
2.27  0.25b

0.91  0.08c
0.98  0.17c
1.14  0.03bc
1.22  0.10bc
1.49  0.14abc
1.42  0.03abc
1.78  0.38ab
1.66  0.41abc
1.21  0.12bc
1.91  0.12a

0.93  0.04b
0.97  0.18ab
0.93  0.05b
1.03  0.13ab
0.97  0.09ab
1.06  0.07ab
1.44  0.21ab
1.53  0.19a
1.13  0.09ab
1.52  0.18ab

0.78  0.01b
0.84  0.10b
0.87  0.05b
0.94  0.05ab
0.95  0.08ab
0.97  0.05ab
1.10  0.17ab
1.09  0.05ab
0.99  0.08ab
1.32  0.14a

0.78  0.03b
0.8  0.11ab
0.83  0.09ab
0.79  0.03ab
0.92  0.12ab
0.89  0.07ab
0.94  0.08ab
0.99  0.08ab
0.76  0.12b
1.16  0.11a

0.71  0.07c
0.78  0.04bc
0.76  0.05c
0.78  0.09bc
1.21  0.19a
0.84  0.04abc
0.92  0.09abc
1.14  0.08ab
1.00  0.09abc
1.04  0.07abc

0.73  0.10a
0.74  0.11a
0.64  0.08a
0.69  0.09a
0.78  0.14a
0.51  0.14a
0.94  0.16a
0.99  0.12a
0.65  0.13a
0.74  0.01a

0.62  0.08a
0.70  0.09a
0.52  0.07a
0.59  0.04a
0.69  0.16a
0.43  0.07a
0.75  0.05a
0.98  NAa
0.56  0.08a
0.79  0.07a

0.51  0.11a
NA
0.37  0.02a
0.48  0.05a
0.79  0.45a
0.34  0.07a
0.4  NAa
0.91  0.07a
0.5  0.11a
0.7  0.08a

Table A2
d13C signature (m) in the experimental treatments at different depths with standard error (n = 4). Superscript letters represent the groups identified by a Tukey’s HSD test (a = 0.1). NA = not available. For treatment codes A–O see
Table 1.
Treatment
A
B
C
F
G
H
J
M
N
O

0–15
25.45  0.31ab
25.72  0.09ab
25.27  0.23a
25.84  0.30ab
26.21  0.11b
25.79  0.24ab
26.65  NAb
25.8  0.12ab
25.42  0.07ab
25.47  0.07ab

15–17.5
25.25  0.53a
25.51  0.21a
25.54  0.10a
25.93  0.23a
26.21  0.35a
25.84  0.13a
26.04  0.41a
25.76  0.13a
25.29  0.03a
25.61  0.25a

17.5–20
25.14  0.30a
25.61  0.10ab
25.78  0.31ab
25.97  0.19ab
26.36  0.20b
25.76  0.26ab
26.29  0.21b
25.76  0.06ab
25.6  0.02ab
25.31  0.14a

20–22.5
25.25  0.24a
25.93  0.15ab
25.16  0.26a
25.73  0.22ab
26.06  0.27ab
25.57  0.09ab
26.33  0.28b
26.04  0.15ab
25.59  0.27ab
25.42  0.18ab

22.5–25
25.25  0.19a
25.9  0.12a
25.27  0.14a
25.24  0.81a
26.1  0.15a
26.06  0.11a
25.99  0.30a
25.9  0.08a
25.54  0.09a
25.96  0.33a

25–27.5
25.00  0.29a
25.83  0.14a
25.31  0.23a
25.39  0.34a
25.6  0.30a
25.8  0.10a
26.08  0.17a
25.61  0.32a
25.21  0.26a
25.36  0.25a

27.5–30
25.65  0.11a
25.22  0.44a
25.53  0.10a
25.04  0.49a
25.65  0.31a
25.61  0.17a
25.85  0.26a
25.95  0.19a
24.88  0.23a
25.52  0.39a

30–35
25.74  0.13a
25.33  0.15a
25.04  0.35a
24.89  0.5a
25.83  0.07a
25.84  0.12a
25.94  0.12a
25.56  0.09a
25.08  0.38a
24.97  1.11a

35–40
25.57  0.17a
25.38  0.17a
25.24  0.29a
24.61  0.88a
25.57  0.23a
25.32  0.13a
25.58  0.14a
25.37  0.27a
25.11  0.23a
25.24  0.02a

40–50
25.67  0.14a
25.53  0.42a
25.25  0.3 +a
24.99  0.02a
25.26  0.2 + a
25.46  0.26a
25.48  0.04a
25.36  NAa
24.59  0.15a
25.23  0.24a

50–60
25.75  0.20a
NA
26.43  0.33a
25.43  0.21a
25.22  0.1 +a
25.74  0.18a
25.78  NAa
25.78  0.07a
24.86  0.14a
25.09  0.47a

L. Menichetti et al. / Agriculture, Ecosystems and Environment 200 (2015) 79–87

Treatment
A
B
C
F
G
H
J
M
N
O

L. Menichetti et al. / Agriculture, Ecosystems and Environment 200 (2015) 79–87

References
Balesdent, J., Mariotti, A., 1996. Measurement of soil organic matter turnover using
13
C natural abundance. In: Yamasaki, S., Boutton, T. (Eds.), Mass Spectrometry of
Soil. M. Dekker, New York, pp. 47–82.
Bååth, E., Anderson, T., 2003. Comparison of soil fungal/bacterial ratios in a pH
gradient using physiological and PLFA-based techniques. Soil Biol. Biochem. 35,
955–963.
Barré, P., Eglin, T., Christensen, B., Ciais, P., Houot, S., Kätterer, T., Van Oort, F., Peylin,
P., Poulton, P., Romanenkov, V., Chenu, C., 2010. Quantifying and isolating stable
soil organic carbon using long-term bare fallow experiments. Biogeosciences 7,
3839–3850.
Bergkvist, P., Jarvis, N., Berggren, D., Carlgren, K., 2003. Long-term effects of sewage
sludge applications on soil properties: cadmium availability and distribution in
arable soil. Agric. Ecosyst. Environ. 97, 167–179.
Bird, M.A., Kracht, O.B., 2003. The effect of soil texture and roots on the stable carbon
isotope. Aust. J. Soil Res. 41, 77–94.
Blagodatskaya, E., Yuyukina, T., Blagodatsky, S., Kuzyakov, Y., 2011. Turnover of soil
organic matter and of microbial biomass under C3–C4 vegetation change:
consideration of 13C fractionation and preferential substrate utilization. Soil
Biol. Biochem. 43, 159–166.
Bolinder, M., Angers, D., Giroux, M., Laverdiere, M., 1999. Estimating C inputs
retained as soil organic matter from corn (Zea mays L.). Plant Soil 85–91 1996.
Bolinder, M., Janzen, H., Gregorich, E., Angers, D., Van den Bygaart, A., 2007. An
approach for estimating net primary productivity and annual carbon inputs to
soil for common agricultural crops in Canada. Agric. Ecosyst. Environ. 118,
29–42.
Börjesson, G., Menichetti, L., Kirchmann, H., Kätterer, T., 2011. Soil microbial
community structure affected by 53 years of nitrogen fertilization and different
organic amendments. Biol. Fertil. 48, 245–257.
Chabbi, A., Kögel-Knabner, I., Rumpel, C., 2009. Stabilized carbon in subsoil horizons
is located in spatially distinct parts of the soil profile. Soil Biol. Biochem. 41,
256–261.
Eglin, T., Ciais, P., Piao, S.L., Barre, P., Bellassen, V., Cadule, P., Chenu, C., Gasser, T.,
Koven, C., Reichstein, M., Smith, P., 2010. Historical and future perspectives of
global soil carbon response to climate and land-use changes. Tellus: Ser. B 62,
700–718.
Eriksson, J.A., 1999. Land-use history in Gamla Uppsala. Laborativ Arkeologi 12,
25–34.
FAO statistical database, FAOSTAT, http://faostat.fao.org/, (accessed 25.03.2013).
Francey, R., Allison, C., Etheridge, D., Trudinger, C., Enting, I., Leuenberg, M.,
Lagenfelds, R., Michel, E., Steele, L., 1999. A 1000-year high precision record of
d13C in atmospheric CO2. Tellus: Ser. B 51, 170–193.
Freschet, G., Aerts, R., Cornelissen, J., 2012. A plant economics spectrum of litter
decomposability. Funct. Ecol. 26, 56–65.
Gerzabek, M., Pichlmayer, F., Kirchmann, H., Habernauer, G., 1997. The response of
soil organic matter to manure amendments in a long-term experiment at
Ultuna, Sweden. Eur. J. Soil Sci. 48, 273–282.
Jenkinson, D.S., Coleman, K., 2008. The turnover of organic carbon in subsoils: part
2 modelling carbon turnover. Eur. J. Soil Sci. 59, 400–413.
Jenkinson, D.S., Poulton, P.R., Bryant, C., 2008. The turnover of organic carbon in
subsoils: part 1. Natural and bomb radiocarbon in soil profiles from the
Rothamsted long-term field experiments. Eur. J. Soil Sci. 59, 391–399.
IUSS Working Group, 2007. World reference base for soil resources 2006, first
update 2007. World Soil Resources Reports No. 103. FAO, Rome.
Jobbagy, E., Jackson, R., 2000. The vertical distribution of soil organic carbon and its
relation to climate and vegetation. Ecol. Appl. 10, 423–436.
Kätterer, T., Bolinder, M., Andrén, O., Kirchmann, H., Menichetti, L., 2011. Roots
contribute more to refractory soil organic matter than aboveground crop
residues: as revealed by a long-term field experiment. Agric. Ecosyst. Environ.
141, 184–192.
Kätterer, T., Bolinder, M., Berglund, K., Kirchmann, H., 2013. Strategies for carbon
sequestration in agricultural soils in northern Europe. Acta Agric. Scand. Sect. A
1–18 May 2013.
Kätterer, T., Börjesson, G., Kirchmann, H., 2014. Changes in organic carbon in topsoil
and subsoil and microbial community composition caused by repeate