Charred organic carbon in German chernozemic soils
M . W . I . S C H M I D T a , J . O . S K J E M S T A D b , E . G E H R T c & I . K OÈ G E L - K N A B N E R a
aLehrstuhl fuÈr Bodenkunde, Technische UniversitaÈt MuÈnchen, 85350 Freising-Weihenstephan, Germany, bCSIRO, Land and Water,
Glen Osmond, SA 5064, Australia, and cNiedersaÈchsisches Landesamt fuÈr Bodenforschung, 30655 Hannover, Germany
Summary
Burning vegetation produces partly charred plant material which subsequently could contribute to the
highly refractory proportion of soil organic matter. The presence of charred organic carbon (COC) was
investigated in 17 horizons originating from nine soils from Germany and the Netherlands using a suite of
complementary methods (high-energy ultraviolet photo-oxidation, scanning electron microscopy, solid-
state 13C nuclear magnetic resonance, lignin analysis by CuO oxidation). Charred organic carbon could
not be detected in the A horizons of an Alisol and a Gleysol, but it contributed up to 45% of the organic
carbon and up to about 8 g kg±1 of the soil in a range of grey to black soils (Cambisol, Luvisol, Phaeozem,
Chernozem and Greyzem). All these soils have chernozemic soil properties (dark colour, A±C pro®le,
high base saturation, bioturbation). A 10-km colour sequence of four chernozemic soils, which were very
similar in chemical and physical properties, showed a strong relation between colour and the content of
COC. This suggests that the COC affects mainly soil colour in the sequence studied. Finely divided COC
seems to be a major constituent of many chernozemic soils in Germany. These results suggest that
besides climate, vegetation and bioturbation, ®re has played an important role in the pedogenesis of
chernozemic soils.
Introduction
Burning vegetation produces large amounts of highly refrac-
tory organic matter consisting of charcoal and partly charred
plant material on the surface and incorporated in the soil. We
call these forms of thermally altered organic carbon `charred
organic carbon' (COC). The COC can have a major impact on
composition, turnover and formation of soil organic matter. It
may increase the amount of aromatic C and contribute a
relatively inert type of carbon to the soil organic matter pool
(Skjemstad et al., 1996). From COC highly aromatic humic
acids can be extracted (Haumaier & Zech, 1995; Skjemstad
et al., 1996).
Using light microscopy several workers have found black-
ish, coal-like particles or ¯akes often with a cellular structure
and diameters between 2 and 10�m in numerous typical
Chernozems from the Central Russian highland forest steppe
on loess±loam (Kubiena, 1938; Yarilova, 1972; Pawluk,
1985), Canadian Chernozems (Pawluk, 1985) and other soil
types with dark A horizons (AltemuÈller, 1992). There is also
indirect evidence that COC produces highly aromatic humic
acids in volcanic ash soils (Hatcher et al., 1989), in
Chernozems (Kononova, 1966; Schnitzer, 1992), in a Vertisol
in Mali (Gehring et al., 1997), in an Argentinian Hapludoll
(Zech et al., 1997), and in Japanese volcanic ash soils (Golchin
et al., 1997). In a systematic study of northern Eurasian soils,
humic acid contents increased in the order Grey forest
soils < Dark grey forest soils < Chernozems (Kononova,
1966). Particles of COC were identi®ed by light microscopy
and proton-spin relaxation editing in 13C nuclear magnetic
resonance (NMR) spectroscopy (Golchin et al., 1997). Using a
combination of high-energy ultraviolet (UV) photo-oxidation,
scanning electron microscopy and solid-state 13C NMR
spectroscopy, Skjemstad et al. (1996) identi®ed COC and
determined its content in Australian soils. Australian grassland
soils which were under aboriginal ®re management, presum-
ably for thousands of years, are characterized by black A
horizons with up to 30% of the soil carbon present as COC,
whereas adjacent forested soils not subjected to regular
burning are grey and contain little COC (Skjemstad et al.,
1997).
To determine whether the black colour of some German
soils could also be attributed to COC, we studied a variety of
Chernozem-like soils. All of these soils had dark grey to black
A horizons, high base saturation, and bioturbation, and were
classi®ed as Chernozem, Greyzem, Phaeozem, Cambisol or
Luvisol (FAO, 1994). For simplicity, we use the more general
term `chernozemic soils' for these soils in this paper. As
references, two non-chernozemic light-coloured soils (Alisol,
Gleysol) were sampled.
R
Correspondence: M. W. I. Schmidt, Max-Planck-Institut fuÈr Biogeochemie,
PO Box 10 01 64, 07745 Jena, Germany.
E-mail: [email protected]
Received 7 April 1998; revised version accepted 21 December 1998
European Journal of Soil Science, June 1999, 50, 351±365
# 1999 Blackwell Science Ltd 351
To investigate the apparent coincidence between black
colour and the presence of COC we examined the chemical
structure of the soil organic matter in a colour sequence in soils
of part of the loess belt between Hildesheim and Braun-
schweig, Germany, where chernozemic soils are typical.
Recently, Gehrt (1999) reported a patchwork-like distribution
of black and grey soils south of Peine. Apart from colour, the
soil properties seemed almost identical, i.e. morphology (depth
of pro®le, horizons, bioturbation, position in the landscape,
water table, crop rotation) and physical and chemical proper-
ties (texture, bulk mineralogy, C and N contents, pH). In
previous studies, the correlation between organic carbon
content and soil colour, expressed as Munsell values, was
found to be predictable for Ap horizons (r2 > 0.9) within soil
landscapes, provided soil texture did not vary widely (Schulze
et al., 1993). However, the patchwork (diameter 5±15 km) of
black and grey chernozemic soils (Chernozems or Greyzems)
is visible in satellite images and in the ®eld. The factors
responsible for this apparent colour pattern remain unde®ned,
but they could be related to the structure of the soil organic
matter.
The content of COC in soils can be determined by the
technique of high-energy UV photo-oxidation in combination
with 13C NMR spectroscopy. Previous studies by Skjemstad
et al. (1993) showed that several materials found in soil,
including wood, lignin and humic acids, could all be destroyed
by the high-energy photo-oxidation process, provided they
were exposed to ultraviolet radiation in the presence of excess
oxygen. In subsequent studies Skjemstad et al. (1996) found
that this method could differentiate between natural soil
organic matter and COC.
In this work we pose two separate questions: (i) what is the
potential contribution of COC to chernozemic soils from
Germany, and (ii) is there a relation between soil colour and
the chemical structure of soil organic matter? To answer these
questions we investigated the contribution of COC to the
chemical composition of these soils, using a suite of
complementary methods, including elemental analysis, CuO
oxidation of lignin compounds, high-energy UV photo-
oxidation, solid-state 13C NMR spectroscopy and scanning
electron microscopy. The soils we chose originate from sites
under agriculture and forest.
Materials and methods
Soils
As references two non-chernozemic soils were sampled. One is
a Haplic Alisol (soil 1) located in the eastern hills in Siggen,
Schleswig-Holstein, under a beech and oak (Melico±Fagetum)
vegetation. The mean age of the stand is 90±100 years, and
forest on this site is documented since the 13th century
(Schimming et al., 1993). The other is a Dystric Gleysol (soil
2) at Wageningen, the Netherlands, in a ®eld that has been
under agriculture for 100 years. Both mineral fertilizers and
cow manure have been applied, with a typical crop rotation of
maize, wheat and potato, to this ®eld.
The chernozemic soils we studied are as follows. A Humic
Cambisol (soil 3) from loess overlying WuÈrmian ¯uvic gravel
was sampled 30 km south of MuÈnchen. It has been under
coniferous vegetation for at least 100 years. The A horizons
display classic chernozemic soil properties, i.e. deep black A
horizons directly overlying the carbonaceous C horizons, high
base saturation and bioturbation. This soil receives 1400 mm
of rain per year on average, which is unusually large for
chernozemic soils. A Humic-stagnic Luvisol from the Baltic
coast close to Groûenbrode and developed from Weichselian
till (Schimming et al., 1993) is our soil 4. It lies in an isolated
patch of black soils on the peninsula neighbouring the island of
Fehmarn. The occurrence of these chernozemic soils is
focused around a few islands in the southwestern Baltic Sea
(Samsoe, Poel, Fehmarn). Because of their apparent unique-
ness, the black soils of Fehmarn have attracted the attention of
soil scientists over many years (Wolff, 1930; Schlichting,
1953). The pedogenesis of this soil has been related to that of
Chernozems, although a steppe vegetation during the Holo-
cene could not be proven by pollen analysis (Schmitz, 1955).
Other explanations were introduced to explain the dark colour,
i.e. clay mineralogy (Schimming et al., 1993) and slow
degradation of organic matter because of a high water table
(Rohdenburg & Meyer, 1968). No conclusive explanation for
the exceptional dark colour of this soil, however, has been
generally accepted. Our soil 5 is a Haplic Phaeozem north of
Halle/Saale in Seeben and has been farmland since the
beginning of this century (Schmidt et al., 1996).
Our sites 6±9 represent a colour sequence changing from
black to grey (Chernozems and Greyzems) developed on loess
in the region of Hildesheim-Braunschweig, northern Germany
(Gehrt, 1999). Except for the variation in colour, soil
properties are very similar. Gehrt (1999) has suggested that
differences in colour might be related to variation in clay
mineralogy arising from parent loess and underlying Cretac-
eous sediments. According to Bailly (1972), variations in soil
colour due to differences in palaeo-vegetation can be excluded.
The soils were described, sampled and classi®ed according
to established procedures, and horizons were designated
according to the German Soil Survey Description (AG-Boden,
1994; FAO, 1994). Table 1 gives a detailed description of the
investigated soils.
Sample pretreatment
Roots and visible plant remains were removed from the
samples where possible. The samples were frozen and freeze-
dried. Soil aggregates were crushed and the fraction > 2 mm
was removed by dry sieving. The pH (in 0.01 M CaCl2) was
measured with a glass electrode in the supernatant suspension
of a 2.5:1 (water:soil by weight).
L
352 M. W. I. Schmidt et al.
# 1999 Blackwell Science Ltd, European Journal of Soil Science, 50, 351±365
R
Ta
ble
1C
hem
ical
and
physi
cal
pro
per
ties
of
the
soil
s
Tex
ture
Dep
thIn
org
anic
CO
rgan
icC
To
tal
NS
and
Sil
tC
lay
Co
lou
rP
reci
pit
atio
n
No
Cla
ssi®
cati
on
aH
ori
zon
b/c
m_________
_______________
_/g
kg
±1
______________
__________
C:N
pH
(CaC
l 2)
__________
______
/mas
s%
______________
__
/Y-C
IEv
alu
e/m
m
1H
apli
cA
liso
lA
h0±14
<0.1
15
.11
.69
3.7
46
.93
7.4
15
.7±
48
0
2D
yst
ric
Gle
yso
lA
h0±20
<0.1
17
.51
.21
55
.48
2.4
15
.73
.3±
80
0
3H
um
icC
amb
iso
lA
xh
20±45
<0.1
37
.42
.31
64
.12
1.0
53
.32
5.7
9.4
14
00
AxhB
v45±55
<0.1
27
.91
.81
64
.02
0.9
48
.83
0.3
10
.0
4H
um
ic-s
tag
nic
Axp
0±30
4.0
15
.71
.21
37
.07
4.0
17
.78
.37
.35
70
Lu
vis
olc
Axh
30±40
1.0
12
.01
.01
26
.96
8.2
21
.61
1.4
7.6
AxhB
t40±60
2.0
4.8
0.4
12
6.9
46
.93
5.4
17
.7±
5H
apli
cP
hae
oze
mA
xp
0±20
<0.1
23
.72
.01
25
.93
1.0
49
.91
9.3
6.0
48
0
Axh
20±50
<0.1
22
.61
.91
26
.13
1.8
49
.01
9.2
6.2
6H
apli
cC
her
no
zem
Axp
0±20
<0.1
23
.62
.11
17
.43
.17
5.2
21
.76
.16
30
Axh
20±60
<0.1
17
.81
.31
47
.61
.67
4.6
23
.95
.5
7H
apli
cC
her
no
zem
Axp
0±20
<0.1
18
.31
.51
27
.23
.08
0.2
16
.87
.96
30
Axh
20±45
<0.1
11
.91
.49
7.2
3.1
79
.31
7.6
8.6
8H
apli
cG
rey
zem
Axp
0±20
9.0
5.0
0.5
10
7.5
3.3
78
.71
8.0
16
.06
30
Axh
20±45
2.0
12
.81
.21
17
.41
.67
9.6
18
.71
7.1
9H
apli
cG
rey
zem
Axp
0±20
<0.1
13
.51
.41
07
.02
.47
9.6
18
.01
4.0
63
0
Axh
20±45
<0.1
13
.31
.11
27
.30
.97
7.0
22
.11
7.0
aA
cco
rdin
gto
FA
O(1
994).
bA
ccord
ing
toA
G-B
oden
(199
4).
cD
ata
for
tex
ture
fro
mS
chim
min
get
al.
(19
93
).
Charred organic carbon in German chernozemic soils 353
# 1999 Blackwell Science Ltd, European Journal of Soil Science, 50, 351±365
Elemental analysis
For elemental analysis, a subsample was milled in a ball mill
for 10 min. Carbon and nitrogen contents were determined by
dry combustion in duplicate with an Elementar Vario EL. The
minimum detectable amounts were 0.1 6 0.3 g kg±1 for C and
N. For the various fractions investigated during the photo-
oxidation procedure, carbon was determined by a modi®ed
chromic acid digestion procedure (Heanes, 1984).
Photo-oxidation
Prior to the photo-oxidation procedure, each soil was saturated
with Na+ to facilitate the separation of silt and clay particles.
Na+ saturation was completed by shaking 10 g of soil in 50 ml
of 5 M NaCl in centrifuge tubes and centrifuging at 850 g for
10 min and ®ltering the supernatant (GF/A glass ®lter paper).
Samples were washed with 1 M NaCl twice and then dialysed
against distilled water until free of salt. Each sample was
transferred to a 250-ml beaker with 100 ml of water and treated
with ultrasonic energy for 10 s using a Branson B-30 soni®er
set to 50% power and 50% pulse and ®tted with a probe
(13 mm diameter). The soni®ed sample was allowed to stand
for 3 min to allow large particles to settle, and the supernatant
was poured through a nest of sieves (200 and 53�m) into a
measuring cylinder. Ultrasonic treatment and decantation was
repeated twice. After the third soni®cation, the remaining
sample was transferred to the cylinder through the sieves, and
the < 53-�m fraction was made to 500 ml.
Aliquots of the < 53-�m suspension, comprising » 2.5 mg
organic C, were placed in quartz test tubes and made up to
20 ml with water. To keep the sample in suspension, air was
introduced to the bottom of each tube through stainless-steel
capillary tubing (50 ml min±1). A stainless-steel cold-®nger
condenser was inserted to maintain room temperature. All
suspensions where photo-oxidized at 2.5 kW for 2 h (Skjem-
stad et al., 1996). After photo-oxidation, suspensions were
transferred to centrifuge tubes, 0.5 ml of saturated Al2(SO4)3
solution added and the samples centrifuged at 850 g for 10 min
and washed twice with water. Several samples (30±50) of each
soil were photo-oxidized and combined prior to HF treatment
(Skjemstad et al., 1994) and then freeze-dried for subsequent13C NMR studies.
Solid-state 13C NMR spectroscopy
Samples of the bulk soils treated with HF and the photo-
oxidized < 53-�m residues were packed in a 7-cylindrical
zirconia rotor (diameter 7 mm) with Kel-F caps and spun at
5 kHz at a frequency of 50.309 MHz in a Doty Scienti®c MAS
probe. Spectra were obtained at a 13C resonance frequency of
50.3 MHz on a Varian Unity 200 spectrometer using magic
angle spinning (MAS) and a standard cross-polarization (CP)
pulse sequence (Wilson, 1987). A contact time of 1 ms was
used, and recycle delay time was 0.3 s to ensure complete
relaxation between scans (recycle delay > 7 T1H). The spectra
were plotted between ±100 and 300 p.p.m. using a Lorentzian
line broadening of 50 kHz and a Gaussian function broadening
of 0.004 s. Chemical shift assignments are given in
Table 2. The NMR spectra were divided into ®ve chemical
shift regions representing alkyl C (0±46 p.p.m.), O±alkyl C
(46±110 p.p.m.), aryl C (110±145 p.p.m.), O±aryl C (145±165
p.p.m.), carbonyl C (165±190 p.p.m.) and aldehyde/ketone C
(190±220 p.p.m.).
Determination of the content of charred organic carbon
Spectra were corrected for spinning sidebands from aromatic C
(230 and 30 p.p.m.) and carbonyl C (273 and 73 p.p.m.) by
integrating the signal intensities between 220 and 250 p.p.m.,
respectively 250 and 295 p.p.m. Assuming that the correspond-
ing sidebands at 30 and 73 p.p.m. are of equal size, integrals
were doubled and added to the aromatic and carbonyl region
and subtracted from the alkyl C and O±alkyl C region.
Spinning sidebands from O±aryl C spectra were not differ-
entiated because of their small contributions to the total signal
intensities. The content of COC is determined by the amount
of aryl carbon corrected for lignin content if following photo-
oxidation a recognizable O±aryl C peak is observed. For the
L
Table 2 Chemical shift assignment of peaks in the solid-state 13C CPMAS NMR spectra which is referenced to tetramethylsilane = 0 p.p.m.
(Wilson, 1987; LuÈdemann & Nimz, 1973)
Chemical shift
range /p.p.m. Assignment
220±190 Aldehyde, ketone
190±160 Carbonyl carbons
160±140 Aromatic COR or CNR carbons
140±120 Aromatic C-H carbons, guaiacyl C-2, C-6 in lignin, ole®nic carbons
120±100 Anomeric carbon of carbohydrates, C-2, C-6 of syringyl units in lignin
100±60 Carbohydrate-derived structures (C-2 to C-5) in hexoses, C-� of some amino acids, higher alcohols
60±45 Methoxyl groups and C-6 of carbohydrates and sugars, C-� of most amino acids
45 to ±10 2°, 3° and 4° carbons in alkyl structures, methyl carbons
354 M. W. I. Schmidt et al.
# 1999 Blackwell Science Ltd, European Journal of Soil Science, 50, 351±365
samples studied here such a correction was considered
unnecessary.
In materials that contain carbon nuclei remote from protons
(> 4±5 bond lengths), effective cross-polarization will not
occur, and these nuclei will not be detected by CPMAS
spectroscopy (Alemany et al., 1983; Snape et al., 1989). As a
result, highly aromatic structures, such as those found in COC,
will be underestimated by CPMAS spectroscopy. Bloch decay
does not require the proximity of protons, and relies on
relaxation processes through the C nuclei. To analyse all
samples by Bloch decay for COC content would be totally
impractical, however, considering the long recycle times
required (typically > 60 s) and the small amount of sample
available from the photo-oxidation method. We therefore
applied an empirically determined correction to the CPMAS
data in an attempt to correct for the underestimation of
aromatic carbon. Skjemstad & Taylor (unpublished data)
compared a number of photo-oxidized samples by both
CPMAS and Bloch decay. These data were used to develop
an equation to transform CPMAS-determined aryl C data to
that determined by Bloch decay (r2 = 0.98). Using this
correction, the COC content (g COC kg±1 soil) was calculated
as equal to the corrected aryl C contribution to the total signal
intensity (arylcorr). Although this approach suffers from some
uncertainty, it does provide a conservative measure of the
content of COC. From the corrected aryl C content (arylcorr) of
the < 53-�m fraction and its relative contribution to the bulk
soil, it is possible to calculate the content of COC in that
fraction normalized to the soil mass. To exclude differences in
organic C content in the different horizons, the COC content
was also normalized to the soil C content.
Scanning electron microscopy
Scanning electron microscopy (SEM) was carried out on a
Cambridge Stereoscan S250 on samples coated with 20±30 nm
of carbon. Elemental characterization was done using a Link
AN1000 EDX analyser.
CuO oxidation
Alkaline CuO/NaOH oxidation was carried out by the method
of KoÈgel-Knabner (1995). Triplicate subsamples containing
» 50 mg organic C were added to 2 M NaOH in Te¯on vials
together with CuO. The vials were sealed under N2 and heated
for 2.5 h at 170°C. After ®ltration, humic acids were removed
by acidi®cation and centrifuging. The supernatant was
extracted with conventional clean-up columns (Baker,
Germany), and silyl derivatives were separated and quanti®ed
by GC-FID (gas chromatography-¯ame ionization detection).
The total yields were determined for vanillyl (vanillin + va-
nillic acid), syringyl (syringealdehyde + syringic acid) and
cinnamyl units (p-coumaric acid + ferulic acid), respectively.
Amounts are given in g per kg of organic carbon (g VSC kg±1
C). Relative standard deviation for the detection of phenolic
products is 9% (n = 10). The p-hydroxyl compounds are not
included in the sum of phenolic oxidation products because
they can also be derived from non-lignin structures. The ratios
of acid-to-aldehyde of vanillyl (ac:al)V and syringyl (ac:al)S
and of syringyl-to-vanillyl (S:V) are calculated as mass ratios.
Colour
The colour of dry soil samples was determined with Munsell
colour charts. For standardization and better resolution, the
degree of darkness was measured with an ASD ®eld spectro-
meter under sunlight, and is reported as Y-CIE value, which
can be converted to Munsell values and corresponds closely to
human observations (Schulze et al., 1993).
Results and discussion
We now report the results obtained from 17 horizons,
originating from nine soils, and discuss them. First, the A
horizons of two non-chernozemic light-coloured soils (soils 1
and 2) are compared. Second, data for A horizons originating
from three chernozemic soils (soils 3±5) are presented, and
®nally a colour sequence of four chernozemic soils (soils 6±9)
with a gradual change from black to grey colour is
investigated.
Non-chernozemic soils
The Ah horizon of the Haplic Alisol (soil 1) contains 15.1 g
organic C kg±1 soil (Table 3). A major proportion of the bulk
soil C (57%) is in the < 53-�m fraction, subsequently
investigated by 13C CPMAS NMR spectroscopy (Figure 1).
The spectrum of the < 53-�m fraction is dominated by peaks in
the alkyl C region (±10 to 45 p.p.m.) with major peaks
representing methylene carbon (33 and 30 p.p.m.). Signals in
the O±alkyl C region (45±110 p.p.m.) indicate the presence of
polysaccharide and alcoholic structures. Signals near 56 p.p.m.
can be attributed to methoxyl groups and to C-�. The shoulder
at 65 p.p.m. is due to C-6 carbon in polysaccharides. Signals
near 74 p.p.m. correspond with ring carbon C-5, C-3 and C-2,
whereas resonances at 104 p.p.m. are due to anomeric C-1
carbon. Some resonances in the aryl C region (145±110 p.p.m.)
are attributable to protonated and alkyl-substituted aryl carbon
(116 and 131 p.p.m.). These signals for aryl C in combination
with O±aryl C structures (165±145 p.p.m.) are typical for the
presence of lignin units, whereas COC would be characterized
by signals in the aryl C region centred around 130 p.p.m. with
few additional peaks in the O±aryl C region. The carbonyl C
region (190±165 p.p.m.) is dominated by a set of resonances at
173 p.p.m. most likely due to amide carbon. The aldehyde/
ketone C region (220±190 p.p.m.) reveals only small signal
intensities, which here can be considered to be of little
importance (Preston & Ripmeester, 1983).
R
Charred organic carbon in German chernozemic soils 355
# 1999 Blackwell Science Ltd, European Journal of Soil Science, 50, 351±365
L
Ta
ble
3O
rgan
icca
rbo
nin
the
inves
tigat
edso
ils
and
frac
tio
ns.
Cco
nte
nt,
C-s
pec
ies
asd
eter
min
edb
y13C
CP
MA
SN
MR
spec
trosc
op
y,
calc
ula
ted
con
ten
to
fch
arre
do
rgan
icca
rbo
n
(CO
C)
and
lig
nin
par
amet
ers
are
com
par
edin
the
bulk
soil
and
inth
e<
53
-�m
frac
tio
nb
efo
rean
daf
ter
hig
h-e
ner
gy
UV
ph
oto
-ox
idat
ion
Lig
nin
par
amet
ers
Sam
ple
:O
rgan
icca
rbon
Csp
ecie
sC
OC
c
bu
lkso
ilY
ield
Rat
ioac
:ad
or
frac
tion
/gkg
±1
/%ket
one
carb
ox
yl
O±
ary
lar
yl
O±
alk
yl
alk
yl
ary
l corr
b/g
kg
±1
/%/g
VC
SS
:V
No
Ho
rizo
na
/�m
of
bulk
soil
___________
______________
______
/%o
fto
tal
sig
nal
inte
nsi
ty___________
______________
______
of
bu
lkso
ilk
g±1
CV
anil
lin
Sy
rin
gy
l
1A
hb
ulk
soil
15.1
100
31
22
74
33
4±
±N
D8
.50
.78
1.3
80
.90
<5
38.6
57
±±
±±
±±
±±
±±
±±
<5
3U
V1.6
11
11
33
54
23
5N
DN
D±
±±
±
2A
pb
ulk
soil
17.5
100
19
28
33
47
±±
ND
11
.70
.18
0.8
22
.24
<5
38.7
50
±±
±±
±±
±±
±±
±±
<5
3U
V1.8
10
27
51
52
94
3N
DN
D±
±±
±
3A
xh
bu
lkso
il37.4
100
11
32
28
30
26
±±
15
2.9
0.2
71
.98
0.3
8
<5
320.3
54
±±
±±
±±
±±
±±
±±
<5
3U
V6.1
16
21
46
63
87
92
5.6
±±
±±
Ax
hB
vb
ulk
soil
27.9
100
11
44
31
23
26
±±
30
1.0
0.2
81
.75
0.3
1
<5
319.4
70
±±
±±
±±
±±
±±
±±
<5
3U
V8.3
30
21
36
59
10
98
97
.4±
±±
±
4A
xp
bu
lkso
il15.7
100
31
34
15
41
24
±±
41
2.2
0.4
10
.57
1.5
6
<5
311.2
71
±±
±±
±±
±±
±±
±±
<5
3U
V1.2
82
11
63
92
51
66
10
.7±
±±
±
Ax
hb
ulk
soil
12.0
100
51
74
21
35
19
±±
15
8.0
0.5
10
.57
1.8
3
<5
38.7
73
±±
±±
±±
±±
±±
±±
<5
3U
V2.3
19
21
07
41
19
21
72
1.7
±±
±±
Ax
hB
tb
ulk
soil
4.8
100
41
44
17
39
22
±±
8±
±±
±
<5
32.7
57
±±
±±
±±
±±
±±
±±
<5
3U
V0.6
13
31
05
34
30
19
62
0.4
±±
±±
5A
xp
bu
lkso
il23.7
100
11
07
18
35
30
±±
91
5.5
0.2
60
.63
1.6
5
<5
317.9
76
±±
±±
±±
±±
±±
±±
<5
3U
V4.1
17
31
39
39
21
16
66
2.7
±±
±±
Ax
hb
ulk
soil
22.6
100
31
36
22
32
24
±±
15
15
.20
.31
0.6
31
.49
<5
316.0
71
±±
±±
±±
±±
±±
±±
<5
3U
V4.0
18
21
28
46
18
14
77
3.1
±±
±±
356 M. W. I. Schmidt et al.
# 1999 Blackwell Science Ltd, European Journal of Soil Science, 50, 351±365
R
Ta
ble
3C
on
tin
ued
Lig
nin
par
amet
ers
Sam
ple
:O
rgan
icca
rbon
Csp
ecie
sC
OC
c
bu
lkso
ilY
ield
Rat
ioac
:ad
or
frac
tion
/gkg
±1
/%ket
one
carb
ox
yl
O±
ary
lar
yl
O±
alk
yl
alk
yl
ary
l corr
b/g
kg
±1
/%/g
VC
SS
:V
No
Ho
rizo
na
/�m
of
bulk
soil
_____________
______________
____
/%o
fto
tal
sig
nal
inte
nsi
ty_____________
_______________
___
of
bu
lkso
ilk
g±1
CV
anil
lin
Sy
rin
gy
l
Co
lou
rse
qu
ence
6A
xp
bu
lkso
il23.6
100
11
45
21
32
26
±±
19
6.6
0.2
00
.69
1.5
1
<5
316.7
71
±±
±±
±±
±±
±±
±±
<5
3U
V4.9
21
11
37
53
16
11
84
4.1
±±
±±
Ax
hb
ulk
soil
17.8
100
01
76
37
22
18
±±
45
1.7
0.1
90
.65
1.5
2
<5
314.6
82
±±
±±
±±
±±
±±
±±
<5
3U
V8.0
45
11
39
65
67
95
7.6
±±
±±
7A
xp
bu
lkso
il18.3
100
21
24
18
37
28
±±
14
11
.20
.21
0.6
71
.57
<5
312.1
66
±±
±±
±±
±±
±±
±±
<5
3U
V4.2
23
31
38
46
13
17
77
3.2
±±
±±
Ax
hb
ulk
soil
11.9
100
11
43
19
36
28
±±
23
2.5
0.2
50
.65
2.1
7
<5
38.3
70
±±
±±
±±
±±
±±
±±
<5
3U
V3.9
33
21
48
51
12
12
82
3.2
±±
±±
8A
xp
bu
lkso
il5.0
100
31
34
14
42
24
±±
51
7.8
0.3
70
.61
1.1
0
<5
32.8
56
±±
±±
±±
±±
±±
±±
<5
3U
V0.7
14
41
07
27
33
19
52
0.4
±±
±±
Ax
hb
ulk
soil
12.8
100
51
45
15
41
21
±±
31
0.4
0.4
50
.46
1.1
6
<5
38.9
70
±±
±±
±±
±±
±±
±±
<5
3U
V1.3
10
61
27
22
34
20
44
0.6
±±
±±
9A
xp
bu
lkso
il13.5
100
21
44
12
40
28
±±
21
4.9
0.3
50
.58
1.2
4
<5
311.2
83
±±
±±
±±
±±
±±
±±
<5
3U
V0.8
61
10
82
43
32
34
80
.4±
±±
±
Ax
hb
ulk
soil
13.3
100
31
44
13
40
27
±±
37
.30
.18
0.7
41
.75
<5
38.7
65
±±
±±
±±
±±
±±
±±
<5
3U
V1.3
10
49
82
63
32
05
00
.7±
±±
±
±,
no
td
eter
min
ed.
ND
,not
det
ecta
ble
.a
Acc
ord
ing
toA
G-B
oden
(1994).
bar
yl
corr
=ar
yl
con
tent
corr
ecte
dfo
rB
loch
dec
ay;
for
det
ails
see
Mat
eria
lsan
dm
eth
od
s.c
CO
C,
char
red
org
anic
carb
on.
Charred organic carbon in German chernozemic soils 357
# 1999 Blackwell Science Ltd, European Journal of Soil Science, 50, 351±365
All labile organic matter is removed by photo-oxidation for
2 h, and only physically or chemically protected matter
remains (Skjemstad et al., 1996). In fact, most of the organic
carbon is removed by this treatment, leaving about 10% of the
bulk soil C (Table 3). Comparing the resulting spectrum with
the spectrum obtained before photo-oxidation, only minor
alterations are apparent in shape and integrated areas.
Although peak heights in the alkyl C and O±alkyl C regions
change, the integrated areas vary only within the error range.
Intensities decrease in the aryl C and ketone C region (±2%)
while intensities for O±aryl and carboxyl C slightly increase
accordingly (+1%). These alterations probably can be
explained by a greater line broadening due to the smaller C
content of the sample after photo-oxidation. As a result, the
organic carbon resistant to photo-oxidation presumably is
protected physically inside organo-mineral complexes, reveal-
ing a bulk chemical structure similar to the carbon fraction
prior to the treatment. Skjemstad et al. (1996) observed similar
results for other soils. Aromatic carbon is present mainly in
lignin-derived phenols, as clearly demonstrated by typical
signals at 130 p.p.m. in combination with peaks in the O±aryl
C region. As a result, the presence of COC cannot be
determined by this technique in the Ah horizon of the Haplic
Alisol.
The Ap horizon obtained from the Dystric Gleysol (soil 2)
comprises 17.5 g C kg±1 dry soil with 50% of the bulk soil C in
the fraction < 53�m. The spectrum of this fraction reveals
signals mainly in the chemical shift regions, as discussed
above. The main differences occur in the aryl C region with a
more distinct peak around 130 p.p.m. with shoulders at 120
and 140 p.p.m. indicating the presence of lignin. Contributions
from aryl C are slightly larger (8% of the total signal intensity)
compared with the Alisol (7%), suggesting greater contribu-
tions of lignin to the organic matter.
The lignin signature, as determined by CuO oxidation,
provides information on the structure of aromatic carbon and
re¯ects the vegetation from which the soil organic matter was
derived (Ertel & Hedges, 1984). Lignin from gymnosperm
wood produces mainly vanillyl-type (V) oxidation products,
whereas lignin from angiosperm wood produces syringyl units
(S) in addition to vanillyl units. Large yields of cinnamyl units
(C) are characteristic of non-woody tissues of both types of
plants. The proportion of lignin-derived phenols (g VSC kg±1
C) and the ratio acid:aldehyde (ac:ad) for syringyl (S) and
vanillyl (V) phenols varies among plant species (Sarkanen &
Ludwig, 1971; Ertel & Hedges, 1984). Large ratios (ac:ad)V
and (ac:ad)S indicate a greater degree of decomposition of
lignin phenols. As detected by CuO oxidation, proportions of
lignin-derived phenols are greater in the Ap horizon of the
Gleysol (11.7 g VCS kg±1 C) than in the Alisol (8.5 g VCS kg±1
C; Table 3), which con®rms conclusions from NMR spectro-
scopy. As in the Ah horizon of the Alisol, 10% of the bulk soil
C resists photo-oxidation. Again, signal intensities in the 13C
CPMAS spectrum of this fraction show only small alterations,
except for an increase in the aryl C region (15%) compared
with the untreated sample (8%). In the Gleysol these signals
can be assigned to lignin phenols, as indicated by typical
shoulders at 153 p.p.m.
In summary, we could not detect COC in the A horizons of
the non-chernozemic soils (Alisol, Gleysol) by the procedure
of photo-oxidation followed by 13C CPMAS NMR spectro-
scopy.
Chernozemic soils (soils 3±5)
The Humic Cambisol (soil 3) shows similar C:N ratios and pH
values in both horizons studied, whereas texture and the
content of organic matter differ. The AxhBv horizon contains
slightly more clay (30.3 mass percentage) compared with the
overlying Axh horizon (25.7 mass percentage). The C and N
contents are greater in the Axh horizon than in the underlying
AxhBv horizon, which probably re¯ects the larger input of
roots and leaves mixed into this horizon by bioturbation. Also
in this soil, the < 53-�m fraction contains a major proportion of
the bulk soil C in both the Axh horizon (54%) and the AxhBv
horizon (70%).
The 13C CPMAS spectra (Figure 2) display similar peak
heights and integrated areas for alkyl C, carboxyl C and ketone
C. Some differences between the spectra can be found in the
O±alkyl C region and in the aryl and O±aryl C regions.
Contributions from O±alkyl C are greater in the Axh horizon
(30%) than in the underlying AxhBv horizon (23%), in
accordance with the larger proportion of polysaccharides
derived from plant litter in the Axh horizon. The contributions
from aryl C (Table 3) are greater in the Axh horizon (28%) and
in the AxhBv horizon (31%) than in soils 1 and 2 (< 8%).
These large contributions can be attributed only partly to lignin
compounds. As demonstrated before, lignin typically produces
signals in the aryl C region in combination with resonances in
the O±aryl C region (165±145 p.p.m.), but in the spectra
obtained from the Humic Cambisol, signals in the O±aryl C
L
Figure 1 Solid-state 13C NMR spectra of the investigated A
horizons from (1) the Haplic Alisol and (2) the Dystric Gleysol
before (left) and after high-energy UV photo-oxidation (right).
358 M. W. I. Schmidt et al.
# 1999 Blackwell Science Ltd, European Journal of Soil Science, 50, 351±365
region are weak, suggesting that lignin contributes only part of
the observed aromatic carbon. This is con®rmed by small
yields of CuO oxidation products (< 2.9 g VSC kg±1 C)
compared with those of soils 1 and 2 (Table 3). These results
indicate the presence of unsubstituted aromatic carbon in both
horizons of the Phaeozem. If this signal were due to COC, the
material would resist photo-oxidation, and the resulting
spectrum would show large signal intensities for aryl C. In
fact, a larger proportion of soil organic carbon resists photo-
oxidation (Axh 16% and AxhBv 30%) than in soils 1 and 2
(< 11%). The spectra of both horizons, after photo-oxidation,
are dominated by aryl C peaks at 130 p.p.m. with spinning
sidebands present at 230 and 30 p.p.m. At 30 p.p.m. these
signals overlap with resonances of alkyl C, and minor signals
indicate the presence of some O±alkyl C (75 p.p.m.) and
carboxyl C (170 p.p.m.). Skjemstad et al. (1993) showed that
carbonyl bands remaining after photo-oxidation are most likely
due to carboxylic acids. The absence of signals in the O±alkyl
C region indicates the absence of carbohydrate carbon which is
typical for COC.
To investigate the nature of this material further we
observed its morphology by scanning electron microscopy
(SEM). Micrographs are shown in Figure 3(a,b). Many of the
larger particles have a morphology characteristic of the xylem
structure of woody material. These and many of the ®ne
particles were probed with an EDX system and found to
contain no elements with an atomic weight greater than
sodium, except for sulphur. Few particles could be identi®ed as
mineral. This suggests that many of the particles are organic
and probably are COC.
If a major proportion of aromatic carbon is present in
unsubstituted structures, this poses questions about the
quantitative reliability of the CPMAS spectra obtained from
soil 3. In cross-polarization (CP) experiments, the polarization
energy for carbon nuclei is transferred via hydrogen nuclei
with an ef®ciency depending on the proximity of these nuclei.
Consequently, the carbon nuclei in unsubstituted structures are
underestimated by CP experiments relative to Bloch decay
experiments, in which carbon nuclei are polarized directly.
The samples contain little carbon. Long recycle times would
therefore be required in Bloch decay experiments, and it would
take several days to obtain a single spectrum which would still
give a poor signal-to-noise ratio. With so many samples Bloch
decay experiments were quite impractical. We therefore
introduced an empirically determined correction for contribu-
tions of unsubstituted aromatic carbon (for details see
Materials and methods). We assumed that the resulting
corrected signal intensities for unsubstituted aromatic carbon
(arylcorr) all originate from COC. For the < 53-�m fractions the
results are expressed relative to the bulk soil, i.e. normalized to
organic C and to soil mass of the bulk soil, respectively (Table
3). In the < 53-�m fractions from both horizons of the
Cambisol, organic carbon is almost exclusively present in
unsubstituted aromatic structures with negligible contributions
from other C species, as revealed by corrected aromaticity
(Table 3). Expressed relative to the bulk soil, COC contributes
30% to the bulk soil C in the AxhBv horizon, which is
equivalent to 7.4 g COC kg±1 bulk soil. In the Axh horizon,
COC contributes less to the bulk soil C (15%) and to the bulk
soil mass (5.6 g COC kg±1 bulk soil).
As a result, the Humic Cambisol contains considerable
amounts of a material which (i) resists photo-oxidation, (ii)
consists of unsubstituted aromatic carbon, and (iii) displays a
morphology typical of cell wall material from plants, which
provides evidence for the presence of charred plant remains or
COC.
Three horizons from a Humic-stagnic Luvisol (soil 4)
with dark A horizons were studied. Progressing from Axp
to AxhBt horizons, the C and N contents decrease, whereas
clay contents increase (8.3 to 17.7 mass percentage clay;
R
Figure 2 Solid-state 13C NMR spectra of the investigated A
horizons from (3) the Humic Cambisol, (4) the Humic-stagnic
Luvisol and (5) the Haplic Phaeozem before (left) and after high-
energy UV photo-oxidation (right). Asterisks (*) indicate spinning
sidebands from carbonyl C (273 p.p.m.) and aromatic C
(230 p.p.m.). For simplicity, the corresponding sidebands at 73 and
30 p.p.m., respectively, were not labelled.
Charred organic carbon in German chernozemic soils 359
# 1999 Blackwell Science Ltd, European Journal of Soil Science, 50, 351±365
L
Fig
ure
3E
lect
ron
mic
rogra
phs
of
frag
men
tsof
char
red
org
anic
carb
on
(mar
ked
wit
han
aste
risk
*)
fro
mth
e<
53
-�m
frac
tio
ns
afte
rh
igh
-en
erg
yU
Vp
ho
to-o
xid
atio
n.
Sam
ple
so
rig
inat
e
fro
m(a
)th
eA
xh
ho
rizo
nan
d(b
)th
eA
xhB
vhori
zon
of
aH
um
icC
amb
iso
l(s
oil
3),
and
(c)
the
Ax
hh
ori
zon
(so
il4
)an
d(d
)th
eA
xh
ho
rizo
no
fa
Hap
lic
Ph
aeo
zem
(so
il5
).S
cale
bar
:
(a)
10�
m;
(b)±
(d)
40�
m.
360 M. W. I. Schmidt et al.
# 1999 Blackwell Science Ltd, European Journal of Soil Science, 50, 351±365
Table 1). The C:N ratios are uniformly small (12±13), and
pH is constantly neutral (6.9±7.0) throughout the pro®le.
For all three horizons the < 53-�m fraction represents
> 57% of the soil C.
The 13C CPMAS spectra (Figure 2) obtained from the Axp
and Axh horizons demonstrate that contributions from aryl C
(15±21% of the total signal intensity) are intermediate between
those detected for the Humic Cambisol (soil 3) (28±31%) and
those observed for the Alisol and the Gleysol (< 8%). Typical
signals for lignin structures in the O±aryl C region indicate that
some of the signal intensities in the aromatic region can be
attributed to lignin structures. This conclusion is supported by
larger proportions of lignin-derived phenols in the Axp (12.2 g
VSC kg±1 C) and in the Axh horizon (8.0 g VSC kg±1 C). After
photo-oxidation, C spectra show a dominant peak at 130 p.p.m.
and a large contribution in the aryl C regions (> 34% of the
total signal intensities). Weak signals in the O±aryl C region
(4%) indicate that lignin structures contribute in only a minor
way to the organic carbon present in this fraction. Microscopic
investigation (Figure 3c) demonstrates the presence of ®nely
divided angular particles with a cellular morphology, char-
acteristic of COC. Contrasting with the samples from the
Humic Cambisol, the particles are much ®ner and are more
dif®cult to attribute exclusively to COC. This is also re¯ected
by the spectra, which in addition to signals in the aryl C region
show peaks at 30, 75, 106 and 175 p.p.m. indicating the
presence of alkyl C, O±alkyl C and carbonyl C. Progressing
through the pro®le, calculated contributions of COC to the
bulk soil C are smallest in the Axp horizon (4% of the bulk soil
C), largest for the Axh horizon (15%) and intermediate (8%) in
the AxhBt horizon. The same pattern is true for the absolute
amounts of COC in these horizons.
As a typical representative of the chernozemic soils in the
Halle-Magdeburg area, a Haplic Phaeozem under agricultural
practice was studied (soil 5). Only minor variations between
Axp and Axh horizons were observed for C and N content, pH
and texture. For both horizons, > 71% of the bulk soil C was
recovered in the < 53-�m fraction for photo-oxidation. The
spectra obtained from these fractions are almost identical to
those obtained from the non-fractionated soils as presented
previously (Schmidt et al., 1996), showing that the fraction
investigated can be regarded as representative for the bulk soil.
Because of the presence of peaks at 153 p.p.m., some of the
signal intensity in the aromatic C region can be attributed to
lignin structures, as con®rmed by large proportions of lignin-
derived phenols (> 15.2 g VSC kg±1 C).
Almost identical amounts of organic C resist photo-
oxidation in the Axp and Axh horizons (17 and 18% of the
bulk soil C, respectively). Also the 13C CPMAS spectra
(Figure 2) are similar, with the largest intensities in the aryl C
region peaking at 130 p.p.m. (Axp 39% and Axh 46% of the
total signal intensity). The absence of distinct signals near
153 p.p.m. suggests that aryl signals can be assigned to COC,
which is supported by scanning electron microscopy of the
fraction obtained from the Axh horizon (Figure 3d). As can be
seen from Table 3, contributions of COC increase with depth,
both normalized to the bulk soil C (Axp 9% and Axh 15% of
the bulk soil C) as well as normalized to the bulk soil mass.
Again, the presence of COC could not be detected in the
light-coloured non-chernozemic soils (Haplic Alisol, Dystric
Gleysol), whereas in chernozemic soils (Humic Cambisol,
Humic-stagnic Luvisol, Haplic Phaeozem) COC contributed
up to 30% to the bulk soil carbon with largest contents in the
subsurface horizons (Axh, AxhBv or Ah) and smallest contents
in surface horizons (Axh, Axp). These proportions of COC are
similar to those reported by Skjemstad et al. (1996) for some
Australian soils.
Colour sequence (soils 6±9)
The apparent coincidence of dark soil colour with the presence
of COC raises questions of a potential relation between these
two variables. Soil colour and organic matter content are more
closely related for soils occurring together in landscapes and
having similar texture and parent material than for soils over a
wide geographic region or widely varying in texture (Schulze
et al., 1993). We therefore studied the relation between them in
a 10-km colour sequence of four chernozemic soils. The soils
studied had developed from almost identical parent material
(loess) and have similar chemical and physical properties, and
they are all under agriculture with similar crop rotations. The
striking feature is a gradual change from black to grey. For a
more accurate quanti®cation of the small differences in colour,
the Munsell value of each soil was determined by measuring
the re¯ectance in the laboratory with a diffuse re¯ectance
spectrometer and expressed as Y-CIE value (Table 1).
The C contents (Table 1) for the ploughed horizons (Axp)
vary between 5.0 and 23.6 g C kg±1 soil, whereas the subsoils
vary less (Axh 11.9±17.8 g C kg±1), and show variations
typically found in this region (Gehrt, 1999). The C:N ratios
are uniformly small (9±14) in all horizons. Also, the texture is
similar in all soils and horizons, with small sand contents
(< 3.3 mass percentage) and typically large silt contents (77.0±
80.2 mass percentage). Clay contents (16.8±23.9 mass
percentage) are often slightly larger in the Axh horizons than
in the overlying Axp horizons. There is a little variation in the
pH (7.0±7.6).
For the sequence investigated (Table 3), the sum of lignin-
derived phenols is typically largest in the Axp horizons (6.6±
17.8 g VSC kg±1 C) and somewhat smaller in the correspond-
ing Axh horizons (1.7±10.4 g VSC kg±1 C). This indicates the
larger input of plant material from agricultural crops in the
ploughed horizon than in the Axh horizons. As demonstrated
by the acid:aldehyde ratios, the degree of lignin decomposition
varies only slightly in the horizons investigated: (ac:ad)V 0.18±
0.45; (ac:ad)S 0.46±0.74. Also, the ratios of syringyl:vanillyl
phenols (S:V) are similar (1.10±2.17), con®rming the larger
input from gymnosperms than from the angiosperm vegetation
R
Charred organic carbon in German chernozemic soils 361
# 1999 Blackwell Science Ltd, European Journal of Soil Science, 50, 351±365
of the forest (soil 1 < 0.38) and mixed deciduous vegetation
(0.90) (Sarkanen & Ludwig, 1971; Ertel & Hedges, 1984). As
a result, the lignin signature suggests that the sources of
organic matter are currently similar in the soils of the colour
sequence.
Taking into account the similar pedogenetic factors and soil
properties, including the similar C content, one could expect a
similar soil colour (Schulze et al., 1993). However, in
progressing from soil 6 to soil 9, the colour changes gradually
from black to grey. The visual assessment is con®rmed by the
spectrometrically determined Y-CIE values, which are an
equivalent to the Munsell values (Table 1). Small Y-CIE
values correspond with dark colours, while larger ones
correspond with lighter colours (Schulze et al., 1993).
In the colour sequence, the < 53-�m fraction comprises a
major proportion of the bulk soil carbon (56±83% of the bulk
soil C). The 13C CPMAS NMR spectra obtained from these
fractions (Figure 4, left side) are fairly uniform, except for the
aromatic C regions. Signals from aryl C (Table 3) are strongest
in the Axh horizon of soil 6 (37% of the total signal intensity)
and weakest in the Axp horizon of soil 9 (12%). After photo-
oxidation only small proportions of carbon are left, consis-
tently decreasing from soil 6 to soil 9 for both Axp (21 to 10%)
and Axh horizons (45 to 10%). As revealed by the corrected
aryl C value (arylcorr, Table 3) in the Axh horizon of soil 6,
almost all (95%) organic carbon is present in unsubstituted
aromatic structures. Also here, scanning electron microscopy
con®rms an abundance of what appears to be COC in this
fraction, which is also true for the Axh horizon from soil 7.
The largest proportion of COC in the colour sequence can be
detected in this Axh horizon from soil 6. Here, COC
contributes 45% of the soil organic carbon, which is equivalent
to 7.6 g COC kg±1 soil.
To visualize a potential interdependence between COC
content (Table 3), lignin compounds (Table 3) and soil
colour (Table 1), data are displayed as a colour sequence
from black to grey in Figure 5. To exclude variations
resulting from different C contents, the contents of COC
and lignin compounds are normalized to organic carbon
content. Progressing from black to grey (from soils 6 to 9)
two patterns for colour are evident. First, the colour of the
soil generally becomes lighter, i.e. the re¯ection expressed
as Y-CIE value increases, except for the almost identical
soils 8 and 9. Second, the Axp horizons are darker than the
corresponding Axh horizons. Parallel to these trends,
contributions of COC to the bulk soil C decrease
L
Figure 4 Solid-state 13C NMR spectra of the investigated Axp and
Axh horizons from the colour sequence before (left) and after high-
energy UV photo-oxidation (right). Asterisks (*) indicate spinning
sidebands from carbonyl C (273 p.p.m.) and aromatic C
(230 p.p.m.). For simplicity, the corresponding sidebands at 73 and
30 p.p.m., respectively, were not labelled.
Figure 5 Content of charred organic carbon (COC) and lignin
compounds detected in the Axp and Axh horizons of the
chernozemic soils 6±9 displayed as a colour sequence. Data for the
COC and lignin are displayed on the left axis; note different units
for these parameters. Re¯ection is given as Y-CIE value which is
an equivalent to the Munsell value.
362 M. W. I. Schmidt et al.
# 1999 Blackwell Science Ltd, European Journal of Soil Science, 50, 351±365
consistently in the Axp horizons (19 to 2%) and in the Axh
horizons (45 to 3%), while proportions of lignin compounds
increase, again with soils 8 and 9 being almost identical.
As a trend, progressing from dark to light soils in the colour
sequence (i) organic matter becomes less resistant to photo-
oxidation, (ii) contributions of COC to the bulk soil C
decrease, whereas (iii) contributions of lignin to the bulk soil C
increase. Following the Axp and the Axh horizons in the
colour sequence, we are tempted to conclude that soil colour
directly depends on the contribution of COC to soil organic
matter, but within the same soil pro®le this conclusion does not
seem to hold. For all the soils investigated, the surface
horizons (Ap, Axp, Axh) are darker than the corresponding
subsoils, except for soil 6 (Table 1), whereas contributions
from COC are always greatest in the subsoils (Table 3). The
lighter colour of the subsoils may be related to the smaller C
content. However, for soils 5 and 8 this explanation does not
hold, because the C contents are similar in the Axp and Axh
horizons. This observation shows that besides COC other
organic and inorganic pigments contribute to the soil's colour.
For the agricultural soils (4±9) and the forest soil (soil 3), the
contributions from COC are always greater in the subsoils than
in the surface horizons, both normalized to organic C content
and normalized to bulk soil mass. The same is true for the clay
content, except for soil 5 (Table 2). The apparent differences
between surface horizons (Ap, Axp, Axh) and subsoils may be
explained by primary differences during sedimentation of the
loess layers or by a subsequent eluviation within the pro®le
induced by cultivation.
Conclusions
The photo-oxidation method allowed us to identify charred
organic carbon (COC) in several chernozemic soils, but none
in the A horizons of the Haplic Alisol and the Dystric Gleysol.
It seems that COC is restricted to chernozemic soils, and this is
the ®rst time that COC has been detected in German
chernozemic soils. The COC contributes up to 45% of the
bulk organic carbon, which is equivalent to about 8 g kg±1 of
the bulk soil. A colour sequence of chernozemic soils, similar
in chemical and physical properties, showed a strong relation
between colour and the content of COC, suggesting that, at
least in the sequence studied, the COC dominates the colour of
the soil.
The presence of COC in chernozemic soils poses some
questions on its possible in¯uence on soil properties. The COC
might affect the soil physical properties (water holding
capacity, structural stability), chemical properties (pH, C
content, cation exchange capacity, sorption) and biological
properties (stability against biological degradation). For
example, standard determinations of the C content include
the highly recalcitrant COC, and so overestimate the content of
native soil organic matter.
The presence of COC also raises questions about the
pedogenesis of Chernozems. It is generally accepted that the
pedogenesis of Chernozems depends on continental dry
climate with little or no leaching. The characteristic deep
black A horizons are considered to develop as a result of
accumulation of organic matter and bioturbation. The
Chernozem belt developed from Siberia to Eastern Europe,
with isolated occurrences in Germany. There, Chernozems are
regarded as relic soils which developed about 8000 years ago
under a continental climate and were preserved under
favourable conditions, i.e. low precipitation or high ground
water table. Without these preserving conditions, Chernozems
are expected to be degraded by gradual decalci®cation and
accumulation of black clay±humus complexes in deeper
horizons, resulting in Phaeozems or Luvisols (MuÈckenhausen,
1977). However, the German chernozemic soils that we
studied display many properties typical of Chernozems,
although the present annual precipitation varies widely
between 480 and 1400 mm.
In the chernozemic soils we studied, the COC can be a
major contributor to the organic carbon, but the origin of this
®re-induced form of organic carbon remains unclear. It could
originate from natural vegetation ®res as well as from human
activity. Because it so recalcitrant, COC can be preserved in
the soil for a long time. Consequently, it could have originated
from ®res in the post-glacial vegetation, i.e. tundra, taiga and
later mixed deciduous forests. Today, the circumpolar taiga
may be an important zone for the production of charred
organic carbon from ®re (Wein, 1993). The COC could also
originate from post-mesolithic human use of ®re for the
clearing of the forests and subsequent agriculture. We sampled
the colour sequence in an area with an apparently random
pattern of patches with black and grey chernozemic soils on a
regional scale of some kilometres. This small-scale pattern
could be related to local human activity rather than to large
natural vegetation ®res. However, we have too little informa-
tion on the ®re and vegetation history to draw many
conclusions. The time of COC formation could be determined
by 14C analysis and provide further information on the history
of the soils.
If ®re was involved in the formation of the charred
proportion of soil organic matter in the chernozemic soils,
it may also be involved in the pedogenesis of other
chernozemic soils. This could explain the reported presence
of black ¯akes in numerous typical Chernozems from
central Russia (Kubiena, 1938; Yarilova, 1972; Pawluk,
1985), the high aromaticity of humic acids extracted from
other chernozemic soils (Kononova, 1966; Schnitzer, 1992;
Zech et al., 1997) as well as the relation between humic
acid content and soil colour for chernozemic soils from
northern Eurasia (Kononova, 1966).
Our results suggest that besides climate, vegetation and
bioturbation, ®re also plays an important role in the
pedogenesis of chernozemic soils.
R
Charred organic carbon in German chernozemic soils 363
# 1999 Blackwell Science Ltd, European Journal of Soil Science, 50, 351±365
Acknowledgements
The work was ®nancially supported by the Deutsche
Forschungsgemeinschaft (Ko 1035/6-1 and 2). A travel
grant was provided by the Massenberg foundation of the
University of Bochum. We thank Claus Schimming for
helping us select soils 1 and 6, and Lothar Beyer
(University of Kiel) for providing reference material from
soil 4. Walter Grottenthaler (Geological Survey of Bavaria)
provided samples and analytical information from soil 3.
We thank also Louis Maratos and Janine Taylor (CSIRO,
Land and Water) for soil fractionations and NMR spectro-
scopy. Experimental assistance for lignin analysis was
provided by Mathias Kempkens and Patrick GaÈrtner
(University of Bochum). Re¯ectance was measured by
Thomas Jarmer (University of Trier), and Hans-JuÈrgen
AltemuÈller (Braunschweig) provided helpful comments and
literature on charred particles in Chernozems.
References
AG-Boden 1994. Bodenkundliche Kartieranleitung. E. Schweizer-
bart'sche-Verlagsbuchhandlung, Stuttgart.
Alemany, L.B., Grant, D.M., Pugmire, R.J., Alger, T.D. & Zilm, K.W.
1983. Cross polarization and magic angle sample spinning NMR
spectra of model organic compounds. 2. Molecules of low or
remote protonation. Journal of the American Chemical Society,
105, 2142±2147.
AltemuÈller, H.-J. 1992. Visuelle Erscheinungs- und Zustandsformen
der organischen Bodensubstanz. Berichte uÈber Landwirtschaft, 206,
30±44.
Bailly, F. 1972. Zum Problem der Schwarzerde-Bildung in
Norddeutschland. Mitteilungen der Deutschen Bodenkundlichen
Gesellschaft, 15, 3±11.
Ertel, J.R. & Hedges, J.I. 1984. The lignin component of humic
substances: Distribution among soil and sedimentary humic, fulvic,
and base-insoluble fractions. Geochimica et Cosmochimica Acta,
48, 2065±2074.
FAO 1994. FAO-Unesco Soil Map of the World, Revised Legend.
FAO, Rome.
Gehring, A.U., Guggenberger, G., Zech, W. & Luster, J. 1997.
Combined magnetic, spectroscopic and analytical-chemical ap-
proach to infer genetic information for a Vertisol. Soil Science
Society of America Journal, 61, 78±85.
Gehrt, E. 1999. Die aÈolischen Sedimente im Bereich der
noÈrdlichen LoÈûgrenze zwischen Leine und Oker und deren
Ein¯uÈsse auf die Bodenentwicklung. Geologisches Jahrbuch F,
34, in press.
Golchin, A., Clarke, P., Baldock, J.A., Higashi, T., Skjemstad, J.O. &
Oades, J.M. 1997. The effects of vegetation and burning on the
chemical composition of soil organic matter in a volcanic ash soil
as shown by 13C NMR spectroscopy. I. Whole soil and humic acid
fraction. Geoderma, 76, 155±174.
Hatcher, P.G., Schnitzer, M., Vassallo, A.M. & Wilson, M.A.
1989. The chemical structure of highly aromatic humic acids in
three volcanic ash soils as determined by dipolar dephasing
NMR studies. Geochimica et Cosmochimica Acta, 53, 125±130.
Haumaier, L. & Zech, W. 1995. Black carbon-possible source of
highly aromatic components of soil humic acids. Organic
Geochemistry, 23, 191±196.
Heanes, D.L. 1984. Determination of total organic-C in soils by an
improved chromic acid digestion and spectrophotometric proce-
dure. Communications in Soil Science and Plant Analysis, 15,
1191±1213.
KoÈgel-Knabner, I. 1995. Composition of soil organic matter. In:
Methods in Applied Soil Microbiology and Biochemistry (eds K.
Alef & P. Nannipieri), pp. 66±78. Academic Press, London.
Kononova, M.M. 1966. Soil Organic Matter. Pergamon Press,
Oxford.
Kubiena, W.L. 1938. Micropedology. Collegiate Press, Ames, IA.
LuÈdemann, H.-D. & Nimz, H. 1973. 13C-Kernresonanzspektren von
Ligninen, 2. Die Makromolekulare Chemie, 175, 2409±2422.
MuÈckenhausen, E. 1977. Entstehung, Eigenschaften und Systematik
der BoÈden der Bundesrepublik Deutschland, 2nd edn. DLG-Verlag,
Frankfurt-am-Main.
Pawluk, S. 1985. Soil micromorphology and soil fauna: problems and
importance. Quaestiones Entomologicae, 21, 473±496.
Preston, C.M. & Ripmeester, J.A. 1983. 13C-labelling for NMR
studies of soils: CPMAS NMR observation of 13C-acetate
transformation in a mineral soil. Canadian Journal of Soil Science,
63, 495±500.
Rohdenburg, H. & Meyer, B. 1968. Zur Datierung und Boden-
geschichte mitteleuropaÈischer Ober¯aÈchenboÈden (Schwarzerde,
Parabraunerde, Kalksteinbraunlehm): SpaÈtglazial oder HolozaÈn?
GoÈttinger Bodenkundliche Berichte, 6, 127±212.
Sarkanen, K.V. & Ludwig, C.H. 1971. Lignins. John Wiley & Sons,
New York.
Schimming, C.-G., Blume, H.-P., Beyer, L., Dultz, S. & Walch, D.
1993. Exkursion A, Landschaften und BoÈden Ostholsteins.
Mitteilungen der Deutschen Bodenkundlichen Gesellschaft, 70,
47±87.
Schlichting, E. 1953. Die BoÈden der Insel Fehmarn. Schriften des
Naturwissenschftlichen Vereins fuÈr Schleswig-Holstein, 26, 176±
187.
Schmidt, M.W.I., Knicker, H., Hatcher, P.G. & KoÈgel-Knabner, I.
1996. Impact of brown coal dust on the organic matter in
particle-size fractions of a Mollisol. Organic Geochemistry, 25,
29±39.
Schmitz, H. 1955. Die Waldgeschichte Ostholsteins und der zeitliche
Verlauf der postglazialen Transgression an der ostholsteinischen
OstseekuÈste. Berichte der Deutschen Botanischen Gesellschaft, 66,
151±166.
Schnitzer, M. 1992. Bedeutung der organischen Bodensubstanz fuÈr die
Bodenbildung, Transportprozesse in BoÈden und die Bodenstruktur.
Berichte uÈber Landwirtschaft, 206, 63±81.
Schulze, D.G., Nagel, J.L., Van Scoyoc, G.E., Henderson, T.L.,
Baumgardner, M.F. & Stott, D.E. 1993. Signi®cance of organic
matter in determining soil colors. In: Soil Color (eds J.M. Bingham
& E.J. Ciolcosz), pp. 71±90. Special Publication No 31, Soil
Science Society of America, Madison, WI.
Skjemstad, J.O., Clark, P., Golchin, A. & Oades, J.M. 1997.
Characterization of soil organic matter by solid state 13C NMR
spectroscopy. In: Driven by Nature: Plant Litter Quality and
Decomposition (eds G. Cadisch & K. E. Giller), pp. 253±271. CAB
International, Wallingford.
L
364 M. W. I. Schmidt et al.
# 1999 Blackwell Science Ltd, European Journal of Soil Science, 50, 351±365
Skjemstad, J.O., Clarke, P., Taylor, J.A., Oades, J.M. & McClure,
S.G. 1996. The chemistry and nature of protected carbon in soil.
Australian Journal of Soil Research, 34, 251±271.
Skjemstad, J.O., Clarke, P., Taylor, J.A., Oades, J.M. & Newman,
R.H. 1994. The removal of magnetic materials from surface soils.
A solid state 13C CP/MAS n.m.r. study. Australian Journal of Soil
Research, 32, 1215±1229.
Skjemstad, J.O., Janik, L.J., Head, M.J. & MacClure, S.G. 1993. High
energy ultraviolet photo-oxidation: a novel technique for studying
physically protected organic matter in clay- and silt-sized
aggregates. Journal of Soil Science, 44, 485±499.
Snape, C.E., Axelson, D.E., Botto, R.E., Delpuech, J.J., Tekely, P.,
Gerstein, B.C. et al. 1989. Quantitative reliability of aromaticity
and related measurements on coals by 13C n.m.r. ± A debate. Fuel,
68, 547±560.
Wein, R.W. 1993. Historical biogeography of ®re: circumpolar taiga.
In: Fire in the Environment: the Ecological, Atmospheric, and
Climatic Importance of Vegetation Fires (eds P.J. Crutzen & J.G.
Goldammer), pp. 267±276. John Wiley & Sons, New York.
Wilson, M.A. 1987. NMR Techniques and Applications in
Geochemistry and Soil Chemistry. Pergamon Press, Oxford.
Wolff, W. 1930. Die Bodenbildung Schleswig-Holstein und ihr
VerhaÈltnis zu den geologischen Bodenarten. Jahrbuch der
Preuûischen Geologischen Landesanstalt Berlin, 51, 141±178.
Yarilova, E.A. 1972. Comparative characteristic of fabric components
in some chernozems of the U.S.S.R. In: Third International
Working-Meeting on Soil Micromorphology Wroclaw, 1969,
Zeszyty Problemowe Postepow Nauk Rolniczych No 123, pp.
357±369.
Zech, W., Guggenberger, G., Zalba, P. & Peinemann, N. 1997. Soil
organic matter transformation in Argentinian Hapludolls.
Zeitschrift fuÈr P¯anzenernaÈhrung und Bodenkunde, 160, 563±571.
R
Charred organic carbon in German chernozemic soils 365
# 1999 Blackwell Science Ltd, European Journal of Soil Science, 50, 351±365