Refractory organic carbon in C-depleted arable soils, as
studied by
13
C NMR spectroscopy and carbohydrate analysis
Rita Kiem
a,*, Heike Knicker
a, Martin KoÈrschens
b, Ingrid KoÈgel-Knabner
aaChair of Soil Science, Technische UniversitaÈt MuÈnchen, 85350 Freising-Weihenstephan, Germany
bDepartment of Soil Research, UFZ-Umweltforschungszentrum Leipzig-Halle GmbH, 062460 Bad LauchstaÈdt, Germany
Abstract
Soil organic matter (SOM) comprises refractory compounds, to which a turnover time of more than 1000 years has been attributed in SOM models. The goal of this study is to characterize the chemical structure of refractory com-pounds of organic carbon in arable soils by means of13C NMR spectroscopy and analysis of carbohydrates. C-deple-ted soils that are expecC-deple-ted to be enriched in refractory compounds are compared with fertilized soils from long-term agroecosystem experiments. In the C-depleted soils, lower proportions ofO/N-alkyl C and higher proportions of aro-matic and carboxyl C compared with the fertilized counterparts are observed. Ratios of alkyl toO-alkyl C are higher in the depleted soils than in the fertilized ones. Along with the overall C-depletion, the absolute amount of all carbon species was reduced. This net decrease is highest for theO/N-alkyl C and smallest for the aromatic C. Yields of wet chemically determined carbohydrates positively correlate with the relative intensities ofO-alkyl C in the NMR spectra, and con®rm the net decrease ofO-alkyl C compounds along with C-depletion. The refractory organic carbon pool in arable soils appears to have a lower contribution ofO/N-alkyl C, and a higher proportion of recalcitrant aromatic structures compared with more labile fractions of organic carbon.#2000 Elsevier Science Ltd. All rights reserved.
Keywords:Soil organic carbon; Refractory/passive pool; Long-term agroecosystem experiments;13C CPMAS NMR spectroscopy;
Carbohydrates; Hydrolysis residue
1. Introduction
SOM is a heterogeneous mixture of compounds, which represent a continuum from fresh plant residues to strongly humi®ed material (KoÈgel-Knabner, 1993). The dierent components in this continuum dier with respect to their turnover time in soil. On the basis of these distinct turnover times, in SOM-models, the total SOM pool is divided into conceptual fractions (Van Veen and Paul, 1981; Parton et al., 1987; Cambardella, 1998). Essentially, SOM in these models has been considered to consist of a compartment ofactive components with a residence time (or turnover time) of years to decades, and a passive (or refractory) fraction remaining in soil for hundreds to thousands of years. In some model approa-ches, ``active'' components are further distinguished into
a labile fraction of plant litter and microbial biomass, and anintermediatefraction of particulate organic matter which is stabilized for a few decades. Factors that deter-mine stability of organic carbon in soil have been sum-marized as chemical recalcitrance of organic molecules against microbial attack,interactionsbetween organic and mineral compounds, and accessibility of organics to microbes and enzymes (Sollins et al., 1996). Among the chemically most resistant compounds are aromatic and paranic structures (Oades, 1995). With regard to inter-actions and accessibility, the stabilization of organic car-bon in soils was suggested to be in¯uenced by the clay content, and those factors that control the aggregation status (Oades, 1995).
Long-term agroecosystem studies provide a means to investigate the eect of management practices on C sequestration in soils. In this study we investigated the chemical structure of at least two contrasting treatments from long-term agroecosystem experiments, comparing the structure of soil organic carbon (SOC) offertilized
plots with that of treatmentsdepletedin organic carbon,
0146-6380/00/$ - see front matter#2000 Elsevier Science Ltd. All rights reserved.
P I I : S 0 1 4 6 - 6 3 8 0 ( 0 0 ) 0 0 0 4 7 - 4
Organic Geochemistry 31 (2000) 655±668
www.elsevier.nl/locate/orggeochem
* Corresponding author: Tel. 8161-713147; fax +49-8161-714466.
i.e. unmanured plots and bare fallows (Fig. 1). In agroe-cosystem experiments usually each treatment is estab-lished on several replicate plots, which are randomly arranged within the experimental ®eld. Due to this design, environmental factors such as changes in geolo-gical substrate, soil texture, or atmospheric input of e.g. mineral matter should randomly aect the various treatments. Mean dierences in SOM observable between the treatments should, therefore, be attributable to the type of soil management.
Fertilized plots receive organic and mineral fertilizers in order to guarantee high crop yields. Thus, high amounts of organic materials enter the soil, as crop residues and farmyard manure. In the plots where no inorganic or organic nutrients are added, crop produc-tion is reduced compared with the fertilized plots. Con-sequently, in the unmanured plots the total input of organic matter into the soil is lower than in the fertilized counterparts. Bare fallows are kept free of any vegeta-tion cover. Thus, these soils do not receive any (or a negligible quantity of) organic matter input. In the fal-lows investigated in this study, weed plants are removed manually avoiding the use of herbicides. The pro-nounced dierences in organic input Ð at the long term Ð result in a dierentiation in the total SOC level of the plots (Fig. 1). Furthermore, the labile and inter-mediate fractions of SOM are expected to be strongly aected by the type of soil management and the amounts of organic input, respectively, whereas the passive/ refractory pool should remain unaected at a time scale of decades (Elliott et al., 1996). Compounds with a residence time of years to decades would be depleted in the plots with low/missing organic input, and as a con-sequence the passive pool should make up a higher proportion of total SOC compared with the fertilized treatments (Fig. 1). The relative accumulation of
refractory compounds in the C-depleted soils is the basic assumption of our experimental approach.
The objective of this study is to assess the chemical composition of refractory organic carbon in arable soils, by comparing C-depleted treatments with fertilized soils, which dier in the relative proportion of refractory com-pounds. This includes the study of the gross chemical composition by means of13C nuclear magnetic resonance (NMR) spectroscopy, and the analysis of carbohydrates as a speci®c compound class, which is quantitatively of great importance for SOM (Lowe, 1978).
2. Materials and methods
2.1. Study sites and soil sampling
We have selected soils from eight European long-term agroecosystem experiments (Table 1). The study sites are located in dierent regions of Central and Eastern Europe, covering a range of climatic features, of geolo-gical substrates and of soil types. At each experimental site, samples were obtained from dierent treatments: plots with the combined addition of mineral and organic fertilization, and unmanured plots with the same crop rotation as in the fertilized treatments but without fer-tilization. In the experiments of Bad LauchstaÈdt and Prague bare fallows were established. The two contrast-ing treatments are denoted as fertilized plots and C
-depletedplots (unmanured soils and bare fallows) (see Fig. 1).
Except for the bare fallows in Prague and Bad LauchstaÈdt, all types of treatments are replicated at least three-fold on separate plots arranged in a randomized way within the experimental ®elds (3±5 replications depending on the experiment). The bare fallow plots of Prague and Experiment A (Bad LauchstaÈdt) are 2±3 m2 in size and have no replications. The bare fallow treatment of Experiment B (Bad LauchstaÈdt) is replicated twice.
Soil samples were collected from a depth 0±20 cm. Soil material collected at 10 sampling points on each of the replication plots of a certain treatment was mixed. Subsequently, samples were air-dried and components >2 mm were removed by dry sieving. For elemental analysis and hydrolysis of carbohydrates, aliquots of the samples were ground using a ball mill.
2.2. Elemental analysis and pH
C and N contents were determined by dry combustion using a Elementar Vario EL Analyzer. Inorganic and organic carbon was dierentiated by determining the amount of carbon before (total carbon) and after igni-tion of the samples at 550C for 3 h (inorganic carbon). Inorganic carbon was not detected in any sample; thus, the total carbon amount is referred to as organic carbon.
Table 1
Site characteristics of the long-term agroecosystem experiments and treatments considered in the present study
Location Start of
1937 Albic Luvisol 2.7 8.6 520 Fertilized NPK+lime+farmyard manure
(15 t haÿ1yearÿ1)
Schnieder (1990)
Unmanured No fertilization
Groû Kreutz (Germany)
1967 Albic Luvisol 4.0 8.9 537 Fertilized N+farmyard manure
(3 t haÿ1yearÿ1)
Asmus (1990)
Unmanured No fertilization
Skierniewice (Poland)
1923 Luvisol 6.0 7.9 527 Fertilized NPK+lime+farmyard manure
(6 t haÿ1yearÿ1)
Mercik et al. (1997)
Unmanured No fertilization
Puch (Germany)
1983a Orthic Luvisol 18 7.9 927 Fertilized NPK+lime+farmyard manure
(10 t haÿ1yearÿ1)
Krauss et al. (1997) Diez et al. (1997)
1953b Bare fallow Without crops and fertilization
Lauterbach 1966 Dystric Cambisol 18 6.3 900 Fertilized NPK+lime+farmyard manure Reichelt (1990)
(Germany) Unmanured No fertilization (but liming)
Bad LauchstaÈdt (Germany)
Haplic Chernozem 23 8.6 490
Experiment A 1902c Fertilized NPK+farmyard manure
(15 t haÿ1yearÿ1)
KoÈrschens and Eich (1990)
1956d Unmanured No fertilization
Bare fallow Without crops and fertilization
Experiment B 1984 Fertilized Farmyard manure
(200 t haÿ1yearÿ1)
KoÈrschens et al. (1998)
Bare fallow Without crops and fertilization
Prague (Cech Republic)
1958 Luvi-haplic
Chernozem
29 8.1 450 Bare fallow+
fertilization
Farmyard manure
(80 t haÿ1yearÿ1)
KubaÂt and NovaÂk (1992)
Bare fallow (1) Without crops and fertilization
Bare fallow with soil tillage (0-20 cm) (2)
a Fertilized treatment since 1983.
b Bare fallow since 1953.
c Fertilized and unmanured plots since 1902.
d Bare fallow since 1956.
The pH values were measured in the supernatant of a soil suspension in 0.01 M CaCl2by using a glass electrode. The ratio of soil material (w) to CaCl2solution (v) was 1:2.5 (Schlichting et al., 1995).
2.3. 13C CPMAS NMR spectroscopy
Sieved samples were treated with 10% hydro¯uoric acid (HF) to remove paramagnetic species and mineral matter, resulting in a concentration of organic material (Schmidt et al., 1997). Fifty millilitres of 10% (v/v) HF were added to 10 g of soil in a polyethylene bottle. The suspension was shaken horizontally for approximately 12 h. After centrifugation, the supernatant was removed. The HF treatment was repeated four times. Finally, the residue was washed with deionzed water and freeze dried.
The solid-state 13C nuclear magnetic resonance (NMR) spectra were obtained on a Bruker DSX 200 spectrometer operating at a13C resonance frequency of 50.3 MHz by using the cross polarization magic-angle spinning (CP-MAS) technique (Schaefer and Stejskal, 1976). Samples were packed into a rotor of zirconium dioxide with a diameter of 7 mm, and spun at a frequency of 6.8 kHz. A pulse delay of 400 ms and a contact time of 1 ms was used. Due to low sensitivity of the sandy soil samples from Thyrow, Groû Kreutz and Skiernie-wice, the spectra of these samples were obtained by using a pulse delay of 250 ms. A ramped 1H-pulse was used during contact time in order to circumvent inexact Hartmann-Hahn conditions (Peersen et al., 1993). After accumulation of 23,000±350,000 scans and prior to Fourier transformation, a line broadening of 100±150 Hz was applied. The chemical shift scale is referenced to tet-ramethylsilane (= 0 ppm). The spectra are divided into four major chemical shift regions, assignable to alkyl C (0±45 ppm), O/N-alkyl C (45±110 ppm), aromatic C (110±160 ppm) and carboxyl/carbonyl C (160±220 ppm). A detailed scheme of most tentative assignment of chemical shift regions to dierent carbon types is given in Table 2. Signal intensities for aromatic and carboxyl carbon were corrected for spinning side bands, adding the intensities of the ranges 276±220 ppm and 0
toÿ50 ppm to those of the aromatic carbon region. One side band of the carboxyl carbon is found in the range 323± 276 ppm. Assuming that the second side band for carboxyl C, between 0 and 45 ppm, is of equal size, the integral of the ®rst side band was doubled and added to the signal 160± 220 ppm; then, the intensity of the ®rst side band (326±276 ppm) was subtracted from the alkyl C region.
In the depleted soils, to quantify the decrease of C in the various shift ranges, the proportions of the individual shift ranges were normalized to the OC of the fertilized soils:
Carbon species % of OCÿ Depleted
OCDepleted
OCFertilized
% of OCFertilized
1
OCDepleted Organic Carbon of the depleted plot
(g kgÿ1)
OCFertilized Organic Carbon of the fertilized plot (g kgÿ1)
2.4. Carbohydrate analysis
Analysis of carbohydrates was carried out according to KoÈgel-Knabner (1995). The method includes acid hydrolysis of carbohydrates followed by a colorimetric determination of sugar monomers by the MBTH (3-methyl-2-benzothiazolinone hydrazone hydrochloride) procedure. In this procedure, monosaccharides are reduced to alditols, followed by an oxidation of the terminal glycol (±CH2OH) groups of the alditols yield-ing two moles of formaldehyde per mole of original monosaccharide. The formaldehyde concentration is determined photometrically at 635 nm after reaction with MBTH (Pakulski and Benner, 1992).
For the hydrolysis of non-cellulosic carbohydrates, soil samples were incubated with 1 M HCl at 105C for 5 h. For the hydrolysis of total carbohydrates (cellulosic and non-cellulosic), the samples were incubated with 12 M H2SO4 at room temperature for 16 h, and subse-quently with 1 M H2SO4at 105C for 5 h. The cellulosic fraction of carbohydrates was determined by calculating the dierence of monosaccharides released by H2SO4
Table 2
Tentative assignment of signals in the13C NMR spectra (from Almendros et al., 1992; Knicker and LuÈdemann, 1995)
Chemical shift range (ppm)
Assignment
0±45 Terminal CH3groups (0±25); CH2groups in chains (30) (lipids, proteins)
45±60 OCH3in aromatic structures (lignin) and in polysaccharides (hemicelluloses);a-amino C (amino acids)
C-6 of some polysaccharides
60±90 Higher alcohols, C-2 to C-5 of hexoses;a-,b-,g-C inb-O-4 linked units (lignin)
90±110 (103±105) C-1 in polymeric carbohydrates (anomeric C) and C-2 and C-6 in syringyl units (lignin)
110±140 Protonated and C-substituted aromatics; ole®nic carbons
140±160 Aromatic COR and CNR groups
(total carbohydrates) and by HCl hydrolysis (non-cellu-losic carbohydrates). Yields of monosaccharides were expressed relative to a calibration curve of glucose. The amount of glucose-Cwas calculated by multiplying the total glucose mass with 0.4. For correction of mass dif-ference between glucose monomers and polysaccharide structures, dominating in SOM, the yields were multi-plied with a factor of 0.9.
All analyses were run in triplicate. The standard deviation relative to the mean value ranged between 2 and 25% for the non-cellulosic carbohydrates, and between 3 and 23% for the measurement of total carbohydrates, respectively.
To assess the net decrease of carbohydrate C in the depleted plots, the content of total carbohydrate C was normalized to the OC of the fertilized plots:
Carbohydrate C % of OCÿ Depleted
OCDepleted
OCFertilized
% of OCFertilized
2
OCDepleted Organic Carbon of the depleted plot
(g kgÿ1)
OCFertilized Organic Carbon of the fertilized plot (g kgÿ1)
3. Results and discussion
3.1. Carbon and nitrogen content, pH values
The contrasting management practices lead to a con-siderable dierence in the content of organic carbon and nitrogen (Table 3). Unmanured plots and bare fallows contain between 40 and 67% of the organic carbon and
between 41 and 67% of the nitrogen of the respective ferti-lized plots. Lacking any input of organic material, the bare fallow of Experiment A (Bad LauchstaÈdt) is more depleted in C and N than the corresponding unmanured soil. As the relative decrease is similar for organic carbon and total nitrogen, in most of the experiments the C/N-ratio of the depleted soils is not subject to change compared with the fertilized plots. Only in 4 experiments, 3 of which are bare fallows, C-depleted soils have slightly higher C/N ratios than the fertilized counterparts (Table 3).
In Prague, soil tillage of the bare fallow results in a fur-ther loss of organic matter compared with the untilled fallow. This may be explained by the eect of tillage on soil structure. Tillage leads to the breakdown of aggre-gates of various size. Upon disintegration of aggreaggre-gates, OM previously protected within the aggregate (``physical protection'') is assumed to become exposed to micro-organisms and degradative enzymes (Christensen, 1996).
In Thyrow, Skierniewice and Puch the lack of lime application in the depleted plots Ð in contrast to the fertilized ones Ð resulted in lower pH values (Table 3). The chernozemic soil of Bad LauchstaÈdt is not given any lime at all, and a similar pH is maintained in the dierently managed plots of this soil. In Groû Kreutz and Lauterbach the fertilized plots show a lower pH value than the depleted, unfertilized soils. One possible explanation of this decrease in pH may lie in the nitri®-cation of ammonium-N present in the organic and inorganic fertilizers, respectively (Paul and Clark, 1989).
3.2. Chemical structure of SOC (13C NMR)
13C nuclear magnetic resonance (NMR) spectra of the
soils are given in Figs. 2 and 3, relative signal intensities
Table 3
Content of organic carbon and total nitrogen, C/N ratios and pH of dierent treatments of long-term agroecosystem experiments
Organic carbon (g kgÿ1) Total nitrogen (g kgÿ1) C/N ratio pH
Skierniewice 8.8 4.4 51 0.76 0.39 51 11 11 6.1 4.5
Puch 12.0 7.0 58 1.44 0.91 63 8 8 6.9 5.2
Lauterbach 48.3 30.1 62 4.39 2.74 62 11 11 6.0 6.6
Bad LauchstaÈdt Experiment A
Unmanured plot 24.0 16.0 67 1.97 1.32 67 12 12 7.2 7.5
Bare fallow 14.9 62 1.12 57 13 n.d.a
Experiment B 41.3 19.7 48 3.68 1.64 44 11 12 7.1 7.1
Prague Bare fallow 29.1 14.5 50 2.62 1.22 47 11 12 7.1 6.4
Bare fallow+tillage 12.8 44 1.07 41 12 6.9
an.d., not determined.
are presented in Table 4. In the spectra, the resonance line discernable around 30 ppm, in the region of alkyl carbon, can probably be assigned to methylene structures. The peak around 56 ppm (in the region ofO- andN -sub-stituted alkyl carbon, 45±110 ppm) may originate from methoxyl groups and N-substituted alkyl carbon. The peaks around 72 and 104 ppm are most probably assigned to carbohydrates. The signal between 110 and 140 ppm peaking at 130 ppm may originate from protonated and C-substituted aryl carbon, as well as from unsaturated alkyl structures. The signal around 175 ppm is attribu-table to carboxyl/amide functional groups.
Dierences in the relative intensity distribution between the spectra of the depleted and the fertilized plots are most apparent for the O/N-alkyl C and the aromatic C, respectively (Table 4). Compared with the fertilized plots, low organic input lead to a relative
decrease of the proportion of O/N-alkyl C, and to a relative enrichment of aromatic C. Except for Puch and Lauterbach, the percentage of aromaticity Ð given as the ratio of aromatic C to the sum of aromatic and total aliphatic C in Table 4 Ð increases by a factor of 1.2 on an average. Aromaticity is exceptionally high in the bare fallow (A) of Bad LauchstaÈdt and the unmanured plot of Thyrow. By comparison, in the majority of the soils of this study, the O/N-alkyl C region is the quantita-tively most important one, as it is reported for soils from dierent regions of the world (e.g. Baldock et al., 1992; Haider, 1992; Guggenberger et al., 1995).
As seen in this study, in long-term experiments at Rothamsted Experimental Station the relative contribu-tion ofO-alkyl C diminished along with reduced organic input (Kinchesh et al., 1995). In a wide range of soils, the O-alkyl C range has been found to be negatively correlated with the aromatic C region (Mahieu et al., 1999). The changes in relative signal distribution found
Fig. 2. 13C CPMAS NMR spectra of soils from three
experi-mental sites diering in treatment.
Fig. 3. 13C CPMAS NMR spectra of soils from two
in our experiments, the concomitant decrease ofO-alkyl C and increase of aromatic C, are in agreement with these correlations.
The total range of aromatic carbon can be further divided into a region ofO-aryl C (140±160 ppm) and of aryl C (110±140 ppm). A prominent signal in theO-aryl C region with a peak around 150 ppm, mainly assignable
to lignin units (LuÈdemann and Nimz, 1974), is usually found in plant materials in ®rst stages of decomposition (KoÈgel et al., 1988). In our soils the contribution ofO -aryl C (140±160 ppm) is generally low, ranging from 6 to 10% of total signal intensity (Fig. 4), indicating that lignin does not constitute a major component of SOC. The aromatic region of most of our soils is characterized
Table 4a
Relative contributions of carbon species to the total signal intensity in13C NMR spectroscopy of soil samples under dierent
treat-ments in long-term agroecosystem experitreat-ments
Experimental site
% Alkyl C (0±45 ppm)
%O/N-Alkyl C
(45±110 ppm)
% Aromatic C (110±160 ppm)
% Carboxyl C (160±220 ppm)
Alkyl C
O/NÿAlkyl C % Aromaticityc
Fert.a 0b Fert. 0 Fert. 0 Fert. 0 Fert. 0 Fert. 0
Thyrow 19 17 35 28 30 36 15 17 0.55 0.62 36 44
Groû Kreutz 22 20 41 38 25 28 11 14 0.55 0.53 28 33
Skierniewice 21 22 38 36 26 30 15 11 0.55 0.66 31 34
Puch 19 22 43 41 27 25 11 12 0.43 0.54 30 28
Lauterbach 22 25 43 43 22 20 13 12 0.51 0.57 25 22
Bad LauchstaÈdt Experiment A
Unmanured plot 27 20 33 33 26 33 14 13 0.81 0.60 30 39
Bare fallow 20 28 34 17 0.72 41
Experiment B 21 19 39 35 28 32 13 14 0.53 0.55 32 37
Prague Bare fallow 20 19 41 35 26 30 13 15 0.50 0.54 30 36
Bare fallow+tillage 18 35 34 12 0.51 39
a Fertilized plots.
b Depleted plots.
c AromaticC 110ÿ160
AromaticC 110ÿ160 AliphaticC 0ÿ110100 (FruÈnd et al., 1994).
Table 4b
Absolute amounts of the carbon species (13C NMR spectroscopy) under dierent treatments in long-term agroecosystem experiments
Experimental site
Alkyl C (0-45 ppm)
(g C kgÿ1)
O/N-Alkyl C
(45±110 ppm)
(g C kgÿ1)
Aromatic C (110±160 ppm)
(g C kgÿ1)
Carboxyl C (160±220 ppm)
(g C kgÿ1)
Fert.a 0b Fert. 0 Fert. 0 Fert. 0
Thyrow 1.33 0.57 2.42 0.92 2.09 1.19 1.00 0.57
Groû Kreutz 2.31 0.83 4.24 1.57 2.59 1.16 1.18 0.59
Skierniewice 1.82 0.97 3.33 1.62 2.32 1.35 1.28 0.49
Puch 2.24 1.58 5.16 2.91 3.22 1.74 3.22 1.38
Lauterbach 10.52 7.39 20.76 12.96 10.57 5.86 6.32 3.49
Bad LauchstaÈdt Experiment A
Unmanured plot 6.42 3.18 7.93 5.28 6.28 5.36 3.31 2.14
Bare fallow 3.04 4.25 5.00 2.56
Experiment B 8.50 3.79 16.01 6.84 11.51 6.31 5.20 2.73
Prague Bare fallow 5.94 2.77 11.90 5.14 7.48 4.41 3.81 2.19
Bare fallow+tillage 2.29 4.54 4.37 1.59
a Fertilized plots.
b Depleted plots.
by a broad signal in the aryl C range (110±140 ppm) with a peak around 130 ppm (Figs. 2 and 3). The increase of aromaticity found in the C-depleted soils (except for Puch and Lauterbach) is mainly attributable to an increase of this aryl C signal (Fig. 4). Sp2-hybridized carbon of various origin contribute to the signal around 130 ppm, for example carbon in condensed aromatic rings (Skjemstad and Dalal, 1987). Possible sources of condensed ring structures in soil are charred plant resi-dues (charred organic carbon) and coal material, both giving a pronounced signal around 130 ppm in13C NMR spectra (Haumaier and Zech, 1995; Skjemstad et al., 1996; Rumpel et al., 1998; Schmidt et al., 1999). Charred organic carbon was identi®ed in chernozemic soils of Australia and Germany (Skjemstad et al., 1996; Schmidt et al., 1999). In these studies it was suggested that the charred material may originate from natural vegetation ®res or from human use of the ®re, e.g. for clearing of forests in Central Europe. There may possibly be a contribution of charred organic carbon in some of the soils investigated in this study. The evalution of this hypothesis by direct measurement of charred material should be a subject of future research. A further source of condensed aromatic rings present in SOM is given by the atmospheric deposition of dust and coal-like particles, emitted by the coal processing industry
(Schmidt et al., 1996). As some of the soils are loca-ted in the surroundings of industrialized areas, the possibility of dust emissions into the soils should also be considered.
Except for Puch and Lauterbach, in the depleted soils a slightly lower intensity of alkyl C is observed com-pared with the fertilized plots (Table 4). In the soils of Puch and Lauterbach, the signal intensity of alkyl C is higher in the depleted than in the fertilized soils. The pro-portion of carboxyl C either increases slightly or remains unchanged in the depleted soils (Table 4). Knicker (1993) showed that the variation associated with phase- and baseline correction of Fourier-transformed spectra is highest for the shift ranges 160±220, 45±60 and 0±45 ppm. The relative standard deviation of the signal intensity attributed to these regions was up to 13%. The relative standard deviation for intensities in the ranges 60±110 and 110±160 ppm was found to be up to 6.5%. Consequently, the dierences in relative signal intensity observed between the treatments of this study can be considered to be more sound for theO/N-alkyl C and aromatic C region than for alkyl and carboxyl C.
According to Baldock et al. (1997), the ratio of alkyl toO-alkyl C can be taken as an indicator for assessing the degree of decomposition of organic materials. In six of the experiments, the ratio of alkyl to O/N-alkyl
Fig. 4. Relative distribution of signal intensity of the aryl C region (110±140 ppm) and theO-aryl C region (140±160 ppm) in fertilized
carbon is higher in the depleted than in the fertilized soils (Table 4). A higher proportion of alkyl C relative to O-alkyl C would be related to a higher degree of decomposition of the residual organic matter in the depleted plots compared with the fertilized ones. An increase of alkyl C relative toO-alkyl C was also noted in cultivated soils in comparison with the corresponding native sites (Oades et al., 1988; Preston et al., 1994).
3.3. Depletion of C associated with dierent C-species
Table 4b shows that all carbon forms are diminished in terms of absolute carbon amount, indicating that all carbon species were aected by degradation processes.
Applying Eq. (1), the carbon amounts of the dierent species are normalized to the OC content of the ferti-lized plots. In Fig. 5, the percentages of carbon nor-malized to the fertilized plots are given for the dierent C-species. Ratios between the percentages of depleted and fertilized treatments are calculated (mean values in Fig. 5). These ratios demonstrate that on an average the extent of carbon decrease follows the orderO/N-alkyl C > alkyl C > carboxyl C > aromatic C. The higher decrease associated with O/N-alkyl C than with aro-matic C compounds is in line with the higher biode-gradability of the ®rst and the higher recalcitrance of the latter, respectively. Setting the total decrease of OC to 100%, the major part of this decrease is accounted for
Fig. 5. Contribution of the four carbon species expressed as percentage of organic carbon of the fertilized plots. The ratio given for each carbon species is the mean value of the 8 experiments.
byO/N-alkyl C (42% of the C loss). Alkyl and aromatic C account for 22% each, and carboxyl C for 14%.
In the bare fallows of Prague, soil tillage lead to a further decrease of all carbon species except the aromatic C, corroborating the refractory nature of aromatic C (Fig. 5). In Experiment A of Bad LauchstaÈdt a gradient in the OC content exists among the 3 plots with dierent management regime, i. e. the fertilized plot, the unma-nured plot, and the bare fallow (see Table 3). In Fig. 6, proportions of the carbon species are normalized to the
OC content of the fertilized plot. The decrease of carbon in the 2 depleted plots is smallest for the aromatic C. When comparing the mostly depleted bare fallow with the unmanured plot, the carbon present in alkyl, aro-matic and carboxyl structures is more resistant against further degradation thanO/N-alkyl carbon compounds. This ®nding again points out the labile nature of O/N-alkyl C.
3.4. Carbohydrates
Total carbohydrate C contents, expressed as a per-centage of OC, are similar in the fertilized and depleted plots of the experiments (Table 5), except for the bare fallow of Experiment A, Bad LauchstaÈdt, which is depleted in carbohydrates compared with the other soils. In both treatments, the carbohydrate pool is dominated by the non-cellulosic fraction (Table 5). But the cellulose proportion of total carbohydrates is slightly smaller in the depleted than in the fertilized soils. This loss of cel-lulose re¯ects the reduced input of plant materials, as cellulose is turned over rapidly in soils (Haider, 1992).
Analogous to the carbon species, the carbohydrate content of the depleted soils was normalized to the OC of the fertilized soils [see Eq. (2)] (Table 5). The depleted plots contain between 39 and 72% of the carbohydrate C present in the fertilized plots (mean value 52%). Comparing the ratios determined for carbohydrate C andO/N-alkyl C (52% or 0.52, and 0.48, respectively), the relative extent of decline in carbohydrate C andO/N -alkyl C are of similar magnitude. The data of Experiment
Fig. 6. Contribution of the four carbon species (13C NMR)
and total carbohydrate C (MBTH method) in Experiment A (Bad LauchstaÈdt), expressed as percentage of organic carbon of the fertilized plot.
Table 5
Amounts of total carbohydrates, proportions of cellulose, and total carbohydrates normalized to OC of the fertilized plots in long-term agroecosystem experiments (% of OC of fertilized plot)
Experimental site Fertilized
Carbo:ÿCDepleted Carbo:ÿCFertilized
(%)
Thyrow 13.9 15.4 11 3 13.9 7.3 52
Groû Kreutz 14.6 15.7 13 n.d. 14.6 6.3 43
Skierniewice 15.5 13.5 14 n.d. 15.5 6.8 44
Puch 15.7 18.5 13 7 15.7 10.8 69
Lauterbach 14.8 13.6 12 n.d. 14.8 8.5 57
Bad LauchstaÈdt Experiment A
Unmanured plot 11.2 12.0 9 4 11.2 8.0 72
Bare fallow 8.2 n.d.a 5.1 46
Experiment B 12.5 10.2 17 13 12.5 4.9 39
Prague Bare fallow 10.5 11.5 n.d. n.d 10.5 5.8 55
Bare fallow+tillage 10.2 n.d. 4.5 43
A (Bad LauchstaÈdt) show that there is an analogous trend for the changes in O/N-alkyl C on the one hand and carbohydrate C on the other hand, with progressive decrease of organic carbon (Fig. 6). In summary, the carbohydrate results con®rm those of 13C NMR regarding the decline ofO/N-alkyl C compounds in the depleted plots.
The contents of carbohydrate C signi®cantly correlate with the signal intensities ofO-alkyl C (60±110 ppm) in the soils (Fig. 7). On an average, carbohydrate C (nor-malized to OC) accounts for 47% of the relative signal intensity in the range 60±110 ppm (O-alkyl C). In the work of KoÈgel-Knabner et al. (1988) a comparable relationship between hydrolyzable carbohydrates and intensity in the shift range of O-alkyl C (50±110 ppm) was found for forest ¯oors. The discrepancy between yields of carbohydrates obtained by hydrolysis and intensities of O-alkyl C in NMR spectra may be explained by several reasons (KoÈgel-Knabner, 1997). First, it may be due to methodological problems asso-ciated with the hydrolysis procedure. As the digestion of polymeric carbohydrates to monosaccharides is a pre-requisite for their determination by the MBTH proce-dure, incomplete breakdown of polysaccharides during hydrolysis leads to an underestimation of colori-metrically determined carbohydrates. As shown by Allard et al. (1997), acid hydrolysis of cell material at high temperatures for several hours lead to the pre-cipitation of a dark-colored residue, which was identi-®ed as melanoidin-like polymer, formed in the presence of both sugars and amino acids. Second, the discrepancy may result from the fact that other alcohol and ether groups present for example in lignin side chains con-tribute to theO-alkyl C signal in13C NMR spectra.
To assess the contribution of both hydrolyzable carbo-hydrates and non-hydrolyzable structures to theO-alkyl C
region,13C NMR spectra were obtained from the resi-dues after H2SO4hydrolysis of four soils (Fig. 8). The spectra show that most of the carbohydrates (with resonances around 72 and 105 ppm) were hydrolyzed. Between 21 and 31% of the signal intensity of the O -alkyl C region of the untreated soils remain in the spec-tra of the hydrolysis residues (Table 6). Apart from the contribution of a small portion of non-hydrolyzable carbohydrates, this residual signal intensity may be derived from lignin side chains, other non-speci®ed alcohol and ether structures, and from overlapping sig-nals of alkyl or aromatic carbon. According to Table 6, between 39 and 51% of the O-alkyl C intensity are identi®ed as carbohydrates by means of hydrolysis and MBTH reaction, leaving between 25 and 40% of the signal intensity as hydrolyzable, but non-identi®ed car-bon. This proportion of hydrolyzable compounds which are not identi®ed by the MBTH method was calculated by dierence, taking the two other ``fractions'' into account. As stated above, the gap between the sum of carbohydrates (MBTH method) plus non-hydrolyzable compounds and the total O-alkyl C signal may be due to: (i) di- or oligomeric carbohydrates in the hydrolyzate which do not form a colored complex during the MBTH procedure, (ii) carbohydrates which are involved in the formation of melanoidin-like polymers, and (iii) mono-saccharides which are lost during the hydrolysis proce-dure (Beudert, 1988).
Fig. 7. Correlation between O-alkyl C signal intensities and
yields of total carbohydrate C in various plots of long-term agroecosystem experiments. The solid line is given by the regression equation.
Fig. 8. 13C CPMAS NMR spectra of soils before and after
hydrolysis with H2SO4.
4. Summary and conclusions
In the experimental approach of this study it was supposed that exhaustive depletion of OC in arable soils leads to a relative accumulation of refractory com-pounds in the total SOM. Comparing C-depleted plots, i. e. unmanured plots and bare fallows, with fertilized plots from long-term agroecosystem experiments, the following trends have been observed.
In comparison with the fertilized plots, the residual SOM in the depleted plots is characterized by: (i) a relative decrease inO/N-alkyl C compounds, (ii) a rela-tive accumulation of aromatic C, especially in aryl C compounds, and (iii) in most cases a slightly higher degree of oxidation as indicated by the relative propor-tion of carboxyl funcpropor-tional groups. The proporpropor-tion of alkyl C either increases or is diminished in the depleted plots. However, alkyl C is found to be enriched relative toO-alkyl C in most of the depleted soils.
Regarding the absolute decline of OC, the amount of all carbon species is diminished in the depleted plots compared with the level present in the fertilized ones. However, the various carbon species dier with respect to the extent of this carbon decrease. The decrease was shown to follow the orderO/N-alkyl C>alkyl C>car-boxyl C> aromatic C. The dierence in the behaviour ofO/N-alkyl C and aromatic C is in line with the degree of biodegradability of these structures. The extent of decrease ofO/N-alkyl carbon is con®rmed by the results obtained from wet chemical analysis of carbohydrates.
Our data demonstrate the value of long-term agroe-cosystem experiments for studying the composition of the slowly turned over SOC pool. According to our results, this pool of organic carbon is relatively depleted in O/N-alkyl C compounds, whereas it is relatively enriched in aromatic carbon in comparison with active/ labile fractions of OC.
Acknowledgements
The work was ®nancially supported by the Deutsche Forschungsgemeinschaft. The authors would like to thank Michael Baumecker from the Experimental Sta-tion at Thyrow (Germany), Dr. Pommer from the Bayerische Landesanstalt fuÈr Bodenkultur und P¯an-zenbau at Freising (Germany), and Prof. Stanislaw Mercik from Warsaw Agricultural University (Poland) for the help in obtaining soil samples from the various experimental sites. We are grateful to Dr. Jaromir KubaÂt from the Research Institute of Crop Production at Prague (Cech Republic) for providing soil samples of the bare fallow experiment.
References
Allard, B., Templier, J., Largeau, C., 1997. Artifactual origin of myobacterial bacteran. Formation of melanoidin-like artifact macromolecular material during the isolation process. Organic Geochemistry 26, 691±703.
Almendros, G., Martinez, A.T., Gonzalez, A.E.,
Gonzales-Vila, F.J., FruÈnd, R., LuÈdemann, H.-D., 1992. CPMAS13C
NMR study of lignin preparations from wheat straw trans-formed by ®ve lignocellulose-degrading fungi. Journal of Agricultural and Food Chemistry 40, 1297±1302.
Asmus, F., 1990. Versuch M4 Groû Kreutz- Wirkung orga-nischer und mineralischer DuÈngung und ihrer Kombination auf P¯anzenertrag und DuÈngung. In: Dauerfeldversuche-UÈbersicht, Entwicklung und Ergebnisse von Feldversuchen mit mehr als 20 Jahren Versuchsdauer, pp. 245±250. Baldock, J.A., Oades, J.M., Nelson, P.N., Skene, T.M., Golchin,
A., Clarke, P., 1997. Assessing the extent of decomposition of
natural organic materials using solid-state13C NMR
spectro-scopy. Australian Journal of Soil Research 35, 1061±1083. Baldock, J.A., Oades, J.M., Waters, A.G., Peng, X., Vassallo,
A.M., Wilson, M.A., 1992. Aspects of the chemical structure Table 6
Distribution of the signal in the chemical shift range 60±110 ppm (13C NMR) in soils from long-term experiments
O-Alkyl C (60±110 ppm)=100%
Soils Hydrolyzable C,
identi®ed as carbohydrate Ca(%)
Hydrolyzable C,
non identi®edb(%)
Non hydrolyzable Cc
(%)
Thyrow 51 25 24
fertilized plot
Bad LauchstaÈdt,Experiment A 47 25 28
fertilized plot
Bad LauchstaÈdt,Experiment B 39 40 21
bare fallow
Bad LauchstaÈdt,Experiment B 44 25 31
fertilized plot
a Total carbohydrate C determined by H
2SO4hydrolysis+MBTH method.
b Calculated values by taking into account the carbohydrate C and the non hydrolyzableO-alkyl C.
of soil organic materials as revealed by solid-state13C NMR
spectroscopy. Biogeochemistry 16, 1±42.
Beudert, G., 1988. Morphologische, naûchemische und 13
C-NMR-spektroskopische Kennzeichnung der organischen Substanz von Waldhumuspro®len nach Dichtefraktionier-ung. PhD Thesis, UniversitaÈt Bayreuth.
Cambardella, C.A., 1998. Experimental veri®cation of simu-lated soil organic matter pools. In: Lal, R., Kimble, J.M., Follett, R.F., Stewart, B.A. (Eds.), Soil Processes and the Carbon Cycle. Advances in Soil Science. CRC Press, Boca Raton, pp. 519±526.
Christensen, B.T., 1996. Carbon in primary and secondary organomineral complexes. In: Carter, M.R., Stewart, B.A. (Eds.), Structure and Organic Matter Storage in Agricultural Soils. Advances in Soil Science. CRC Press, Boca Raton, pp. 97±165.
Diez, T., Beck, T., Brandhuber, R., Capriel, P., Krauss, M.,
1997. VeraÈnderungen der Bodenparameter im
Inter-nationalen Organischen StickstodauerduÈngungsversuch
(IOSDV) Puch nach 12 Versuchsjahren. Archiv fuÈr Acker-und P¯anzenbau Acker-und BodenkAcker-unde 41, 113±121.
Elliott, E.T., Paustian, K., Frey, S.D., 1996. Modeling the measurable or measuring the modelable: a hierarchical approach to isolating meaningful soil organic matter frac-tionations. In Powlson, D.S., Smith, P., Smith, J.U. (Eds.), Evaluation of Soil Organic Matter Models Using Existing Long-term Datasets. NATO ASI Series, vol. I 38. Springer Verlag, Berlin, pp. 161±179.
FruÈnd, R., Haider, K., LuÈdemann, H.-D., 1994. Impact of soil management practices on the organic matter structure Ð
investigations by CPMAS 13C NMR-spectroscopy.
Zeit-schrift fuÈr P¯anzenernaÈhrung und Bodenkunde 157, 29±35. Guggenberger, G., Zech, W., Haumaier, L., Christensen, B.T.,
1995. Land-use eects on the composition of organic matter in particle-size separates of soils: II. CPMAS and solution
13C NMR analysis. European Journal of Soil Science 46,
147±158.
Haider, K., 1992. Problems related to the humi®cation pro-cesses in soils of temperate climates. In: Stotzky, G., Bollag, J.-M. (Eds.), Soil Biochemistry, vol. 7. Marcel Dekker, New York, pp. 55±94.
Haumaier, L., Zech, W., 1995. Black carbon Ð possible source of highly aromatic components of soil humic acids. Organic Geochemistry 23, 191±196.
Kinchesh, P., Powlson, D.S., Randall, E.W., 1995.13C NMR
studies of organic matter in whole soils: II. A case study of some Rothamsted soils. European Journal of Soil Science 46, 139±146.
Knicker, H., 1993. Quantitative15N- und13
C-CPMAS-FestkoÈr-per- und 15N-FluÈssigkeits-NMR-Spektroskopie an
P¯anzen-komposten und natuÈrlichen BoÈden. PhD Thesis, UniversitaÈt Regensburg.
Knicker, H., LuÈdemann, H.-D., 1995. N-15 and C-13 CPMAS and solution NMR studies of N-15 enriched plant material during 600 days of microbial degradation. Organic Geo-chemistry 23, 329±341.
KoÈgel, I., Hemp¯ing, R., Zech, W., Hatcher, P.G., Schulten, H.-R., 1988. Chemical composition of the organic matter in forest soils: 1. Forest litter. Soil Science 146, 124±136. KoÈgel-Knabner, I., 1993. Biodegradation and humi®cation
processes in forest soils. In: Bollag, J.-M., Stotzky, G. (Eds.),
Soil Biochemistry, vol. 8.. Marcel Dekker, New York, pp. 101±135.
KoÈgel-Knabner, I., 1995. Composition of soil organic matter. In: Nannipieri, P., Alef, K. (Eds.), Methods in Applied Soil Microbiology and Biochemistry. Academic Press, New York, pp. 66±78.
KoÈgel-Knabner, I., 1997.13C and15N NMR spectroscopy as a
tool in soil organic matter studies. Geoderma 80, 243±270. KoÈgel-Knabner, I., Zech, W., Hatcher, P.G., 1988. Chemical
composition of the organic matter in forest soils: the humus layer. Zeitschrift fuÈr P¯anzenernaÈhrung und Bodenkunde 151, 331±340.
KoÈrschens, M., Eich, D., 1990. Der Statische Versuch Lauch-staÈdt. In: Dauerfeldversuche- UÈbersicht, Entwicklung und Ergebnisse von Feldversuchen mit mehr als 20 Jahren Ver-suchsdauer, pp. 7±23.
KoÈrschens, M., Weigel, A., Schulz, E., 1998. Turnover of soil organic matter (SOM) and long-term balances Ð tools for evaluating sustainable productivity of soils. Zeitschrift fuÈr P¯anzenernaÈhrung und Bodenkunde 161, 409±424. Krauss, M., Pommer, G., Beck, R., Brandhuber, R., Capriel,
P., 1997. Dauerversuche mit Fruchtfolgen und DuÈngung am
Staatsgut Puch, Oberbayern. Auswirkungen auf
denmerkmale. Archiv fuÈr Acker- und P¯anzenbau und Bo-denkunde 42, 193±199.
KubaÂt, J., NovaÂk, B., 1992. C- und N-Dynamik im Boden unter Schwarzbrache im Modellversuch in Prag-RuzyneÂ. In: Tagungsbericht zum Symposium Dauerfeldversuche und NaÈhrstodynamik. pp. 70±72.
Lowe, L.E., 1978. Carbohydrates in soil. In: Schnitzer, M., Khan, S.U. (Eds.), Soil Organic Matter. Elsevier, Amster-dam, pp. 65±93.
LuÈdemann, H.-D., Nimz, H., 1974.13C-Kernresonanzspektren
von Ligninen, 2: Buchen- und Fichten-BjoÈrkmann-Lignin. Die Makromolekulare Chemie 175, 2409±2422.
Mahieu, N., Powlson, D.S., Randall, E.W., 1999. Statistical analysis of published carbon-13 CPMAS NMR spectra of soil organic matter. Soil Science Society of America Journal 63, 307±319.
Mercik, S., Stepien, W., Gebski, M., 1997. Results of 75 years of continuous fertilization experiments in Skierniewice as a contribution to modi®cation of fertilization. Archiv fuÈr Acker- und P¯anzenbau und Bodenkunde 42, 201±210. Oades, J.M., 1995. An overview of processes aecting the
cycling of organic carbon in soils. In: Zepp, R.G., Sonntag, C. (Eds.), Role of Nonliving Organic Matter in Earth's Car-bon Cycle. John Wiley and Sons, New York, pp. 293±303. Oades, J.M., Waters, A.G., Vassallo, A.M., Wilson, M.A.,
Jones, G.P., 1988. In¯uence of management on the compo-sition of organic matter in a red- brown earth as shown by
13C nuclear magnetic resonance. Australian Journal of Soil
Research 26, 289±299.
Pakulski, J.D., Benner, R., 1992. An improved method for the hydrolysis and MBTH analysis of dissolved and particulate carbohydrates in seawater. Marine Chemistry 40, 143±160. Parton, W.J., Schimel, D.S., Cole, C.V., Ojima, D.S., 1987.
Analysis of factors controlling soil organic matter levels in great plains grasslands. Soil Science Society of America Journal 51, 1173±1179.
Paul, E.A., Clark, F.E., 1989. Soil Microbiology and Bio-chemistry, Academic Press, San Diego.
Peersen, O.B., Wu, X., Kustanovich, I., Smith, S.O., 1993. Variable-amplitude cross-polarization MAS NMR. Journal of Magnetic Resonance 104, 334±339.
Preston, C.M., Newman, R.H., Rother, P., 1994. Using 13C
CPMAS NMR to assess eects of cultivation on the organic matter of particle size fractions in a grassland soil. Soil Sci-ence 157, 26±35.
Reichelt, H., 1990. Dauerversuche auf Berglehm-Braunerde im Erzgebirge. In: Dauerfeldversuche- UÈbersicht, Entwicklung und Ergebnisse von Feldversuchen mit mehr als 20 Jahren Versuchsdauer. pp. 275±318.
Rumpel, C., Knicker, H., KoÈgel-Knabner, I., Skjemstad, J.O., HuÈttl, R.F., 1998. Types and chemical composition of organic matter in reforested lignite-rich mine soils. Geo-derma 86, 123±142.
Schaefer, J., Stejskal, E.O., 1976. Carbon-13 nuclear magnetic resonance of polymers spinning at magic angle. Journal of the American Chemical Society 98, 1031±1032.
Schlichting, E., Blume, H.-P., Stahr, K., 1995.
Bod-enkundliches Praktikum, Blackwell Wissenschafts-Verlag, Berlin.
Schmidt, M.W.I., Knicker, H., Hatcher, P.G., KoÈgel-Knabner, I., 1996. Impact of brown coal dust on the organic matter in parti-cle-size fraction of a Mollisol. Organic Geochemistry 25, 29±39. Schmidt, M.W.I., Knicker, H., Hatcher, P.G., KoÈgel-Knabner,
I., 1997. Improvement of13C and15N CPMAS NMR spectra
of bulk soils, particle size fractions and organic material by treatment with 10% hydro¯uoric acid. European Journal of Soil Science 48, 319±328.
Schmidt, M.W.I., Skjemstad, J.O., Gehrt, E., KoÈgel-Knabner, I., 1999. Charred organic carbon in German chernozemic soils. European Journal of Soil Science 50, 351±365. Schnieder, E., 1990. Die Dauerversuche in Thyrow. In:
Dauer-feldversuche- UÈbersicht, Entwicklung und Ergebnisse von Feldversuchen mit mehr als 20 Jahren Versuchsdauer. pp. 205±229.
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 Science 34, 251± 271.
Skjemstad, J.O., Dalal, R.C., 1987. Spectroscopic and chemical dierences in organic matter of two vertisols subjected to long periods of cultivation. Australian Journal of Soil Research 25, 323±335.
Sollins, P., Homann, P., Caldwell, B.A., 1996. Stabilization and destabilization of soil organic matter: mechanisms and con-trols. Geoderma 74, 65±105.
