SEARCH

SEARCH BY CITATION

Keywords:

  • black carbon;
  • C sink;
  • steppe soils;
  • Mollisols;
  • radiocarbon dating;
  • 13C natural abundance

Abstract

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Materials and Methods
  5. 3. Results and Discussion
  6. 4. Conclusions
  7. Acknowledgments
  8. References
  9. Supporting Information

[1] Black carbon (BC) is the product of incomplete burning processes and a significant component of the passive soil organic carbon (SOC) pool. The role of BC in the global carbon cycle is still unclear. This study aimed to quantify and characterize BC in major grassland ecosystems of the world. Twenty-eight representative soil profiles (mainly Mollisols) were sampled in the Russian Steppe, the U.S. Great Plains, the Argentinian Pampa, the Manchurian Plains in China, and the Chernozem region in central Germany. Black carbon contents were estimated using benzene polycarboxylic acids (BPCA) as a molecular marker, and indications about the origin of the BC were derived from bulk and compound-specific δ13C analyses and radiocarbon dating of bulk soil organic matter (SOM). Our findings suggest that between 5% and 30% of SOC stocks consist of BC. Maximum BC contributions to SOC frequently were found at deeper parts of the A horizon with 14C ages younger than 7000 years BP; that is, incorporation of C as charred particles accompanied ecosystem development since the mid-Holocene. Most of this BC formed from local vegetation, as indicated by a 13C isotope signature similar to that of bulk SOM. At some sites, also nonlocal sources contributed to soil BC, e.g., fossil fuel BC inputs at the German sites. Black carbon stocks were highest in Chernozems and lowest in Kastanozems. The Russian Steppe and Chinese Manchurian sites stored about 3–4 times more BC (around 3 kg m−2) than did the other sites because of thicker A horizons that were rich in BC. On a global scale, we estimate that steppe ecosystems contain between 4 and 17 Pg BC.

1. Introduction

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Materials and Methods
  5. 3. Results and Discussion
  6. 4. Conclusions
  7. Acknowledgments
  8. References
  9. Supporting Information

[2] When vegetation burns, charred black organic matter remains as incomplete combustion residue in soil. Condensed aromatic black carbon (BC) structures found in soils also may originate from fossil fuel combustion or from coal dust [e.g., Andreae, 1991; Laskov et al., 2002; Brodowski et al., 2007]. All these types of BC are of purely terrestrial origin, are prone to fluvial and atmospheric transport [Forbes et al., 2006; Hockaday et al., 2007; Guggenberger et al., 2008], and occur ubiquitously in the environment [Goldberg, 1985; Clark, 1988; Masiello and Druffel, 1998; Schmidt and Noack, 2000]. Estimates of global BC production range from 40 to 600 Tg year−1 [Schmidt and Noack, 2000], which might explain, at least in part, the current discrepancy between bottom-up and top-down estimates of the annual C budget for Europe (135–205 Tg year−1) [Janssens et al., 2003]. As more than 80% of BC produced is deposited in soils [Kuhlbusch and Crutzen, 1995], BC may add significantly to the stable soil organic matter (SOM) pool [Skjemstad et al., 1996; Schmidt et al., 1999; Glaser et al., 2001].

[3] Mollisols are typical soils of native grasslands (steppe, prairie, pampa) and, as such, a globally significant carbon pool. The Mollisols in the USA contain, on average, 11.2 ± 8 kg soil organic C m−3 (SOC) and 5.6 kg inorganic C m−3 (SIC) [Guo et al., 2006]. Russian Chernozems contain 349–462 Mg ha−1 SOC and 91–242 Mg ha−1 SIC [Mikhailova and Post, 2006a, 2006b]. Among the steppe soils, the C-rich Chernozems are known to preserve large amounts of BC (10–45% of total SOC in the mollic horizons [Schmidt et al., 1999; Glaser and Amelung, 2003; Rodionov et al., 2006]. The presence of BC frequently correlates with a diagnostic black color in the A horizon (mollic epipedon) [Spielvogel et al., 2004; Eckmeier et al., 2007]. Both natural and anthropogenic (i.e., for agricultural management) vegetation fires may be involved in Chernozem genesis [Schmidt and Noack, 2000; Glaser and Amelung, 2003; Rodionov et al., 2006; Eckmeier et al., 2007]. Some Chernozems in the vicinity of industrial areas are also heavily contaminated with BC from fossil energy sources [Rethemeyer et al., 2004; Brodowski et al., 2007]. However, Chernozems with low BC contents also have been reported [Kleber et al., 2003]. Hence, the true contribution of BC to SOM pools in these terrestrial ecosystems is still open to speculation [e.g., Czimczik and Masiello, 2007]. It is therefore necessary to extend the number of BC measurements in Chernozems and other steppe soils to a world-wide database that not only includes the surface but also subsoil horizons, and that may allow for differentiating between BC of on- and off-site origins.

[4] Soils with a mollic epipedon like Chernozems are found mainly in the semiarid loess and siltstone plains of temperate grassland ecosystems of the world with a typical steppe climate (mean annual precipitation between 300 and 600 mm) [Krasilnikov and Calderon, 2006]. Such ecosystems occur in Russia, central Europe, in the Argentinean Pampa, in the North-American Prairies, and in the Chinese Manchurian region (Figure 1). The potential vegetation usually is composed of long and short grasses, with a variable contribution of shrubs and trees. Most of the plant species have C3 photosynthesis, though the proportion of C4 species increases as temperature rises [Teeri and Stowe, 1976; Boutton, 1996]. As both types of photosynthetic pathways are associated with different degrees of 13CO2 discrimination, assessing the 13C natural abundance of SOM has been used to identify the origin of SOM within such ecosystems [Boutton, 1996; Boutton et al., 1998]. If extended to specific compounds such as BC, this technique also might help to identify whether charred organic matter is derived from local plant inputs, from other ecosystems with different proportions of C3 and C4 grass, or even from fossil fuels which are uniquely derived from C3 plants [Kuhlbusch and Crutzen, 1995; Clark et al., 2001; Krull et al., 2003].

image

Figure 1. Locality of the study sites in the major Mollisol areas of the world (map redrawn from Scheffer and Schachtschabel [2002, p. 538]).

Download figure to PowerPoint

[5] Several techniques have been developed to quantify BC in soils and sediments, each of which covers a different range of BC structures occurring in the environment [Hammes et al., 2007]. To quantify the total amount and also the properties of BC (the latter to indicate the origin), analysis of benzene polycarboxylic acids (BPCA) has been recommended [Glaser et al., 1998; Brodowski et al., 2005a; Hammes et al., 2007]. Using this method, condensed aromatic BC moieties are oxidized to BPCA; that is, with increased degree of ring condensation a higher number of carboxylic groups dominate the BPCA pattern. Soot may not be oxidized completely [Hammes et al., 2007]. Nevertheless, it has been possible to reconstruct the input of BC from fossil energy sources by increased portions of mellitic acid in the BPCA pattern of the surface soils [Brodowski et al., 2007]. An even better source assignment might be achieved by linking the BPCA patterns to compound-specific δ13C analyses [Glaser and Knorr, 2008].

[6] The objective of this study is to assess the contribution of BC to SOM in the major grassland ecosystems of the world. For this aim we used the sum of benzene polycarboxylic acids (BPCA) produced by hot HNO3 oxidation as molecular markers for the charred organic matter [Glaser et al., 1998]. Radiocarbon dating of bulk SOM and compound-specific δ13C analyses of BC were used to obtain a deeper insight into the time frame of BC formation and its origin, respectively. The analyses were not restricted to surface soils but were extended to different soil depths which are known to exhibit different radiocarbon ages [Scharpenseel et al., 1986; Margolina et al., 1988; Pessenda et al., 1996]. This allowed us to investigate the significance of BC accumulation over the time span of soil development and thereby possibly throughout the Holocene history of these grassland ecosystems.

2. Materials and Methods

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Materials and Methods
  5. 3. Results and Discussion
  6. 4. Conclusions
  7. Acknowledgments
  8. References
  9. Supporting Information

2.1. Soil Samples and Collection Sites

[7] Samples from typical grassland ecosystems of the world (Figure 1 and Table 1) [Soil Survey Staff, 2000] were collected in October 1991 (USA, 6 soil profiles) [Amelung et al., 1997], August 1995 (Russia, 7 soil profiles) [Rodionov et al., 1998, 2006], September 1999 (Germany, 5 soil profiles) [Kahle et al., 2002; Brodowski et al., 2007], March 2000 (Argentina, 5 soil profiles plus one topsoil from an ancient farm) and June 2000 (China, 5 soil profiles) from each genetic soil horizon. The samples were air-dried, passed through a 2-mm sieve and stored air-dry for analysis. According to the Food and Agriculture Organization (FAO) [1988] classification (in brackets classification according to Soil Survey Staff [2000]; see Table 1), the native soils comprised Greyzems (Argiudolls) under deciduous forest, Phaeozems and Chernozems (Hapludolls) under meadow steppe, Chernozems and Kastanozems in the dry steppe (Paleustolls, Argiudolls, and Argiustolls) and semi-desert steppe (Calciustolls). Bulk density values were available for most samples from previous field protocols or local soil mapping; for a comparison of BC stocks across all sites, missing bulk density values were approximated by calculating the average of bulk density of the other sites at respective soil depth which amounted to 1.34 g cm−3 for the lower A, and 1.42 g cm−3 for the B horizons (Data Set S1). All sites were level to gently sloping at < 2%. Basic site parameters are summarized in Table 1. In China, Argentina, and Germany, the soils were used for long-term arable cropping; details regarding cropping and fertilization history are partly uncertain. All other sites were still under native grassland or forest vegetation.

Table 1. Soils Under Study
Site Identification (Longitude; Latitude)NumberSoil Type FAO [1988]/Soil Survey Staff [2000]MAPa (mm)MATb (°C)Site HistorySOC Stockc (kg m−2)Range
BCd (g kg−1 soil)δ13C Soil (‰)
  • a

    MAP, mean annual precipitation.

  • b

    MAT, mean annual temperature.

  • c

    SOC, soil organic carbon, stocks given for whole profile sampled (until lower B or upper C horizon; see Data Set S1).

  • d

    BC, black carbon calculated from BPCA-C content multiplied with a conversion factor of 2.27.

  • e

    Basic soil data reported by Kahle et al. [2002].

  • f

    Major soil formation under former udic moisture regime; sites drier today.

  • g

    Sites not classified as Mollisols and thus not included in Figure 1.

Russia
Tula (37°30′ E; 54°00′ N)R1haplic Greyzem/typic Hapludoll7154.1native deciduous forest28.89.7 to 1.0−25.7 to −27.1
Kursk (36°10′ E; 51°45′ N)R4haplic Phaeozem/typic Hapludoll5735.3native steppe33.75.6 to 1.6−25.0 to −27.4
Kursk (36°10′ E; 51°45′ N)R5haplic Chernozem/typic Hapludoll5735.3native steppe36.15.5 to 2.2−24.9 to −26.2
Kursk (36°10′ E; 51°45′ N)R6haplic Chernozem/typic Hapludoll5735.3native oak forest29.87.2 to 2.7−25.6 to −26.3
Novo-Annenskij (43°45′ E; 50°35′ N)R7calcic Chernozem/calcic Hapludoll4505.5steppe; arable land since 189019.55.1 to 2.9−24.7 to −25.7
Frolovo (44°30′ E; 50°00′ N)R8haplic Kastanozem/typic Paleustoll3006.0steppe; arable land since 189022.85.9 to 3.8−24.9 to −25.8
Dubovka (44°50′ E; 49°10′ N)R9calcic Kastanozem/petrocalcic Paleustoll3007.0steppe; arable land since 18905.52.9 to 1.1−22.2 to −23.4
         
China
Hulan (Harbin) (126°67′ E; 46°13′ N)C1haplic Phaeozem/typic Hapludoll5453.6>150 years plowed47.73.9 to 1.7−21.4 to −23.9
Daan (Fuyu) (124°38′ E; 45°39′ N)C3calcic Chernozem/calcic Hapludoll4234.5>200 years plowed12.11.8 to 0.4−19.5 to −25.2
Wulanhaotu (Baicheng) (122°04′ E; 46°10′ N)C4haplic Chernozem/typic Hapludoll4104.2>150 years plowed13.22.1 to 1.1−20.7 to −22.5
Taonan (Baicheng) (122°12′ E; 45°43′ N)C6calcic Kastanozem/petrocalcic Paleustoll3784.9>250 years plowed32.43.0 to 1.2−20.1 to −22.2
Tuquan (Baicheng) (121°74′ E; 45°48′ N)C7calcic Kastanozem/petrocalcic Paleustoll3884.9>50 years plowed14.05.5 to 1.8−19.1 to −20.5
         
Argentina
Bahia Blanca (62°11′ W; 38°45′ S)A1calcic Kastanozem/petrocalcic Paleustoll58215.6arable land since 18908.21.1 to 0.5−20.3 to −23.5
Tres Arroyos (60°16′ W; 38°23′ S)A2calcic Chernozem/calcic Hapludoll73114.5arable land since 18908.82.1 to 0.4−19.5 to −22.8
Ancient farm soil (60°16′ W; 38°23′ S)A*calcic Chernozem/calcic Hapludoll73114.5park area, never plowed5.15.1n.d.
Balcarce (58°18′ W; 37°46′ S)A3luvic Chernozem/petrocalcic Argiudoll84313.8arable land since 189015.12.5 to 1.1−19.5 to −22.7
Balcarce (58°18′ W; 37°46′ S)A4Brunizem/typic Argiudoll84313.8arable land since 189017.64.6 to 0.6−19.3 to −21.5
Cascalares (60°22′ W; 38°22′ S)A5haplic Chernozem/typic Hapludoll73114.5arable land since 189010.73.4 to 0.5−19.8 to −28.0
         
Germany
Barnstädte (11°38′ E; 51°19′ N)G1ahaplic Phaeozem/typic Haplustoll4848.8>200 years arable land14.54.5 to 0.5−25.2 to −26.5
Bad Lauchstädte (11°59′ E; 51°29′ N)G1bhaplic Chernozem/typic Haplustollf4808.7arable land since 189516.95.2 to 0.6n.d.
Leimbache (11°32′ E; 51°21′ N)G2haplic Phaeozem/typic Haplustollf4808.7>200 years arable land8.84.0 to 0.3−24.9 to −26.5
Leimbach (11°31′ E; 51°21′ N)G3chromic Luvisol/typic Argiustollf4508.7>200 years arable land10.73.6 to 0.5−23.9 to −25.9
Leimbach (11°30′ E; 51°21′ N)G4haplic Luvisol/typic Haplustalfg4848.7>200 years arable land5.92.2 to 0.2−24.9 to −26.7
Halle (11°52′ E; 51°23′ N)G5haplic Phaeozem/typic Haplustoll4909.0arable land since 189510.92.6 to 0.4n.d.
         
United States
Elmont (95°36′ W; 39°04′ N)U1Luvic Phaeozem/typic Argiudoll83912.6pasture17.54.5 to 1.2−14.0 to −16.9
Topsey (97°58′ W; 31°10′ N)U2calcic Kastanozem/typic Calciustoll82920.0rangeland16.14.7 to 0.9−14.5 to −20.5
Hollister (100°13′ W; 33°12′ N)U3Luvic Kastanozem/pachic Paleustoll50717.7rangeland8.22.3 to 0.6−11.7 to −20.4
Blackpipe-4 (103°47′ W; 44°00′ N)U4Luvic Kastanozem/typic Argiustoll4976.8pasture6.22.1 to 0.7−18.6 to −20.8
Zatoville (106°14′ W; 46°16′ N)U5calcic Cambisol/borollic Camborthidg2746.5pasture6.12.4 to 0.4−21.1 to −22.6
McAllen (98°54′ W; 27°57′ N)U6haplic Calcisol/aridic Ustochreptg44623.4rangeland8.12.9 to 1.1−22.5 to −23.2

2.2. Carbon and Nitrogen Analyses

[8] Total C was determined as CO2 after dry combustion using a CHNS analyzer (Elementar Analysensysteme GmbH, Hanau, Germany). Organic C was determined by dry combustion after pretreatment with 0.1 M HCl to remove carbonates. Inorganic C was calculated as the difference between total C and organic C.

2.3. Black Carbon Analysis

[9] Black carbon was analyzed using BPCA as molecular marker [Glaser et al., 1998; Brodowski et al., 2005a]. An aliquot of 0.5 g of ground sample was digested with 10 mL of 4 M trifluoroacetic acid (TFA) for 4 h at 105°C. Subsequently, the dried residue was oxidized with 2 mL of 65% HNO3 for 8 h at 170°C at elevated pressure. The digested solution was diluted to 10 mL with deionized water to reduce the acid concentration. A 2-mL aliquot of the diluted digest was subjected to a cation exchange cleanup procedure using a Dowex 50WX8, 200–400 mesh column. After freeze-drying, the aromatic acids were analyzed as trimethylsilyl (TMS) derivatives by capillary gas chromatography (GC) using HP 6890 instrument equipped with HP-5 capillary column (30 m x 0.32 mm x 0.25 μm film thickness) and a flame ionization detector (FID). Citric acid (1 mg mL−1 in water) was used as internal standard 1, added prior to the sample cleanup; 2.2′-biphenyldicarboxylic acid (1 mg mL−1 methanol, internal standard 2) was added prior to derivatization. The sum of the yields of BPCA after nitric acid oxidation is a relative measure of BC [Glaser et al., 1998]. However, the actual amount of BC is difficult to assess because it exists as a continuum of thermally altered material [Hedges et al., 2000], whereas analytical methods rely on operational definitions with clear-cut boundaries [Schmidt et al., 2001]. Conversion factors have been suggested for BC estimates, and only C but not O and H atoms of BPCA are used for BC estimations. Here we applied a conversion factor of 2.27 to estimate BC contents from BPCA-C concentrations as commercial charcoals yielded, on average, 44% BPCA-C upon nitric acid digestion [Glaser et al., 1998]. Calculated BC amounts therefore correspond to fresh rather than “weathered charcoal” equivalents. As discussed by Brodowski et al. [2005a], different types of BC yielded different conversion factors but all of them exceeded 2.27. The use of this factor provides a conservative minimum estimate of the true BC contents in soil. All analyses were performed in duplicate; the coefficient of variation averaged below 10%. The absolute limit of detection corresponded to approximately 1% BC referred to SOC in the deeper subsoil, beyond the respective absolute amounts only a nonlinear quantification of BPCA can be performed [Brodowski et al., 2005a], which was not done here.

[10] It is currently in dispute whether humification processes may form melanoidins. Poirier et al. [2000] suggested that soils may contain (pseudo-)melanoidins, which, according to Brodowski et al. [2005a], also contain chemical structures similar to BC (13 mg BPCA-C per g melanoidin C, on average). Nevertheless, clear evidence for the existence of (pseudo-) melanoidins in soils and an unambiguous distinction of melanoidins from BC is still lacking. The main BPCA detected so far from artificial melanoidins were tetra- and pentacarboxylic acids [Brodowksi et al., 2005a], whereas the main BPCA found in this study was benzene hexacarboxylic acid (Data Set S1). We conclude that pseudo-melanoidin formation during humification processes does not account for the majority of BC in our samples, and if melanoidins have been generated during vegetation fires, these melanoidin-derived BPCA are directly assigned to BC.

2.4. Bulk and Compound-Specific δ13C Analyses

[11] An aliquot of each soil was treated with 1 M HCl to remove pedogenic and lithogenic carbonates to determine the stable carbon isotope composition (δ13C) of SOM. In addition to the bulk organic C, we determined the natural 13C abundance of the sum of BPCA. For this purpose, 1.0 g of ground sample was digested with 4 M TFA and oxidized with HNO3 as described above. Thereafter the extracts were washed several times with de-ionized water and dried at 120°C. For δ13C analysis, both the bulk organic C and the dried aromatic acids were analyzed by dry combustion on a Carlo Erba CN 2500 analyzer coupled with a Deltaplus continuous-flow isotope ratio mass spectrometer (EA-IRMS; Thermo Finnigan MAT, Bremen, Germany) via a Conflow II interface (Thermo Finnigan MAT, Bremen, Germany). Sucrose ANU (IAEA, Vienna, Austria) and CaCO3 (NBS 19, Gaithersburg, USA) were used as calibration standards. Natural abundance of 13C was expressed as δ13C, which represents the ratio of 13C/12C relative to V-PDB. δ13C was determined in duplicate for each sample and analytical precision averaged ±0.15‰.

2.5. The 13C Nuclear Magnetic Resonance Spectroscopy of the Sum of Benzene Polycarboxylic Acids

[12] Purity of the dried aromatic acid fractions used for compound-specific δ13C analyses was checked by 13C nuclear magnetic resonance (NMR) spectroscopy. Twelve replicate samples of the mollic epipedon of the Kursk Chernozems were oxidized with HNO3 as described above to obtain sufficient material for solution-state 13C NMR spectroscopic analysis. The combined extracts were evaporated to dryness at 60°C. The residue, which was supposed to contain solely BPCA, was dissolved in 3 mL of 1.5 M NaOD and analyzed on a Bruker DRX 500 spectrometer: spectrometer frequency, 125 MHz; inverse-gated decoupling; acquisition time, 0.16 s; delay time, 1.84 s; line-broadening factor, 100 Hz.

2.6. Radiocarbon 14C Age Dating

[13] For 14C analyses, acid and base insolubles (humin) were isolated from selected samples (for further details, see Glaser and Zech [2005] and references therein). The radiocarbon 14C was measured with Accelerator Mass Spectrometry (AMS) at the Department of Physics, University of Erlangen-Nürnberg, Germany (target Erl lab numbers from 4292 to 4298). AMS radiocarbon ages were corrected according to the simultaneous recording of the 13C/12C and background 14C/13C ratios. Conventional radiocarbon ages (years BP) were calculated using international standards (NBS oxalic acid). Both conventional and recently calibrated 14C ages are given in Table 2. Since the older studies cited here report uncalibrated 14C data, only the conventional data were used for comparisons with earlier radiocarbon datings [Schlütz and Zech, 2004]. True calibrated 14C ages may become older when recent calibrations are used.

Table 2. Radiocarbon Ages of Mollisol Samples
SitesDepth (cm)Lab Number14C Years BP, Conventional14C Years BP, Calibratedaδ13C (‰)
Russia
Kursk (R4)b,c0–10 1000 ± 40 −27.39
 20–30 1440 ± 50 −26.91
 30–40 2400 ± 50 −27.15
 50–60 4040 ± 60 −25.43
 70–80 4580 ± 60 −25.16
Kursk (R5)b,c10–20 1680 ± 60 −25.75
 30–40 2950 ± 80 −26.20
 50–60 2970 ± 110 −25.88
 70–80 4020 ± 90 −24.92
 120–130 6100 ± 200  
 140–150 6700 ± 100  
Kursk (R6, forest)b,c3–11 1250 ± 40 −26.33
 12–18 1600 ± 40 −25.96
 70–80 5165 ± 60 −25.73
Orlovc10–20 1020 ± 70  
 50–60 2680 ± 80  
 110–120 4720 ± 60  
 240–250 12470 ± 360  
Tambovc10–20 2120 ± 130  
 30–40 2800 ± 120  
 70–80 2970 ± 130  
 120–130 6780 ± 150  
Charkovc10–20 1190 ± 60  
 50–60 2650 ± 70  
 110–120 5920 ± 140  
      
Argentina
Balcarce (A3)70–82Target Erl-42923588 ± 553890 ± 192−21.56
 >82Target Erl-42934828 ± 875526 ± 202−21.70
Chasico60–70 1400 ± 305d  
 90–100 3300 ± 100d  
Oriente110–115 4627 ± 76d  
      
China
Wulanhot (C4)160–180Target Erl-42945939 ± 726768 ± 200−20.10
 300–320Target Erl-42959473 ± 8610806 ± 303−22.89
Hulan (C1)175–192Target Erl-42966181 ± 617077 ± 169−21.36
      
Germany
Bad Lauchstädt (G1a)50–58Target Erl-42973659 ± 683977 ± 244−26.30
 90–117Target Erl-42984722 ± 675454 ± 135−24.04
Söllingene10–20 1210 ± 70  
 20–30 2070 ± 80  
 30–40 2560 ± 90  
 40–50 2310 ± 90  
 50–60 2830 ± 80  
 60–70 3020 ± 80  
 70–80 4800 ± 100  
 80–90 4000 ± 80  
      
United States
Akron, nativef0–10 193 ± 118  
 10–20 1528 ± 128  
 20–30 2398 ± 22  
 30–45 4068 ± 528  
Akron, cultivatedf30–60 5153 ± 333  
 60–90 6160 ± 285  
 90–120 7015 ± 450  
Mandanf0–15 1200 ± 75  
 15–30 2155  
 0–15 56  
 15–30 1825  
Haysf0–15 645 ± 5  
 15–30 1215 ± 55  

3. Results and Discussion

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Materials and Methods
  5. 3. Results and Discussion
  6. 4. Conclusions
  7. Acknowledgments
  8. References
  9. Supporting Information

3.1. Soil Organic Carbon Contents

[14] Chernozems, Phaeozems, Greyzems, and Kastanozems are the dominant soil types in the major grassland regions of the world (Figure 1). They are known for their high SOM contents and for their high soil fertility. The surface A horizons of the dark Greyzems under forest (Russia) contained the highest SOC contents (80 g SOC kg−1 soil; profile R1). Chernozems and Phaeozems (grassland ecosystems in Russia) frequently had SOC contents above 50 g SOC kg−1 soil. The SOC contents in Argentina and the USA were lower (<43 g SOC kg−1), followed by those in Germany and China (<25 g SOC kg−1; see Data Set S1). The lowest SOC contents of all A horizons occurred at great soil depths in the deeply developed Chinese Chernozems (3.8 g SOC kg−1 soil at 1.1 m depth in profile C3; see Data Set S1). Carbon stocks in China and Russia were more than twice as high as the stocks found at the other sites (Figure 2a; calculation possible from Data Set S1). When whole soil profiles are considered, as opposed to considering the A horizons only, the same picture emerges. Because of deeper soil formation rather than due to more efficient C storage within a given soil horizon, SOC stocks in the Asian sites exceeded, by far, those of the American ones (Figure 2a).

image

Figure 2. (a) Soil organic carbon (SOC) stocks and (b, c) contents and stocks of black carbon (BC) in the A horizons of Mollisols at the different geographic regions of the world (units in parentheses; error bars represent the standard deviations).

Download figure to PowerPoint

3.2. Black Carbon Contents

[15] Kastanozems mainly occur in the drier grassland areas. Kastanozems had low SOC stocks and also the lowest BC content of the soils under study (Table 1). The minimum BC contents of Kastanozems' mollic A horizons amounted to about 20 g BPCA-C kg−1 SOC (profile C6, China, and U4, USA; see Data Set S1). Even lower BPCA-C contents were found for the three soils of this study that were not Mollisols, i.e., the haplic Luvisol in Germany as well as the Cambisol and Calcisol in the United States (profiles G4, U5, U6; 10–15 g BPCA-C kg−1 SOC). The profiles G1a, G2, G3, G4 represent a sequence of advanced Chernozem degradation as a result of enhanced leaching and increasing Podzoluvisol formation. This was accompanied by an increase in lightness of soil color (Munsell value) in the order G1a = G2 < G3 < G4, which correlated with decreasing BC contents in these samples as suggested by Eckmeier et al. [2007]. Our data are consistent with this observation, and also with the idea that BC accumulates in soil as soil genesis proceeds [Czimczik and Masiello, 2007].

[16] In contrast to Kastanozems, the maximum contributions of BPCA-C to SOM of the A horizons reached 120 g BPCA-C kg−1 SOC in the deeper A horizons of Chinese Chernozems (profile C3, 60–70 cm soil depth; see Data Set S1). Between 4% (Kastanozems) and almost 30% of the soil C (Chernozems, Greyzems) consisted of BC when calculated with the conversion factor of 2.27 for minimum estimation of BC. On average, using the BPCA method, 11–15% of SOC was attributable to BC (Figure 2b), with Chernozems containing significantly (P < 0.05) higher amounts of BC than other soils under study. Schmidt et al. [2002] claimed that significant amounts of BC in chernozemic soils of central Europe result from ancient biomass burning; such burning events also are relevant for other grassland ecosystems of the world. Other sites having less SOC stored than do the Chernozems also exhibited lower BC contents (Figure 3). This correlation, previously reported only for U.S. sites [Glaser and Amelung, 2003], thus extends across different geographic regions. We may discount the possibility of an artificial BPCA formation from fresh plant material, as spiking different plant residues to soils did not result in additional BPCA formation using our modified BPCA procedure [Brodowski et al., 2005a]. Isotopic labeling experiments proved that less than 25% of BPCA may have been derived from unknown sources (melanoidins, other still unknown artifacts, biological sources) [Glaser and Knorr, 2008]. The pigments of Aspergillus niger may indeed form BPCA precursors in soils. Therefore, we cannot rule out the possibility that some biological processes contributed to structures similar to BC [Glaser and Knorr, 2008]. However, as outlined by Brodowksi et al. [2005a, p. 1308] “only 0.26 g highly aromatic carbon per kg organic C can be derived from aspergillin (without conversion factor), that being the case only if we assume a living fungal biomass of 5 mg C kg−1 soil, i.e., 10 times higher than supposed by Dunger [1974], 20 g kg−1 organic C in soil, a worst-case contribution of 10% of Aspergillus niger to the fungal biomass, and an enrichment factor of 1000 for the selective preservation of aspergillin during humification and litter decomposition.” Other biological sources of BPCA have not yet been identified. As in this study the lowest BPCA-C contents were 8–10 g BPCA-C kg−1 SOC (Data Set S1), we assume that the biological contribution to BC structures was small; that is, if BPCAs were detected, then BC was the most likely precursor for aromatic SOM formation.

image

Figure 3. Relationship between the stocks of soil organic carbon (SOC) and black carbon (BC) across the whole sampling depths in the Mollisols under study.

Download figure to PowerPoint

[17] Different rates of BC input and different soil-specific mechanisms of BC preservation may have contributed to higher BC contents in Chernozems than in the other soils studied. Low BC contents may be the result of: 1) low BC inputs from vegetation burnings (relatively infrequent fires); 2) low net primary production (biomass production) [Glaser and Amelung, 2003]; 3) more frequent re-burning of surface-accumulated BC at the drier sites [Czimczik and Masiello, 2007]; or 4) less efficient burial of BC into deeper soil horizons by earthworms [Czimczik and Masiello, 2007] at such sites relative to Chernozems. This also would result in lower aggregate stability and hence less subsequent stabilization within aggregates [Brodowski et al., 2006]. All above mentioned processes affect both the storage of total SOC as well as its storage in BC forms; that is, they sustain the correlation found between total BC and SOC stocks of the profiles (Figure 3).

[18] We failed to detect extremely high BC contributions to SOM in mollic horizons as reported, e.g., for a few German sites (up to 45%) [Schmidt et al., 1999]. We assume this is due to both an overestimation of BC with the NMR method [Hammes et al., 2007] used by Schmidt et al. [1999] and also due to anthropogenic input of BC from fossil fuel combustion in the industrialized areas of the former East Germany [Brodowski et al., 2007]. Only the deeper subsoil horizons (C horizons) of some German sites exhibited BC contributions > 130 g kg−1 SOC (Data Set S1) [see also Brodowski et al., 2007].

[19] When total SOC contents are low, the actual contribution of BC to SOC gives little information about the potential of soils to sequester C in charred forms. We calculated BC stocks from available and estimated bulk density values. Figure 2c reveals that highest BC stocks are found in the Russian sites, closely followed by the soils of the Manchurian steppe because of their thick A horizons. The BC stocks of the soils from other geographic areas were lower by a factor of three. In these loess-derived soils, BC stocks are building up over time so that thicker A horizons go along with higher BC contents (see also below when depth profiles are discussed). Extrapolation of these data to the total area of Mollisols in the world, 5.48 million square kilometers [Eswaran et al., 1993], suggests that between 4 and 17 Pg of BC are already stored in Mollisols. This is orders of magnitude larger (2000 to 8000 times) than the total amount of C that could be sequestered annually in the U.S. by converting to no-till agriculture [Bernacchi et al., 2005]. It is also 20–100 times larger in magnitude than the current gap of the C sink in Europe [Janssens et al., 2003].

[20] It should be recalled that the estimations of BC stocks are based on a conversion factor of 2.27, which is the lowest conversion factor ever reported for the BPCA method [Brodowski et al., 2005a] (see also method section). Therefore, the estimates given above yield minimum BC stocks in soils. True BC contents are likely to be higher than this estimate, which underlines the significance of charred C forms for C storage in grassland ecosystems. Changing this BC pool, e.g., by management, could thus also affect the global C balance. Nevertheless, although BC can be microbially degraded [e.g., Hamer, 2004; Hammes et al., 2008; Steinbeiss et al., 2009], it still has not yet been considered as a major part of the active (or slow) soil C pool [Skjemstad et al., 2004; Preston and Schmidt, 2006; Flessa et al., 2008].

3.3. Variation of BC With Soil Depth

[21] Soil BC contents did not change gradually with soil depth. In many sites, the lower part of the A horizon had a relatively high ratio of BC to SOM compared to the over- and underlying horizons (Figure 4, illustrated for selected soils only; also see Data Set S1). The depth of this “BC maximum” varied among the different geographic regions, although in general the thickness of the A horizon (China > Russia > other sites) determined the depth at which the BC maximum occurred (Figure 4). The reasons for these findings remain unclear.

image

Figure 4. Representative depth profiles of the contents of black carbon (BC) in major grassland regions of the world (see Data Set S1 for more details).

Download figure to PowerPoint

[22] Recent input of SOC from the growing vegetation should be nearly free of BC; a lower contribution of BC to the SOM of the surface soil layer may, in part, be attributed to recent dilution effects in the surface soil layer. Leaching of oxidized BC forms and biological mixing (e.g., transport by earthworms) may have contributed to the relative BC enrichment in the deeper soil horizons [Mitra et al., 2002; Knicker, 2007; Czimczik and Masiello, 2007]. Further, at the soil's surface, BC can be susceptible to physical breakdown [Patterson et al., 1987] and to particulate transport similar to fine clay particles [Ohta et al., 1986]. Both processes should selectively affect the less stable, poorly condensed BC. As outlined below, such BC structures were enriched in the lower soil horizons, although primarily at the dry sites.

3.4. Radiocarbon Analyses

[23] Selected soil samples were radiocarbon dated and compared with other published radiocarbon ages of loess soils in the study areas (Figure 5). Compared to all data sets (r2 = 0.81; see Figure 5), either the oldest data point from our sites or from the data acquisition in 1985 [Chichagova, 1985] (Table 2) was slightly biased which may be attributed to the lack of, e.g., suitable local dendrochronological calibrations to these conventional radiocarbon ages. Nevertheless, SOM ages in general increase with increasing soil depth in all soils under study (Figure 5), confirming Scharpenseel et al. [1986] that intensive bioturbation failed to homogenize radiocarbon ages within these soils. Progressive leaching, if it was present, failed to homogenize radiocarbon ages. Most of the SOM found at a certain soil depth likely reflected the age of its formation. This is likely true in particular for the BC pool, which is more stable against microbial degradation than many other SOM constituents.

image

Figure 5. Radiocarbon ages of bulk soil organic matter from selected soil horizons of Mollisols from different continents (see data in Table 2).

Download figure to PowerPoint

[24] The deeper the maximum of BC contribution to total SOM (Figure 4a), the older was the respective organic material (Table 2 and Figure 5). In no case was this BC maximum correlated with a maximum of SOC contents at this depth (Data Set S1 and Figure 4b). This supports the above argument that there was no significant biological BPCA production with increasing humification and SOM decomposition. If this were the case, then we might expect that the contribution of BPCA-like structures to total SOM would have continued to increase with increasing soil depth. It therefore seems reasonable to speculate that specific site conditions promote the accumulation of BC. The variation of radiocarbon data at a certain soil depth exceeded 1500 years for the upper 60 cm soil (Figure 5) and 3000 years below that depth, reflecting that different time periods in the Holocene favored Ah formation in the different geographic regions.

[25] At all sites, SOM accumulation started during the Holocene, but intensity of BC and SOC accumulation occurred during different time intervals. In the Southern hemisphere, most SOM accumulated during the last 3 ka (probably influenced by human activity), and also during more humid intervals (i.e., 5700 years BP and 9000–10,000 years BP) [Zech et al., 2009]. In Europe it was moist during the Atlantic (ca. 4500–7500 years BP) [Hintermaier-Erhard and Zech, 1997]. At all sites the BC maxima occurred at a depth of younger age. Two reasons might then account for the deeper BC maximum in Asia. First, the loess layers are thicker in Asia; burrowing animals could go deeper down the profile during winter, thus transporting more BC to greater soil depth, even if not successful in homogenizing BC gradients with depth. Second, soils of the Asian loess plateau are known to have developed mainly from mid-Holocene dust emissions (as shown, e.g., for the Guanzhong Basin [Zhao et al., 2007] and for Qilian Shan Mountains [Küster et al., 2006]). Such deposits reached a maximum accumulation rate of 16 cm ka−1, which could explain the deeper horizon with maximum BC enrichment found at the Chinese sites. In addition to aeolian deposits, colluvial and/or alluvial phenomena might have contributed to the formation of thick A horizons in Manchuria (profiles C1, C4), which exceeded 2 m (Data Set S1). Compared to the Chinese sites, Holocene dust accumulation and translocation in Germany was of much lower intensity. Charcoal-enriched Anthrosols in central Amazonia [Glaser et al., 2001] and the anthropogenically influenced black earths in the lower Rhine area, Germany [Eckmeier et al., 2007], are assumed to be influenced by tilling. More research will be required to confirm or reject the hypothesis that in the Manchurian region frequent tilling after charring, or even char applications by ancient populations, promoted the formation of fertile Chernozems. Similarly, for the Argentinian profiles, more research is needed to support the hypothesis that their specific locality was affected by colluvial impacts during early land uses.

[26] Burning of harvest residues is still a common agricultural practice in China. The BC from C4 plant remains such as maize can then later be plunged into the A horizon and thus contribute to a storage of C in charred forms. All other sites were conventionally plowed with the straw being frequently left in the field (Russia, Germany, USA).

3.5. Black Carbon Quality

[27] The BPCA marker method has one distinct advantage over other BC procedures: the pattern of BPCA produced gives additional insight into the properties of the relevant BC. In general, the more condensed the BC is, the higher the proportions of mellitic acid [Glaser et al., 1998]. In most of the soils, the BPCA pattern was dominated by mellitic acid ( = b6 in Data Set S1), followed by benzene pentacarboxylic acid (b5 in Data Set S1), benzene tetracarboxylic acids (b4 in Data Set S1), and benzene tricarboxylic acids (b3 in Data Set S1)). Overall, high portions of mellitic acid were observed only when the BC contents were high, and it appeared that a minimum mellitic acid contribution of 30% to total BPCA found was a prerequisite that BC stocks could exceed 0.3 kg m−2 (Figure 6). As condensed aromatic structures are hard to break down by the soil microbial community, the higher portions of mellitic acid thus confirm that stable BC forms may have persisted in the soils and could potentially be used to reconstruct ancient fire histories, especially at lower soil horizons not yet contaminated by fossil fuel BC input [Brodowski et al., 2007; Flessa et al., 2008].

image

Figure 6. Relationship between the degree of BC condensation (indicated by the contribution of mellitic acid C to total BPCA-C) and the BC stocks.

Download figure to PowerPoint

[28] As soil depth increased, many profiles exhibited increasing ratios of b4/b6 and b5/b6 (e.g., R1, R8, and R9; C1–C7; A1, A3, A4; G1–G2, G5; U4, U5). This suggests a preferential loss of condensed BC forms during early stages of BC decomposition and/or a preferential gain of less condensed BC in the subsurface soil. The sites in question are not predominantly located in mesic areas but rather are located in the drier areas. It seems unlikely that this enrichment over depth is caused by leaching. It also seems unlikely that the specific BC produced and accumulating in dry ecosystems had lower degree of condensation than that of more moist areas, because this still would not explain why we observe a BC maximum at intermediate soil depth. In the surface soil of the German sites, Brodowski et al. [2007] have shown that BC is enriched in mellitic acid as a result of the input of fossil fuel residues. However, this input did not extend below 30–40 cm soil depth. For the subsoil horizons with higher b4/b6 and b5/b6 ratios, we therefore suggest that lower portions of mellitic acid reflect advanced stages of BC decomposition. Increasing ratios of b4/b6 and b5/b6 in the surface soils for the German transect of Chernozem degradation in the order of G1a (0.43; 0.66) < G2 (0.51; 0.70) < G3 (0.76; 0.75) < G4 (0.74; 0.83) support this finding. In general, BC is oxidized from the surface of the particles [Brodowski et al., 2005b; Cheng et al., 2008], thus producing either free mellitic acid [Möller et al., 2000], which escapes our analytical window, or lower ranked BPCA upon our chemical HNO3 oxidation of the remaining BC. Reconstructing the origin of ancient BC in soils from the BPCA pattern under aerobic site conditions may therefore be difficult and should consider BC degradation kinetics which are currently poorly understood.

3.6. Natural 13C Isotope Signature

[29] The δ13C signal in plant biomass and SOM increases with increasing temperature and the associated shift from C3 to C4 vegetation [Boutton, 1996]. Within a given soil profile there was little change in δ13C, expressed as slightly elevated δ13C values with increasing soil depth. This enrichment of 13C natural abundances barely exceeded 2‰ (Data Set S1). The limited degree of enrichment may be attributed to microbial 13C discrimination with advanced SOM transformation and decay [Ehleringer et al., 2000]. The absence of larger isotopic variations suggests, however, that there have not been major shifts in vegetation type (C3 versus C4) during SOM formation in any of the soils under study. An exception were some sites in China where the surface SOM was slightly enriched in 13C compared to the subsoil horizons (Data Set S1). This may reflect a higher portion of C4-derived C in the mollic epipedons than in subsoil SOM, possibly due to remnants from, e.g., former maize cropping, or due to larger portions of isotopically light C3 plant species in the past.

[30] Among the different sites, the δ13C values increased in the order Russia (−25.5 ± 1.0‰) = Germany (−25.6 ± 0.7‰) < China (−21.9 ± 1.5‰) = Argentina (−21.2 ± 1.6‰) < USA (−18.1 ± 3.5‰). We interpret this as an indication that the origin of SOM changed from mainly C3-derived C to increasing proportions of C4-derived C in the same geographic order. In the U.S. prairies, C4 grasses (sites U1–3) or mixed C3–C4-type vegetation (U4–6) was the dominant belowground C input since 5000–8000 years BP, the radiocarbon age of the subsoil in that region (Table 2) [Paul et al., 1997]. This is in accordance with Clark et al. [2001] who showed a demise of woody vegetation followed by a fluctuating dominance of grasses (40% C4) and forbs during a period of maximum aridity at 8000–4000 years BP in east North Dakota. The modern forest-prairie ecotone in northeast Kansas has occupied the same position since about 5000 years BP [Kurmann, 1985]. Reversal of post-glacial warming, to become cooler and wetter, is reported to have probably occurred about 5000 [Wayne, 1991; Kelly et al., 1993] to 7000 years BP [Wright, 1970], causing the prairie to retreat to the west. Afterwards, an increase of C4 grass populations was recorded for paleosols, indicating a warmer climate [Kelly et al., 1993; Leavitt et al., 2007]. If the BC in these soils was mainly formed in situ, its δ13C should reflect these shifts in vegetation composition.

[31] To determine the δ13C signature of BC, we isolated BPCA after HNO3 digestion by drying the final solutions. As a purity check, one isolation was repeated several times from site Kursk to obtain sufficient material for solution-state 13C NMR spectroscopic analysis. The 13C-NMR spectrum of the residue after HNO3 digestion indicated only the presence of aromatic (110–160 ppm) and carboxyl carbon (170–180 ppm). The dominant signals in the NMR spectrum were attributable to benzene hexacarboxylate (135.1 ppm) and benzene pentacarboxylate (135.6 and 137.6 ppm), which agrees with the results of the gas chromatography analysis (Figure 7). We conclude that the isolated BPCA material reliably represented the fraction of interest for 13C natural abundance measurements; that is, the isolations were suitable for EA-IRMS assessment of bulk δ13C values of all BPCA isolates.

image

Figure 7. Solution 13C nuclear magnetic resonance (NMR) spectroscopic characterization of the benzene polycarboxylic acids from soil organic matter of the Russian plain (site R5, typic Hapludoll, 30–40 cm). Characteristic chemical shifts are shown for the aromic ring carbons in benzene hexacarboxylic acid (135.1 ppm) and pentacarboxylic acid (135.6 and 137.6 ppm) with the associated signals in the carboxylic C range at 176.5 and 176.7 ppm for benzene pentacarboxylic acid and at 177.0 ppm for carboxyl groups in benzene hexacarboxylic acid, respectively.

Download figure to PowerPoint

[32] The isotope measurements showed that the δ13C value of the BC markers was not systematically different to that of bulk SOM (Figure 8), despite the fact that burning itself may induce some minor 13C discrimination [Roscoe et al., 2000; Krull et al., 2003]. In general, these BC-specific δ13C values also increased in the order Russia (−25.9 ± 0.5‰) ≤ Germany (−23.9 ± 2.3‰) ≤ China (−22.4 ± 1.1‰) ≤ Argentina (−20.9 ± 1.3‰) < USA (−16 ± 0.9‰). We conclude that the majority of BC was indeed formed by in situ vegetation burning in most of the investigated sites. The originating C3/C4 vegetational composition of BC varied across the continents according to that of bulk SOM. However, the deviations from the 1:1 line in Figure 8 provide some clues to the origin of BC. In Northern and Southern America as well as in China, the δ13C values of BC were both higher or lower than that of bulk SOM. This could be due to the fact that in these large areas long distance transport contributed to BC pools from sources with other mixed C3/C4 plant composition [e.g., Teeri and Stowe, 1976; Tieszen, 1991; Kuhlbusch et al., 1998]. In Russia, such variable BC sources were not evident, since both the forests and major grasses were predominantly of C3 type. In Germany, however, there was a strong bias to C3-derived materials, mainly in the surface soils. We have additional evidence that these German sites are heavily contaminated with BC from fossil energy sources [Rethemeyer et al., 2004; Brodowski et al., 2007; Flessa et al., 2008]. Up to 50% of total C in these soils may originate from fossil C sources. However, it is not the BC but the total C which was biased toward lighter δ13C values (Figure 8). It seems likely, therefore, that the fossil C sources did not enter the soil as BC, but perhaps as lignite dust poor in BPCA [Laskov et al., 2002] from surrounding open cast mining. Such additional lignite is of C3 origin, because C4 plants did not exist at the time when lignite was formed, and such impacts certainly overlay past (and unknown) land-use effects. We did not find any systematic evidence, however, that significant additions of fossil C3-C might also have contributed to the BC content in the other mollic epipedons under study.

image

Figure 8. Relationship between δ13C values of black C with those of bulk SOM (unit of both axes in ‰ relative to PDB standard).

Download figure to PowerPoint

4. Conclusions

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Materials and Methods
  5. 3. Results and Discussion
  6. 4. Conclusions
  7. Acknowledgments
  8. References
  9. Supporting Information

[33] This study confirms that black Chernozems and Phaeozems in particular store significant amounts of BC in the surface horizons. Plant biomass production at these sites is generally higher than in other grassland ecosystems of the world. The large standing crop of grass biomass in these ecosystems means that a relatively high amount of BC is produced when the vegetation burns. The moist conditions at these sites likely prevent rapid re-burning of BC during dry periods. Moreover, these conditions facilitate leaching and biogenic burial and thus storage of BC at greater soil depths.

[34] Overall, 4–30% of SOC found in grassland ecosystems could be assigned to BC. The BC contents were similar in the different mollic epipedons from different continents. Nevertheless, the A horizons of Mollisols of the Chinese and Russian steppe contained almost three times as much BC as those in the American prairies, the Argentinian pampa, and German croplands. This was due to higher BC contents recovered in the subsoil; that is, the success of the soils to sequester C in charred forms is driven not only by different BC inputs and storage in the mollic surface horizon but mainly by the ability to store that BC at greater soil depths.

[35] As soil depth increased, some of this BC was less condensed which we attributed to advanced in situ degradation. In most cases the stable C isotope signature of this BC still resembled that of the vegetation currently found at each site, mainly consisting of C3 plants in Russia, with increasing contributions of C4 vegetation (Germany, China, Argentina, USA) to pure C4 grasses (USA). For some of the American and Chinese sites the bulk δ13C signals of BC varied, suggesting that a proportion of BC at these sites may have been from nonlocal sources, possibly derived after long-distance transport from other vegetation burnings. Only the BC of the surface A horizons of the former East-German sites was biased systematically toward depleted δ13C values, likely due to the input of additional fossil C such as lignite dust.

Acknowledgments

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Materials and Methods
  5. 3. Results and Discussion
  6. 4. Conclusions
  7. Acknowledgments
  8. References
  9. Supporting Information

[36] We thank Sonja Brodowski and two anonymous reviewers for constructive comments and Brendon Ladd and Joan Sandeno for their assistance in editing the manuscript. The Deutsche Forschungsgemeinschaft is gratefully acknowledged for financial support.

References

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Materials and Methods
  5. 3. Results and Discussion
  6. 4. Conclusions
  7. Acknowledgments
  8. References
  9. Supporting Information

Supporting Information

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Materials and Methods
  5. 3. Results and Discussion
  6. 4. Conclusions
  7. Acknowledgments
  8. References
  9. Supporting Information

Auxiliary material for this article contains a data set with soil properties and sets of horizon-specific original measurements of soil organic carbon, soil inorganic carbon, black carbon, bulk density, benzene polycarboxylic acids with 3–6 carboxylic groups (b3–b6), mellitic acid proportion, soil organic carbon stock, black carbon stock, 13C isotopic signature of soil organic carbon, and 13C isotopic signature of black carbon.

Auxiliary material files may require downloading to a local drive depending on platform, browser, configuration, and size. To open auxiliary materials in a browser, click on the label. To download, Right-click and select “Save Target As…” (PC) or CTRL-click and select “Download Link to Disk” (Mac).

See Plugins for a list of applications and supported file formats.

Additional file information is provided in the readme.txt.

FilenameFormatSizeDescription
gbc1692-sup-0001-readme.txtplain text document2Kreadme.txt
gbc1692-sup-0002-ds01.txtplain text document22KText S1. Soil properties under study.
gbc1692-sup-0003-t01.txtplain text document5KTab-delimited Table 1.
gbc1692-sup-0004-t02.txtplain text document2KTab-delimited Table 2.

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.