14C-labeled organic amendments: Characterization in different particle size fractions and humic acids in a long-term field experiment.

Knowledge about the stabilization of organic matter input to soil is essential for understanding the influence of different agricultural practices on turnover characteristics in agricultural soil systems. In this study, soil samples from a long-term field experiment were separated into silt- and clay-sized particles. In 1967, 14C labeled farmyard manure was applied to three different cropping systems: crop rotation, monoculture and permanent bare fallow. Humic acids (HAs) were extracted from silt- and clay-sized fractions and characterized using photometry, mid-infrared and fluorescence spectroscopy. Remaining 14C was determined in size fractions as well as in their extracted HAs. Yields of carbon and remaining 14C in HAs from silt-sized particles and Corg in clay-sized particles decreased significantly in the order: crop rotation > monoculture ≫ bare fallow. Thus, crop rotation not only had the largest overall C-pool in the experiment, but it also best stabilized the added manure. Mid-infrared spectroscopy could distinguish between HAs from different particle size soil fractions. With spectroscopic methods significant differences between the cropping systems were detectable in fewer cases compared to quantitative results of HAs (yields, 14C, Corg and Nt). The trends, however, pointed towards increased humification of HAs from bare fallow systems compared to crop rotation and monoculture as well as of HAs from clay-sized particles compared to silt-sized particles. Our study clearly shows that the largest differences were observed between bare fallow on one hand and monoculture and crop rotation on the other.

Introduction

The turnover of carbon pools in soil under differen class="Chemical">nt agricultural management systems is a key issue, particularly with regard to CO2 emissions from different agricultural practices. The global reservoir of soil organic matter (SOM) is estimated to be approx. 1600 Gt (Hedges and Oades, 1997). Knowledge about the distribution of carbon pools in particle size fractions in soil is crucial because they represent pools of differing stability. Many studies have shown that different cultivation practices have an effect on the quantity and quality of soil organic carbon (SOC) in particle size fractions (e.g. Covaleda et al., 2006; Guggenberger et al., 1994; Mikha and Rice, 2004; Six et al., 2000).

The advantage of following the decomposition of n class="Chemical">14C-labeled material is its excellent “determinability” because it is a β-emitter at 0.155 MeV and has a half-life of 5730 yr (Jenkinson, 1971). Early studies addressing the decomposition of 14C labeled organic matter (OM) in soil investigated 14C labeled straw with (Sørensen, 1963) and without (Szolnoki and Vágó, 1959) determination in extracted SOM. Extraction of SOM has involved a variety of methods and a long history. The fate of amended organic material in alkaline extracted organic material was evaluated early on (Sauerbeck and Führ, 1970; Szolnoki and Vágó, 1959; Zeller et al., 1966), but only for bulk soils and amended plant materials. Schnitzer and Schuppli (1989) compared 0.5 M NaOH- and 0.1 M Na4P2O7-extracted HAs from selected soils and particle-size fractions, applying 13C-CPMAS NMR spectroscopy. For both extracting agents, higher aromaticity values and lower aliphaticity values were obtained for medium silt vs. coarse clay (Schnitzer and Schuppli, 1989). As far as we are aware, no study involving the extraction of humic acids (HAs) from particle size soil fractions containing 14C-labeled organic amendments has been carried out — unlike studies dealing with HAs extracted from particle size fractions of unlabeled soil. Extraction using 1 M NaOH was successfully applied (Tatzber et al., 2007, 2008, 2009a,b) to evaluate the characteristics of SOM originating from bulk soils of different tillage or cropping treatments. In this context, this fraction exhibited a very dynamic behavior. Furthermore, Stemmer et al. (2000) investigated 14C in different particle size fractions in the same long-term field experiment and obtained the highest proportions of remaining 14C in the silt and clay fractions. This makes an investigation of the dynamic fraction of HAs in these particle size fractions promising.

OM extraction requires additional effort but provides valuable information (von Lützow et al., 2007). Due to the time-consuming methodology and very low yields of n class="Chemical">fulvic acid extractions, we focused on the SOM fraction of HAs, which are known indicators for soil properties. Analyses of organic matter extracted from soil have a long history, but also very recent applications (Dos Santos et al., 2010; Fernández-Getino et al., 2010; Grasset et al., 2010; Vergnoux et al., 2011).

In 1967, an open-land, long-term field experiment was established that included three differen class="Chemical">nt cropping systems and application of 14C-labeled farmyard manure. The 14C-measurements of Tatzber et al. (2009a) revealed that the HAs consisted mostly of rapidly depleted 14C-labeled material, with a very small proportion having a significantly longer half-life. Further, HAs were successfully investigated with a selection of spectroscopic methods (mid-infrared and fluorescence spectroscopy) to follow and characterize their decomposition behavior (Tatzber et al., 2009b). These methods enabled consistently following the characteristics of different mid-infrared bands with time, as was also the case for fluorescence spectroscopy (Tatzber et al., 2009b).

The present work was designed to study the transfer of n class="Chemical">14C-labeled material to different size fractions of silt and clay and their HAs, because most organic carbon is stored in these particles. Both 14C and spectroscopic measurements should provide valuable information about the dynamics of this SOM-pool in the different particle size soil fractions under the different cropping systems.

Materials and methods

Experimental field

In 1967, a long-term field experiment with n class="Chemical">14C-labeled farmyard manure was established with three different cropping systems, i.e. crop rotation (sugar beet (Beta vulgaris L.), winter wheat (Triticum aestivum) and spring barley (Hordeum distichum)), monoculture (spring wheat (T. aestivum)) and permanent bare fallow. The experiment is located in Fuchsenbigl, Lower Austria, approx. 45 km east of Vienna. Before the start of the experiment, the field site had been used as farmland. The average annual temperature at the field site is 9.4 °C and the mean annual precipitation 529 mm. The soil is described as Haplic Chernozem according to WRB (World Reference Base of Soil Resources) (2006: clay: 15%, silt: 41%, sand: 54%, pH(CaCl2): 7.40, organic C: 1.53%, inorganic C: 1.61%, total N: 0.16%; values were measured before the beginning of the experiment; investigated soil depth: 0–20 cm).

The experiment consisted of six plots (2.56 m² each, each being divided in class="Chemical">nto four sub-plots for soil sampling), which were separately tilled, cropped, harvested and analyzed, leading to four sub-plot replicates for each treatment. The sub-plots were separated by eternit-plates (0.6 m high and entering to a soil depth of 0.5 m) and the sub-plots were separately tilled, cropped and harvested; no radioactivity was detectable beside the plots (Oberländer and Roth, 1974). 14C-labeled material was applied as straw or farmyard manure to each of the three plots, treated with three different cropping systems each: crop rotation, monoculture and permanent bare fallow. The farmyard manure was produced by feeding a cow with 14C-labeled plant material under veterinary supervision. For details about the production of the 14C-labeled farmyard manure, see Zeller et al. (1968). The labeled farmyard manure was applied once in November 1967 (equivalent to 5 t dry matter ha− 1) to all cropping systems. Organic residues were removed after harvest; accordingly, the annual plant input through plant biomass in crop rotation and monoculture resulted from remaining roots and micro organisms. These inputs were calculated at 2.21 t ha− 1 for winter wheat, 1.34 t ha− 1 for spring barley and 0.92 t ha− 1 for sugar beet in crop rotation and 1.29 t ha− 1 for spring wheat in monoculture (Rampazzo Todorovic, 2008, personal communication; Steen and Lindén, 1987).

For the present study, only the farmyard manured plots were sampled. Detailed descriptions of the experimen class="Chemical">nt are given in previous works (Oberländer and Roth, 1974; Tatzber et al., 2009a; Zeller et al., 1968); results on the depletion of the different labeled organic manures were reported by Oberländer and Roth (1980). The production of the 14C-labeled wheat straw was described by Oberländer and Roth (1968), and the generation of farmyard manure and the design of the long-term field experiment were described in detail by Oberländer and Roth (1974).

Soil sampling

Soil samples were taken in April 2004 from the plots at a soil depth of 0–20 cm. For every sub-plot, twenty samples were taken, mixed, 2 mm-sieved and air-dried. Three of the four sub-plots were sampled in this way, resulting in three sub-plot replicates.

Separation into silt and clay fractions

The approach for separating silt and clay fractions was based on the work of Stemmer et al. (1998) with some modifications. Ultra-sonication of the soil was carried out with a Bandelin ultrasonic homogenizer (Sonoplus n class="Disease">HD 2200) and a VS 70 T probe. The ultrasonic homogenizer included a titanium flat tip TT13 of 13 mm diameter and had an HF-output Weff of max. 200 J s− 1. The insertion depth of the probe was 10 mm, its diameter 13 mm, the length ca. 126 mm and the amplitude 153 μmSS (peak to peak). An aliquot (100 g) of 2 mm sieved and air-dried soil suspended in 200 mL of water was sonicated (2 min) at 50% of the maximum power. As the purpose was to destroy the macro-aggregates, no calorimetric calibration of absorbed power was performed; attention was paid to the reproducibility of the experiment. Subsequently, the silt and clay fractions were separated from the bulk soil samples by sieving through a 63 μm sieve, with a vibration amplitude of 2 μm, rinsing with 600 mL of water and collected. Four bottles were filled with 200 mL suspended silt and clay each, placed in the centrifuge-rotor and centrifuged (1 min) at 185 ×g (acceleration and break times not included). For better separation from clay, each pellet (silt) was suspended with water (by filling up to 200 mL and shaking) and centrifuged in the same way. This “washing” procedure was repeated (3 ×).

After transferring and combining the silt pellets in a separate bottle, the clay suspensions were centrifuged at 3900 × g for 30 min. This was repeated (3 ×) un class="Chemical">ntil all remaining suspensions were separated from clay. The next step was to transfer the combined pellets into a separate bottle (analog to silt) and to freeze-dry the fractions. The isolated fractions were stored in airproof bottles.

Extraction of HAs

HA extraction was carried out according to Tatzber et al. (2007) and Tatzber et al. (2009a) with some modifications. Freeze-dried silt (30 g) or clay (7 g; due to different organic n class="Chemical">carbon (Corg) content) was shaken overnight with 1 M NaOH (volume of soil and NaOH: 200 mL) in a 250 mL plastic bottle. On the next day, the suspension was centrifuged for 10 min at 12,400 × g. The extraction was replicated twice, followed by precipitation of the HAs with 37% HCl. The first washing of the HAs was performed with 20 mL 1 M NaOH, refilling with water to a volume of 200 mL and centrifuging this solution (25,900 ×g, ≥ 30 min) to remove the fine clay particles (not done for the bulk soil samples). Subsequently, this alkaline solution was acidified with 3 mL 37% HCl after separation from the centrifuged pellet. The second washing was carried out with 10 mL 1 M NaOH and 1.5 mL 37% HCl and for the third washing, the precipitated HAs were suspended with water and acidified with 10 drops of 37% HCl from a Pasteur pipette. The purified HAs were transferred with a small amount of H2O to 50 mL plastic bottles and freeze-dried. The dried samples were stored in small air-proof bottles.

Determination of total C, Corg and total nitrogen (Nt)

Total C and Nt content of soil fractions and HAs were measured after dry combustion at 1050 °C using a Carlo-Erba Total Analyser. Subtraction of inorganic carbon from total C yielded the Corg content (ÖNORM L 1080, 1999). The carbonate content of the samples was measured using the Scheibler method and the content of inorganic carbon was calculated. About 0.5–1 g dry soil sample was treated with 10 mL of 10% HCl and connected to the gas-tight Scheibler apparatus. Released CO2 was measured volumetrically. The apparatus was calibrated with pure CaCO3, which was treated in the same way (ÖNORM L 1084, 1999).

Measurement of 14C in bulk soil samples and HAs

The freeze-dried silt and clay samples were powdered in an agate mill. 14C was measured using a mixture of 0.5 g soil and 0.2 mL Combust-Aid® which was ashed in a Sample Oxidizer Packard 1415 in pure oxygen flow. The produced CO2 was reduced to carbamine using Carbo-Sorb® and adsorbed in Szintillator Permafluor®. The measurements were performed with a WALLAC 1415 3×30 LSC apparatus (efficiency: 80%, blank value: 20 cpm). Subsequently, each scintillation cocktail was mixed with 50 μL Spec-Check (4355 cpm) and remeasured.

HA samples were ashed in a Sample Oxidizer Packard 307 in pure oxygen flow. In the case of very low extraction yield, 5 or 2 mg HAs were mixed with ca. 50 mg standard soil of known background activity.

Mid-infrared measurements of HAs

Mid-infrared measurements were based on Tatzber et al. (2008) and Tatzber et al. (2009b). Mid-infrared spectra were obtained from KBr pellets, which were produced by mixing an exactly weighed amoun class="Chemical">nt (ca. 5 mg) of a HA/KBr suspension (10% (w/w) extracted HAs in KBr) with pure KBr (FT-IR grade) to a total weight of ca. 200 mg in a swinging mill. The mixture was pressed to obtain a transparent pellet. The FT-IR measurements were performed with a Perkin Elmer Paragon 500 Spectrometer in the mid infrared area from 400 to 4000 cm− 1. Recording was performed with a resolution of 4 cm− 1, weak apodization and 16 scans per sample. Each spectrum was corrected against pure KBr and the ambient air as a background spectrum. Band evaluations were performed with start- and end-basis points and are consistent with the method of tangential baselines. The observed dimension of the units of the peak area was A cm− 1 or A cm− 1 (mg HA–C)− 1 after normalization of the values based on mg C in the dried HA (mg HA–C). Detailed interpretation of the mid-infrared bands can be found in Table 3. The bands were selected based on the greatest differences between the soil fractions and cropping systems. They were band number (BN) I (3050 cm− 1, aromatic groups; Senesi et al., 2003), BN II (1700 cm− 1, carbonyl vibrations of carboxyl groups, aldehydes and ketones; Senesi et al., 2003; Hesse et al., 2005), BN III (1500 cm− 1, amide II band and C═C vibrations; Senesi et al., 2003; Hesse et al., 2005), BN IV (1450 cm− 1, aliphatic C–H deformation; Hesse et al., 2005; Senesi et al., 2003), BN V (1420 cm− 1, amide III band and carboxylates; Hesse et al., 2005; Senesi et al., 2003), BN VI (1370 cm− 1, CO–CH3 groups and nitrates; Hesse et al., 2005), BN VII (1315 cm− 1, sulfone groups and possibly one of two ester bands; Francioso et al., 1998; Hesse et al., 2005; Senesi et al., 2003) and BN VIII (870–732 cm− 1, aromatic C–H groups; Hesse et al., 2005; Senesi et al., 2003).

Table 3

Bands in mid-infrared spectra, including possible assignments.

Basepoints for tangential base-line corrected band integration	Band number for evaluation	Conventional band assignment(s)
3110–3019	BN I	Aromatic groups (Senesi et al., 2003)
2998–2880		Aliphatic groups (Senesi et al., 2003)
2880–2821		Aliphatic groups (Senesi et al., 2003), methoxy groups (Hesse et al., 2005)
1765–1698	BN II	Carbonyl groups (Senesi et al., 2003)
1698–1610		C=O of amides (amide I) and quinone ketones (Senesi et al., 2003), α-, β-unstaurated ketones, amide I and II bands (Hesse et al., 2005)
1610–1576		C=C, carboxylates (Senesi et al., 2003)
1556–1526		NH and C=N (amide II band), C=C (Senesi et al., 2003)
1526–1486	BN III	NH and C=N (amide II band), C=C (Senesi et al., 2003), NH (Hesse et al., 2005)
1477–1439	BN IV	C–H (Senesi et al., 2003, Hesse et al., 2005)
1439–1397	BN V	Amide III (Senesi et al., 2003), carboxylates (Hesse et al., 2005)
1393–1352	BN VI	–CO–CH3, nitrates (Hesse et al., 2005)
1348–1312	BN VII	Sulfone groups and/or esters (2 bands between 1330 and 1050 cm− 1) (Hesse et al., 2005)
1290–1251		C–O of phenolic groups (Francioso et al., 1998), R–O–NO2, =C–O–C, P–O–Aryl (Hesse et al., 2005)
1251–1195		C–O and OH from COOH, C–O from aryl ethers and phenols (Senesi et al., 2003)
1141–1104	BN VIII	C–OH from aliphatic alcohols (Senesi et al., 2003), sulfone groups (Hesse et al., 2005)
1099–1067		C–O of alcohols and aliphatic ethers (Senesi et al., 2003)
1064–1004		C–O of polysaccharides or similar substances, Si–O of silicates (Senesi et al., 2003), possible contributions: S=O, P–O–alkyl and/or =C–O–C (Hesse et al., 2005)
870–732	BN IX	C–H of aromatic systems (Senesi et al., 2003); less substituted systems appear at lower wave numbers (Hesse et al., 2005)

Fluorescence spectroscopic measurements of HAs

Fluorimetric measurements were performed according to Tatzber et al. (2008) and Tatzber et al. (2009b). HA (3 mg) in KBr (ca. 10% HA) were dissolved in 3 mL of 0.3 M n class="Chemical">Na2CO3 −/0.3 M NaHCO3 buffer (pH 9.5). The solution was transferred to a fluorescence cuvette and measured with a Perkin Elmer LS50B Luminescence Spectrometer. Constant pH was maintained by dissolving the HAs in the buffer solution. Spectra of selected samples were recorded in 3D. The stability of the maxima of the investigated regions was verified. Based on these data, emission- or excitation-scans with fixed excitation or emission were chosen to compare the intensities. Three scans were accumulated for each measurement and background correction was applied. We used an exact concentration of 100 mg L− 1 of NaOH-extracted HAs and performed the measurements immediately after dissolving the HAs in the buffer.

Based on preliminary work (Tatzber et al., 2009b), the following peaks were taken into account for characterizing the HAs (λexc max/λem max; λexc means wavelength maximum of excitation and λem means wavelength maximum of emission):

Max at 392/470 nm (region 1), scan at λexc 392 nm, λem from 460 to 485 nm.

Max at 435/514 (region 2) and 459/517 nm (region 3), scan at λem 516 nm, λexc from 430 to 470 nm.

Additionally, a reference HA was always measured at the beginning and the end, and sometimes in between, a series of single scans to verify RFI signal stability. The signal was always stable. All data are based on Corg in HA dry weight. Due to best data quality (lower standard deviation) of region 3, data from this region were evaluated for the determination of fluorimetric characteristics.

Photometric measurements

Photometric measurements were conducted in a concen class="Chemical">ntration of 100 mg L− 1. Six-milligram suspensions of 10% HAs in KBr were dissolved in 6 mL 0.3 M Na2CO3 −/0.3 M NaHCO3 buffer (pH 9.5). The absorbance was measured at 400 and 600 nm with an Agilent 8453 UV–visible Spectrophotometer plus sipper system and flow cell.

Statistical evaluation

For the determination of significant differences in quan class="Chemical">ntitative parameters (e.g.: yield, 14C content), infrared, fluorimetric and/or photometric characteristics, univariate analyses of variance (ANOVAs) were calculated for differentiating the cropping systems separately in fractions of silt- and clay-sized particles and also between these two sizes. Bonferroni post hoc tests were calculated to differentiate between the cropping systems. The significance level for all the calculations was assumed at 0.05. The software used for these calculations was PASW Statistics 18 (a recent version of SPSS).

Results

Corg and Nt in bulk soil, soil size fractions and HAs including HA yield

Basic data of bulk soils (Corg and n class="Chemical">Nt) and HAs (Corg, Nt, yields and ash contents) plus their explanations are provided in Tatzber et al. (2009a,b). Significant differences between the cropping systems were obtained for bulk soil Corg (with permanent bare fallow being significantly lower from crop rotation and monoculture), bulk soil Nt (where permanent bare fallow was significantly lower than crop rotation; monoculture did not differ significantly from one of these groups) and yields of HA–C (where all three groups differed significantly from each other).

Table 1 includes the data for separated fractions of silt- and clay-sized particles. The proportions of the fractions were consistent for each plot. n class="Chemical">Corg was significantly higher in the clay- vs. silt fraction; furthermore, there were three significantly different treatments (crop rotation > monoculture > bare fallow) of Corg content in the clay-sized particles. No significant differences between the different plots were detected in the silt-sized fraction. Nt behaved similarly in clay and silt; in the clay fraction, however, only two significant groups were detected: the permanent bare fallow was significantly reduced vs. the two other cropping systems. The HA content (based on carbon inside the HAs yielded by extractions, units for this were g HA–C per kg soil; see Table 1) was significantly higher in the silt-sized particles than in clay, and cropping significantly impacted the HA yield from silt. Referring to HA contents in clay-sized particles, the bare fallow plot was significantly more depleted than the crop rotation and monoculture plots. In the HAs from both particle-size soil fractions, carbon concentration was lower in the bare-fallow plots (significantly in clay-sized particles). HAs from silt-sized particles contained significantly more carbon than those from clay. Both particle-sizes were not significantly different with respect to Nt in HAs; in both, the values were significantly lower for permanent bare fallow plots compared to crop rotation; for clay-sized particles the difference was also significant between monoculture and bare fallow. The ash content for both particle-sizes was higher for permanent bare fallow, quite clearly in the case of clay-sized particles.

Table 1

Basic data for separated silt and clay fractions. Note that except for Nt in HAs (where significances were not provided) and ash contents (where significances were not calculated) all determined variables showed significant differences between the particle size fractions. For bulk soil data see Tatzber et al. (2009a,b).

Plot	Soil fraction in % of bulk soil1	Corg(g kg− 1)1	Nt(g kg− 1)1	Yield of HA–C2 in soil fraction(g kg− 1)1	Corg in HAs(%)1	Nt in HAs(%)1	Ash contents of HAs(%)
Silt
Crop rotation	35 ± 1a	12.4 ± 1.3 a	1.2 ± 0.1 a	1.70 ± 0.02 a	50.26 ± 0.11a	4.7 ± 0.2 a	2.1 ± 0.6
Monoculture	34.1 ± 0.6 a	12 ± 4 a	1.4 ± 0.3 a	1.55 ± 0.06 b	50.7 ± 1.8 a	4.37 ± 0.03ab	2.2 ± 0.7
Permanent bare fallow	34.8 ± 0.6 a	9.2 ± 0.8 a	0.9 ± 0. 1 a	0.75 ± 0.05c	48.4 ± 0.3 a	4.20 ± 0.09 b	3.8 ± 0.7
Clay
Crop rotation	9.6 ± 0.7 a	45.5 ± 1.1 a	4.56 ± 0.13 a	0.74 ± 0.06 a	44.3 ± 1.8 a	4.70 ± 0.36 a	2.8 ± 1.2
Monoculture	8.9 ± 0.6 a	41.6 ± 0.3b	4.41 ± 0.15 a	0.63 ± 0.04 a	46.0 ± 0.4 a	4.88 ± 0.07 a	2.1 ± 1.6
Permanent bare fallow	9.2 ± 0.5 a	35.9 ± 1.2 c	3.44 ± 0.15 b	0.49 ± 0.05 b	40.0 ± 1.4 b	4.01 ± 0.13 b	12 ± 9

Different letters label significantly different groups (P < 0.05); a, b and c identify groups in the cropping systems, x and y differences between particle sizes.

Yield of HAs × Corg of HAs.

14C in bulk soil, separated fractions of silt and clay and their HAs

Table 2 provides an overview of all the 14C results. For the n class="Chemical">14C-data of bulk soils and their HAs see also Tatzber et al. (2009a). When determining the remaining farmyard manure in the particle-size fractions, significantly more material was stored in silt, without significant trends between the cropping systems. For HAs from bulk soil, significantly more 14C-labeled material was found in the crop rotation and monoculture plots than in the bare fallow. The most distinct differences were obtained for HAs extracted from silt- and clay-sized particle fractions. Approx. ten times more 14C-labeled material was stored in HAs from silt- vs. clay-sized particles. For silt, 14C in HAs differed significantly between cropping systems. HAs from the clay in the bare fallow plot contained significantly less 14C vs. crop rotation; HAs from monoculture clay-sized particles did not significantly differ from the other two cropping systems, resulting in intermediate values.

Table 2

Remaining 14C labeled material in different soil pools. For 14C-data in bulk soils and HAs from bulk soils, see also Tatzber et al. (2009a). Note that for 14C-labeled material in both soil materials and HAs, differences were significant between the particle size fractions.

Plot	Remaining 14C-labeled material in bulk soil1 or coming from particle size soil fraction2 (% of original application1)	Remaining 14C-labeled material in bulk soil fraction of HAs3 or coming from particle size soil fraction of HAs4 (‰ of original application1)
Bulk soil5
Crop rotation	8.7 ± 0.3 a	7.3 ± 0.4 a
Monoculture	8.5 ± 0.5 a	7.0 ± 0.4 a
Permanent bare fallow	7.5 ± 0.9 a	4.2 ± 0.3 b
Silt
Crop rotation	3.9 ± 0.8 a	6.5 ± 0.2 a
Monoculture	4.0 ± 1.6 a	5.69 ± 0.17 b
Permanent bare fallow	4.2 ± 1.8 a	4.1 ± 0.4 c
Clay
Crop rotation	2.2 ± 0.3 a	0.43 ± 0.04 a
Monoculture	1.8 ± 0.3 a	0.36 ± 0.03 ab
Permanent bare fallow	1.67 ± 0.12 a	0.32 ± 0.04 b

100% of original application = 0.7603 mg 14C labeled farmyard-C per g soil.

Calculation of values : x = 14C labeled material per g soil fraction × % share of the fraction in bulk soil / 0.7603 × 100%.

Calculation of values : x = 14C labeled material in HAs (mg mg− 1) × 100 / % Corg (in HAs) × yield of HA–C (mg g− 1) / 0.7603 × 1000‰.

Calculation of values : x = 14 C labeled material in HAs (mg mg− 1) x 100 / % Corg (in HAs) × yield of HA–C (mg g− 1) × % of fraction in soil / 0.7603 × 1000‰.

Data referring to bulk soils are published as Tatzber et al. (2009a,b).

Mid-infrared spectroscopy of HAs

Fig. 1 shows one mid-infrared spectrum of HAs of every soil particle-size fraction of all involved cropping systems plus the highlighted bands which were evaluated. Differences between the spectra of HAs from silt and clay-sized particles in the fingerprin class="Chemical">nt area are clearly visible. Detailed band interpretations are presented in Table 3. Figs. 2a–e and 3a–d show the results for all band areas chosen for study (see Table 3). The criteria for choosing bands were clear assignments and significant differences between particle-sizes and/or cropping systems. For instance, the bands in the aliphatic stretching region showed no significant trends, either between particle-sizes or between cropping systems (data not shown). Fig. 2a and b show two aromatic bands, one in the stretching region (Fig. 2a) and one in the deformation region (Fig. 2b). For both bands, HAs from clay showed significantly stronger signals than for silt. No significances were apparent between the cropping systems. Carbonyl bands (Fig. 2c) were significantly more intense for HAs from clay; no significant differences were apparent for the cropping systems. Two mid-infrared bands assignable to amide groups (Fig. 2d, e) were significantly lower in HAs from clay- than from silt-sized particles; these bands also showed no significant trends regarding cropping systems. The band representing aliphatic deformation vibrations (BN IV, Fig. 3a) was significantly larger in the HAs from silt- than from clay-sized particles, without showing significant differences between the investigated cropping systems. For the band representing acetyl-groups (CO–CH3) and nitrates (BN VI, Fig. 3b), the HAs from clay-sized particles provided more intense signals. Furthermore, the signal intensities for bare fallow were significantly higher than for crop rotation. The band representing sulfone and/or ester groups (BN VII, Fig. 3c) provided significantly higher signals for the HAs from silt vs. clay. Finally, HAs from the silt fraction differed significantly between crop rotation and permanent bare fallow. Another significant difference was obtained between the HAs from the two particle-size classes for the band assignable to aliphatic alcohol groups (BN VIII, Fig. 3d). The HA signals from clay-sized particles were significantly lower than from silt-sized particles. For the HAs from clay, the bare fallow plot had significantly lower signals than either crop rotation or monoculture.

Fig. 1

Mid-infrared spectra of HAs from silt- and clay-sized particles. The bands with the highest differentiation potential are highlighted. SSP, silt-sized particles; CSP, clay-sized particles. For band numbers (BN) and their assignments see Table 3.

Fig. 2

Mid-infrared band areas of signals assignable to aromatic stretching and deformation vibrations (BN I (a) and IX (b)), carbonyl (BN II (c)) and amide bands (BN III (d) and V (e)) of mid-infrared measurements of HAs. Values given as means and standard deviations (n = 3). SSP, silt-sized particles; CSP, clay-sized particles; a and b indicate significantly different groups between cropping systems. All signals in this figure showed significant differences between silt- and clay-sized particles.

Fig. 3

Band areas of aliphats (BN IV, deformation vibrations (a)), CO–CH3 and nitrate groups (BN VI (b)), sulfone and/or ester groups (BN VII (c)) and a band consisting of aliphatic alcohols and sulfones (BN VIII (d)) of mid-infrared measurements of HAs. Values given as means and standard deviations (n = 3). SSP, silt-sized particles; CSP, clay-sized particles; a and b indicate significantly different groups between cropping systems. All evaluated signals in this figure showed significant differences between silt- and clay-sized particles.

Fluorescence spectroscopy and photometry of HAs

Fig. 4 gives the results of photometric (Fig. 4a, b) and fluorescence spectroscopic (Fig. 4c) measurements. For all three approaches, HA signals from silt-sized particles were significantly lower than those from clay-sized particles. For photometry at 400 nm, HAs from bare fallow plots showed more intense signals than those from crop rotation with respect to the silt-sized particle fractions. The HAs from silt, measured both fluorimetrically and photometrically at 600 nm, showed significantly higher signals for bare fallow plots than for both crop rotation and monoculture plots.

Fig. 4

Evaluation of photometric (at 400 and 600 nm) and fluorimetric (Region 3 in Fig. 4) characteristics. Values given as means and standard deviations (n = 3). SSP, silt-sized particles; CSP, clay-sized particles; a and b indicate significantly different groups between cropping systems. All signals in this figure showed significant differences between silt- and clay-sized particles.

Discussion

Corg and Nt in bulk soil, particle-size fractions and HAs, including yields

Recalculating the values in Table 1 to pool size (multiplying Corg and n class="Chemical">Nt values by the soil fraction in % of bulk soil) showed that most Corg and Nt were in silt- and clay-sized particles, which is in line with previous work (Stemmer et al., 2000). Similar findings have been reported from other sites (e.g.: Benoit and Preston, 2000; Gerzabek et al., 2006; Guggenberger et al., 1995; Schulten et al., 1993). The significant differences in Corg (as well as Nt, in a smoothed way) between the cropping systems for the clay-sized fraction suggest this pool to be a deposit of a carbon pool built up and influenced by the supply of fresh material, resulting from the type of cropping system. The results show that mainly clay-sized particles contribute to the significant differences in Corg and Nt for bulk soils between cropping systems.

In the silt-sized fraction, Corg tended to behave similarly in the differen class="Chemical">nt crop systems. Although the differences between the systems were bigger for silt than for clay, they were not significant. Beedy et al. (2010) investigated the impact of a gliricidia/maize intercropping system on SOM: the SOM content was significantly correlated with the clay content for the gliricidia/maize and sole maize systems. In the present study, clay also contributed the highest proportion of SOM (both unlabeled and 14C-labeled) to the bulk SOM pool, as a result of its higher SOM content (see Table 1). In a study by Gerzabek et al. (2006), the influence of different amendments was investigated and the Corg content of different particle-size fractions was quantified for a long-term field experiment in Ultuna, Sweden. There, Corg and Nt contents changed significantly for both silt- and clay-sized particles treated differently for 44 years with fallow, no N, Ca(NO3)2, animal manure and grassland. The soil type in the Ultuna experiment is quite different (Eutric Cambisol, pH 6.6), and the distribution of particle-sizes (silt: 52.6%, clay: 23.6%) differs somewhat from our soils.

Yields of HA carbon were significantly higher for the silt- than for the clay-sized fractions. This underlines the comparably young character of the HAs extracted with 1M NaOH; since we previously found inverse relation between HA-yield and humification degree of SOM (see Tatzber et al., 2008; 2009a,b). The higher number of obtained significant differences between the HAs from the different cropping systems and from silt-sized particles supported this impression of the young character. Moreover, HA yields for the bare fallow system were also significantly lower than for both other cropping systems with respect to clay-sized particles. Remarkably, the behavior of the HA pools differed completely from that of the Corg pools in the different particle-sizes. This shows that the fraction of HAs extracted with 1 M NaOH was a sub-pool of SOM with a special behavior.

Schnitzer and Schuppli (1989) investigated HAs from particle-size fractions, extracted with 0.5 M NaOH and 0.1 M n class="Chemical">Na4P2O7. Importantly, the analysis involved coarse clay (0.2–2.0 μm) and medium silt (5.0–20.0 μm) fractions of two different soil samples (Bainsville (pH 6.3): 13.5% coarse clay, 3.23% C; 22.7% silt, 4.36% C; Melfort (pH 5.72): 40.0% coarse clay, 6.29% C; 15.3% medium silt, 3.49% C). Using 0.5 M NaOH, more HAs were obtained from the Bainsville medium silt fraction than from the coarse clay fraction. An inverse result was obtained for the Melfort soil (Schnitzer and Schuppli, 1989). Clapp and Hayes (1999) characterized amendments of corn-residues via HAs from silt- and clay-fractions. There, HA-yield in the clay was clearly higher than from silt. Chefetz et al. (2002) structurally characterized (among others) HAs using pyrolysis-GC/MS and 13C NMR spectroscopy. There, the size fraction < 2 μm contained less HAs than the 2–50 μm fraction. Thus, the presently determined ratio between silt and clay in terms of HA-yield is clearly site specific and not universally valid.

14C in bulk soils, silt and clay fractions and their HAs

The results for remaining 14C show that significantly more farmyard manure-C was stored in silt (Table 2). No significant differences between the cropping systems were detected (see also Tatzber et al., 2009a). Bruun et al. (2008) investigated the incorporation of 14C-labeled barley straw in two soils after 40 days and 40 years: the silt and clay fractions contained a large proportion of 40-yr-old C and a small proportion of 40-day-old C. This agrees with the present study, investigating the stabilization of 14C-labeled material after almost 40 years. Gerzabek et al. (2001) evaluated the incorporation of organic amendments in particle-size fractions using 13C measurements. Organic carbon in these fractions responded significantly to the different amendments; most Corg was in the silt, in agreement with the present study.

The HAs from particle-size fractions contained significan class="Chemical">ntly different amounts of 14C for the different cropping systems. The more distinct differentiation between the cropping systems in the silt-sized fraction reflects the different C turnover and inputs in this fraction, respectively. Clearly, the turnover of the HA pool differs from that of the bulk SOM, leading to more significant differences between the cropping systems in the HA pool. We found no similar study dealing with 14C-labeled farmyard manure with respect to HAs, especially not for particle-size soil fractions. Note that the 14C concentration in the HAs was clearly higher for the silt HAs than the clay HAs. The concentrations of 14C-labeled material in the HAs were 4.19 ± 0.03 for crop rotation, 4.16 ± 0.17 for monoculture and 5.7 ± 0.2 for bare fallow for silt-sized particles, and 2.05 ± 0.10 for crop rotation, 2.24 ± 0.05 for monoculture and 2.16 ± 0.02 for bare fallow for clay-sized particles (all values in μg farmyard C per mg HAs). Thus, the highly significant differences between the different origins (particle-size fractions and cropping systems) of HAs reflect both C pool size and 14C concentration within the pools.

The mid-infrared spectra in Fig. 1 show qualitative differences between HAs from silt- and clay-sized particles. The best examples are in the fingerprint region, especially bands V–VIII. The exact evaluation of mid-infrared band areas (Figs. 2a–e and 3a–d) shows that differen class="Chemical">ntiation of HAs between particle-size fractions is clearly possible (significantly different groups of particle-sizes for every evaluated band). Cropping systems influenced the mid-infrared spectra less: only 2 bands differed significantly. This arrangement suggests that both pools – HAs from silt-sized particles and HAs from clay-sized particles – formed certain molecular characteristics connected with the different cropping systems over the period 1967 to 2004. These results, as well as the tendency for higher aromatic-C and carbonyl-C in the silt-sized fraction of bare fallow (aromatic bands I and IX, Fig. 2a, b; carbonyl band BN II, Fig. 2c) indicate a more humified nature of HAs than for the two other cropping systems. According to Zech et al. (1997), increasing humification is indicated by increases in COOH and aliphatic groups, a slight increase in aromatic C and a decrease of O-alkyl groups. Taking these findings as benchmarks, both the clay-sized fractions and the bare fallow systems contained more humified material, although no differences in the aliphatic region were detected.

Fluorescence spectroscopy and photometry of extracted HAs

Aromatic properties of the HAs were investigated using fluorescence spectroscopy (Fig. 4c), photometry (Fig. 4a, b) and mid-infrared spectroscopy via bands I and IX (Fig. 2a, b). There was agreement between the applied methods: aromatic signals for HAs extracted from clay-sized particles were significantly higher than for HAs from silt. For HAs from silt, the bare fallow system had significantly higher aromatic signals than both other cropping systems for fluorescence spectroscopy and photometry at 600 nm; photometry at 400 nm showed significant differences only for crop rotation and bare fallow. For the two mid-infrared bands assignable to aromatic characteristics (Fig. 2a, b), these trends were present but not significant. This indicates a higher sensitivity of fluorescence spectroscopy and photometry than mid-infrared spectroscopy with respect to aromatic characteristics. The aforementioned work of Schnitzer and Schuppli (1989) includes data for HAs extracted with 0.5 M NaOH and characterized via 13C-NMR spectroscopy and photometry. For both investigated soil samples, higher aromaticity was obtained for the HAs from silt than for those from clay for E4/E6 as well as 13C-NMR spectroscopy. The difference obtained in the present study underpins the site and soil type specificity of the results. With respect to aromaticity, the results of the aforementioned work of Chefetz et al. (2002) are also in line with the present study.

Examining the data in view of organic matter turnover and composition revealed some interesting aspects. The fraction of HAs extracted with 1 M NaOH was dynamic based on a number of properties: the yields, carbon and nitrogen contents as well as spectroscopic data. The strongest differences were detected for the 14C-labeled material contained in the HA fraction between both particle-size fractions and cropping systems. This fraction also depleted faster (yield of HA–C) than the Corg (Table 1), which further supports its dynamic character. Although the fraction of HAs represents only one tenth of the organic carbon in the particle-size soil fractions, it serves as a useful indicator for influences on the soil system not only for the spectroscopic but also for quantitative parameters (Tables 1 and 2).

With regard to turnover, silt apparently contained mostly younger material, with clay being a kind of end-deposit for SOM. In previous work (Tatzber et al., 2009a), the HA fraction seemed to contain two pools: a larger, dynamic one and a very short half-life and a smaller one with a comparably long half-life. The present study suggests that the first, larger sub-pool is located in the silt and the second, very small pool in the clay. This is supported by the spectroscopic results, which always differentiated significantly between HAs of the two different particle-size soil fractions. These spectroscopic results showed that materials extracted from clay were more humified than those extracted from silt. These results are in line with the work of Tatzber et al. (2009b) where a time-sequence of HAs from the bare fallow plot was investigated with the same spectroscopic methods (mid-infrared spectroscopy, fluorimetry and photometry). This study leaves the impression that the stabilization processes are ongoing and lead from the SOM in the silt fraction to an end-deposit in the clay fraction in the investigated long-term field experiment. Our results could once more show the ability of the HA fraction for reflecting SOM-characteristics quite sensitively.

Limitations of approaches used

Some of the methods applied in this study are being discussed in the soil science community, calling for some comments here on their applicability. First, HAs refer to a fraction of SOM that can be altered by the extraction with 1 M n class="Chemical">NaOH. This should mainly influence phenolic and carboxylic characteristics, ester groups, but less aliphatic and aromatic characteristics because the latter are too inert to react with 1 M NaOH at room temperature. The infrared bands shown in Figs. 2c, e and 3c could be affected due to Kolbe–Schmitt-like reactions. However, the significant differences due to different particle-size fractions and sometimes also cropping systems were also reflected by the bands in Figs. 2c, e and 3c, which shows that these groups were still indicating soil characteristics significantly. This could be due to the fact that the reactions with 1 M NaOH were not occurring (e.g. the Kolbe–Schmitt-reaction is reported to run at high pressure and temperature) or that they influenced the functionalities of all different cropping systems equally, resulting in maintained significant differences. Hence, especially functional groups which could be affected by the extractant may be cautiously taken as indicators for differences in soil characteristics; however, they do not allow for making assumptions about absolute contents of functional groups in the original SOM from which the HAs were extracted. The indicative properties could be shown in previous works and influences from 1 M NaOH were already discussed in some of them also (Tatzber et al., 2007, 2008, 2009a,b). As a consequence, more legitimate approaches are comparisons between the extracted HAs, their pool-sizes and semi-quantitative comparisons. This should be kept in mind whenever applying HA extractions.

Second, mid-infrared spectroscopy is not appropriate to draw conclusions about absolute quantities of functional groups in SOM due to band overlaps in the spectra. n class="Chemical">Nonetheless, this method can also be used to semiquantitatively differentiate between samples of different origins. The tangential base line correction in context with band integrations allows for excluding most band interactions and hence obtaining semiquantitatively useable results. In fluorescence spectroscopy, inner filter effects are important, requiring working with defined concentrations. A connection of aromatic signals from mid-infrared spectroscopy with photometric and fluorimetric signals enhances the plausibility of the obtained results. Finally, previous studies show that mid-infrared bands and fluorescence measurements significantly reflect differently tilled or cropped soil samples (Tatzber et al., 2008, 2009b). Thus, the authors expressively recommend the applied methodology for other scientific questions but draw the readers' attention to certain limitations.

Conclusions

The spectroscopic approaches differentiated between HAs from differen class="Chemical">nt particle-size soil fractions. All observed infrared bands showed significant differences between the HAs from the silt and clay fractions. Two mid-infrared bands showed significant differences between cropping systems. In both cases the signals from HAs of bare fallow had lower signal intensities than the other cropping systems. Fluorescence spectroscopy and photometry were more sensitive for differentiating between cropping systems. For these methods, all observed signals delivered significant differences not only between HAs from silt and clay, but also between HAs from bare fallow and crop rotation. In all significant cases, the aromaticity for HAs was higher from bare fallow systems.

Spectroscopic results indicated greater humification for HAs from clay-sized particles than from silt-sized fractions and for bare fallow systems vs. monoculture and n class="Disease">crop rotation.

The turnover characteristics of carbon pools differed in the three cropping systen class="Disease">ms. The largest differences were obtained for the bare fallow system compared with monoculture and especially with crop rotation; differences between the latter two were also obtained: The crop rotation system stored significantly more Corg in clay-sized particles, significantly more HA carbon in bulk soils and silt-sized particles (Table 1) and significantly more remaining 14C labeled material in HAs from silt-sized particles, compared with the monoculture system (Table 2).

The three different cropping systen class="Disease">ms could be differentiated significantly with respect to a) yields of carbon extracted with HAs from bulk soils (see also Tatzber et al., 2009a,b) and silt-sized particles, b) Corg of clay-sized particles and c) remaining 14C-labeled material in the HAs from silt-sized particles. The remaining 14C-labeled material in the HAs extracted from clay was clearly lower than for bulk soils or silt-sized particles. This suggests that the stabilization process leading from the silt 14C-pool (expected to contain younger material) to the clay 14C-pool (expected to contain older material) had not yet reached its equilibrium in the year 2004, which means a time-period of more than 36 years.

Carbon stabilization was ranked as follows in the three cropping systen class="Disease">ms: crop rotation > monoculture ≫ permanent bare fallow. Thus, crop rotation not only had the largest overall C-pool in the long-term experiment (Corg in bulk soils and particle-size fractions), but it also stabilized the added farmyard manure best (14C data of bulk soils (see also Tatzber et al., 2009a,b) and particle-size fractions, see Table 2). These results were mirrored by the extracted HAs.

Further investigations of other SOM-pools, especially the humin fraction, could add to our overall understanding of fluxes between the OC-pools in the experiment.