• Open Access

‘Bioenergy from cattle manure? Implications of anaerobic digestion and subsequent pyrolysis for carbon and nitrogen dynamics in soil’

Authors


Abstract

Cattle manure can be processed to produce bioenergy, resulting in by-products with different physicochemical characteristics. To evaluate whether application of such bioenergy by-products to soils would be beneficial compared with their unprocessed counterpart, we quantified differences in greenhouse gas emissions and carbon (C) and nitrogen (N) dynamics in soil. Three by-products (15N-labeled cattle manure, from which anaerobic digestate was obtained, which was subsequently pyrolysed) were applied to a loess and a sandy soil in a laboratory incubation study. The highest losses of soil C from biological activity (CO2 respiration) were observed in manure treatments (39% and 32% for loess and sandy soil), followed by digestate (31% and and 18%), and biochar (15% and and 7%). Emissions of nitrous oxide (N2O) ranged from 0.6% of applied N from biochar to 4.0% from manure. Isotope labeling indicated that manure N was most readily mineralized, contributing 50% to soil inorganic N. The anaerobic digestate was the only by-product increasing the mineral N pool, while reducing emissions of N2O compared with manure. In biochar treatments, less than 18.3% of soil mineral N derived from the biochar, while it did not constrain mineralization of native soil N. By-products of anaerobic digestion and pyrolysis revealed soil fertility in addition to environmental benefits. However, the reported advantages lessen when the declining yields of C and N over the bioenergy chain are considered.

Introduction

Recent concerns that bioenergy production from conventional crops might negatively affect food production has shifted research toward the development of ‘second generation’ biofuels from alternative sources of biomass (Koh & Ghazoul, 2008). A potentially large contribution could come from crop residues and animal manures (Woolf et al., 2010). However, crop residues and manures fulfill an important role in the maintenance of soil quality. Lal (2005) estimated that the removal of 30 to 40% of crop residues from land can deplete the soil organic matter (SOM) pool and cause land productivity to decline. However, a mature bioenergy industry will generate considerable quantities of bioenergy by-products, which might be returned to the soil instead (Cayuela et al., 2010; Taheripour et al., 2010). A shift in products used as soil amendments will translate into effects on greenhouse gas (GHG) emissions, carbon storage, and soil fertility, which should be included in the overall assessments of bioenergy chains.

Animal manures may be processed to improve their properties and to derive energy. For example, they may be physically separated in a liquid and a solid fraction, or be subject to anaerobic digestion (Bertora et al., 2008). The digestion process decreases the feedstock's carbon : nitrogen ratio and chemical and biological oxygen demand, while yielding biogas, a mixture of methane (CH4), carbon dioxide (CO2) and trace gases (Bousfield et al., 1979; Ward et al., 2008). Most research suggests that especially the easily mineralizable carbon (C) and nitrogen (N) are digested, resulting in lower emissions of GHGs, such as CO2 and nitrous oxide (N2O) and greater stability of C in the soil following their application (Amon et al., 2006; Marcato et al., 2009; Möller & Stinner, 2009), although other workers found no differences (Clemens et al., 2006; Bertora et al., 2008).

Another bioenergy platform that currently gains much attention is pyrolysis (Laird et al., 2009). Its solid by-product, ‘biochar’, is rich in stable aromatic compounds and has been identified as means to enhance soil fertility and sequester photosynthetically fixed C in soils (Lehmann et al., 2006; Gaunt & Lehmann, 2008). The N present in biochars is also predominantly present in persistent, heterocyclic compounds (Knicker, 2010). Recent studies indicate that soils amended with biochar may even reduce soil N2O emissions, most likely by affecting the rates of nitrification and denitrification (Clough & Condron, 2010; Zhang et al., 2010). Hence, there are multiple treatment or conversion options to change the nature of C and N in biomass residues and the nature and rate of soil processes following their application to soil. Especially when applied in sequence, the processing steps increasingly derive energy from the residue and change its composition. Recent work suggests that pyrolysis of anaerobically digested matter results in properties beneficial for soil application (Inyang et al., 2010). However, such a sequence also increasingly reduces the total and easily degradable C and N content of the residue. We question whether these modifications encourage the use of bioenergy by-products as soil amendments; their impact on SOM quality and nutrient availability remain largely unknown.

The comparison of different residues in terms of C and N dynamics may be assisted by stable isotope techniques. Stable isotopes may be used as biomarkers to follow the evolution of certain pools and pathways in the soil more closely, for example by distinguishing between mineralization of applied organic matter vs. originally present SOM. There are many studies using 15N to follow the mineralization of residues (Sørensen, 2001; Bol et al., 2003; Kuzyakov et al., 2009; Hilscher & Knicker, 2011) or fertilizers (Van et al., 1985; Schulten & Schnitzer, 1997; Bengtsson et al., 2003; Burger & Jackson, 2003) in soil. However, tracer studies comparing manures and digestates are scarce, and to our knowledge no soil incubations have been performed of a sequence of labeled manure bioenergy by-products.

The aim of the present study was to quantify GHG emissions and evaluate C and N dynamics of soils amended with cattle manure and digestate and biochar derived from that manure. A 15N labeled cattle manure was prepared, from which we obtained an anaerobic digestate that was subsequently pyrolysed. In theory, these three products represent increasingly stabilized organic matter, decreasing the bioavailability of C and N. Specifically, we tested the hypotheses that, compared with manure, (1) the digestate and the biochar have increasingly higher C sequestration potential; (2) N mineralization of the digestate and biochar are progressively lower, so that these residues will be less effective as short-term fertilizers; and (3) emissions of N2O from amended soils decline over the sequence.

Materials and methods

Soils and residues used for incubations

Two typical agricultural soils in the Netherlands with different properties were selected for the incubation experiments: a loess soil and a sandy soil. The loess soil (20% sand, 61% silt, and 19% clay, pH 6.4; C : N ratio 18) was collected at the arable farm ‘Wijnandsrade’ (50°54′N, 5°52′E). The sandy soil (75% sand, 23% silt, and 2% clay; pH 4.7; C : N ratio 11) was collected at the experimental farm ‘Droevendaal’ (51°59′N, 5°39′E). Both soils were sampled from the 0–25 cm layer of the arable field. Air-dried soils were sieved (<7 mm) and stored (20 °C) until the beginning of the experiments.

Residues of three successive stages of biomass processing were used (details in Table 1):

Table 1. Main chemical properties of residues used for incubation with soil
 ManureDigestateBiochar
  1. TOC, total organic carbon; TN, total nitrogen.

TOC (% dry weight)37.1430.3631.73
TN (% dry weight)1.522.441.71
TOC : TN24.412.518.6
Atom% 15N0.5510.7590.656

15N labeled cattle manure (0.551 atom%15N);

15N labeled anaerobic digestate from the above manure (0.759 atom%15N);

15N labeled biochar from the above digestate (0.656 atom%15N);

The 15N labeled cattle manure was prepared by feeding 15N labeled rye grass (Lolium perenne L.) to a nonlactating cow (Powell & Wu, 1999). The 15N enrichment of the rye grass was 18.4 atom% excess. The excrements of the cow (urine and feces) were collected over a 7-day period, and the urine and feces mixed.

A 10 L subsample of the 15N labeled manure was anaerobically digested at 35 °C in an 11 L continuously stirred tank reactor for 44 days at the Department of Environmental Technology, Wageningen University, the Netherlands.

A 5 L subsample of the digestate was air-dried at 25 °C for 2 days and consecutively pyrolysed (flash pyrolysis) at 500 °C (PyRos-process patented by TNO NL99/00688) in the Laboratory of Thermal Engineering, University of Twente, the Netherlands.

Manure and digestate were freeze-dried. All residues were ground and sieved (<0.5 mm) before application to avoid a particle size effect.

Experimental setup

The incubation experiment was carried out with units consisting of 250 g (oven-dry weight basis) soil in 500 ml polypropylene jars (Sarstedt, Nümbrecht Germany) at 20 °C. Before the start of the incubation, soils were adjusted to ca. 60% of water holding capacity (WHC) and pre-incubated at 20 °C for 7 days. Subsequently, ground manure, digestate, and biochar were added and thoroughly mixed with soil. Application rate was equivalent to 50 mg N kg−1 soil. Soils without amendment were included as controls. De-ionized water was added to adjust moisture to approximately 70% (loess soil) and 60% (sandy soil) of WHC [total water added: 64.5 g (loess), 37.5 g (sandy soil)]. The jars were covered with a woven black polyethylene cloth to allow gaseous exchange, but retard evaporation and prevent exposure to light. The units were maintained in a climatic room with a constant temperature (20 °C) and air humidity (40%).

During the first 30 days, the soil moisture was kept constant by biweekly gravimetrical adjustments with de-ionized water for each individual unit, at least 12 h before a gas measurement. The average evaporation was 4 g kg−1 soil day−1. Subsequently a dry-wet cycle was simulated. From day 30 to 48, soil moisture adjustment was aborted, after which the WHCs of the loess soil and sandy soil were about 50% and 30%, respectively. On day 49, water was added to reach about 80% of WHC. On day 58, a freeze-thaw cycle was simulated by storing the jars for 4 days in a freezer (−18 °C). After 48 h of thawing, the original moisture regime of two gravimetrical adjustments per week to 70% and 60% of WHC was regained during 22 days to conclude the experiment on day 84.

The experiment was laid out as a randomized block design with five replicates per treatment. Parallel incubations with three replicates per treatment were set up for destructive soil sampling after 2, 8, 29, and 84 days.

Gaseous emissions

Emissions of CO2 and N2O were measured using the flux chamber technique and an acoustic gas monitor (Innova AirTech 1314, Ballerup, Denmark). Measurements took place every day in the first week, after which the frequency was decreased (2–3 times per week). For 4 and 5 days after rewetting and after thawing, respectively, daily measurements were performed again. Measurements were conducted by sealing each unit with gas-tight polypropylene screw caps for 20–40 min. A soda-lime filter was used during N2O emissions to minimize interference by CO2 (Velthof et al., 2002).

C and N extractions

Moist soils were extracted by shaking three replicates of each treatment (1/10 w/v, dry weight basis) with 0.01 m CaCl2 for 2 h. Extracts were centrifuged (15 min, 2345 g) and filtered (0.45 μm) before analysis. Soil-extractable C and N were analyzed in a TOC-TN analyzer (Skalar Analytical, Breda, the Netherlands). Soil mineral N (NH4+, NO3 and NO2) was determined spectrophotometrically by a continuous flow analyzer (Brann en Luebbe TrAAcs 800 Autoanalyzer; Skalar Analytical B.V. Breda, the Netherlands). Soil microbial biomass N was determined on day 84 by the fumigation-extraction method (Brookes et al., 1985), using 0.5 m K2SO4 (1/4 w/v, dry weight basis) as extractant.

Isotopic signature

The 15N enrichment of inorganic N was determined using a microdiffusion technique based on Van Groenigen et al. (2005). In short, the technique relies on the collection of a known amount of N by pushing the N pool of interest to ammonia, that precipitates in a Glassfiber microfilter (GF/A, 6 mm diameter, Whatman Nederland B.V., AH‘s-hertogenbosch, the Netherlands) spiked with 2 m KHSO4. The microfilters were packed in Teflon tape to seal the filter from the solution, but enabling diffusion of ammonia. An amount of soil extract containing about 100 μg N was collected in a 100 mL container and supplemented with KCl (aq) to obtain 100 mL 1 m solution, leaving about 5 mL headspace. The solution received 4 mg MgO to raise the pH to approximately 10 and 4 mg Devarda's alloy to convert all NO3 into NH4+. The containers were closed airtight and stored for 7 days at 20 °C with periodic shaking of the samples, after which the sealed filters were collected, washed with de-ionized water, unsealed, dried, and placed in tin capsules for analysis. Inorganic N and its 15N enrichment were analyzed using an automated C/N analyzer-isotope ratio mass spectrometer PDZ Europa ANCA-GSL elemental analyzer interfaced to a PDZ Europa 20-20 isotope ratio mass spectrometer (Sercon Ltd., Cheshire, UK) at UC Davis Stable Isotope Facility, USA.

Calculations and statistical analysis

The N2O and CO2 fluxes were calculated assuming linear accumulation over time during lid closing, which has been confirmed in earlier incubations (e.g. Velthof et al., 2002). Cumulative emissions were calculated assuming linear changes between subsequent measurements.

The percentage of dissolved inorganic N in soil originating from the by-products was calculated from the isotopic signature assuming 0.3663% 15N abundance of soil native N.

Kinetics of CO2 release from amended soils was fitted with a two-compartment model according to the following equation:

display math(1)

where f is the fraction (%) of the labile C pool, k1 is the decomposition rate constant for the labile pool, k2 for the stable pool, and t is the incubation time in days. The fraction of fresh organic matter that remains in soil after 1 year (the humification coefficient h), was calculated based on an average annual temperature of 10 °C (Bradbury et al., 1993). We used the complete data set (including the dry-wet and freeze-thaw cycles) for calculating h, given that such disturbances are common in field situations.

All results are expressed on an oven-dry basis (105 °C, 24 h). For comparison of treatments, anova was used followed by Tukey-HSD test at the 95% level of probability using R Statistical Package.

Results

Carbon dioxide emissions and C sequestration potential

The C decay patterns in soil following the incorporation of the different by-products were fitted to the two-pool model (Fig. 1, Table 2). Manure addition led to highest CO2 emissions (39% and 32% relative to applied C in loess and sandy soil, respectively), followed by digestate (31% and 18%) and biochar (15% and 7%). Emissions of CO2 from biochar amended soils leveled off around day 40, whereas digestate and manure continued releasing CO2. Fitting the model resulted in low decay rates for the stable C pool in biochar in both soils (k2 < 1E-8). As a consequence, the humification coefficient (the percentage of C remaining after 1 year at 10 °C) was highest for biochar (85% and 93%). By contrast, k2 of manure and digestate ranged between 0.0022 and 0.0038, resulting in C retention of 53–72% after 1 year.

Figure 1.

Remaining fraction of organic C in soil amended with manure, digestate, and biochar; (a) loess soil and (b) sandy soil. Comparison of experimental results (symbols) with the two-pool model fitting.

Table 2. Parameters obtained from the fitted two-pool model
SoilAmendmentf (%) k 1 k 2 χ 2 h (%)
  1. f is the fraction of organic matter in the labile C pool, k1, and k2 are the decomposition rate constants of the labile and stable pool, respectively, and h is the humification coefficient (proportion of C remaining in soil after 1 year according to the two-pool model with rate constants corrected for an average annual temperature of 10 ºC).

LoessManure23.30.06850.002911.9252.5
 Digestate11.40.06690.003424.7956.8
 Biochar15.50.0870-4.9384.5
SandyManure10.90.10500.00387.6954.7
 Digestate4.30.19200.002210.7371.9
 Biochar6.60.06933.55 × 10−92.3293.4

In all treatments, a CO2 flush was observed in the first 2 weeks of the incubation, after which emission rates gradually declined (Fig. S1, S2). Effects of rewetting and thawing varied between soils: in loess, rewetting yielded no clear pattern whereas thawing resulted in an emission peak that was especially pronounced for manure. In sandy soil, manure produced the biggest peak after rewetting.

Nitrous oxide emissions

Figure 2 shows cumulative N2O emissions from the amended loess soil during the incubation period. Fluxes of N2O were highest for manure; 0.8% of manure N was lost as N2O in the first 3 weeks, whereas emissions from digestate and biochar were not different from the control. Rewetting followed by a freeze-thaw cycle caused cumulative emissions from manure amended loess to increase to 3.7%. Rewetting and thawing also led to discernible N2O emissions from loess amended with digestate (1.1% of added N) and biochar (0.6% of added N). These were, however, not significantly different from the control (P < 0.05).

Figure 2.

Cumulative N2O emissions following the application of manure, digestate, and biochar to loess soil. Letters show significant differences (after 78 days) according to the Tukey test (P < 0.05).

No significant N2O emissions were measured from sandy soils (Fig. S4).

Extractable organic carbon

The CaCl2-extractable organic C (EOC) of the control was relatively stable over time. All amendments resulted in significant increases of the EOC pool compared with the control (Fig. 3). The EOC of amended soils decreased after day 2, but remained significantly higher than the control, except for the digestate amended loess soil. At the end of the incubation, biochar amended soil resulted in highest EOC levels, followed by manure, digestate, and control (P < 0.05 between all treatments in sandy soil). When corrected for the amount of added C (Table 1), amendment of the digestate resulted in the highest levels of EOC in both soils per g of added C (Table S1).

Figure 3.

CaCl2-extractable organic carbon (EOC) in loess soil (a) and sandy soil (b), 2, 8, 29, and 84 days after application of manure, digestate, and biochar. Letters show significant differences between treatments on each separate day according to the Tukey test (P < 0.05).

Extractable nitrogen

Inorganic N (NO3, NO2, and NH4+) increased in loess from about 40 to 80 mg N kg−1 during the experiment, and in the sandy soil from about 20 to 40 mg N kg−1 (Fig. 4). In all treatments, the largest increases occurred between day 29 and 84, when the dry-wet and freeze-thaw cycles took place. These increases were all significant at the 99% confidence level, except for manure amended loess soil (Table S2).

Figure 4.

Inorganic nitrogen (NH4+, NO2, and NO3) in loess soil (a) and sandy soil (b), 2, 8, 29, and 84 days after application of manure, digestate, and biochar. The gray segments indicate the amount of inorganic N derived from mineralization of the amended residue as determined by isotopic tracing. Letters show significant differences between treatments on each separate day according to the Tukey test (P < 0.05).

The proportion of inorganic N originating from the by-product (gray segments in Fig. 4), as identified by isotopic analysis, was consistently highest for manure, followed by digestate and biochar (Table 3). In manure amended soils, inorganic N originating from the residue remained largely constant at about 50% in loess and 78% in sandy soil. In digestate amended soils, inorganic N derived from the residue remained constant until day 29, followed by significant increases between day 29 and 84 (P < 0.001). Inorganic N derived from biochar was only 7% in loess and 18% in sandy soil. Inorganic N values in loess were always higher than in sandy soil; inorganic N mineralized from each by-product, however, was very similar in both soils.

Table 3. Percentage of inorganic N (NO3, NO2, and NH4+) derived from the added by-product as deduced from the isotopic dilution (relative to the total mineral N extracted from soil).
TreatmentIncubation time (days)
282984 anova
  1. One-way analysis of variance (anova) was applied to detect significant differences among treatments and time. Values in the same column followed by the same letter are not significantly different according to the Tukey test (P < 0.05).

  2. ns, not significant.

  3. a

    P < 0.001;

  4. b

    P < 0.01;

  5. c

    P < 0.05.

Loess soil
Manure54.1 Aab47.5 Aa52.8 Aab57.4 Ab c
Digestate16.2 Ba19.5 Ba19.2 Ba42.2 Bb a
Biochar3.1 Ca4.5 Cab6.4 Cbc7.4 Cc b
anova a a a a  
Sandy soil
Manure78.1 Aa79.6 Aa74.7 Aa78.8 Aans
Digestate36.3 Ba36.8 Ba36.8 Ba64.2 Bb a
Biochar7.2 Ca8.8 Ca10.3 Ca18.3 Cb a
anova a a a a  

The share of NH4+ in the inorganic N pool was consistently negligible in this experiment (data not shown). Extractable organic nitrogen (EON) was relatively low (Table S3), but amended soils had higher EON than controls. Expressed as percentage of total extractable N, EON reached 14% (digestate) to 22% (manure) in sandy soil on day 2, compared with a maximum of 7% in loess (biochar).

The amount of N in microbial biomass, recorded in the end of the incubation, was highest for manure amended soils (Table S4). However, differences were not significant (P > 0.05).

Discussion

Stability of carbon

Anaerobic digestion and pyrolysis increasingly stabilized organic matter by reducing the easily biodegradable C fractions. As a result, the mineralization rate of remaining C was low, which confirms the findings of previous independent studies on digestates and biochars (Clemens et al., 2006; Zimmerman, 2010). Although C in digestates of animal slurries is typically found to be more stable than C in untreated slurries (Marcato et al., 2009), differences are not always seen (e.g. Bertora et al., 2008). Cayuela et al. (2010) found anaerobic digestion of pig slurry to reduce C loss from 57% to 40% over 60 days, but recorded no difference between cattle manure and its digestate (38% and 36% of added C lost as CO2, respectively). The observed variability in recalcitrance to mineralization of digestion residues may therefore be explained by dissimilar feedstock properties and (suboptimal) digestion conditions (Angelidaki & Ahring, 2000).

Previous incubation studies have shown high recalcitrance of biochar to degradation in soil (Spokas & Reicosky, 2009; Zimmerman, 2010). However, flash pyrolysis chars are known to display higher reactivity than those prepared by slow pyrolysis (Zhang et al., 2009). Expressed as percentage of C added, biochar amendment resulted in biochar C losses of 14% in loess and 6% in sandy soil during our experiment (78 days), which agrees with the findings of Bruun et al. (2012) for flash pyrolysis biochar.

The present study did not discriminate between evolution of biochar C (BC) and native soil organic carbon (SOC). Studies that separate BC degradation from the effect of biochar on SOC degradation by isotopic labeling (Kuzyakov et al., 2009), or by subtracting abiotic degradation of BC from measured C loss in soil (Spokas & Reicosky, 2009), show that degradation of BC itself is very low. The role of biochar as means for C sequestration, however, requires that SOC mineralization should not be enhanced instead. So far, the great diversity of biochars in a wide range of different circumstances has not conclusively settled this issue (e.g. Wardle et al., 2008; Singh et al., 2010), largely because of the lack of long-term data. Recently, an extensive incubation study by Zimmerman et al. (2011) led to the hypothesis that as biochars mature in soil, positive priming (i.e. enhanced SOC mineralization) declines and is dominated by negative priming (i.e. sorption of SOC onto biochar surfaces), leading to a net stabilization of biochar plus SOC.

The dissolved organic carbon (DOC) pool in soil is usually strongly and positively correlated with CO2 respiration (Haynes, 2005). Despite this, typically only 10–40% of DOC is observed to be readily degradable, probably owing to soluble humic substances that are relatively recalcitrant (ibid.). In our experiment, biochar amendment increased EOC (a measure of DOC) to the same extent as the other treatments (Fig. 3, day 2). However, the modest decline in EOC of biochar amended soils over the incubation period hints at a relative recalcitrance of extractable BC to degradation.

Nitrogen mineralization and nitrous oxide emissions

In accord with our second hypothesis, the proportion of soil inorganic N that originated from the residue was highest for manure, followed by the digestate and the biochar. Despite these significant differences, only the digestate resulted in increases of total inorganic N compared with nonamended soils (P < 0.05 in sandy soil). For manure amended soils, net mineralization (residue + soil) was equal to that of the control, whereas high amounts (57% and 79%, loess/sandy soil) derived from the residue. Conversely, biochar application did not prevent mineralization of native N, and mineralization of N from biochar, although small, was purely additive to soil N mineralization.

The evolution of N after digestate amendment exhibited a combination of both patterns. The apparent ‘suppression’ of soil N mineralization by manure and digestate amendment suggests commutability of sources of N for microbial assimilation. Immobilization as a fate for nitrogen upon residue amendment should also be considered, as microbial biomass N tended to be elevated in the manure treatments. The absence of significance between microbial biomass of the different treatments may be the result of a release of inorganic N by cell lysis or microbial osmoregulation following the rewetting and thawing events, which is supported by the high increases of inorganic N between day 29 and 84 (Davidson, 1992; Unger et al., 2010).

In our experiment, N evolution from the added residue was comparable between soils, whereas the contribution from the soil was consistently smaller in the sandy soil. The C : N ratio of the sandy soil is high compared with the loess soil (18 and 11, respectively). This could have increased the relevance of added residue N for total microbial N turnover (Booth et al., 2005). In addition, the loess soil exhibited patterns of higher microbial activity overall, as measured mineralization rates were always higher (for both C as N).

The influence of biochar amendment on soil N2O emissions is hitherto controversial. Several authors found that biochar soil application can significantly reduce N2O emissions (Spokas & Reicosky, 2009; Cayuela et al., 2010; Van Zwieten et al., 2010; Zhang et al., 2010). An increase in N2O production after biochar amendment has been also reported (Clough et al., 2010; Kammann et al., 2011). The mechanisms involved are still rather speculative and could include many biotic and abiotic factors. In our study, the most probable mechanism behind N2O production was denitrification in the loess soil. Probably the highest C availability in the manure led to higher denitrification rates than in the other treatments. Hence, emissions of N2O were reduced over the bioenergy sequence, which is in agreement with our third hypothesis.

There are not many studies reporting the effects of flash pyrolysis biochar on N2O emissions. In a recent study, Bruun et al. (2011) found that flash pyrolysis biochar from wheat straw increased N2O emissions from a loamy soil at high moisture conditions. The higher reactivity of flash pyrolysis biochar in comparison with slow pyrolysis biochar could explain the increase of N2O losses.

Nitrous oxide emissions from the sandy soil were negligible, which we relate to its higher aeration due to its texture and to its moisture level, which was kept lower (60% of WHC) than in the loess soil (70% of WHC). N2O emissions are enhanced in moist soils, as long as nitrification is not inhibited by oxygen limitations due to reduced aeration (Linn & Doran, 1984). However, in anoxic conditions the denitrifying community is known to cause N2O fluxes when nitrate availability is sufficient (Beare et al., 2009). In line with this, N2O emissions in loess declined during the drying phase, and subsequently peaked after rewetting to 80% WHC and after thawing the units.

Considerations for cattle manure management

Demographic and dietary developments globally cause an increase in animal production systems, generating a growing flow of secretion products that require disposal. During manure storage, transport and field application, odorous substances and GHGs are released to the environment, and cases of eutrophication are widespread (Oenema & Tamminga, 2005). Treatment of animal residues can considerably reduce the burden on the environment in these successive stages (Amon et al., 2006; Clemens et al., 2006; Bertora et al., 2008; Möller & Stinner, 2009; Kaparaju & Rintala, 2011). Certain treatments may yield energy, such as biogas in the case of anaerobic digestion, which results in further GHG abatement through reduced demands for fossil fuel (Cantrell et al., 2008; Holm-Nielsen et al., 2009). In the use of agricultural residues for bioenergy production, only the chemical energy present in reduced carbon compounds is of interest; other compounds may be used to restore the fertility of soils. However, to balance depletion of SOM by agricultural practices, it is also desired to return a significant amount of C to the soil (Lal, 2005). In principle, this study showed that soil amendment with bioenergy by-products could compensate for the SOM inputs no longer provided when manure amendment is aborted. The stability of SOM in biochar and digestate treatments were high when considering their reduced C : N ratios, and an increased N supply was observed in the case of digestate.

Biochar is rapidly gaining recognition as a soil amendment, improving the fertility of soil through a range of short- and long-term processes (Antal & Grönli, 2003; Joseph et al., 2010). Importantly, its recalcitrance is cause for the consideration of using biochar as a means for widespread carbon sequestration in soils. For most sources of organic matter, the scope for C sequestration is limited by the capacity of clay particles to stabilize SOM on the one hand (Six et al., 2002), and the balance of C and N inputs and outputs on the other (Schlesinger, 2000; Khan et al., 2007). Biochar has a much greater inherent stability (Masiello, 2004; Forbes et al., 2006; Laird et al., 2008), and modeled C balance studies have shown convincing net benefits (Gaunt & Lehmann, 2008; Woolf et al., 2010). The present study raises, however, some questions on its use. In our study, the biochar C yield was approximately 33%. The total C of the cattle manure feedstock was first reduced by 50% through anaerobic digestion, which means that the amount of biochar C is 17% from the initial quantity, of which 84–93% is recalcitrant. Therefore, the C sequestration potential of cattle manure via the digestion-pyrolysis route, as judged by the 1-year humification coefficient obtained in this study, was only 15% compared with 53–55% for untreated cattle manure. This underpins the importance of long-term behavior of residues in soil when evaluating the benefits of competing C pathways.

The results found in this laboratory experiment cannot be upscaled to field conditions. Hence, our results should be considered as a relative rather than an absolute approach. However, our study clearly shows that bioenergy production from manure has important implications on the recalcitrance of C and N in the by-products. Through N tracing, we demonstrated that the release of N in soil derived from the by-product decreases with anaerobic digestion and even more after pyrolysis, which has important parallel effects, such as the reduction of N2O emissions.

Acknowledgments

The authors are very grateful to Kealan Gell and Tania Fernandes for their valuable contribution in the anaerobic digestion experiment at the Department of Environmental Technology at WUR. Many thanks to Ali Imran and Gerrit Brem for the production of biochar at the University of Twente, to Gerard Ros for his assistance with the micro-diffusion procedure, to Gonzalo Gonzalez-Barberá for his valuable help with statistics and to Ron de Goede for supplying the labeled cattle manure. Maria Luz Cayuela was supported through a European Community Marie Curie Fellowship (Intra-European Fellowship for career development: FP7-PEOPLE-2007-2-1-IEF-Proposal No. 220868 – BEST) under which this study was performed. Thanks to two anonymous reviewers for their constructive comments.

Ancillary