Methanogenic community changes, and emissions of methane and other gases, during storage of acidified and untreated pig slurry




Acidification with concentrated H2SO4 is a novel strategy to reduce NH3 emissions from livestock slurry. It was recently found that also CH4 emissions from acidified slurry are reduced. This study investigated the microbiological basis and temporal stability of these effects.

Methods and Results

Pig slurry from two farms, acidified by different techniques or untreated, was stored for 83 days in a pilot-scale facility. Methanogens were characterized before and after storage by T-RFLP and qPCR targeting mcrA. Emissions of NH3 and CH4 during storage were quantified. Acidified slurry pH was nearly constant at values of 5·5 and 6·5. Ammonia losses were reduced by 84 and 49%, respectively, while CH4 emission with both acidification techniques was reduced by >90%. T-RFLP fingerprints showed little effect of acidification or storage time. A major T-RF of 105 bp could represent methanogens related to Thermoplasmata (Tp). No treatment effects on gene copy numbers were seen with universal methanogen primers, whereas effects were found with Tp-specific primers.


Methane emissions were reduced >90% during storage. Thermoplasmata-related methanogens could be involved in CH4 emissions from pig slurry.

Significance and Impact of the Study

The effect of acidification on CH4 emissions during storage of pig slurry was quantified for the first time. Acidification with sulphuric acid holds promise as a novel greenhouse gas mitigation strategy for confined livestock production.


Based on nitrogen flows, between 50 (national inventories) and 78% (RAINS database) of pig production within EU27 is based on liquid manure (slurry) management (Leip et al. 2010). Pig slurry is a mixture of excretal returns, water for washing and cooling, and bedding material, as determined by housing design and management. Concentrations of ammoniacal N, and hence the potential for loss of ammonia (NH3) via ventilation air or during storage, are typically high (Philippe et al. 2011), resulting in negative environmental impacts and loss of N for crop production (Sommer et al. 2006).

As the concentration of un-ionized, and therefore volatile, NH3 is a function of pH and temperature, lowering of pH can effectively reduce NH3 emissions (e.g., Panetta et al. 2005; Kai et al. 2008; Petersen et al. 2012) and improve N-use efficiency (Schils et al. 1999; Fangueiro et al. 2009). Effects of acidification on emissions of CH4 during storage, in contrast, were only recently documented (Ottosen et al. 2009; Eriksen et al. 2010; Petersen et al. 2012). In a short-term incubation, Ottosen et al. (2009) observed a 98% reduction in CH4 emissions from pig slurry acidified to 5·5–5·7 compared to untreated pig slurry. Petersen et al. (2012), in a 3-month laboratory study, found that adjustment of cattle slurry pH to 5 with sulphuric acid reduced CH4 emissions by 67–87%. Possible mechanisms of inhibition were discussed, but direct effects on the methanogenic community are unknown. Most methanogenic Archaea characterized so far are grouped within six taxonomic orders (Methanobacteriales, Methanococcales, Methanomicrobiales, Methanosarcinales, Methanopyrales and Methanocellales), but recent reports suggest that in several environments, methanogens related to the class Thermoplasmata are also important (Borrel et al. 2012; Dridi et al. 2012; Paul et al. 2012; Poulsen et al. 2013). Hydrogenotrophy predominates in well-characterized orders with the exception of Methanosarcinales, in which members of all three physiological groups (hydrogenotrophs, acetotrophs and methylotrophs) are well represented. In contrast, methylotrophy may be widespread among members of Thermoplasmata (Borrel et al. 2012; Dridi et al. 2012; Poulsen et al. 2013).

The source of methanogens in slurry is faecal matter. In ruminants, methanogenesis is traditionally considered to be dominated by hydrogenotrophs (Janssen and Kirs 2008), but it was recently shown that Thermoplasmata-related methanogens are also likely key players (Poulsen et al. 2013). Methanobacteriales spp. have been reported to predominate in pig faeces (Dridi et al. 2009; Mao et al. 2011), but already several years ago, uncultured Thermoplasmatales-like 16S clones isolated from pig waste storage pits were proposed to represent novel, at the time unidentified, methanogenic Archaea (Whitehead and Cotta 1999).

In this study, the effects of acidifying pig slurry with sulphuric acid on methanogens and CH4 emissions during long-term storage were investigated for the first time. Methanogenic communities in untreated and acidified slurry were characterized before and after storage. Based on previous observations, we hypothesized that acidification of pig slurry would reduce CH4 emissions during storage and that this would be accompanied by quantitative and/or qualitative changes in the community of methanogens.

Materials and methods

Slurry materials

Slurry materials were collected from two farms with finishing pigs in Western Denmark in early May 2011. Farm A had two pig houses with separate collection tanks; one of the houses had equipment for in-house acidification with sulphuric acid (96%). The use of acid on the farm was 10–11 kg t−1. Acidified slurry from Farm A was collected from below slatted floors in the pig house, that is, this material was <1 month old. Due to access constraints, untreated slurry could not be collected from inside the pig house without acidification, and instead untreated slurry was obtained from the outside storage tank after mixing, that is, this material ranged in age from fresh to >6 months. On Farm B, both untreated and acidified slurry came from the same batch of slurry which was collected during 3 weeks in April after the storage tank had been emptied for field application. Untreated slurry was sampled for the storage experiment and then pH of the remaining slurry was adjusted with 96% sulphuric acid during vigorous mixing; the recommended use of acid for this technique is 6 kg t−1 slurry.

Design of storage experiment

Storage of untreated and acidified slurry took place in the pilot-scale storage facility described by Petersen et al. (2009), which includes eight cylindrical 6·5-m3 units (2 m diam., 2 m height incl., 70 cm above-ground) constructed from 40 mm high-density polyethylene. Each unit has eight air inlets and a single outlet in a removable cover which is connected to a main ventilation duct. All eight stores were covered for the duration of this storage experiment and ventilated at a fixed rate of 138 m3 h−1. All settings and readings were logged every 15 min using LabView (National Instr.; Hørsholm, Denmark).

The four slurry materials (untreated and acidified slurry from Farm A and Farm B) were stored in each of two randomized blocks. The storage experiment lasted from 19 May to 10 August 2011, a total of 83 days. On 6 July, all slurries were mixed to break up surface crusts which had formed.

Slurry analyses

At the beginning and end of the experimental storage period, duplicate samples of slurry were collected from each storage unit after thorough mixing. These samples were subsampled for determination of dry matter (DM) and volatile solids (VS), total N (TN) and total ammoniacal N (TAN), electrical conductivity (EC) and pH (Table 1). Other subsamples were taken for analysis of methanogens (see below). Slurry pH was determined with a Sentron 3001 pH-meter (Roden, the Netherlands) and electrical conductivity with a Radiometer conductivity meter (Copenhagen, Denmark). Dry-matter content was determined after drying of 10 g subsamples for 24 h at 105°C, and volatile solids after an additional 4 h at 550°C. Total N and ammoniacal N were determined by Kjeldahl digestion (Kjeltec 1030, Foss Analytical AB; Höganäs, Sweden). Slurry temperature at 5 cm depth was determined at each sampling.

Table 1. Selected properties of acidified and untreated slurry from two different pig farms on 18 May (day 0) and 11 August (day 83) of the pilot-scale storage experiment. Acidification had taken place on the farm using two different technologies, either in-house (Farm A) or in an outside collection tank (Farm B). Duplicate samples were taken from each of the two replicate pilot-scale stores; mean analytical results were used to calculate treatment means
dS m−1kg Mg−1
  1. a

    EC: electrical conductivity, DM: dry matter, VS: volatile solids, TN: total nitrogen, TAN: total ammoniacal nitrogen.

  2. b

    *P < 0·05; **P < 0·01; ***P < 0·001.

Farm A18 MayAcidified5·640·175·454·86·54·7
11 AugAcidified5·994·572·651·96·65·1
Farm B18 MayAcidified6·629·137·624·64·73·6
11 AugAcidified6·866·63925·24·73·5
Farm (F)b  ******************
Treatment (T)  ******************
Date (D)  ******    
F × T  ******************
F × D  ****    
T × D   **    
F × T × D  ****    

DNA extraction

DNA was extracted using a standard phenol : chloroform : isoamyl alcohol-based protocol adapted from Griffiths et al. (2000). Slurry samples (two replicates for each storage tank) of 250 μl were transferred to 2-mL ceramic bead tubes (MoBio; Carlsbad, CA). Phenol : chloroform:isoamyl alcohol (500 μl, pH 8·0) and CTAB extraction buffer (500 μl) were added, and the samples beadbeaten for 2 × 20 s at room temperature, with 60 s on ice in between. The samples were centrifuged (5 min at 16 000 g, 4°C) and the supernatants (300–325 μl) transferred to new 2-ml tubes on ice. Phenol was removed by secondary addition of 300–325 μl chloroform : isoamyl alcohol (24 : 1), followed by centrifugation (5 min at 16 000 g, 4°C) and transfer of the aqueous phase (~300 μl) to new tubes. Volumes of 600–650 μl 30% PEG and 1 μl glycogen were added to the samples and the tubes incubated on ice for 2 h. The samples were centrifuged (30 min, 14 600 rev min−1, 4°C), the supernatant removed and the pellet re-suspended in 100 μl DPEC-treated dH2O. The samples were stored on ice until DNA purification was performed using Nucleospin Extract II (Macherey-Nagel; Düren, Germany), according to the manufacturer's protocol, with the exception that the final elution of DNA was done using 35 μl elution buffer. The DNA quality and quantity of each sample was checked by agarose gel electrophoresis (3 μl DNA + 2 μl loading buffer) and Nanodrop® ND-1000 spectrophotometry (Thermo Fisher Scientific; Wilmington, DE), respectively. DNA extracts were stored at −20°C until further analysis.

Terminal-restriction fragment length polymorphism (T-RFLP) analysis

T-RFLP analysis, targeting the mcrA gene of potentially all methanogens, was performed using a forward primer (5′-GGT GGT GTM GGD TTY ACH CAR TA), modified from the mlas primer of Steinberg and Regan (2008), and the mcrA-rev reverse primer (5′-FAM-CGT TCA TBG CGT AGT TVG GRT AGT; 30). The PCR master mix consisted of 2 μl of each primer (stock conc. 10 pmol μl−1), 2 μl dNTP (5 nmol μl−1), 0·5 μl Dynazyme (2U μl−1), 5 μl DyNazyme 10X buffer, 36·5 μl dH2O and 2 μl DNA extract as template. The samples were PCR amplified on a Doppio thermo-cycler (VWR Int.; Radnor, PA) using the following protocol: an initial denaturation step at 95°C for 3 min, followed by five cycles of denaturation at 95°C for 30 s, annealing at 48°C for 45 s and extension at 72°C for 30 s, with a ramp rate of 0·1°C s−1 from the annealing to the extension temperature. These initial five cycles were followed by 30 cycles of denaturation at 95°C for 30 s, annealing at 55°C for 45 s and extension at 72°C for 30 s, followed by a final extension step at 72°C for 10 min. The approx. 470-bp PCR amplicons were verified by gel electrophoresis and purified by QIAquick® PCR Purification kit (Qiagen GmbH; Hilden, Germany) according to the manufacturer's manual. From each sample, a total of 0·2 μg purified PCR amplicon, quantified using the Nanodrop® ND-1000 spectrophotometer (Thermo Fisher Scientific; Wilmington, DE), was digested by restriction endonuclease TaqI (New England Biolabs; Ipswich, MA). Digested fragments were precipitated using 3 mol l−1 sodium acetate and 96% ice-cold ethanol, washed with 70% ice-cold ethanol, vacuum dried and re-dissolved in 2 μl ddH2O. Each sample was mixed with 10 μl formamide and 0·2 μl MegaBACETM ET900-R size standard (GE Healthcare; Princeton, NJ), and analysed on an ABI 3730XL Capillary Sequencer (ABI by Life Technologies; Carlsbad, CA) equipped with a 3730xl DNA analyzer capillary array (length: 36 cm; O.D.: 150 μm; I.D.: 50 μm) and using the POP-7 polymer (run time: 35 min; run voltage: 8·5 kV). The electropherograms were analysed using the ABI Peak ScannerTM Software ( Archaea genotypes (genera and orders) representing the terminal-restriction fragments (T-RFs) observed were tentatively identified by in silico TaqI enzymatic digests of a collection of approx. 300 mcrA gene sequences (uncultured clones excluded) obtained from NCBI GenBank and The Functional Gene Pipeline & Repository (, and using the Microbial Community Analysis (MiCA 3) software (

Quantitative PCR (qPCR)

Quantification of methanogen mcrA gene copies was made using a Roche Lightcycler 480 (Roche; Mannheim, Germany) and applying either the universal methanogen primer set used for the T-RFLP analysis (see previous section) or a second primer set, referred to as Thermoplasmata primers (Tp-mcrA-F: 5′-GAY RAC ATC CTB GAR GAY TA-3′; Tp-mcrA-R: 5′-RTC GWA WCC RTA GAA TCC GAG-3′), designed specifically to target mcrA genes of Thermoplasmata-related methanogens. Specificity of the Tp primers was verified through cloning and sequencing of PCR-amplified fragments. No amplification of nontarget fragments was detected. Each reaction mixture (20 μl) consisted of 10 μl of SYBR® Green I Master Mix (Roche; Mannheim, Germany), 1 μl of each of the forward and reverse primers (10 pmol μl−1), 7 μl of ddH2O and 1 μl of DNA template at a concentration of approx. 10 ng μl−1 (DNA concentration in crude DNA extracts was quantified using a NanoDrop-1000, Thermo Fisher Scientific, Wilmington, DE). Running conditions were 95°C for 3·5 min, followed by 40 cycles of 92°C for 30 s, 55°C for 45 s and 72°C for 30 s. Acquisition of fluorescent signal was done at 83°C (15 s) after each cycle. Product specificity and size was confirmed by melting point analysis and by agarose gel electro-phoresis, respectively. Prior to the qPCR, fragments of the mcrA gene were amplified using the universal and the Thermoplasmata-specific primer set, and the relevant PCR programme on a conventional Doppio thermocycler. Fragments were purified as previously described and cloned into Escherichia coli JM109 competent cells using the pGEM®-T Easy Vector System (Promega; Madison, WI). Plasmids were extracted using the Plasmid Mini Kit (Qiagen GmbH; Hilden, Germany), and mcrA gene inserts were verified through sequencing (Eurofins MWG Operon; Eberberg, Germany). Extracted plasmids with mcrA gene inserts were linearized using EcoRV (New England Biolabs Inc., Ipswich, MA) and used to establish the standard curves.

Emission measurements

For the duration of the storage experiment, ventilation air from each store was sampled at 15 ml min−1 using a peristaltic pump. Ammonia emissions were measured continuously using acid traps with 80 ml 20 mmol l−1 H3PO4 that were placed in each gas sampling line near the point of subsampling. Acid traps were replaced at approximately weekly intervals. For analysis, the volume was adjusted to 100 ml and concentrations of NH4+ determined colorimetrically (Keeney and Nelson 1982).

Emissions of CH4 and N2O were measured during 24-h periods at approximately weekly intervals. Via a manifold with three-way solenoid valves, subsamples of ventilation air from each store were taken every 5 min and pooled in 3-L Tedlar gas sampling bags (SKC Ltd., Dorset, UK). Gas samples were analysed using an Agilent 7890 gas chromatograph with CTC CombiPal autosampler. It had a 2-m backflushed precolumn with Hayesep P connected to a 2-m main column with Poropak Q which, in turn, was connected to a four-port valve diverting the gas stream to either an electron capture detector (ECD) for N2O analysis or a flame ionization detector (FID) for CH4 analysis. The carrier gas was N2 at a flow rate of 45 ml min−1. For the ECD, Ar-CH4 (95% of 5%) at 40 ml min−1 was used as make-up gas. The FID was supplied with 45 ml min−1 H2, 450 ml min−1 air and 20 ml min−1 N2. Temperatures of injection port, columns, ECD and FID were 80, 80, 325 and 200°C, respectively.

As a proxy for odour emissions, concentrations of total reduced S were determined directly in the headspace of the ventilated stores on 12 occasions during the 83-day storage period. Total reduced sulphur (TRS) was measured with a portable Jerome Hydrogen Sulfide Analyzer (Model 631-X, Arizona Instruments; Phoenix, AZ) which had a detection limit of 0·3 ppm at the operating range selected. The instrument has a response of 100% to H2S and of 0–45% to 11 other reduced sulphur gasses (Koelsch et al. 2004); the overall response is expressed in H2S equivalents.

Data analysis

The storage experiment included four stores with acidified slurry and four with untreated slurry, but only two true replicates per treatment as pig slurry came from two different pig farms and different stages of manure management. With an expected variability in emissions of around 20%, as judged from the previous study with cattle slurry (Petersen et al. 2012), treatment effects of 60% or higher would be significant at P < 0·05 (Mead et al. 1993).

The T-RFLP Analysis Expedited (T-REX) software was used for the processing and analysis of T-RFLP data (Culman et al. 2009). Individual T-RFs detected in pig slurry from Farms A and B, respectively, were compared across treatments (untreated, acidified) and storage times (before, after) by two-way anova; table-wise error rates were controlled using the sequential Bonferroni approach (Holm 1979). To further analyse the T-RFLP data, the additive main effects and multiplicative interaction model (AMMI) of T-REX, also known as doubly-centred PCA (Culman et al. 2009), was applied. The model uses anova to first partition the variation into main effects and interactions and then applies PCA to the interactions to create interaction principal components axes (IPCAs). T-REX interfaces with matmodel 3.0 (Gauch 2007) to run the AMMI analysis.

Slurry characteristics, as well as qPCR results for mcrA gene copy numbers (log-transformed), were compared across farms, treatments and storage times with a mixed model that included two- and three-way interactions, and with block in the pilot-scale storage facility as random effect. These analyses were conducted with sas 9.2. Cumulated emissions were estimated assuming rates changed linearly between sampling days. Treatment effects were identified by one-way anova and Tukey's HSD tests.


Storage environment

With permanent covers and forced ventilation, temperature was the main environmental variable to influence storage conditions (Fig. 1). Daily mean air temperature ranged from 9·5 to 20·2°C, with an average of 14·8°C. Slurry temperature at 5 cm depth followed the seasonal trends in air temperature and remained within 15–17°C during the last several weeks of the storage experiment.

Figure 1.

Air temperature during the period 15 May to 15 August 2011 (line), and slurry temperature measured at 5 cm depth in the liquid phase (open circles).

Slurry characteristics

Slurry materials were characterized at the beginning and end of storage (Table 1). The acidified pig slurry from Farm A, collected directly from a slurry pit inside the pig house, was more concentrated than the other three materials, as indicated by DM and VS, but also EC and N content. Some crust formation was observed with both untreated and acidified slurries within 2 weeks, most pronounced in the acidified slurry from Farm A which had the highest DM content (Table 1). Changes in TN during the 83-day storage period were not significant, but declining trends were indicated for untreated slurry which would correspond to losses of 1·8 and 0·5 kg N for Farm A and Farm B, respectively.

Acidified slurry from Farm A had a pH of 5·5 at the start of the storage experiment, a reduction of 2 pH units compared to the untreated slurry. At Farm B, the single adjustment of slurry pH in the on-farm storage tank had only lowered slurry pH by 0·5–1 unit (Table 1). The pH of both untreated and acidified slurry materials remained largely constant during the 3-month storage period (Fig. 2a–b).

Figure 2.

Untreated (open circles) and acidified pig slurry (closed circles) from two different farms with, respectively, in-house acidification (Farm A) and in-store acidification (Farm B) was stored for 83 days. The plots show the temporal dynamics of pH (a, b), NH3 emissions (c, d) and CH4 emissions (e, f) in untreated and acidified slurry. The data represent mean ± standard error (= 2).

T-RFLP analysis of methanogens

Based on relative abundance, the three dominating amplicons in pig slurry from Farm A were those of 57, 105 and 154 bp (Table 2). Similarly, three T-RFs accounted for 75% of the Farm B amplicons, namely 105, 154 and 387 bp, with a clear predominance of the 105 bp T-RF (Table 3). For Farm A, the T-RFLP fingerprints of untreated slurry (from the main storage) and acidified slurry (from the slurry pit) were clearly different: the relative abundance of T-RF 105 bp in acidified slurry was 10-fold lower than in the untreated slurry, and lower abundances were also observed for T-RFs 65, 184 and 438 bp, whereas the relative abundances of T-RFs 57, 90 and 158 bp were higher in acidified slurry. For Farm B, the amplicon of 105 bp was prominent in the T-RFLP fingerprint of both acidified and untreated slurry originating from the same batch of slurry. This suggests that either this population of methanogens was excreted by the pigs or it emerged during the first few weeks of storage. The T-RF of 105 bp could represent Thermoplasmata-related methanogens, but also different members of Methanosarcinaceae, whereas T-RF 154 and 387 bp could represent members of the Methanomicrobiales order. The in silico analysis did not identify sequences representing the T-RF 57 bp.

Table 2. Relative abundancea of mcrA terminal-restriction fragments (T-RFs) in slurry from Farm A
T-RFbAcidifiedUntreatedEffectscMethanogen genera and ordersd potentially represented by the T-RFse
bp18 May 201111 Aug 201118 May 201111 Aug 2011 T D T × D
  1. a

    Per cent of total peak height for each sample. Numbers represent mean values of duplicates.

  2. b

    Only T-RFs of >3·0% relative abundance for at least one treatment in one of the farms are included.

  3. c

    Effects of treatment (T) and sampling date (D). *P < 0·05; **P < 0·01.

  4. d

    Taxonomic orders: Methanobacteriales (Mb), Methanococcales (Mc), Methanomicrobiales (Mm), Methanosarcinales (Ms), Methanopyrales (Mp), Methanocellales (Mce), Thermoplasmata-related (Tp).

  5. e

    Identified by in silico analysis of database mcrA gene sequences.

  6. f

    The listed genera are represented in T-RFs of 471 bp or larger.

393·50·42·84·2   Methanobrevibacter (Mb), Methanopyrus (Mp), Methanosaeta (Ms), Methanomassiliicoccus (Tp)
650·10·17·212·6*  Methanogenium (Mm)
1053·82·940·624·4*  Methanococcoides (Ms), Methanohalophilus (Ms), Methanosarcina (Ms), Cand. Methanomethylophilus (Tp), Tp
15426·126·423·516·8   Methanoculleus (Mm), Methanomicrobium (Mm), Methanoplanus (Mm)
1840·50·23·41·3*** Methanocella (Mce)
2973·23·83·81·2   Methanohalophilus (Ms), Methanosarcina (Ms), Tp
3873·12·13·14·2   Methanocorpusculum (Mm), Methanofolis (Mm), Methanogemium (Mm), Methanospirilum (Mm)
4380·60·21·92·5**  Methanoplanus (Mm)
4711·61·41·96·1*****Methanobrevibacter (Mb), Methanocella (Mce), Methanocaldococcus (Mc), Methanococcus (Mc), Methanocorpusculum (Mm), Methanosaeta (Ms), Methanosalsum (Ms), Methanosphaera (Mb), Methanonothermococcus (Mc), Methanotorris (Mc)f
Table 3. Relative abundancea of mcrA terminal-restriction fragments (T-RFs) in slurry from Farm B
T-RFbAcidifiedUntreatedEffectscMethanogen genera and ordersd potentially represented by the T-RFse
bp18 May 201111 Aug 201118 May 201111 Aug 2011 T D T × D
  1. a

    Per cent of total peak height for each sample. Numbers represent mean values of duplicates.

  2. b

    Only T-RFs of >3% relative abundance for at least one treatment in one of the farms are included.

  3. c

    Effects of treatment (T) and sampling date (D). *P < 0·05.

  4. d

    Taxonomic orders: Methanobacteriales (Mb), Methanococcales (Mc), Methanomicrobiales (Mm), Methanosarcinales (Ms), Methanopyrales (Mp), Methanocellales (Mce), Thermoplasmata-related (Tp).

  5. e

    Identified by in silico analysis of database mcrA gene sequences.

  6. f

    The listed genera are represented in T-RFs of 471 bp or larger.

391·11·00·92·5   Methanobrevibacter (Mb), Methanopyrus (Mp), Methanosaeta (Ms), Methanomassiliicoccus (Tp)
651·74·31·87·0 * Methanogenium (Mm)
10555·656·656·339·7   Methanococcoides (Ms), Methanohalophilus (Ms), Methanosarcina (Ms), Cand. Methanomethylophiius (Tp), Tp
1549·68·06·010·9   Methanoculleus (Mm), Methanomicrobium (Mm), Methanoplanus (Mm)
1841·91·41·33·5   Methanocella (Mce)
2972·61·41·92·3   Methanohalophilus (Ms), Methanosarcina (Ms), Tp
38712·814·214·212·2   Methanocorpusculum (Mm), Methanofolis (Mm), Methanogemium (Mm), Methanospirilum (Mm)
4383·64·54·77·1   Methanoplanus (Mm)
4710·1ndnd0·2   Methanobrevibacter (Mb), Methanocella (Mce), Methanocaldococcus (Mc), Methanococcus (Mc), Methanocorpusculum (Mm), Methanosaeta (Ms), Methanosalsum (Ms), Methanosphaera (Mb)f

qPCR of methanogens

The mcrA gene copy numbers determined with universal primers showed significant effects of farm, treatment and date (Table 4). Gene copy numbers were also determined using the Tp-mcrA primer set assumed to be specific for Thermoplasmata-related methanogens. The Tp primer set had been constructed based on Thermoplasmatales-related sequences from humans and animals available at the time (2012). It was checked for mismatches against species of Methanobrevibacter, Methanosarcina, Methanocella, Methanosphera, Methanomicrobium and Methanolinea. So far, all sequenced mcrA clones (>75) retrieved from a variety of animals with this primer set have been Tp-positive. Gene copy numbers with Tp-mcrA primers were lower than with universal primers, as expected if representing a subset of the methanogen community. It is notable that the Tp primer-based gene copy number in acidified slurry from Farm A was 25 times lower than in untreated slurry from the same farm, and in slurry from Farm B. There was a significant correlation (r = 0·77, P < 0·05) between the relative abundance of the T-RF of 105 bp and Tp-mcrA gene copy numbers.

Table 4. Levels of methyl–coenzyme M reductase A (mcrA) genes as determined by quantitative PCR (qPCR) using universal primers or primers targeting Thermoplasmata-related methanogens (Tp). Numbers represent average ± SE (= 2)
 mcrA gene copies per μg DNA (log10)Tp vs universal (%)
Universal primersTp primers
  1. *P < 0·05; **P < 0·01; ***P < 0·001.

  2. a

    One replicate lost.

  3. b

    0·05 < P < 0·1.

Farm A
18 May 20118·39 (0·11)5·82 (0·06)0·3 (0·1)
11 Aug 20118·27 (0·15)5·86 (0·79)0·9 (0·8)
18 May 20118·05 (0·05)7·21 (0·31)16 (6·3)
11 Aug 20117·96 (0·089)6·2 (0·24)2 (1·0)
Farm B
18 May 20118·01a7·26 (0·04)19
11 Aug 20117·72 (0·36)7·31 (0·21)56 (41)
18 May 20118·15 (0·43)7·43 (0·2)21 (7·6)
11 Aug 20117·97 (0·03)7·56 (0·09)40 (7·4)
Farm (F)
Treatment (T)********
Date (D)*** 
F × T*  b
F × D * 
T × D   b
F × T × D**  

In untreated slurry from Farm A, the proportion of Tp-derived gene copies (Table 4) declined on average from 16 to 2% during the storage experiment, while the proportion in acidified slurry from Farm A remained low at <1%; these trends, though not statistically significant (0·05 < P < 0·1), were in accordance with observed CH4 emission rates (see below). For Farm B, the proportions of Tp-derived gene copies increased in both untreated and acidified slurry.

Ammonia emissions

As expected, emissions of NH3 were effectively reduced by the acidification to pH <6 of slurry from Farm A (Fig. 2c–d), and less so with slurry from Farm B where pH of both acidified and untreated slurry was closer to neutrality. Mixing of slurry stores in mid-July only briefly stimulated NH3 emissions, showing that crust formation was not responsible for the reduction in NH3 emissions in acidified slurry. The reduction in NH3 loss due to acidification corresponded to 1·6 and 0·7 kg N for Farm A and Farm B, respectively, during the 83-day storage period, in accordance with the higher recovery of TN in acidified slurry.

Methane emissions

There was a dramatic reduction in CH4 emissions with both acidification techniques (Fig. 2e–f). Cumulated CH4 emissions were reduced by 99 and 94% in slurry treated by in-house (Farm A) and in-store acidification (Farm B), and there were no indications that inhibition was alleviated by the end of the 83-day storage period. Hence, the effects of acidification on CH4 emissions were robust. The time course of CH4 emissions from untreated slurry from Farm A and Farm B was very different, that is, the slurry obtained from a long-term storage tank at Farm A had initially high CH4 emission rates which dropped rather abruptly, whereas emissions from Farm B were initially low, but then accelerated during July.

N2O and TRS emissions

In most storage units, a crust developed that can support growth of nitrifying and denitrifying bacteria, resulting in emissions of N2O (e.g., Sommer et al. 2000; Petersen et al. 2013). In the present study, no N2O emissions were observed, possibly because surface crusts were destroyed by the mixing before populations had developed.

Emissions of H2S are evidence that sulphur transformations, including sulphate reduction, take place in the slurry. Total reduced S (mainly H2S) emissions from slurries collected at Farm A remained low, but with consistently lower emissions from acidified slurry (Fig. 3). In contrast, H2S emissions from both fresh and acidified pig slurry collected at Farm B increased during the storage experiment, initially with higher emissions from acidified slurry. A brief release of H2S was observed immediately after mixing of the slurries on 6 July, indicating that some H2S was trapped below the surface crust.

Figure 3.

Temporal dynamics of total reduced S (mainly H2S) in untreated (open circles) and acidified slurry (closed circles). The data represent mean ± standard error (= 2).


Comparing effects of management practices on manure microbiology is complicated by site-specific differences in management and storage conditions (Hensen et al. 2006; Sneath et al. 2006). A pilot-scale storage facility, as used in this study, enables side-by-side comparison of treatment and management practices under well-defined and semi-realistic conditions. Slurry pH levels were temporally stable, also after mixing of slurries in July, indicating that pH adjustment of pig slurry under practical storage conditions will be effective for several months.

Methanogens in acidified and untreated slurries

The high abundance of a few T-RFs indicated that methanogenic communities were dominated by relatively few species/genera, as also reported in other studies (Whitehead and Cotta 1999; Zhu et al. 2011; Cardinali-Rezende et al. 2012). Still, the total number of T-RFs in each fingerprint suggested that species richness of the methanogenic community was significant. One of the prominent T-RFs (154 bp) was tentatively associated with Methanoculleus spp., confirming observations by Whitehead and Cotta (1999) and Hook et al. (2009) in studies of stored swine slurry. This T-RF, however, did not show any response to acidification in the present study, although it should be emphasized that community analyses were conducted at DNA level which may not reveal dynamics in activity.

Virtually no effects of acidification or storage time were seen in slurry from Farm B (Table 3, Fig. 4). In contrast, slurry from Farm A had very different T-RFLP fingerprints of untreated and acidified slurry, and reduction during storage (though not statistically significant) was suggested for the concentration of one of the major T-RFs, that is 105 bp. Possible identifications for this T-RF are Methanosarcinales spp. (Table 2 and 3), as well as Thermoplasmata-related methanogens. In a previous storage experiment with pig slurry (unpublished data), a mcrA gene clone library had been prepared in which Thermoplasmata-related sequences were the only ones with T-RFs of 105 bp, that is, no Methanosarcinales-related sequences represented a T-RF of this size. Thus, T-RF 105 bp could be a marker for Thermoplasmata in pig slurry. It is notable that the T-RF of 105 bp was significantly related to Tp-mcrA gene copy number.

Figure 4.

Principal component analysis (PCA) of the T-RFLP data. Data represent slurry samples taken from the pilot-scale facility (two replicates per treatment) on 18 May 2011 (open symbols) and 11 August 2011 (closed symbols). The slurry from Farm A was acidified in the slurry pit below slatted floors, while untreated slurry was from the storage tank of a separate production line; slurry from Farm B was acidified in a storage tank following removal of a batch (untreated slurry) for the experiment. For further details, see 'Materials and methods'. (○●) Farm A, acidified; (▵▲) Farm A, reference; (□■) Farm B, acidified; (▿▼) Farm B, reference.

Thermoplasmata-related methanogens could represent novel methanogenic Archaea of the suggested seventh taxonomic order of methanogens, Methanoplasmatales (Dridi et al. 2012; Paul et al. 2012; Poulsen et al. 2013). So far, conclusions are based mainly on investigations of uncultured clones and metagenomes, and only one Thermoplasmata-related methanogen, Methanomassiliicoccus luminyensis, has been isolated and characterized (Dridi et al. 2012), which makes precise taxonomic grouping difficult. From a physiological point of view, there are strong indications that Thermoplasmata-related genomes harbour genes for at least methylotrophic and hydrogenotrophic methanogenesis (Paul et al. 2012; Poulsen et al. 2013).

Quantitative changes

Changes in mcrA gene copy numbers during storage were small or insignificant despite highly variable CH4 emission rates. Neither have previous studies found any clear relationships between methanogenesis and methanogen numbers as determined by qPCR targeting either 16S rRNA or mcrA genes (Hook et al. 2011; Barret et al. 2012). If only a subset of the methanogens was actually responsible for the observed fluctuations in emissions, then dynamics of this group may not be detectable in the overall gene pool. Tp-derived gene copy numbers showed stronger dynamics, and these were in most cases consistent with changes in CH4 emission (Fig. 2). It suggests a link between Thermoplasmata-related methanogens and CH4 emissions, but this is complicated by the fact that high Tp-mcrA gene copy numbers were found in both acidified and untreated slurry from Farm B despite a > 90% reduction in CH4 emissions in acidified slurry (Table 5). The nature of inhibition of methanogens at pH 5·5 (Farm A) and near-neutral pH (Farm B) may differ. At near-neutral pH, successful competition for H2 by sulphate-reducing bacteria in the sulphate-rich environment of acidified slurry may account for the reduction in CH4 emissions (Uberoi and Bhattacharya 1995). According to Chen et al. (2008), methanogens are sensitive to hydrogen sulphide rather than total sulphide, and therefore the potential for toxic effects was much higher in slurry from Farm A acidified to pH 5·5.

Table 5. Cumulated fluxes of CH4, N2O and NH3, as well as total GHG emissions expressed as CO2 equivalents, in the acidification experiment (83 days). The different slurry materials were stored with a cover and active ventilation for the duration of the experiment; NH3 was measured continuously, gas samples for GHG analyses only during 24-h periods
g m−3g m−2g m−2kg CO2 eq m−3
  1. BD: below detection limit.

Farm A
Untreated5181 715·9131·8
Farm B
Acidified377 247·010·1
Untreated6720 507·8169·5

Emissions during storage

Acidifying pig slurry from Farm A to pH 5·5 gave a reduction in accumulated NH3 emissions of 84%, while the adjustment of pH in pig slurry from Farm B to 6·5 gave a 49% reduction in NH3 losses (Fig. 2). The moderate pH adjustment at farm B could be due to insufficient mixing and degassing of CO2 during pH adjustment (Ni et al. 2009; Dai and Blanes-Vidal 2013). The dramatic reduction in CH4 emissions corroborates a previous study where acidification of fresh and aged cattle slurry reduced CH4 emissions by 67 and 87%, respectively (Petersen et al. 2012). In that study, it was also determined that CH4 emissions from aged slurry were reduced by 78% due to the pH adjustment alone (adding HCl to avoid sulphate), by 48% due to the sulphate amendment (adding K2SO4 to avoid acidification) and by 69% when adding methionine (to avoid stimulating sulphate-reducing bacteria). This confirms that the inhibition of methanogenesis involves different mechanisms, and it helps explain the effectiveness and long-term stability of inhibition over a wide pH range, which may also involve toxic effects of protonated volatile fatty acids (VFA; Ottosen et al. 2009).

Where CH4 emissions from acidified slurry remained low throughout the 83-day storage period, the untreated slurries from Farms A and B showed very different temporal dynamics with respect to CH4 emissions. Fresh excreta have a low capacity for methanogenesis unless inoculated by a microbial consortium adapted to the slurry environment (Zeeman et al. 1988; Sommer et al. 2007). As untreated slurry from Farm A was a mixture of material collected during autumn, winter and spring, it probably contained an adapted methanogenic community that could support CH4 emissions from the beginning of the experimental storage. The decline after 5–6 weeks most likely reflected a depletion of degradable organic matter (Petersen et al. 2013). In view of the minor changes in methanogen communities during storage (Tables 2-4), the dramatic increase in CH4 emissions observed with slurry from Farm B around mid-July could not be explained by growth of methanogens. Massé et al. (2008) observed a 6-month delay in CH4 emissions from a cattle slurry until high, and presumably inhibitory, concentrations of VFA started to decline. We did not measure slurry VFA content in the present study, but inhibitory concentrations could have delayed CH4 emissions from the less than 1-month-old slurry from Farm B during the first several weeks of storage.

Total reduced S emissions did not suggest a dramatic effect of slurry acidification on odour emissions from pig slurry during storage. A moderate reduction in pH may result in higher emissions of methanethiol which has an odour threshold one order of magnitude below H2S (Devos et al. 1990), and attention to the effect of acidification on slurry odour characteristics is thus warranted.

Acidification as GHG mitigation strategy

Several storage experiments have demonstrated that CH4 emissions during summer storage dominate the greenhouse gas (GHG) balance of liquid manure management (Clemens et al. 2006; VanderZaag et al. 2009; Petersen et al. 2013). The overall GHG balance for each slurry material is presented in Table 5, including NH3 as an indirect source of N2O. Direct N2O emissions during storage were not observed in this study. Total GHG balances were also in this study dominated by CH4, and overall GHG mitigation due to acidification was 98·5 and 94% with slurry from Farms A and B, respectively. Evidently, acidification at an early stage of manure handling is needed for effective GHG mitigation. The reduction in NH3 emissions during storage can then also lead to higher N-use efficiency (Webb et al. 2013).

With the achieved reduction in CH4 emissions from pig slurry of more than 90% over nearly 3 months using two different commercial technologies for on-farm acidification, this study provides strong evidence that slurry acidification is an effective GHG mitigation option that also conserves ammoniacal N for crop production. Molecular analyses of methanogenic communities in acidified and untreated slurries indicated that taxonomic changes during storage were moderate. However, a 105-bp T-RF was significantly suppressed in slurry acidified to pH 5·5, but not at pH 6·5, and this T-RF declined in untreated pig slurry where substrates for methanogenesis had been depleted. Results from qPCR suggested that this T-RF belonged to Thermoplasmata-related methanogenic Archaea, but this needs direct confirmation. More studies are also needed to verify a causal relationship between methanogens represented by this T-RF and emissions of CH4 during slurry storage.


The technical assistance of Claudia Nagy, Bodil Steensgaard, Karin Durup, Karin Dyrberg, Morten Skov and Arne Grud is greatly appreciated. This study was supported by the Danish Ministry of Food, Agriculture and Fisheries via the research programme Livestock Production for the Future.

Conflict of Interest

No conflict of interest declared.