Introduction
Reducing greenhouse gas emissions from agricultural systems is vital in developing sustainable food production practices (Gagnon et al., Reference Gagnon, Ziadi, Rochette, Chantigny, Angers, Bertrand and Smith2016). Intensively managed grassland systems recycle considerable amounts of nitrogen (N) (Deenen, Reference Deenen1994) and retain inherent amounts of labile organic carbon (C) (Baggs et al., Reference Baggs, Rees, Smith and Vinten2000; Jérôme et al., Reference Jérôme, Beckers, Bodson, Heinesch, Moureaux and Aubinet2014). Both N and C are mainly sourced from fertilizer, excreta and mineralization of soil organic matter (OM) (Franzluebbers et al., Reference Franzluebbers, Stuedemann, Schomberg and Wilkinson2000; Soussana and Lemaire, Reference Soussana and Lemaire2014). Globally, grasslands cover about 26% of the ice-free areas (Steinfeld et al., Reference Steinfeld, Gerber, Wassenaar, Castel, Rosales and De Haan2009). Because of their geographical positioning, grasslands are often subjected to high soil moisture conditions, which reduces soil aerobicity, thus increasing the potential for denitrification (Jarvis et al., Reference Jarvis, Hatch, Pain and Klarenbeek1994; Steinfeld et al., Reference Steinfeld, Gerber, Wassenaar, Castel, Rosales and De Haan2009). Denitrification is a bacterially mediated process whereby nitrate (NO3−) is transformed to NO2−, NO, N2O and finally to N2, under limited oxygen (O2) as most denitrifying bacteria are facultative anaerobes (Robertson and Groffman, Reference Robertson, Groffman and Paul2007). Most denitrifying bacteria couple NO3− reduction with organic C oxidation to gain energy, making a supply of readily available C a usual requirement for denitrification to occur, a process which further produces CO2 (Knowles, Reference Knowles1982; Beauchamp et al., Reference Beauchamp, Trevors, Paul and Stewart1989).
Livestock manures are returned to the land to recycle nutrients for plant growth (Watson et al., Reference Watson, Atkinson, Gosling, Jackson and Rayns2002; Tittonell et al., Reference Tittonell, Rufino, Janssen and Giller2010). Studies on the effects of manure applications on the N and C cycles are found in the literature, especially related to soil processes such as denitrification (Morley and Baggs, Reference Morley and Baggs2010; Soussana and Lemaire, Reference Soussana and Lemaire2014). The source and concentration of C in relation to NO3− and O2 have been documented to control denitrification rates (Beauchamp et al., Reference Beauchamp, Trevors, Paul and Stewart1989). Early studies demonstrated the effect of cattle manure applications on denitrification rates and explanations for the resulting increase were found to be related to the appearance of anaerobic microsites for the denitrification process (Guenzi et al., Reference Guenzi, Beard, Watanabe, Olsen and Porter1978; Beauchamp et al., Reference Beauchamp, Trevors, Paul and Stewart1989).
Manure mainly comprises undigested materials, which are not always easily decomposable (Rufino et al., Reference Rufino, Rowe, Delve and Giller2006; Gómez-Brandón et al., Reference Gómez-Brandón, Juárez, Domínguez, Insam and Matovic2013). Carbon and N dynamics are linked during the decomposition of animal manure (Chantigny et al., Reference Chantigny, Rochette and Angers2001). Animal slurries undergo decomposition processes, and the products are mainly volatile fatty acids, which are available C sources for the soil micro-organisms (Chantigny et al., Reference Chantigny, Angers and Rochette2002; Hossain et al., Reference Hossain, Rahman, Biswas, Miah, Akhter, Maniruzzaman, Choudhury, Ahmed, Shiragi and Kalra2017). Among these, acetic, propionic and butyric acid are the most common (Mathur et al., Reference Mathur, Owen, Dinel and Schnitzer1993; Zhu and Jacobson, Reference Zhu and Jacobson1999). Grazed grassland soils may also contain a wide range of C compounds, mainly sourced from plant material and animal excreta (Chen et al., Reference Chen, Harrison, Liao, Elliott, Liu, Brown, Wen, Solana, Kincaid and Stevens2003). These may include cellulose, glucose (Chen et al., Reference Chen, Harrison, Liao, Elliott, Liu, Brown, Wen, Solana, Kincaid and Stevens2003), glucosamine (Sradnick et al., Reference Sradnick, Murugan, Oltmanns, Raupp and Joergensen2013); vanillin (Yamamoto et al., Reference Yamamoto, Futamura, Fujioka and Yamamoto2008), benzoic acid (Dijkstra et al., Reference Dijkstra, Oenema, Van Groenigen, Spek, Van Vuuren and Bannink2013), stearic acid (Hristov et al., Reference Hristov, Vander Pol, Agle, Zaman, Schneider, Ndegwa, Vaddella, Johnson, Shingfield and Karnati2009) and phytic acid (Burkholder et al., Reference Burkholder, Guyton, McKinney and Knowlton2004). Based purely on water solubility, it is expected that glucose (Gunina and Kuzyakov, Reference Gunina and Kuzyakov2015) and glucosamine (Roberts et al., Reference Roberts, Bol and Jones2007) are more available to the micro-organisms in the soil, than vanillin and cellulose (Brown et al., Reference Brown, Kauri, Kushner and Mathur1988), which are less soluble (Table 1).
Examples of slurry characteristics, as reported in the literature and the slurries from this study
The bottom two values for labile and non-labile fractions are the sum (respectively) of the reported results Dendooven et al. (Reference Dendooven, Bonhomme, Merckx and Vlassak1998a) and Dendooven et al. (Reference Dendooven, Bonhomme, Merckx and Vlassak1998b).
Published data on the composition of organic manures such as pig slurry show that the majority of the components are readily decomposable, and that cellulose and lignin are only minor components (Dendooven et al., Reference Dendooven, Bonhomme, Merckx and Vlassak1998b). The NO3− content in animal slurries is usually very low due to the anaerobic environment, with almost all of the inorganic nitrogen in the ammonium form, and organic N representing between 40 and 60% of total N in slurries (Chadwick et al., Reference Chadwick, Pain and Brookman2000a).
A large amount of C compounds contained in biomass are released through complex decomposition processes and used in denitrification (Chen et al., Reference Chen, Wen, Zhou and Vymazal2014). The availability of C substrates for microbial processes is linked to their decomposability and ability to support microbial growth (Tusneem, Reference Tusneem1970). Denitrification is carried out by facultative anaerobes and free energy, N2 and CO2, which are produced as a result of electron transfer between NO3− and C (Tusneem, Reference Tusneem1970; Hume et al., Reference Hume, Fleming and Horne2002). This process is highly dependent upon the supply of C and accounts for about 37% of the CO2 produced in the soil respiration system (Ingersoll and Baker, Reference Ingersoll and Baker1998; Rastogi et al., Reference Rastogi, Singh and Pathak2002).
It has been proposed that the electron supply per mole of C from various substrates can affect the efficiency of denitrification (Beauchamp et al., Reference Beauchamp, Trevors, Paul and Stewart1989). A positive correlation has been observed between denitrification and water-soluble organic C (Burford and Bremner, Reference Burford and Bremner1975). Sainju et al. (Reference Sainju, Jabro and Caesar-TonThat2010) found peak CO2 fluxes immediately after substantial precipitation events (above 10 mm), which further highlights the role of denitrifying conditions on CO2 production. Furthermore, Paul and Beauchamp (Reference Paul and Beauchamp1989) reported a strong correlation between CO2 produced under anaerobic conditions and total denitrification.
Carbon sources supply the electrons for these processes in the presence of glucose, as shown in the below reaction:
This reaction indicates a molar ratio of NO3− reduction to CO2 production of 0.8. Differences in this value can be attributed to the presence of other electron acceptors or other factors such as sources of CO2. Ratios of ~0.7 have been attributed to labile compounds, while ratios lower than 0.2 to non-labile compounds (Beauchamp et al., Reference Beauchamp, Trevors, Paul and Stewart1989; Kumar and Sarma, Reference Kumar and Sarma2018).
The majority of research on N losses from agricultural soils has overlooked the role and ranking of different C compounds found in manure for potential denitrification; thus, there exists a knowledge gap, to understand the availability of varying C compounds found in manure in promoting denitrification. The aim of this study was to evaluate the availability of different C compounds in manures and their role in promoting denitrification by (i) establishing the reactivity of varying C compounds found in cattle and pig slurry through their availability for denitrification and (ii) exploring C-quality effects in promoting denitrification.
Materials and methods
Site description
Soil samples were collected from an experimental site at Rothamsted Research, North Wyke, Devon, UK (50°46′10″N, 3°54′05″E). This area is situated at an altitude ranging between 157 and 177 m above sea level, receives a mean annual precipitation of 1040 mm, and has a 30-year mean annual temperature of 10.1°C (Orr et al., Reference Orr, Murray, Eyles, Blackwell, Cardenas, Collins, Dungait, Goulding, Griffith and Gurr2016). The soils are classified as a clayey pelostagnogley of the Hallsworth series (Clayden and Hollis, Reference Clayden and Hollis1985), or an FAO dystric gleysol (FAO, 2006). The top 10 cm are characterized by 36.6% clay, 47.7% silt and 13.9% fine sand and 1.8% coarse sand in the inorganic fraction. The soil pH was 5.7, and organic C was 5.3% (Armstrong and Garwood, Reference Armstrong and Garwood1991; Scholefield et al., Reference Scholefield, Hawkins and Jackson1997; Harrod and Hogan, Reference Harrod and Hogan2008). The experimental site is part of the Rowden Drainage Experiment at North Wyke, which consists of 1 ha paddocks divided into ten equal sections, which were either only grazed, only cut or grazed and/or cut and were used for pasture production. Soil samples for this experiment were collected from two plots receiving no-fertilizer N and without tile drains, which were also grazed during the summer season. Samples were collected in a W-pattern, mixed to form a composite sample and then split into different vessels for treatment application.
Experimental design
Two experiments were carried out to rank different C compounds based on their capability to promote denitrification: (i) experiment 1: incubation of eight different standard C compounds found in manures to rank their availability for potential denitrification and (ii) experiment 2: cattle and pig slurry treatments were included in the incubation and their C availability for denitrification compared to the four highest-ranked C compounds from experiment 1. Changes in time were measured for soil NO3− and NH4+ and the production of CO2. The ranking was based on the molar ratio of CO2 evolved to NO3− reduced after incubation using a low N and C soil medium. The ratio was calculated by dividing CO2 evolved by NO3− reduced between days 1 and 3 of the experiment. The determination of NH4+ was used to help in the interpretation of the results, particularly in the changes in NO3−.
Experiment 1
A range of carbon compounds was selected to provide a variety of molecular weights and structures that may be typically found in animal manure (Dendooven et al., Reference Dendooven, Bonhomme, Merckx and Vlassak1998a; Bertora et al., Reference Bertora, Alluvione, Zavattaro, van Groenigen, Velthof and Grignani2008; Velthof and Mosquera, Reference Velthof and Mosquera2011). These compounds were individually applied as dry materials in combination with NO3− to a soil/sand mixture. Thus, the treatments were: Control – No N or C (CO), N only (CO + N), N + glucose-C (GLU), N + glucosamine-C (GLU-INE), N + cellulose-C (CELL), N + stearic acid-C (STEA), N + benzoic acid-C (BEN), N + lignin-C (LIG), N + vanillin-C (VAN) and N + phytic acid-C (PHY).
A mixture of inert sand and sieved soil (<2 mm) was used as a matrix for the incubation to provide a low N and C content medium. For each vessel (250 ml), the soil/sand mixture comprised of 60 g of dry sand (acid washed) with 40 g of field moist soil, and water was added to achieve 80% water-filled pore space (WFPS) from the initial 30% WFPS. Nitrate (15 mg, equivalent to 150 mg N/kg soil) and 30 mg of C (equivalent to 1000 mg C/kg soil) were added to each vessel and mixed with a spatula in each of the incubation vessels. After that, the soil in each vessel was then pressed to a bulk density of 1 g/cm3. Thereafter, 15 replicates of each treatment were incubated under standard temperature conditions of 15° C for 12 days. CO2 production, NO3− depletion and NH4+ appearance were measured from each vessel on days 1, 3, 6, 9 and 12 after incubation, using the methods described below.
Experiment 2
Based on the results of the previous experiment, we selected a series of C compounds to study the effect of C quality in promoting denitrification. Pig and cattle slurry treatments were included, and their C availability compared with four standard C compounds; glucose, acetic acid, vanillin and cellulose. Thereafter, the treatments were: Control – No N or C (CO), N only (CO + N), N + glucose-C (GLU), N + cellulose-C (CELL), N + acetic acid-C (ACETIC), N + vanillin-C (VAN), N + cattle slurry-C (CS) and N + pig slurry-C(PS) (Table 2).
Percentage of added carbon (C) lost as carbon dioxide (CO2) following C and slurry additions in experiment 2
‡All values are mean ± standard error (s.e.) (n = 3).
No observations were made in the control treatment; hence was omitted.
The procedure for soil medium preparation was the same as for the first experiment. Slurries were obtained from commercial farms and applied at the equal C loading of 30 mg C per vessel (the equivalent of 1000 mg C/kg soil). The same incubation procedures were followed as in the first experiment, with the exception that twice as much NO3− (30 mg per vessel, the equivalent of 300 mg N/kg soil) was applied to each vessel than in the first experiment to ensure an adequate N-source for the entire 9-day incubation period. Slurries were added to the soil/sand medium at the same C loading rate as the C compounds. Gravimetric moisture content was determined, and water was added to achieve 80% WFPS. The moisture content of the slurry was accounted for when adjusting the water content to 80% WFPS. Additionally, CO2 production and NO3− depletion were measured from each vessel on days 1, 3 and 9 after application, using the methods described below.
Carbon dioxide production measurements
The amount of CO2 produced during the incubation period was determined by CO2 adsorption into sodium hydroxide (NaOH). Immediately after adding the C and NO3− amendments, open top 25 ml vials containing 10 ml of 0.5 m NaOH solution were placed in each vessel before sealing the vessels with a screw cap. The experiments were designed with sufficient replication to allow three replicates of each treatment to be destructively sampled on days 1, 3, 6 and 9 (for experiment 2), and additionally on day 12 (for experiment 1) after amendment application. During each destructive sampling, the vial of NaOH was carefully removed, and a 5 ml aliquot was added to 5 ml of 10% BaCl2 in a flask and three drops of 1% phenolphthalein were added. This was then titrated against 0.2 m HCl, and the volume of HCl recorded when the solution became colourless. The amount of CO2 respired during the period was then calculated using the below equation:
where B = standard HCl used to titrate NaOH in the blank (ml), V = the standard HCl used to titrate NaOH in treatment (mL), N = normality of HCl (1.00 N), 22 = equivalent weight of CO2 and W = dry weigh of soil per vessel (mg).
Soil nitrate and ammonium analysis
At the end of each CO2 trapping period, the soil in each vessel was analysed for inorganic N. This was done by adding 200 ml of 2 m KCl to each vessel and mixing for 1 h employing a rotary shaker. The contents of each vessel were then filtered through Whatman no. 4, and the filtrate was analysed for NO3− and NH4+ using a colorimetric automated flow injection technique after Kamphake et al. (Reference Kamphake, Hannah and Cohen1967) and Searle (Reference Searle1984), respectively.
Statistical analysis
To estimate the ratios between nitrate depletion and CO2 production, we used the change in concentration and flux, respectively, between days 1 and 3 for experiment 1 and days 1 and 6 for experiment 2. This was because NO3− had been completely depleted in some treatments at days 3 and 6 for experiments 1 and 2, respectively. Data were tested for significant effects (P < 0.05) of C sources on NO3− reduction, CO2 production, NH4+ appearance, percentage of C evolved as CO2 and molar ratios of CO2 produced : NO3− reduced. This was performed by a two-way analysis of variance followed by the Tukey's post-hoc test using the Statistical Analysis System (Version 9.4, Cary, North Carolina, USA). All values in results and discussion are presented as means (n = 3) and graphs were prepared using SigmaPlot (Version 14, Systat Software Inc., CA, USA).
Results
Experiment 1. Ranking of the different carbon compounds
Soil mineral nitrogen
The soil NO3− reduction followed a zero kinetics order described by Swerts et al. (Reference Swerts, Merckx and Vlassak1996a), with NO3− concentrations decreasing rapidly, reaching almost zero in all treatments at day 6 after incubation (Table 3). NO3− reduction ranged from 276.3 ± 6.59 to 337.6 ± 1.24, below the detection limit (bdl) to 171.6 ± 9.29, and bdl to 6.43 ± 7.12 mg/kg during the 1st, 3rd and 6th days after incubation, respectively. Rapid NO3− reduction was associated with GLU and GLU-INE treatments, with all NO3− having been utilized on the 3rd day after incubation. Nitrate had disappeared in all treatments on the 6th day after incubation, except for the CO + N treatment, which still contained 6.43 ± 7.2 mg/kg of soil.
Soil nitrate (NO3−) reduction, ammonium (NH4+) appearance, cumulative carbon dioxide (CO2) and % C evolved as CO2 at days 1, 3, 6, 9 and 12 after incubation in experiment 1
Treatment description: CO (Control – No N or C), CO + N (N only), GLU (N + glucose-C), GLU-INE (N + glucosamine-C), CELL (N + cellulose-C), STEA (N + stearic acid-C), BEN (N + benzoic acid-C), LIG (N + lignin-C), VAN (N + vanillin-C) and PHY (N + phytic acid-C). ‡All values are mean ± standard error (s.e.) (n = 3).
§bdl = below the detection limit, and NO3− and NH4+ were not determined at day 9 and 12.
Ammonium evolution was detected from day 1 after incubation (Table 3), and all treatments had an increase until day 6. The largest amount of NH4+ was observed during day 6 in all treatments ranging from 101.1 ± 7.5 to 272.2 ± 10.1 mg/kg. The GLU-INE treatment resulted in significantly (P < 0.05) greater NH4+ compared to the other treatments at each sampling date. Correlations between disappeared NO3 and evolved NH4+ were found to be positive and significant on day 1 (r = 0.13; P = 0.01); negative and significant on day 3 (r = −0.51; P = 0.02) and negative and not significant on day 6 (r = −0.35; ns).
Carbon dioxide production
There were significant differences in cumulative CO2 production (between days 1 and 6) by the different C compounds on different days (Table 3). On most days, GLU-INE resulted in the greatest CO2 production, but they were not significantly different from GLU and BEN (P > 0.05). On day 6, BEN had the greatest cumulative CO2 production (944.8 ± 0.0 mg C-CO2/kg), but this was not significantly different from the GLU and GLU-INE treatments (P > 0.05). On days 9 and 12, cumulative CO2 production followed the same trend as day 6, and no significant differences were observed between the highest treatments; GLU, GLU-INE and BEN.
The CO2 production during the 12-day incubation represented about 60% of the C added in the GLU-INE treatment (Table 3). The treatments had the following order, in terms of the CO2 production as a percentage of added C:
Carbon dioxide : nitrate molar ratios
Glucosamine (1.17 ± 0.04) and GLU (1.07 ± 0.04) had a significantly higher reactivity compared to the other treatments (P < 0.05), but they were not necessarily significantly different from each other (Fig. 1). We grouped the compounds according to the mean differences, and they fall into three reactivity groups based on the values obtained:
(i) ‘high’ CO2 production/‘fast’ NO3− reduction; GLU, GLU-INE;
(ii) ‘intermediate’ CO2 production/‘intermediate’ NO3− reduction; BEN, VAN, LIG;
(iii) ‘low’ CO2 production/‘slow’ NO3− reduction; STEA, CELL.
Ranking of C reactivity based on the ratio of CO2 produced : NO3− reduced in experiment 1. Treatment description: PHY (N + phytic acid-C), STEA (N + stearic acid-C), CELL (N + cellulose-C), VAN (N + vanillin-C), BEN (N + benzoic acid-C), LIG (N + lignin-C) CO + N (N only), GLU-INE (N + glucosamine-C) and GLU (N + glucose-C). Vertical lines represent standard error of each treatment mean (n = 3). No reactions were observed in the CO treatment; hence it was omitted.
Experiment 2. Carbon quality effect in promoting denitrification
Slurry composition
The main properties of the slurries selected for the experiment are summarized in Table 1. The CS had a slightly higher dry matter and total N content than the PS, while the total C was marginally lower in the CS (Table 1). The total C and C : N ratio of the slurries used in the current experiment were similar to slurries reported by Risberg et al. (Reference Risberg, Cederlund, Pell, Arthurson and Schnürer2017), and Velthof et al. (Reference Velthof, Kuikman and Oenema2003), for the pig and cattle slurry, respectively.
Soil nitrate
The highest rate of NO3− reduction was observed from the ACETIC and GLU treatments (Fig. 2). Indeed, all of the NO3− had been depleted in the ACETIC treatment by day 6, and by day 9 in the GLU treatment. The apparent rate of NO3− removal from the PS and CS treatments was similar to that of the CO + N treatment.
Average soil NO3− reduction following C and slurry additions in experiment 2. Treatment description: ACETIC (N + acetic acid-C), GLU (N + glucose-C), VAN (N + vanillin-C), CELL (N + cellulose-C), PS (N + pig slurry-C), CS (N + cattle slurry-C) and CO + N (N only). Vertical lines represent standard error of each treatment mean (n = 3). No reactions were observed in the CO treatment; hence it was omitted.
Carbon dioxide production
There were significant differences between treatments for cumulative CO2 and % C produced as CO2 from days 1 to 9 (Table 2 and Fig. 3). Cumulative CO2 production ranged from 267.6 ± 18.6 to 920.9 ± 18.6 mg/kg, with the ACETIC treatment (920.9 ± 18.6 mg/kg), resulting in the highest CO2 output from all the treatments (P < 0.05) (Fig. 3). The % of C evolved as CO2 ranged from 4.8 ± 1.9 to 61.3 ± 1.9%, with the ACETIC treatment (61.3 ± 1.93%) resulting in the highest value (P < 0.05) followed by GLU (42.6 ± 2.5%) (Table 2). The PS (24.4 ± 1.2%) resulted in a higher percentage of CO2 production than the CS treatment (12.0 ± 1.1%).
Average cumulative CO2 production after incubation following C and slurry additions in experiment 2. Treatment description: ACETIC (N + acetic acid-C), GLU (N + glucose-C), VAN (N + vanillin-C), CELL (N + cellulose-C), PS (N + pig slurry-C), CS (N + cattle slurry-C), CO + N (N only) and CO (Control-No N or C). Vertical lines represent standard error of each treatment mean (n = 3).
Carbon dioxide : nitrate molar ratios
Based on the molar ratios, the PS (0.79 ± 0.11) ranked significantly higher (P < 0.05) as a C-source for promoting denitrification compared to other treatments, except for ACETIC treatment which had a molar ratio of 0.72 ± 0.23.
Discussion
Ranking the reactivity of carbon compounds
Nitrate had disappeared entirely from the GLU and GLU-INE treatments by day 3 of the incubation period (Table 3). The rapid NO3− reduction upon incubation with GLU and GLU-INE C compounds was likely due to their high solubility and immediate availability to provide electrons for NO3− reduction by soil micro-organisms (Robertson and Groffman, Reference Robertson, Groffman and Paul2007; Gunina and Kuzyakov, Reference Gunina and Kuzyakov2015). The results of the current study are in agreement with reports by Beauchamp et al. (Reference Beauchamp, Trevors, Paul and Stewart1989) and Geisseler et al. (Reference Geisseler, Horwath, Joergensen and Ludwig2010), that concluded that glucose is an immediately available electron source for microbial utilization of NO3− in the soil. Also, an immediate reduction of N in the presence of elevated glucosamine in grazed grassland soils was reported by Roberts and Jones (Reference Roberts and Jones2012).
Significant amounts of NH4+ were observed in all treatments, including the CO (Table 3) treatment where no NO3− and/or C amendments were applied, which could have been because of residual N from the soil medium but most likely from mineralization of the soil OM especially after sieving at <2 mm (Cookson et al., Reference Cookson, Abaye, Marschner, Murphy, Stockdale and Goulding2005). This was found in a study by Gill et al. (Reference Gill, Jarvis and Hatch1995). In the case of the GLU-INE treatment, the largest amount of NH4+ recorded could have been due to its release from the glucosamine molecules (Tiedje, Reference Tiedje and Zehnder1988; Currey et al., Reference Currey, Johnson, Sheppard, Leith, Toberman, Van Der Wal, Dawson and Artz2010). This process occurs in anaerobic conditions; however, its importance in soils remains uncertain (Yonebayashi and Hattori, Reference Yonebayashi and Hattori1980). The significant correlations between NO3− reduction and NH4+ recovery are in agreement with findings by Page et al. (Reference Page, Dalal and Menzies2003) and Tusneem (Reference Tusneem1970), that reported that NO3− is reduced to NH4+ under O2-limited conditions, similarly to the current experiment. This process requires an electron source generally in the form of C; hence there are significant differences in NH4+ appearance and NO3− reduction as influenced by different C sources (Mohan and Cole, Reference Mohan, Cole, Bothe, Edward and Ferguson2007; Soussana and Lemaire, Reference Soussana and Lemaire2014). Our findings are in agreement with the observations by Mohan and Cole (Reference Mohan, Cole, Bothe, Edward and Ferguson2007) at days 1 and 3 after incubation; however, the relationship became insignificant at day 6. This phenomenon might be the result of rapid microbial NO3− reduction in the presence of C electron sources than would naturally occur (Chantigny et al., Reference Chantigny, Rochette and Angers2001). The amount of NO3− that disappeared was much larger than the amount of NH4+ produced (Table 3): a decrease in NO3− between 1.4 (for GLU-INE) to 4.2 (for BEN) times larger than the increase in NH4+. It seems that the reduction of NO3− could have been only due to denitrification. The relatively high soil moisture conditions of our experiment agree with denitrification as the dominant process occurring (Davidson et al., Reference Davidson, Keller, Erickson, Verchot and Veldkamp2000).
High CO2 production from grassland soils was associated with the GLU and GLU-INE compounds (Table 3). The higher CO2 production from GLU and GLU-INE treatments were in agreement with other authors, particularly Beauchamp et al. (Reference Beauchamp, Trevors, Paul and Stewart1989), Roberts and Jones (Reference Roberts and Jones2012) and Hossain et al. (Reference Hossain, Rahman, Biswas, Miah, Akhter, Maniruzzaman, Choudhury, Ahmed, Shiragi and Kalra2017), all reporting that higher soil respiration rates were associated with elevated glucose and glucosamine levels in soils.
GLU and GLU-INE were the most reactive C sources, whereas PHY and STEA were the less labile ones (Fig. 1). These results are in agreement with earlier findings by Swerts et al. (Reference Swerts, Merckx and Vlassak1996a) and Wang et al. (Reference Wang, Feng, Liao, Zheng, Butterbach-Bahl, Zhang and Jin2013) where CO2 to NO3− ratios above 0.7 were associated with highly labile C compounds. A study by Swerts et al. (Reference Swerts, Merckx and Vlassak1996b) showed a CO2 to NO3− ratio of 0.80 when the C : N (as glucose and NO3−) application was 14 : 1. This ratio is smaller than the value from the current experiment, perhaps due to the lower amount of NO3− available (proportionally) in the current study compared to Swerts et al. (Reference Swerts, Merckx and Vlassak1996b). The ranking of the C compounds reactivity was based on the assumption that NO3− depletion was mostly due to NO3− reduction under anaerobic conditions (Ellis et al., Reference Ellis, Dendooven and Goulding1996), but other processes are reported to convert NO3− to NH4+ such as nitrate ammonification (DNRA) (Butterbach-Bahl et al., Reference Butterbach-Bahl, Baggs, Dannenmann, Kiese and Zechmeister-Boltenstern2013). Still, this is considered to be a small source when high rates of N are applied (Baggs, Reference Baggs2008). Additionally, in the relatively high soil moisture conditions the current experiment, it is not expected that nitrification will occur in soil (Davidson et al., Reference Davidson, Keller, Erickson, Verchot and Veldkamp2000).
Carbon quality effects on potential denitrification
The two slurries included in this study were relatively similar in overall composition (Table 4). However, it is possible there were differences in the proportions of labile v. non-labile C fractions that we did not assess in the current study. Dendooven et al. (Reference Dendooven, Bonhomme, Merckx and Vlassak1998a) and Dendooven et al. (Reference Dendooven, Bonhomme, Merckx and Vlassak1998b) reported pig slurry composition with labile fractions of the order of 78–92%. In cattle slurry, Fangueiro et al. (Reference Fangueiro, Pereira, Macedo, Trindade, Vasconcelos and Coutinho2017) reported labile C in the order of 50%, which was higher than observations of 35.5% by Köster et al. (Reference Köster, Cárdenas, Bol, Lewicka-Szczebak, Senbayram, Well, Giesemann and Dittert2015). Therefore, it is possible that the PS in the current experiment contained greater quantities of labile C compared to the CS. This is reflected in the higher NO3− : CO2 evolution in the PS treatment compared to the CS, which could have been due to the presence of carbonates in pig slurries as reported by other studies (Sommer and Husted, Reference Sommer and Husted1995). However, this was not measured in the experiments of the current study. Differences in the slurry composition are also found between years as Velthof and Mosquera (Reference Velthof and Mosquera2011) report. These authors analysed N in pig and cattle slurries but did not report C contents.
Utilization of carbon (C, %) and carbon : nitrogen (C : N) ratios in slurries reported by some authors in the literature compared to the current study
PS, pig slurry; CS, cattle slurry; ND, not determined.
The different C sources significantly influenced NO3− reduction in the current experiment (Fig. 2). Nitrate was depleted faster upon incubation with the ACETIC and GLU treatments, with all NO3− reduced at days 6 and 9 for the two C-sources, respectively. These results were consistent with findings by Takai and Kamura (Reference Takai and Kamura1966) and Swerts et al. (Reference Swerts, Merckx and Vlassak1996a), indicating that glucose and acetic acid had increased availability as electron sources for soil microbial NO3− reduction reactions. On the other hand, the relatively slower NO3− depletion in the PS and CS treatments could be the result of their C being less decomposed (Rochette et al., Reference Rochette, Angers, Chantigny, Bertrand and Côté2004) compared to the C found in the GLU and ACETIC treatments. Furthermore, the slow NO3− reduction in the CO + N treatment verifies the significance of C in NO3− reducing reactions since organic C has been well documented to stimulate such responses from soils (Sommer and Husted, Reference Sommer and Husted1995; Meijide et al., Reference Meijide, Díez, Sánchez-Martín, López-Fernández and Vallejo2007; Soussana and Lemaire, Reference Soussana and Lemaire2014).
The percentage of C evolved as CO2 in slurries observed in the current study, was generally lower than that reported by other authors (Bertora et al., Reference Bertora, Alluvione, Zavattaro, van Groenigen, Velthof and Grignani2008; Risberg et al., Reference Risberg, Cederlund, Pell, Arthurson and Schnürer2017). The lower CO2 evolved in the CS, PS, VAN and CELL treatments compared to the GLU and ACETIC treatments was likely the result of their C not being easily accessible during the denitrification process (Chantigny et al., Reference Chantigny, Angers and Rochette2002; Rochette et al., Reference Rochette, Angers, Chantigny, Bertrand and Côté2004; Hossain et al., Reference Hossain, Rahman, Biswas, Miah, Akhter, Maniruzzaman, Choudhury, Ahmed, Shiragi and Kalra2017); a significant CO2 producing reaction. The coupling of N and C in the release of atmospheric CO2 has been recorded by several studies (Bertora et al., Reference Bertora, Alluvione, Zavattaro, van Groenigen, Velthof and Grignani2008; Morley and Baggs, Reference Morley and Baggs2010; Hossain et al., Reference Hossain, Rahman, Biswas, Miah, Akhter, Maniruzzaman, Choudhury, Ahmed, Shiragi and Kalra2017). The low CO2 production from the CO + NI treatment in the current study agrees with the coupling effect of soil N and C in CO2 production (Rastogi et al., Reference Rastogi, Singh and Pathak2002; Van Groenigen et al., Reference Van Groenigen, Osenberg and Hungate2011; Risberg et al., Reference Risberg, Cederlund, Pell, Arthurson and Schnürer2017) (Table 5). During the denitrification process, bacteria utilize NO3− as a terminal electron acceptor in the absence of O2 during respiration (Robertson and Groffman, Reference Robertson, Groffman and Paul2007). Denitrifying organisms use C compounds as electron donors for energy; thus, denitrification is highly dependent on the amount and availability of C compounds (Beauchamp et al., Reference Beauchamp, Gale and Yeomans1980; Aulakh et al., Reference Aulakh, Doran, Mosier and Stewart1992). The degradation and transformation of C contained slurries may result to several carbonates, and C compounds with varying availabilities and solubility (Rochette et al., Reference Rochette, Angers, Chantigny, Bertrand and Côté2004); some of which may stimulate CO2 production rates upon reaction with NO3− (Aulakh et al., Reference Aulakh, Doran, Mosier and Stewart1992; Rochette et al., Reference Rochette, Angers, Chantigny, Bertrand and Côté2004).
Molar ratio of carbon dioxide (CO2) produced : nitrate (NO3−) reduction as a ranking for carbon (C) availability for potential denitrification for the first 6 days in experiment 2
‡All values are mean ± standard error (s.e.) (n = 3).
No observations were made in the control treatment; hence was omitted.
Although the slurries did not have a rapid NO3− reduction and CO2 production, the pig slurry emerged with the highest (P < 0.05) CO2 : NO3− molar ratio of 0.79 ± 0.11 (Table 5). This could be because the pig slurry might have contained more NH4+, which might have been nitrified, producing NO3− before inducing denitrifying conditions during the incubation. Table 1 shows some examples from other studies (Knowles, Reference Knowles1982; Skiba, Reference Skiba, Jorgensen and Fath2008). The slurries would also provide NH4+ from their inorganic pool but because of the anaerobic soil conditions, it was not expected that this would have increased the NO3− pool due to nitrification. Subsequent denitrification of this additional NO3− under the favourable conditions of the experiment (temperature and moisture) would have masked the actual rate of reduction of the added NO3− (Jarvis et al., Reference Jarvis, Hatch, Pain and Klarenbeek1994; Griffin et al., Reference Griffin, Honeycutt and He2002), but this was not the case in the current study.
Reactivity of slurries from livestock is complex not only because of the nature of the C they contain but also because they contribute with microbial populations themselves (Acea and Carballas, Reference Acea and Carballas1988; Clemens and Huschka, Reference Clemens and Huschka2001). Changes in slurry composition during storage makes the original non-labile C available (Bertora et al., Reference Bertora, Alluvione, Zavattaro, van Groenigen, Velthof and Grignani2008), increasing the potential for NO3− leaching, increased denitrification and production of N2O. Studies to develop country-specific N2O emission factors for dung and urine deposited during grazing, use freshly collected material that is preserved before use (Cardenas et al., Reference Cardenas, Misselbrook, Hodgson, Donovan, Gilhespy, Smith, Dhanoa and Chadwick2016; Thorman et al., Reference Thorman, Nicholson, Topp, Bell, Cardenas, Chadwick, Cloy, Misselbrook, Rees and Watson2020). The results would likely differ if aged slurries were applied, and possibly emissions would be more substantial. Taylor et al. (Reference Taylor, Parkinson and Parsons1989) stated that the ratio of C to N (C : N) is useful as the first proxy of OM decomposability, with greater C : N generally leading to slower decomposition in slurries. Usually, labile compounds (i.e. soluble sugars and unshielded cellulose) are preferentially lost during the initial phase of decay, and then lignin progressively becomes the dominant constituent of decomposing OM (Berg, Reference Berg2014). In the case of manure, due to the different quality, the dynamics of its chemical composition and its regulation on decomposition process may vary if compared to more standard materials as in the case of plants (Eldridge et al., Reference Eldridge, Chen, Xu, Chan, Boyd, Collins and Meszaros2017). Markewich et al. (Reference Markewich, Pell, Mbugua, Cherney, van Es, Lehmann and Robertson2010) and Bhogal et al. (Reference Bhogal, Williams, Nicholson, Chadwick, Chambers and Chambers2016) stated that manures with relatively low C : N ratios, generally mineralize rather than immobilize mineral N. It would be expected in the current study that, the CS would have more potential to mineralize compared to the PS due to its lower C : N ratio. This ratio has also been found to have a positive relationship with decomposition rate even when the C : N was relatively high and within a wide range (Chen et al., Reference Chen, Xu, Cusack, Castellano and Ding2019), so in the current study, the cattle slurry would decompose slower in agreement with the lower CO2 : NO3− ratio (Table 5).
Conclusions
The results of the molar ratio of CO2 evolution to NO3− reduction ranked glucose and glucosamine as highly reactive C-compounds, and the pig slurry and acetic acid as good quality C-sources, to promote potential denitrification. The results of this study show the importance of characterization of the carbon quality of slurries and, if possible, for multiple years. These results could be useful for improving the accuracy of newly developed mitigation and emission factors and for feeding into models.
Financial support
The British Council is acknowledged for awarding Jerry Dlamini the Newton-Researcher Links Travel Grant (Grant No. 2017-RLTG9-10691) for the visit to Rothamsted Research at North Wyke, United Kingdom between June and July 2018. The BBSRC is also acknowledged for providing a grant to José Martinez during his visit to Rothamsted Research at North Wyke, United Kingdom. This paper was supported by BBSRC grant IO320.
Conflict of interest
None.
Ethical standards
Not applicable.
