Effects of the Addition of Different Additives before Mechanical Separation of Pig Slurry on Composition and Gaseous Emissions

: The treatment of animal slurry is used to improve management on a farm scale. The aim of this laboratory study was to assess the effects of the addition of the additives biochar, alum and clinoptilolite before the mechanical separation of whole pig slurry (WS) on the characteristics and emission of NH 3 , N 2 O, CO 2 and CH 4 from solid (SF) and liquid fractions (LF). The additives were mixed with WS (5% w / w ), followed by separation, in a total of 12 treatments with 3 replicates, including the controls and WS with additives. Gaseous emissions were measured for 30 d by a photoacoustic multigas monitor, and initial characteristics of the slurries were assessed. The results indicated that the separation of the WS modiﬁed the initial physicochemical characteristics and increased the GWP emissions of the SF and LF, but not the NH 3 losses. However, the addition of additives before separation increased the nutrient value and reduced the GWP emissions from the SF and LF. Additionally, just the additive alum was effective in the reduction of E. coli . The additives led to signiﬁcant reductions in NH 3 and N 2 O emissions, with higher reductions in NH 3 losses for alum (51% for NH 3 ) and similar N 2 O losses for all additives (70% for N 2 O) observed, whereas the CO 2 and CH 4 emissions were reduced by biochar (25% for CO 2 and 50% for CH 4 ) and alum (33% for CO 2 and 30% for CH 4 ) but not by clinoptilolite. Although the additives had a positive effect on slurry management, it can be concluded that the addition of alum before mechanical separation has the potential to be the best mitigation measure because it improves the nutrient content and sanitation and decreases gaseous losses from slurry management.


Introduction
High livestock densities and the subsequent generation of large quantities of animal slurry (liquid manure) in certain areas of the world generate hotspots of increased environmental risks through ammonia (NH 3 ), carbon dioxide (CO 2 ), methane (CH 4 ) and greenhouse gases emissions (nitrous oxide (N 2 O)) [1]. The key solution to minimise or avoid environmental and health concerns associated with animal slurry is to ensure the appropriate management through the entire slurry chain from animal housing, storage, treatment operations and the application to soil [1,2].
Animals excrete most of the non-metabolised N as urea (CO(NH 2 ) 2 ), but birds excrete uric acid, which rapidly hydrolyses under the influence of the ubiquitous enzyme, urease, into ammonium (NH 4 + ). Ammonium is in equilibrium with the NH 3 that is lost by volatilisation, also leading to CO 2 emissions by the dissociation of ammonium carbonate ((NH 4 ) 2 CO 3 ) into NH 4 + , CO 2 and H 2 O [3]. The hydrolysis reaction consumes H + with an increase in pH, consequently leading to an imbalance in the NH 4 + /NH 3 balance, increasing the volatilisation of NH 3 [3]. The solid fraction is more likely to be the source of CH 4 and CO 2 production by anaerobic decomposition of the organic matter, and to some extent, N 2 O by nitrification and denitrification processes [2,3].
Previous studies [4][5][6] have compiled and fully described most of the Best Available Techniques (BAT) for mitigation of the environmental impacts (namely NH 3 and greenhouse gas (GHG) emissions) associated with animal slurry management. The technical report prepared by Foged et al. [7], the guidance document from the UNECE Task force on reactive nitrogen [8] and the UNECE Guidance document on sustainable nitrogen management [9] are also good summaries of BAT. Under the generic denomination of slurry additives is a group of products made up of different compounds that interact with the slurry, changing its chemical, biological and physical characteristics and properties [10]. The following positive effects are claimed and described to different degrees on the label of every product: a reduction in the emission of several gaseous compounds (NH 3 and H 2 S); a reduction in unpleasant odours; a change in the physical properties of the manure to make it easier to handle; an increase in the fertilising value of the slurry; a stabilisation of pathogenic microorganisms. Several additives are marketed to reduce the NH 3 emission and odours from stored slurry but are not listed in the UNECE guidance document due to limited evidence of their efficacy and to clear independence during their testing process [8,10,11].
Mechanical separation of animal slurries on European farms, into a liquid and a relatively solid fraction, is often the first manure management step adopted on farms with nutrient excesses [12,13]. Slurry separation allows the concentration of dry matter, organic N and phosphorus (P) in the solid fraction, which can be used to target other parts of the farm where soil P status is suitable or exported from the farm to areas with nutrient deficiencies. The liquid fraction contains the largest fraction of the NH 4 + and potassium (K) content of the original slurry and is often stored on the farm until used as an organic fertilizer in proximate soils [14]. The separation of the liquid fraction from dry matter reduces the requirement for expensive storage and improves the manageability of the liquid during pumping and soil application [15]. Furthermore, treatments of the slurry have been proposed to alter the chemical and physical characteristics of the separation influent, e.g., acidification, flocculation or coagulation to increase the efficiency of the separation treatment [16,17]. Thus, the addition of different additives before mechanical separation of pig slurry could modify the composition of the separated fractions and then further reduce gaseous losses.
The aim of this laboratory study was to assess, during short term storage, the effects of the addition of additives biochar, alum and clinoptilolite before the mechanical separation of pig slurry on the characteristics and emission of NH 3 , N 2 O, CO 2 and CH 4 from the resulting solid and liquid fractions.

Slurries and Additives
Whole pig slurry was obtained via the intensive fattening of a pig from a commercial farm located in Viseu, Portugal. Each one of the additives biochar, (Bioc), alum (Alum) and clinoptilolite (Clin), were mixed into a sample (20 kg) of whole slurry (WS) at a rate of 5% (w/w), using closed plastic barrels at 20 • C for 24 h. Another sample (20 kg) of whole slurry (WS) without any additive was retained in a similar barrel and storage conditions. Then, 4 kg of WS with each one of the three additives (WS + Bioc, WS + Alum, and WS + Clin), as well as the same amount of WS without additive (WS), were kept in closed plastic barrels and subsamples were retained for analysis, before the start of the experiment. Additionally, the remaining samples (16 kg) of each one of the four slurries with and without additives were subjected to sieving through a 1.0 mm screen, generating a solid (SF) and a liquid fraction (LF), with the following separation yields (w/w): 26.3% for SF and 73.7% for LF; 28.4% for SF + Bio and 71.6% for LF + Bio; 29.7% for SF + Alum and 70.3% for LF + Alum; 29.2% for SF + Clin and 70.8% for LF + Clin. The sieving of WS through a 1.0 mm screen was to mimic the commercially mechanical separators used on commercial farms.

Gaseous Emissions
The experiment was carried out using a system of twelve Kilner jars (H = 230 mm, Ø = 105 mm, volume = 2.0 L) filled with 1.0 kg (H = 105 mm) of each treatment (with three replications per treatment) along 30 d and at constant airflow rate and temperature (20 • C), such as those used by Pereira et al. [21,22]. Briefly, one air inlet and one air outlet were inserted in the jar lid with a Teflon tube (Ø = 3 mm) through one of the septa, with the end kept above the slurry surface (H = 20 mm). The airflow through the headspace of each jar was achieved by a pump (KNF, model N010.KN.18, Neuberger GmbH, Freiburg, Germany), with a flowrate (2.5 L min −1 ) regulated by a needle valve coupled to a flow meter (AalborgTM FT10201SAVN, Aalborg, Denmark). The inlet air was subjected to NH 3 -trapping filters with oxalic acid and the outlet air of the Kilner jars was exhausted out of the climatic room by a fume hood. The concentrations of NH 3 , N 2 O, CO 2 and CH 4 were measured in the exhaust air with a photoacoustic multigas monitor (INNOVA 1412i-5, Lumasense Technologies, Ballerup, Denmark) and air samples collected, in sequence (2 min intervals), through one sampling point (Ø = 3 mm) per Kilner jar, by a multipoint sampler (INNOVA 1409-12, Lumasense Technologies, Ballerup, Denmark) provided with PTFEfilters (0.001 mm pore size, Whatman, Ome, Japan). The photoacoustic multigas monitor was equipped with an optical filter for water vapour (filter type SB0527) and the detection limits for NH 3 (filter type UA0973), N 2 O (filter type UA0985), CO 2 (filter type UA0982) and CH 4 (filter type UA0969) were, respectively, 0.1521, 0.0589, 2.9471 and 0.2864 mg m −3 . The photoacoustic multigas monitor was calibrated by the manufacturer before the beginning of the experiment and operated in a mode that compensated for water interference and cross interference.
For each experiment, individual samples of WS with and without additives and their corresponding fractions were thawed (over 24 h at 4 • C) and then brought to 20 • C immediately before being inserted in the Kilner jar. The temperature was monitored without interruption by sensors (CS107, Campbell Scientific, Loughborough, UK) connected to a micrologger (CR3000, Campbell Scientific, Loughborough, UK).

Data Analysis
The NH 3 , N 2 O, CO 2 and CH 4 concentrations were used to determinate means per hour and day per each outlet sampling point. Then, the emission of these four gases was determined (per hour) using a mass balance as described in Equation (1): where E is the gas emission (mg m −2 h −1 ), F is the air flowrate in the Kilner jar (m 3 h −1 ), OUT is the outlet gas concentration (mg m −3 ), IN is the inlet gas concentration (mg m −3 ) using the following background coefficients: 0.00266 mg m −3 for NH 3 , 0.58942 mg m −3 for N 2 O, 628.71429 mg m −3 for CO 2 and 1.07411 mg m −3 for CH 4 , and A (m 2 ) is the emitting surface area of the Kilner jar.
The reduction efficiencies (RE, %) of NH 3 , CO 2 , CH 4 and N 2 O emissions from slurries and their fractions with additives, comparatively to untreated slurries, were determined as described in Regueiro et al. [17] using the Equation (2): where AD is the mean value of individual or cumulative gas values from slurries and their fractions with additives, and UN is the mean value of individual or cumulative gas values from untreated slurries. The cumulative emissions of NH 3 , N 2 O, CO 2 and CH 4 were determined by averaging the flux between two sampling occasions and multiplying by the time interval between the measurements [21,22]. Then, cumulative emissions were expressed as % of total N or C applied in each slurry and fraction. The global warming potential (GWP) for each Kilner jar was assessed using the global warming potential coefficients for direct greenhouse gas emissions (265 for N 2 O, 1 for CO 2 and 28 for CH 4 ) and indirect N 2 O emissions (1% of NH 3 -N volatilised for N 2 O-N) [21][22][23].
To assess the effect of the separation process on gaseous emissions, the sum of the corrected cumulative emissions from slurries and their fractions with additives were compared with their respective unseparated slurries. The sum of the emissions was calculated by Equation (3): where GS is the sum of the corrected cumulative emissions from separated slurries, GLF and GSF are the corrected cumulative emissions for liquid and solid fractions, respectively, and ALF and ASF are the proportions of liquid and solid fractions, respectively, obtained after the separation of additive and non-additive slurries.
The data obtained was analysed by two-way analysis of variance (ANOVA) to test the effects of dependent parameters (composition and gaseous emissions of slurries and their fractions with and without additives), followed by Tukey's significant difference test (p < 0.05) comparisons of means tests (for the factor (slurries or additives) or interaction effects), using the statistical software package STATISTIX 10.0 (Analytical Software, Tallahassee, FL, USA).

Composition of the Slurries
At the beginning (0 d) of the study, the main characteristics of the treatments that received slurries (WS, SF and LF) with and without additives (Bioc, Alum and Clin) are provided in Table 1. The initial pH values (0 d) did not differ significantly (p > 0.05) among treatments WS and SF (pH = 7.1), being significantly higher (p < 0.05) in treatment LF (pH = 7.7) ( Table 1). In addition, the initial pH values of slurry treatments with the additive Alum (pH < 4.2) decreased significantly (p < 0.05) when compared with all other treatments (pH > 6.9) ( Table 1). The initial dry matter content (0 d) did not differ significantly (p > 0.05) among treatments WS and LF (DM < 2.3%), being significantly higher (p < 0.05) in treatment SF (DM = 13.8%) ( Table 1). The initial DM content increased significantly (p < 0.05) in almost all additive treatments (Bioc, Alum and Clin) when compared with the same treatments without additives (WS, SF and LF) (2.3% for WS against 8.5% for WS + Bioc), with higher values for treatments WS and LF with additive Bioc ( Table 1).
The initial total C (0 d) was significantly higher (p < 0.05) in treatment SF relative to treatments WS and LF (149.9 vs. 48.8 g total C kg −1 ) ( Table 1). The initial total C increased significantly (p < 0.05) in treatments WS and LF with the additive Bioc when compared with all other treatments. (Table 1). The initial total N (0 d) was significantly higher (p < 0.05) in treatment SF relative to treatments WS and LF (6.8 vs. 3.3 g total N kg −1 ) ( Table 1). The initial total N did not increase significantly (p > 0.05) in all additive treatments (Bioc, Alum and Clin) when compared with the same treatments without additives (WS, SF and LF) ( Table 1). The initial NH 4 + (0 d) was significantly higher (p < 0.05) in treatment SF relative to treatments WS and LF (3.3 vs. 1.8 g total N kg −1 ) ( Table 1). The initial NH 4 + did not increase significantly (p > 0.05) in all additive treatments (Bioc, Alum and Clin) when compared with the same treatments without additives (WS, SF and LF) ( Table 1). The initial NO 3 − (0 d) was significantly higher (p < 0.05) in treatments WS and SF relative to treatment LF ( Table 1). The initial NO 3 − did not increase significantly (p > 0.05) in all additive treatments (Bioc, Alum and Clin) when compared with the same treatments without additives (WS, SF and LF) ( Table 1). The initial NH 4 + /total N ratio (0 d) did not differ significantly (p > 0.05) among treatments WS, SF and LF ( Table 1). The initial NH 4 + /total N ratio did not increase significantly (p > 0.05) in all additive treatments (Bioc, Alum and Clin) when compared with the same treatments without additives (WS, SF and LF) ( Table 1). The initial C/N ratio was significantly higher (p < 0.05) in treatment SF relative to treatments WS and LF (C/N = 22 for LF against C/N = 14 for WS or LF) ( Table 1). The initial C/N ratio (0 d) did not increase significantly (p > 0.05) in almost all additive treatments (Bioc, Alum and Clin) when compared with the same treatments without additives (WS, SF and LF) ( Table 1).
The separation yields of the SF increased significantly (p < 0.05) in all slurries with additives (Bioc, Alum and Clin) relative to WS, with higher values for Alum (approximately 30%), in agreement with previous studies that reported an enhancement of separation yields due the addition of additives before the mechanical separation [12,14,16,17].
The additives interact with the whole slurry, changing its chemical, biological and physical characteristics and properties, with the following positive effects: reduction in the emission of several gaseous compounds, change in the physical properties of the manure to make it easier to handle, increase in the fertilising value of the manure and stabilisation of pathogenic microorganisms [3,[8][9][10]. Biochar is a porous carbonaceous material largely containing C jointly with the inorganic components of the biomass utilised, such as alkali and alkaline earth metals, and its addition to slurry before separation increases the pH, the C/N ratio, cation-exchange capacity and microbial activities [22]. Clinoptilolite are crystalline, hydrated aluminosilicates of alkali and alkaline earth cations with high porosity, ion exchange and adsorption capacity for NH 4 + retention, and its addition to slurry before separation reduces the dissolved NH 4 + by adsorbing on ion exchange sites [24]. Alum acts by acidification of the slurry at pH < 5.0, conserving NH 3 , and its addition to slurry before separation improves fertilizer value and sanitisation [25].
The initial E. coli (0 d) did not differ significantly (p > 0.05) among treatments WS, SF and LF (Table 1). In addition, the initial E. coli of slurry treatments with the additive Alum decreased significantly (p < 0.05) when compared with all other additive treatments (Bioc and Clin) (1.0 colony-forming units (CFU) mL −1 for Alum) ( Table 1). Results of this study (Table 1) did not show evidence that the addition of biochar and clinoptilolite might be effective at reducing E. coli, corroborating with the literature concerning the reduced effectiveness of these additives on bacterial activity [20,22]. On other hand, the addition of alum was effective on the reduction in E. coli (Table 1), being in line with previous studies where acidification was able to achieve sanitisation to pH < 5.0 [25].

Nitrogen Emissions
On most measurement days, the daily NH 3 fluxes decreased progressively in treatments throughout the 30 d of experiment (from 980 to 30 mg m −2 h −1 ) and are shown in Table 2. Additionally, on day 30 of the experiment, significantly higher (p < 0.05) NH 3 fluxes were observed in the following order: LF > WS > SF, ( Table 2). Comparative to the WS treatment, the SF treatment significantly reduced (p < 0.05) the daily NH 3 fluxes by 54% whereas the LF treatment increased these fluxes by 54% (Table 2). During most measurement days, the daily NH 3 fluxes were significantly reduced (p < 0.05) in all additive treatments (Bioc, Alum and Clin) when compared with the same treatments without additives (WS, SF and LF), with reductions of 50% for additives Alum and Clin and of 38% for additive Bioc ( Table 2). The cumulative NH 3 emissions, expressed in g m −2 , increased significantly (p < 0.05) in the following order: LF > WS > SF, with a reduction of 53% for the SF treatment and an increase of 45% for the LF treatment when compared with the WS treatment ( Table 2). The cumulative NH 3 emissions, expressed in g m −2 , were significantly reduced (p < 0.05) in all additive treatments (Bioc, Alum and Clin) relative to the same treatments without additives (WS, SF, LF), with reductions of 52% for the additives Alum and Clin and of 38% for the additive Bioc ( Table 2). The cumulative NH 3 emissions, expressed as % of total N applied, increased significantly (p < 0.05) in the following order: LF > WS > SF, with a reduction of 78% for the SF treatment and an increase of 28% for the LF treatment when compared with the WS treatment ( Table 3). The cumulative NH 3 emissions, expressed as % of total N applied, were significantly reduced (p < 0.05) in all additive treatments (Bioc, Alum and Clin) relative to the same treatments without additives (WS, SF and LF), with reductions of 51% for the additive Alum and of 36% for the additives Bioc and Clin (Table 3).
As can be seen in Table 3, the NH 3 emissions were reduced by 36% by the addition of biochar or clinoptilolite and by 51% by the addition of alum, which could be related with saturation of the capacity of adsorption of NH 4 + by biochar or clinoptilolite and the maintenance of low and stable pH by alum [24,26]. The reduction of NH 3 losses by biochar were due to the high specific surface area and the high cation exchange capacity of these additives, which enhance the NH 4 + and NH 3 binding [26]. Previous studies [22,27] reported that the addition of biochar (1-12% w/w) to animal manure reduced NH 3 emissions between 12 and 77%, in the same range than the present study (36% NH 3 reduction for 5% w/w biochar). The addition of clinoptilolite increases the number of NH 4 + exchange sites, decreasing the quantity of dissolved NH 4 + and, thus, the quantity of equilibrated NH 3 gas available for NH 3 volatilisation [24]. In this study, the reduction of NH 3 emissions (36% NH 3 reduction for 5% w/w of clinoptilolite) by the addition of clinoptilolite was in line with emissions (26-50% NH 3 reduction for 2.50-6.25% w/w of clinoptilolite) reported in other studies [22,24] for animal slurry. The addition of alum was effective in conserving NH 3 because the percentage of total solution ammoniacal N (NH 4 + and NH 3 ) that was dissociated as NH 3 gas is approximately 0.006% at a pH of 5.0 and temperature of 25 • C [24]. Previous studies [17,24] reported that the addition of alum (2.0-2.5% w/w) to animal slurry reduced NH 3 emissions by between 60 and 67%, being comparable with emissions of the present study (51% NH 3 reduction for 5% w/w of alum).
The daily N 2 O fluxes follow the same trend in treatments, with a small variation throughout the 30 d of experiment, except in SF treatments with and without additives with a progressive increase in the last 20 d of the experiment (Table 4). Comparative to other treatment slurries, the daily N 2 O fluxes were significantly higher (p < 0.05) in the first 11 d of the experiment for the LF treatment, and from day 16 until the end of the experiment for the SF treatment (Table 4). Comparative to treatments without additives, the daily N 2 O fluxes were significantly reduced (p < 0.05) in the first 11 d of the experiment for the WS and SF treatments, and from day 12 until the end of the experiment for the SF treatment ( Table 4). The cumulative N 2 O emissions, expressed in g m −2 , were not significantly different (p > 0.05) among the WS and LF treatments, being lower by approximately 70% than the SF treatment ( Table 4). The cumulative N 2 O emissions, expressed in g m −2 , did not differ significantly (p > 0.05) among the WS and LF treatments with and without additives (Bioc, Alum and Clin), whereas these treatment additives were significantly reduced (p < 0.05) by 69% when compared with the SF treatment without additive ( Table 4). The cumulative N 2 O emissions, expressed as % of total N applied, did not differ significantly (p > 0.05) between the WS and LF treatments, but were significantly lower (p < 0.05) relative to the SF treatment (1.2% of total N applied for WS or LF treatments and 2.3% of total N applied for the SF treatment) ( Table 3). The cumulative N 2 O emissions, expressed as % of total N applied, were significantly reduced (p < 0.05) by approximately 70% in SF treatments with additives (Bioc, Alum and Clin) when compared with the same treatment without additive, whereas no significant reductions (p > 0.05) were observed in the WS or LF treatments with additives (Table 3). Table 2. Average ammonia fluxes (mg m −2 h −1 ) and total flux (mg m −2 ) from each treatment (mean ± standard deviation) (n = 3).

Treatments Days of Experiment
The N (NH 3 + N 2 O) emissions, expressed as g m −2 or as % of applied N, increased significantly (p < 0.05) in the following order: SF > WS > LF, with a reduction of approximately 60% for the SF treatment and an increase of approximately 36% for the LF treatment, when compared with the WS treatment ( Table 3). The cumulative NH 3 emissions, expressed as % of total N applied, were significantly reduced (p < 0.05) in all additive treatments (Bioc, Alum and Clin) relative to the same treatments without additives (WS, SF and LF), with reductions of 52% for the additive Alum and 36% for the additives Bioc and Clin ( Table 3).
The NH 3 emissions did not differ significantly (p > 0.05) among the separated fractions together (SF and LF) and the WS, which is not in agreement with previous studies [13,29] that state that NH 3 emissions could increase when raw slurry was separated. Comparative to the application of WS (100% emission), slurry separation alone (LF vs. SF) did not significantly increase (p > 0.05) NH 3 , N 2 O and N emissions ( Table 5). The combination of the slurry separation with the additives (Bioc, Alum and Clin) did not significantly reduce (p > 0.05) NH 3 , N 2 O and N emissions (Table 5). Table 5. Effect of different additives and slurry separation on the balance of gaseous losses compared with the whole slurry (as % of emissions observed in the whole slurry) (mean ± standard deviation) (n = 3).

Parameters
Whole

Carbon Emissions
Comparative to other treatment slurries, the daily CO 2 fluxes were significantly higher (p < 0.05) in the first 2 d of experiment for the LF treatment and between day 9 and the end of the experiment for the WS treatment ( Table 6). The daily CO 2 fluxes were reduced, but not always significantly, in treatments with the additives (Bioc, Alum and Clin) when compared with the same treatments without additives (Table 6). On most measurement dates, the daily CO 2 fluxes from treatments with additives were reduced significantly (p < 0.05) in the following order: Alum < Bioc < Clin, with a reduction of approximately 20% for treatments with Bio and Alum ( Table 6). The cumulative CO 2 emissions, expressed in g m −2 , were not significantly different (p > 0.05) among SF and LF treatments, being significantly lower (p < 0.05) by approximately 25% than the WS treatment ( Table 6). The cumulative CO 2 emissions, expressed in g m −2 , were reduced in all additive treatments (Bioc, Alum and Clin) relative to the same treatments without additives (WS, SF and LF), with a significant decrease of 22% for the additives Bioc and Alum ( Table 6). The cumulative CO 2 emissions, expressed as % of total C applied, were reduced significantly (p < 0.05) in the following order: SF < LF < WS, with a reduction of 61% in the SF treatment relative to WS ( Table 3). The cumulative CO 2 emissions, expressed as % of total C applied, were significantly reduced (p < 0.05) by 25% in all treatments with the additive Bioc and by 33% in the SF treatment with Alum ( Table 3).
The two main sources of CO 2 emissions are the microbial degradation of organic matter and urea hydrolysis [33]. In addition, it will be expected that the CO 2 emissions are higher for SF since these losses seem higher in slurry fractions with high amounts of C [13], but this patten is not always reported in other studies [12,34]. The high CO 2 emissions obtained in WS and LF relative to SF could be related with the release of the CO 2 dissolved in the slurry itself and/or bicarbonate and carbonate present in the slurries [35]. Moreover, the SF had dissolved CO 2 and very low amounts of water-soluble C together with the reduction in volume by water loss and aerobic condition by sample compaction [17]. As can be seen in Table 3, the CO 2 emissions were reduced significantly by 25% by the addition of biochar. The results of this study are lower than previous studies [21,32], which reported that CO 2 emissions from animal slurry were reduced by between 34 and 50% by the addition of biochar (5-10% w/w), due to either sorption onto the biochar or a reduction in the labile C availability. However, the additive clinoptilolite appears to have had no effect on CO 2 emission in this study (Table 3), which is in line with a previous study [21] that reported the absence of significant effect of this additive (2.5% w/w) on CO 2 reduction. In this study, the decrease in CO 2 emission by alum added to SF (33% CO 2 reduction for 5% w/w of alum) was because most of the dissolved CO 2 is lost during the acidification process [30], which is in line with Regueiro et al. [17], who reported that the SF of pig slurry amended with alum (2% w/w) reduced CO 2 loss by 41%.
The cumulative C (CO 2 + CH 4 ) emissions, expressed in g m −2 , were not significantly different (p > 0.05) among SF and LF treatments, being significantly lower (p < 0.05) by approximately 45% than the WS treatment ( Table 3). The cumulative CH 4 emissions, expressed in g m −2 , were not significantly different (p > 0.05) among WS and LF treatments, being significantly lower (p < 0.05) by approximately 74% than the SF treatment ( Table 3). The cumulative C (CO 2 + CH 4 ) emissions, expressed as % of total C applied, were reduced significantly (p < 0.05) in the following order: SF < LF < WS, with a reduction of 60% in the SF treatment relative to WS ( Table 3). The cumulative C (CO 2 + CH 4 ) emissions, expressed as % of total C applied, were significantly reduced (p < 0.05) by 26% in treatments with the additive Bioc, when compared with all other treatments with or without additives ( Table 3). The cumulative GWP emissions, expressed as CO 2 eq. m −2 , were significantly increased (p < 0.05) by approximately 67% in the SF treatment, when compared with WS and LF treatments ( Table 3). The cumulative GWP emissions, expressed as CO 2 eq. m −2 , were significantly reduced (p < 0.05) by approximately 28%, respectively, in treatments with the additives Bio and Alum when compared with all other treatments with or without additives (Table 3).
Comparative to the application of WS (100% emission), slurry separation alone (LF vs. SF) significantly reduced (p < 0.05) CO 2 or C emissions by 29%, and CH 4 by 40% ( Table 5). The separation alone significantly increased (p < 0.05) the GWP emissions of separated fractions together (SF and LF). The combination of the slurry separation with the additives (Bioc, Alum and Clin) did not significantly reduce (p > 0.05) CH 4 emissions, whereas the CO 2 and C emissions were significantly reduced (p < 0.05) by approximately 40% with the additive Alum (Table 5). However, when additives (Bioc, Alum and Clin) were applied before separation, the GWP emissions of the fractions combined together (SF and LF) were significantly lower (p < 0.05) than from WS.

Conclusions
The results indicated that the mechanical separation of the WS modified the initial physicochemical characteristics and increased the GWP emissions of the two separated fractions together (solid and liquid fractions), but not the NH 3 losses. However, the addition of the additives (biochar, alum or clinoptilolite) before mechanical separation increased the fertilizer value and reduced the GWP emissions from the solid and liquid fractions. Additionally, just the additive alum was effective in the reduction of E. coli. The addition of the three additives led to significant reductions in NH 3 and N 2 O emissions, with higher reductions in NH 3 losses for alum observed and similar N 2 O losses for all additives, whereas the CO 2 and CH 4 emissions were reduced by biochar and alum, but not by clinoptilolite.
Globally, the addition of alum before mechanical separation has the potential to be the most effective mitigation measure because it improved the fertilizer value and sanitation and decreased the gaseous losses from pig slurry management when compared with biochar and clinoptilolite. Thus, farm scale studies are needed to validate these results under real conditions.