Effects of cattle-slurry treatment by acidification and separation on nitrogen dynamics and global warming potential after surface application to an acidic soil

Journal of Environmental Management 162 (2015) 1e8

Contents lists available at ScienceDirect

Journal of Environmental Management journal homepage: www.elsevier.com/locate/jenvman

Research article

Effects of cattle-slurry treatment by acidification and separation on nitrogen dynamics and global warming potential after surface application to an acidic soil
Pereira b, c, Andre
Bichana a, So
nia Surgy a, Fernanda Cabral a, David Fangueiro a, *, Jose d ~o Coutinho Joa a

LEAF, Instituto Superior de Agronomia, Universidade de Lisboa, Tapada da Ajuda, 1349-017 Lisboa, Portugal CI&DETS, ESA de Viseu, Instituto Polit
ecnico de Viseu, Quinta da Alagoa, 3500-606 Viseu, Portugal
s-os-Montes e Alto Douro, Quinta de Prados, 5000-801 Vila Real, Portugal CITAB, Universidade de Tra d
s-os-Montes e Alto Douro, Quinta de Prados, 5000-801 Vila Real, Portugal Centro de Química, Universidade de Tra b c

a r t i c l e i n f o

a b s t r a c t

Article history: Received 19 May 2015 Received in revised form 10 July 2015 Accepted 14 July 2015 Available online xxx

Cattle-slurry (liquid manure) application to soil is a common practice to provide nutrients and organic matter for crop growth but it also strongly impacts the environment. The objective of the present study was to assess the efficiency of cattle-slurry treatment by solideliquid separation and/or acidification on nitrogen dynamics and global warming potential (GWP) following application to an acidic soil. An aerobic laboratory incubation was performed over 92 days with a Dystric Cambisol amended with raw cattle-slurry or separated liquid fraction (LF) treated or not by acidification to pH 5.5 by addition of sulphuric acid. Soil mineral N contents and NH3, N2O, CH4 and CO2 emissions were measured. Results obtained suggest that the acidification of raw cattle-slurry reduced significantly NH3 emissions (
88%) but also the GWP (
28%) while increased the N availability relative to raw cattle-slurry (15% of organic N applied mineralised against negative mineralisation in raw slurry). However, similar NH3 emissions and GWP were observed in acidified LF and non-acidified LF treatments. On the other hand, soil application of acidified cattle-slurry rather than non-acidified LF should be preferred attending the lower costs associated to acidification compared to solideliquid separation. It can then be concluded that cattle-slurry acidification is a solution to minimise NH3 emissions from amended soil and an efficient strategy to decrease the GWP associated with slurry application to soil. Furthermore, the more intense N mineralisation observed with acidified slurry should lead to a higher amount of plant available N and consequently to higher crop yields. © 2015 Elsevier Ltd. All rights reserved.

Keywords: Ammonia Methane Nitrous oxide Carbon dioxide Slurry acidification Solideliquid separation

1. Introduction In most European countries, the dairy production suffered a strong transformation over the last decades and is now mainly based on industrial farm units with a slurry (liquid manure) based system (Bittman et al., 2014). Furthermore, the end of the milk quota in EU-27 will open room to large dairy production units and slurry management should become a key point for farms sustainability. Cattle-slurry application to soil provides nutrients to plant growth and increases soil organic matter content. However, slurry application to soil has a well-known environmental impact that affects all ecological compartments: air, soil, water and biosphere * Corresponding author. E-mail address: [email protected] (D. Fangueiro). http://dx.doi.org/10.1016/j.jenvman.2015.07.032 0301-4797/© 2015 Elsevier Ltd. All rights reserved.

(FAO, 2006). It contributes namely to emissions of ammonia (NH3) and greenhouse gases (GHG) as nitrous oxide (N2O), methane (CH4) and carbon dioxide (CO2) that strongly impact the environment but such N emissions also decrease the slurry fertilising value (EEA, 2013, 2014). Cattle-slurry treatment before soil application can be seen as a solution for these problems related to slurry management. Namely, solideliquid separation has been proposed to improve slurry management during storage and after soil application (Hjorth et al., 2010). It leads to a solid fraction rich in organic matter that can easily be transported to other farms and to a liquid fraction (LF) that can be used on-farm. Such technique is now widely used at farm scale in European countries but the interest for the resulting LF is still scarce and many farmers mix it back in the slurry pit. Indeed, even if the concentration of ammonium ðNH4 þ Þ N in LF is relatively

2

D. Fangueiro et al. / Journal of Environmental Management 162 (2015) 1e8

high, the organic matter and organic N content of LF is low relative to raw slurry (Hjorth et al., 2010; Fangueiro et al., 2012). Nevertheless, the soil infiltration of LF can be enhanced, in some conditions, by its lower dry matter content and viscosity relative to raw slurry and consequently, lower NH3 emissions can be expected following LF application relative to raw slurry (Sommer and Hutchings, 2001; Sommer et al., 2003; Nyord et al., 2013) even if some studies reported the opposite (Fangueiro et al., 2015b). Another solution to minimise NH3 emissions consists in decreasing the slurry pH by acid addition (Stevens et al., 1989, 1992). Such practice is very effective to prevent NH3 emissions but little is known about the global impact of acidified slurry application to soil (Kai et al., 2008; Fangueiro et al., 2015a). A combined treatment -separation followed by LF acidificationappears as an efficient solution to improve the slurry value through SF export to other farms and acidified LF application to soil at farm, with low impact on environment. Alternatively, raw slurry acidification could be a cheaper solution in farms with enough soil area to receive all the slurry produced since solideliquid separation implies a significant investment in machinery and daily energy costs. In this case acidified slurry could be applied on soil surface rather than injected as recommend in most EU countries. Furthermore, in countries from South Europe where most slurry is still applied by broadcast followed by incorporation, band application of acidified slurry appears as a good alternative since it precludes incorporation. Since more ammonium N remained in soil after acidified slurry or LF application, an intense nitrification/denitrification process might be expected and the potential risk of N2O emissions might be higher compared to non-acidified slurry. Furthermore, previous studies by Fangueiro et al. (2009, 2010b, 2013) indicated that soil application of acidified pig slurry decreases soil respiration, nitrification, and microbial-biomass-C values, relative to soil application of raw pig slurry. The aim of the present study was to assess, in microcosms, the efficiency of cattle-slurry treatment by acidification and/or solideliquid separation on nitrogen dynamics, GHG emissions and global warming potential (GWP) following application to a sandy loam soil. 2. Materials and methods 2.1. Soil and slurry preparation The soil used in our study was classified as a Haplic Cambisol (Dystric) (IUSS Working Group WRB, 2006) with a sandy texture (380 g kg
1 coarse sand (0.2e2 mm), 380 g kg
1 fine sand (0.02e0.2 mm), 150 g kg
1 silt (0.002e0.02 mm) and 90 g kg
1 clay (<0.002 mm). The soil was collected from the upper layer (0e200 mm) of an agricultural field located at the Agricultural Polytechnic School of Castelo Branco, Portugal (39 490 1700 N 07 270 4400 W). At laboratory, soil was manually sieved (<2 mm) and then stored at room temperature before the beginning of the experiment. The physical-chemical properties of the soil were: bulk density: 1.52 g cm
3, pH (H2O): 5.05, electrical conductivity: 845 mS cm
1, water holding capacity (WHC): 305 g kg
1, total C: 12.4 g kg
1 dry soil, total N: 0.5 g kg
1 dry soil, available P: 100.4 mg kg
1 dry soil, NH4 þ eN: 11.5 mg kg
1 dry soil, NO3
eN: 28.5 mg kg
1 dry soil. The cattle-slurry used in this study was sampled from the concrete pit of a commercial dairy farm. Five kg of raw slurry (S) was subjected to solideliquid separation by centrifugation at 1509 g during 15 min to obtain two fractions, a liquid (LF) and a solid fraction. Then, an acidified cattle-slurry (AS) and an acidified liquid fraction (ALF) were obtained by adding concentrated sulphuric acid

to S and LF, respectively, until pH 5.5 was reached, following the procedure described by Fangueiro et al. (2013). All the treated and untreated slurries were preserved at 4  C until required. The resulting materials were then characterized and the main parameters are shown in Table 1. A comprehensive description of the methods used to assess the physical-chemical properties of the soil and the four cattle slurries studied (S, AS, LF and ALF), can be found in Fangueiro et al. (2010b, 2013). Extractions with 0.01 M CaCl2 (1:10 w/v) was performed to quantify total soluble N by chemiluminescence detection in a Formac Skalar analyzer, soluble organic C by elemental analysis, using a NIRD detector (Formac, Skalar, Breda, NL) and total soluble P by molecular absorption spectrophotometry in a SanPlus (Skalar, Breda, NL) segmented flow analyzer. 2.2. Experimental design Three independent aerobic incubations were performed simultaneously under controlled conditions (25  C) to follow the (i) soil nitrogen dynamics, (ii) NH3 emissions, and (iii) GHG emissions in soil amended with raw and treated slurry (S, AS, LF and ALF). Slurries were manually applied to soil surface to simulate band application at a rate of 80 mg total N kg
1 dry soil (Table 1). The corresponding rates of C and NH4 þ applied are shown in Table 1. Two additional treatments were considered: soil only (Control) and S application on soil surface followed by soil incorporation (depth ¼ 20 mm) with the help of a small rake on all surface area. The final moisture content of amended and non-amended soil was adjusted at 60% WHC by adding deionised water. A total of six treatments were considered in this study: 1. 2. 3. 4. 5. 6.

Non amended soil (Control); Surface application of S followed by soil Incorporation (S-I); Surface application of S (S-S); Surface application of AS (AS-S); Surface application of LF (LF-S); Surface application of ALF (ALF-S).

2.2.1. Soil nitrogen dynamics An amount of 50 g dry soil (depth ¼ 50 mm) was packed into PVC containers (Ø ¼ 25 mm, H ¼ 70 mm). Slurry was then applied as above described. A total of 52 (13 sampling dates  4 replicates) PVC containers were prepared for each treatment. Each container was loosely caped and placed in a temperature controlled cabinet. Moisture content was checked regularly by weighing and adjusted to 60% WHC by deionised water addition whenever necessary. Table 1 Main characteristics of the slurries used and soil application rates (N ¼ 3). Parameters

Slurries S

pH (H2O) Electrical conductivity (mS cm
1) Dry matter (g kg
1) Total C (g C kg
1) Soluble organic C (g C kg
1) Total N (g kg
1) Soluble organic N (g N kg
1) NH4 þ eN (g N kg
1) NO3
eN (mg N kg
1) NH4 þ /total N ratio soluble P (g kg
1) C/N Application rate mg C kg
1 dry soil mg NH4 þ eN kg
1 dry soil

AS a

LF b

ALF a

7.2 14.6d 113a 74.7a 3.2ab 3.2a 1.6a 1.3a 0.1b 0.41b 36c 23.3a

5.5 21.3a 116a 76.0a 1.9c 3.2a 1.7a 1.3a 1.6ab 0.41b 245a 23.7a

7.2 15.3c 21b 15.4b 2.4bc 2.3b 1.5a 1.2a 0.9ab 0.52a 21c 6.7b

5.5b 17.9b 25b 16.5b 3.5a 2.3b 1.7a 1.2a 2.1a 0.52a 51b 7.2b

1879b 32.7b

1912a 32.7b

550c 42.9a

589c 42.9a

Within rows, values presented with different superscripts are significantly different (P < 0.05) by Tukey test

.

D. Fangueiro et al. / Journal of Environmental Management 162 (2015) 1e8

On days 0, 4, 7, 14, 21, 28, 42, 49, 56, 63, 77 and 92, three PVC containers of each treatment were used to determine soil pH and mineral N content. Soil pH values were determined after 2-h of contact with occasional agitation in a soil/deionised water (1:5 w/ v) suspension (European Standards, 1999). Ammonium-N and NO3
eN were assessed by extraction with 2 M KCl at a 1:5 ratio (Houba et al., 1995). The net N mineralisation (NNM) and total N mineralisation over the incubation (Min tot) were calculated by Eqs. (1) and (2), respectively.

  NNMðtÞ mg N kge1 soil ¼ min NðtÞ
min Nðt0 Þ

(1)

3

the incubation. For this, cumulated GHG emissions were expressed as CO2-equivalents using the conversion factors of 298 and 25 for N2O and CH4, respectively (Forster et al., 2007). 2.3. Statistical analysis The data collected in our study were subjected to statistical analysis of variance, followed by Tukey test for comparison of means among treatments, at P < 0.05 using the software STATISTIX 7.0.

  ðNNMð92ÞÞ treatment
ðNNM ð92ÞÞcontrol  1000 Min tot mg N ge1 N applied ¼ total N applied

(2)

where, min N (t): mineral N content at time t, min N(t0): mineral N content at the beginning of the experiment.

3. Results and discussion

2.2.2. Measurement of gas fluxes To evaluate the NH3 emissions in each treatment, Kilner jars (L ¼ 100 mm, H ¼ 210 mm) were filled with 0.75 kg dry soil. The same six treatments (Control, S-I, S-S, AS-S, LF-S and ALF-S) were applied using the same procedure previously described and three replicates were considered per treatment. WHC was adjusted at 60% and controlled regularly. Ammonia fluxes were measured during the first 14 days of incubation using the acid traps method. An acid trap, containing 50 mL of 0.05 M orthophosphoric acid, was hung behind the soil inside the Kilner jars. Acid solution in each trap was replaced after 5, 22, 29, 45, 69, 94, 166, 220 and 333-h of the beginning of the experiment. Acid solutions collected at each sampling date, were analysed for total ammonium N (NH4 þ eN þ NH3eN) content using automated segmented-flow spectrophotometry (Houba et al., 1995). A second independent set of Kilner jars (L ¼ 100 mm, H ¼ 210 mm) was prepared as above described to follow the emissions in each treatment (3 replicates). Nitrous oxide, CO2 and CH4 fluxes from each Kilner jar were measured using modified lids fitted with two septa (Ø ¼ 10 mm) and a Teflon tube (Ø ¼ 4 mm, H ¼ 100 mm) to allow air sampling. Gas measurements were carried out at days 1, 3, 4, 5, 8, then every two days up to day 36 and once a week up to day 92. For each measurement, the Kilner jar was hermetically sealed by replacing the lid. The first gas sample (25 mL) was immediately taken (time-zero (T0) sample) using a syringe and flushed through evacuated 20 mL gas vials. After 0.5-h (T1) and 1-h (T2) of closure, the headspace of each Kilner jar was sampled again following the same sequential order to ensure that the same time had elapsed between sampling in each Kilner jar. The concentrations of the gas samples stored in vials were measured by gas chromatography (GC) using a GC-2014 (Shimadzu, Japan) equipped with an electron capture 63Ni detector (ECD), a thermal conductivity detector (TCD) and a flame ionization detector (FID) for N2O, CO2 and CH4 analysis, respectively. The GC accuracy was 50 ppbe100 ppm for N2O, 1 ppm to 1% for CO2 and 0.1 ppm to 1% for CH4. Gas fluxes were calculated by fitting linear regressions through the data collected at T0, T1 and T2 and then corrected for temperature and the amount of soil in each Kilner jar. Cumulative emissions were estimated by averaging the flux between two sampling occasions and multiplying by the time interval between the measurements. The global warming potential of the treatments considered here was compared based on the total amount of GHG emitted during

The concentration of NH4 þ in the Control treatment was lower than 12 mg NH4 þ eN kg
1 dry soil throughout the 92 days of incubation (Fig. 1A). The addition of S, AS, LF and ALF to soil led to a significant increase (P < 0.05) of the NH4 þ concentrations (54 mg NH4 þ eN kg
1 dry soil in S-S and AS-S, and 42 mg NH4 þ eN kg
1 dry

3.1. Soil nitrogen dynamics

Fig. 1. Soil NH4 þ (A) and NO3
(B) concentrations in the treatments during the experiment. Vertical bars represent the standard error of the mean (N ¼ 4).

4

D. Fangueiro et al. / Journal of Environmental Management 162 (2015) 1e8

soil in LF-S and ALF-S) in the first 14 days after application, which then declined to background levels by day 21 as nitrification occurred (Fig. 1B). Similar NH4 þ concentrations were observed in acidified (AS-S and ALF-S) and non-acidified (S-S and LF-S) treatments until day 7. Then, the NH4 þ concentrations decreased slower (P < 0.05) in acidified treatments relative to non-acidified treatments (Fig. 1A). From day 8 until the end of the incubation, the NH4 þ eN concentrations in AS-S and ALF-S treatments were significantly higher (P < 0.05) than in S-S and LF-S treatments, respectively. A similar effect of the acidification of pig and cattle slurry has been reported in previous laboratory studies with different soils (Fangueiro et al., 2010b, 2013) while, at field scale, no significant differences were observed between acidified and non-acidified pig and cattle slurry (Sørensen and Eriksen, 2009). It is still to note that the transformation of NH4 þ as via nitrification was faster in ALF-S than in AS-S treatment even if the amount of NH4 þ applied with ALF was higher than with AS. No significant differences (P > 0.05) were observed between S-I and S-S in terms of NH4 þ concentration during the whole incubation. At the beginning of the experiment, the NO3
concentration was close to 30 mg NO3
eN kg
1 in all treatments including Control (Fig. 1B). The NO3
concentrations increased during the 92 days of the experiment in all treatments, being significantly lower (P < 0.05) in S-S and AS-S treatments than in ALF-S or LF-S (Fig. 1B). The NO3
concentration in acidified and non-acidified treatments did not differ significantly (P > 0.05) during the experiment, although lower mean values were observed in ALF-S relative to LF-S (Fig. 1B). The effect of acidification on nitrification is not always evident since nitrification relies on the activity of nitrifying prokaryotes but also on the amount of available NH4 þ (Gandhapudi et al., 2006). It is generally assumed that nitrification becomes strongly reduced at pH lower than 6, due to a lower activity of the nitrifying prokaryotes that can be inhibited (Gandhapudi et al., 2006). However, in the present study, the effect of acidified slurry application on pH was not so noteworthy, since values of soil pH were always higher than 5, what can explain the low effect observed on nitrification (Fig. 2). Nevertheless, significantly lower (P < 0.05) pH values were observed in AS-S than in S-S at several sampling dates, while no significant differences (P > 0.05) were observed between LF-S and ALF-S.

Fig. 2. Soil pH evolution over the experiment. Vertical bars represent the standard error of the mean (N ¼ 4).

Higher NO3
content was observed in S-S relative to S-I in most sampling dates and NO3
contents in S-I and Control were similar in a few sampling dates, indicating that raw slurry incorporation has a significant impact on nitrification. Nitrogen immobilisation occurred in all treatments during the first 28 days followed by N mineralisation until the end of the experiment, except for S-I treatment, where N mineralisation occurred only in the last days of the incubation (Fig. 3A). No differences (P > 0.05) were observed between AS-S and ALF-S in terms of N mineralisation but a more intense N mineralisation was observed in LF-S relative to S-S during almost all the incubation period. Over the whole incubation, 32 and 10% of mineral N applied (equivalent to 14 and 4% of total N applied) was immobilised in the S-I and S-S treatments, respectively, whereas more than 9% of total N applied was mineralised in LF treatment (Fig. 3B). These results were expected since both LF and ALF have a C:N ratio (see Table 1) lower than 15, and according to Chadwick et al. (2000), such material should lead to N mineralisation while materials with higher C:N ratios should induce N immobilisation. Furthermore, the finest particle-size fractions in slurries are generally associated with N mineralisation and the coarser to N immobilisation (Fangueiro et al., 2008, 2010a). Hence, N mineralisation in LF might also be explained by the existence of finer particles in LF with a significantly lower dry matter, total C and C:N ratio relative to S (Fangueiro et al., 2012). On the other hand, acidification of raw slurry led to an increase (P < 0.05) of N mineralisation in most sampling dates but a decrease of N mineralisation was observed in some dates with acidification of LF. At the end of the experiment, N mineralisation in AS and ALF treatments was significantly higher (P < 0.05) than in all other

Fig. 3. Net N mineralisation (A) and potential N mineralisation (B) from treatments during the experiment. Vertical bars represent the standard error of the mean. Values presented with different letters are significantly different (P < 0.05) by Tukey test (N ¼ 4).

D. Fangueiro et al. / Journal of Environmental Management 162 (2015) 1e8

treatments (Fig. 3A). The impacts of slurry acidification on N mineralisation are not clear (Fangueiro et al., 2015a). Previous studies performed in our research group in controlled conditions, showed that acidification of pig slurry or derived LF decreases N immobilisation relative to raw slurry after soil application (Fangueiro et al., 2009, 2010b) but application of acidified cattle slurry rather than raw slurry led to a decrease of N mineralisation (Fangueiro et al., 2013). In the present study, the increase of N mineralisation in AS-S relative to S-S could be explained by an increase of finest particles in AS relative to S due to some complex dissociations and OM degradation (Fangueiro et al., 2015a). It can then be concluded that acidification of cattle-slurry or derived LF increases the potential organic N mineralisation and consequently, the amount of available N for crops. 3.2. Nitrogen emissions The daily NH3 fluxes increased significantly (P < 0.05) in S-I relative to Control, namely in the first 2 days after soil application and, significantly higher (P < 0.05) daily NH3 fluxes were observed in S-S treatment relative to all other treatments during the whole measurement period (Fig. 4A). The daily NH3 fluxes from LF-S, ALFS and AS-S treatments remained similar (P > 0.05) to those observed in Control during all the measurement period. Cumulative NH3 emissions from S-I and S-S were significantly higher (P < 0.05) than those observed in the remaining treatments that led to similar values (Table 3). Amounts of applied NH4 þ eN lost as NH3 in S-I (1.5%) and S-S (5.2%) treatments were significantly higher (P < 0.05) than in the other amended treatments where less

Fig. 4. Ammonia (A) and N2O (B) fluxes following the application of each treatment. Vertical bars represent the standard error of the mean (N ¼ 3).

5

than 0.7% of NH4 þ eN applied was lost (Table 3). Ammonia emissions from LF were significantly reduced (P < 0.05) relative to raw slurry. Slurry separation by centrifugation is the most efficient process to separate dry matter (Hjorth et al., 2010) since it removes the largest particles of undigested material from the derivate LF. Hence, the infiltration of LF in soil should be greater than infiltration of raw-slurry because the dry matter content controls greatly the physical processes ruling the movement of slurry into and within the soil and consequently NH3 emissions (Sommer et al., 2003, 2004; Ndegwa et al., 2008). Once in the soil, NH4 þ ions derived from LF would have been subsequently immobilised on cation exchange sites, thus reducing the potential for NH3 volatilisation (Sanz-Cobena et al., 2011). As already reported in several studies (Frost et al., 1990; Kai et al., 2008; Fangueiro et al., 2015a), soil application of acidified slurry reduced significantly (P < 0.05) NH3 emissions relative to raw slurry (Table 3). By lowering the slurry pH to 5.5, the ratio NH4 þ :NH3 is strongly modified, with about 98.00e99.98 % of NH4 þ (Fangueiro et al., 2015a). Our results are in agreement with those reported by Stevens et al. (1992) who observed a 75% decrease in NH3 emission after soil application of acidified cattle-slurry (pH 6.5) and Fangueiro et al. (2015b) who found a 81% decrease in NH3 emission from acidified cattle-slurry and LF (pH 5.5). A significant (P < 0.05) increase of N2O fluxes relative to Control was observed immediately (day 1) after application in all amended treatments (Fig. 4B). The highest increase was observed in S-I treatment and the lowest in AS-S while all other treatments led to similar increases. Nitrous oxide fluxes in all amended treatments decreased strongly during the first 5 days and then peaked twice: first on days 8e14 and then on days 29e32. The highest N2O fluxes were observed in S-I and AS-S on the first peak, and in LF-S and ALF-S for the second peak. Furthermore, the second N2O peak occurred on day 29 in LF-S, ALF-S and S-S but only on day 32 in S-I and AS-S, probably due to a delay/decay in the nitrification process. It is still to note that the first N2O peak was observed when a more intense nitrification started in all treatments, while the second N2O peak occurred when NO3
concentration in most treatments reached a plateau indicating that nitrification cessed or slow down significantly. Nitrous oxide fluxes in acidified treatments (AS-S and ALF-S) were generally lower than in non-acidified treatments (S and LF) even if differences were not always statistically different (P > 0.05) (Fig. 4B). Furthermore, N2O emissions in S-I treatments were always significantly higher or similar to those from S-S indicating that slurry incorporation might enhance N2O emissions due to anaerobic conditions that favours denitrification. Cumulative N2O emissions from amended treatments were significantly higher (P < 0.05) than in Control (Table 3). The cumulative N2O emissions from AS-S treatment were significantly lower (P < 0.05) in ca. 26% relative to S-S treatment. However, no differences (P > 0.05) were observed between ALF-S and LF-S treatment neither between S-I and S-S concerning cumulative N2O emissions (Table 3). It is known that N2O emissions from agricultural soils are originated from the nitrification and denitrification processes, and denitrification is the main source of N2O fluxes from soil amended with animal slurry (Kool et al., 2011). Both slurry and derived LF supplied high concentrations of NH4 þ , organic N and different amounts of readily available organic C that increased N2O emissions relative to Control (Butterbach-Bahl et al., 2013; Müller and Clough, 2013) (Table 1). The total N2O emissions from LF-S, S-S and S-I were similar (Table 3); higher emissions were expected from S-S and S-I treatments than LF-S, since the supply of organic compounds as C source was higher in raw slurry application relative to LF (Table 1). The absence of differences on N2O emissions between LF-S and S-S or S-I is not clear but might be due to NH4 þ immobilisation (Fig. 3B)

6

D. Fangueiro et al. / Journal of Environmental Management 162 (2015) 1e8

Table 2 Carbon dioxide fluxes (mg C kg
1 day
1 dry soil) following the application of each treatment (N ¼ 3). Treatments

Time (days) 1

Control S-I S-S AS-S LF-S ALF-S

3 c

4 22a 21a 4c 15ab 8bc

4 bc

3 2c 4bc 9ab 8abc 11a

5 c

2 13ab 10abc 10abc 16a 7bc

8 c

1 13ab 2bc 19a 6bc 0c

b

3 45a 39a 16a 35a 10a

10

12

15

17

20

22

25

29

32

a

a

a

a

a

a

a

a

a

3 9a 34a 22a 16a 8a

9 13a 4a 16a 10a 3a

8 19a 14a 16a 6a 2a

0 0a 9a 17a 3a 1a

2 14a 17a 14a 4a 4a

1 30a 20a 13a 6a 4a

2 5a 6a 5a 3a 3a

0 8a 9a 8a 4a 3a

0 16a 9a 8a 3a 0a

36e92 1a 1a 1a 1a 1a 1a

Within columns, values presented with different superscripts, are significantly different (P < 0.05) by Tukey test.

Table 3 Cumulative gaseous emissions during the experiment (N ¼ 3). Parameters

Treatments Control

S-I

S-S

AS-S

LF-S

ALF-S

NH3 emissions (mg N kg
1 dry soil) (% NH4 þ applied) (% N applied) N2O emissions (mg N kg
1 dry soil) (% N applied) N emissions (mg N kg
1 dry soil) (% N applied) CO2 emissions (mg C kg
1 dry soil) (% C applied) CH4 emissions (mg C kg
1 dry soil) (% C applied) C emissions (mg C kg
1 dry soil) (% C applied)

0.1c N.A. N.A. 0.2d N.A. 0.3d N.A. 143.0b N.A. 0.0b N.A. 143.0b N.A.

0.5b 1.5b 0.6b 1.6ab 2.0ab 2.1bc 2.6b 500.1a 26.6b 1.3a 0.1a 501.3a 26.7b

1.7a 5.2a 2.1a 1.3bc 1.7bc 3.0a 3.8a 459.7a 24.5b 0.6a 0.0a 460.3a 24.5b

0.2c 0.6c 0.3c 1.0c 1.2c 1.2c 1.5c 428.2a 22.4b 0.1b 0.0a 428.3a 22.4b

0.1c 0.2c 0.1c 1.5ab 1.9ab 1.6bc 2.0b 311.6a 56.6a 0.0b 0.0a 311.5a 56.6a

0.1c 0.2c 0.1c 1.8a 2.3a 1.9bc 2.4b 150.9b 25.6b 0.0b 0.0a 150.9b 25.6b

Within rows, values presented with different superscripts are significantly different (P < 0.05) by Tukey test.

and lower NO3
concentrations in S-S and S-I relative to LF-S that might have decreased the amount of substrate for denitrifiers. In our study, N2O emissions from acidified cattle-slurry were significantly reduced relative to non-acidified cattle-slurry, but no differences (P > 0.05) were observed between acidified LF and nonacidified LF, even if NO3
content in LF-S was higher than in ALF-S while similar NO3
contents were observed in AS-S and S-S treatment over the incubation (Table 3). Our results are consistent with a previous study (Fangueiro et al., 2010b) where a similar decrease of N2O emissions after soil application of acidified pig slurry was reported and attributed to a delay or inhibition of the nitrification process. Furthermore, the lower N2O emissions from AS-S relative to S-S observed here may also be explained by the CO2 losses that occurred during the acidification process (Fangueiro et al., 2013), which reduced the amount of soluble organic C (Table 1) available for denitrification and consequently reduced the N2O emission. On the other hand, the high CO2 emissions observed immediately after S or LF application to soil (not observed in ALF-S or AS-S) might have led to O2 depletion, creating better conditions for denitrification and N2O emissions. The total N (NH3 þ N2O) emissions from Control were significantly lower (P < 0.05) relative to all other treatments (Table 3). A significantly (P < 0.05) lower total N emission was observed in AS-S relative to S-S treatment (less 60%). However, similar values (P > 0.05) were observed in LF-S and ALF-S. More than 2.6% of the applied N was released in the S-I and S-S treatments against less than 1.5% in AS-S treatment (Table 3). Thus, application of acidified slurry rather than raw slurry seems of great interest to minimise N losses but a combined treatment separation þ acidification did not bring any benefit relative to NH3 and N2O emissions. Even if no significant differences were observed between AS-S and LF-S in terms of N emissions, soil application of AS rather than LF might be motivated by the lower costs associated to acidification compared to solideliquid separation. Furthermore, one has to bear in mind

that AS contains higher amounts of other nutrients as phosphorous or potassium than LF (Fangueiro et al., 2009). 3.3. Carbon emissions Higher CO2 fluxes were observed over the first 22 days of the experiment in all treatments, with 45e73% of the total CO2 emissions occurring during this period, followed by a reduction until the end of the measurements (Table 2). The daily CO2 fluxes from all S amended soils (S-I, S-S and AS-S) were higher relative to all LF amended soils (LF-S and ALF-S). Furthermore, CO2 fluxes were lower in ALF-S and AS-S than in LF-S and S-S, respectively, even if differences were not always significant (Table 2). A peak of CO2 emission was observed in S-I, S-S and LF-S on day 8 while it occurred on day 10 in AS-S and no peak was observed in ALF-S. The cumulative CO2 emissions from S amended soil (S-I, S-S and AS-S) were significantly higher (P < 0.05) than in Control with increases between 200 and 250% (Table 3). Although not statistically significant (P > 0.05), cumulative CO2 emissions were lower (P < 0.05) in ca. 32% when S-S was applied relative LF-S. Nevertheless, the cumulative CO2 emissions were lower (P < 0.05) in ca. 52% when ALF-S was applied relative to LF-S amendment (Table 3). Cattle-slurry application increases the soil microbial activity and consequently CO2 emissions after amendment due to the organic matter mineralisation (Pereira et al., 2010). In our study, a reduction of CO2 emissions was observed in LF-S relative to S-S (Table 3). This difference is in agreement with the significant lower amount of total C added by LF (0.5 g kg
1 dry soil in LF against 1.6 g kg
1 dry soil in S) (Table 1). The CO2 emissions from AS-S and ALF-S were reduced in 7 and 52% relative to S-S and LF-S, respectively (Table 3). Such decrease might be attributed to a decrease of microbial activity and consequently oxygen consumption in soil amended with acidified slurry (Ottosen et al., 2009). The effect of acidification on CO2 emissions

D. Fangueiro et al. / Journal of Environmental Management 162 (2015) 1e8

was more pronounced in LF, in agreement with results from Fangueiro et al. (2013) who observed a stronger effect of acidification on CO2 emissions in cattle-slurry with low dry matter content than in cattle-slurry with high dry matter content. Furthermore, a great amount of CO2 was released during the acidification process (Stevens et al., 1989; Fangueiro et al., 2015a) and consequently lower C was available after soil application in acidified treatments. Measurable CH4 fluxes were observed only during the first 4 days of experiments in S-I and S-S with values of 0.3e0.4 mg C kg
1 day
1soil in S-I and 0.2 mg C kg
1 day
1soil in SS, while no CH4 emissions were detected in all the other treatments (data not shown). Previous studies reported that CH4 emissions occurred for a short period (few days) after cattle-slurry application to soil, and are related with volatilisation of the CH4 produced during storage and initially dissolved in the cattle-slurry rather than from methanogenesis in soil (Pereira et al., 2010; Chadwick et al., 2011; Fangueiro et al., 2015a). Indeed, the incubation was performed in aerobic conditions and at a low value of soil WHC what ensures a good aeration of soil and should have minimised CH4 production. Animal slurry treatment by acidification reduces the methanogenesis and consequently CH4 emissions during storage (Fangueiro et al., 2015a), what explained the lower CH4 emissions observed in AS-S relative to S-S. The absence of CH4 emissions in LF-S or ALF-S treatments might be explained by the release during the separation process of most of the dissolved CH4 initially present in slurry. The total C emissions (CO2 þ CH4) from S and LF amended soils were significantly higher (P < 0.05) relative to Control, with increases of 120 and 220%, respectively (Table 3). It is to note that the C emissions from AS-S were similar (P > 0.05) to S-S treatment while the C emissions from ALF-S treatment were significantly lower (P < 0.05), in ca. 52%, than in LF-S. More than 50% of the applied C was released in the LF-S treatment whereas in all the other amended treatments, less than 35% of the applied C was released (Table 3). 3.4. Global warning potential and implications at farm scale The GWP in amended treatments was significantly higher (P < 0.05) than in Control with increases between 330 and 670% (Fig. 5). The GWP from AS-S treatment was significantly lower (P < 0.05) than in S-S (less 28%), while the GWP from ALF-S was similar (P > 0.05) to that of LF-S treatment (Fig. 5).

Fig. 5. Global warming potential estimated in each treatment during the experiment period. Values presented with different letters are significantly different (P < 0.05) by Tukey test (N ¼ 3).

7

The results obtained here at laboratory scale indicated clearly that there is no drawback in applying acidified cattle-slurry rather than raw slurry relative to GHG emissions with the strong advantage of avoiding or minimising NH3 emissions. Furthermore, when considering the GWP, it is clear that slurry acidification can be more than a solution to minimise NH3 emissions at field scale. Nevertheless, the results obtained here with an acidic soil might be different if considering a neutral or alkaline soil. The other strategies proposed here to minimise NH3 emissions were equivalent to acidification in terms of NH3 emissions but not so efficient to decrease N2O emissions relative to raw slurry. Furthermore, acidification should be competitive relative to solideliquid separation even in terms of costs if acidification is performed in the storage tanks immediately before soil application (Fangueiro et al., 2015a). Further studies at farm scale will be required to validate our results and assess the most cost effective cattle-slurry acidification treatment. Acknowledgements This work was supported by national funds by FCT e Portuguese Foundation for Science and Technology, under the project UID/ AGR/04033/2013, project PEst-OE/AGR/UI0528/2014 and project PTDC/AGR-PRO/119428/2010. David Fangueiro has received a grant from the FCT (SFRH/BPD/84229/2012). References Bittman, S., Dedina, M., Howard, C.M., Oenema, O., Sutton, M.A. (Eds.), 2014. Options for Ammonia Mitigation: Guidance from the UNECE Task Force on Reactive Nitrogen. Centre for Ecology and Hydrology, Edinburgh, UK. Butterbach-Bahl, K., Baggs, E.M., Dannenmann, M., Kiese, R., ZechmeisterBoltenstern, S., 2013. Nitrous oxide emissions from soils: how well do we understand the processes and their controls? Philos. T. R. Soc. B 368 (1621), 20130122. Chadwick, D.R., John, F., Pain, B.F., Chambers, B., Williams, J., 2000. Plant uptake of nitrogen from the organic nitrogen fraction of animal manures: a laboratory experiment. J. Agric. Sci. 134, 159e168. Chadwick, D., Sommer, S., Thorman, R., Fangueiro, D., Cardenas, L., Amon, B., Misselbrook, T., 2011. Manure management: implications for greenhouse gas emissions: a review. Anim. Feed Sci. Tech. 166e167, 514e531. EEA, 2013. Annual European Union Greenhouse Gas Inventory 1990e2011 and Inventory Report 2013; Submission to the UNFCCC Secretariat. Copenhagen, Denmark. http://www.eea.europa.eu//publications/european-union-greenhouse-gas-inventory-2013 (accessed: 07.15.). EEA, 2014. Ammonia (NH3) Emissions (APE 003). Copenhagen, Denmark. http:// www.eea.europa.eu/data-and-maps/indicators/eea-32-ammonia-nh3-emissions-1/assessment-4 (accessed: 07.15.). European Standards, 1999. Soil Improvers and Growing Media. European Standards (ES) 13037. Determination of pH. European Committee for Standardization, Brussels. Fangueiro, D., Bol, R., Chadwick, D., 2008. Assessment of the potential N mineralization of six particle size fractions of two different cattle slurries. J. Plant Nutr. Soil Sci. 171, 313e315. Fangueiro, D., Ribeiro, H., Vasconcelos, E., Coutinho, J., Cabral, F., 2009. Treatment by acidification followed by solideliquid separation affects slurry and slurry fractions composition and their potential of N mineralization. Bioresour. Technol. 10, 4914e4917. Fangueiro, D., Gusm~ ao, M., Grilo, J., Porfírio, G., Vasconcelos, E., Cabral, F., 2010a. Proportion, composition and potential N mineralisation of particle size fractions obtained by mechanical separation of animal slurry. Biosyst. Eng. 106, 333e337. Fangueiro, D., Ribeiro, H., Coutinho, J., Cardenas, L., Trindade, H., Cunha-Queda, C., Vasconcelos, E., Cabral, F., 2010b. Nitrogen mineralization and CO2 and N2O emissions in a sandy soil amended with original or acidified pig slurries or with the relative fractions. Biol. Fertil. Soils 46, 383e391. Fangueiro, D., Lopes, C., Surgy, S., Vasconcelos, E., 2012. Effect of the pig slurry separation techniques on the characteristics and potential availability of N to plants in the resulting liquid and solid fractions. Biosyst. Eng. 113, 187e194. Fangueiro, D., Surgy, S., Coutinho, J., Vasconcelos, E., 2013. Impact of cattle slurry acidification on carbon and nitrogen dynamics during storage and after soil incorporation. J. Plant Nutr. Soil Sci. 176, 540e550. Fangueiro, D., Hjorth, M., Gioelli, F., 2015a. Acidification of animal slurry e a review. J. Environ. Manage. 149, 46e56. Fangueiro, D., Pereira, J., Macedo, S., Trindade, H., Vasconcelos, E., Coutinho, J., 2015b. Nitrogen use efficiency of surface application of treated cattle slurry as alternative to slurry injection. Biol. Fertil. Soils (under review).

8

D. Fangueiro et al. / Journal of Environmental Management 162 (2015) 1e8

FAO, 2006. World Reference Base for Soil Resources. A Framework for International Classification, Correlation and Communication. Food and Agriculture Organization of the United Nations, Rome, Italy. World Soil Resources Reports 103. Forster, P., Ramaswamy, V., Artaxo, P., Berntsen, T., Betts, R., Fahey, D.W., Haywood, J., Lean, J., Lowe, D.C., Myhre, G., Nganga, J., Prinn, R., Raga, G., Schulz, M., Van Dorland, R., 2007. Changes in Atmospheric Constituents and in Radiative Forcing. Cambridge University Press, Cambridge, United Kingdom. Frost, J.P., Stevens, R.J., Laughlin, R.J., 1990. Effect of separation and acidification on ammonia volatilization and on the efficiency of slurry nitrogen for herbage production. J. Agric. Sci. 115, 49e56. Gandhapudi, S.K., Coyne, M.S., Angelo, E.M.D., Matocha, C., 2006. Potential nitrification in alum treated soil slurries amended with poultry manure. Biores. Technol. 97, 664e670. Hjorth, M., Christensen, K.V., Christensen, M.L., Sommer, S.G., 2010. Solid-liquid separation of animal slurry in theory and practice. A review. Agron. Sustain. Dev. 30, 153e180. Houba, V.J.G., Van der Lee, J.J., Novozamsky, I., 1995. Soil Analysis Procedures e Other Procedures, Part 5B, sixth ed. Wageningen Agricultural University, The Netherlands. IUSS Working Group WRB, 2006. World Reference Base for Soil Resources. FAO, Rome. World Soil Resources Reports 103. Kai, P., Pedersen, P., Jensen, J.E., Hansen, M.N., Sommer, S.G., 2008. A whole-farm assessment of the efficacy of slurry acidification in reducing ammonia emission. Eur. J. Agron. 28, 148e154. Kool, D.M., Dolfing, J., Wrage, N., Van Groenigen, J.W., 2011. Nitrifier denitrification as a distinct and significant source of nitrous oxide from soil. Soil Biol. Biochem. 43, 174e178. Müller, C., Clough, T.J., 2013. Advances in understanding nitrogen flows and transformations: gaps and research pathways. J. Agric. Sci. 152 (S1), 34e44. Ndegwa, P.M., Hristov, A.N., Arogo, J., Sheffield, R.E., 2008. A review of ammonia emission mitigation techniques for concentrated animal feeding operations.

Biosyst. Eng. 100, 453e469. Nyord, T., Liu, D., Eriksen, J., Adamsen, A.P.S., 2013. Effect of acidification and soil injection of animal slurry on ammonia and odour emission. In: Proceedings from the 15th RAMIRAN Conference, Versailles, France. In: http://www.ramiran. net/doc13/Proceeding_2013/documents/S9.14.pdf (accessed: 07.15.). Ottosen, L.D.M., Poulsen, H.V., Nielsen, D.A., Finster, K., Nielsen, L.P., Revsbech, N.P., 2009. Observations on microbial activity in acidified pig slurry. Biosyst. Eng. 102, 291e297. Pereira, J., Fangueiro, D., Chadwick, D.R., Misselbrook, T.H., Coutinho, J., Trindade, H., 2010. Effect of cattle slurry pre-treatment by separation and addition of nitrification inhibitors on gaseous emissions and N dynamics: a laboratory study. Chemosphere 79, 620e627. Sanz-Cobena, A., Misselbrook, T., Camp, V., Vallejo, A., 2011. Effect of water addition and the urease inhibitor NBPT on the abatement of ammonia emission from surface applied urea. Atmos. Environ. 45, 1517e1524. Sommer, S.G., Hutchings, N.J., 2001. Ammonia emission from field applied manure and its reduction e invited paper. Eur. J. Agric. 15, 1e15. Sommer, S.G., Genermont, S., Cellier, P., Hutchings, N.J., Olesen, J.E., Morvan, T., 2003. Processes controlling ammonia emission from livestock slurry in the field. Eur. J. Agron. 19, 465e486. Sommer, S.G., Hansen, M.N., Søgaard, H.T., 2004. Infiltration of slurry and ammonia volatilisation. Biosyst. Eng. 88, 359e367. Sørensen, P., Eriksen, J., 2009. Effect of slurry acidification with sulphuric acid combined with aeration on the turnover and plant availability of nitrogen. Agric. Ecosyst. Environ. 131, 240e246. Stevens, R.J., Laughlin, R.J., Frost, J.P., 1989. Effect of acidification with sulphuric acid on the volatilization of ammonia from cow and pig slurries. J. Agric. Sci. 113, 389e395. Stevens, R.J., Laughlin, R.J., Frost, J.P., 1992. Effects of separation, dilution, washing and acidification on ammonia volatilization from surface-applied cattle slurry. J. Agric. Sci. 119, 383e389.