Nitrogen fluxes from irrigated common ‐ bean as affected by mulching and mineral fertilization

The objective of this work was to measure the fluxes of N2O‐N and NH3‐N throughout the growing season  of  irrigated  common‐bean  (Phaseolus vulgaris),  as  affected  by mulching  and mineral  fertilization. Fluxes of N2O‐N and NH3‐N were evaluated in areas with or without Congo signal grass mulching (Urochloa ruziziensis) or mineral fertilization. Fluxes of N were also measured in a native Cerrado area, which served as reference. Total N2O‐N and NH3‐N emissions were positively related to the increasing concentrations of moisture, ammonium, and nitrate in the crop system, within 0.5 m soil depth. Carbon content in the substrate and microbial biomass within 0.1 m soil depth were  favoured by Congo signal grass and  related  to higher emissions  of  N2O‐N,  regardless  of  N  fertilization.  Emission  factors  (N  losses  from  the  applied  mineral nitrogen) for N2O‐N (0.01–0.02%) and NH3‐N (0.3–0.6%) were lower than the default value recognized by the Intergovernmental Panel on Climate Change. Mulch of Congo signal grass benefits N2O‐N emission regardless of N fertilization.


Introduction
Common-bean (Phaseolus vulgaris L.) is an important staple food in Brazil.During the dry season in the Brazilian Cerrado, between May and September, irrigated common-bean crop covers 182 thousand hectares (National Company of Food Supply, 2013), commonly cultivated under no-tillage system on the crop residues of maize or Congo signal grass (Urochloa ruziziensis).Center pivot is the most used irrigation system, by which yields can reach up to 2.5 Mg ha -1 , usually 40% higher than nonirrigated ones (National Company of Food Supply, 2013).Besides irrigation, high inputs of mineral N (mainly urea, with total amounts reaching 245 kg ha -1 ) also explain the higher yields of this crop system (Posse et al., 2011).
Environmental impacts of such an intensive crop system can be high, and its evaluation is a important step towards sustainable production.
Quantifying the emission of greenhouse gases (GHG) from crop systems is becoming increasingly important, considering worldwide trends towards full carbon accounting.Nitrous oxide (N 2 O) is one of the GHG which contributes to climate change.In the Brazilian agriculture, emissions of N 2 O-N increased in 24% from 1994 to 2005, and agricultural soils are responsible for more than 95% of N 2 O-N emissions, including direct N 2 O-N emission sources -mineral fertilizers, crop residues, grazing animals, animal off-field produced manure, biological fixation, etc. -and indirect ones -leaching, runoff, and atmospheric deposition (Cerri et al., 2009).Local assessment of GHG emissions from Brazilian agriculture systems have been reported (Carvalho et al., 2006;Metay et al., 2007;Jantalia et al., 2008;Zanatta et al., 2010;Cruvinel et al., 2011); however, the relative scarcity of data makes it difficult to assess the overall impact of irrigated common-bean under no-tillage on N 2 O-N and NH 3 -N fluxes in the Brazilian Cerrado.
The effect of no-tillage on decreasing CO 2 emissions is largely known (Piva et al., 2012).There are, however, many constrains related to the exact role of this cropping system in mitigating N 2 O-N/ NH 3 -N emissions, especially because the impact of management on soil microbial activity (Bardgett et al., 2008) is still largely unknown.N Losses via volatilization can be high under no-tillage because the fertilizer usually has a restricted contact with the soil, due to the presence of mulch or crop residues (Da Ros et al., 2005).Additionally, large amounts of N can be lost via N 2 O-N fluxes under irrigation, where denitrification is stimulated by anaerobic conditions.Rochette (2008) reported that no-tillage generally increased N 2 O-N emissions in poorly-aerated soils, and Woodward et al. (2009) suggested that adopting reduced or no-tillage systems, in some cases, enhances the emissions of N 2 O-N from soil, especially due to increased denitrification.However, even within presumably homogenous regions, spatial and temporal variability and range of N 2 O-N emissions is large (Pérez et al., 2001); hence, local assessments of N 2 O-N emissions are indispensable to estimate the overall impact of agricultural systems on GHG emissions (Reay et al., 2012).
The objective of this work was to measure fluxes of N 2 O-N and NH 3 -N, throughout the growing season of irrigated common-bean, as affected by mulching and mineral fertilization.

Materials and Methods
The N-fluxes measurements were made within an area cultivated with common-bean under no-tillage, irrigated via center pivot, on a clayey Rhodic Ferralsol, located at Embrapa Arroz e Feijão, in Santo Antônio de Goiás, GO,Brazil (16°29'17"S and 49°17'57"W).Manual static chambers were used to measure N 2 O-N fluxes, and manual open static chambers were used to measure NH 3 -N fluxes.Congo signal grass was desiccated with glifosate (0.5 kg i.a.ha -1 ) 15 days prior to the seeding of common-bean, in order to form mulch.In the treatments without Congo signal grass mulch, common-bean was cultivated on crop residues of maize.Common-bean ('BRS Agreste') was sown in June 9 th and harvested in September 20 th , 2008.From June to August 2008, there was no precipitation, air temperature ranged from 13º to 34ºC, and average daily evaporation was 154 mm.
The effects of mulching with Congo signal grass (Urochloa ruziziensis) and of mineral fertilization (NPK) on N-fluxes were evaluated.The evaluated treatments were: common-bean cultivated with Congo signal grass mulch and fertilizer; common-bean cultivated with mulch, but not fertilized; common-bean cultivated without mulch, but fertilized; and common-bean cultivated without mulch or fertilizers.Each treatment was assessed with six manual static chambers, used to measure the N-fluxes, each chamber corresponding to a replicate.The manual static chamber used to measure N 2 O-N fluxes consisted of a metal base (0.38 m wide x 0.58 m long) covering a soil area of 0.22 m 2 and a plastic cap (0.1 m height) fixed on the metal base, similar to the chamber used by Alves et al. (2012).When closed, the volume of the chamber was 19.8 L.
Soil chemical and physical properties were determined in 1 m trenches according to Silva (2009) (Table 1).Fertilized treatments received 400 kg ha -1 fertilizer (5-30-15 N-P-K) applied at planting.Urea, 200 kg ha -1 (45% N) was applied via fertigation by the center pivot system: 100 kg ha -1 in the 27 th day after sowing (DAS), and 100 kg ha -1 at 41 DAS.During fertigation, the chambers and corresponding area for soil sampling were covered with plastic to prevent any chance of contamination of the treatments without N fertilization.For these treatments, the same amount of water without N was applied via irrigation.Total irrigation (via center pivot), throughout the growing period, was 419 mm (about 9 mm per event at 29 hour intervals).
Fluxes of N 2 O-N were measured after irrigation events at June 10,11,12,13,14,17,23;July 1,7,8,9,11,12,13,14,16,21,22,23,24,25,26,30;August 6,13,20,28;and September 3,10,18.Gas samples were taken between 9:00 to 11:00 a.m., as recommended by Alves et al. (2012).Gases accumulated in the static chamber in a period of 20 min were collected using a manual vacuum pump.From one chamber in each treatment, samples were taken at five-minute intervals, during the collection period of 20 min.Additionally, three air samples were taken to be used as controls.Soil temperature at 0.05 m soil depth next to the chambers was measured simultaneously with N 2 O-N flux sampling.Concentration of N 2 O inside each chamber was calculated as the difference between N 2 O concentration in the air and N 2 O concentration inside the chamber.Gas samples were analyzed by gas chromatography with an electron capture detector (ECD) auto system XL (Perkin Elmer, São Paulo, SP, Brazil), calibrated with certified N 2 O standards of 350 and 1,000 ppb.Fluxes of N 2 O-N were calculated according to Alves et al. (2012).When estimating total emitted N 2 O-N, negative fluxes (lower than the concentration of the air), were set to zero.Total emissions were calculated by interpolating and integrating mean fluxes over time.Unfertilized treatments were used to calculate emission factor.
Fluxes of NH 3 -N were quantified with manual open static chambers, each one covering 0.008 m 2 of ground.These chambers were made from 2 L plastic bottles of 0.1 m diameter, from which bottoms have been removed, according to Jantalia et al. (2012).Inside each bottle, a 70 mL plastic pot was hung and contained a polyethylene foam strip, moistened with 40 mL sulfuric acid solution (1.5 mol L -1 H 2 SO 4 + 4% glycerol).These chambers were installed in the sowing lines, immediately after sowing, and were regularly replaced, when the foam strips were changed, allowing the soil under the chambers to receive the treatment with urea and irrigation.Foam strips were changed 19 times at June 11,14,18,26;July 2,8,11,14,16,21,23,25,30;August 6,13,20,28;and September 3,10.Fluxes of NH 3 -N were calculated based on N recovery of 57% for field conditions, according to Araújo et al. (2009).
Soil moisture, ammonium (NH 4 + ) and nitrate (NO 3 -) concentrations were determined from 100 g soil samples collected within 0.5 m soil depth simultaneously with N 2 O-N flux sampling.Around 10 g of soil was weighed, before and after drying in an oven for 24 hours at 105ºC.Soil moisture (cm 3 cm -3 ) was calculated considering the soil bulk density (g cm -3 ) determined for each treatment (Table 1).The available NH 4 + and NO 3 -were determined according to Mulvaney (1996).
Table 1.Soil properties of a clayey Rhodic Ferralsol cultivated with irrigated common-bean under no-tillage with or without mulching, and in a noncultivated area (Cerrado), at 0-0.1 m, 0.1-0.3, and 0.3-0.5 soil depths.Soil microbial biomass was quantified from 200 g soil samples collected within 0.1 m soil depth, during common-bean flowering stage (56 DAS).N and C in the soil microbial biomass were determined by fumigation and extraction methods (Brookes et al., 1985;Vance et al., 1987).
Seventeen days after sowing the common-bean, maize crop residues or Congo signal grass mulch were collected in four spots of 0.08 m 2 within each treatment, in order to determine total dry matter mass, and total C and N content, using an elemental analyzer 2400 Series II CHNS/O (Perkin Elmer, São Paulo, SP, Brazil).
The analysis of variance was done using the generalized linear model procedure (Proc GLM) of SAS/STAT (SAS Institute, Cary, NC, USA).To contrast treatment effects, we used the Tukey's Studentized range test (at 5% probability).Correlations were determined between variables related to soil management and total emission of N 2 O-N and NH 3 -N; and a linear model was fitted to test whether the concentration of N 2 O inside the chambers was dependent on the time of incubation.The magnitude of correlations and linear adjustment were assessed with the squared Pearson correlation coefficient between observed and predicted values (R 2 ).Parameter estimates are shown with respective nominal significance value (p-value) and standard error of estimates.

Results and Discussion
Approximately 34% of the N 2 O-N measured fluxes were negative (lower than concentration of N 2 O in the air).Most (54%) of the negative fluxes were measured in unfertilized areas (34% in Cerrado and 20% in unfertilized treatments).The concentration changes of N 2 O over time inside the chambers were not significant for all treatments, except for the treatments fertilized with mulching and without mulching or fertilizer (Figure 1).Apart of the low fluxes itself, damping of flux diffusion control, pressure effects, and leaking -a source of error when using static chambers (Kroon et al., 2008) -are possible reasons for the observed nonsignificance.However, manual static chambers were reported to be the best method to get a direct measurement of small-scale spatial variability because it works well under all climate conditions and it has low cost, although it requires high workload (Drösler  , 2008).Fluxes of N 2 O-N (mg m -2 per day) ranged from -0.63±0.09(Cerrado) to 1.74±0.84(unfertilized mulching), and the ones of NH 3 -N, from 4.66±2.44(Cerrado) to 62.85±12.52(fertilized mulching) (Figure 2).Throughout the growing season, significantly higher N 2 O-N fluxes occurred during 3 (1.71±1.01),31 (0.47±0.34), and 79 (0.24±0.07) days after sowing event (DAS), in the fertilized mulching; and during 3 (1.74±0.85)and 79 (0.18±0.21)DAS in the unfertilized mulching (Figure 2).Increasing soil moisture and available mineral N were the main causes for higher N 2 O-N fluxes during 3 and 31 DAS in fertilized mulching.Passianoto et al. ( 2003) also observed higher N 2 O-N losses in a nontilled soil covered by mulch, than in tilled soil, during the first week following mineral fertilization.However, cumulative fluxes of N 2 O-N during period 3, after fertilization and irrigation (42 to 50 DAS), were significantly higher than during period 1 (1 to 21 DAS), in the fertilized treatment without mulching, and were higher than during period 4 (57 to 100 DAS) in all treatments (Figure 3).Fluxes of NH 3 -N, to the contrary, were more equally distributed over the growing season, with no significant difference for cumulative fluxes among periods for fertilized mulching.Moreover, no significant differences for NH 3 -N fluxes between fertilized and unfertilized treatments were detected.Cumulative fluxes were higher in period 3 than in period 2 for the fertilized treatment without mulching (Figure 3).Cumulative fluxes of NH 3 -N were significantly higher in period 4 (57 to 92 DAS), during crop maturity, than in periods 1, 2, and 3, in the unfertilized treatments.
Total emitted NH 3 -N was not statistically different between fertilized and unfertilized treatments; however, it was significantly higher in the fertilized mulching than in the unfertilized treatment without mulching.Total emitted N 2 O-N was significantly higher with Congo signal grass mulch than without it (Table 2).Total emission of both N 2 O-N and NH 3 -N and the contents of moisture and nitrate within 0.5 m soil depth were significantly lower in the Cerrado than in the crop system (Table 2).Total emission of N-N 2 O and NH 3 -N was positively correlated with contents of soil moisture, nitrate, and ammonium within 0.5 m soil depth.Total N 2 O-N emission was positively correlated with total C added by mulch or crop residues, and with C and N in microbial biomass within 0.1 m soil depth Figure 3. Cumulative N 2 O-N and NH 3 -N fluxes within four periods over growing season of irrigated common-bean under no-tillage, with or without mulching and mineral fertilization (NPK).Periods after fertilization and irrigation are represented by the periods 1, 2 and 3; and crop maturity, by period 4, when no fertilization was applied.Bars represent standard deviation of mean (n=6).Values followed by the same letter are not significantly different, according to Tukey's Studentized test, at 5% probability.
Table 2. Total N 2 O-N and NH 3 -N emissions, soil moisture within 0.5 m soil depth, concentration of soil nitrate (N-NO 3 -) and ammonium (N-NH 4 + ) 0.5 m depth, carbon (CMB) and nitrogen (NMB) in soil microbial biomass within 0.1 m depth, and total carbon and nitrogen in mulch and crop residues, in an irrigated common-bean crop system under no-tillage (1) .(Table 3).Carbon in microbial biomass for treatments with mulch was equivalent to that observed in the Cerrado (Table 2), meaning that the soil under this mulch was as biologically active as the Cerrado soil.Because of the irrigation, however, soil moisture was higher in the crop system.
In an irrigated crop system, the amount of water applied can be associated with the magnitude of N 2 O-N emission (Scheer et al., 2012).In the same Ferralsol of the present study, Metay et al. (2007) observed that N 2 O-N fluxes were positively related to increasing water-filled pore space.Zanatta et al. ( 2010) found the highest fluxes of N 2 O-N during 45 days after soil management associated with high water-filled pore space and increased organic C and mineral N, which accounted for 30% of N 2 O-N emissions in a year.During two years of field experiment, Jantalia et al. (2008) observed the highest fluxes of N 2 O-N after periods of high rainfall associated with mineral fertilization.However, the C content in the mulch and microbial activity, favoured by the Congo signal grass, induced a higher emission of N 2 O-N in treatments with mulching, regardless of N fertilization.The presence of mulch has been considered important for high N 2 O-N emissions in no tillage systems (Gomes et al., 2009).Baggs et al. (2003) reported that surface-mulching residues of rye resulted in higher N 2 O-N losses than with the incorporation of residues.They attributed this result to the degradable C of rye in the presence of anaerobic conditions under mulch.Available organic C, associated with increasing water-filled pore space, can induce anaerobic sites, which can enhance denitrification of the available nitrate in soil, resulting in high N 2 O-N fluxes in no tillage system (Giacomini et al., 2006).
Nitrogen losses (N 2 O-N + NH 3 -N) from the applied mineral N (emission factor) varied from 0.01% (fertilized without mulching) to 0.02% (fertilized mulching) for N 2 O-N, and from 0.3% (fertilized mulching) to 0.6% (fertilized without mulching) for NH 3 -N.These values were much lower than the default value proposed by the Intergovernmental Panel on Climate Change (IPCC), which is 1% for N 2 O-N, ranging from 0.3% to 3%; and 10% for NH 3 -N, ranging from 3 to 30%.In an irrigated common-bean crop system in the Brazilian Cerrado, with broadcasted urea, Cruvinel et al. (2011) found an emission factor for N 2 O-N of 0.20%.The emission of N 2 O-N found in the present study was equivalent to that measured by Metay et al. (2007) 2012) observed 1.9% to 2.4% NH 3 -N losses of the urea broadcasted in an irrigated system.Turner et al. ( 2012) found NH 3 -N losses ranging from 1.8 to 23% of the added N fertilizers in different cropping systems.They attributed the variability of these results to soil-climate conditions, such as soil moisture, temperatures, and wind speed.Due to the wide range of values that can be obtained, estimation of N losses from fertilizer application in crop systems is challenging.

Conclusions
1. Carbon content of the mulch and soil microbial activity, both favoured by Urochloa ruziziensis, enhances the emissions of N 2 O-N in treatments with mulching, regardless of N fertilization.
2. Total N 2 O-N and NH 3 -N emissions are positively related to the concentrations of moisture, ammonium, and nitrate within 0.5 m soil depth, when mineral fertilization is used.
3. Emission factors for N 2 O-N and NH 3 -N under the conditions of this study were lower than the default value recognized by the Intergovernmental Panel on Climate Change.
Table 3. Pearson correlation coefficients representing the relationship between total emission of N (N 2 O-N + NH 3 -N) and soil management variables in an irrigated common-bean crop system under no-tillage (1) .

Figure 1 .
Figure 1.N 2 O accumulation into manual static chambers placed in an irrigated common-bean crop system under no-tillage: A, with Congo signal grass mulch and fertilizer; B, with Congo signal grass mulch, without fertilizer; C, without Congo signal grass mulch and with fertilizer; D, without Congo signal grass mulch and without fertilizer; and E, noncultivated area (Cerrado).Black bars indicate the standard deviation of mean (n=19); dotted lines represent the fitted linear model; values between brackets represent the standard error of estimate.*Significant at 10% probability.

Figure 2 .
Figure2.Fluxes of N 2 O-N and NH 3 -N throughout the growing season of irrigated common-bean under no-tillage: with Congo signal grass mulch and fertilizer (■Mulch+NPK); with mulch no fertilizer (Mulch); without mulch, and with fertilizer (•NPK); without mulch of Congo signal grass, and without fertilizer (○No mulch or NPK); and from a noncultivated area (▲Cerrado).Dotted lines 1, 2, and 3 indicate the periods after mineral N fertilization; and period 4 indicates crop maturity, when no fertilization was applied.Bars represent standard deviation of means (n=6).*Significantdifferences among treatments, except for the noncultivated area (Cerrado), within each DAS, according to Tukey's Studentized test, at 5% probability.
Values followed by the equal letters are not significantly different, according to Tukey's Studentized test, at 5% probability.
of 0.03%, in a rainfed rice crop system under no tillage, with Congo signal grass mulch, in a Cerrado Ferralsol.Literature information on NH 3 -N fluxes is rather variable, since they are often obtained through different methodologies.Using semi open static chambers, Da Ros et al. (2005) observed 17% NH 3 -N losses of the urea broadcasted on crop residues, in a subtropical Acrisol, in Southern Brazil.Using open chambers, Jantalia et al. (