POSTHARVEST NITROUS OXIDE EMISSIONS FROM A SUBTROPICAL OXISOL AS INFLUENCED BY SUMMER CROP RESIDUES AND THEIR MANAGEMENT ( 1 )

Nitrous oxide (N2O) is the most important non-CO2 greenhouse gas and soil management systems should be evaluated for their N2O mitigation potential. This research evaluated a long-term (22 years) experiment testing the effect of soil management systems on N2O emissions in the postharvest period (autumn) from a subtropical Rhodic Hapludox at the research center FUNDACEP, in Cruz Alta, state of Rio Grande do Sul. Three treatments were evaluated, one under conventional tillage with soybean residues (CTsoybean) and two under no-tillage with soybean (NTsoybean) and maize residues (NTmaize). N2O emissions were measured eight times within 24 days (May 2007) using closed static chambers. Gas flows were obtained based on the relations between gas concentrations in the chamber at regular intervals (0, 15, 30, 45 min) analyzed by gas chromatography. After soybean harvest, accumulated N2O emissions in the period were approximately three times higher in the untilled soil (164 mg m-2 N) than under CT (51 mg m-2 N), with a short-lived N2O peak of 670 mg m-2 h-1 N. In contrast, soil N2O emissions in NT were lower after maize than after soybean, with a N2O peak of 127 g m-2 h-1 N. The multivariate analysis of N2O fluxes and soil variables, which were determined simultaneously with air sampling, demonstrated that the main


INTRODUCTION
The atmospheric concentrations of the three main greenhouse gases (GHGs) -carbon dioxide (CO 2 ), nitrous oxide (N 2 O), and methane (CH 4 ) -have increased rapidly in the last few decades, which is a phenomenon associated to anthropogenic activities (Mosier et al., 1991;1998).Worldwide, it is estimated that agriculture contributes with approximately 22 % to the total CO 2 emissions, 80 % to the total N 2 O and 55 % to the total CH 4 emissions (IPCC, 2007).On the other hand, it is estimated that 75 % of CO 2 , 94 % of N 2 O, and 91 % of the CH 4 emissions in Brazil come from agricultural activities (Embrapa, 2006).Although emitted in smaller amounts, the global warming potential of N 2 O is 296 times higher than of CO 2 (IPCC, 2007).Thus, the marked contribution of N 2 O emission by agriculture and the potential of this gas to trap thermal infrared radiation justify the need of characterization of the soil management practices with a way to reduce soil N 2 O emission (Six et al., 2004).
Most N 2 O is produced by nitrification and denitrification processes in the soil (Firestone & Davidson, 1989).Nitrification, which requires aerobic conditions, depends on NH 4 + supply and is mediated by autotrophic bacteria, whereas denitrification is performed by anaerobic heterotrophic bacteria, which depend on the availability of labile organic C (Rice et al., 1988) and NO 3 -.Their activity is intensified in anaerobic environments or soil microsites with low O 2 availability (Moreira & Siqueira, 2006;Aita & Giacomini, 2007).In addition to these soil variables, temperature, water-filled pore space (WFPS), and pH have also been identified as N 2 O-production-controlling variables in agricultural soils, since they affect the activity of nitrifying and denitrifying bacteria as well (Maag & Vinther, 1996;Weitz et al., 2001;Dalal et al., 2003).
Among the agricultural practices, N fertilizers, soil tillage, and legume grain crops have been cited regarding their effect on soil N 2 O emissions.The inefficient use of N fertilizers and continuous legume cropping are mainly responsible for stimulating soil N 2 O emissions to the atmosphere since they increase the soil mineral N content and influence the soil labile C supply (Dalal et al., 2003;Gomes et al., 2009).In turn, soil tillage affects the majority of the N 2 Oemission-controlling soil variables and higher emissions of this GHG have usually been reported in soils under NT than under CT (Rice & Smith, 1982;Liu et al., 2007).This finding has been attributed to a lower O 2 diffusion due to soil compaction (Vinten et al., 2002;Baggs et al., 2003;Liu et al., 2006), mainly in periods of frequent and intense rainfall (Jantalia et al., 2006), and to a higher microbial biomass activity, consuming O 2 and forming anaerobic microsites (Baggs et al., 2006).However, most results regarding N 2 O emission were obtained in temperate soils, and limited information is available for tropical and subtropical Oxisols, where favorable soil physical properties of soil structure and aeration as well as rapid water-infiltration may restrict N 2 O production and emission by denitrification (Jantalia et al., 2006;Metay et al., 2007;Jantalia et al., 2008).
The main objective of this study was to evaluate the long-term effect of soil management systems on N 2 O emissions from a Rhodic Hapludox in the postharvest period of summer grain crops (soybean and maize) under the subtropical climatic conditions of the state of Rio Grande do Sul, Brazil.

Experimental site characterization
The study was carried out in a 22-year experiment conducted at the Fundação Centro de Experimentação e Pesquisa Fecotrigo (FUNDACEP), Cruz Alta, state of Rio Grande do Sul (RS), Brazil.The research center FUNDACEP (latitude 28 o 36 ' S, longitude 53 o 40 ' W, 409 m asl) lies in the center of the plateau of the state (Rio Grande do Sul, 1994).The soil is classified as typic dystrophic Red Latosol by the Brazilian Soil Taxonomy System (Embrapa, 2005), and as Rhodic Hapludox by the USDA Soil Taxonomy (USDA, 1999).The regional climate is humid, type Cfa 2a, according to Köppen's classification (Moreno, 1961), with a mean annual rainfall of 1,755 mm (mean from 1974 to 2006 -FUNDACEP weather station) with an annual drainage volume of 513 mm and evenly distributed rainfall during the year, slightly concentrated in the spring (Maluf, 2000).The annual mean temperature is 18.7 °C (mean from 1998 to 2006 -FUNDACEP weather station), with a minimum temperature of 8.6 °C in July and maximum of 30 °C in January (Moreno, 1961).

Treatments
Soil N 2 O emissions were evaluated in CR-1, in May 2007 in three plots in this crop rotation system.Two were assessed after soybean harvest, in succession to black oat, under NT and CT, and the third after maize harvest, in succession to black oat + common vetch, under NT.Maize was harvested on February 28, 2007, and radish oil was planted immediately after the harvest, while soybean was harvested on April 18, 2007.The cover crop used after soybean harvest was black oat + common vetch, sown with a seed drill SHM 1517.Conventional tillage (CT) consisted of plowing to a depth of 0.20 m, followed by harrow disking to a depth of 0.15 m, within an interval of two days between plowing and harrow disking.In the NT System, the winter cover crop was sown on the maize and soybean residues and soil tilling was restricted to the crop rows.In each treatment, an area of 2 x 2 m was delimited for soil N 2 O emission and soil sampling.

Measurements of nitrous oxide emissions
Soil N 2 O emissions were evaluated eight times in 24 days with sampling intervals ranging from 1 to 5 days, immediately after soil tillage.Air samples for N 2 O analysis were taken using polyvinyl chloride (PVC) static chambers (height 0.20 m, diameter 0.25 m).Immediately before air sampling, the atmosphere inside the chamber was homogenized with an internal fan for 30 s and the internal temperature monitored with a digital thermometer.The chambers were placed on an aluminum base which was inserted 0.05 m deep into the soil and maintained in the experimental plot during the entire evaluation period.During sampling, the internal chamber atmosphere was isolated by filling water into a channel at the upper part of the base upon which the PVC chamber was supported.Three static chambers were used per treatment.
The air was sampled from 10 to 11 a.m., as proposed by Jantalia et al. (2008), 0, 15, 30, and 45 min after the closure of the chambers, using 20 mL polypropylene syringes equipped with valves in the top of the chambers.Air samples were cool-stored and taken to the Environmental Biogeochemical Laboratory of the Federal University of Rio Grande do Sul-Porto Alegre -RS, Brazil, where they were analyzed by gas chromatography within less than 24 hours after sampling.N 2 O was determined using a gas chromatograph Shimadzu GC 2014 Greenhouse model, equipped with an electron capture detector (ECD), and a Porapack Q column at 70 °C, using N 2 as the carrier gas at a flow of 26 mL min -1 , injector temperature of 250 °C and detector temperature of 325 °C.N 2 O flows were calculated by the equation /dt is a change in the gas concentration within the chamber in the time interval (L min -1 ); MW is the molecular weight of the gas; V and T are the volume (L) and the inner temperature (°K) of the chamber; P is the pressure within the chamber, assumed as 1 atm and R is the universal gas constant (L atm K -1 mol -1 ).The variation of the gas concentration over time was obtained by the angular coefficient of the adjusted linear regression.The daily mean N 2 O emission was calculated by the mean value of the three chambers installed in each treatment and the emission of the evaluation period was obtained by the integration of the area below the curve, which was established by the interpolation of the daily emission values.

Soil analysis
Simultaneously to air sampling, the soil temperature (0.10 m), gravimetric moisture, and the NH 4 + and the NO 3 -contents (0-0.10 m) were monitored.The soil temperature was evaluated by geothermometers and soil moisture calculated by drying a sub-sample under forced air circulation at 105 ºC for 24 h.The soil NH 4 + and NO 3 -were extracted from the soil samples using 1.0 mol L -1 KCl solution, determined by distillation in semi-micro Kjeldahl and acid titration (Tedesco et al., 1995).The microbial biomass activity was assessed based on the quantification of the soil CO 2 -C flux (Baggs et al., 2003), determined by the concomitant CO 2 and N 2 O analysis of the same air samples.In a GC 2014 chromatograph, CO 2 was reduced to CH 4 and quantified by a flame ionization detector (FID) at 250 °C.WFPS was calculated based on the values of soil bulk density and gravimetric moisture.

Statistical analysis
The effect of the soil management systems on N 2 O emissions was evaluated by descriptive data analysis based on the mean standard deviation.Soil N 2 O emission, soil temperature, microbial biomass activity, WFPS, and the NO 3 -and NH 4 + soil contents were subjected to multivariate analysis based on Principal Coordinate Analysis (PCA), using software MULTIV (Pillar, 1997).The data of each variable were transformed by vectors (centering and normalization) and the product moment was used as similarity measure between the management systems on the different evaluation dates.Only coordinates with eigenvalues > 1 and explaining > 30 % of variability were considered.Coefficients of correlation > 0.60 were included in the interpretation of the principal coordinates.The scatter diagram was complemented with cluster analysis, which allowed the verification of the similarity patterns among soil management systems in each field evaluation.

Meteorological conditions during the study period
Mean daily air temperatures ranged from 4 to 18 °C during the study period, and rainfall from 0.1 to 69 mm (Figure 1).On average, the weather conditions were colder and wetter than in the normal mean.In May, the rainfall volume exceeded the normal average of the period by 24 % (179 mm), while the mean air temperature was 13 % (14 °C) lower than the normal average in May.

Soil N 2 O emission
Soil N 2 O emissions were influenced by the tillage systems and the soybean and maize post-harvest residues (Figure 1).From the soybean crop residues, N 2 O emissions were higher in NT soil (164 mg m -2 N) than CT soil (51 mg m -2 N).In the NT soil, the N 2 O emissions peaked in the first two evaluations, with a maximum of 670 mg N m -2 h -1 recorded 19 days after soybean harvest.Afterwards, there was an exponential decrease of N 2 O emission, although in general slightly higher than the N 2 O emissions recorded in the CT soil (Figure 1), with a N 2 O peak (143 g m -2 h -1 N) on the 29 th day after soybean harvest, which was more than four times lower than the respective value in NT soil (Figure 1).In an Oxisol, also in Rio Grande do Sul, Jantalia et al. (2008) found fluxes in the range of 80 -150 mg m -2 h -1 N for both NT and CT in the first three weeks after sowing.The N 2 O emission ranges were therefore similar in the cited and the present study, although Jantalia et al. (2008) reported lower differences in soil N 2 O fluxes between tillage systems and a lower N peak emission compared to our results.On the whole-year basis, Jantalia et al. (2008) observed mean fluxes ranging from 8.1 to 18.7 mg m -2 h -1 N. In a three-year study, Rochette et al. (2008) found that emissions from NT plots in a poorly drained heavy clay soil were rarely different (p > 0.05) from those of moldboard plow plots when fluxes were low, but were greater during almost all episodes of high emissions, similarly to the trend observed here.In the study of Rochette et al. (2008), a few peaks of N 2 O fluxes in NT plots reached exceptionally high levels between 2.000 and 16.000 mg m -2 h -1 N. The peaks of N 2 O fluxes from temperate, poorly drained soil (Gleysoil) were much higher than reported in our study with tropical, welldrained soil (Oxisol).
Other studies, in agreement with the present findings, also showed higher soil N 2 O emission from NT than from CT soils (Rice & Smith, 1982;Smith et al., 2001;Yamulki & Jarvis, 2002;Liu et al., 2006;Liu et al., 2007;Rochette et al., 2008;Gomes et al., 2009).The main reasons, according to those authors, is related to soil compaction and higher WFPS, resulting from the higher soil moisture, and in some cases, lower soil porosity in NT.These variables are inversely related to soil O 2 diffusion, therefore favoring N 2 O production by denitrifying bacteria activity.Nevertheless, in the soil investigated, the small difference in bulk density in the NT soil (1.23 Mg m -3  in the 0-0.10 m layer) in comparison to CT soil (1.20 Mg m -3 at the same depth), and the considerable similarity of the WFPS values (Figure 2) cannot explain the 10 times higher N 2 O emissions from NT than from CT soil in the first sampling (Figure 1).This suggests the existence of other soil and crop management factors controlling N 2 O emission in NT soil, which will be approached by the multivariate analysis.At this same site, Chavez et al. (2009) and Pes (2009) found that the seeding operation in NT increased soil gas exchanges.These authors suggest that even the minimal soil disturbance by the driller was enough to increase GHG emission.In addition, Varner et al. (2003) reported that mortality of fine roots, due to their fast turnover, has the potential to contribute significantly to N 2 O production and fluxes.Keller et al. (2005) found that soil compaction also increased the N 2 O flux from a North Brazilian Oxisol.After maize harvest, the mean soil N 2 O emission in NT soil was 34 mg m -2 N in the evaluation period, with an N 2 O peak of 127 mg m -2 h -1 N (Figure 1), which was approximately five times below the maximum observed after soybean harvest.The lowest N 2 O soil emission in NT after maize harvest indicates, mainly, the microbial biomass assimilation of mineral N for high C:N maize residue decomposition, mainly of NO 3 -(Figure 2), decreasing the substrate to the N 2 O production by denitrification and nitrification processes in soil.
According to Aita & Giacomini (2007), the decomposition of N-rich crop residues, eg., of soybean, is less influenced by the lower soil-crop residue contact in NT soils than by a high C:N ratio of the crop residues.This would explain the rapid increase in the soil mineral N contents observed under NT after soybean harvest (Figure 2).Consequently, the input of residues with a low C:N ratio represents a high input of labile C, which stimulated microbial activity and may have resulted in the considerable O 2 consumption, creating anaerobic microsites in the soil surface layer and favoring N 2 O production in the NT soil (Smith et al., 2003).On the other hand, in CT soil this would have not occurred at the same intensity since plowing and harrowing tillage harmonized the crop residue distribution in the 0-0.20 m layer, disfavoring the formation of such microsites.
The positive correlation between CO 2 and N 2 O fluxes (p < 0.007, n = 8, r = 0.52) supports this hypothesis, indicating that the highest soil N 2 O emission occurred simultaneously with the peak of soil microbial activity (Rice et al., 1988).In studies carried out in England and Germany, respectively, Baggs et al. (2003) and Sehy et al. (2003) also verified positive relations between the CO 2 and N 2 O fluxes, emphasizing that mulch on the NT soil surface favored moisture conservation, besides gradually labile C input, essential in the denitrification process.The relation between labile C and the denitrifying activity in NT soil results in an increased N 2 O flux, as reported previously by Rice et al. (1988) and Liu et al. (2007).
The N 2 O emission peaks observed in this study were superior to those reported in earlier studies in a South Brazilian Alfisol (Gomes et al., 2009) and Oxisol (Jantalia et al., 2006;Metay et al., 2007;Jantalia et al., 2008), in which soil N 2 O emission rates were approximately five times lower than those found in this study.However, these studies were carried in spring-summer, and the first with maize was grown after common vetch (Vicia sativa L.), and the second and fourth during soybean growth.Therefore, in both experiments the growing plants absorbed available mineral soil NO 3 -and because it uncorroborated to N 2 O emission.This study conducted after soybean harvest possibly explains the higher soil N 2 O emission rates, since in this period, emission peaks of up to 3,500 mg m -2 h -1 N have been verified elsewhere (Yang & Cai, 2005;Ciampitti et al., 2005;2008), resulting from the large supply of labile organic C by the crop residues and from the exudates and products of nodule and root senescence, which are promptly metabolized by soil microorganisms (Rochette et al., 2004;Ciampitti et al., 2005;Yang & Cai, 2005).In a Canadian heavy clay soil, Rochette et al. (2004) reported a short period of high emissions (up to 480 mg m -2 h -1 N) after soybean harvest, presumably resulting from the high denitrification following the input of available C and N substrates when the soil water content was high.In central Iowa (USA), soybean fields exhibited N 2 O peak fluxes of 630 mg m -2 h -1 N, after rainfall events (Parkin & Kaspar, 2006).These results are similar to those found in our study.

Multivariate analysis of the main variables that control the soil N 2 O flux
The PCA analysis was used to identify the driving variables of N 2 O emissions.Through the ordination analysis of the values of N 2 O emissions and the soil variables, a scatter diagram was obtained for the three treatments (Figure 3).Axis 1 explained 36 % of the data variability, and axis 2 explained 34 % (Table 1).The variables with significant loading on axis 1 were N 2 O emission, microbial activity (CO 2 -C), soil temperature, and WFPS.On axis 2, soil NO 3 -content had positive loading and soil NH 4 + content had a significant negative loading (Table 1).
The soil N 2 O emission of the group III (1 st , 2 nd , 4 th , and 5 th samplings in NT soybean postharvest and in the 4 th in NT maize postharvest) were the highest and were associated with soil temperature, microbial biomass activity, WFPS, and soil NO 3 -content (Figure 3).The relation of these variables with N 2 O flux in this group indicated N 2 O production by denitrification, which was favored by the high WFPS values (Smith et al., 2003;Liu et al., 2007) and the consumption of soil O 2 due to the microbial activity oxidizing labile organic C (Figure 2) in NT soil, and consequent formation of anaerobic microsites.These results were similar to those reported by Khalil and Baggs (2005), who claimed that the denitrification process contributes considerably to agricultural soil N 2 O emissions and that NO 3 -availability is the major driving variable regulating the magnitude of these emissions.On the other hand, in CT soil, the occurrence of anaerobic microsites is not probable even under soybean residues.In the other samplings (Groups I, II and IV), there was no prevailing control factor in the N 2 O emission, which may explain the lowest emission rates.In a previous study carried out in South Brazil, Jantalia et al. (2008) observed a poor relationship between WFPS and N 2 O flux.These authors associate this result to frequently low WFPS, of less than 40 %, despite heavy rain events (above 60 mm day) due to fast drainage into the Oxisol.
The findings of this study evidence the need of enhanced crop rotations under NT after legume grain crops, in particular soybean, to mitigate soil N 2 O emissions.For this purpose, the improvement of the synchronism between N release by the legume crop residues and the demand of the subsequent crop (graminea or crucifera) with high N demand may be a pormising strategy to reduce mineral N availability, and consequently, reduce N 2 O emissions.For the evaluation of the effect of soil management on reduction of soil N 2 O emissions in future research, the whole crop rotation cycle should idealistically be contemplated, taking the emission of the other GHGs into account as well as soil C accumulation and the CO 2 -C costs of crop production systems.

CONCLUSIONS
The N 2 O emission from soil after soybean harvest is higher under no-tillage than under conventional tillage.Results indicate that the denitrification process is the major source of N 2 O in the wet and cool autumn of Rio Grande do Sul, and that the main driving variables of N 2 O emissions are microbial biomass activity, soil temperature, water-filled pore space (WFPS), and soil NO 3 -content.The maximum soil N 2 O emissions found in this study after soybean harvest were higher than previously reported for Brazilian Oxisols.Nitrous oxide emissions can be reduced by the cultivation of maize in succession to soybean, and by the sowing of high-N-demanding winter cover crops immediately after soybean harvest.

Figure 1 .
Figure 1.Rainfall, air temperature in May, and nitrous oxide (N 2 O) emission from a subtropical Oxisol in autumn (May 7 to 30, 2007), in Southern Brazil as influenced by tillage system (conventional CT and notillage NT) after soybean (CTsoybean and NTsoybean) and after maize harvest (NTmaize), in the rotation sequence black oat/soybean/black oat + common vetch/maize/radish oil/wheat/soybean.Bars represent the mean standard error.

Figure 2
Figure 2. Microbial activity, contents of NO 3 -and NH 4 + , WFPS, and soil temperature of a subtropical Oxisol in autumn (May 7 to 30, 2007), in Southern Brazil as influenced by tillage systems (conventional tillage-CT and no-tillage-NT) after soybean (CTsoybean and NTsoybean) and after maize harvest (NTmaize), in the rotation sequence black oat/soybean/black oat + common vetch/maize/radish oil/wheat/ soybean.

Figure 3 .
Figure 3. Scatter diagram of the variables and tillage systems (conventional-CT and no-tillage-NT) evaluated after soybean (CTsoybean and NTsoybean) and after maize harvest (NTmaize), in eight samplings (1-8), in the crop rotation black oat/soybean/black oat+common vetch/maize/radish oil/wheat/soybean under Rhodic Hapludox during autumn (May 7 to 30, 2007), in Southern Brazil.The clusters formed by the similarity between the management systems are indicated as I, II, III, and IV.