Straw management effects on global warming potential and yield-scaled greenhouse gas emissions in a subtropical rice ecosystem

: Global warming potential (GWP) of rice paddies depends on straw management. This study evaluated methane (CH 4 ) and nitrous oxide (N 2 O) emissions and soil C stocks to determine GWP and yield-scaled GWP under different strategies and intensities of rice straw management in a subtropical climate. We hypothesized that decreasing soil management intensity and straw incorporation in the soil would reduce GWP. Methane fluxes were substantially higher during the rice growing season than in the off-season, while the opposite was observed for N 2 O fluxes. The cumulative emissions of CH 4 during the growing season among the straw management strategies evaluated ranged from 165.8 to 586.0 kg ha -1 . Annual CH 4 emissions were lower when soil and straw received some type of management compared to no-tillage. Daily N 2 O fluxes ranged from -2.8 to 201.7 g ha -1 day -1 ; cumulative N 2 O emissions during the off-season ranged from 455.2 to 2816.5 g ha -1 . During the off-season, strategies to reduce N 2 O emissions include post-harvest straw incorporation using a disc harrow, winter straw removal, and ryegrass cropping. Soil organic C stocks ranged from 35.96 to 38.36 Mg ha -1 . Straw management using a disc harrow reduced soil organic C stocks, with more adverse effects than straw removal. Soil and rice straw management did not affect rice grain yield, with an average of 10.4 Mg ha -1 . Methane emissions were the main component of GWP in all straw management systems. The contribution of N 2 O emissions to GWP was small and mostly (>85 %) determined by off-season emissions. Yield-scaled GWP ranged from 0.64 to 1.06 Mg CO 2 eq Mg -1 yield and was lower when soil and straw management systems occurred shortly after the rice harvest. Our results indicate that soil and straw management immediately after rice harvest reduces CH 4 emissions, GWP, and yield-scaled GWP.


INTRODUCTION
Rice, which is world cultivated on almost 165 M ha (Faostat, 2021), serves as a staple food for more than 3 billion people, and demand for cereal, including rice, is expected to increase in the coming years (Alexandratos and Bruinsma, 2012;Zhu et al., 2018).Flooded rice comprises an area of approximately 93 M ha and is responsible for more than 75 % of the world's rice production (Rao et al., 2017).Areas in which rice is grown under flood irrigation are among the main sources of methane (CH 4 ) released to the atmosphere (Smith et al., 2014), as the anoxic conditions created by flooding encourage the decomposition of organic material and thus favor methanogenesis (Le Mer and Roger, 2001).Additionally, rice paddies can significantly contribute to nitrous oxide (N 2 O) emissions, especially during the rice off-season (Zschornack et al., 2018), a period in which the soil is not flooded and is subjected to wetting and drying cycles that favor the production of N 2 O by nitrification and denitrification processes (Congreves et al., 2018).Although rice production currently accounts for only approximately 11 % of total global agricultural GHG emissions (Pachauri et al., 2014), in southern Brazil, CH 4 from rice fields accounts for 20 % of GHG emissions from the agricultural sector (MCTI, 2016).Strategies to mitigate these emissions are crucial to the sustainability of rice production in this subtropical ecosystem (Bayer et al., 2014).
In rice fields, the timing of tillage or straw management has been considered essential to reduce the global warming potential (GWP) resulting from soil CH 4 and N 2 O emissions (Sander et al., 2014;Yang et al., 2018).Bayer et al. (2015) showed that anticipation of tillage with straw incorporation into the soil in the autumn, shortly after the rice harvest in the Southern Hemisphere, was effective in reducing seasonal CH 4 emissions by 24 % compared to tillage and incorporation of straw into the soil in spring, just before sowing and soil flooding.According to the authors, with the anticipation of management, straw decomposes under aerobic conditions, producing CO 2 as a final product of microbial respiration, thus decreasing the amount of substrate for methanogenesis during the rice growing season under flood conditions.In Rio Grande do Sul state, the anticipation of tillage with straw incorporation and no-tillage is carried out in approximately 61 % of the area, while conventional tillage is still used in 30 % of the area; pregerminated rice fields occupy 9 % (Sosbai, 2018).An aspect that deserves attention is the combination of different implements and times for carrying out soil preparation and its impacts on the annual emissions of CH 4 and N 2 O, encompassing the growing season and off-season periods.Most studies focus on the impact of agricultural practices on GHG emissions during the growing season without considering that off-season emissions can compromise potential environmental benefits in terms of annual emissions (Yang et al., 2018;Zschornack et al., 2018).
Soil tillage and straw management in areas under flooded rice can also affect soil organic C stocks, but the rates at which this occurs are not well-known in subtropical ecosystems.Although the effect of maintaining crop residues on the soil surface in the no-tillage system on the accumulation of C in the soil is recognized (Pandey et al., 2014), rice cropping in a subtropical Gleisol for 10 years in conventional tillage and no-tillage does not promote differences in organic C stocks in the 0.00-0.20 m layer (Nascimento et al., 2009).Little is known about the impacts of soil and straw management (soil surface or incorporation) on the storage of C in subtropical soils subjected to irrigated rice cultivation.Another aspect to be highlighted is that greater stocks of C in the soil can also mean greater emissions of CH 4 in the long term (Liu et al., 2014).Therefore, the benefits of soil and straw management practices on GHG emissions should ideally be evaluated by their global warming potential (GWP), which considers the emissions of the three main GHGs and their respective forcing indices (1 for CO 2 , 28 for CH 4 and 265 for N 2 O) according to Pachauri et al. (2014), and yield-scaled GWP, which yields the GWP per unit of rice grain yield (Mosier et al., 2006;Bayer et al., 2015).
Rev Bras Cienc Solo 2023;47nspe:e0220134 Although soil CH 4 and N 2 O emissions in paddy rice fields have been documented (Bayer et al., 2014(Bayer et al., , 2015;;Moterle et al., 2013;Zschornack et al., 2016), to the best of our knowledge, this study is a pioneering one in assessing the effect of soil and straw management on net soil CO 2 fluxes (using changes in soil C as a proxy) and on seasonal GHG (CH 4 and N 2 O) emissions in a subtropical paddy rice ecosystem in southern Brazil.We hypothesized that decreasing soil management intensity and straw incorporation in the soil would decrease the GWP of the field.This study aimed to evaluate the potential of different soil and straw management practices to mitigate GWP and yield-scaled GWP in a subtropical paddy rice field.

Site and experiment description
The experiment was carried out over two years at the Federal University of Santa Maria, located in Rio Grande do Sul State, southern Brazil (29° 45' S, 53° 42' W, approximately 95 m altitude).The climate of the experimental area is humid subtropical, Cfa2 according to Köppen's classification system.The monthly mean air temperature varies from 14 °C for the coldest month (June) to 25 °C for the hottest month (January).The mean annual rainfall is 1,600 mm without a well-defined dry season.Air temperature and rainfall data were obtained from an automated weather station located 500 m from the experimental site.The soil was classified as Planossolo Háplico (Santos et al., 2018), corresponding to an Albaqualf (Soil Survey Staff, 2014).The soil layer of the 0.00-0.10m presented the following physical and chemical properties at the installation of the experiment: 210 g kg -1 of sand, 220 g kg -1 of clay, 4.7 g kg -1 of carbon, pH(H 2 O) (1:1 soil:water) 5.9, 16.2 mg dm -3 of P, 144 mg dm -3 of K, 6.8 cmol c dm -3 of Ca 2+ , 1.5 cmol c dm -3 of Mg 2+ and bulk density 1.42 Mg m -3 .Prior to the experiment, the site was planted with rice-fallow succession for two years.plot surface immediately after grain harvest.The amount of rice straw that was returned and distributed on the soil surface of all treatments at the beginning of the experiment was adjusted to 6.5 Mg ha -1 in 2010 and 11.3 Mg ha -1 in 2011, with the exception of the SR treatment, which retained the plant culms (0.10 m) that were not cut by the harvester.

Soil and Straw Management
In the NT treatments, the straw was not impacted by handling.In NT+R, ryegrass was sown using 60 kg ha -1 of seeds.Ryegrass received only topdressing nitrogen fertilization of 30 kg ha -1 at the tillering stage.During ryegrass cultivation, cuts were made, and the residue was removed from the plots to simulate the effect of animal grazing.Straw management using KR treatments was carried out with a knife roller at a water depth of approximately 0.10 m.In the DH treatments, straw management was carried out with a disc harrow in moist soil.The use of the disc harrow in August in the treatments DH PH + DH PS and KR PH + DH PS seeks to improve soil leveling, which is a common practice among rice farmers in southern Brazil.The schedule of operations performed during the two years for soil and straw management systems is summarized in table 1.
The cultivar Puitá Inta-CL was sown in both seasons in October, using 90 kg -1 ha -1 of seeds.The plant population was adjusted to 250 plants m -2 .At the time of sowing, all plots were fertilized at doses equivalent to 300 kg ha -1 of the formula N-P 2 O 5 -K 2 O 5-20-20.
Nitrogen topdressing was carried out three times: the 1st time-before the flood (50 kg ha -1 ), four leaves stage; the 2nd time-at the beginning of rice tillering (35 kg ha -1 ); and the 3rd time-at the beginning of rice panicle differentiation (35 kg ha -1 ).At sowing, the amount of straw remaining on the soil surface in the NT treatment was 1.5 and 1.1 Mg ha -1 in 2010/11 and 2011/12, respectively, and in the NT+R treatment, it was 3.6 and 2.2 Mg ha -1 .The soil was flooded 20 days after rice emergence and remained with a 0.10 m water depth during the entire cultivation period.Irrigation was suspended 7 and 5 days before harvest in 2010/11 and 2011/12, respectively.In both seasons, the grain yield in each plot was evaluated in a central area equivalent to 2.55 m 2 (10 rows × 1.5 m length), for which the grain yield was expressed at 13 g kg -1 of moisture.After the productivity evaluation at each harvest, the self-propelled harvester was passed to harvest the remaining rice plants in the experimental area.

Soil N 2 O and CH 4 emissions
Nitrous oxide and CH 4 were measured using the static chamber method (Mosier, 1989) between April 18, 2010, andMarch 12, 2012.The apparatus consisted of galvanized steel chambers (0.40 × 0.40 × 0.20 m, length × width × height) with a 32 L mean air volume per chamber attached to a metal base.The metal base was inserted into the soil to a depth of 0.12 m, covering two rows of rice plants, and removed only for sowing and harvest operations.Each base had an open bottom channel on the sides to allow water to flow freely, which was sealed before each sampling event.Additional extensors were stacked on the bases as the rice plants grew taller, and the chamber volume was considered in the calculations.Measurements included 64 air sampling events in 2010/11, with 44 and 20 events during the off-season and season, respectively; and 65 air sampling events in 2011/12, with 47 and 18 events during the off-season and season, respectively.The intervals between air samplings varied from 1 to 15 days, depending on the time frame required for straw management, soil tillage, sowing, harvesting and N applications.The flux chambers were simultaneously closed for all treatments, and air samples were manually collected between 9 and 12 a.m.Air samples were collected 0, 8, 16 and 24 min after the chambers were closed.Prior to collection, the air inside the chambers was homogenized for 30 sec using an electrical ventilator attached to the chamber wall, and the internal temperature was measured.The syringes were closed, placed in a refrigerated box and immediately sent to the laboratory for analysis.Air samples in the syringes were analyzed for N 2 O and CH 4 concentrations on a gas chromatograph (GC-2014, Shimadzu Corp., Kyoto, Japan) equipped with electron capture (63Ni ECD) and flame ionization (FID) detectors.The GC was equipped with three packed columns (HayeSep Q 80/100) set at 70 °C, N 2 as the carrier gas flowing at 26 mL min -1 , an injector with a 1 mL sample loop for direct injection at 250 °C, and ECD and FID detectors at 325 °C and 250 °C, respectively.The air samples were analyzed 24 h after their arrival at the laboratory.

Soil sampling and analysis
In the off-season, soil samples from the 0.00-0.10m layer were collected to determine soil moisture and soil inorganic N content.Soil moisture was determined by drying the samples at 105 °C for 48 h.Inorganic N in the soil was extracted by stirring 20 g of moist soil with 80 mL of KCl solution for 30 min.After decanting, the supernatant was collected and kept frozen until analysis.The contents of NH 4 + -N and NO 3 --N were determined by Kjeldahl distillation after sequential additions of MgO and Devarda's alloy, respectively, and titration with H 2 SO 4 0.0025 mol L -1 (Keeney and Nelson, 1983).The bulk soil density was determined after the rice harvest using the volumetric ring method (Blake and Hartge, 1986).
After rice harvest, soil samples from the layers o 0.00-0.05,0.05-0.10 and 0.10-0.20 m were collected at the end of the experiment in April 2012.In each plot, four subsamples were collected, which were air-dried, sieved through a 2 mm mesh and finely ground in a ball mill.The organic C content in the samples was determined by dry combustion in an elemental analyzer (FlashEA 1112, Thermo Electron Corporation, Milan, Italy).

Calculations
Carbon stocks in the 0.00-0.20 m soil layer were calculated using the soil equivalent mass methodology (Ellert and Bettany, 1995), using the soil density of the DH treatment Rev Bras Cienc Solo 2023;47nspe:e0220134 as a reference.The amounts of soil NH 4 + and NO 3 -(kg ha -1 ) in the 0.00-0.10m layer were calculated by multiplying the NH 4 + and NO 3 -content (mg kg -1 ) by the soil mass determined from bulk density in each treatment.Total porosity of the soil was estimated considering the bulk density and assuming a particle density of 2.65 kg dm -3 (Danielson and Sutherland, 1986).The water-filled pore space (WFPS) in the 0.00-0.10m soil layer was estimated by dividing the volumetric soil water content by the total porosity (Robertson and Groffman, 2015).Soil CH 4 and N 2 O fluxes were calculated using equation 1 as follows: in which: f is the gas production rate (g m -2 h -1 ); ∆Q/∆t is the variation in gas concentration (mol h -1 ); P is the atmospheric pressure within the chamber (1 atm); V is the chamber volume (L); R is the ideal gas constant (0.08205 atm L mol -1 K -1 ); T is the temperature within the chamber (K); M is the molar gas mass (g mol -1 ); and A is the basal chamber area (m -2 ).Cumulative gas emissions were obtained by integrating the daily fluxes between sampling events.
The GWP (Mg CO 2 eq ha -1 ) was calculated considering the annual emissions of N 2 O and CH 4 and net CO 2 emissions, according to equation.
in which: CO 2 denotes the net annual CO 2 fluxes; N 2 O and CH 4 represent the annual fluxes; and 265 and 28 represent their respective forcing indices used for the conversion of N 2 O and CH 4 to CO 2 (Pachauri et al., 2014), respectively; and CO 2 costs associated with agricultural operations (sowing, agrochemical and fertilizer applications, irrigation and harvest) and chemical inputs (agrochemicals and fertilizers) were obtained from Lal (2004).The net annual CO 2 fluxes were calculated using the variation in SOC stocks in soil and straw management systems in relation to the reference treatment (DH) as a proxy, according to equation 3.
Net annual CO 2 fluxes = (SOC system -SOC reference ) × 44/12 Eq. 3 in which: CO 2 denotes the annual net CO 2 emission in soil of management systems (SOC system ) relative to soil in the reference treatment (SOC reference ) multiplied by a conversion factor of the molecular mass of C to CO 2 (44/12).The reference treatment was DH.
Yield-scaled GWP was calculated by dividing the GWP of each treatment by the respective crop grain yield, according to equation 4.

Data analysis
The results of seasonal emissions of N 2 O and CH 4 in the growing season and off-season, the variations in soil C stocks, the GWP and the yield-scaled GWP were submitted to analysis of variance (ANOVA).Straw management and soil management factors were considered fixed effects, while block and growing season were considered random effects.When the factors were significant at the 5 % level, treatment means were compared using the Scott-Knott test at the 5 % level.

Soil water-filled pore space and inorganic nitrogen
Soil WFPS in the off-season ranged from 49 to 87 % in 2010 and from 43 to 87 % in 2011, with WFPS greater than 60 % in 78 % of samplings (Figures 2a and 2b).In 2011,

CH 4 emissions
Methane fluxes were low throughout the off-season in the two years evaluated.In this period, CH 4 fluxes ranged from -0.01 to 1.84 kg ha -1 day -1 in 2010 (Figures 3a and 3b) and from -0.02 to 1.15 kg ha -1 day -1 in 2011 (Figures 4a and 4b).Although negative fluxes were observed in some evaluations, all treatments represented a source of CH 4 to the atmosphere during the off-season period (Table 2).Cumulative CH 4 emissions in the off-season differed between treatments only in the first year (Table 2), with the highest and lowest emissions evidenced in the treatments with KR and without soil disturbance (NT and SR), respectively.In the KR and KR+DH treatments, the cumulative CH 4 emissions were significantly higher in the first year than in the second year (Table 2).
During irrigated rice cultivation, CH 4 fluxes were substantially higher than in the off-season.Methane fluxes gradually increased after flooding and reached emission peaks of 18.7 kg ha -1 day -1 in 2010/11 (Figures 3a and 3b) and 13.3 kg ha -1 day -1 in 2011/12 (Figures 4a and 4b after the suspension of water entry into the crop, the daily CH 4 fluxes of all treatments were close to zero. The cumulative emissions of CH 4 among the straw management strategies evaluated ranged from 334.5 to 586.0 kg CH 4 ha -1 in 2010/11 and from 165.8 to 346.5 kg CH 4 ha -1 in 2011/12.The treatment with straw incorporation with a disc harrow immediately following rice harvest showed the lowest CH 4 emissions in 2010/11 but did not differ from the other treatments in 2011/12.However, it did differ from those without straw management.In the two evaluated seasons, the cumulative CH 4 emissions in the NT and NT+R treatments without straw management were higher than in the treatments with some management type (KR, DH and SR).The CH 4 emissions drove the total annual emissions of each management during the growing season, which represented more than 95 % of the annual emissions.In all treatments, the accumulated emissions of CH 4 were higher in the 2010/11 season than in the 2011/12 season.

N 2 O emissions
The N 2 O fluxes ranged from -2.63 to 94.55 g ha -1 day -1 in 2010 (Figures 3c and 3d) and from -2.80 to 201.70 g ha -1 day -1 in the 2011 off-season (Figures 4c and 4d).Several N 2 O emission peaks were observed during soil wetting and drying cycles in both off-season periods.Soil N 2 O cumulative emissions differed between treatments only in the second year during the off-season.Off-season N 2 O emissions were higher in 2011 than in 2010, except for the DH and DH+DH treatments (Table 2).In the 2010 off-season, straw management after harvest did not affect soil N 2 O emissions.In the 2011 off-season, cumulative N 2 O emissions were reduced in three ways: straw incorporation after rice harvest using a disc harrow, straw removal, and ryegrass cultivation.Most N 2 O emitted to the atmosphere came from the off-season in the two years evaluated.
In 2010/11, the highest N 2 O fluxes were observed a few days before the soil was flooded.
In the second year (2011/12), soil N 2 O fluxes were close to zero throughout the period.
Negative cumulative N 2 O emissions observed during the harvest period indicated N 2 O consumption by the soil (Table 2).On average, across treatments, cumulative N 2 O emissions in 2011/12 were 175 % higher than those in 2010/11.

Soil carbon stocks
At the end of the two rice seasons, the soil C organic stocks (0.00-0.20 m) varied from 35.96 Mg ha -1 in the DH+DH treatment to 38.36 Mg ha -1 in the NT+R treatment (Figure 5).Soil C stocks in all treatments with the use of a disc harrow for straw management were lower than those in treatments without the use of a disc harrow.
Straw removal had a less adverse effect on soil C stocks than straw management operations involving disc harrows.Furthermore, straw removal did not affect C stocks compared to the NT treatment.Though not significantly different from the NT, KR and SR treatments, in absolute terms, the C stock tended to be higher in the NT+R treatment soil.
Compared to the disc harrow reference system (DH), the annual rates of change in soil organic C stocks ranged from -0.01 Mg ha -1 yr -1 in the DH+DH and KR+DH treatments to 1.19 Mg in the NT+R treatment.

Rice grain yield, GWP and yield-scaled GWP
Rice grain yield was not affected by soil and rice straw management systems or by year (Table 3).The average yield was 10.4 Mg ha -1 in both seasons.Straw management in autumn, right after rice harvest, reduced the GWP of the systems in relation to spring management, just before rice sowing.The highest GWP was observed in the treatment without straw management after harvest and in the treatments with two soil operations just before rice seeding.Methane emissions were the main component of the GWP in all straw management systems.The contribution of N 2 O emissions to GWP was small and mostly (>85 %) determined by emissions in the off-season.In straw management systems without the use of disc harrows, the increase in soil C stocks neutralized part of the CH 4 and N 2 O emissions, reducing GWP.The yield-scaled GWP ranged from 0.64 to 1.06 Mg CO 2 eq Mg -1 yield and was lower in soil and straw management systems shortly after rice harvest compared to management applied in spring (Table 3).

Effects of soil and straw management on soil carbon stocks
The soil C stocks in the 0.00-0.20 m layer observed in our study are in agreement with values reported in other locations in subtropical climates, which range from 35.9 to 41.4 Mg ha -1 (Nascimento et al. al., 2009;Huang et al., 2016;Yang et al., 2018).
After two years, SOC stocks were lower in treatments using disc harrows and higher in treatments without soil tillage.In addition, our results suggest that the implement used for straw management (disc harrow or knife roller) has a greater effect on soil C stocks than the tillage time.Pandey et al. (2014) showed that soil C losses are proportional to the intensity and frequency of soil tillage and that even the recalcitrant fraction of soil C cannot be considered inert to long-term biodegradation.Soil tillage significantly reduces the stability and proportion of macroaggregates, which are broken down into microaggregates or individual particles, thus indicating that tillage reduces the structural quality of soils (Jiang et al., 2011).Liu et al. (2021) suggest that paddy soils that have been continuously tilled for many years have depleted soil carbon stocks because tillage reduces the physical and chemical protection of C from microbial attacks through organo-mineral associations and increases soil C losses associated with erosion.
On the other hand, some studies have not observed differences in soil C stocks in different tillage systems in irrigated rice production systems (Nascimento et al., 2009;Huang et al., 2016).Nascimento et al. (2009) observed a higher stock of C in the surface layer of the soil under no-tillage, which suggests that physical protection has little effect on the stabilization of C in flooded soils and, therefore, the tillage systems do not differ in terms of C stocks.However, Tang et al. (2020) observed a higher mass of large aggregates and SOC content in superficial soil layers (0.00-0.05 m) under no-tillage than conventional tillage.Soil C stocks were higher in the treatment with straw removal than those with soil tillage using disc harrows (usual management).These findings indicate that the breakdown of aggregates affecting soil structure under frequent harrowing is more detrimental to soil C stocks than the removal of rice straw, but this conclusion has to be viewed with caution.Pampolino et al. (2008) suggest that residues from roots and algae that proliferate in aquatic environments may be sufficient to maintain C and N stocks in soils that remain flooded for most of the year.

Effects of straw management on N 2 O emissions
The off-season showed several peaks of N 2 O emissions, as observed in other studies (Zhang et al., 2016a;Yang et al., 2018).During the off-season, as a result of irrigation suppression, the topsoil is exposed to several cycles of wetting and drying, driven by rainfall and evapotranspiration.Some studies have shown that soil wetting and drying cycles increase substrate availability for nitrification and denitrification (Borken and Matzner, 2009;Guo et al., 2014).In addition, rewetting reduces soil oxygen availability, resulting in N 2 O and N 2 emissions; if flooding is prolonged, these gases remain trapped in the soil and are released as the soil dries (Congreves et al., 2018).(Yang et al., 2018) and that N uptake by a winter crop (Bayer et al., 2015) contributes to reducing available N in the soil and, consequently, N 2 O emissions in rice paddies.Removal of rice straw reduces the availability of organic C in the soil, which can influence denitrification under anaerobic conditions during the off-season.Wu et al. (2018) compared the effect of straw on N 2 O fluxes in two irrigation regimes and observed that straw removal reduced annual N 2 O emissions, especially in the system where the soil does not remain flooded throughout the period.On the other hand, other studies have found that straw removal can increase soil N 2 O emissions, as it serves as a substrate for microbial growth, possibly generating strong competition for NH 4 + between different groups of microorganisms (Sander et al., 2014;Zhang et al., 2017).
During the growing season, N 2 O fluxes were mostly low.Several studies report this same pattern of N 2 O emissions during the period of soil flooding (Sander et al., 2014;Wu et al., 2018;Zschornack et al., 2018).Soil flooding can reduce N 2 O emissions in two ways: first, it restricts nitrification, which, in addition to being one of the processes responsible for N 2 O emissions, also provides NO 3 -as a substrate for denitrification; second, the low availability of O 2 in the soil establishes conditions that favor the final reduction of N 2 O to N 2 (Chapuis-Lardy et al., 2007).In some samplings, N 2 O influxes were observed, suggesting that soil acted as an N 2 O sink during the growing season.Several other studies have also observed influxes of N 2 O during soil flooding (Bayer et al., 2014;Wu et al., 2018;Zschornack et al., 2018), which can be attributed to the intensely reduced soil redox potential.Cumulative N 2 O emissions throughout the years evaluated, though especially the second year, were similar to those in other studies carried out in the Philippines (Sander et al., 2014), China (Zou et al., 2005) and India (Tirol-Padre et al., 2016) but slightly higher than those in other studies (Ahmad et al., 2009;Zhang et al., 2016;Wu et al., 2018).

Effects of straw management on CH 4 emissions
The results of our study confirm our hypothesis that rice straw management shortly after harvest decreased soil CH 4 fluxes in the subsequent growing season.Soil CH 4 fluxes gradually increased after soil flooding in all treatments due to changes in soil redox potential (Zschornack et al., 2016;Bertora et al., 2018).Soil flooding induces an intense soil redox potential reduction, and once other final electron acceptors, such as Fe 3+ , Mn 4+ , SO 4 2and NO 3 -, have been consumed, the environment becomes favorable for CH 4 production (Kögel-Knabner et al., 2010); in addition, the presence of crop residues further reduces the soil redox potential and provides a substrate for methanogenesis (Bertora et al., 2018).As the flooding period and the development of rice plants advance, conditions become even more favorable for the production and emission of CH 4 .Several studies indicate maximum peaks of CH 4 emission observed in the reproductive stages of the crop cycle (Bhattacharyya et al., 2012;Bayer et al., 2015;Zhang et al., 2017).According to Le Mer and Roger (2001), in this period, CH 4 fluxes are related to the increase in root exudation and organic matter inputs due to root exfoliation and plant senescence.The maximum CH 4 emission rates during rice cultivation observed in our study ranged from 9 to 18.7 kg CH 4 ha -1 day -1 and are in agreement with the CH 4 emission rates reported in Rev Bras Cienc Solo 2023;47nspe:e0220134 other studies carried out in the same region (Bayer et al., 2015;Camargo et al., 2018;Zschornack et al., 2018).
Without management, the maintenance of rice and/or ryegrass crop residues on the soil surface did not reduce CH 4 emissions during the rice crop.These results disagree with other studies that report lower CH 4 emissions in no-tillage, which is attributed to the maintenance of plant residues on the soil surface compared to their incorporation into the soil by tillage (Ahmad et al., 2009;Bayer et al., 2014).In our study, the absence of tillage and the maintenance of straw on the soil surface possibly reduced straw decomposition during the winter.The daily CH 4 fluxes partially support this hypothesis during the season.In treatments without crop residue management, the first peak of CH 4 emissions was observed in the first air sampling after soil flooding, which can be attributed to the decomposition of easily mineralizable organic matter (Le Mer and Roger, 2001).This effect was even more evident when ryegrass was cultivated during winter, increasing the amount of substrate for methanogenesis at the time of soil flooding in spring/summer.
On the other hand, removing or incorporating straw in the soil soon after the rice harvest possibly reduced the amount of C available for anaerobic decomposition and, consequently, the CH 4 fluxes.In agreement with these results, Bertora et al. (2018) also demonstrated that both early management and straw removal are effective strategies to reduce CH 4 emissions in rice fields because they reduce substrate availability for methanogenesis.When soil is drained during the off-season, straw management soon after harvest allows straw decomposition to be carried out under aerobic conditions, reducing substrate availability for methanogenesis in the rice growing season (Bayer et al., 2015;Bertora et al., 2018).In this sense, Sander et al. (2014) recommend draining the area in the off-season as a strategy to reduce CH 4 emissions during the growing season.
The CH 4 emissions observed during the rice growing season corresponded to more than 95 % of the annual emissions.Cumulative emissions of CH 4 were similar to studies carried out in Asia (Ahmad et al., 2009;Sander et al., 2014) and slightly higher than those observed by Tirol-Padre et al. (2016).Although our results indicate a small increase in CH 4 fluxes during the off-season when crop residues were managed with a knife roller in the first year, this trend was not repeated in the second year, and CH 4 emissions in the off-season were close to zero.Several other studies have indicated that the rice growing season period is mainly responsible for CH 4 emissions (Zhang et al., 2016a;Yang et al., 2018;Zschornack et al., 2018).Thus, management strategies aimed at reducing CH 4 emissions should be focused on the rice growing season.

Effects of straw management on GWP and yield-scaled GWP
The annual GWP ranged from 6.0 to 10.4 Mg CO 2 eq ha -1 yr -1 between soil and straw management systems, which are close to the values observed in other studies carried out in subtropical climate regions (Bayer et al., 2014(Bayer et al., , 2015;;Zhang et al., 2016b;Zschornack et al., 2018).Methane emissions were the main determinant of GWP in all management systems but were partially neutralized by soil C sequestration in treatments where the soil was not tilled.In a meta-analysis study, Liu et al. (2014) showed that flooded soils are effective in increasing C stocks, but this increase can also lead to higher CH 4 emissions and GWP.Our results suggest that straw removal might be a strategy to reduce the GWP of systems in the short term, but it is necessary to evaluate the effects of straw removal on soil C stocks in the long term.Furthermore, as noted by Bertora et al. (2018), the costs involved in removing straw as well as the nutrient cycling promoted by straw, make this practice unattractive.According to the authors, similar results regarding GWP can be obtained by increasing the temporal distance between straw management and soil flooding.In fact, our results demonstrate that the anticipation of straw management is an effective strategy to reduce GWP.
According to Sosbai (2018), in 61 % of areas cultivated with rice in the state of Rio Grande do Sul, straw management is carried out immediately after harvest, while 30 % of the areas are managed with conventional tillage.Based on our GWP estimates and considering that 30 % of the cultivated area corresponds to 300,000 ha -1 , managing straw immediately after the rice harvest could generate a C-saving of 561 Gg CO 2 eq yr -1 .Managing the straw using a knife-roller right after rice harvest could have even greater benefits.The low GWP of this system is a result of the combination of low CH 4 emissions and increased soil C stocks.Thus, our results suggest that management practices with reduced soil disturbance that allow aerobic straw decomposition during the winter period and that promotes increases in soil C stocks are the most suitable to reduce the GWP of rice paddies.However, long-term studies should assess the possibility of soil C saturation and reduced ability to neutralize CH 4 emissions.Yield-scaled GWP was directly related to GWP since the grain yield did not differ between the straw management systems.Thus, the anticipation of straw management reduces GHG emissions per Mg of grain produced, that is, it makes the system more efficient.

CONCLUSIONS
Soil tillage with a disc harrow to incorporate rice straw, regardless of the time of year, reduces soil C stocks.However, different straw management strategies do not affect rice grain yield.The off-season period is responsible for most annual N 2 O emissions, while CH 4 emissions occur mostly during the rice growing season.Straw incorporation using a disc harrow, straw removal, or ryegrass cropping in the winter are strategies to reduce N 2 O emissions in the off-season.Straw management immediately after rice harvest efficiently reduces CH 4 emissions during the rice growing season and, consequently, reduces GWP and yield-scaled GWP.

Figure 1 .
Figure 1.Schematic representation of the soil and straw management systems evaluated regarding greenhouse gas balance for two years (rice season and off-season) in a subtropical ecosystem in southern Brazil.The treatments were applied to the respective plots in the two agricultural years.

April 1 -
No tillage (NT) 2 -Knife-roller post-harvest (KR PH) 3 -Straw removal (SR) 4 -Disc harrow post-harvest + disc harrow pre-sowing (DH PH + DH PS) 5 -No tillage + ryegrass (NT + RG) 6 -Disc harrow pre-sowing (DH PH) 7 -Disc harrow pre-sowing (DH PS) 8 -Knife-roller post-harvest + disc harrow pre-sowing (KR PH + DH PS) Bras Cienc Solo 2023;47nspe:e0220134 soil WFPS was evidently greater in the NT treatments than in the other treatments.2e and 2f) soil contents fluctuated during the off-season; no clear patterns emerged for mineral N levels.With few exceptions, the amount of NH 4 + -N and NO 3 --N in the soil was below 5 kg ha -1 in most assessments carried out in 2010 and 2011.

Figure 5 .
Figure 5. Carbon stocks in the 0.00-0.20 m soil layer in different straw management systems.Soil Carbon Stocks (Mg C ha -1 )

Table 1 .
Schedule of field operations performed over two years

Table 2 .
Cumulative N 2 O and CH 4 emissions in the off-season and rice-growing season Means followed by the same letter are not statistically different within each year, according to Scott-Knott test (α<0.05).*indicates that differences between years are statistically significant according to Scott-Knott test (α<0.05).NT: no tillage; NT + R: no tillage + ryegrass; KR PH: knife-roller post-harvest; DH PH: disc harrow post-harvest; SR: straw removal; DH PS: disc harrow pre-sowing; DH PH + DH PS: disc harrow post-harvest + disc harrow pre-sowing; KR PH + DH PS: knife-roller post-harvest + disc harrow pre-sowing.Rev Bras Cienc Solo 2023;47nspe:e0220134

Table 3 .
Zschornack et al. (2018)4 emissions, variation in C stocks (Δ soil C), operating costs, estimated global warming potential (GWP) and yield-scaled GWP Average of two growing seasons.Means followed by the same letter are not statistically different within each year according to Scott-Knott test (α<0.05).*indicates that differences between years are statistically significant according to Scott-Knott test (α<0.05).NT: no tillage; NT + R: no tillage + ryegrass; KR PH: knife-roller post-harvest; DH PH: disc harrow post-harvest; SR: straw removal; DH PS: disc harrow pre-sowing; DH PH + DH PS: disc harrow post-harvest + disc harrow pre-sowing; KR PH + DH PS: knife-roller post-harvest + disc harrow pre-sowing.Rev Bras Cienc Solo 2023;47nspe:e0220134The peaks of N 2 O emissions observed during the off-season represented more than 85 % of annual emissions.Zschornack et al. (2018)reinforce the need to include assessments of N 2 O emissions during the rice crop off-season, a period that accounted for up to 90 % of annual N 2 O emissions.According to the authors, strategies to mitigate N 2 O emissions should be focused on in this period.Our results indicate that incorporating straw with disc harrows, cultivating ryegrass in winter, and removing straw after the rice harvest are potential strategies to reduce N 2 O emissions in the off-season because they possibly restrict substrate availability for nitrification and denitrification.Some studies have suggested that N immobilization caused by rice straw incorporation with a high C/N ratio