Differences in net global warming potential and greenhouse gas intensity between major rice-based cropping systems in China

Double rice (DR) and upland crop-single rice (UR) systems are the major rice-based cropping systems in China, yet differences in net global warming potential (NGWP) and greenhouse gas intensity (GHGI) between the two systems are poorly documented. Accordingly, a 3-year field experiment was conducted to simultaneously measure methane (CH4) and nitrous oxide (N2O) emissions and changes in soil organic carbon (SOC) in oil rape-rice-rice and wheat-rice (representing DR and UR, respectively) systems with straw incorporation (0, 3 and 6 t/ha) during the rice-growing seasons. Compared with the UR system, the annual CH4, N2O, grain yield and NGWP were significantly increased in the DR system, though little effect on SOC sequestration or GHGI was observed without straw incorporation. Straw incorporation increased CH4 emission and SOC sequestration but had no significant effect on N2O emission in both systems. Averaged over the three study years, straw incorporation had no significant effect on NGWP and GHGI in the UR system, whereas these parameters were greatly increased in the DR system, i.e., by 108% (3 t/ha) and 180% (6 t/ha) for NGWP and 103% (3 t/ha) and 168% (6 t/ha) for GHGI.

and individual field studies have also reported variable CH 4 emissions among cropping systems 12,13 . However, no field study to date has simultaneously addressed CH 4 and N 2 O emissions from different rice cropping systems. Moreover, to our knowledge, the differences in NGWP and GHGI between different rice cropping systems have not been documented.
Straw return has been widely recommended for agricultural ecosystems in China. Chinese agriculture produces approximately 620 Tg of crop straw every year, with an increasing trend of an annual rate of 1.4% 17 , and approximately 25% of the straw is currently returned to the field 18 . Indeed, straw incorporation is a common practice in rice production, as it helps to maintain soil quality and recycle mineral nutrients 19 . Straw incorporation also has a considerable influence on CH 4 and N 2 O via changes in soil properties, such as the porosity, temperature and moisture 20,21 . In general, straw incorporation can enhance carbon sequestration, resulting in improved soil productivity and air quality, and thus offset GHG emissions from rice fields. However, a significant stimulation of CH 4 emission due to straw incorporation in rice fields has been well documented 22,23,24 . In contrast, straw incorporation inhibits 7,12 or has no significant effect 25 on N 2 O emission from rice fields. Nonetheless, the mechanism by which straw addition affects carbon sequestration as well as CH 4 and N 2 O emissions and GHGI in different rice cropping systems remains unknown.
Based on previous studies, we hypothesize that (1) different rice cropping systems may differ greatly in CH 4 and N 2 O emissions due to drastic flooding periods and (2) straw incorporation may result in different influences on CH 4 and N 2 O emissions from different rice cropping systems. To test these hypotheses, a field experiment was established to measure CH 4 and N 2 O emissions and SOC changes between the two major rice cropping systems in China. The objectives were to gain insight into the differences in grain yield, NGWP and GHGI between UR and DR systems as affected by straw application.

CH 4 emission.
Analysis of variance (ANOVA) indicates that annual CH 4 emission depended strongly on the cropping system, straw incorporation, and their interactions (Table 1). Inter-annual variation was also observed (Table 1). In the UR system, similar seasonal patterns of net CH 4 flux were observed for all treatments. The net CH 4 flux was significant during the rice-growing season but was negligible during the wheat-growing season (Fig. 1, Table 2), ranging from − 1.45 to 36.2 mg C m −2 h −1 during the three annual rotations. In addition, the net CH 4 flux increased after rice transplantation and decreased dramatically during the midseason drainage; after reflooding, the flux increased again to a low emission peak and then decreased gradually to a negligible amount until harvest. In the DR system, all plots served as minor sinks or sources for atmospheric CH 4 during the oil rape-growing season (Table 2), and all plots served as net sources of atmospheric CH 4 during the early and late rice-growing seasons ( Table 2). The patterns of CH 4 flux observed during the early and late rice-growing seasons were similar to those of the rice-growing season in the UR system. Compared with the UR-S0 plot (104 kg CH 4 ha −1 year −1 ), the annual CH 4 emission increased significantly by 76.9% in the DR-S0 treatment (185 kg CH 4 ha −1 year −1 ).
Straw incorporation significantly stimulated CH 4 emission. The highest CH 4 fluxes, i.e., 24.8, 34.5 and 36.2 mg C m −2 h −1 , were observed in 2009, 2010 and 2011, respectively, in the UR-S2 plot. The annual CH 4 emissions averaged over the three years were 104, 208 and 303 kg CH 4 ha −1 year −1 for the UR-S0, UR-S1 and UR-S2 plots, respectively ( Table 3). The average annual CH 4 emissions were 99.7% and 191% higher in the UR-S1 and UR-S2 plots, respectively, than in the UR-S0 plot. In comparison to UR-S1, the annual CH 4 emission was significantly increased in the UR-S2 treatment (by 45.5%). Similar to the UR year −1 for the DR-S0, DR-S1 and DR-S2 plots, respectively. Compared with the DR-S0 plots, annual CH 4 emissions were significantly increased by 150% and 280% in DR-S1 and DR-S2, respectively. Furthermore, straw incorporation enhanced the differences between the DR and UR systems ( Table 3). The increase due to straw incorporation in the DR system was 150% and 280% for DR-S1 and DR-S2, respectively, obviously higher than the 99.7% and 191% observed for UR-S1 and UR-S2, respectively in the UR system. N 2 O emission. The majority of N 2 O emission occurred during the wheat-growing season in the UR system and the oil rape-growing season in the DR system (Fig. 2), and the N 2 O fluxes were primarily driven by fertilizer application and precipitation. No peaks of N 2 O flux were observed during the 2009 wheat-and oil rape-growing seasons because almost no precipitation occurred after the application of the basic fertilizer. However, during the rice-growing season, N 2 O flux peaks were observed in response to both N fertilizer application and midseason aeration. Straw incorporation tended to decrease N 2 O emission during the rice-growing season in both systems; however, the effects were not statistically significant ( Table 2).
The annual N 2 O cumulative emissions averaged over the three years were 2.26, 2.08 and 2.19 kg N ha −1 year −1 for the UR-S0, UR-S1 and UR-S2 plots and 2.77, 3.13 and 3.04 kg N ha −1 year −1 for the DR-S0, DR-S1 and DR-S2 plots, respectively (Table 3). It is apparent that the annual N 2 O emissions were significantly increased in the DR system relative to the UR system ( Table 3). Analysis of variance (ANOVA) indicated that the annual N 2 O emission was strongly dependent on the cropping system but not influenced by straw incorporation. Although inter-annual variation was observed, no interactions were found (Table 1).

SOC sequestration.
The soil organic C content was 14.6 g kg −1 upon establishment of the field experiment in November 2008, and SOC in the topsoil (0-20 cm) increased in all treatments over the three years of the study. After the three cycles of field experiment, the SOC contents reached 14.7 g kg −1 , 15.7 g kg −1 , 16.7 g kg −1 , 15.2 g kg −1 , 16.3 g kg −1 and 17.4 g kg −1 in UR-S0, UR-S1, UR-S2, DR-S0, DR-S1 and DR-S2, respectively. From November 2008 to November 2011, the rate of SOC increase ranged from 0.03 g C kg −1 yr −1 for the UR-S0 plot to 0.95 g C kg −1 yr −1 for the DR-S2 plot. The topsoil SOC density was estimated based on the topsoil SOC content and bulk density, with the SOC sequestration rate (SOCSR) ranging from 0.08 t C yr −1 for UR-S0 to 2.42 t C yr −1 for DR-S2 (Table 3). Compared with the UR-S0 plot, SOCSR tended to increase in the DR-S0 plot, but this effect was not statistically significant. Straw incorporation enhanced SOCSR in both the UR and DR systems, but the differences among UR-S1 and DR-S1, and UR-S2 and DR-S2 were not statistically significant (Table 3).  Yield, NGWP and GHGI. Over the three years, the annual yields were strongly dependent on the cropping system and year as well as their interaction ( Table 1). The annual yield was significantly increased in the DR-S0 plot compared with the UR-S0 plot (Table 3). However, straw incorporation had no significant effect on the seasonal grain yield, except that the late rice yield from DR-S2 was increased by 14.7% compared with DR-S0 (Table 2); straw incorporation tended to increase the annual grain yield of both systems but not to a statistically significant extent (Table 3). Relative to the UR system, the annual grain yield of the DR system was significantly increased when straw was incorporated (Table 3). NGWP was significantly influenced by the cropping system, straw incorporation, and the year, varying significantly as a result of interactions between the cropping system and straw incorporation as well as straw incorporation and the year. GHGI also strongly depended on the cropping system, straw incorporation, and the year as well as the interaction between the cropping system and straw incorporation ( Table 1). Relative to the UR system, the annual NGWP of the DR system increased markedly while no significant difference between the UR-S0 and DR-S0 treatments was observed for GHGI (Table 3). Although straw incorporation had no significant effect on the annual NGWP and GHGI of the UR system, these parameters significantly increased in the DR system, and compared with a moderate rate, the incorporation of straw at a high rate resulted in further annual NGWP and GHGI increases in the DR system (Table 3).

Discussion
In the present study, the annual CH 4 emission from the DR system was significantly higher than that from the UR system (Table 3). Due to the long period of flooding, double rice cropping systems emit more CH 4 than single rice cultivation systems 26 . In fact, water management has been recognized as one of the most important practices affecting CH 4 emission from paddy fields 27 . When plots are flooded, oxygen cannot diffuse into the soil, and strong anaerobic conditions may develop, thus favouring the growth of methanogen 28 .
In our study, annual N 2 O emissions were significantly increased in the DR system, which is not in agreement with the results of previous pot experiments 16 . However, differences between field and pot experiments may produce different results with regard to N 2 O emissions. Nonetheless, the inter-annual variation in N 2 O emissions was significant (Table 1), and the cumulative N 2 O emissions were considerably lower during the non-rice season in both the UR and DR systems in 2009 compared with 2010 and 2011 (Table 2). When soil is not maintained under flooded conditions, the water-filled pore space (WFPS) and the available N content are the two most important factors affecting N 2 O emissions 7,29,30 . In addition, high N 2 O emissions during the non-rice season generally occurred after the application of basal fertilizer and precipitation events in 2010 and 2011, consistent with a previous study 21 . Indeed, precipitation events can create suitable soil moisture conditions for N 2 O production via microbial processes 31 , yet almost no N 2 O flux peaks were observed during the non rice-growing season in both the UR and DR systems in 2009 because no precipitation occurred after the basal fertilizer application. Similar results were observed by Ma et al. 30 in the same region. Because NGWP is dominated by CH 4 emissions in both systems, high CH 4 emissions resulted in a significantly higher NGWP for the DR system compared with the UR system (Table 3). NGWP was also significantly higher in DR-S0 than in UR-S0, which was accompanied by an annual grain yield that was dramatically higher in DR-S0 compared with UR-S0. Consequently, there was no significant difference between the UR-S0 and DR-S0 treatments with respect to GHGI (Table 3). These two major rice cropping systems are equally appropriate for sustainable rice production on the basis of per unit of yield.
Straw incorporation significantly increases CH 4 emissions because of the additional C that is available for methanogenesis during the rice-growing season, which has been widely demonstrated in previous studies 12,14,23 . CH 4 emissions were highest in the S2 treatment, followed by S1 and S0, during the rice-growing seasons in both the UR and DR systems ( Table 2). CH 4 is typically produced under strictly anaerobic conditions with a low soil redox potential 27 , and the net CH 4 flux is the balance between methanogenic and methanotrophic processes 32 . Thus, organic amendment and the water regime during the rice-growing season are the top two variables controlling CH 4 flux 33 .
Straw incorporation tended to decrease N 2 O emissions during the rice-growing season in both the UR and DR systems (Table 2), a finding also supported by previous studies 7,34 . N 2 O is naturally produced in soil through nitrification and denitrification 2 , which are generally regulated by the availability of organic C and the availability and forms of N in the soil under anaerobic or aerobic conditions 35,36 . The observed decreases in N 2 O during the rice-growing season in the presence of straw incorporation may be explained by the following: the decomposition of crop residues with a high C:N ratio (> 40) can enhance microbial N immobilization, resulting in less available N for nitrification and denitrification and consequently decreased N 2 O emissions 12,22 . Furthermore, our previous study proved that straw incorporation during the rice-growing season can decrease the soil redox potential (Eh) and increase the concentration of Fe 2+ , thus facilitating the further reduction of N 2 O to N 2 and resulting in decreased N 2 O emissions 7 .
Although straw incorporation had no effect on NGWP and GHGI in the UR system, significant increases were observed in the DR system (Table 3). This finding is primarily related to the high amount of CH 4 emissions induced by straw incorporation. In addition, straw incorporation enhanced the difference in NGWP between the UR and DR systems (Table 3) because the incorporation occurred in both the early and late rice seasons in the latter system. As the annual NGWP driven by straw incorporation outweighed the benefits of grain yield and SOCSR increases, the annual GHGI was significantly increased in DR-S1 and DR-S2 compared with UR-S1 and UR-S2, respectively. Therefore, considering the annual NGWP and GHGI, direct straw incorporation during the rice-growing season in China is beneficial for the UR system but is not a good strategy for the DR system.  Table S1. The region is characterized by a typical subtropical climate with an annual average air temperature of 15.7°C and precipitation of 1050 mm. The daily mean air temperatures and precipitation during the experiment were collected from a nearby weather station, as shown in Fig. S1. The soil of the experimental field has a bulk density of 1.28 g cm −3 , pH 5.7, organic C content of 14.6 g kg −1 , and total N content of 1.32 g kg −1 .

Experimental site.
Experimental treatments and field management. Two crop rotation systems were included in this experiment, i.e., a wheat-rice rotation, which represents the UR system, and an oil rape-early rice-late rice rotation, which represents the DR system. Straw (0, 3 and 6 t/ha) was incorporated during the rice season in the UR system (UR-S0, UR-S1, UR-S2) and during both the early and late rice seasons in the DR system (DR-S0, DR-S1, DR-S2) before rice transplantation (Table S2). Nitrogen fertilizer (urea) was applied at a rate of 200 kg N ha −1 for the early and late rice seasons and 250 kg N ha −1 for the other crop seasons. A total of six field experimental treatments with three replicated plots (4 m × 5 m) were established using a randomized block design. The nitrogen fertilizer (urea) was split broadcast at a ratio of 4:3:3 as a basal fertilizer and two topdressings. Phosphate and potassium fertilizers were applied uniformly as a basal fertilizer to the different treatments at 60 kg P 2 O 5 ha −1 and 120 kg K 2 O ha −1 . Field management followed the local agronomic practices, including cultivation, irrigation, fertilizer application and pest and weed control.
All field plots were drained in the winter season. Consistent with the water management of local winter crop-rice systems, flooding was initiated 2-3 days before rice transplantation and was maintained for 30-40 days until midseason drainage for one week. A final drainage event occurred approximately 15 days before rice harvesting in all treatments.

Measurements of CH 4 and N 2 O fluxes. CH 4 and N 2 O emissions were measured from November
2008 to November 2011 using static opaque chambers and gas chromatography. One chamber was placed within each treatment replicate to achieve three replicate gas flux measurements for each observation time. The chamber, which was 0.5 or 1.1 m tall, having been adapted for the crop growth and plant Scientific RepoRts | 5:17774 | DOI: 10.1038/srep17774 height, covered a field area of 0.2025 m 2 (45 × 45 cm) and was placed on a fixed PVC frame in each plot. To minimize air temperature changes inside the chamber during sampling, the chamber was wrapped with a layer of sponge and aluminium foil. For each flux measurement, four gas samples were collected from 9:00 to 11:00 am using a 25-mL syringe at 0, 10, 20, and 30 min after the chambers were placed on the fixed frames. Over the three annual cycles, CH 4 and N 2 O fluxes were generally measured once a week in triplicate plots for all treatments, but samples were collected more frequently after a precipitation event, fertilizer application and rice transplantation.
The flux (F) of CH 4 and N 2 O was calculated using the following equation: where F is the flux of greenhouse gas (mg•m -2 •h -1 ), ρ is the density of CH 4 (0.536 g•L -1 ) or N 2 O (1.25 g•L -1 ), V is the volume of the static opaque chamber (m 3 ), A is the cover areas of the fixed PVC frame (m 2 ), dc/dt is the rate at which the concentration of CH 4 or N 2 O changes with time, and T is the temperature inside the static opaque chamber (°C). The gas samples were analysed for CH 4 and N 2 O concentrations using a gas chromatograph (Agilent 7890A, Shanghai, China) equipped with an electron capture detector (ECD) for N 2 O analysis and a hydrogen flame ionization detector (FID) for CH 4 analysis (CO 2 was first reduced by hydrogen to CH 4 in a nickel catalytic converter at 375°C). N 2 O was separated using two stainless steel columns packed with 80-100 mesh Porapak Q. One column was 2 m long with an inner diameter of 2 mm; the other column was 3 m long with an inner diameter of 2 mm. The carrier gas was argon-methane (5%) at a flow rate of 40 ml min −1 . The temperatures of the columns and the ECD detector were maintained at 40°C and 300°C, respectively, and the oven and FID were operated at 50°C and 300°C, respectively. The detection limits for CH 4 and N 2 O in this study are 0.023 mg C m -2 h -1 and 1.72 μ g N m -2 h -1 , respectively.

Measurements of changes in SOC.
Soil samples were collected when the field experiment was initiated in November 2008 and after three years in November 2011. A composite sample for each plot was obtained by randomly collecting five or six soil cores at a depth of 20 cm (3 cm diameter) and mixing them thoroughly. Any visible roots, stones, or organic residues were removed manually after air-drying the samples at room temperature. The samples were then ground to pass through a 2-mm sieve, and a portion was subsequently ground in a porcelain mortar to pass through a 0.15-mm sieve for SOC measurement. The total SOC was analysed following wet digestion with H 2 SO 4 -K 2 Cr 2 O 7 . The minimum change in SOC that can be detected by this method is 0.01 g kg −1 .
The soil organic carbon sequestration rate (SOCSR) was calculated as follows: where the numbers 34 and 298 represent the IPCC factors for the conversion of CH 4 and N 2 O to CO 2 equivalents, respectively 1 .

Statistical analysis.
Statistical analyses were performed using JMP 9.0 (SAS Institute Inc., Cary, USA). Three-way factorial analysis of variance (ANOVA) was used to test the effect of cropping system, straw incorporation and year on annual CH 4 and N 2 O emissions, yields, NGWP and GHGI.