Introduction

Modern changes in global climate, including greater atmospheric concentrations of greenhouse gases (CO2, CH4, N2O, etc.), temperature shifts, higher frequency of extreme events, such as drought or heavy rainfall, and heat waves, are directly or indirectly associated with the combustion of fossil fuels, land use change and other human activities that have occurred since the Industrial Revolution1. Such changes have greatly affected terrestrial ecosystems. For example, extreme events could lead to negative and even disastrous effects on plants2.

Rice (Oryza sativa L.) is the most widely consumed staple food crop in the world, and is particularly important in Asia. The food safety and national security of some countries depends on its production. Rice cultivation is also a significant source of greenhouse gas emissions, primarily methane (CH4) and nitrous oxide (N2O). Annual CH4 emissions from rice fields have been estimated to account for about 5–19% of global CH4 emissions, and agricultural N2O emissions increased by nearly 17% between 1990 and 2005, and now account for 60% of global anthropogenic N2O emissions3. The flooded environment, created during rice cultivation, provides anaerobic conditions that favor CH4 production by methanogens. The resulting CH4 can be oxidized by methanotrophs under aerobic conditions (e.g., in the rhizosphere and at the soil-water interface) and is finally emitted to the atmosphere through soil- or water-atmosphere interfaces and by the rice plant aerenchyma4. Nitrogen fertilization and water management (e.g. alternating wetting and drying) facilitates N2O emission via the processes of nitrification and/or denitrification in rice paddies4, hence, when applied appropriately, fertilizer and management interventions can play important roles in effectively controlling CH4 and N2O emissions during rice cultivation5,6,7,8,9. However, rice varieties differ significantly in terms of total CH4 and N2O emissions10,11,12,13. Selecting a rice variety that results in low CH4 and N2O emissions may therefore be a promising way to mitigate greenhouse gas emissions from rice paddies.

Drought is a serious limiting factor on crop production and the most damaging stressor in modern agriculture14,15. Rice consumes 70–90% of the total amount of irrigation water used in agriculture16,17, and as roughly a half of the world’s 158 million ha of rice land is paddy rice18, the production of this crop is very susceptible to water stress19. A series of agricultural practices, such as the use of saturated soil cultures and alternate wetting and drying, are recommended to reduce water input from irrigation and to enhance water use efficiency in rice cultivation. These practices could greatly reduce the water input required. However, they also lower rice grain yields to a certain extent18. Breeding new varieties of rice with drought tolerance may be an effective way of sustainably addressing the water scarcity issue20. These new varieties are called Water-saving and Drought-resistance Rice (WDR), and are characterized by having a similar yield potential and grain quality as wild type varieties, but require less water (50% less water use). In water-limited environments, they show higher drought resistance and minimize yield loss20. However, to our knowledge, little is known about the CH4 and N2O emissions from rice paddies by WDR under different climate conditions.

The annual mean surface air temperature for China over the past 97 years has experienced a warming of 0.79 °C, with a warming rate of 0.08 °C/10a, which is slightly larger than the global or northern hemispheric average as given by IPCC Third Assessment Report21. In addition, in the experimental region, the mean air temperature during the rice growing season in the 1990 s was 0.2 °C higher than that in the 1960–80s. However, the mean seasonal air temperature between 2001 and 2008 increased by 1.1 °C relative to that in the 1960–80s22. Climate models project that the increase in global surface air temperatures may exceed 1.5 °C by the end of the 21st century relative to the average between 1850 and 190054. Elevated air temperature can result in rice grain yield loss, mainly through reduced photosynthesis caused by chloroplast damage, spikelet sterility caused by decreased pollen production, and increased energy consumption caused by higher respiration demand25. Moreover, there was a significant interaction between rice variety and year (P < 0.001, Table 3).

The average CO2-eq emissions in the NA100% plots was 378 kg CO2-eq t−1 with a range of 170 to 736 kg CO2-eq t−1 (Table 3), which was within the 33–557 kg CO2-eq t−1 range reported in previous studies conducted in the same region9,32,33,34,35. In most treatments, the detected CO2-eq emissions were higher in the ‘warm and dry’ season than in the normal season. However, the difference between the two seasons was not significant (Table 3). Higher CH4 emissions and lower rice grain yields from the treatments led to greater CO2-eq emissions in the ‘warm and dry’ season relative to the normal season. Interestingly, there was a significant difference in CO2-eq emissions between Huayou14 and Hanyou8 (P < 0.05, Table 3). This suggests that the CO2-eq emissions vary considerably with rice growing season or rice variety. Furthermore, CO2-eq emissions from the NA70% and NA30% plots of both rice varieties (excluding H8-NA30% in the 2014 normal season) were potentially depressed when compared to the NA100% plot (Table 3). The decreases in CH4 emissions from the NA70% and NA30% plots were the main cause of the effective depression in CO2-eq emissions, especially during the ‘warm and dry’ season. However, in the normal season, due to the higher N2O emissions, reductions in the amount of irrigation had little effect on CO2-eq emissions.

In conclusion, CH4 and N2O emissions strongly differed according to rice variety and irrigation management between the two rice growing seasons under contrasting climate conditions. The amount of irrigation water was significantly higher in the ‘warm and dry’ season than that in the normal season. Although the same amount of irrigation water was applied to the two rice varieties, the SSWDs in the plots planted with Hanyou8 were higher than Huayou14, due to lower water demand from Hanyou8. The CH4 emissions by Huayou14 and Hanyou8 increased 93% and 161% in the ‘warm and dry’ season (2013), respectively, compared with that in normal season (2014). Moreover, the CH4 emissions from Hanyou8 were higher than from Huayou14 in both seasons. Reducing the amount of irrigation water can effectively reduce the CH4 emissions, regardless of the rice variety and climate conditions. However, less irrigation during the ‘warm and dry’ season greatly decreased the Huayou14 grain yield, but had little impact on Hanyou8 yield. In contrast, compared to the effect of rice variety, N2O emission depended more on fertilization and surface standing water depth when the fertilizer was applied. Under the global warming scenarios, feasible reductions in the amount of irrigation water applied and the suitable selection of rice varieties would be a promising way to mitigate greenhouse gas emissions as well as maintain rice grain yield.

Materials and Methods

Study site and experimental design

The study was conducted in an experimental field at the Shanghai Engineering Research Center of Low-carbon Agriculture, which is a part of the Zhuanghang Experimental Station (30°53′N, 121°23′E), and is located in the Yangtze River delta zone in east China. A rice-wheat crop** rotation system is the typical practice in this area. The paddy field soil was plowed to a depth of ~15 cm, and its chemical and physical properties were as follows: soil organic C (SOC) 13.7 g·kg−1, total N 1.4 g·kg−1, bulk density 1.4 g·cm−3, and pH (H2O) 7.6.

Each experimental plot was 60 m2 and an impermeable membrane was buried vertically in the soil around each plot at 1.1 m depth to prevent lateral seepage between the experimental plots. Then a concrete wall (30 cm width × 60 cm height) was built around the experimental plots. It was half buried into the soil over the impermeable membrane. A centrifugal pump (SW100-160, 100 m3·h−1, Shanghai Sanxing Supply and Drainage Equipment Co., Ltd., Shanghai, China) and polyethylene pipes were used to transport river water to each plot for irrigation. A meteorological station was established nearby in 2012, which provided data about the air/soil temperature, dry/wet precipitation, evaporation, solar radiation, wind speed/direction etc.

Two rice varieties (Oryza sativa L. Huayou14 and Hanyou8) and three types of irrigation management were employed in this study. Each treatment was replicated three times, resulting in a total of 18 plots in this experiment (i.e., 2 rice varieties × 3 irrigation levels × 3 replicates). Huayou14 is a high-yielding hybrid that is often cultivated by local farmers and Hanyou8 was recently developed by the Shanghai Agrobiological Gene Center for its water-saving and drought-resistant traits. The three types of irrigation management applied were normal amount of traditional irrigation management (NA100%), 70% of normal (NA70%), and 30% of normal (NA30%). The performance of NA100% was consistent with conventional irrigation practice for meeting rice growth demand. The other two types were applied proportionately during every irrigation event when little or no surface-standing water was observed in the NA100% plot for Huayou14. The irrigation was mainly carried out between 30 and 100 days after transplanting.

Climate conditions and agricultural practices

Seasonal changes in daily mean air temperature, daily precipitation, and soil temperature at 5 cm depth in the 2013 and 2014 rice growing seasons are shown in Fig. 1. The mean seasonal air temperature was 26.7 °C and ranged from 14.8 to 33.5 °C during the 2013 rice growing season (Table 1 and Fig. 1a), whereas it was 24.4 °C (ranging from 14.7 to 30.2 °C) in 2014 (Table 1 and Fig. 1b). The mean seasonal air temperature (24.4 °C) in 2014 was similar to the normal value (24.7 °C) reported by Su et al. in this region55. The mean seasonal air temperature in 2013 was 2.3 °C higher than that in 2014. Seasonal variations in soil temperature were similar to the daily mean air temperature (Fig. 1). Total precipitation in the 2013 and 2014 seasons was 492.1 and 762.7 mm, respectively (Table 1). In this region, annual precipitation is about 1200 mm, and about 60% of precipitation occurs between May and September56. In the 2013 season, precipitation between June and September (i.e., a duration from transplanting to the grain-filling stage) was only 271.8 mm, and this season was considered as a ‘warm and dry’ season. In contrast, in 2014, the total precipitation between June and September was similar to the average for the area, and this season was regarded as normal season.

Rice plants were transplanted at a density of 20 hills per m2 on June 14/16 and harvested on October 21/22 in 2013 and 2014, respectively. The N fertilizer application rate was 225 kg·ha−1, which was applied at a ratio of 5:3:2 (w/w/w) as base, tillering and heading applications, respectively. The base fertilizer was applied in the form of a compound fertilizer at 1–2 days before transplanting. The tillering and heading fertilizers were applied in the form of urea at about 1 week and 7 weeks after transplanting, respectively. Phosphorous (P2O5) fertilizer was applied as a base, compound fertilizer at a rate of 112.5 kg·ha−1, and 44% potassium (K2O) fertilizer was applied as a base, compound fertilizer at a rate of 255 kg·ha−1. The remaining of potassium (K2O) fertilizer was applied as a heading fertilizer in the form of commercial potassium chloride (KCl).

There were 12 and three irrigation events in the 2013 and 2014 seasons, respectively. Mid-season drainage (MD) is a conventional agricultural practice during the rice growing season. The MD began on July 19/21 and finished on July 29/August 4 in 2013 and 2014, respectively.

Measurements

The samples used to determine CH4 and N2O concentrations were taken using a static transparent chamber consisting of a plexiglass base frame (50 cm length × 40 cm width × 20 cm height) and a plexiglass lid (50 cm × 40 cm × 50 cm) equipped with a battery-driven 12 V fan at the center of its inner top. Other plexiglass frames were used to extend the lid height, by 20, 40, or 60 cm depending on the height of rice plants. The base frames were inserted approximately 15 cm into the soil, and four hills of rice plants were transplanted. One base frame was placed in each plot. Four gas samples were collected from each chamber at 6-min intervals using an auto gas sampler attached to four aluminum foil gas bags (1 L, Dalian Delin Gas Packing Co., Ltd., Dalian, China) at each sampling time. The auto gas sampler was composed of a 12 V rechargeable battery (NP7-12, YUASA Battery (Guangdong) Co., Ltd, Guangzhou, China), a gas pump (FAY4002, 2 L min−1, Chengdu Qihai E&M Manufacturing Co., Ltd., Chengdu, China), a box containing a circuit board (Nan**g Weina Electronic Co., Ltd., Nan**g, China), and a series of compact direct-operated 2-port solenoid valves (VDW23-6 G-1, SMC Pneumatics Ltd., Tokyo, Japan).

The gas samples from all the plots were collected between 08:00 and 10:00 and immediately taken to the laboratory. The concentrations of CH4 and N2O were determined by a gas chromatograph (7820 A, Agilent Technologies, Inc., Santa Clara, CA, USA) equipped with a flame ionization detector and an electron capture detector, respectively. The sampling frequency was once a week. However, whenever there was a fertilizer application, an MD, or irrigation after MD, a higher sampling frequency (once every 2 days) was used and daily sampling lasted for one week. CH4 and N2O flux were calculated by examining the linear increases of CH4 and N2O concentrations in the headspace of the chambers over time. The seasonal total CH4 and N2O emissions from all plots were calculated directly from the fluxes.

The surface-standing water depth (SSWD) was measured directly using a ruler after each gas-sampling event. Soil Eh (Oxidation-Reduction Potential) was detected at 5 cm soil depth with a pH/NO3/Eh meter (PRN-41, DKK-TOA Co., Tokyo, Japan). Soil temperature was monitored at 5 cm soil depth using a moisture meter (HH2, Delta-T Devices Ltd, Cambridge, UK) during each gas-sampling event around the base frames. At the end of each rice growing season, the plants in each plot were manually harvested. The dry weight of the rice grains was determined using an oven at 75 °C. Finally, the rice grain yield of each plot was calculated on a rice grain dry weight basis using the equation:

where RY is the rice grain yield (t·ha−1), DW is the dry weight of rice grains (t·ha−1), and 14.5% was used as the standard moisture content for storage of the rice varieties used in this study.

The equivalent CO2 (CO2-eq) emission for total CH4 and N2O emissions (greenhouse gas intensity) was calculated on a rice grain yield basis using the equation:

where TCO2-eq is the total amount of equivalent CO2 emission (kg CO2-eq·t−1), TCH4 is the total amount of CH4 emission (kg·ha−1), TN2O is the total amount of N2O emission (kg·ha−1), 25 and 298 are the multiples of GWP (global warming potential) for CH4 and N2O versus CO2 over 100 years3, and RY is the rice grain yield (t·ha−1).

Statistical analysis

The impacts of the three parameters (irrigation management, rice variety, and year) on CH4 and N2O emissions from rice paddies were examined. Their effects were analyzed using the general linear model for analysis of variance along with the least significant difference test. The significance level for both tests was 5%. SPSS 20.0 statistical software (IBM Co., New York, USA) was used to conduct the analysis. The figures were prepared using Sigmaplot 12.5 software (Systat Software Inc., San Jose, CA, USA).

Additional Information

How to cite this article: Sun, H. et al. A two-year field measurement of methane and nitrous oxide fluxes from rice paddies under contrasting climate conditions. Sci. Rep. 6, 28255; doi: 10.1038/srep28255 (2016).