Increase energy use efficiency and economic benefit with reduced environmental footprint in rice production of central China

Identifying an energy-efficient system with low energy use, low global warming potential (GWP), and high profitability is essential for ensuring the sustainability of the agro-environment. Given the global importance of China’s rice production, this study determines energy, environmental, and economic performances of transplanted (TPR) and direct-seeded rice system (DSR) in central China. The results showed that total energy inputs for TPR and DSR were 31.5 and 22.8 GJ ha-1 across two growing seasons, respectively. Higher energy input for TPR primarily resulted from extra energy use of the nursery beds and transplanting. Higher energy output of DSR (202.5 GJ ha-1) over that of TPR (187.7 GJ ha-1) was due to a slightly higher yield from DSR. Therefore, DSR exhibited significantly higher energy use efficiency than that of TPR. Lower specific energy for DSR (2.78 MJ kg-1) relative to TPR (4.02 MJ kg-1) indicated that the energy used to produce per unit of rice grain could be reduced by 30.8% by adopting DSR. On average, GWP of DSR was reduced by 5.6% compared with TPR. Moreover, DSR had a 55.8% higher gross return and a 25.7% lower production cost than those of TPR. Overall, compared with TPR, DSR has the potential to increase gross economic return and energy output with reduced energy input and emissions. Therefore, this study suggests that DSR is an environmentally-sound and economically-viable production system. As such, DSR is noted as an energy-efficient and climate-smart production system that could be used by policymakers and farmers to achieve not only improvements in the environment but also financial benefits.


Introduction
The agriculture sector has been extensively shown to face a great challenge in producing sufficient food for growing population under the pressures of decreased cropland area, climate change, and the need to protect environment (Gordon et al. 2005;Chen et al. 2020). Notably, energy input, environmental impact, crop productivity, and economic benefit are closely linked in agriculture (Mohammadi et al. 2017;Rautaray et al. 2020). Energy input is critical to agriculture, and it has increased substantially throughout the years in response to growing food demand (Yuan and Peng 2017a). However, increased agricultural inputs not only cause high financial burdens but affect the economic return to farmers (Pokhrel and Soni 2017;Wu et al. 2017). Agricultural production is a critical source of greenhouse gas (GHG) emissions, accounting for 14-17% of global anthropogenic emissions (Maraseni and Qu 2016;Meena et al. 2017). Intensive energy input in forms of irrigation water, fertilizer, machinery, diesel oil, and pesticides have dramatically increased GHG emissions (Yuan et al. 2019), which in turn threaten the sustainability of agriculture by contributing to global warming (Maraseni et al. 2015;Chen 2016). Therefore, reducing energy expenditure and enhancing energy use efficiency (EUE) are crucial to increasing production and minimizing environmental impact (Mohammadi et al. 2014). These concerns are particularly strong for rice production in China.
Rice is one of the most important staple crops in the world; more than half of the global population depends on rice for food (Godfray et al. 2010). China is the largest rice producer and consumer worldwide, accounting for approximately 28% of total global rice production (FAOSTAT 2020). It is forecasted that approximately 60% more rice than the current level will be needed by the year 2025 (Normile 2008). However, there is a growing concern regarding this increase, as rice cultivation is a significant contributor to climate change because it aggravates global warming potential (GWP) (Groenigen et al. 2013;Thanawong et al. 2014). Therefore, identifying a sustainable rice cultivation system with high crop productivity and reduced energy input and GHG emissions is fundamental for ensuring food and ecological security (Pokhrel and Soni 2017;Coltro et al. 2017;Pittelkow et al. 2014;Maraseni et al. 2018).
Rice is generally grown in transplanted (TPR) or directseeded system (DSR). As reported previously, TPR is the dominant rice cropping system in China (Peng et al. 2009). Compared with DSR, TPR requires more inputs, particularly at the time of land preparation and transplanting (Bouman 2009;Farooq et al. 2011), which not only leads to high energy use, and thereby more environmental issues but also reduces economic benefits for farmers . Given the advantage of reducing labor inputs (Kumar and Ladha 2011;Pathak et al. 2013), increasing attention has been paid to DSR. It was reported that DSR had a lower energy input and higher EUE than TPR according to data collected from local rice farmers in Iran (Eskandari and Attar 2015). Additionally, GWP of DSR was approximately 68% lower than that of TPR (Tao et al. 2016). A meta-analysis revealed that DSR was an alternative method with yield advantages and higher economic return relative to TPR (Chakraborty et al. 2017). Moreover, it was noted that grain yield of DSR was similar to or higher than TPR with optimized crop management practices (Xu et al. 2019), while reducing emissions ).
However, little information has been paid on the complex linkages among energy use, productivity, emissions, and financial return for DSR and TPR in China. Clearly, an environmentally and economically viable cultivation system is needed, and this system must be assessed based on knowledge of their energy, GWP, and profitability. Exploring this interaction can not only fill this knowledge gap in the literature but also provide scientific information for policymakers to promote an energyefficient system. As such, this study was conducted with the following specific objectives: (i) assessing the energy input and its use efficiency in DSR and TPR, (ii) identifying the environmental and financial performances of DSR and TPR, and (iii) estimating the eco-efficiency difference between DSR and TPR.

Site description
This experiment was conducted in Wuxue County, Hubei Province, China. The site is located at 29°51′ N and 115°33′ E at an altitude of 23 m ASL (meters above sea level). Hubei Province is located in central China in the basin of the Yangtze River and represents a typical agricultural region in central China. The total arable land in Hubei is 3.4 Mha, of which 2.0 Mha is cultivated with rice crops (Statistical Bureau of Hubei Province 2020). The research field is located in a subtropical monsoon climate region, where summers are hot and rainy and winters are cool and dry. During the last decade, the mean annual rainfall, mean daily solar radiation, and mean annual temperature were 1310 mm, 12.5 MJ m -2 , and 18.1°C, respectively. The monthly distributions of precipitation, solar radiation, and minimum and maximum temperatures in 2014 are depicted in Fig. 1. The soil (0-20 cm) of the experimental site had a clay loam texture with pH of 5.35, organic matter of 22.4 g kg -1 , total N of 2.09 g kg -1 , available P of 22.2 mg kg -1 and available K of 108.7 mg kg -1 .

Crop management
In this area, rice is grown under both DSR and TPR. To attain overviews of the EUE, GWP, and financial benefit involved in rice production, a replicated field experiment was conducted to study the impact of rice cultivation system on the energy, environmental, and economic performance. The experiment was performed during the early (April to July) and late (July to November) rice-growing seasons in 2014. Six rice varieties that are commonly grown in central China were used during each growing season.
For TPR, pregerminated rice seeds were sown in a nursery, and 30-day-old and 20-day-old rice seedlings were transplanted into well-prepared paddy soil on May 3 and July 29 during the early and late rice-growing seasons, respectively. Three seedlings per hill were transplanted at a hill spacing of 20.0×20.0 cm for both seasons. For DSR, pregerminated seeds were manually and uniformly broadcasted into standing water using the recommended seeding rate of 90 kg ha -1 on April 13 for early-growing season and on July 15 for late-growing season. Chemical fertilizers were applied in accordance with the recommended fertilizer dose for rice production in this area. In this study, 176 kg ha -1 N, 72 kg ha -1 P 2 O 5 , and 113 kg ha -1 K 2 O were applied to the two systems during both growing seasons. Approximately 48% of N, all of P 2 O 5 , and 40% of K 2 O were applied as basal fertilizer. The remaining N was top-dressed in two split doses during middle tillering and panicle initiation stages, and the remaining 60% of K 2 O was top-dressed at panicle initiation.

Energy analysis
The energy analysis presented in this study compared energy input, output, and EUE between DSR and TPR. Energy input was estimated using direct and indirect energy inputs. Direct energy inputs include energy from diesel, water, and labor used in crop production. Indirect energy inputs consist of seed energy and energy used in the production of machinery, fertilizers, and pesticides (Yuan and Peng 2017b). Energy inputs are also classified as renewable and nonrenewable forms. Renewable energy includes labor, seed, and water for irrigation, while nonrenewable energy consists of machinery, fuel, fertilizers, and pesticides (Kazemi et al. 2015).
Energy equivalents shown in Table S1 were used to estimate input and output energy. All of the inputs used in rice production, including fertilizers, seeds, pesticides, diesel, labor, water, and machinery, were determined and quantified. The grain and straw yield were considered as components of output energy. Energy input and output were calculated by multiplying inputs and outputs with the corresponding energy equivalents and summation of all corresponding components. Net energy (NE), EUE, specific energy (SE), energy productivity (EP), and energy profitability (EPB) were calculated as follows (Pokhrel and Soni 2017;Lal et al. 2015): Energy use efficiency EUE ð Þ Energy profitability EPB ð Þ¼ Net energy GJ ha −1 À Á

Environmental analysis
The environmental impact was assessed by determining GWP. Thus, the total amount of GHG emissions in relation to CO 2 -eq was calculated from CO 2 , N 2 O, and CH 4 released directly and indirectly during rice production. There are three major sources of GHG emissions in rice production: (i) GHGs emissions from production, packaging, and transportation of agricultural inputs, (ii) N 2 O emissions due to nitrogen (N) application, and (iii) CH 4 emission from rice paddy field (Yuan et al. 2019). The GHGs emissions from each of agricultural inputs and direct and indirect N 2 O emissions due to nitrogen fertilizer application were estimated using emission factors given in Table S2. The CH 4 emission from rice cultivation was estimated following 2019 Refinement to the 2006 IPCC Guidelines for National Greenhouse Gas Inventories (IPCC 2019), using a daily emission factor of 1.32 kg CH 4 ha -1 day -1 , using scaling factors accounting for water regime during rice cultivation period, water regime in the pre-season before rice cultivation, and type and amount of organic amendment applied. As reported previously (IPCC 2007), N 2 O and CH 4 were converted into CO 2 -eq using GWP equivalent factors of 298 and 25, respectively, over a 100-year time horizon. The GWP from each system was then calculated by multiplying input by its corresponding CO 2 -eq emission factor. Yield-scaled GWP (GWPi; kg CO 2 -eq Mg -1 grain), also known as GHG intensity, was calculated as follows: Economic analysis This study applied economic input-output analysis to explore economic benefits in each system. This inventory was similar to that for energy input-output analysis, and the same inputs and outputs per hectare were applied. Production cost and gross return were calculated by multiplying inputs and outputs with the corresponding prices and summation of all corresponding components (Table S3). The prices of all the inputs and output were defined by market prices in US$ that prevailed in 2014. Total production cost and gross return were determined by multiplying inputs and output by the corresponding market price and summation of all the corresponding components, respectively. The net return and benefit-cost ratio were estimated as follows: Additionally, eco-efficiency defined as the ratio of economic benefit to environmental impact was investigated as follows (Kumar et al. 2016):

Statistical analysis
Data on the energy, environmental, and economic metrics in the two cultivation systems were subjected to statistical analyses of variance, and means were compared using the least significant difference (LSD) at the 0.05 level of probability. Statistix 8.0 (Analytical Software, FL, USA) was used as the statistical software. All the figures were generated by SigmaPlot 12.5 (SPSS Inc., Point Richmond, CA, USA).

Energy analysis
The quantities of agricultural inputs and energy input from each item in DSR and TPR are presented in Table 1. It showed that the difference in energy use between the two systems was significant during both seasons. In accordance with previous study (Chaudhary et al. 2017), TPR exhibited higher energy input than DSR. In the present study, total energy input in TPR and DSR was 33.2 and 22.7 GJ ha -1 in early-growing season and 29.9 and 23.0 GJ ha -1 in late-growing season, respectively (Table 1). Higher energy used under TPR was primarily due to extra energy input for nursery and transplanting, which were reported to be large energyconsuming farm operations (Pathak et al. 2013;Chaudhary et al. 2017). Energy consumption of rice production determined in this study was higher than that in Nepal ( The results clearly suggested that other than seeds and pesticide, energy input from machinery, diesel, labor, fertilizer, irrigation water, and plastic film were higher in TPR than DSR (Table 1). Compared with TPR, higher energy use from pesticide was primarily related to more intensive use of herbicide because of aerobic condition in DSR, which favored weed growth (Eskandari and Attar 2015;Canakci et al. 2005). Moreover, seeding rate for DSR in this area was higher than the recommended (Sun et al. 2015). Thus, there is a room to reduce energy use from seed and herbicide use in DSR by optimizing crop management practices.
As illustrated in Table 1, energy input from fertilizer contributed the highest proportion to total energy use, accounting for an average of 53.3% (16.8 GJ ha -1 ) and 60.5% (13.8 GJ ha -1 ) in TPR and DSR, respectively. Notably, some studies stated that fertilizer input was excessive for paddy fields in this area (Peng et al. 2010;Xu et al. 2016). Regarding high energy expenditure from fertilizer, adopting improved nutrient management, such as site-specific nutrient management and integrated soil-crop system management (Peng et al. 2006;), in particular for N fertilizer, is a crucial step towards decreasing total energy input and reducing environmental pollution (Mohammadi et al. 2014;Kazemi et al. 2015). On average, diesel was second in the order of importance and accounted for 20.9% and 17.0% of total energy input in TPR and DSR, respectively (Table 1). Next, came irrigation water, which was responsible for an average of 13.5% and 11.2% of total energy input for TPR and DSR, respectively (Table 1). Energy from fertilizer, diesel, and irrigation contributed more than 85% to total energy input, which indicated that research could be oriented to reduce energy input from these three major sources.
As shown in Fig. 2, indirect and nonrenewable energy was higher than direct and renewable energy, respectively. On average, the share of indirect energy to total energy consumption was 63.2% and 69.7% in TPR and DSR, respectively. The results showed that 17.9% of total energy input in rice production was renewable, while the contribution of nonrenewable energy was 82.1%. These results were in conjunction with previous studies that share of nonrenewable energy was much higher than renewable energy (Singh et al. 2007;Aghaalikhani et al. 2013), which further indicated that rice production was heavily based on nonrenewable energy. It has to be noted that renewable and nonrenewable energy forms in DSR declined by 19.9% and 29.2% compared with TPR, respectively. This finding clearly suggested that DSR can effectively reduce energy consumption, especially for energy from nonrenewable form. This result is helpful for promoting the sustainable development of rice production.
Differences in grain and straw yields were observed between the two cultivation systems and, thus, resulted in different energy outputs (Fig. 3). Total energy outputs for TPR vs. DSR were 181.3 vs. 201.6 GJ ha -1 during early-growing season and 194.0 vs. 203.4 GJ ha -1 during late-growing season, respectively (Fig. 3). The finding of higher energy output for DSR than TPR in the current study was in contrast with that reported by Eskandari and Attar (2015). Total energy outputs were reportedly 114.7 GJ ha -1 for TPR and 98.7 GJ ha -1 for DSR in Iran. In our study, higher energy output of DSR was due to slightly higher crop yield than TPR (Fig. 3). Notably, several previous studies found a positive yield response in DSR based on on-station, on-farm, and farmer household survey studies (Chakraborty et al. 2017;Tabbal et al. 2002;Singh et al. 2001). The yield increase in DSR could be the result of earlier planting to achieve timelier crop establishment than TPR (Quilty et al. 2014).
As a result, the differences in NE, EUE, SE, EP, and EPB between the two systems were significant (Table 2). On average, NE and EUE were 156.1 GJ ha -1 and 5.98 in TPR, which were significantly lower than a respective of 179.7 GJ ha -1 and 8.87 in DSR. Higher NE and EUE in DSR were due to higher energy output and lower energy input relative to TPR. A higher EUE in DSR than TPR was also reported by others (Eskandari and Attar 2015). The EUE derived in this study was higher than that reported in many studies (Kazemi et al. 2015;Quilty et al. 2014), but still in the range of 1.5-11.0 for rice determined by Mushtaq et al. (2009) andPishgar-Komleh et al. (2011). Averaged over growing seasons, EP was 362.1 kg GJ -1 in DSR and 251.8 kg GJ -1 in TPR, which suggested that substantially lower energy expenditure would be required to produce equal amounts of rice grain in DSR compared with TPR. More importantly, this further indicated that DSR is an alternative system to reduce energy consumption

Environmental analysis
In this study, GWP and GWPi were significantly influenced by rice cultivation system during both growing seasons. The DSR emitted an average of 5.0 Mg CO 2 -eq ha -1 , compared with 5.4 Mg CO 2 -eq ha -1 in TPR (Table 3). A 5.6% lower GWP in DSR was mainly due to lower emissions from agricultural inputs (including N 2 O from nitrogen application), though CH 4 emissions from DSR were higher due to longer crop growth duration in the main field as compared with TPR. The environmental impact was also expressed per mass of rice grain produced in this study. On average, GWPi was 618 kg CO 2 -eq Mg -1 grain in DSR, which was slightly lower than that of 684 kg CO 2 -eq Mg -1 grain in TPR (Fig. 4). Lower GWPi in DSR was mainly due to decreased GWP as compared with TPR.
Across growing seasons and cultivation systems, GWP from in-season CH 4 emissions contributed between 56 and 66% to total GWP, which was in accordance with several previous studies showing that CH 4 accounted for the majority of GWP from rice production (Groenigen et al. 2013;Kumar et al. 2016), suggesting the importance of mitigating CH 4 emissions from rice production. CH 4 emissions are impacted by such several factors as water management, application of inorganic and organic fertilizers, rice cultivars, and properties of paddy soil (Gutierrez et al. 2013). This study indicated that more attention should be paid on determining the role played by these factors in CH 4 emission under TPR and DSR conditions with the goal of reducing GWP. The results also implied that agricultural inputs such as N fertilizer and diesel were principal sources of CO 2 and/ or N 2 O emissions (Mohammadi et al. 2014;Chaudhary et al. 2017). Thus, reducing fertilizer and diesel consumptions through increasing N use efficiency and mechanical efficiency are critical for both transplanting and directseeding rice. Overall, the results of environmental analysis indicated that DSR could be an important approach for reducing GWP from rice farming.

Economic analysis
The economic input and output of rice production under DSR and TPR are shown in Table 4. The results revealed that the differences in total production cost, gross return, net return, and benefit-to-cost ratio were significantly different between the two systems during both growing seasons. Based on the average of two seasons, DSR reduced total production cost by 25.7%, corresponding to 504.3 US$ ha -1 (Table 4). The costs of production estimated in our study were similar to those reported for rice production in China by other studies Liu et al. 2014). Lower production cost for DSR was primarily due to lower cost of machinery, diesel, and labor due to the omission of transplanting relative to TPR. A slightly higher gross return of 140.8 US$ ha -1 in DSR than TPR was observed.
Average net return for DSR (1800.3 US$ ha -1 ) was 55.8% higher than that of TPR (1155.2 US$ ha -1 ) ( Table 4). Higher net return for DSR mainly resulted from its lower production cost compared to TPR. This was in line with result that the financial benefit of DSR (1568.6 US$ ha -1 ) was 49.5% higher than TPR as determined by Chi et al. (2008). A significantly  higher benefit-to-cost ratio was also found in DSR (2.13 and 2.36 in early-and late-growing season, respectively) than in TPR (1.47 and 1.72). As one of the major financial indicators, benefit-to-cost ratio has been well-documented in the literature, and it was reported to range from 0.8 to 3.1 for rice production (Pokhrel and Soni 2017;Gathala et al. 2015).
Higher net return and benefit-to-cost ratio of DSR may explain the gradual transition from TPR to DSR as occurred in China.
In this study, eco-efficiency was estimated as the ratio of product or service value to the corresponding environmental cost. We note that US$ 329 and 381 were obtained from 1 Mg of CO 2 -eq emissions in early-and late-growing season under DSR, respectively, which was 81% and 52% higher than those observed under TPR (Fig. 4). This finding implied that an average of around US$ 140 more was obtained from one Mg of CO 2 -eq emissions in DSR than TPR. Clearly, DSR was an eco-efficient cultivation system. Eco-efficiency estimated in this study was higher than US$ 82-134 per Mg of CO 2 -eq reported in northeastern Thailand due to lower yield and higher GWP in that study (Thanawong et al. 2014), but lower than an average of 720 US$ per ton of CO 2 -eq across various rice-based cropping systems in Nepal (Pokhrel and Soni 2017). The disparity indicated that the economic return per unit of GHGs emissions could be further improved by innovating cropping systems.
Scaling up of this eco-efficient technology can be realized by issuing policies. The government could place more Table 3 Global warming potential expressed in CO 2 -eq from methane, nitrous oxide, and agricultural inputs (including diesel, fertilizers, machinery, and pesticides) for transplanted (TPR) and direct-seeded (DSR) rice cultivation systems during early-and late-growing seasons  Fig. 4 Yield-scaled global warming potential (A and B, kg CO 2 -eq Mg -1 grain) and eco-efficiency (C and D, US$ Mg -1 CO 2 -eq) for transplanted (TPR) and direct-seeded (DSR) rice cultivation systems during early-(A and C) and late-growing seasons (B and D). Data are means ± standard error emphasis on providing farmers with information on DSR technology through increasing investments in farmers' training and education. On the one hand, prioritizing R&D programs for DSR research plays an important role in solving constraints occurred presently and developing advanced agricultural technology, including reducing fossil energy input, increasing the usage of renewable energy, and improving the efficiency of agricultural machinery. On the other hand, researcher and agricultural extension staff should work together to help farmers adopt existing crop management practices and further optimize farming operations of DSR in agricultural fields to reduce external inputs with the goals of reducing energy use and environmental impact. Overall, direct-seeded rice has a potential of decreasing energy input and emissions while increasing economic return, without sacrificing productivity and energy output. There are still constrains for DSR adoption including (a) poor crop establishment due to a range of abiotic stresses, especially low temperature, drought, and waterlogging, and (b) high weed infestation especially in dry field conditions, and (c) lodging associated with decreased breaking resistance of the rice internode due to high plant density . Again, we discussed energy balance and economic and environmental performance between the two cultivation systems based on the results obtained from the two rice growing seasons in a year. The food-energy-emission nexus performance of DSR and TPR needs to be confirmed by more filed experiments in multiple environments. From the perspective of energy-saving and production cost-reducing as well as emission-mitigation, DSR technology holds good promise to increase energy use efficiency and economic benefit with reduced environmental impact in the context of energy crisis and environment degradation. Therefore, findings derived from the current study are relevant for other lowland rice systems in the world, especially in regions with high production cost and substantial resources inputs.

Conclusions
The aim of this study was to identify a sustainable rice cultivation system using energy, environmental, and financial analyses. Energy input was higher in TPR than DSR, which demonstrated that adopting DSR was an effective way to reduce energy expenditure, especially for nonrenewable energy. Furthermore, DSR showed significantly higher NE, EUE, and EP compared with TPR, which was primarily due to its higher energy output. Lower SE for DSR indicated that the energy used to produce one unit of grain can be reduced by 30.8% by DSR compared with TPR. The environmental analysis showed that GWP and GWPi were both lower in DSR than TPR. This result revealed that DSR was a clean production technology and a promising system to reduce emissions from rice production. Moreover, lower production cost and higher gross return indicated that DSR was an economically sound system. This result was also confirmed by the evaluation of eco-efficiency, which noted DSR as an eco-efficient system. Adopting DSR may help rice-producing farmers with limited agricultural inputs to reduce production cost without compromising productivity and energy output.

Declarations
Ethics approval and consent to participate Not applicable

Consent for publication Not applicable
Competing interests The authors declare no competing interests.