Energy budget and carbon footprint in a wheat and maize system under ridge furrow strategy in dry semi humid areas

The well-irrigated planting strategy (WI) consumes a large amount of energy and exacerbates greenhouse gas emissions, endangering the sustainable agricultural production. This 2-year work aims to estimate the economic benefit, energy budget and carbon footprint of a wheat–maize double cropping system under conventional rain-fed flat planting (irrigation once a year, control), ridge–furrows with plastic film mulching on the ridge (irrigation once a year, RP), and the WI in dry semi-humid areas of China. Significantly higher wheat and maize yields and net returns were achieved under RP than those under the control, while a visible reduction was found for wheat yields when compared with the WI. The ratio of benefit: cost under RP was also higher by 10.5% than that under the control in the first rotation cycle, but did not differ with those under WI. The net energy output and carbon output followed the same trends with net returns, but the RP had the largest energy use efficiency, energy productivity carbon efficiency and carbon sustainability among treatments. Therefore, the RP was an effective substitution for well–irrigated planting strategy for achieving sustained agricultural development in dry semi-humid areas.


Results
Productivity and economics. The wheat and maize grain yields ranged from 4.18 to 9.16 Mg ha −1 season −1 to 8. 40-10.23 Mg ha −1 season −1 during the two rotation cycles (Fig. 1). The WI and RP strategies significantly increased grain yields of both wheat (119.0% and 64.4%, respectively) and maize (21.8% and 18.3%, respectively) relative to those under the control. The average annual wheat yield was significantly lower by 24.9% under RP than that under WI, whereas no significant difference was observed between the WI and RP strategies.
Across the 2 rotation cycles, the WI and RP improved the system productivity by 50.9% and 32.1%, respectively, relative to those under the control (Fig. 2a). The average annual gross return and net return ranged from 28.78 to 43.44 × 10 3 Yuan ha −1 to 14.59-22.86 × 10 3 Yuan ha −1 with the trends of C < RP < WI (Fig. 2b,c). The average annual benefit: cost ratio was 2.03, 2.11 and 2.16 under the control, WI and RP strategies, and no significant difference existed between each strategy for the benefit: cost ratio during the two rotation cycles (Fig. 2d).
The annual average energy output from wheat and maize grains under RP was up to 101,090 MJ ha -1 and 146,168 MJ ha -1 , respectively, which was visibly higher by 64.4% and 18.3% than that under the control, while lower by 24.9% and 2.8% than that under WI (Table 2), respectively. As to the entire rotation cycle, the annual average energy outputs of crop production under RP increased by 33.6% relative to that under the control, while reduced by 13.3% relative to that under the WI (Fig. 3a). The energy output under RP was significantly higher than those under the control, while lower than those under WI in 2012-2013 and 2013-2014, respectively (Fig. 3a). The net energy output under RP was sharply enhanced by 48.9% and 31.8% relative to those under the control in 2012-2013 and 2013-2014, respectively, while had no significant difference with those under WI over 2 rotation cycles (Fig. 3b). The energy use efficiency under RP was higher by18.3% and 7.5% than those under the control, and by 31.2% and 27.0% than those under WI in 2012-2013 and 2013-2014, respectively (Fig. 3c). Meanwhile, the energy productivity had the same trends with the energy use efficiency (Fig. 3d).
Carbon footprint. The annual average CF under RP was obviously higher by 30.9% and 23.8% than those under the control for wheat and maize production, respectively (Table 3). However, there existed no significant difference between WI and RP for maize production, and a 15.4% reduce was found under WI for wheat production ( Table 3). The annual average CF under RP increased by 27.2% relative to the control, while reduced by 6.8% relative to the WI in the entire rotation cycle ( Table 3). The 165 and 1908 kg CO 2 -eq ha -1 was more from uses of farm machinery and plastic film under RP than those under both the control and WI, while 2785 kg CO 2 -eq ha -1 was less from uses of electricity for irrigation under RP than that under WI. Over 2 rotation cycles, the use of fertilizers and electricity for irrigation occupied 36.6% and 33.4% of the total emissions, followed by N 2 O emissions based on estimation (20.8%).
The carbon input under RP was significantly higher by 16.1% and 16.4% than those under the control, while lower by 16.2% and 13.5% than those under WI in 2012-2013 and 2013-2014, respectively (Fig. 4a). The carbon output under RP was significantly higher by 44.8% and 43.9% than those under the control, while lower by 12.3% and 11.5% than those under WI in 2012-2013 and 2013-2014, respectively (Fig. 4b). Meanwhile, the carbon efficiency under RP was significantly higher by 24.7% and 23.7% than those under the control, and slightly higher by 4.7% and 2.2% than those under WI in 2012-2013 and 2013-2014, respectively (Fig. 4c). Additionally, the Table 2. Effect of different planting strategies on annual average energy inputs and outputs (MJ ha -1 ) of wheat-maize cropping system. C, conventional rain-fed flat planting; WI, well-irrigation planting; RP, ridgefurrow planting with plastic film mulch over the ridge. Data are averaged over the two growing cycles. www.nature.com/scientificreports/ carbon sustainability index under RP was significantly higher by 29.6% and 29.0% than those under the control, and slightly higher by 5.5% and 2.6% than those under WI in 2012-2013 and 2013-2014, respectively (Fig. 4d).

Discussion
In the present study, significantly higher grain yields for both wheat and maize were achieved under RP than those under the control in both years ( Fig. 1). However, remarkable reduction was only found for wheat grain yields when compared with the WI over the 2 rotation cycles ( Fig. 1). Those results implied that adopting the RP could substantially promote grain yields under the wheat-maize cropping system in dry semi-humid areas, and that maize yields under RP reached a plateau close to the yield potential ceiling without drought stress 27 . The high grain yields under RP were mainly attributed to the superiority of RP in adjusting soil moisture and temperature to match the needs of crop production 17 . Similar results are also reported by Hu et al. 28 in sub-humid drought-prone and semi-arid regions. Additionally, the maize yields in 2014 with a rainfall of 331 mm did not show any improvement over those in 2013 with a rainfall of 219 mm, although the rainfall increased by 51.1%. This phenomenon was mainly because the larger rainfall before the silking stage in 2013 (Fig. 5), resulting in a dramatically higher soil water storage to promote maize growth than those in 2014 23 . What's more, more solar radiation for improving maize photosynthesis and growth, because the rainy days after silking in 2013 were lower than that in 2014. The total cost of wheat production ranged from 6.266 × 10 3 Yuan ha -1 under the control to 10.466 × 10 3 Yuan ha -1 under WI (Table 1), falling well within the range of 2.402 -10.814 × 10 3 Yuan ha -1 for wheat production reported by recent studies in China 10,29,30 . Likewise, the total cost of maize production ranged from 8.276 × 10 3 Yuan ha -1 under the control to 10.076 × 10 3 Yuan ha -1 under WI (Table 1), which also fell well within the range of 3.185-11.925 × 10 3 Yuan ha -1 reported by Zheng et al. 31 and Liang et al. 32 . Regarding the entire rotation cycle of wheat and maize, the total cost under RP was lower than that under WI. Those phenomena indicated that adopting the RP could reduce the cost of production when compared with the acknowledged high-yield production strategy of WI. Cost incurred for different component of cost analysis for the RP followed the order of farm machinery > fertilizer > labour > plastic film > seeds > irrigation/plant protections ( Table 1). The order and share of different components were changed under the control and WI, because of the changes in costs involved in farm machinery, plastic film, irrigation, and labour. Due to the adoption of supplemental irrigation and water-saving www.nature.com/scientificreports/ measures, the gross returns under the WI and RP were significantly higher than those under the control throughout the two rotation cycles (Fig. 2). The gross return under the control was in proximity to the total economic production gained in the relative drought years, but lower than those in the relatively humid years reported by Lu and Liao 10 . However, the gross returns under WI and RP were also higher than those achieved by Lu and Liao 10 , whether in drought or a humid year. The gross returns from the RP were similar to that (38.122 × 10 3 Yuan ha -1 ) reported under irrigated plots by Cui et al. 29 . Similarly, the net returns under the control in our study throughout the two rotation cycles were significantly lower than those from the WI and RP, and were below the net return values reported by Lu and Liao 10 . These results mainly attributed to the lower rainfall in our study. The net returns under the WI and RP in 2012-2013 also exceeded those gained under irrigated plots by Cui et al. 29 , but the net returns in 2013-2014 had a contrary tendency. The reason for those phenomena might be that the rainfall was not in step with crop growth needs in 2013-2014 (Fig. 5). The ratio of benefit: cost under RP was visibly higher than that under the control in 2012-2013, but did not differ with other treatments over 2 rotation cycles. Consequently, the results confirmed that adopting ridge-furrow planting with plastic film mulch over the ridge was a promising and economical option substitution for supplemental irrigation to produce wheat and maize grain in a dry semi-humid area of China.
The study has showed that the annual energy inputs of wheat production were ranged from 28,395 to 60,255 (Table 2). However, the total energy inputs of wheat production varied from 10,800 MJ ha −1 to 57,800 MJ ha −1 in other studies 8,33,34 . The values has exceeded the reported total energy inputs of wheat production due to the energy inputs from irrigation under WI (Table 2). In previous studies, the energy inputs of irrigation, nitrogen fertilizers, and farm machinery accounted for 23.5-32.1%, 24.0-38.3%, and 30.8-60.2% of the total energy inputs for raising wheat [34][35][36] . But the highest energy inputs under WI, control and RP were irrigation, fertilizer and fertilizer, respectively, which occupied over 40% of total energy inputs of wheat production. In addition, the use of plastic film contributed more than 10% to the total energy inputs under RP. The apparent discrepancy may result from different irrigation strategies and other field managements as well as edaphic and climatic conditions. The total energy inputs of maize production in the study were fairly high compared to other studies of 4200-10,400 MJ ha −1 in Bertocco et al. 37 and of 12,700-23,000 MJ ha −1 in Amaducci et al. 38 . Similar to wheat production, irrigation, fertilizers, and farm machinery were also the main contributors of the energy inputs. In the entire rotation cycle, the total energy inputs showed: WI > RP > control (Table 2), which revealed that the total energy inputs of crop production under RP increased by 18.5% relative to that under the control, while reduced by 32.6% relative to that under the WI. Furthermore, the energy input derived from the irrigation is on the increase due to the decline of groundwater level 39 . This condition approved that adopting energy-save irrigation strategies, such as the ridge-furrow planting with plastic film mulch over the ridge, is urgent to supersede the supplemental irrigation to produce wheat and maize grain in a dry semi-humid area of China.
Values for energy output from wheat grains under RP and WI in the present study were higher than those previously reported values 34,40 , which was mainly due to the higher grain yields under RP and WI. Meanwhile,     www.nature.com/scientificreports/ The obtained net energy outputs under RP and WI were higher than that reported by Singh et al. 41 . Additionally, the energy use efficiency and energy productivity under RP was higher than those under the control and WI in the entire rotation cycle. but the specific energy under RP was lower than those under the control and WI. Those results implied that adopting the RP could reduce direct energy input, offsetting the decreased system productivity and energy output from grain yield, and that adopting the RP can be expected to achieve identical results with those under well-irrigation planting in dry semi-humid regions due to better soil water conservation 23,42 .
As to the entire rotation cycle, the annual average CF showed: WI > RP > control ( Table 3). The primary factors triggering significant differences in the CF among planting strategies were the different uses of farm machinery, plastic film, and electricity for irrigation. The use of fertilizers and electricity for irrigation occupied over 30% of the total emissions under two rotation cycles, which differed from the concept that 75.0% of GHG emissions derived from N fertilizer application during crop production 43,44 . This discrepancy could be because the Loess Plateau of China is susceptible to water scarcity with the evapotranspiration significantly exceeds the available precipitation 45 . Thus, electricity consumption for irrigation water from low groundwater levels per unit amount is larger than other regions. A similar result was also found in the North China Plain 4 . Thus, the RP can be considered as a viable planting strategy for practicing low-carbon agriculture in a dry semi-humid area of China.
The carbon input and carbon output under RP was significantly higher than those under the control, while lower than those under WI in two rotation cycles. Those results indicated the higher input produced more carbon output. For anthropogenic GHG emissions and their resulting global climate change, the sustainability of crop production increases with the increasing use efficiency of Carbon-based inputs 12 . The carbon efficiency and carbon sustainability index under RP was significantly higher than those under the control, and slightly higher than those under WI in two rotation cycles (Fig. 4); which exhibited that the RP was an effective substitution for supplemental irrigation for the mitigation of climate change and the achievement of sustained agricultural development in an intensive maize-wheat cropping system in a dry semi-humid area of China.
Although our study indicated that RF practice have lower carbon footprint and higher carbon efficiency, the use of plastic film can cause a series of environmental problems, for example white pollution, microplastic pollution and soil pollution 46 . After the plastic film was used in farmland, the plastic film cannot be completely removed and recycled and most of it remain in the soil for long time 47 . Which affects soil structure and mechanical tillage, resulted in environment pollution and mechanical damage. With the rapid promotion and application of plastic film in China, plastic film was covered in 19 million ha cropland and reached 2.7 million tons 48 . Fortunately, biodegradable film has similar properties to plastic film and reduce polyethylene residue in soil and plastic pollution 46 . This can be a good option to alternative plastic film and worth futher study for agricultural sustainable development and environmental protection. In addition, although the study and some others similar studies accomplished over a 2-years period [49][50][51] , some studies are more than 2 years, such as 4 or 6 years 1,20 , to reduce the effect of weather variability from year to year on crop growth, yield, irrigation and energy budget, carbon footprint 1,20 . Thus, this study needs to be conducted over a long period of time for further refine the results.

Conclusions
This 2-year study assessed the impacts of different planting strategies on productivity, economic benefit, energy consumption and carbon footprint in an intensive wheat-maize cropping system to identify carbon friendly and cleaner planting technologies in a dry semi-humid area of China. The data showed that grain yields ranged from 3.22 to 9.31 Mg ha −1 for wheat and from 7.6 to 11.6 Mg ha −1 for maize, respectively, with the lowest yields under the control, followed by RP and WI. The gross return and net return had the same trends as those of grain yields, but the benefit: cost ratio was close between the WI and RP. The RP increased the net energy output, energy use efficiency, and energy productivity, but reduced the specific energy relative to the control. The annual average CF under RP increased by 27.2% relative to the control, while reduced by 6.8% relative to the WI. The carbon output under RP was significantly higher by 44.8% and 43.9% than those under the control, while slightly lower by 12.3% and 11.5% than those under WI in 2012-2013 and 2013-2014, respectively. The RP had the largest carbon efficiency and carbon sustainability. Therefore, shifting from planting strategies with supplemental irrigation to the ridge-furrow planting with plastic film mulch over the ridge increases the energy use efficiency and carbon efficiency, and thus provides potential solutions for the development of C-friendly planting technologies in dry semi-humid areas of China or other countries with similar agro-meteorology in the world. Nevertheless, the environment hazards of ridge-furrow planting with plastic film mulch over the ridge also needs to be concerned, for example, "white pollution" from plastic film. The innovation of covering material development and the formulation of related policies urgently need to solve this problem for better agricultural environment.

Methods
Experimental site and climate. The experiment was conducted at the Doukou Experimental Station of Northwest A&F University (34°36′N, 108°52′E) from October 2012-October 2014 in Sanyuan, Shaanxi Province, China. The study area has a temperate, dry semi-humid continental monsoon climate liable to drought with hot summers and cold winters. Based on 30 years' climatic data, the annual average sunshine duration, temperature, and frost-free period was 2096 h, 13.4 °C, and 215 d, respectively. The annual average rainfall was 517.7 mm with 75% occurring from July to September. Precipitation data were recorded using standard weather station (Vantage Pro2, USA) on the experimental site. The daily maximum/minimum air temperature and precipitation distribution during experimental period are presented in Fig. 5. The amounts of precipitation were 183 and 222 mm during wheat growing season, and were 219 and 331 mm during maize growing season in 2012-2013 and 2013-2014 rotation cycles, respectively. The soil is classified as loamy clay 23 . The initial soil (0-20 cm) contained 17.77 g kg −1 SOM, 1.26 g kg −1 total N, 259.48 mg kg −1 available K, 22.08 mg kg −1 Olsen P with a pH of 8.45 (soil/water = 1:1) and a bulk density of 1.20 g cm −3 . www.nature.com/scientificreports/ Experimental details. The field experiment included: conventional rain-fed flat planting (control, C), well-irrigation planting (WI), and ridge-furrow planting with plastic film mulch over the ridge (RP); the detail description was in Li et al. 23,52 . The treatments were applied in 6.4 m × 8 m plots in a randomized complete block design with four replications. The ridge-furrow planting systems were built by changing soil surface into alternating ridges and furrows with 30 and 55 cm in width. The ridges' height was nearly 15 cm. The crops were sown in two rows in the furrows. The cultivars of wheat and maize were Xinong 979 and Zhengnong 9.
To ensure better seedling establishment, the control and RP plots were irrigated with 980 and 1180 m 3 ha −1 at 8 days after sowing (DAG) during the second wheat period, and with 980 and 880 m 3 ha -1 at 12 DAG during the first maize period and 3 days after sowing during the second maize period, respectively. No other supplemental irrigation was performed under control and RP plots. The WI plots were irrigated with 1200, 1100, 1100 and 1000 m 3 ha −1 at 6, 89, 153 and 179 DAG during the first wheat period, with 1180, 1100, 1000 and 1000 m 3 ha −1 at 8, 95, 160, and 180 DAG during the second wheat period, with 980 and 1000 m 3 ha −1 at 12 and 50 DAG during the first maize period, and with 980, 790 and 980 m 3 ha −1 at 3, 33 and 49 DAG during the second maize period, respectively.
During the wheat and maize periods, all of the treatments were fertilized with 90 kg N ha −1 and 50 kg P ha −1 and 30 kg K ha −1 by hand via broadcasting before sowing and then incorporated into the 0-20 cm soil layer with rotary tillage. Additionally, the plots were treated with 67.5 kg N ha −1 during the elongation and heading stages of wheat, and the elongation and tasseling stages of maize, respectively. The N topdressing was performed before raining or irrigation. All of wheat and maize straw were smashed (< 10 cm long) with a residue chopper after harvested with combine-harvesters. The chopped straw was incorporated into the soil by rotary tillage before ridge-furrow tillage. Other field management practices, including field preparation, sowing, harvesting, and the application of insecticides, herbicides and fungicides, followed the locally recommended practice in both years. The inputs are shown in Table S1.
Yield measurements. At maturity, maize and wheat grains were manually harvested in duplicate from the center (6 and 2 m 2 for each crop) of each plot every year. After air-drying, portions of grain were oven-dried at 60 °C for grain determination. System productivity in term of wheat equivalent yields (WEY) was estimated to compare the effects of different treatments on crop performances by converting grain yields of both crops into the WEY on the basis of market price followed with the Eq. (1): where WEY is the system productivity; M p and W p are the market price of maize and wheat grains. During the study period, the annual average maize and wheat grain prices were 2.40 and 2.06 Yuan kg -1 , respectively. Economic analysis. The economic analysis was computed by assessing a range of components, including the cost of cultivation (C tot ), gross revenue (GR), economic profit (EP), and the ratio of net income to cost (RIC). These analyses were conducted based on the prevailing market price of the inputs, outputs, and services, and were followed with the equations [Eqs. (2)-(5)] suggested by Lu and Liao 10 .
where, RIC is the ratio of net income to cost. Energy analysis. The energy inputs and outputs of each treatment were estimated based the complete record of all inputs (Table S1) and outputs (grain yields).
where, EO is the total energy out (MJ ha −1 ). Y is the grain yields (Mg ha −1 , OW). EC is the corresponding energy coefficient of grain yields.
where, NEO is net energy out (MJ ha −1 ).
where, EUE is the energy use efficiency (%).
where, EP is the energy productivity. WEP is the system productivity.
Carbon footprint (CF). The CF was been used to assessed environmental impacts of different planting patterns, because the CF can be as a powerful tool to know and build more environmentally friendly crop production systems 53,54 . The CF is the total amount of GHG emissions (CO 2 and N 2 O, CO 2 equivalents) throughout the crop growth 55 . Because of CH 4 emission was often negligible in dry semi-humid regions, our recent study only considered the N 2 O and CO 2 gases. The N 2 O was converted into 265 CO 2 equivalents 3 . The corresponding emission coefficients, which was presented in Table S3, were used to calculated the GHG emissions of the field operation and inputs. In fields, ammonia volatilization was determined from fertilizer-N using rates of 23% and 26% for wheat and maize, respectively 56 . Nitrate leaching was determined from fertilizer-N using rates of 14% and 16% for wheat and maize, respectively 43 . Direct N 2 O emissions came from 1.25% of fertilizer-N 56 . Indirect N 2 O emissions were estimated by 1% of ammonia-N and 2.5% of nitrate-N, respectively 56 . The carbon footprints (CF, kg CO 2 -eq ha -1 ) was obtained using Eq. (11): where, CF is the energy productivity.
Carbon output, carbon efficiency, and carbon sustainability index. The carbon output is the total carbon equivalent of grain, straw, stubble and root biomass produced by the crop 57 . The below-ground root biomass represented 22% and 23% of wheat and maize straw biomass, respectively 58 . The proportions of stubble to straw biomass were estimated to be 20% and 10% for wheat and maize, respectively. The carbon content was assumed to be 40% for both wheat and maize biomasses. Carbon efficiency was calculated as the ratio of carbon output to carbon input, and the carbon sustainability index was estimated by computing the difference between carbon output and carbon input and dividing it by carbon input 1,12,59,60 .
Statistical analysis. Statistical analyses were performed by using Excel 2013 and SPSS 19.0 (SPSS Inc., Chicago, IL, US). The mean differences among treatments were determined by the Duncan multiple range test at P < 0.05.