Carbon and Nitrogen Footprints of Major Cereal Crop Production in China:A Study Based on Farm Management Surveys

XU Chunchun,CHEN Zhongdu,JI LongLU Jianfei

(1College of Agriculture,Yangzhou University,Yangzhou 225009,China;2China National Rice Research Institute,Hangzhou 310006,China;#These authors contributed equally to this work)

Abstract:Greenhouse gas (GHG) emissions and reactive nitrogen (Nr) releases are central environmental problems,which are closely linked to climate change,environmental ecology and crop production.Sustainable development of agriculture plays an important role in GHG emissions and Nr loss.The life cycle assessment (LCA) method was used to calculate the product and farm carbon footprints (CFs) and nitrogen footprints (NFs) in rice,wheat and maize production in China based on farm survey data.The results pinpointed that the CFs of rice,wheat and maize were 0.87,0.30 and 0.24 kg/kg.Meanwhile,the computed NFs were 17.11,14.26 and 6.83 g/kg,respectively.Synthetic nitrogen fertilizer applications and methane (CH4) emissions were dominant CF sources,while ammonia (NH3) volatilization was the main NF contributor.Moreover,significant decreases in CF and NF by 20%-54% and 33%-61%,respectively,were found in large-size farms (＞ 20 hm2) when compared to small-size farms (＜ 0.7 hm2).Furthermore,the significantly positive relationships between CF and NF indicated the potential for simultaneous mitigation in the regions with high agricultural inputs,like amounts of fertilizer.Based on our results,some effective solutions would be favorable toward mitigating climate change and eutrophication of the major cereal crop production in China,especially optimizing fertilizer use and farm machinery operation efficiencies,as well as developing large-size farms with intensive farming.

Key words:carbon footprint;nitrogen footprint;life cycle assessment;grain crop;sustainability

Climate change and eutrophication pollution are two of the most crucial environmental concerns (Davis et al,2017) that seriously threaten the well-being of humankind and other organisms on our planet.Agriculture is one of the principal contributors to anthropogenic greenhouse gas (GHG) emissions,especially non-carbon dioxide (CO2) emissions,such as methane (CH4) and nitrous oxide (N2O) (Stocker et al,2013;Yang et al,2021).GHG emissions are further exacerbated by economic development,population growth,increasing energy,and use of chemical fertilizers,pesticides and agricultural films instituted to maintain the grain yield.Moreover,a significant portion of nitrogen (N) fertilizers applied each year is released into the environment as reactive nitrogen (Nr,all N species except N2),which causes a cascade of environmental problems,including air pollution,eutrophication and stratospheric ozone depletion (Galloway et al,2008;van Grinsven et al,2013).In addition,the indirect effects of secondary air pollutants,like secondary particulate matter from Nr deposition,are more of a concern to human health and surrounding ecosystems (Moldanová et al,2011).Quantifying and assessing the magnitude of the impacts of carbon (C) and Nr emissions on agroecosystems can facilitate the identification of potential solutions to mitigate climate change and further environmental issues.The results will also help to improve the policymakers understanding of C and N pollution in agricultural production of China,so as to strengthen the development and application of environment-friendly technologies.

Footprint measures are increasingly used to incorporate considerations about sustainability in the selection and development of alternate products and processes.The concept of footprint originated in the early 1990s when the term ‘ecological footprint’ was coined by Rees (1992).Carbon footprint (CF) is widely used to compare the impacts of different products on climate change,and to explore mitigation measures for GHG emissions (Yan et al,2015).The nitrogen footprint (NF) is an indicator expressing the total amount of Nr lost to the environment due to human activities (Leach et al,2012).Since then,measures of CF and NF have joined the indicator family for sustainable decision-making (Wiedmann and Minx,2008;Leach et al,2012).Even though such measures have been around for about three decades and used widely by corporations,governments and nongovernmental organizations,activities causing unprecedented environmental and ecosystem degradation have continued (Xue et al,2016).Huge disruptions in the C and N cycles are beyond the safe operating limits of the earth system (Lorenz and Lal,2009).Recent studies have attempted to simultaneously estimate the CF and NF of crop production (Xue et al,2016).Integrated assessments are preferred when seeking to understand tradeoffs or synergies and possible simultaneous mitigation practices.For example,ammonia (NH3) volatilization,N2O emission and N leaching can be promoted linearly or exponentially,through N fertilization (Chen et al,2014),and N fertilizer production is also an important contributor to CFs (Zhang et al,2013).Moreover,N fertilization has been reported to either enhance the activities of methanotrophs or promote inputs of root exudates,by stimulating crop growth (Chen et al,2016),making it important in regulating CH4emissions,a hotspot sector of CF for rice production (Chen et al,2014).In the UK,for example,quantifying the CFs of major staple crops under different farming systems has identified N fertilizer as a main emission source in crop production (Hillier et al,2009).Similarly,N fertilizer additions are known to promote the releases of various Nr species,linearly or exponentially (Cui et al,2013).Evaluating the CF and NF of staple food production can assist farmers and consumers to understand the overall environmental effects of food production (Xia and Yan,2012).

Globally,agriculture is one of the main non-CO2GHG emission sources of China,accounting for about 16% of the total emissions,which includes 40.5% of the total CH4and 65.5% of the total N2O emissions.It is the third largest contributor of GHG after energy and industry in China (Chen et al,2021).In 2001-2010,China used about 2.7 × 1010kg of N fertilizer per year for crop production,mainly to produce rice,wheat and maize (Huang and Tang,2010).At present,the large input and low efficiency of resources and energy in food production aggravate the deterioration of climate and environment (Chen et al,2014).In any case,the use of various related resources,such as pesticides and fertilizers,is likely to increase continuously (Ray et al,2012;Shen et al,2013).Simply put,many Chinese farmers may buy and use more agricultural materials,but without significantly improving the net economic benefits (Zhang et al,2012).Meanwhile,investigation has occurred concerning the CF and NF of Chinese agriculture but only a few environmental analyses have considered both the carbon CF and NF in producing staple foods at the farmer scale in a region.This lack of knowledge hinders the exploration of regional-or/and crop-oriented mitigation options.Little information exists about the CF assessment of staple grain production,and no information has been presented about the NF of food staples.Hence,this study used the life cycle assessment (LCA) approach to:(1) assess the CFs and NFs of rice,wheat and maize using farm survey data,(2) analyze the prime driving forces of CFs and NFs of three grain crops on provincial levels,and (3) compare the CFs and NFs of crop production between climate regions and farm sizes.

RESULTS

Farm size,grain yield and agricultural input

Differences were observed in grain yield,farm size and input of agricultural capital.The surveyed farm size in 0.1-0.5 hm2for rice,wheat and maize accounted for about 90% of total farmers,showing the great fragmentation of China’s croplands.The farm sizes of rice and wheat were larger than those of maize.The highest grain yield was found in maize.Grain yield of rice was higher in Jiangxi than in Hunan.Wheat yield between Jiangsu and Anhui and maize yield between Hebei and Jilin both had no significant differences.The input from diverse forms of synthetic fertilizer followed the order:N fertilizer ＞ K2O fertilizer ＞ P2O5fertilizer for rice and wheat.N fertilizer use ranged from 141.7 kg/hm2to 460.6 kg/hm2.The mean N fertilizer use was the highest in rice (363.2 kg/hm2) and the lowest in maize (172.8 kg/hm2).For wheat,N fertilizer was applied at a higher level in Jiangsu than in Anhui.For maize,the N application was higher in Jilin than in Hebei (Table 1).Diesel oil,a large input of agricultural resources,was used by over 80% of the total farms surveyed in the range of 61.8-163.3 kg/hm2.Film was not used for crop production much,except in rice,where 5.5-8.5 kg/hm2films were used in Jiangxi and Hunan.

GHG and Nr releases from agricultural inputs

The average GHG emission from agricultural inputs was rice ＞ maize ＞ wheat (Table 2).The GHG emissions of synthetic fertilizer production and application (including N,P2O5and K2O fertilizers) were the largest contribution to the total agricultural inputs.In addition,GHG emissions from N fertilizer were the highest in rice compared to wheat and maize.For P2O5fertilizer,maize ranked higher as source of GHG emissions,followed by rice and wheat.Following synthetic fertilizers,diesel oil was the second largest contributor to GHG emissions,accounting for 36.9%,47.2% and 40.9% for rice,wheat and maize,respectively.The GHG emissions from seeds were greater in rice than in maize and wheat.The GHG emissions from pesticide,associated with herbicide,insecticide and fungicide,were the lowest sources.The Nr emissions related to agricultural inputs for wheat were slightly higher than that for maize,but lower than that for rice.Different to GHG emissions,the Nr emissions of diesel oil consumption shared the largest percentage of agricultural inputs (Table 2).Next to diesel oil consumption,synthetic fertilizers emitted 278.8,227.3 and 222.7 g/hm2per year for rice,wheat and maize,respectively,accounting for 31.2%,28.5% and 30.7%,respectively.The pesticides were still the least contributor of Nr emissions in all grain crops,accounting for less than 2% of total Nr emissions from agricultural inputs.

Carbon footprint (CF) and nitrogen footprint (NF) of grain crop production

The CF of rice was 2.8 and 3.6 times higher than those of wheat and maize,respectively.This CF was largely attributable to higher CH4emissions from rice,accounting for 63%,and the second largest was hidden GHG emissions,accounting for 26.7%,while the N2O emissions from rice had a small impact on CF.Although secondary contributor to the CF in rice,agricultural inputs were the largest contributor in wheat and maize,accounting for 64.5% and 75.0%,respectively (Table 3).The NF for rice,wheat and maize at yield-scale were 17.11,14.26 and 6.83 g/kg per year,respectively.The NF of maize was obviously less than that of wheat and rice,and values were similar between wheat and rice (Table 3).NH3volatilization from fields associated with N fertilizer for all grain crops,accounting for 95.8%,33.6% and 31.9% for rice,wheat and maize,respectively.TheNH4+leaching from maize had a small impact on NF,measured at only 18.4 mg/kg per year at yield-scale.

Table 1.Life cycle inventory dataset of farm size,grain yield,agricultural inputs of rice,wheat and maize production in surveyed regions.

Table 2.Average hidden greenhouse gas and reactive nitrogen emissions from agricultural inputs of rice,wheat and maize production in China.

Table 3.Average carbon footprint and nitrogen footprint of rice,wheat and maize production based on farm survey in China.

Impact of farm size on carbon and nitrogen footprints

The CF and NF distributions were strongly correlated with farm size.In all the three crops,a decreasing trend of CF and NF was observed with the increase in crop farm size (Table 4).For rice,large-size household farms of Jiangxi had a significantly lower CF (by 20%) than small-size household farms.For wheat,large-size household farms in semiarid Jiangsu and Anhui was produced with lower CF (by 40%) compared to middle-size household farms.Meanwhile,no CF difference was observed between large-and middle-size household farms for maize.The mean NF was 20.7 g/kg for rice in Jiangxi for all the surveyed farms,which was produced with a significantly higher NF than those in Hunan by 48%.Similarity on NF from farm size was observed between wheat and rice.Greater NF of wheat was observed from small-size farms compared to large-size farms in Jiangsu and Anhui.In particular,Jiangxi for rice,Jiangsu for wheat,and Hebei for maize showed higher NF than the other provinces (Table 4).

Table 4.Variations of product carbon footprint and nitrogen footprint with farm size classes.

Relationships among carbon and nitrogen footprints

The NFs of the three staple foods were linearly correlated with CFs and the linear relationship between CF and NF was stronger in wheat (R2=0.6890) and maize (R2=0.5206) compared to rice (R2=0.4551) (Fig.1).It is speculated that all the surveyed farms that produced higher GHG emissions also had higher Nr discharges.The significant linear relationship between CF and NF of food production from all the surveyed farms can be attributed to the large contribution of N fertilizer use to both Nr and GHG releases (Fig.1).N fertilizer additions areknown to promote the releases of various Nr species,linearly or exponentially,and it is widely accepted that N fertilizer use is a substantial source of GHG emissions during the life cycle of cereal grain production.The synthetic N fertilizer inputs contributed more to the CF of wheat and maize rather than that of rice.

DISCUSSION

Carbon and nitrogen footprints from grain crop production

The CFs of surveyed farms for rice,wheat and maize ranged from 0.84 to 0.90,0.27 to 0.34,and 0.23 to 0.26 kg/kg,respectively.The corresponding CFs for wheat and maize grain production in China were similar to those in Canada [0.3 kg/kg for wheat (Gan et al,2012) and 0.3 kg/kg for maize (Jayasundara et al,2014)].The estimated CF for rice in China was lower than that in India (1.22 kg/kg),where rice production was comparatively low-yielding and high energy cost for irrigation (Pathak et al,2010),and was a little higher than that in Japan (0.8 kg/kg) (Breiling et al,1999).This observation may be due to generally larger levels of agricultural inputs in China.The CFs of rice production in China were previously determined in Guangdong,Hunan,Heilongjiang,Sichuan and Jiangsu provinces as 2.50,2.33,1.89,1.54 and 1.34 kg/kg,respectively (Xu et al,2013).The average CF was estimated at about 1.36 kg/kg for rice production based on National Statistical Dataset in China (Cheng et al,2015).Different CF values of crop could be primarily attributed to differences of the source and quality of data collection,system boundary,the emission factor of agricultural inputs and the calculation method among studies.For example,different provinces have different requirements for irrigation.Compared with the agricultural areas in northern China where water resources are scarce,the Yangtze River basin has a small demand for irrigation due to its natural superior climate resources,thus reducing CF caused by irrigation electricity.In addition,the superior geographical features and climatic conditions of the Yangtze River basin generally result in higher crop yields compared to other agricultural areas,resulting in a low CF per unit yield.Average GHG emissions from agricultural inputs were higher in rice than in wheat and maize.Seasonal difference of GHG emissions may be due to higher use of diesel oil,electricity,seed,fertilizer and film for rice production process,despite higher pesticide use in wheat and P2O5fertilizer use in maize (Table 2).Additionally,rice cultivation is a primary contributor to global CH4emissions (Stocker et al,2013),of which we considered CH4emissions from rice fields as a primary component of CF,similar to Cheng et al (2015).

The average regional NF of rice,wheat and maize among regions were 17.11,14.26 and 6.83 g/kg per year,respectively.All the values were higher than those in the Gulf of Mexico using the LCA method,which estimates NF of cereal production at about 2.65 g/kg (Xue and Landis,2010),but the values were lower than those in Austria estimated at about 21.9 g/kg based on the input-output analysis (Pierer et al,2014).These studies used different calculation methods and N management in the cereal production process which could also be a possible cause resulting in the difference of Nr loss.Additionally,this study found that NH3volatilization was the main NF source during grain crop production,which has been reported by Leip et al (2014).The NH3volatilization linearly increases with the N fertilizer application in the production of rice,wheat and maize (Wang et al,2012).Furthermore,higher levels of NH3volatilization during rice growing seasons have been attributed as a cause of higher NF in rice than in wheat and maize (Chen et al,2014).This trend could be attributed to higher moisture content and urea use during rice growing season that facilitates soil urease activity,resulting in higher soil NH4+-N concentration in rice fields (Wang et al,2012).

Farm size impacts on carbon and nitrogen footprints of crop production

This study showed the difference in CF and NF based on farm sizes.The CF and NF in large-size farms were significantly lower than those in small-size farms both for wheat and maize.Generally,this difference could be caused by a higher level of farmland management acumen of farmers with large-size farms allowing them to have more efficient application of agricultural materials,including optimal use of water and fertilizer.Huang (2016) has proposed that planting scale has a negative impact on the fertilizer application,and land transfer should be increased to promote the concentration of land to some farmers to reduce the fertilizer application per unit area.This agreed with the findings of Feng et al (2011) about the topsoil organic carbon storage being 30% higher in large-size farms (＞ 0.7 hm2) than in small-size ones (＜ 0.7 hm2).Using a questionnaire survey data,Sefeedpari et al (2013) found that farms in central Iran for rainfed wheat production having less than 1 hm2in size have a higher total energy input by 17%,21% and 34% compared to land sizes of 1-4,4-10,and ＞ 10 hm2,respectively.The CF and NF could be more sensitive to management practices in wheat than in maize.The role of the fragmentation of farmlands needs to be further addressed as present crop production is operated primarily under a household management system in China (Huang and Wang,2008).Increasing farm size with cooperative small-size householders or aggregating small-size farms into large-scale units can offer an opportunity to help mitigate GHG and Nr emissions from crop production in China,without additional technical inputs.Therefore,we strongly recommend the followings to promote low-carbon and green production of crops in China:(1) large-size operation of farms,(2) strengthening of land transfer policy formulation and management,(3) reduction of additional inputs in agricultural production,and (4) avoidance of excessive energy loss.

Mitigation scenarios and realization possibility

The implementation of N reduction,combined with the increasing grain yields and the reducing CH4emissions,could lessen CF and NF (Table 1).Reduction of CH4emissions in rice fields could be an efficient solution toward lowering the CF of rice.Adoption of appropriate farming practices could reduce CH4emissions from rice cultivation,such as optimizing tillage practice and improving water and fertilizer managements.Compared with continuous flooding in rice growing season,rational water resource management must be adopted to reduce CH4emissions,such as intermittent irrigation,intermittent irrigation-drainage in mid-season-frequent waterlogging,non-waterlogging-drainage in mid-season-intermittent irrigation (Hou et al,2012).To cut N inputs and enhance grain yields,there is a need to greatly improve the nitrogen partial factor productivity on a national scale (Zhang et al,2012).Chen et al (2014) found that,based on knowledge-based managements,such as the precision managements of nutrients,water and solar resources,the nitrogen partial factor productivity in Chinese main agro-ecological areas for rice,wheat and maize production could attain 54,41 and 56 kg/kg N,respectively,which are 3.6,2.9 and 2.5 times higher than the current values of 15.1,13.9 and 22.8 kg/kg N.In addition,the total amount of N fertilizers applied in rice should be split into four applications:basal fertilization,early tillering,panicle initiation and heading stages,which have been proven efficient in maintaining,or even enhancing rice yields,with 20%-30% N savings (Ma et al,2013;Zhao et al,2015).For wheat and maize,‘right N time and right N placement’ management,including the implementation of two top-dressings (with one topdressing applied at the later growth stage of wheat and maize),instead of the current one top-dressing practice,and popularizing fertilizer deep placement for maize top-dressing,have all been shown to greatly increase grain yields,even under reduced rates of N application (Zhang et al,2012;Chen et al,2014).

Main uncertainties of this study

We estimated the uncertainties of this study using a farmer survey to determine CF and NF in cereal grain production in select provinces in China.The various data sources and empirical models used presented some limitations in CF and NF analysis.Spatial heterogeneities exist in the climate,soil,cropping system and farming practices across China.We only estimated the CF and NF of cereal production in one season in 2016-2017.The limited samples in two years and in two seasons might cause some uncertainty.In this study,we adopted an LCA method to estimate CF and NF at the farm level based on the survey data.When scaled up to a simulation of provincial,regional and national levels,the spatial heterogeneity of the data variability may have introduced uncertainties into the results (Yu et al,2013).Then,we adopted the irrigation norms of crops as irrigation water inputs because the real water volume cannot be obtained from the field survey given to farmers.The irrigation norm might lead to an overestimate of the CF and NF of crop production in China.In addition,NH3volatilization loss rate under the same grain cropping system used the same loss rate in the NF calculation of farmers’ survey in each province,which may lead to some differences from the actual value due to the influence of soil properties,climatic conditions and farm management practices between regions (Sommer et al,2004).Moreover,a possible difference of the amount of N fertilizer use is observed in different studies (Wang et al,2012),which can lead to the difference of NH3volatilization.Despite the above limitations,trends in NH3contributions would likely not change for the NF of all the grain crops.Finally,GHG and Nr emissions are often affected by multiple factors of climate,soil,cropping system and farming practice (Yan et al,2015).These factors and their interactions make it difficult to generalize when considering suitable rules (Yu et al,2013),which must be refined based on an extensive monitoring network involving different agricultural regions and environmental conditions (Pandey and Agrawal,2014).The results in this study indicated the need for more spatial data and emission factors to improve the CF and NF analysis for crop production.Otherwise,the parameters from other developed countries applied to rice systems in China may lead to an underestimation of the total GHG and Nr emissions because of the lower energy use efficiency.To avoid this possibility,we used higher coefficients for agricultural inputs that were suitable for China.The conversion coefficients of CO2and Nr equivalents for most inputs were used from the Chinese Life Cycle Database (CLCD v0.7,IKE Environmental Technology CO.,Ltd,China).Energy from human labor was not considered in the study;however,there was considerable human labor involved in the current agricultural management conditions of rice production.As agricultural mechanization accelerates in China,more energy will be derived from diesel oil rather than human labor (Zhou et al,2020).

This study used a field national-scale systematic survey for agricultural management practices of farmers.Based on the survey data,CF and NF were quantified in rice,wheat and maize production in China.In contrast to global production,the greater contributions of NF mean that cereal production depends more on NH3volatilization in China.Meanwhile,applications of synthetic N fertilizer and CH4emissions were dominant CF sources.Furthermore,the significantly positive relationships between CF and NF indicate the potential for simultaneous mitigation in the regions with high agricultural inputs,like the amounts of fertilizers.Therefore,optimal use of synthetic fertilizers is necessary to reduce the NF of cereal production.In all the three crops,a decreasing trend of CF and NF was observed with the increase in crop farm size.An adjustment of planting zones for cereal production cropping is another possible measure to balance both CF and NF among regions.Resultsalso suggested a strong need to implement potential measures for climate change mitigation through lowering environmental pollution in China at the farm level by employing environmentally sound practices involving efficient fertilizer management,reducing soil tillage,and developing large-size farms.

METHODS

Study region

The sites selected represent the major crop production areas of China (Fig.S1).The sites for rice were in Jiangxi and Hunan provinces,southern China with a warm and humid climate.Generally,the sites practice double rice cropping systems,including the early and late rice cultivation.The local water management entailed normal irrigation of rice under intermittent flooding regime with midseason drainage.Winter wheat areas were in Jiangsu and Anhui provinces,while maize areas were in Jilin and Hebei provinces (Fig.S1).Maize is a typical system of grain crop production in areas with a humid climate,while summer maize and winter wheat in rotation were typical in Hebei with a semi-humid climate.

System boundary

Within the boundary,GHG and Nr emissions were accounted for at the provincial scale and were based on available data coupled with empirical models.The boundary of the life cycle inventory was assessed for the entire production chain of crops (Fig.2).GHG and Nr emissions included the followings:(1) electricity generation,gasoline and diesel emissions from mechanical activities such as tilling,seeding,irrigating harvesting and packing;(2) manufacturing,storage,and transportation of agricultural materials,including fertilizers [(N,phosphate (P) and potassium (K)],pesticides,seeds and films;(3) total CH4and N2O seasonal emissions from fields,as well as NH3volatilization,and nitrate (NO3-) and ammonium (NH4+) ions leaching during crop growing periods.Generally,most of crops were sold as a commodity by individual farmers after harvest.Therefore,the equivalent crop grain yields (kg/kg per year and g/kg per year) were defined as the functional units in this study.

Data sources

The farm management survey was a multiphase survey of major cereal crop farms in six provinces of China.Stratified random sampling was used in the study.The questionnaire consisted of four parts:(1) amounts of NPK fertilizers and pesticides used for each crop production;(2) farm mechanical operations like the methods of soil tillage and harvesting;(3) water management practices such as tube or well irrigation;and (4) farm area and grain yield of each crop.The questionnaire was pre-tested to 40 representatives at the Swan village,Ningxiang,Hunan Province,China to test the reasonability of the questionnaire.Finally,the questionnaire was revised based on comments and suggestions during the pretest,including removing financial-related questionnaire.

We randomly selected two townships in each county and two villages in each township for field surveys.Within each village,20 farm households were randomly selected,and we interviewed the household head (farmer) face to face.Household farms were divided into three size categories:small (＜ 0.7 hm2),medium (0.7-20.0 hm2),and large (＞ 20.0 hm2) based on the farm size data obtained from the survey and referenced on the land planning standards of the Ministry of Agriculture and Rural Affairs of the People’s Republic of China.Of the 600 interview questionnaires conducted among farmers from all six provinces,555 surveys were fully completed and analyzed.

CF calculation

We estimated the CF and NF of crop production primarily using the LCA method (Pandey and Agrawal,2014).Following the farm gate principle generally accepted for LCA in agriculture,the system boundary was set from seeding to harvesting of a cereal crop.The GHG emissions were estimated using the global warming potential (GWP) over a time span of 100 years (Ma et al,2013).Based on the life cycle inventory,for each crop in each province,CF (kg/kg) was calculated using the following equation:

whereCEtis the GHG emissions for 100 years of all the trace gases with radiative forcing (IPCC,2006),associated with the entire life cycle of grain production [kg/(kg·hm2)],Yis the grain yield (kg/hm2).

The CH4and N2O emissions were estimated according to IPCC (2006).The direct CH4emissions were estimated using the following equation:

where4CHCFis the annual CH4emission of crop cultivation (kg/hm2).EFi,j,kis a daily emission factor [kg/(hm2·d)].ti,j,kis the crop growing period (d).Letters i,j and k represent water regimes,type and amount of organic amendments,and additional conditions which CH4emissions may vary under different ecosystems.The value 25 is the relative molecular warming forcing of CH4in a 100-year horizon (IPCC,2006).

The direct N2O emissions from N fertilizer application were estimated with the following equation:

whereNis the quantity of N fertilizer applied in a single crop season.εis the default emission factor of N2O emission of applied N fertilizer.Emission factors of synthetic N fertilizer use in dry crops and submerged rice paddies were adopted respectively from Zou et al (2007) (Dry cropland,0.01 kg/kg fertilizer-N;Rice,0.0073 kg/kg fertilizer-N).The 44/28 is the molecular conversion factor of N2to N2O,and 298 is the GWP of N2O relative to CO2over a 100-year horizon.

NF calculation

In this study,NF was used as an indicator of the total direct N-losses to the environment that occur to produce one unit of (food) product,measured in food product (g/kg).The eutrophication potential was chosen to assess the impact associated with Nr emissions and losses during grain crop production.Additionally,the NF of grain crop produced was calculated according to ISO (2006).

whereNEtis the total Nr emission associated with the entire life cycle of grain crop production (g/hm2) that includes the Nr emissions during the process of production of kinds of agricultural inputs and the process of grain crop production in the field.

whereφis the NH3volatilization loss coefficient (0.338 for rice,0.275 for wheat,and 0.226 for maize) (Wang et al,2012).The σ is the NO3-leaching coefficients (0.305,0.606 and 0.175 for rice,wheat and maize,respectively).The γ is the NH4+leaching coefficients (0.339,0.190 and 0.043 for rice,wheat and maize,respectively).The 17/14,62/14 and 18/14 are the molecular weight ratios of NH3to NH3-N,NO3-to NO3--N,and NH4+to NH4+-N,respectively.The values 0.833,0.476,0.238 and 0.786 represent the eutrophication potential factors of NH3(kg/kg),N2O (kg/kg),NO3-(kg/kg),and NH4+(kg/kg),respectively.The applied eutrophication potential factors were sourced from the CML2002 method (Guinée et al,2002),and the multiplier 1 000 is a unit conversion factor (g/kg).

Statistical analysis

All statistical analyses were conducted using IBM SPSS Statistics,Windows Version 22 (IBM Corp.,Armonk,USA).One-way analysis of variance and the least significant difference test were used to check the differences between farm size classes and regions.

ACKNOWLEDGEMENTS

This study was supported by the Natural Science Foundation of Zhejiang Province,China (Grant No.LQ21C130002),the Engineering Science and Technology Development Strategy Consulting and Research Project of China (Grant No.Js2019-zd01),and the Central Public-Interest Scientific Institution Basal Research Fund of China (Grant No.CPSIBRF-CNRRI-202130).The authors thank Accdon,LCC for its linguistic assistance of this manuscript.

SUPPLEMENTAL DATA

The following materials are available in the online version of this article at http://www.sciencedirect.com/journal/rice-science;http://www.ricescience.org.

Fig.S1.Geographical distribution of sites surveyed in China.

Table S1.Emission factors of farm inputs and soils for rice,

wheat and maize production in China.