Modelling to increase the eco-efficiency of a wheat–maize double

Modelling to increase the eco-ef ficiency of a wheat –maize double cropping systemZhigan Zhao a ,b ,Xin Qin a ,b ,Enli Wang b ,*,Peter Carberry c ,Yinghua Zhang a ,Shunli Zhou a ,Xiying Zhang d ,Chunsheng Hu d ,Zhimin Wang a ,**aDepartment of Agronomy and Biotechnology,China Agricultural University,Yuanmingyuan West Road 2,Beijing 100193,China bCSIRO Agriculture Flagship,GPO Box 1666Clunies Ross Street,Canberra,ACT 2601,Australia cCSIRO Agriculture Flagship,203Tor Street,Toowoomba,QLD 4350,Australia dThe Center for Agricultural Resources Research,IGDB,CAS,Hebei,ChinaA R T I C L E I N F OArticle history:Received 27January 2015Received in revised form 17May 2015Accepted 18May 2015Available online 26May 2015Keywords:Winter wheat Summer maize RotationNorth China Plain APSIMA B S T R A C TThe winter wheat-summer maize rotation is the main cropping system in the North China Plain (NCP).While its productivity has been constantly increased,the sustainability of the double cropping system has been increasingly questioned due to the high resource use and negative environmental impact.This paper combined field data with scenario modelling to investigate the productivity,resource (water and nitrogen)use ef ficiency and environmental impact of the wheat –maize rotation in response to management changes at a representative site Wuqiao in NCP.In addition the potential options to increase the system ’s eco-ef ficiency,i.e.,producing more with less input and less negative impact on the environment were explored.The results show that 180kg/ha/year of nitrogen could be considered as the optimal N rate for summer maize.For the wheat –maize double cropping rotation,225mm of irrigation water for wheat and 330kg N/ha (150for wheat and 180for maize)are required to maintain the potential grain yield at 18t/ha and to have minimum impact on environment in terms of N leaching to the groundwater and N loss due to denitri fication.These derived rates are 175mm/year less and 120kg N/ha less than the highest irrigation and N rates currently practiced.The reduced water and N inputs could reduce the annual totals of drainage,N leaching and denitri fication N loss to a minimum of 117mm,12kg N/ha and 28kg N/ha,respectively.Our simulation results also enabled derivation of yield response curves to water and nitrogen inputs,which can be used to optimise N applications under various levels of irrigation water inputs.ã2015Elsevier B.V.All rights reserved.1.IntroductionThe winter wheat-summer maize double cropping system in the North China Plain (NCP)has been considered as an example of successful intensi fication of agricultural systems (Wang et al.,2012).The system produces two crops per year with the growing season of wheat from October to June and of maize from June to early October.The total grain production in NCP accounts for >50%and >33%of the nation ’s wheat and maize productions (Liang et al.,2011),respectively.While its productivity has been constantly increased in the last four decades,the intensive use of water for irrigation and chemical fertilizers has led to serious negative impacts on the environment (Wang et al.,2008),leading tocessation of surface water flows,depletion of ground water resources,pollution to surface and ground water,and increase of greenhouse gas emissions (Ju et al.,2009).Therefore,the sustainability of the double cropping system has been increasingly questioned.Considerable experimental work has been conducted to investigate possible ways of increasing grain yield for both winter wheat and summer maize crops in the NCP.In order to maximise grain yield of winter wheat,irrigation needs to be applied and the amount of water needed for irrigation changes with the inter-annual variation of rainfall in the wheat season.One to three times of irrigation (each 75–80mm)was recommended in the middle to northern part of the NCP (Li et al.,2005;Wang et al.,2000;Wang et al.,2004b;Zhang et al.,2005;Zhang and Yu,2003).Pre-sowing irrigation of winter wheat is considered to be essential for achieving both high yields and high water use ef ficiency (WUE)(Lan and Zhou,1995;Wang et al.,2000;Wang et al.,2004b;Zhang et al.,2006).For the summer maize crop,since its growing period*Corresponding author.Tel.:+61262465964;fax:+61262465965.**Corresponding author.Tel.:+861062734011.E-mail addresses:Enli.Wang@csiro.au (E.Wang),Zhimin206@ (Z.Wang)./10.1016/j.agee.2015.05.0050167-8809/ã2015Elsevier B.V.All rights reserved.Agriculture,Ecosystems and Environment 210(2015)36–46Contents lists available at ScienceDirectAgriculture,Ecosystems and Environmentj o u r n a l h o m e p a g e :w w w.e l s e v i e r.c o m /l o c a t e /a g eecoincides with the summer rainy season,irrigation is not necessary in most years.To maintain the maximum yield level of both crops, the recommended nitrogen(N)application rate ranges from150to 240kg N/ha/year for wheat and90to360kg N/ha/year for maize, respectively(Du et al.,2009;Liang et al.,2011;Lüet al.,2011;Qin, 2012;Wang et al.,2009;Wang,2004b;Wu,2005;Zhang,2009; Zhao et al.,2006).While such recommendations aim to target the maximum crop productivity,high rates of water and N inputs often result in low resource use efficiency.Furthermore,local farmers often apply irrigation and N fertilizers at higher rates than recommended,exacerbating the water and nitrogen problems and the negative impact on the environment.Further experimental research in recent years has focused on understanding of the processes controlling crop water and nitrogen uses both temporally and spatially.Due to the contrast in climatic conditions and in the length of the growing seasons, along with hereditary characteristics,winter wheat can grow much deeper roots(up to200cm)than maize(up to120cm)(Wu, 2005;Wu et al.,2009;Zhou et al.,2008).The shallower maize root system together with the concentrated summer rainfall allows water and nitrogen to move below the maize root zone,leading to potentially significant water and nitrogen losses into the ground-water system.However,due to the coincidence of the end of maize growth and the end of the rainy season,the water and nitrogen in those deeper layers do not move further.The deep-rooted winter wheat crop is able to take up a large fraction of the water and nitrogen back before the next rainy season comes(Wu et al.,2009; Zhou et al.,2008).Such differences in use of water and nitrogen by the two crops across time and soil depth create an opportunity for better management of water and nitrogen applications for improved efficiency.However,this result was only based on experimental work in a couple of years,an understanding of the impact of the long-term inter-annual climate variability is still lacking.Such understanding is essential for development of management strategies to improve water and nitrogen use efficiency under the variable climate.Agricultural systems modelling has been proven to be an effective means to investigate the potential impacts of climate variability and management interventions on crop productivity, resource use efficiencies,and environmental impact of farming systems.In recent years,the farming systems model APSIM (Keating et al.,2003;Wang et al.,2002)has been intensively used as a tool to analyse the yield and resource use efficiency of the wheat-maize system in NCP.Several studies showed that once the model was properly calibrated,it was able to predict the biomass growth,grain yield,crop water and nitrogen uptake in response to water and nitrogen supply(Chen et al.,2010a,b,c),therefore it can be used to explore strategies for increasing water and nitrogen use efficiencies in NCP(Chen et al.,2010a;Fang et al.,2010).One modelling study that focused on possible ways to increase the eco-efficiency of the double cropping system has indicated real potential to increase production while reducing environmental risks(Carberry et al.,2013).The concept of eco-efficiency is to promote eco-friendly production systems,higher eco-efficiency means a system can produce more grains with less input and less impact on the environment.In spite of these previous studies, there is still a lack of systematic investigation on how application of water and nitrogen interacts to determine production and environmental outcomes in the wheat-maize double cropping systems in NCP,which is essential for understanding the impact of declining groundwater resources for irrigation,and its subsequent consequences.In this study,we combinefield measurement data,cropping systems modelling,and scenario analysis to investigate the productivity,resource(water and N)use efficiency and environ-they respond to management changes at the representative site of Wuqiao in NCP.Availability of irrigation water is considered as a main driver influencing systems performance,because it is the most limiting factor at the study site.Environmental impact is expressed as water and nitrogen losses that could otherwise contribute to the productive use of resources.The aim is to explore potential options to increase the eco-efficiency of the wheat-maize system in the future,i.e.producing more grain with less water and N inputs and less negative impact on the environment.2.Materials and methods2.1.Study site,climate and experimental dataThe study was carried out at Wuqiao(WQ)site(37 410N, 116 370E,altitude20m above sea level,groundwater Table6–9m) in the middle of Heilonggang Catchment in Hebei Province,China. The study site was characterised by a summer monsoon climate, with average annual temperature of12.9 C and annual rainfall of 550mm(1961–2010).About64%of the annual rainfall is received in the summer months from July to September.For the wheat–maize double cropping system,the growing season of wheat is from mid-October to early June,and of maize from mid-June to early October.The soil at the site is classified as a Calcaric Fluvisol (FAO,1990)with a sandy clay loam texture and a deep soil profile down to at least200cm.The topsoil(0–20cm)had a pH of7.8and contained about11.7g kgÀ1organic matter,0.78g kgÀ1total N, 0.02g kgÀ1available P,and0.30g kgÀ1available K.Detailed soil profile characterisation can be found in Zhao et al.(2014) Data used for further model verification(for wheat)was from field experiments(Table1)conducted at the Wuqiao site in 2009–2010wheat growing seasons(Exp1)and2009–2010wheat-maize rotation seasons(Exp2).Exp1covered the impact offive N application rates on wheat growth under sufficient water supply, while Exp2covered impact of four irrigation levels under enough N supply.All the experiments were conducted with randomized complete block design with irrigation water supply ranging from 75mm to375mm per season and fertilizer-N(urea-N)application rates ranging from0to330kg N/ha,each with three replicates.In Exp1,the wheat cultivar‘Shijiazhuang150was sown on12October with plant densities570plants mÀ2.In Exp2,the wheat cultivar ‘Shijiazhuang80was sown on12October with plant densities 600plants mÀ2.Weeds,insect pests and diseases were properly controlled and the crops were not limited by other nutrients.Crop samples were collected5–7times from0.2m2quadrates to measure leaf area index(LAI),above-ground biomass,grain yield and yield I was measured using LI-COR3000c.All plant samples were oven dried at70 C to constant weight to measure biomass.At maturity stage,plants from2m2were harvested for the determination of grain yield.Daily climate data from1961–2012including daily maximum and minimum temperature,rainfall,and sunshine hours were obtained from the weather station of Wuqiao County.Daily solarTable1Water and nitrogen treatments in experiments conducted at WuqiaoExperiment Irrigationtreatments aN application b(kg N/ha)Wheat cultivarExp1W30,123,192,261,330Shijiazhuang15 Exp2W1,W2,W3,W5158Shijiazhuang8a W1–W3represents one to three times of irrigation,each of75mm applied at sowing,jointing andflowering,respectively.W5–75mm applied at sowing, upstanding,jointing,booting and mid grainfilling.b All the N was applied at sowing if N rate was less than123kg N/ha.Otherwise,Z.Zhao et al./Agriculture,Ecosystems and Environment210(2015)36–4637radiation was calculated with daily sunshine hours using the Angstrom equations(Angstrom,1924).2.2.APSIM model parameterization and validationThe APSIM model(Keating et al.,2003;Wang et al.,2002) version7.5was used to simulate above-ground biomass,grain yield,and water and nitrogen use efficiency of wheat and maize crops at the study site.In APSIM,phenological development of wheat and maize crop from emergence towards maturity is driven by accumulation of thermal time,with the rate of accumulation modified by vernalisation and photoperiod for wheat and photoperiod for maize beforefloral initiation.Growth of above-ground biomass is simulated using stage-dependent radiation use efficiency(RUE)together with the intercepted radiation.RUE is further modified by suboptimal temperatures and stresses of water (Wang et al.,2004a)and nitrogen if the water and/or nitrogen supply is not sufficient to meet the crop demand.APSIM simulates the dynamics of water and nitrogen in soil on a daily time-step.Soil water dynamics is modelled using a simple tipping bucket model,with the key parameters of saturation water content(SAT),drained upper limit(DUL),and crop lower limit(LL) for each soil layer.Water above DULflows quickly to the layer below,and water passing the crop root zone is treated as deep drainage.The APSIM-SoilN simulates both carbon and nitrogen dynamics in soil,including decomposition of soil organic matter,N mineralization and immobilisation,nitrification,denitrification and N leaching(Probert et al.,1998).N leaching loss is calculated based on the N concentration in soil solution in the deepest layer and the deep drainage rate.Denitrification N loss is a function of amount of nitrate,soil temperature,and soil water in each soil layer.For general performance of APSIM,we rely on the model validation of Chen et al.(2010a);Wang et al.(2012);Zhang et al. (2012);Liu et al.(2012)where they showed that the APSIM model was able to predict the biomass growth,grain yield,crop water and nitrogen uptake of wheat and maize in response to water and nitrogen supply in NCP and northeast China.Chen et al.(2010a) found that with the original model parameterisation the APSIM-Wheat module significantly underestimated growth of wheat(in terms of LAI and biomass)during early spring time while the APSIM-Maize module underestimated both maize biomass and grain yield under the NCP conditions.Modifications to the temperature response of thermal time calculation and RUE in APSIM-Wheat based on Wang and Engel(1998);Porter and Gawith (1999)led to improved simulations of the dynamics of LAI, biomass,and grain yield of wheat crop.Increasing the RUE for maize from1.6g/MJ in the original model to1.8g/MJ according to Bastiaanssen and Ali(2003)and Tao et al.(2005)also improved the performance of APSIM-Maize(Chen et al.,2010a).These mod-ifications were adopted here.Due to the importance of accurate modelling of wheat response to both water supply(irrigation)and N application in the current study,we further tested the performance of the modified APSIM-Wheat(Fig.1)to simulate wheat growth and grain yield as affected by the water and N treatments in the experiments listed in Table1.Cultivar parameters for the two wheat cultivars and root growth parameters for wheat crop in NCP were derived in Zhao et al.(2014).These parameters were also adopted here.In all simulations,maximum root depth of winter wheat and summer maize were set to2m and1.2m respectively,according to the findings of Zhou et al.(2008)and Wu et al.(2009).For the long-term simulations described below,the maize cultivar‘981’was used with cultivar parameters from Chen et al.(2010a).All the modifications to model parameters used in this study are listed in Table2.2.3.Modelling the system performance in response to irrigation and nitrogen applicationsWith the calibrated APSIM model,the performance of the wheat–maize double cropping rotation was simulated under long-term climate variation from1961to2012.This was done in two steps as described below:Step1:Derivation of the optimal nitrogen application rate for summer maize:Because the maize crop does not require irrigation in most of the years due to the coincidence of its growing season with summer monsoon rainfall,a continuous maize system was simulated with representative sowing date(18June)from1961to 2012and with nitrogen application rates ranging from0to 300kg/ha(30kg/ha interval)and no irrigation.The results were used to derive the yield response curve to N application rate and the optimal N rates for maize.On maize sowing date each year,soil nitrogen was reset to be70kg/ha,according to the average measured soil nitrogen content in0–2m soil profile before maize sowing,to avoid the impact of nitrogen accumulation in the soil. The derived optimal N rate for maize will be used as afixed N rate for maize in the subsequent modelling of the wheat–maize double cropping system,because it represents the suitable N rate for maize to maximise the crop water use and yield during the maize growing season.Step2:Modelling the system responses to supply of water and nitrogen to wheat crop:Water is the most limiting factor for the double cropping systems at the study site and most part of the northern NCP due to the high water demand of the wheat crop before the rainy season.Performance of the whole system heavily depends on the availability of irrigation water to the wheat crop, which in turn determines the nitrogen managementstrategies. 38Z.Zhao et al./Agriculture,Ecosystems and Environment210(2015)36–46Therefore,the performance of the wheat–maize double cropping rotation was modelled with a combination of different levels of irrigation and nitrogen application for wheat crop.In the simulations,wheat was sown on12October withfive irrigation scenarios:rainfed(W0),W1(75mm water at sowing time),W2 (W1+75mm water at jointing stage),W3(W2+75mm water at flowering stage)and W4(W3+75mm water at start of grainfilling stage).Among these irrigation treatments,W1,W2and W3were recommended in NCP(Li et al.,2005).For each irrigation scenario, the impact of11nitrogen rates ranging from0to300kg N/ha (30kg/ha interval)were modelled.In the rotation,maize was planted after wheat harvest and180kg N/ha was added each year (which was the derived optimal N rate from the previous modelling step).No irrigation was applied for maize in most years,except that50mm of water was applied in a few years when the surface soil layer(0–40cm)was too dry for maize emergence.Grain yield,water and nitrogen uptake,N loss due to denitrification,drainage and nitrogen leaching below the crop root zone were simulated to analyse the systems productivity and environmental performance in response to irrigation and nitrogen inputs.The results from step2were used to derive the optimal N rate for wheat and the corresponding wheat yield under eachTable2Modifications in model parameters used in the simulation.Crop Parameters APSIM7.5APSIM Modified wheat Rate of rooting advance(mm/d)a 5.0,30,3010.0,50,50Root/shoot ratios b Root branching rate 1.0,0.3,0.3,0.08,00.02,0.1,1.01.0,0.17,0.10,0.05,00.01,0.2,1.0maize RUE(g/MJ) 1.60 1.80a Values for key stages of germination,emergence and after emergence until end of juvenile.b Values for key stages of after emergence,floral initiation,flag leaf,flowering and start grainfilling.Z.Zhao et al./Agriculture,Ecosystems and Environment210(2015)36–4639irrigation level.The optimal N rate for wheat and maize together, under each irrigation level for wheat,was calculated as the optimal N rate for wheat plus180kg N/ha for maize(which was the derived optimal N rate from step1).The total yield of wheat and maize,N loss by leaching and denitrification,and drainage water loss under each irrigation level of winter wheat were then calculated based on the simulation results.Water use efficiency(WUE),partial fertiliser productivity(PFP) and N recovery were calculated as follows:PFP(kg/kg N)=grain yield/nitrogen application rate.N recovery=nitrogen in biomass/nitrogen application rate.WUE for rainfall and irrigation=grain yield/(rain+irrigation).WUE for irrigation water(I)=grain yield/irrigation.Fraction of reused soil water=water uptake from120to200cm soil/drainage below120cm.Fraction of reused soil nitrogen=Nitrogen uptake from120to 200cm soil/leached N below120cm.3.Results3.1.APSIM performance to simulate wheat response to water and nitrogenThe performance of the APSIM model is shown in Fig.2.In general,simulated above-ground biomass using both the original and modified parameterisation of the APSIM-Wheat followed the measured pattern well at Wuqiao site.At wheat harvestingtime Fig.3.Simulated performance of continuous maize system in response to nitrogen input rates at Wuqiao study site(1963–2012).The box plots show the0,10,25,50,75,90, 40Z.Zhao et al./Agriculture,Ecosystems and Environment210(2015)36–46(early June),both versions of the model captured thefinal biomass reasonably.However,the original parameterisation led to overes-timation of aboveground biomass during the early stages of winter wheat,especially before early May.At jointing stage(early April), the simulated above-ground biomass with the original parameter-isation was about two times higher than measured data.The new parameterisation of the model improved the simulation of biomass dynamics during the early stages of winter wheat(Fig.2a–d)and thefinal biomass across different treatments(Fig.2e),though the accuracy for simulating grain yield remained similar(Fig.2f).More accurate simulation of biomass dynamics under various conditions will likely be achieved through improving simulations of crop water and nitrogen uptake,soil water and N dynamics,and their feedbacks to crop growth.Therefore,the model with the new parameterisation was used for subsequent modelling.3.2.Optimal nitrogen application rate for summer maizeFig.3shows the simulated response of different aspects of the continuous maize system at the study site.Maize biomass,grain yield,water and N uptake,and water use efficiency(WUE) increased with N application rates,with WUE reaching a plateau at N rate of180kg N/ha/year,and very little increase in biomass and yield at higher N rates.At N rate of180kg,the average maize biomass and grain yield were14t/ha and9.8t/ha,respectively (Fig.3a and c),with the corresponding water and N uptake of 355mm and175kg N/ha(Fig.3g and b).Increasing N rates alsoled Fig.4.Simulated responses of wheat biomass,yield,water and nitrogen use to irrigation water and nitrogen inputs(during the wheat season)when planted in rotation withZ.Zhao et al./Agriculture,Ecosystems and Environment210(2015)36–4641to increased inter-annual variability of the simulated maize yield, water and nitrogen uptake,implying that inter-annual climate variability had the maximum influence on maize yield under sufficient N supply,this needs to be taken into account in optimising management practices.This is understandable because N application had the biggest impact on yield in years with sufficient rainfall,while the least impact when rainfall is limited.The grain yield increase per unit increase in N application decreased almost linearly with increased N rates(Fig.3d),leading to decreasing NUE(data not shown)and N recovery(Fig.3f).When N rate was higher than180kg N/ha,grain yield increase per unit increase in N application was reduced from around18kg/kg N to a very small value1.2kg/kg N close to zero(Fig.3d),indicating no further gain from increased N application.Higher N application rates will also lead to increased N leaching(Fig.3j)and denitrification N losses(Fig.3h).The average WUE of ET increased from8to23kg/mm with nitrogen input increased from0to180kg/ha,then maintained stable value of23kg/mm when nitrogen input was more than 180kg/ha(Fig.3e).Nitrogen recovery was2.0at nitrogen input level of30kg/ha,and decreased when nitrogen input increased (Fig.3f).The higher N recovery(>1.0)was due to the uptake of N that was already in soil.The average drainage below the root zone of maize decreased from93to58mm with nitrogen rate increase from0to180kg/ha, then maintained nearly stable at56mm/year(Fig.3i).Moreover, the N leaching and denitrification N loss increased with increase of nitrogen input,especially when it>120kg/ha(Fig.3h and i).BasedFig.5.Simulated responses of biomass,yield,water and nitrogen losses of the wheat-maize double cropping system to irrigation water and nitrogen inputs(during the wheat 42Z.Zhao et al./Agriculture,Ecosystems and Environment210(2015)36–46on these results,180kg/ha/a of nitrogen could be considered as the optimal N rate for summer maize at the study site.3.3.Response of wheat to irrigation and nitrogen applicationsFig.4shows the simulated response of wheat to levels of irrigation and N applications when it was planted in rotation with maize and at an N rate of180kg N/ha for maize.The results clearly show that both the wheat yield and the demand for nitrogen increased with the amount of irrigation water supply.The maximum wheat biomass and grain yield were simulated to be around16and9.5t/ha respectively under the supply of300mm irrigation water and270kg N/ha for the wheat season(Fig.4a and c).However,90%of the maximum biomass and grain yield were achieved at225mm of irrigation water supply with150kg N/ha (Fig.4a and c),beyond which yield increase per unit increase in either water or nitrogen dropped markedly(Fig.4g and h).WUE of irrigation water and NUE decreased with the amount of irrigation water and nitrogen input(Fig.4e and f).The maximum WUE of irrigation decreased from65to27kg/mm when irrigation water amount increased from75to300mm(Fig.4e),while the NUE decreased from112to27kg/kg N when nitrogen input increased from30to300kg/ha(Fig.4f).3.4.The responses of wheat-maize double cropping system to water and nitrogen inputsFig.5shows the simulated responses of the wheat-maize double cropping rotation to different levels of N applications and irrigation water supplies for wheat crop.Total biomass and grain yield increased with irrigation water amount and N application rates,and all reached a plateau at N rate of180kg N/ha/year for wheat.The average maximum total biomass and grain yield(wheat plus maize)were28.5t/ha and18t/ha respectively(Fig.5a and b) under the supply of300mm irrigation water and270kg N/ha for the wheat season,with the corresponding water and N uptake of 740mm/year and425kg N/ha(Fig.5c and d).However,96%of the maximum biomass and grain yield were achieved at300mm of irrigation water supply with180kg N/ha for the wheat season (Fig.5a and b).Increasing N rates reduced the drainage(Fig.5e),and also led to increased N loss from N leaching and denitrification,especially when it>150kg/ha(Fig.5f and h).Increasing irrigation water amount led to increased drainage(Fig.5e),but reduced the runoff (Fig.5g)and the N loss from N leaching and denitrification(Fig.5f, h),due to increased N uptake by crops with irrigation water amount(Fig.5d).3.5.Generalized responses of systems productivity and environmental performanceThe generalised response of crop yield to irrigation and nitrogen inputs(Fig.6)can be used to estimate N requirements and total productivity under various levels of irrigation.For example,if there is225mm of water available for irrigation to fully meet the wheat crop water demand,the optimal N rate for wheat is150kg/ha (Fig.6a).The corresponding wheat yield is8t/ha(Fig.6b)and the total grain yield of wheat plus maize is17t/ha(Fig.6c).If nitrogen can be managed at the optimal level,the total crop yield of wheat and maize can also be directly predicted from the irrigation input to wheat(Fig.6d).For winter wheat crop,the amount of N fertiliser required to achieve a certain yield level increased linearly with irrigation water supply(Fig.6a).Wheat grain yield increased with irrigation and N input linearly until it reached a yield level of 8.2t/ha at irrigation of225mm and N rate of150kg N/ha(Fig.6b).in grain yield.The total yield of both wheat and maize increased linearly with wheat yield,could reach a maximum of17.5t/ha (Fig.6c).Under the current situation in NCP,nitrogen supply is not limited.The total productivity of the double cropping systems is almost linearly proportional to the irrigation water input(Fig.6d).Similarly,the environmental impact(water and nitrogen losses to the environment)could also be evaluated(Fig.7).For example, under225mm of irrigation,the optimal N rate for wheat and maize together is330kg/ha(Fig.7a).The corresponding N loss due to denitrification and leaching is40kg/ha(Fig.7b).The water loss through drainage is117mm(Fig.7c).From the total N rate of240kg N/ha(60kg N/ha for wheat,180kg N/ha for maize),N loss due to both leaching and denitrification increased slowly with N rates, reaching a total N loss of40kg N/ha(12kg N/ha leaching and28kg N/ha denitrification)at N input of150+180kg N/ha(Fig.7b). Thereafter,any additional N input will lead to minimum increase in productivity(Fig.6),but much more significant increase in total N loss to the environment(Fig.7b).Total drainage increased linearly with irrigation amount(Fig.7d),corresponded closely to N lossby Fig. 6.Summarised response of the productivity of the wheat-maize double cropping system to irrigation and nitrogen inputs at Wuqiao in North China Plain, generated based on the average of50years’simulated results from1963to2012.Fig.7.Summarised environmental impact of the wheat-maize double cropping systems to irrigation and nitrogen inputs at Wuqiao in North China Plain.GeneratedZ.Zhao et al./Agriculture,Ecosystems and Environment210(2015)36–4643。