Balancing the economic, social and environmental dimensions of agro-ecosystems: An integrated modeling approach

Agriculture, Ecosystems and Environment 131 (2009) 263–273

Balancing the economic, social and environmental dimensionsof agro-ecosystems: An integrated modeling approach

Yongping Wei *, Brian Davidson, Deli Chen, Robert White

School of Resource Management, The University of Melbourne, Parkville 3010, Australia

A R T I C L E I N F O

Article history:

Received 25 August 2008

Received in revised form 27 January 2009

Accepted 30 January 2009

Available online 3 March 2009

Keywords:

Sustainability of agro-ecosystems

Integrated biophysical-economic modeling

Trade-off curve

Sustainability indicator

Policy assessment

Cropping system

A B S T R A C T

There is an increasing need to view agro-ecosystems and to identify remedial management practices in a

holistic way. An integrated model based on the driving force–pressure–state–impact–response

approach was developed as a tool to assess the effects of policies for improving decision making for

the sustainability of agro-ecosystems. An economic model was linked to a process-based biophysical

model by a meta-model. Then, a holistic indicator-based impact assessment system was linked to the

integrated model to assess policy instruments. The integrated model was applied in the intensive

irrigated wheat–maize cropping system of the North China Plain in which water and nitrogen fertilizer

management are known to be critical issues for sustainable resource management. The results show

there is a trade-off relationship between economic return and environmental outcome. It was shown

that water pricing is a more effective policy instrument for improving the sustainability of agro-

ecosystem than increasing the price of nitrogen fertilizer. When the water price is raised to 1.0 Yuan/m3

under a two-tariff system, the sustainability indicators for the irrigation water use efficiency was found

to increase from 0.37 to 0.77, groundwater use sustainability increased from 0.05 to 0.60, nitrate

leaching increased from 0.48 to 0.55 while the indicators for the farm gross margin, food self-sufficiency,

and soil nitrogen balance remain unchanged. The results suggest the modeling approach developed here

is very useful for evaluating policy options for complex natural resource management issues.

� 2009 Elsevier B.V. All rights reserved.

Contents lists available at ScienceDirect

Agriculture, Ecosystems and Environment

journal homepage: www.e lsev ier .com/ locate /agee

1. Introduction

Agriculture contributes 24% of global GDP and providesemployment to 1.3 billion people or 22% of the world’s population(Smith et al., 2007). It is a critical sector of the world economy.Meanwhile, agriculture is arguably the most important managedecosystem in the world. As the ways in which agro-ecosystems aremanaged and evaluated are heavily dependent on human values,the economic and social components of agro-ecosystems havebeen overemphasized in the past. This has caused malfunctioning(dis-services) of agro-ecosystems like land degradation, green-house gases emission, loss of bio-diversity, nitrate leaching towater bodies and depletion of groundwater (Conway, 1985; Daleand Polasky, 2007).

There is an increasing need to view agro-ecosystems and toidentify the remedial management practices in a holistic way(Pacini et al., 2004). Since the publication of the Brundtland report,the concept of sustainability has received increasing attention in

* Corresponding author at: Australia-China Centre on Water Resources Research/

Uniwater, The University of Melbourne, Parkville, Melbourne, Victoria 3010,

Australia. Tel.: +61 3 83449799; fax: +61 3 83446215.

E-mail address: [email protected] (Y. Wei).

0167-8809/$ – see front matter � 2009 Elsevier B.V. All rights reserved.

doi:10.1016/j.agee.2009.01.021

agricultural research. There would appear to be some consensusthat sustainability has three basic features: environmentalsoundness, economic viability and social acceptability (Dumanskiand Pieri, 2000). Pannell and Schilizzi (1999) argue that sustain-ability indicators are a practical and reasonable vehicle forattempting to deal with the multifaceted nature of the ambiguousterm ‘sustainability’. As understanding of the complex relationshipbetween agriculture and environment increases, many indicatorsof agricultural sustainability, environmental sustainability and theeffect of agriculture on natural resources and the environmenthave been developed (Wei et al., 2007c). However links betweensustainability indicators and agricultural management practiceson one hand, and economic policies on another hand, are not welldefined. As a consequence, farmers, policy makers and adminis-trators do not have enough information to alter managementsystems according to environmental needs (Ahuja, 2003).

Research of agro-ecosystems requires the use of models—thequestion is what kind? Models of agricultural systems have beendeveloped and have evolved since the 1960s. Prior to the mid-1980s most of the modeling work focused on individual processesof agricultural systems (e.g. Saeki, 1960; Monteith, 1965). Thensome multi-process models which describe the processes withinan agro-ecosystem appeared including RZWQM (Ahuja et al.,2000), EPIC (Williams, 1995) and DNDC (Li et al., 1992). When a

Y. Wei et al. / Agriculture, Ecosystems and Environment 131 (2009) 263–273264

multi-process model is found not to represent spatial hetero-geneity at a regional scale, some spatially referenced models likeAGNPS (Young et al., 1987), SWAT (Arnold et al., 1990) and WNMM(Li et al., 2007) have been developed. These models allow users toevaluate alternative practices and scenarios in large agro-ecosystems. However, they do not provide answers to thequestions of how farmers’ management behaviour could bechanged to introduce new management practices (Wei et al.,2005). To overcome this, a rapidly growing number of researchprojects are integrating economic and biophysical processes intomodels (Janssen and van Ittersum, 2007). Some of the morenotable integrated models are ECECMOD (Vatn et al., 1999),FASSET (Berntsen et al., 2003) and SAM (Belcher et al., 2004). Mostof these integrated models have focused on component parts of thesystem rather than the agro-ecosystem as an entire unit.Furthermore, many of the studies that advocate a systemsapproach lack a holistic interpretation of the sustainability ofthe agro-ecosystem. The environmental impacts that are actuallymodeled are often limited in number and aggregate in depiction(e.g. only total pesticide use and nitrogen losses are assessed). Theomission of many environmental aspects can lead to serious errorsin a multi-objective policy-making process and conflicts betweendifferent government programs or regulations.

Given this backdrop, the purpose of this paper is to provide anintegrated-modeling policy analysis tool for improving thesustainability of agro-ecosystems, in which a holistic impactassessment system is adopted. The intensively cropped ecosystemof the North China Plain is taken as the case study area.

2. Methods

2.1. The study area

The North China Plain (NCP) is in north-east China with an areaof 350,000 km2 (Fig. 1). It is the largest agricultural zone in Chinacontaining 34% of the nation’s population, 30% of the irrigated landand 40% of the total grain production. The NCP lies in a semi-arid tosemi-humid continental monsoon zone. The average cultivatedarea per person is 0.095 ha and the average water resource perperson is less than 500 m3. Irrigation is applied intensively and

Fig. 1. North C

extensively and agricultural water use accounts for approximately70% of total water consumption in this region. In addition to water,farmers in the region use fertilisers intensively in order tomaximize crop yields. However, there is some evidence thatfarmers over-fertilise and over-irrigate their crops in this area(Chen et al., 2005). A rotation with two-crops per year is practised.The winter wheat-spring maize rotation accounts for 85% of allcropping patterns in this region.

The long-term excessive use of water and fertilizer has hadnegative side-effects on the natural resource base and the physicalenvironment of the NCP. Groundwater resources are over-exploited. The average depletion rate of groundwater is 1–1.5 mper year over an area of 150,000 km2 (Xu et al., 2005). Morenoteworthy, 25% of all groundwater used is from deep aquifers,where the capacity for recharge is limited. Evans (2002) arguedthat few regions in the world had such vast areas wheregroundwater resources are over-exploited. The excessive use offertiliser and water for irrigation has led to leaching of nutrientsinto groundwater aquifers. High rates of fertiliser and inappropri-ate irrigation practices have also led to increased emission of thegreenhouse gas nitrous oxide (N2O) (Zheng et al., 2004).

Fengqiu County (348140–358530N) is located in the central-southern part of the NCP. It has a temperate, semi-humid,monsoonal climate with an average annual rainfall of 609 mm.The total land area of the county is 1224.6 km2 and supports apopulation of 700,000 people distributed in 25 towns with 606villages. Topographically, Fengqiu County is relatively homoge-neous. Four main soils exist in the study site. They are Typ–Light–Typ–Och–Aquic Cambisol, Typ–Ust–Sandic Entisol, Typ–MdmSit–Typ–Och–Aquic Cambisol and Typ–Cla–Typ–Och–Aquic Cambisol(Chinese Soil Taxonomy Research Group, 1995). The winterwheat–spring maize rotation accounts for 85% of the croppedarea in the county.. Fengqiu County has biophysical and socio-economic conditions similar to the wider NCP (Wei et al., 2007b).In addition, the Chinese Academy of Sciences has an agriculturalexperimental station there. Extensive experiments of land usepractices have been conducted since the 1960s includingfertilization, irrigation, soil salinity and nitrate contaminationprevention practices. The existing research results provide avaluable resource for testing and calibrating simulation models of

hina plain.

Fig. 2. Changes of crop yields with changes of water and nitrogen applied.

Y. Wei et al. / Agriculture, Ecosystems and Environment 131 (2009) 263–273 265

the agro-ecosystem. Therefore, Fengqiu County was taken as a casestudy area.

The average actual total yield of the wheat and maize rotationin Fengqiu County is about 120,000 kg/ha consisting of 6200 kg/ha of wheat and 5800 kg/ha of maize according to a farmersurvey (Wei et al., 2007a). The ranges of crop yields for a wheatand maize rotation in relation to changes in the amount of waterand nitrogen applied are shown in Fig. 2. It can be seen that waterplays a more important role in influencing crop yield thannitrogen.

2.2. The integrated-modelling framework

The conceptual framework was formulated under what can betermed the ‘driving forces–pressures–state–impacts–responsestructure of indicators, hereafter DPSIR (EEA, 2003). A diagram-matic representation of the DPSIR-based framework is presentedin Fig. 3. As the name suggests, there were five components to theframework: driving forces, pressures, states, impacts andresponses. Holding these together is a system-wide model.

Fig. 3. A conceptual framework for the evaluation of policy o

2.2.1. Policy scenario design

The basis of the model is information and data on biophysicalfactors, technological factors and policy interventions. Somebiophysical factors, such as climate change, are items that cannotbe controlled. Technological changes can be the result of scientificresearch and/or successful extension. While crucial, technology byitself is unlikely to result in an agro-ecosystem that is economicallyand environmentally sustainable. It needs innovative farmers who,in turn, are influenced by government policies. Thus, policychanges in this study are considered to be the main driving forcefor the sustainability of an agro-ecosystem.

As outlined before, the main issues of sustainability of the agro-ecosystem on the NCP are groundwater depletion, nitrate leachingand greenhouse gases emission. In the case of non-point sourcepollution (nitrate leaching and N2O emission), policy instrumentscan work by reducing the transport of the emission (water cycling)and/or by reducing the amount of potentially transportablepollutants (nitrogen cycling) (Morari et al., 2004). Because ofhistorical, cultural or institutional factors, and because of thenature of the policy instrument themselves, it is believed that

ptions for improving the agro-ecosystem sustainability.

Y. Wei et al. / Agriculture, Ecosystems and Environment 131 (2009) 263–273266

input-based tax instruments to influence fertiliser and irrigationwater prices become the best choice, as they can be easilymonitored and are a good and direct indicator of the pollution thatneeds to be regulated (Sterner, 2003). In the case of groundwaterover-exploitation, water pricing is, at least in theory, an effectivemethod of regulating scarce water resources (Wei et al., 2007a).

Scenario analysis is an essential element in anticipating thepossible impact of agro-environmental policies. To compare policyinstruments a set of scenarios was constructed as follows: (1) Basescenario: current agricultural policy (2003), 0.1 Yuan/m3 for waterand 3.3 Yuan/kg for nitrogen, that is used to replicate the effect ofcurrent irrigation and fertilisation practices on agriculturalsustainability. This represents the no regulation case. (2) Waterprice increases. The current water price is multiplied by a factorfrom 1.0 to 20. (3) Nitrogen input-based taxes are applied. Thecurrent nitrogen price is multiplied by a factor from 1.0 to 10.

2.2.2. Integrated biophysical-economic model

The kernel of the conceptual framework is an integratedeconomic-biophysical model. In the model farmers’ agriculturalpractices and the influences of these practices on social, economicand environmental conditions are simulated. WNMM is an existingprocess-based water and nitrogen management model which wasfirst developed and calibrated on data collected from FengqiuCounty to simulate intensive cropping production on the NCP (Liet al., 2007). The WNMM model is comprehensive and simulatesthe key processes of crop growth within the water and nitrogencycles on the surface and in the (subsurface) soils includingevapotranspiration, canopy interception, water movement,groundwater fluctuations, heat flow, solute transport, mineralisa-tion of fresh crop residue and soil organic nitrogen, formation ofsoil organic nitrogen, immobilisation in biomass, nitrification,ammonia volatilisation, denitrification, N2O emissions, as well asthe agricultural management practices (crop rotation, irrigation,fertilisation, harvest and tillage). Rather than putting resourcestowards building a new biophysical model the WNMM model waschosen as the biophysical model of water and nitrogen manage-ment in this study.

A central component of any models’ description of farmers’production behaviour is the characterization of how farmers makeland use and management decisions (Stoorvogel et al., 2004). Inthis study, decision making was simulated using a linearprogramming model, in which the objective function was tomaximize the expected farm gross margin.

So, the problem can be formulated as follows:

TGM ¼ MaxX 1

ð1þ rÞk

!

� ðPiYi jk � PwWi jk � PNNi jk � C jk � Tik þ SikÞA j (1)

Subject toX

A j � A (2)

where TGM is the total farm gross margin (Yuan), Pi is the price ofthe ith crop price (in Yuan/kg), Yijk is the yield for crop i under soil j

and in year k. As the wheat–maize rotation system accounts for85% of crop acreage in the study area, only wheat and maize wereconsidered. Wijk is irrigation amount for crop i under soil j and inyear k (in mm/ha), Pw is the price of water (Yuan/mm/ha). Nijk is theapplication rate of nitrogen for crop i under soil j and in year k (inkg N/ha), and PN is price of nitrogen fertiliser (in Yuan/kg).Furthermore, Cik is the fixed costs of crop i in year k (Yuan/ha). Cijk

includes seed costs, pesticide costs, labour costs (for sowing,ploughing, irrigating, applying fertiliser and harvesting), machinerent costs, fuel costs and other fertiliser costs. Tik is the existingagricultural taxes for crop i in year k (Yuan/ha) Sik is the subsidiesfor crop i in year k (Yuan/ha). A discount rate r is included and is

applied over k years. Aj is the area for j soil (ha), and A is the totalcultivation area in Fengqiu County (ha).

Meta-modelling is an analytical procedure that has beendeveloped in order to gain insights into the behaviour of complexsimulation models. It can be used to replace the original simulationmodel and placed in the estimated economic model of farmhousehold behaviour (Kruseman and Bade, 1998). In this study, ameta-modelling approach, employing regression techniques, wasused to link the economic and biophysical models. Crop yield wasconsidered to be the key link between the two models. A quadraticfunction form was chosen as the crop production function as itallows estimation of the effect of increasing input levels anddiminishing marginal returns, particularly when a wide range ofinputs need to be incorporated (Hexem and Heady, 1978; Overmanand Scholtz, 2002). The independent variables considered to beimportant in estimating the production functions of wheat andmaize were nitrogen fertiliser and irrigation applications. Inaddition, timing of management operations is also considered asindependent variables because it is an important characteristicdetermining to a large extent whether water and nutrient losseswill occur or not.

The regression equation adopted for wheat for each soil isdescribed as follows:

Ywheat ¼ b0 þ b1W þ b2W2 þ b3N þ b4N2 þ b5WN þ b6T1

þ b7T21 þ b8T2 þ b9T2

2 (3)

where Ywheat is wheat yield (kg/ha), W is the irrigation amount forthe whole wheat season (mm/ha), N is the nitrogen fertiliserapplication amount for the whole crop season (kg/ha), T1 is the firstirrigation time after sowing (Julian days), T2 is the second irrigationtime after sowing (Julian days), b0 is the intercept, and b1, b2, b3, b4,

b5, b6, b7, b8 and b9 are regression coefficients.The regression equation adopted for maize for each soil type is

described as follows:

Ymaize ¼ b0 þ b1W þ b2W2 þ b3N þ b4N2 þ b5WN þ b6T

þ b7T2 (4)

where Ymaize is crop yield (kg/ha), W is the irrigation amount for thewhole crop season (mm/ha), N is nitrogen fertiliser applicationamount for the whole crop season (kg/ha), T is the irrigation time(Julian days), b0 is the intercept, and b1, b2, b3, b4, b5, b6, b7 areregression coefficients.

2.2.3. Assessment of policy impact

In the evaluation module of the DPSIR framework impacts of thedriving forces are assessed so that the effect of various alternativestrategic measures can be compared. Based on it, the policy optionsfor improving the sustainability of the agro-ecosystem can bedeveloped. This is the final stage of the DPSIR, i.e. the responses tothe system pressures.

An indicator-based system was used to assess the impacts ofthe policy changes. In general the indicators should be selected toreflect the three basic features of the agricultural sustainability:environmental soundness, economic viability and social accept-ability (e.g. OECD, 1998). Specifically, the chosen indicators shouldcover the effects of irrigated cropping production on resources andenvironment caused by modifications in water and nitrogenagricultural practices. These indicators should reflect the mainenvironmental threats occurring on the NCP. They should bederived from integrated economic-biophysical modelling in orderto be linked with policy makers on one end, and the farmers onanother end. In this study, the indicators and their value rangeswere identified from published research on the NCP (Zhang, 1995;Zhang et al., 1996; OECD, 1998; Wang and Xin, 1998; CAS, 2000;

Table 1An indicator-based sustainability evaluation system.

Indicator Range of values Reference sources

Environmental soundness

Water use efficiency 0.8–2.0 (kg/m3) CAS (2000)

Nitrate leaching 35–70 (kg N/ha) Pacini et al. (2004), Zhen et al. (2005), and Sapek (2006)

Groundwater use sustainability 450–850 (mm/ha) Fengqiu Water Resources Bureau (2004)

Fertiliser use efficiency 10–30 (kg yield/kg N) Chen et al. (2005)

Soil nitrogen balance 0–50 (kg/ha) Wang and Xin (1998)

N2O emission 0.6–14.8 (kg N/ha) Scott et al. (2002)

Economic viability

Farm gross margin �4500–6030* (Yuan/ha) The Ministry of Chinese Agriculture (2005) and Fengqiu Agricultural Bureau (2004)

Social acceptance

Food self-sufficiency 250–370 (kg/per person) Ma (1999)

Note: Farmer labour cost was 4500 Yuan/ha/per year in Fengqiu County in 2003.

Y. Wei et al. / Agriculture, Ecosystems and Environment 131 (2009) 263–273 267

Zhang, 2000; Chen et al., 2005; Zhen et al., 2005) and fromdiscussions with local extension officers, scientists and the farmersthemselves. The selected indicators are site-specific and presentedin Table 1. Six indicators for environmental soundness were chosenin relation to the increasingly serious water resources shortage andenvironmental externalities occurring in the North China Plain.

Water use efficiency is a measure of crop productivity per unitof water used. It measures how efficiently crops acquire andtranspire available soil water. It serves as a benchmark forproductivity. Nitrate leaching is a measure of how the pollutionthat arises from various agricultural practices affects groundwaterquality. It is an indicator of water quality. The sustainable yield ofgroundwater is a water stress indicator which provides anindication of the effect of agricultural practices on water resourcerecycling. Nitrogen use efficiency is a common measure of nitrogenresource use efficiency and is the crop yield divided by the totalnitrogen applied. Nitrogen balance in soil is an indicator thatreflects soil quality related to water and fertiliser application rates.The emission of nitrous oxide (N2O) is an indicator that reflects thecontribution of over fertilisation to air quality, i.e. the greenhousegases effect. It should be noted that few researches and literatureare found on the emission of nitrous oxide (N2O) on the NCP. Itsvalue range is taken from the average level of farming systems inthe world (Scott et al., 2002).

Farm gross margin is one of the primary indicators ofagricultural sustainability as it reflects not only whether the farmenterprise stays in business, but also whether there is surplusincome to devote to resource conservation or development. Foodself-sufficiency is an indicator reflecting the social acceptance ofagricultural sustainability. This is very important from both agovernment and an individual farmer’s perspective in China.

When single indicators are aggregated into one sub-index, asimple linear relationship between the physical units and utilitieswas adopted for normalization in this study. This was donebecause precisely quantifying the relationship for each indicator

Fig. 4. Operation of the integrated modeling framework.

requires additional research beyond the scope of this study.Furthermore, other works have also adopted the linear relationship(e.g. Krajnc and Glavic, 2005).

When the impact of the indicator on the sustainability ispositive:

Xi ¼ðXi � XiminÞðXimax � XiminÞ

(5)

While the impact of the indicator on the sustainability is negative,it is given with

Xi ¼ 1� ðXi � XiminÞðXimax � XiminÞ

(6)

where Xi the indicator for evaluating the sustainability of croppingproduction, Ximax and Ximin is its threshold value. Ximax is assumedto be the value of the indicator when that indicator is in very goodcondition. If the Xi > Ximax, then Xi = Ximax. Similarly, Ximin is thevalue of the indicator when the indicator is in very poor condition.If Xi < Ximin, then Xi = Ximin.

Based on the theory of agro-ecosystem sustainability, eachcomponent – social, economic and environmental – are equallyimportant and in this study each indicator was given the sameweight.

2.2.4. Operation of the integrated modelling framework

As shown in Fig. 4, in the pre-running stage, the data forestimating the crop yield response functions (water and nitrogen)for each crop in each soil were generated by pre-running WNMMhundreds of times with a range of timing and amount of water andnitrogen fertiliser applications. Then the crop yield responsefunctions were estimated with the multi-variate non-linearregression technique in the meta-model. In the stage of modelsimulation, the economic model was driven by choosing differentpolicy scenarios (changing farmers’ decision), then the effect offarmers’ decisions on the sustainability of the agro-ecosystem wassimulated with the WNMM model, finally the impact assessmentsystem was used.

3. Results

3.1. Base scenario

The status of the wheat–maize agro-ecosystem on the NCPunder the biophysical, social and economic context in 2003 is givenin Fig. 5. It was found that there were high indices for food self-sufficiency, nutrient balance in soil and greenhouse gases emissionin Fengqiu County in the base scenario. The indicator value for farmgross margin was 0.73. There was the moderate sustainability innitrogen use efficiency, nitrate leaching and irrigation water use

Fig. 5. Initial status for sustainability indicators of agro-ecosystem.

Y. Wei et al. / Agriculture, Ecosystems and Environment 131 (2009) 263–273268

efficiency. Finally, the indicator value for groundwater sustain-ability was estimated to be 0.05, which means that the ground-water use was totally unsustainable.

3.2. Effects of changes of water and nitrogen prices

3.2.1. Effects of changes of water and nitrogen prices on individual

indicators

Very small changes in irrigation use efficiency were foundunless the water price was increased to more than 1.5 Yuan/m3,above which it increases rapidly (Fig. 6a). An increase in nitrogenprice was found to have a negative influence on irrigation wateruse efficiency but the effect was very small in comparison with thepositive effect of water price. It was estimated that for every1 Yuan/kg of increase in nitrogen prices, the irrigation water useefficiency only decreased 0.09 kg/m3 (Fig. 6b). As found later inSection 3.2.3, when the water pumped is less than 445 mm/ha, thegroundwater use will be sustainable. It was found that thegroundwater reached its sustainable yield when the water pricewas increased to 1.0 Yuan/m3 (Fig. 6c). From the integratedmodelling, the water applied is not affected by a change of nitrogenprice. So, any increase in nitrogen price will not influencegroundwater sustainability (Fig. 6d).

Water acts as a vehicle by which nitrogen is transmittedthrough the agro-ecosystem. Increasing the price of water had apositive impact by reducing the amount of nitrogen leaching(Fig. 6e). It was found that for every 0.1 Yuan/m3 increase in waterprice, the nitrate leaching decreased by 1.48 kg N/ha. Substantialincreases in nitrogen prices tend to have a positive impact byreducing nitrate leaching (Fig. 6f). The nitrogen price elasticity ofnitrate leaching was estimated to be very low. This can beinterpreted as a 1 Yuan/kg increase in nitrogen prices would lead toonly a 0.34 kg N/ha fall in nitrogen leaching through the soil.Increasing water price has a small negative impact on nitrogen useefficiency (Fig. 6g). It was found that for every 0.1 Yuan/m3

increase in water price, the nitrogen use efficiency fell by 0.47 kg/ha. So, the effect of water price on the nitrogen use efficiency isvery limited. As expected, nitrogen fertiliser use efficiency willimprove with increasing nitrogen prices. However, the elasticity isvery limited. For every 1 Yuan/kg increase of nitrogen price thenitrogen fertiliser use efficiency only improves by 0.13 kg/kg N(Fig. 6h).

It was found that there was about 150 kg N/ha surplus in thesoil in the base scenario, representing a substantial increase in soilN over the two crops grown over 1 year (Fig. 6i). Greatly increasingwater price will slightly increase the nitrogen surplus. In otherwords, because there is too much nitrogen surplus in the soil, anincrease in water price cannot influence the soil nitrogen balance.As expected, an increase of nitrogen price could significantlydecrease the accumulation of nitrogen in soil. However, even if thenitrogen price is increased by a factor of ten (to 33 Yuan/kg),surplus nitrogen would still accumulate in the soil (Fig. 6j). Thus,

increasing the nitrogen price within the ranges undertaken in thisstudy would not have a negative influence on soil nitrogen balance.

The effect of water price on N2O emission is presented in Fig. 6k.The amount of N2O emission is very small, ranging from 0.47 to0.98 kg N/ha when the water prices change. It should be noted thatthe relationship between water price and N2O emission is complexand it is difficult to determine the correct policy response in thiscase. N2O emission was found to be negatively affected by anincrease in nitrogen price. However, the elasticity is very low. Forevery 1 Yuan/kg increase of nitrogen price the N2O emissiondecreases 0.01 kg N/ha (Fig. 6l).

An increase in water prices was found to have a negative impacton the total farm gross margins (Fig. 6m). The relationship betweenwater prices and the total farm gross margins was found to be quiteelastic, at �3.02. This means that a 10% increase in water priceswould reduce the total farm gross margin by more than 30%. Whenthe water price is greater than 1.3 yuan/m3, the farm total grossmargin becomes negative. An increase in nitrogen prices will alsohave a negative impact on total farm gross margins. If the price ofnitrogen rises above 13.2 Yuan/kg the total farm gross marginbecomes negative (Fig. 6n). Over the positive range of the estimatesthe elasticity was found to be quite elastic at �1.64. This clearlyillustrates that the profitability of wheat–maize farming on theNCP is very sensitive to the prices of water and nitrogen fertiliser.An increase in water price was found not to influence the indicatorof food self-sufficiency until water price is increased to 1.4 Yuan/m3, a level at which farming would already be unprofitable(Fig. 6o). Over a range of nitrogen price change, the indicator offood self-sufficiency was found not to change (Fig. 6p).

3.2.2. Effects of changes of water and nitrogen prices on the social,

economic and environmental sub-indices

When water price increases from 0.1 Yuan/m3 to 1.6 Yuan/m3

the environmental indicator increases from 0.45 to 0.89. Theeconomic indicator decreases from 0.73 to 0.22 (Fig. 7a). The socialindicator decreases from 1.0 to 0.89. It should be noted that thesocial indicators remained unchanged at 1.0 when water priceincreases from 0.1 to 1.4 Yuan/m3. When nitrogen price increasesfrom 3.3 Yuan/kg to 16.5 Yuan/kg, the index of economic sustain-ability decreases from 0.73 to 0, while the index of environmentalsub-index only increases from 0.53 to 0.54 (Fig. 7b). The sub-indexof social sustainability remains unchanged.

3.2.3. Trade-off between economic losses and environmental benefits

that arise from increasing the prices of water and nitrogen

Increasing water prices have a positive impact on environ-mental indicators and a negative impact on the economicindicators. The trade-off between the economic sub-index andenvironmental sub-index is shown in Fig. 8a. This trade-off issimilar to that which economists refer to as a transformation curve.The rate of transformation between the environment andeconomic indicators was found to average (slope) of �0.89. Thatmeans when the economic sub-index increases 0.1, the environ-mental sub-index will decrease 0.089. However it should be notedthat the marginal rate of transformation varies along the curve butis never positive. A positive marginal rate of transformation wouldsignify a range of complementary points. However in generalcompetitiveness between environment and economic indicators ismost common and would be expected in most cases. It is moreusual to find that improvement in one indicator can only achievedby deterioration in the other indicator

The trade-off between indices of environmental and economicimpacts reveals a number of interesting observations when nitrogenprice changes (Fig. 8b). Only those points corresponding to thepositive estimates of the total gross farm margins are plotted. Therange of observations for the environmental index is relatively small,

Fig. 6. Effect of water and nitrogen price increases on irrigation water use efficiency (a, b), groundwater depletion (c, d), nitrate leaching (e, f), nitrogen fertiliser use efficiency

(g, h), soil nitrogen balance (i, j), N2O emission (k, l), total gross margins (m, n) and food self-sufficiency (o, p).

Y. Wei et al. / Agriculture, Ecosystems and Environment 131 (2009) 263–273 269

Fig. 6. (Continued ).

Y. Wei et al. / Agriculture, Ecosystems and Environment 131 (2009) 263–273270

ranging from 0.464 to 0.474 as the curve is very flat. The marginalrate of transformation was estimated to average out at only�0.006.This marginal rate of transformation ranged along the curve from�0.03 to 0.04. Finally, over part of the range observed, therelationship is complementary. Within this range it is possible toincrease both environmental and economic outcomes simulta-neously by raising the price of nitrogen, but to a very small extent.

It can be concluded from this analysis that increasing nitrogenprices are not in the interests of farmers and would appear to

Fig. 7. Effect of water and nitrogen price increases on the three sub-indices of agricu

deliver little benefit to society as a whole. Therefore, it is not reallyan acceptable policy option. Increasing water prices is the betterpolicy instrument.

3.3. Increasing prices of water

As discussed in Section 3.1, it appears that the best approach toimprove the sustainability of the agro-ecosystem in FengqiuCounty is to increase water prices. However, when water prices

ltural sustainability: (a) water price increases and (b) nitrogen price increases.

Fig. 8. Trade-off between economic sub-index and environmental sub-index when water and nitrogen price increases: (a) water price increases and (b) nitrogen price

increases.

Fig. 9. Change of each indicator after water price raises 1.0 Yuan/m3 in two-tariff

system.

Y. Wei et al. / Agriculture, Ecosystems and Environment 131 (2009) 263–273 271

increase the farm gross margin greatly decreases and this affectsthe sustainability of the whole agro-ecosystem (Fig. 6m). Apossible solution is to introduce a ‘two-tariff’ water price system.In this system, farmers would pay the current water price whenthey use less than 1/2 of the current water application rates. Abovethat application rate, farmers would have to pay a higher price. Thethreshold value of 1/2 of current water applications is determinedbecause that is indicated from both experimental and modelingresults in that farmers on the NCP overuse irrigation water byaround 1/2 (Zhen et al., 2005; Chen et al., 2005).

The change in each indicator when the water price is raised to1.0 Yuan/m3 in a two-tariff system is shown in Fig. 9. Thegroundwater sustainability indicator would be improved from 0.05to 0.60 and water use efficiency would be improved from 0.34 to0.77. Considered from the balance of the whole system, onlynitrogen use efficiency and nitrate leaching would be relativelylow (0.48 and 0.55).

4. Discussion and conclusions

Estimating the effects of a policy measure requires theidentification of the causal links between the implementation ofthe measure and its impact on human activities and theenvironment. The conceptual framework proposed in this studybased on the driving force–pressure–state–impact–responseapproach can be used to examine those links. An integratedbiophysical and economic model can capture the stochastic,interconnective, nonlinear interactions and spatial and temporaldifferentiation of ecological systems.

The advantages of farm level approaches adopted in this studyhave been extensively discussed by Pacini et al. (2004) andWossink et al. (1992). One major advantage of this approach is thedegree to which actual situations can be reflected in the decisions

on production taken at the farm level. A major drawback is that atthe farm level there is no mechanism to control the supply anddemand for inputs and outputs, which has consequences for inputand output prices. In addition, the policy recommendations fromfarm level approaches might to some extent contradict policiesfrom the national or regional contexts. For example, whilst therehas been a strong national imperative in China to increase foodproduction, the water pricing recommended in this study wouldmaintain and not greatly decrease the food security at farm level.What is the impact of the recommended policy on the foodsecurity from the national context? Further investigation isneeded here.

The model developed in this study was a static representation,i.e. using an annual time scale. Limiting the timeframe to shortperiods can potentially affect the evaluation of these responses, oreven prevent the introduction of some indicators into themodelling framework, such as soil organic matter. Continuingresearch into modelling irrigated agriculture should be conducted,especially on the long-term assessments of the response andadaptation to policy changes.

The traditional micro-economic model of farm productionconsiders profit maximization (or net farm income, farm grossmargin) as the primary objective that farmers are assumed topursue. This may happen when it is assumed that farmers are riskneutral and have perfect information about prices, field resourcesand technology availability. However, from survey-based results itwas found that farmers have a diverse set of objectives and riskprofiles, and farmer’s lack of knowledge in assessing water andnitrogen application rates (Sonntag and Norse, 2005; Zhen et al.,2005). Further research is needed to investigate the influences ofthese factors on farmers’ decisions to improve the presentmodelling effort.

A holistic indicator-based evaluation system was used in thisstudy, which enables a comprehensive evaluation of farmperformance. A set of environmental sustainability thresholdswas included in this evaluation system. These are important inextending the range of tools available for the evaluation ofsustainability. However such an approach is still in its infancy.Continuing efforts should be put into revising the indicators andtheir respective thresholds and determining the relationshipbetween the physical units and their utility value when normal-izing these into a single indicator. In addition, when the system isapplied at a large scale, spatially differentiated indicators areneeded.

We have shown that water pricing is the most effective policyoption for improving the sustainability of the wheat–maize agro-

Y. Wei et al. / Agriculture, Ecosystems and Environment 131 (2009) 263–273272

ecosystem of the NCP. However, this finding may be tempered byboth an institutional framework and social context. Price signal isonly one of the policy strategies. It is widely recognized that policystrategies for sustainable agriculture are best promoted byunderstanding of the perceptions, attitudes, behaviors andopinions of those involved in the process (Rahman, 2003; Weiet al., 2007a). A two-tariff water price system as proposed here maybe acceptable to farmers whilst achieving improved environmentaloutcomes. Further, it was found that farmer’s lack of knowledge inassessing water and nitrogen application rates is one of mainreasons farmers overuse them (Sonntag and Norse, 2005). Finally,from the point of view of policy instruments, all of them rely tosome extent on education and information provision (Cocklin et al.,2007). Therefore, it could be safely concluded that raising waterprice in combination with provision of training and extensionservices will improve the sustainability of agro-ecosystem on theNCP.

Acknowledgements

The authors are indebted to the Australian Centre of Interna-tional Agricultural Research (ACIAR) (project no: LWR/2003/039)and the Australia-China Special Fund (project no: CH06136) fortheir financial support.

References

Ahuja, L.R., 2003. Quantifying agricultural management effects on soil propertiesand processes. Geoderma 116, 1–2.

Ahuja, L.R., Rojas, K.W., Hanson, J.D., Shaffer, M.J., Ma, L. (Eds.), 2000. Root ZoneWater Quality Model: Modelling Management Effects on Water Quality andCrop Production, Water Resources Publications, LLC.

Arnold, J.G., Williams, J.R., Nicks, A.D., Sammons, N.B., 1990. SWRRB: a basin scalesimulation model for soil and water resources management. Texas A&M Uni-versity Press, College Station, TX.

Belcher, K.W., Boehm, M.M., Fulton, M.E., 2004. Agroecosystem sustainability: asystem simulation model approach. Agricultural Systems 79, 225–241.

Berntsen, J., Petersen, B.M., Jacobsen, B.H., Olsen, J.E., Hutchings, N.J., 2003. Eval-uating nitrogen taxation scenarios using the dynamic whole farm simulationmodel FASSET. Agricultural Systems 76 (3), 817–839.

Chen, D., White, R.E., Li, Y., Edis, R., Wei, Y.P., Zhang, J.B., Li, B.G., Zhang, Y.M., Cai,G.X., 2005. Conservation of water and nitrogen management in the North ChinaPlain—using a GIS based water and nitrogen management model (WNMM) andagricultural decision support tool (ADST). In: Willett, I.R., Gao, Z. (Eds.), Agri-cultural Water Management in China. ACIAR Proceedings, vol. 123, ACIAR,Canberra, pp. 153–160.

Chinese Academy of Sciences (CAS), 2000. Summary report of efficient utilisationand management of agricultural resources, a case in Ningjin County. ShangdongProvince, Beijing, PR China (in Chinese).

Chinese Research group of China Soil Taxonomy System, 1995. China Soil TaxonomySystem. China Agricultural Science and Technology Press, Beijing, 218 pp.

Cocklin, C., Mautner, N., Dibden, J., 2007. Public policy, private landholders: per-spectives on policy mechanisms for sustainable land management. Journal ofEnvironmental Management 85, 986–998.

Conway, G.R., 1985. Agroecosystem analysis. Agricultural Administration 20, 31–55.

Dale, V.H., Polasky, S., 2007. Measures of the effects of agricultural practices onecosystem services. Ecological Economics 64 (2), 286–296.

Dumanski, J., Pieri, C., 2000. Land quality indicators: research plan. Agriculture,Ecosystems and Environment 81, 93–102.

European Environment Agency (EEA), 2003. EEA Core Set of Indicators, Germany.Evans, R.S., 2002. A groundwater management strategy for the North China Plain. In:

Balancing the groundwater budget/International Groundwater Conference.Darwin, May 12, 2002 (CD).

Fengqiu County Agricultural Bureau, 2004. Annual report of agricultural economy inFengqiu County in 2003 (in Chinese).

Fengqiu County Water Resources Bureau, 2004. Annual report on water resource inFengqiu County in 2003 (in Chinese).

Hexem, R.W., Heady, E.O., 1978. Water production functions for irrigated agricul-ture. Iowa State University Press, Ames.

Janssen, S., van Ittersum, M., 2007. Assessing farm innovations and responses topolicies: a review of bio-economic farm models. Agricultural Systems 94, 622–636.

Krajnc, D., Glavic, P., 2005. A model for integrated assessment of sustainabledevelopment. Resources, Conservation and Recycling 43, 189–208.

Kruseman, G., Bade, J., 1998. Agrarian policies for sustainable land use: bio-eco-nomic modelling to assess the effectiveness of policy instruments. AgriculturalSystems 58, 465–481.

Li, C.S., Frolking, S., Frolking, T.A., 1992. A model of nitrous oxide evolution from soildriven by rainfall events: 1. Model structure and sensitivity. Journal of Geo-physical Research 97, 9759–9776.

Li, Y., White, R., Chen, D., Zhang, J.B., Li, B.G., Zhang, Y.M., Huang, Y.F., Edis, R., 2007. Aspatially referenced water and nitrogen management model (WNMM) forirrigated intensive cropping systems in the North China Plain. EcologicalModelling 203, 395–423.

Ma, X., 1999. China’s policies on self-sufficiency of grain and food security. Agri-culture in China and OECD countries: past policies and future challenges. OECD.

Ministry of Agriculture, 2005. Tabulation of production cost and return of agricul-tural products. Ministry of Agriculture, Beijing.

Monteith, J.L., 1965. Evaporation and the environment. Symposium of the Society ofExperimental Biology 19, 205–224.

Morari, F., Lutago, E., Borin, M., 2004. An integrated non-point source model-GISsystem for selecting criteria of best management practices in the Po Valley,North Italy. Agriculture, Ecosystems and Environment 102, 247–262.

OECD, 1998. Sustainable management of water in agriculture, issue and policies.The Athens Workshop, Paris.

Overman, A.R., Scholtz, R.V., 2002. Mathematical Models of Crop Growth and Yield.Marcel Dekker, New York, pp. 23–35.

Pacini, C., Wossink, A., Giesen, G., Vazzana, C., Huirne, R., 2004. Evaluation ofsustainability of organic, integrated and conventional farming systems: afarm and field-scale analysis. Agriculture, Ecosystems and Environment 92,273–288.

Pannell, D.J., Schilizzi, S., 1999. Sustainable agriculture: a question of ecology, ethics,economic efficiency or expedience? Journal of Sustainable Agriculture 13 (4),57–66.

Rahman, S., 2003. Environmental impacts of modern agricultural technology diffu-sion in Bangladesh: an analysis of farmers’ perception and their determinants.Journal of Environmental Management. 68, 183–191.

Saeki, T., 1960. Interrelationship between leaf amount, light distribution, and totalphotosynthesis in a plant community. Botany Magazine 73, 55–63.

Sapek, A., 2006. Agricultural activities as a source of nitrates in groundwater. In:Razowska-Jaworek, L., Sadurski, A.A. (Eds.), Nitrates in Groundwater. BalkemaPublishers, pp. 1–12.

Scott, M.J., Sands, R.D., Rosenberg, N.J., Cesar Izaurralde, R., 2002. Future N2O fromUS agriculture: projecting effects of changing land use, agricultural technologyand climate on N2O emission. Global Environmental Change 12, 105–115.

Smith, P., Martino, D., Cai, Z., Gwary, D., Janzen, H., Kumar, P., McCarl, B., Ogle, S.,O’Mara, F., Rice, C., Scholes, B., Sirotenko, O., Howden, M., McAllister, T., Pan, G.,Romanenkov, V., Schneider, U., Towprayoon, S., 2007. Policy and technologicalconstraints to implementation of greenhouse gas mitigation options in agri-culture. Agriculture, Ecosystem & Environment 118, 6–28.

Sonntag, B.H., Norse, D., 2005. Impediments to sustainable agriculture in China. In:Sonntag, B.H., Huang, J.K., Rozelle, S., Skeritt, J.H. (Eds.), China’s Agriculturaland Rural Development in the Early 21st Century.. Australian Centre ForInternational Agricultural Research, Canberra.

Sterner, T., 2003. Policy Instruments for Environmental and Natural ResourcesManagement Resources for the Future. The World Bank, Washington, pp.152–158.

Stoorvogel, J.J., Antle, J.M., Crissman, C.C., Bowen, W., 2004. The tradeoff analysismodel: integrated biophysical and economic modelling of agricultural produc-tion systems. Agricultural System 80, 43–66.

Vatn, A., Bakken, L., Botterweg, P., Romstad, E., 1999. ECECMOD: an interdisciplinarymodelling system for analysing nutrient and soil losses from agriculture.Ecological Economics 30, 189–205.

Wang, J.F., Xin, S.Z., 1998. Negative effects of fertilisation and countermeasures.Journal of Conservation of Agricultural Environment 17, 40–43 (in Chinese).

Wei, Y.P., Chen, D., Li, B.G., 2007a. Effectiveness of agricultural water pricing onsolving the water shortage of North China Plain. Journal of Resource Science 27,40–45.

Wei, Y.P., Davidson, B., Chen, D., White, R.E., Zhang, J.B., 2007b. Can contingentvaluation method be effective to measure the in situ value of groundwater inthe North China Plain? Water Resources Management 21 (10), 1735–1749.

Wei, Y.P., White, R.E., Chen, D., Davidson, B., Zhang, J.B., 2007c. Farmers’ perceptionof sustainability indicators for crop production on the North China Plain. Journalof Sustainable Agriculture 30 (2), 129–147.

Wei, Y.P., Chen, D., Davidson, B., White, R.E., 2005. Bio-economic strategy to combatnon-point pollution in China. Pedosphere 15 (2), 156–163.

Williams, J.R., 1995. The EPIC model. In: Singh, V.P. (Ed.), Computer Models ofWatershed Hydrology. Water Resources Publications, Highlands Ranch, CO,USA.

Wossink, G.A.A., de Koeijier, T.J., Renkema, J.A., 1992. Environmental-economicpolicy assessment: a farm economic approach. Agricultural Systems 39, 421–438.

Xu, Y.Q., Mo, X.G., Cai, Y.L., 2005. Analysis on groundwater table drawdown by landuse and the quest for sustainable water use in the Hebei Plain in China.Agricultural Water Management 75, 38–53.

Young, R.A., Onstad, C.A., Bosch, D.D., Anderson, W.P., 1987. AGNPS: a non-pointsource pollution model for evaluating agricultural watersheds. Journal of Soiland Water Conservation 44, 168–173.

Zhang, B.Z., 2000. Sustainable Agricultural Development in China. Shandong Pressof Sciences and Technology, China (in Chinese).

Zhang, W.L., 1995. Investigation of nitrate pollution in groundwater due to nitrogenfertilisation in agriculture in northern China. Journal of Plant Nutrition andFertiliser Science 1, 80–87 (in Chinese).

Y. Wei et al. / Agriculture, Ecosystems and Environment 131 (2009) 263–273 273

Zhang, W.L., Tian, Z.X., Zhang, N., Li, X.Q., 1996. Nitrate pollution of groundwaterin northern China. Agriculture, Ecosystem and Environment 59, 223–231.

Zhen, L., Routray, J.K., Zoebisch, M.A., Chen, G.B., Xie, G.D., Cheng, S.K., 2005. Threedimensions of sustainability of farming practices in the North China Plain: A

case study from Ningjing County of Shandong Province, PR China. Agriculture,Ecosystem and Environment. 105, 507–522.

Zheng, S.H., Han, Y.S., Wang, Y., 2004. N2O emissions from the agriculture in China.In: The Third International N Conference, Nanjing, October 12–16, http://www.issas.ac.cn/n2004.