Ecosystem health pattern analysis of urban clusters based on emergy synthesis: Results and implication for management

Energy Policy 59 (2013) 600–613

Contents lists available at SciVerse ScienceDirect

Energy Policy

0301-42http://d

n CorrE-m

journal homepage: www.elsevier.com/locate/enpol

Ecosystem health pattern analysis of urban clusters based on emergysynthesis: Results and implication for management

Meirong Su a,n, Brian D. Fath b,c, Zhifeng Yang a, Bin Chen a, Gengyuan Liu a

a State Key Joint Laboratory of Environment Simulation and Pollution Control, School of Environment, Beijing Normal University, Beijing 100875, Chinab Biology Department, Towson University, Towson, MD 21252, USAc Advanced Systems Analysis Program, International Institute for Applied Systems Analysis, A-2361 Laxenburg, Austria

H I G H L I G H T S

� The use of integrated emergy synthesis and set pair analysis model was standardized.
� The integrated model was applied on the scale of an urban cluster.� Health patterns of different urban clusters were compared.� Policy suggestions were provided based on the health pattern analysis.
a r t i c l e i n f o

Article history:Received 17 December 2011Accepted 4 April 2013Available online 29 April 2013

Keywords:Ecosystem health patternChinese urban clusterEmergy

15/$ - see front matter & 2013 Elsevier Ltd. Ax.doi.org/10.1016/j.enpol.2013.04.015

esponding author. Tel.: +86 10 58801763; fax:ail address: [email protected] (M. Su).

a b s t r a c t

The evaluation of ecosystem health in urban clusters will help establish effective management thatpromotes sustainable regional development. To standardize the application of emergy synthesis and setpair analysis (EM–SPA) in ecosystem health assessment, a procedure for using EM–SPA models wasestablished in this paper by combining the ability of emergy synthesis to reflect health status from abiophysical perspective with the ability of set pair analysis to describe extensive relationships amongdifferent variables. Based on the EM–SPA model, the relative health levels of selected urban clusters andtheir related ecosystem health patterns were characterized. The health states of three typical Chineseurban clusters – Jing-Jin-Tang, Yangtze River Delta, and Pearl River Delta – were investigated using themodel. The results showed that the health status of the Pearl River Delta was relatively good; the healthfor the Yangtze River Delta was poor. As for the specific health characteristics, the Pearl River Delta andYangtze River Delta urban clusters were relatively strong in Vigor, Resilience, and Urban ecosystemservice function maintenance, while the Jing-Jin-Tang was relatively strong in organizational structureand environmental impact. Guidelines for managing these different urban clusters were put forwardbased on the analysis of the results of this study.

& 2013 Elsevier Ltd. All rights reserved.

1. Introduction

Cities play an important role in regional, national, and eveninternational development (Huang, 1998). Clearly, the health ofurban ecosystems can influence economic prosperity and socialprogress at regional and larger scales; therefore, urban ecosystemhealth should be a priority for both researchers and managers.However, a holistic understanding and diagnosis of urban ecosys-tem health has not been emphasized and the emergent andaccumulated environmental impacts in recent years now jeopar-dize economic development and the continued provisioning ofecosystem services within the urban support area. Our goal is to

ll rights reserved.

+86 10 58800397.

apply integrated approaches for the assessment of urban ecosys-tem health to better understand and improve the structure andfunction of urban ecosystems.

First used by the World Health Organization (WHO) (Hancockand Duhl, 1988), urban ecosystem health assessments have beendeveloped to focus on health concepts, indicators, standards, andmodels. This approach has been applied to independent urbanecosystems (Chen et al., 2010; Liu et al., 2011; Spiegel et al., 2001;Zhang et al., 2008), urban ecosystem regions (Liu et al., 2009; Suet al., 2009b), and different urban ecosystems in the same urbancluster (Bi and Guo, 2007; Jin and Zheng, 2010; Su et al., 2011b). Areview of the development of these assessments over the past 23years shows that two main problems remain. First, the conceptsand methodology still lack a full integration of social, economic,environmental, and ecological factors with human welfare. Sec-ond, methods need to be established to define the health

www.elsevier.com/locate/enpol

www.elsevier.com/locate/enpol

http://dx.doi.org/10.1016/j.enpol.2013.04.015



http://crossmark.dyndns.org/dialog/?doi=10.1016/j.enpol.2013.04.015&domain=pdf



mailto:[email protected]


M. Su et al. / Energy Policy 59 (2013) 600–613 601

standards of urban ecosystems since there is not an absoluteand well-acknowledged urban ecosystem health standard(Su et al., 2010).

To address the first point above, a systemic analysis frameworkand method is being developed to integrate multiple factors of urbanecosystem health, elucidating the features of urban ecosystems. Forexample, conceptual models of PSR (pressure–state–response), DPSIR,and DPSEEA (driving force–pressure–state–exposure–effects–action)were applied by Zeng et al. (2005), Zou and Wu (2011), and Spiegelet al. (2001), respectively, to organize the urban ecosystem healthindicators. Emergy synthesis has also been introduced to establishhealth indicators and describe ecosystem health levels from theperspective of the biophysical features of urban ecosystems (Liu et al.,2009; Su et al., 2009b) considering the importance of energy andmaterial flows for urban ecosystems and the merit of emergy asan embodied energetic equivalent and sustainability indicator forcomplex ecological economic systems (Brown and Ulgiati, 2011;Campbell and Garmestani, 2012; Lei et al., 2012; Odum, 1983;Ulgiati and Brown, 2009). Moreover, emergy synthesis has been

Fig. 1. The basic procedure to assess e

successfully applied to characterize urban development from thescale of urban clusters (Bi and Guo, 2007; Cai et al., 2009) to mega-cities (Huang et al., 2007; Lan and Odum, 1994; Ulgiati et al., 2011;Zhang et al., 2011), as well as mid-sized cities (Geng et al., 2010; Luet al., 2003; Whitfield, 1994). Emergy synthesis is regarded as aneffective tool for urban ecosystem health assessment because it linksmultiple factors in urban ecosystems, unifying them by acommon unit.

To address the second point, methods are being developed toresolve the existing problems in urban ecosystem health assess-ment. To account for the intrinsic relativity of ecosystem health(Campbell et al., 2004), set pair analysis is a helpful method tolink multiple factors and understand their interactive relation-ships (Zhao, 1989). Based on set pair analysis, an optimalreference set can be derived by integrating the best values ofmultiple macroscopic health indicators, rather than subjectivelysetting the health standard. Therefore, the approach is internaland data-driven rather than imposed as an arbitrary user-definedclassification.

cosystem health of urban clusters.

M. Su et al. / Energy Policy 59 (2013) 600–613602

The formation and expansion of urban clusters is an inevitable partof contemporary urban evolution. Different cities in an urban clustershare similar geographical bases, climatic conditions, infrastructurefacilities, and the advantages of industrial aggregation; however, theyalso compete for local resources and markets (Cai et al., 2009), all ofwhich leads to increasingly close economic and cultural linkagesamong cities. Since phenomena at all scales are interconnected, andoccurrence at one scale is usually controlled by mechanisms operatingon a larger scale, a fundamental energy systems concept (Odum, 2002)is needed. We believe that analysis of the ecosystem health of urbanclusters will promote a greater understanding of the overall structuraland functional performance of different urban clusters and maysuggest ways to reduce the adverse impact of environmental changesand promote sustainable urban development.

A procedure for integrating emergy synthesis and set pairanalysis (EM–SPA) is established in this paper, which summarizes,standardizes, and generalizes the application of the EM–SPA modelto urban ecosystem health assessment. We characterize therelative health levels of different urban clusters based on energyand material flows and related ecosystem health patterns usingthe EM–SPA procedure and matrix analysis. Analysis results forthree Chinese urban clusters are presented, along with suggestionsto improve their ecosystem health. We provide some discussionand make general conclusions at the end of the paper.

2. Methods

2.1. Procedure for ecosystem health assessment of urban clusters

Building on the established protocols for urban ecosystemhealth assessment (Su et al., 2009b), we develop a generalEM–SPA evaluation procedure for urban clusters (see Fig. 1). Thismethod combines the ability of emergy synthesis to characterizethe condition of urban ecosystems with the ability of set pairanalysis to generate a relative health reference set. As shown inFig. 1, we define seven steps for the assessment of the ecosystemhealth of urban clusters, which can be summarized as follows:(1) selecting the cases, (2) defining boundaries and collecting data,(3) drawing an emergy systems diagram, (4) making an emergyaccounting table, (5) establishing and calculating the emergy-based health indicators, (6) applying set pair analysis to assessthe status of urban ecosystem health; and (7) analyzing the healthpatterns of different urban clusters. Steps 1, 5–7 will be furtherexplained in the following sub-sections; for the other steps werefer to our previous study (Su et al., 2009b) and for the procedureof emergy synthesis see Odum (1996).

2.2. Case study selection (Step 1)

Initially, urban cluster assessment should correspond to theresearch interest or problem under investigation. The most famous

Table 1The identity and main characteristics of three Chinese urban clusters.

Urban cluster Geographical position/included districts Area(km2)

Jing-Jin-Tang Beijing, Tianjin and eight cities in Hebei Province 52,636

Yangtze RiverDelta

Shanghai, eight cities in south of Jiangsu Province, andseven cities in north of Zhejiang Province

110,821

Pearl River Delta Twenty-five cities and three counties in GuangdongProvince

41,600

Data source of area, population density and per capita GDP: Zhang (2007)..

urban clusters in China, Jing-Jin-Tang, Yangtze River Delta, andPearl River Delta, whose combined gross domestic product (GDP)will amount to two-thirds of the national total in 2020 are chosenas the urban clusters to be assessed in this study. The maincharacteristics of the three urban clusters are summarized inTable 1.

In order to demonstrate the methodology, we chose a fewtypical urban ecosystems that have the necessary data andboundary clarity to represent the features of the urban clusters.This choice enables the application of emergy synthesis to theseurban clusters. Taking factors such as the geographical position,economic status, natural resources, quality of life and orientationtoward urban development into account, nine urban ecosystemsincluding Beijing, Tianjin and Tangshan; Shanghai, Nanjing andHangzhou; and Guangzhou, Shenzhen and Zhuhai (see Fig. 2) wereselected as representative for the Jing-Jin-Tang, the Yangtze RiverDelta, and the Pearl River Delta urban clusters, respectively.

After the representative urban ecosystems for concerned urbanclusters have been selected, the interval for set pair analysis,denoted as S, is defined.

2.3. Emergy-based urban ecosystem health indicators (Step 5)

The common emergy indicators for systems analysis can beevaluated, after establishing the urban ecosystem boundary basedon administrative divisions and data availability, collecting thedata for multiple internal and external system flows (e.g., naturalconditions, agricultural and industrial production, energy con-sumption and import and export trade) through field investiga-tions and standard yearbooks compiled by local government,drawing the energy systems diagram, and making the emergyaccounting table (including the renewable resources flowing intothe system, the non-renewable resources, locally renewableresources, resources and services flowing across the system'sboundaries, as well as the corresponding money flows). Further-more, the essential urban ecosystem health factors that refer tothe classical ecosystem health assessment framework (Guo et al.,2002; Rapport, 1989, 1998) are still employed to organize theemergy indicators in a systematic way. This allows certain systemindicators to be redefined and expressed in emergy terms andthen incorporated into the framework of the urban ecosystemhealth assessment. As shown in Table 2, the identified character-istics of urban ecosystem health, vigor, organizational structure,resilience, urban ecosystem service function maintenance, andenvironmental impact are incorporated to represent an emergy-based urban ecosystem health state; and the specific indicators(M1–M14) for vigor (F1), organizational structure (F2), resilience(F3), urban ecosystem service function maintenance (F4), andenvironmental impact (F5) and their meanings are listed inTable 3. Once the urban ecosystem health indicators are chosen,the set of indices for set pair analysis, denoted as M, is defined.

Resources Populationdensity(pop/km2)

Per capitaGDP ($)

Scarce water resources, abundant land and biologicalresources, relatively ample minerals

763 4340

Dense river and lake resources, ample plants andanimals, scarce mineral and energy

746 5136

Abundant water network, plenty of plants andanimals, deficient minerals

1035 5249

Fig. 2. The locations of the nine primary cities in the three urban clusters. BJ: Beijing; TJ: Tianjin; TS: Tangshan; SH: Shanghai; NJ: Nanjing; HZ: Hangzhou; GZ: Guangzhou;SZ: Shenzhen; ZH: Zhuhai.

Table 2Development of concerned factors of ecosystem health assessment.

Items Factors and their meanings

Natural ecosystem healthassessment*

Vigor: a functional characteristic of systems which is measured in terms of activity, metabolism or productivity (Mageau et al.,1995)Organization: number and diversity of interactions among system components (Rapport et al., 1998a)Resilience: a system's ability to maintain its structure and pattern of behavior in the presence of stress (Mageau et al., 1995;Rapport, 1989)Maintenance of ecosystem service: a system's ability to provide a suite of goods and services (Rapport et al., 1998b)Reduced subsidies: capability for maintaining productivity with minimal and non-increasing subsidy (Milsky, 2004; Rapportet al., 1998b)Management options: maintenance of management options allows flexibility for various future needs (Rapport et al., 1998b)Damage to neighboring systems: sustaining the flow of ecosystem services without negative impact on the neighboring systems(Milsky, 2004; Rapport et al., 1998b)

Traditional urban ecosystem healthassessment

Vigor: urban economic productivity and circulation efficiency of energy and materials (Guo et al., 2002)Organizational structure: system structure and diversity of components (Guo et al., 2002)Resilience: elasticity and sustainability of systems usually measured in terms of ability of waste treatment and recycling rate ofmaterials (Guo et al., 2002)Ecosystem service function maintenance: good environmental quality and high living convenience (Guo et al., 2002)Population health: physical health and cultural level of residents (Guo et al., 2002)

Emergy-based urban ecosystem healthassessment

Vigor: basic function of the system (urban economic productivity based on energy utilization)Organizational structure: component of the system (stability and diversity of energy and materials sources)Resilience: sustainability of the system (elasticity and renewal ability based on energy and resources)Urban ecosystem service function maintenance: services for the dominant population in the system (energy welfare for humanliving)Environmental impact: relationship between the system and surroundings (pressure from energy depletion and exchange)

n Generally, only the first three factors (vigor, organization, resilience) are considered.


It is indicated that both the merit of emergy-based indicatorsand the characteristics of urban ecosystems are considered whenselecting the urban ecosystem health indicators. Correspondingly,the selected indicators can be classified into two groups. Onegroup includes M1, M3, M4, M5, M7, M12, M13, and M14, whichmainly presents the intrinsic merit of emergy synthesis. Theseindicators can be applied into other systems such as agriculturalsystem, forest system, and fishery system. The other groupincludes M2, M6, M8, M9, M10, and M11, which mainly presentsthe specific characteristics of urban systems. Additionally, thetraditional health indicators for natural systems are referred toby the greatest extent. For example, empower density (M1) issimilar to gross primary production (GPP) and total systemthroughput (TST), emergy diversity index (M5) is similar to

Shannon's diversity index, and the ratio of waste-to-renewableemergy (M8) is similar to the ratio of photosynthesis per unit ofrespiration (P/R).

Choosing the difficult factor-resilience-as an example, theprocess of health indicator selection is explained in more detail.Resilience usually refers to a system's ability to absorb stress,maintain its structure and pattern of behavior in the presence ofstress (Mageau et al., 1995). The study on natural ecosystemindicated that “resilience is not experimentally measured directly”and “measuring resilience of a system is difficult because it impliesthe ability to predict the dynamics of that system under stress”(Mageau et al., 1995; Milsky, 2004). When it comes to the morecomplex urban ecosystem, it is much more difficult to measure theresilience. However, resilience of urban ecosystem can still be

Table 3Emergy-based urban ecosystem health indicators.

Factor layer Index layer Expression Meaning

F1 Vigor M1 Empower density U/area Emergy gradient available for economic or ecological useM2 Emdollar ratio U/GDP Intensity of economic activity

F2 Organizational M3 Emergy self-sufficiency (R+N0+N1)/U Fraction of total emergy from natural resourcesstructure M4 Ratio of concentrated to rural use (F+G+P2I3+N1)/(R+N0) The ratio of concentrated to dispersed emergy use

M5 Emergy diversity index −∑ðUiU Þ � lnðUi

U Þ emergy use variety

F3 Resilience M6 Carrying capacity density based onrenewable emergy

(R�Pop)/(U� area) Supportable population density based on renewableresources

M7 Fraction of locally non-renewableemergy used

(N0+N1)/U Utilization of locally non-renewable resources

M8 Ratio of waste-to-renewable emergy W/R Pressure of waste dischargeF4 Urban ecosystem servicefunction maintenance

M9 Per capita emergy used U/Pop Residents’ standard of living in the urban ecosystem

M10 Per capita fuel emergy used Fuel/Pop Fuel usage due to production and living consumptionM11 Ratio of electricity emergy used el/U Degree of development of the urban ecosystem

F5 Environmental impact M12 Environmental loading ratio (U−R)/R Potential damage for the concerned urban ecosystemcaused by resources depletion

M13 Ratio of export to import (N2+B+P1E3)/(F+G+P2I3) Equity of trade between the concerned urban ecosystemand other systems

M14 Ratio of import to indigenousemergy

(F+G+P2I3)/(R+N0+N1) Degree of dependence and impact on the outside

R: renewable emergy flow; N: indigenous non-renewable flows; N0: indigenous dispersed rural flow (e.g., soil loss and soil formation); N1: indigenous concentrated use (e.g.,oil, coal, gas, nuclear electricity, hydroelectricity, rock, and iron ore); N2: raw materials exported without use; F: imported fuel; G; imported goods; I3: dollars paid forimported service; B: exported products; E3: dollars received for exported service; P2: world emdollar ratio; P1: local emdollar ratio; U: total emergy used (U¼R+N0+N1+F+G+P2I3, Ui, respectively, denotes each component of U); W: waste; Fuel: fuel used; el: electricity used; GDP: gross domestic products; Pop: population.


described indirectly when regarding the sources of stress andrecovery ability. Population growth, resources depletion, andenvironmental pollution are main stress sources (only endogenousstresses are emphasized herein), while the recovery ability comesfrom the supporting system, e.g., resources (especially indigenousrenewable and non-renewable resources), technology and man-agement. Therefore, such emergy-based indicators as carryingcapacity density based on renewable emergy (M6), fraction oflocally non-renewable emergy used (M7), and the ratio of waste-to-renewable emergy (M8) are selected to represent resiliencefrom aspects of population carrying capacity, resource utilization,and elasticity on pressure of waste discharge.

2.4. Set pair analysis for urban ecosystem health assessment (Step 6)

Set pair analysis is a valuable method to integrate variousobjects and describe the relationships among different objects.With respect to a given problem, a set pair is formed by puttingtwo interrelated sets together. Based on the analysis of thefeatures of this set pair, a connection degree formula describingthe relationship of the two sets can be set up, including the degreeof identity, the degree of discrepancy and the degree of contra-diction. This implies it is possible to conduct relative evaluationsfor different objects (i.e., sets) under certain problematicconditions.

For urban ecosystem health assessment, the problem spacebased on set pair analysis can be defined as

Q ¼ fS;M;Hg ð1Þ

S¼ fskg ðk¼ 1;2;…; pÞ ð2Þ

M¼ fmrg ðr¼ 1;2;…;nÞ ð3Þ

H¼ ðhkrÞp�n ð4Þ

where S is the assessed interval set composed of several selectedurban ecosystems from the urban clusters of interest, and skrepresents the kth urban ecosystem. M is the set of indices(including positive and negative indices) composed of the

emergy-based urban ecosystem health indicators, and mr denotesthe rth indicator. H denotes the decision-making matrix aboutProblem Q based on set pair analysis, and hkr is the attribute valueof indicator mr in the interval sk.

By obtaining the best value for each indicator, the optimalevaluation set can be generated, denoted as U¼{u1, u2,…un}, whilethe worst set is denoted as V¼{v1, v2,…vn}, where ur and vrrespectively represent the best and worst values of the index mr.Then, in the domain Xr¼{hkr, ur, vr} (k¼1,2,…p), the identity andcontradiction degree of the set pair {hkr, ur} marked as akr and ckr(choosing positive index as an example here) can be calculated byEqs. (5) and (6), implying the approaching and diverging degreesbetween hkr and ur.

akr ¼hkr

ur þ vrð5Þ

ckr ¼urvr

ður þ vrÞhkrð6Þ

considering the weight of each index (the information entropymethod is applied here to confirm the indices’ weights), theaverage identity and contradiction degree marked as ak and ckcan be counted in the comparative interval [U, V], respectively,describing the approaching and diverging extents between skand U. Next, the approximate degree between sk and U, markedas rk, can be expressed as

rk ¼ak

ak þ ckð7Þ

A larger value of rk indicates a stronger health status of the kthurban ecosystem. More details about the procedure and equationsof set pair analysis can be found in Su et al. (2009a).

2.5. Ecosystem health pattern analysis of urban clusters (Step 7)

In addition to comprehensive health assessment, set pairanalysis can also be applied to the scale relevant to assess eachfactor related to urban ecosystem health to describe the specificcharacteristics of urban ecosystems. By summarizing the perfor-mance of typical urban ecosystems within an urban cluster, the

Table 4Indicator weight and value for nine urban ecosystems in three urban clusters.

Index Weight Value

Beijing Tianjin Tangshan Shanghai Nanjing Hangzhou Guangzhou Shenzhen Zhuhai

M1 0.0600 20.65 35.59 9.06 125.24 32.07 6.93 39.49 99.76 36.14M2 0.0123 3.93 9.16 4.82 6.94 7.00 3.13 4.55 3.15 7.68M3 0.0820 0.09 0.27 0.88 0.08 0.07 0.15 0.12 0.08 0.18M4 0.0870 77.80 10.09 3.64 12.78 23.10 5.94 21.25 11.10 4.59M5 0.0014 1.23 1.44 1.09 1.35 1.12 1.41 1.39 1.40 1.65M6 0.0428 4.69 62.95 72.41 137.53 19.07 45.26 23.35 71.24 74.42M7 0.1163 0.08 0.20 0.75 0.01 0.05 0.03 0.10 0.01 0.04M8 0.1386 93.74 4.75 59.74 4.96 33.86 4.86 10.31 0.75 0.84M9 0.0316 22.02 40.60 16.83 58.37 35.41 17.41 39.09 107.11 68.08M10 0.0324 10.53 14.12 0.73 15.50 25.57 5.07 20.19 20.08 12.03M11 0.0170 0.16 0.09 0.28 0.11 0.11 0.27 0.11 0.22 0.10M12 0.0490 198.98 12.92 6.43 14.60 46.48 7.79 42.25 12.07 6.13M13 0.0493 0.43 0.92 0.68 1.58 0.54 2.13 0.90 4.41 2.69M14 0.2803 10.31 2.71 0.13 12.16 12.99 5.79 7.34 11.10 4.59

The indices’ weights are determined based on the information entropy method. More details about calculation of information entropy weight can be found in Su et al.(2009b).

Table 5The relative health states of nine urban ecosystems in three urban clusters based on the emergy health indicators.

Value Beijing Tianjin Tangshan Shanghai Nanjing Hangzhou Guangzhou Shenzhen Zhuhai

ak 0.1544 0.1755 0.4627 0.2821 0.1332 0.1766 0.1433 0.4827 0.3248ck 0.6236 0.2191 0.4394 0.4113 0.5248 0.3347 0.3244 0.3604 0.2524rk 0.1984 0.4447 0.5129 0.4069 0.2025 0.3454 0.3064 0.5726 0.5627

Fig. 3. Relative health states of nine urban ecosystems and three urban clustersbased on the emergy indicators of urban ecosystem health.


health patterns of those urban clusters can be described using Eq. (8):

R1 R2 ⋯ Rj ⋯ Rq

h i

¼ 1p

r11 r12 ⋯ r1j ⋯ r1qh i�

þ r21 r22 ⋯ r2j ⋯ r2qh i

þ⋯

þ rk1 rk2 ⋯ rkj ⋯ rkqh i

þ⋯

þ rp1 rp2 ⋯ rpj ⋯ rpqh i�

ð8Þ

where Rj is the relative health status of the jth assessment factor forthe urban cluster of interest, q is the total number of assessmentfactors for urban ecosystem health; p is the number of selected urbanecosystems in an urban cluster; and rkj is the relative health status ofthe jth assessment factor for the kth urban ecosystem in the urbancluster of interest.

When the health matrices of different urban clusters areobtained, their ecosystem health patterns can be analyzed andcompared. This method provides a good description of urbancluster health and allows related environmental management tobe suggested.

3. Results

3.1. Relative health of urban clusters in terms of the emergy-basedhealth indicators

After collecting the required data for 2005, including descrip-tions of the natural ecosystem condition, agriculture, industry,import and export trade and environmental protection, theemergy flows of selected urban ecosystems were analyzed (Exam-ple calculations for Beijing, Shanghai, and Shenzhen, prominentcities in each of the three urban clusters are given in AppendicesA–C). From tables like these the values of emergy-based health

indicators were determined for all nine cites (see Table 4). Com-bining these values with the information entropy weight of eachindicator (see Table 4), the approximate degree of urban ecosys-tem health relative to the optimal evaluation set was obtainedthrough set pair analysis (Table 5).

The relative health status of each urban cluster based on theemergy indicators was then calculated by averaging the data fromthe selected typical urban ecosystems in the urban cluster. Asindicated in Fig. 3, the relative health state of the Jing-Jin-Tangurban cluster was in the middle of the three clusters. Specifically,within the cluster, Tianjin and Tangshan were medium and Beijingwas worst among the nine urban ecosystems. The relative level ofhealth in the Yangtze River Delta urban cluster was low, with thehealth states of Shanghai and Hangzhou being relatively low andthat of Nanjing being only slightly better than Beijing. The relativehealth status of the Pearl River Delta urban cluster was high, withthe health levels of Zhuhai and Shenzhen being highest among the

Fig. 4. Relative health states of three urban clusters based on each factor of theemergy health indicators (F1: vigor; F2: organizational structure; F3: resilience; F4:urban ecosystem service function maintenance; F5: environmental impact).


nine urban ecosystems examined, although that of Guangzhouwas relatively low.

3.2. Ecosystem health characteristics of urban clusters

To further describe and analyze the differences and relationshipsamong the three urban clusters regarding their ecosystem healthpattern, set pair analysis was also conducted on the five factors used inthe emergy-based urban ecosystem health assessment, including vigor(F1), organizational structure (F2), resilience (F3), urban ecosystemservice function maintenance (F4) and environmental impact (F5).

3.2.1. Ecosystem health performance of urban ecosystems based oneach factor of the emergy-based health indicator

(1)
The Jing-Jin-Tang urban clusterEach city's health performance, relative to the optimal evalua-tion set was calculated as follows:
F1 F2 F3 F4 F5Beijing : ½0:4201 0:5808 0:0545 0:4620 0:0290�Tianjin : ½0:4878 0:4055 0:5228 0:5523 0:3569�

Tangshan : ½0:1724 0:4712 0:1063 0:2232 0:9104�

The vigor measures for Beijing, Tianjin and Tangshan were allrelatively poor, because of their low empower density and highemdollar ratio, especially that of Tangshan. The health states ofthe three Jing-Jin-Tang urban ecosystems were relatively strongin organizational structure, because of their relatively diverseemergy sources and components. The health states for resilienceof Beijing and Tangshan were poor, because of relatively largewaste discharge, small carrying capacity based on renewableemergy and large environmental loading ratio, while that ofTianjin was medium. For urban ecosystem service functionmaintenance, the relative health state of Tangshan was the worstamong the nine urban ecosystems, while those of Beijing andTianjinwere slightly better. For Environmental impact, the healthstate of Tangshan was relatively strong, attributed to a relativelysmall emergy investment ratio and a large ratio of export toimport, while the state of Beijing was relatively poor and thestate of Tianjin was moderate.

(2)
The Yangtze River Delta urban clusterThe relative health performance for Shanghai, Nanjing and Hang-zhou, in terms of the five health factors, was calculated as follows:
F1 F2 F3 F4 F5Shanghai : ½0:8579 0:1790 0:9207 0:6645 0:1244�Nanjing : ½0:4916 0:2590 0:2532 0:6540 0:0431�

Hangzhou : ½0:1705 0:1823 0:7722 0:3948 0:3271�

The vigor measures for Shanghai and Nanjing were relativelystrong, while that of Hangzhou was relatively poor. The healthstates of the three Yangtze River Delta cities for organizationalstructure were all relatively poor. The health states of resilience forShanghai and Hangzhou were relatively strong, while that ofNanjing was relatively poor. For urban ecosystem service functionmaintenance, the health states of the three cities were relativelystrong. For environmental impact, the health states of Nanjing andShanghai were relatively poor, except that of Hangzhou, whichwas moderate.

(3)
The Pearl River Delta urban clusterThe relative health performance for Guangzhou, Shenzhen andZhuhai, in terms of the five factors, was calculated as follows:
F1 F2 F3 F4 F5Guangzhou : ½0:6250 0:3592 0:3890 0:6313 0:1007�Shenzhen : ½0:8845 0:1738 0:9829 0:8520 0:2287�Zhuhai : ½0:5198 0:1790 0:9437 0:6400 0:4372�

For vigor, the health level of Shenzhen was the highest amongthe nine cites examined, while the vigor measures for Guangz-hou and Zhuhai were moderate. The relative health state fororganizational structure of Guangzhou was at a moderatelevel, while the state of organizational structure measures inShenzhen and Zhuhai were relatively poor. The heath states ofShenzhen and Zhuhai based on resilience were relativelystrong, while that of Guangzhou was moderate. For urbanecosystem service function maintenance, the health states ofall three cities were relatively strong. The relative health statesfor environmental impact of Shenzhen and Zhuhai weremoderate, while that of Guangzhou was relatively poor.

3.2.2. Ecosystem health pattern analysis of urban clustersAfter determining the ecosystem health performance of the

typical urban ecosystems, the final health pattern matrix of eachurban cluster was obtained using Eq. (6) as follows:

F1 F2 F3 F4 F5The Jing� Jin� Tang urban cluster : 0:3601 0:4859 0:2279 0:4125 0:4321TheYangtzeRiverDeltaurbancluster : 0:5067 0:2068 0:6487 0:5711 0:1648ThePearlRiverDeltaurbancluster : 0:6765 0:2373 0:7719 0:7078 0:2555

To understand the health patterns of these three urban clusters,Fig. 4 was drawn using the above health pattern matrixes. Thisindicates that the Pearl River Delta urban cluster, with the highesthealth level, was strong in terms of vigor, resilience, and urbanecosystem service function maintenance, while the Jing-Jin-Tangurban cluster, with a moderate health status, was strong in termsof organizational structure and environmental impact.

Certain characteristics of the three urban clusters related to theflow of energy and materials can be further summarized. TheYangtze River Delta urban cluster and the Pearl River Delta urbancluster showed more similarity, in contrast to the Jing-Jin-Tangurban cluster. The Pearl River Delta urban cluster, affected by alarge amount of resource ingestion from the surrounding area (asshown by relatively weak environmental impact), contained greatwealth (as shown by relatively high vigor), and consequentlyhuman needs were largely satisfied (as shown by relatively highurban ecosystem service function maintenance). There was a greatdeal of investment in eco-environmental protection and construc-tion (as shown by the relatively high Resilience) but with anunsecured supporting system because of unsustainable energysources and unreasonable usage structure (as shown by therelatively poor Organizational structure value). For the YangtzeRiver Delta urban cluster, the relative health levels based on allfactors of the emergy indicators were slightly lower than those ofthe Pearl River Delta urban cluster, while the relative health levels


based on most of the factors were higher than those of theJing-Jin-Tang urban cluster except for organizational structureand Environmental impact.

4. Discussion

4.1. What influences the health levels of urban clusters?

This analysis indicates that the clusters order, from relativelyhigh to relatively low levels: Pearl River Delta, Jing-Jin-Tang, andlastly Yangtze River Delta. In terms of the specific health patterns,the Pearl River Delta urban cluster was the strongest in vigor,resilience, and urban ecosystem service function maintenance andrelatively strong in organizational structure and environmentalimpact. The Jing-Jin-Tang urban cluster was strong in organiza-tional structure and environmental impact and the weakest invigor, resilience, and urban ecosystem service function mainte-nance. The Yangtze River Delta urban cluster was relatively strongin vigor, resilience, and urban ecosystem service function main-tenance and the weakest in organizational structure and environ-mental impact. Based only on the integrated health levels and thehealth patterns, it is difficult to determine what most influencesthe health states of urban clusters.

Analysis of the indicator weights may provide another view-point on the problem, since the information entropy weight canreflect the contribution of each indicator to the relative healthstates and therefore relative ability of an indicator to distinguishdifferent states of health. According to the weights, environmentalimpact, resilience, and organizational structure had relativelystrong effects on the health results compared with vigor andurban ecosystem service function maintenance. This suggests thataspects such as the natural resources, environmental pressure, andthe interaction between urban ecosystems and their surroundingsall played a role in the final health state. This implies that multiplefactors should be considered when establishing management thataims to improve the health level of urban ecosystems.

4.2. Management guidelines for urban clusters

According to the characteristics of the flow of energy andmaterials, appropriate management can be implemented toimprove the state of ecosystem health in the urban clusters. Forexample, for the Yangtze River Delta urban cluster, managementshould be established to improve the diversity and stability of theenergy and resources structure and reduce the environmentalimpact on the surroundings, based on the knowledge that theregional economy has rapidly developed to satisfy human needsby largely depending on outside energy and material inputs. Forthe Jing-Jin-Tang urban cluster, management can be established toimprove vigor, urban ecosystem service function, and resilience byameliorating regional economic development to increase both the

Fig. 5. The relative health states amongst five factors of the emergy health indicatoorganizational structure; F3: resilience; F4: urban ecosystem service function maintenan

productivity and the degree to which human requirements aresatisfied; there should also be greater investment in environmen-tal protection.

4.3. The health performances of urban ecosystems in the same urbancluster

From the above results regarding emergy-based urban ecosys-tem health indicators, we conclude that the health performancesof different urban ecosystems within the same urban cluster showsimilarity at both the scale of the integrated health indicators andfor each factor. The similarity of those in the same urban cluster isparticularly obvious in terms of organizational structure, with theorganizational structure of nearly all cities within the same urbancluster being similar except for that of Guangzhou, which isrelatively far away from the other two in the Pearl River Deltaurban cluster. This pattern of similarity may be partly because thestructure indicator includes emergy self-sufficiency, the ratio ofconcentrated to rural use, the emergy diversity index, and dependsmore on the urban natural condition than on other indicators. Thedifferent urban ecosystems in the same urban cluster share similarnatural conditions and therefore are similar in terms of healthstates based on structure.

Even though the relative health levels of some factors fordifferent urban ecosystems within the same urban cluster arelargely similar, this is not consistently observed across all factors.For example, the health level of Beijing based on organizationalstructure was relatively high while the levels based on resilienceand environmental impact are relatively low; the health level ofTangshan, in the same urban cluster, based on environmentalimpact was relatively high while the levels based on vigor,resilience and urban ecosystem service function maintenancewere relatively low. This suggests that there should be differentmanagement policies for different urban ecosystems in the sameurban cluster according to the specific situation.

4.4. Specific management for different urban ecosystems

To promote regional sustainable development, management ofparticular urban clusters based on the health patterns of the urbanclusters has previously been mentioned. It helps to understand thedifferences among urban clusters to determine the appropriatemanagement policies. Administrative management is usually notdirectly implemented at the urban cluster scale, but at the cityscale; the related policy for the urban cluster has to be negotiatedby different urban government councils. Therefore, managementpolicies for each city within the urban cluster need to be putforward according to the specific conditions, while referring to themanagement direction for the whole urban cluster.

To analyze the individual condition of each urban ecosystem,the relative health states among each health factor can be clearly

rs for the Jing-Jin-Tang urban cluster and three urban ecosystems (F1: vigor; F2:ce; F5: environmental impact).


described through a normalization process, as defined in Eq. (9):

dj ¼rj

∑5j ¼ 1rj

ð9Þ

where dj (j¼1, 2, …, 5) represents the characteristic value of thejth factor (vigor, organizational structure, resilience, urban ecosys-tem service function maintenance and environmental impact) fora given urban ecosystem when compared with the relative healthstates of the other health factors, rj denotes the above-defined rkvalue of the jth factor for the given kth urban ecosystem. Based onthis normalization process, the health status of a given urbanecosystem can be described using five dimensions (factors), asshown in Fig. 5. Taking the three cities in the Jing-Jin-Tang urbancluster as an example, certain management measures aimed atimproving resilience and environmental impact should be imple-mented for Beijing, while for Tianjin, regulation should be imple-mented to improve organizational structure, environmentalimpact and vigor, and a suitable management plan should beimplemented to improve the vigor, resilience and urban ecosystemservice function maintenance for Tangshan.

Fig. 6. The main progress and potential improvement of urba

4.5. Cross scale management of cities and urban clusters

It should be pointed out that all these regulations for urbanecosystems are in the regulative framework of the whole urbancluster. The relationship between the management guidelines forthe urban cluster and the specific management for an urbanecosystem is similar to that between national policy and localimplementation. In other words, urban managers should under-stand and consider the guidelines at the urban cluster scale, andthen establish suitable management schemes according to thespecific conditions of the urban ecosystem. Taking Beijing as anexample, decision-making concerned with improving the healthlevel of the ecosystem should be performed in two steps. First,using the development direction of the Jing-Jin-Tang urban clusteras a macroscopic guideline, it is implied that the regional economyneeds to be further ameliorated, human requirements furthersatisfied, and environmental protection continuously strength-ened to improve the vigor, urban ecosystem service functionmaintenance and resilience of the system. This will help realizechanges in the external environment, such as increasing the use ofenergy and materials by the urban cluster, which implies Beijing

n ecosystem health assessment based on EM–SPA model.


must develop an efficient approach for using limited energy understronger competition. Second, according to the situation in Beijingcity in 2005 shown in Fig. 5(b) and the dynamic analysis resultsfrom 1986 to 2005 shown in our previous study (Su et al., 2011a),the health states in terms of resilience and environmental impactmust be improved by effective management measures, such ascontrolling population growth, recycling waste and decreasingwaste discharge, while reducing dependence on outside inputs ofenergy and materials.

4.6. Potential improvement of the EM–SPA model

4.6.1. Urban ecosystem health indicatorsEmergy synthesis provides a platform for measuring various

flows (e.g., energy, materials, population, and information) in theurban ecosystem by a common unit. It enables integrating social,economic and ecological factors in the urban ecosystem, charac-terizing the urban ecosystem structure and functions and indir-ectly reflecting the properties of urban ecosystem health. Referringto the factors of concern in natural ecosystem health assessmentand traditional urban ecosystem health assessment (see Fig. 6aand b), related emergy-based indicators are identified to evaluateurban ecosystem health in terms of vigor, organizational structure,resilience, urban ecosystem service function maintenance, andenvironmental impact (see Fig. 6c). On the one hand, the healthfactors and indicators are different between natural ecosystemsand urban ecosystems due to their different characteristics. Forexample, the indicator expressions of vigor, organizational struc-ture and resilience for urban ecosystems are different from theclassical expressions for natural ecosystems (Berkes et al., 2003;Holling, 1973; Mageau et al., 1995; Rapport et al., 1998a; Walkeret al., 2004), and the services provided for human beings must beemphasized for urban ecosystem health since people are thedominant controller in urban ecosystems. On the other hand, theemergy-based health indicators can be improved or expanded byincorporating other concepts and indicators based on specificmethods. For example, ecological footprint (Rees, 1992; Reesand Wackernagel, 1996) can be incorporated into indicatorsof resilience due to their resonance with carrying capacity. Interms of urban ecosystem service function, the emergy-basedindicators used in this study did not cover some importantfunctions of supporting, provisioning, regulating and culturalservices (Costanza et al., 1997; MEA, 2005), especially some vitalsocial benefits in urban ecosystems such as healthcare, education,and culture welfare; and for environmental impact, certain im-proved sustainability indicators (Campbell and Garmestani, 2012;Li et al., 2011; Tilley, 2011) and indicators based on ecologicalnetwork analysis (Fath and Patten, 1999; Patten, 1985) may becombined to better denote such relationships as competition andmutualism between urban ecosystems and their surroundings orrelationships among different urban ecosystems in the same urbancluster.

4.6.2. Urban ecosystem health standardMany studies on confirming ecosystem health standard have

been conducted since the 1990s (Calow, 1993; Rapport, 1992, 1993;Waltner-Toews, 2004). Until now, it is still difficult to establish anacknowledged standard of ecosystem health (Su et al., 2010).Therefore, set pair analysis, which measures the inherent uncer-tainty and extensive relation among different variables, is appliedto represent the relativity of urban ecosystem health and indir-ectly define the health standard. With more data of more cities(especially those with experience of great natural, social andeconomic events or changes) over a long period, dynamic analysismay help observe the potential threshold (critical point or

interval) of urban ecosystem health and identify the key driversand variables (Walker et al., 2012) that induce the shift betweenhealthy and unhealthy states of urban ecosystem health.

4.6.3. Combination of different methods of urban ecosystem healthassessment

As stated by Campbell (2000), energy system theory provides abasis for measuring and interpreting ecosystem health. For a healthyecosystem, the maximum empower principle indicates that allprocesses operating within system components are functioningoptimal to produce maximum empower based on the underlyingemergy signature. This unified basis allows certain emergy indicators(e.g., the ratio of the emergy required to produce a unit of energyflowing along the pathways in an ecosystem to that unit of energy,the empower needed to maintain a given quantity of stored emergyin structure, the ratio of the emergy produced to the emergyconsumed in an ecosystem) to be calculated to represent the stateof ecosystem health as a whole. But the concerned problem occurs, i.e., how to make sure the proximity between simulated model andreal system especially for complex urban ecosystem and even urbanclusters; and these holistic abstract indicators are unable to reflectmanifold properties of the system that are concerned and under-standable by the public and managers.

In this paper, emergy synthesis is applied in the viewpoint ofunderlying energy basis of the system while multiple indicatorsfrom five factors are used to represent manifold properties ofurban ecosystem health. It can give people more concrete cogni-tion of urban ecosystem health and provide more practicalmanagement options. But multiple factors cause a new problem,i.e., how to reconcile different factors of health. For example, ahighly exploited urban ecosystemwith intensive economic activitybut few green areas may perform stronger in vigor but poorer inresilience and environmental impact when comparing with alowly exploited urban ecosystem. Although different weights havebeen assigned to different factors and indicators, the determina-tion of weight is still an open question.

Without an absolutely perfect method, the attempt of combin-ing different methods may be hopeful for methodologicalimprovement of urban ecosystem health assessment in terms ofhealth definition, measurement and interpretation.

5. Conclusions

By combining the merit of emergy as a quality normalizedavailable energy-based equivalent for calculating sustainabilityand other indicators for complex socio-ecological systems andthe ability of set pair analysis to describe extensive relationships,the EM–SPA model was further developed, and a procedure for itsuse on urban clusters was established in this paper. This methodaims to summarize, standardize, and generalize the application ofthe EM–SPA model in urban ecosystem health assessment. Giventhe rapid pace of urban cluster development, the EM–SPA modelwill help better understand the differences among urban clusters’health characteristics and will contribute to the establishment ofeffective urban management that promotes regional sustainabledevelopment. This approach can also be applied to ecosystemhealth assessment at other scales, in which case, the selection ofsystems and related boundaries should be modified.

In this paper, the EM–SPA model was applied to investigate thehealth state in three typical urban clusters, located in the east ofChina, using three urban ecosystems (cities) as representative ofeach of the three urban clusters. The integrated relative healthstates of the urban clusters and their specific health patterns weresummarized and compared. We found that aspects such as theavailability of natural resources, environmental pressures, and the

Table A1Emergy synthesis table of resource and economy flow for Beijing in 2005.

No. Item Raw data Solar transformity(sej/unit)

Refs. Solaremergy (sej)

Emdollarvalues (US$)

Renewable sources1 Sunlight (J) 5.87E+19 1 Odum (1996) 5.87E+19 8.45E+062 Wind, kinetic (J) 8.80E+15 2.45E+03 Odum et al. (2000), Folio #1 2.15E+19 3.09E+063 Rain, geopotential (J) 2.64E+15 4.70E+04 Odum et al. (2000), Folio #1 1.24E+20 1.78E+074 Rain, chemical (J) 3.33E+16 3.05E+04 Odum et al. (2000), Folio #1 1.01E+21 1.45E+085 Earth cycle (J) 2.38E+16 5.80E+04 Odum (2000), Folio #2 1.38E+21 1.99E+08

Indigenous renewable energy6 Hydroelectricity (J) 5.04E+15 3.36E+05 Odum (1996) 1.69E+21 2.43E+087 Agriculture production (J) 3.58E+16 3.36E+05 Brown and McClanahan (1996) 1.20E+22 1.73E+098 Livestock production (J) 3.70E+16 3.36E+06 Brown and McClanahan (1996) 1.24E+23 1.78E+109 Fisheries production (J) 2.69E+14 3.36E+06 Brown and McClanahan (1996) 9.04E+20 1.30E+08

Non-renewable sources from within Beijing10 Coal (J) 2.86E+17 6.69E+04 Odum (1996) 1.91E+22 2.75E+0911 Mineral (g) 4.54E+12 1.43E+09 This work 6.50E+21 9.35E+0812 Soil losses (g) 9.97E+11 1.68E+09 Odum (1996) 1.68E+21 2.42E+0813 Net top soil losses (J) 3.38E+14 7.40E+04 Brown and Bardi, (2001), Folio#3 2.50E+19 3.60E+06

Imports and outside sources14 Goods (US$) 2.32E+10 1.06E+12 Jiang et al. (2009) 2.45E+22 3.53E+0915 Services (US$) 4.84E+09 1.06E+12 Jiang et al. (2009) 5.11E+21 7.35E+0816 Tour (US$) 1.99E+10 5.88E+12 This work 1.17E+23 1.68E+1017 Fuels (J) 1.71E+18 9.51E+04 This work 1.62E+23 2.33E+10

Exports18 Goods (US$) 1.71E+10 6.95E+12 Jiang et al. (2009) 1.19E+23 1.71E+1019 Services (US$) 1.78E+09 6.95E+12 Jiang et al. (2009) 1.24E+22 1.78E+09

Resource consumed20 Fuels (J) 1.98E+18 8.18E+04 This work 1.62E+23 2.33E+1021 Electricity (J) 2.04E+17 2.69E+05 Odum (1996) 5.49E+22 7.90E+09

Waste produced22 Solid waste (J) 5.23E+16 3.02E+06 Huang et al. (1995) 1.58E+23 2.27E+1023 Waste water (J) 6.40E+14 1.12E+06 Huang et al. (1995) 7.17E+20 1.03E+08

Dollar flow24 GDP (US$) 8.61E+10 6.95E+12 Jiang et al. (2009) 5.98E+23 8.61E+10

Table B1Emergy synthesis table of resource and economy flow for Shanghai in 2005.




Renewable sources1 Sunlight (J) 1.99E+19 1 Odum (1996) 1.99E+19 2.86E+062 Wind, kinetic (J) 6.64E+15 2.45E+03 Odum et al. (2000), Folio #1 1.63E+19 2.35E+063 Rain, geopotential (J) 2.81E+14 4.70E+04 Odum et al. (2000), Folio #1 1.32E+19 1.90E+064 Rain, chemical (J) 3.93E+16 3.05E+04 Odum et al. (2000), Folio #1 1.20E+21 1.73E+085 Tide (J) 8.77E+16 7.39E+04 Odum et al. (2000), Folio #1 6.48E+21 9.32E+086 Wave (J) 8.71E+17 5.10E+04 Odum et al. (2000), Folio #1 4.44E+22 6.39E+097 Earth cycle (J) 9.19E+15 5.80E+04 Odum (2000), Folio #2 5.33E+20 7.67E+078 Typhoon (J) 3.86E+16 6.89E+04 Huang (2004) 2.66E+21 3.83E+08

Indigenous renewable energy9 Agriculture production (J) 3.73E+16 3.36E+05 Brown and McClanahan (1996) 1.25E+22 1.80E+0910 Livestock Production (J) 1.08E+16 3.36E+06 Brown and McClanahan (1996) 3.62E+22 5.21E+0911 Fisheries production (J) 1.48E+15 3.36E+06 Brown and McClanahan (1996) 4.98E+21 7.17E+0812 Wood (J) 6.20E+13 2.21E+04 Romitelli (2000) 1.37E+18 1.97E+05

Non-renewable sources from within Shanghai13 Natural gas (J) 2.35E+16 5.88E+04 Romitelli (2000) 1.38E+21 1.99E+0814 Oil (J) 1.52E+16 8.90E+04 Odum (1996) 1.35E+21 1.94E+0815 Soil losses (g) 1.01E+21 1.68E+09 Odum (1996) 1.70E+21 2.45E+0816 Net top soil losses (J) 3.44E+14 7.40E+04 Brown and Bardi (2001), Folio#3 2.54E+19 3.65E+06

Imports and outside sources17 Goods (US$) 9.56E+10 1.06E+12 Jiang et al. (2009) 1.01E+23 1.45E+1018 Services (US$) 8.04E+10 1.06E+12 Jiang et al. (2009) 8.49E+22 1.22E+1019 Tour (US$) 1.64E+10 6.95E+12 This work 1.14E+23 1.64E+1020 Fuels (J) 5.22E+18 8.32E+04 This work 4.34E+23 6.24E+10


Table B1 (continued )





Resource consumed23 Fuels (J) 2.79E+18 7.57E+04 This work 2.11E+23 3.04E+1024 Electricity (J) 3.32E+17 2.69E+05 Odum (1996) 8.93E+22 1.28E+10



Table C1Emergy synthesis table of resource and economy flow for Shenzhen in 2005.




Renewable sources1 Sunlight (J) 7.31E+18 1 Odum (1996) 7.31E+18 1.05E+062 Wind, kinetic (J) 1.49E+15 2.45E+03 Odum et al. (2000), Folio #1 3.65E+18 5.25E+053 Rain, geopotential (J) 3.51E+15 4.70E+04 Odum et al. (2000), Folio #1 1.65E+20 2.37E+074 Rain, chemical (J) 2.07E+16 3.05E+04 Odum et al. (2000) Folio #1 6.31E+20 9.08E+075 Tide (J) 1.14E+16 7.39E+04 Odum et al. (2000), Folio #1 8.44E+20 1.21E+086 Wave (J) 2.76E+17 5.10E+04 Odum et al. (2000), Folio #1 1.41E+22 2.03E+097 Earth cycle (J) 2.83E+15 5.80E+04 Odum (2000), Folio #2 1.64E+20 2.36E+078 Typhoon (J) 1.19E+16 6.89E+04 Huang (2004) 8.20E+20 1.18E+08

Indigenous renewable energy9 Agriculture production (J) 6.97E+14 3.36E+05 Brown and McClanahan (1996) 2.34E+20 3.37E+0710 Livestock production (J) 2.35E+15 3.36E+06 Brown and McClanahan (1996) 7.88E+21 1.13E+0911 Fisheries production (J) 3.53E+14 3.36E+06 Brown and McClanahan (1996) 1.19E+21 1.71E+08

Non-renewable sources from within Beijing12 Soil losses (g) 1.03E+10 1.68E+09 Odum (1996) 1.73E+19 2.49E+0613 Net top soil losses (J) 3.48E+12 7.40E+04 Brown and Bardi (2001), Folio#3 2.58E+17 3.71E+04

Imports and outside sources14 Goods (US$) 8.13E+10 1.06E+12 Jiang et al. (2009) 8.58E+22 1.23E+1015 Services (US$) 9.17E+09 1.06E+12 Jiang et al. (2009) 9.68E+21 1.39E+0916 Fuels (J) 4.44E+17 9.81E+04 This work 4.35E+22 6.26E+0917 Tour (US$) 1.98E+10 2.00E+12 This work 3.97E+22 5.71E+09


Resources consumed20 Fuels (J) 4.09E+17 8.94E+04 This work 3.65E+22 5.25E+0921 Electricity (J) 1.58E+17 2.69E+05 Odum (1996) 4.26E+22 6.13E+09




interaction between the urban ecosystems and their surroundingsplayed an important role in the final health state of the systems.Given data on more cities in different urban clusters, especiallythose located in the west of China and other countries, the relativehealth states of urban clusters in different regions and theircharacteristics regarding energy and material flows could beanalyzed in more detail, which will help to establish more specificand effective measures that promote regional sustainabledevelopment.

Based on the pattern of health indicators in the urban clusters,general management guidelines that aimed to improve the healthlevels of the urban clusters from a biophysical perspective weresuggested. As a general principle, we propose coordination betweenthe people developing specific management guidelines for each city

within an urban cluster to address the each city's health indicators andconsistent policies that take into account the management direction ofthe whole urban cluster.

Acknowledgements

Financial support is provided by the National Natural ScienceFoundation of China (Grant No. 40901269), Program for NewCentury Excellent Talents in University (NCET-09-0226), theNational Science Foundation for Innovative Research Group (No.51121003), and China Postdoctoral Special Foundation (Grant No.201003063). The authors would also thank the help of the editor


and the comments of the reviewers, which significantly improvedthe quality of this paper.

Appendix A

See Appendix Table A1.

Appendix B

See Appendix Table B1.

Appendix C

See Appendix Table C1.

References

Berkes, F., Colding, J., Folke, C, 2003. Navigating Social-ecological Systems: BuildingResilience for Complexity and Change. Cambridge University Press,Cambridge416.

Bi, D.S., Guo, X.P., 2007. An evaluation on the urban ecosystem health of ChangjiangDelta. Ecological Economy 2, 327–330, in Chinese.

Brown, M.T., Bardi, E., 2001. Folio #3: emergy of ecosystems. Handbook of EmergyEvaluation: A Compendium of Data for Emergy Computation Issued in a Seriesof Folios. Center for Environmental Policy, University of Florida, Gainesville, FL,USA.

Brown, M.T., McClanahan, T.R., 1996. Emergy analysis perspectives of Thailand andMekong River dam proposals. Ecological Modelling 91, 105–130.

Brown, M.T., Ulgiati, S., 2011. Understanding the global economic crisis: a biophy-sical perspective. Ecological Modelling 223, 4–13.

Cai, Z.F., Zhang, L.X., Zhang, B., Chen, Z.M., 2009. Emergy-based analysis of Beijing–Tianjin–Tangshan region in China. Communications in Nonlinear Science andNumerical Simulation 14, 4319–4331.

Calow, P., 1993. Ecosystems not optimized. Journal of Aquatic Ecosystem Health 2(1), 55.

Campbell, D.E., 2000. Using energy systems theory to define, measure and interpretecological integrity and ecosystem health. Ecosystem Health 6 (3), 181–204.

Campbell, D.E., Cai, T.T., Olsen, T.W., 2004. Ecosystem health: energy indicators. In:Cleveland, C.J. (Ed.), Encyclopedia of Energy. Elsevier, Boston, pp. 131–142.

Campbell, D.E., Garmestani, A.S., 2012. An energy systems view of sustainability:emergy evaluation of the San Luis Basin, Colorado. Journal of EnvironmentalManagement 95, 72–97.

Chen, K.L., Su, M.X., Li, S.C., Lu, J.X., Chen, Y.Y., Zhang, W., Liu, Z.J., 2010. The healthassessment of the urban ecosystem of Xining city. Geographical Research 29(2), 214–222, in Chinese.

Costanza, R., d'Arge, R., de Groot, R., Farberk, S., Grasso, M., Hannon, B., Limburg, K.,Naeem, S., O'Neill, R.V., Paruelo, J., Raskin, R.G., Sutton, P., van den Belt, M., 1997.The value of the world's ecosystem services and natural capital. Nature 387,253–260.

Fath, B.D., Patten, B.C., 1999. Review of the foundations of network environ analysis.Ecosystems 2, 167–179.

Geng, Y., Zhang, P., Ulgiati, S., Sarkis, J., 2010. Emergy analysis of an industrial park:the case of Dalian, China. Science of the Total Environment 408 (22),5273–5283.

Guo, X.R., Yang, J.R., Mao, X.Q., 2002. Primary studies on urban ecosystem healthassessment. China Environmental Science 22 (6), 525–529, in Chinese.

Hancock, T., Duhl, L.J., 1988. Promoting Health in the Urban Context. WHO HealthyCities Papers no.1.

Holling, C.S., 1973. Resilience and stability of ecological systems. Annual Review ofEcological Systems 4, 1–23.

Huang, S.L., 1998. Urban ecosystems, energetic hierarchies, and ecological econom-ics of Taipei metropolis. Journal of Environmental Management 52, 39–51.

Huang, S.L., 2004. Energy Basis for Urban Ecological Economic System. Chan's ArchBooks Co. Ltd, Taipei271, in Chinese.

Huang, S.L., Kao, W.C., Lee, C.L., 2007. Energetic mechanisms and development of anurban landscape system. Ecological Modelling 201, 495–506.

Huang, S.L., Wu, S.C., Chen, W.B., 1995. Ecosystem, environmental quality andecotechnology in the Taipei metropolitan region. Ecological Engineering 4 (4),233–248.

Jiang, M.M., Chen, B., Zhou, J.B., Yang, Z.F., Ji, X., Zhang, L.X., Chen, G.Q., 2009. Anecological evaluation of Beijing ecosystem based on emergy indices. Commu-nications in Nonlinear Science and Numerical Simulation 14, 2482–2494.

Jin, C.F., Zheng, G.Z., 2010. Urban ecosystem health assessment for urban clustersalong river in Jiangsu Province. Environment and Sustainable Development 6,13–17, in Chinese.

Lan, S.F., Odum, H.T., 1994. Emergy evaluation of the environment and economy ofHongkong. Journal of Environmental Science 6 (4), 432–439.

Lei, K.P., Hu, D., Zhou, S.Q., Guo, Z., 2012. Monitoring the sustainability and equity ofsocioeconomic development: a comparison of emergy indices using Macao,Italy and Sweden as examples. Acta Ecologica Sinica 32, 165–173.

Li, L.J., Lu, H.F., Ren, H, Kang, W.L., Chen, F.P., 2011. Emergy evaluations of threeaquaculture systems on wetlands surrounding the Pearl River Estuary, China.Ecological Indicators 11, 526–534.

Liu, G.Y., Yang, Z.F., Chen, B., Ulgiati, S., 2009. Emergy-based urban health evaluationand development pattern analysis. Ecological Modelling 220 (18), 2291–2301.

Liu, G.Y., Yang, Z.F., Chen, B., Ulgiati, S., 2011. Monitoring trends of urbandevelopment and environmental impact of Beijing, 1999–2006. Science of theTotal Environment 409, 3295–3308.

Lu, H.F., Ye, Z., Zhao, X.F., Peng, S.L., 2003. A new emergy index for urban sustainabledevelopment. Acta Ecologica Sinica 23 (7), 1363–1368.

Mageau, M.T., Costanza, R., Ulanowicz, R.E., 1995. The development and initialtesting of a quantitative assessment of ecosystem health. Ecosystem Health 1(4), 201–213.

Millennium Ecosystem Assessment (MEA), 2005. Ecosystems and Human Well-Being: Synthesis, DC. Island Press, Washington155.

Milsky, D.J., 2004. A Conceptual Analysis of Ecosystem Health. Doctoral Disserta-tion. Chicago, IL, USA: University of Illinois at Chicago.

Odum, H.T., 1983. System Ecology: An Introduction. JohnWiley & Sons, New York370.Odum, H.T., 1996. Environmental Accounting: Emergy and Environmental Decision

Making. John Wiley & Sons, New York384.Odum, H.T., 2000. Folio #2: emergy of global processes. Handbook of Emergy

Evaluation: A Compendium of Data for Emergy Computation Issued in a Series ofFolios. Center for Environmental Policy, University of Florida, Gainesville, FL, USA.

Odum, H.T., 2002. Explanations of ecological relationships with energy systemsconcepts. Ecological Modelling 158, 201–211.

Odum, H.T., Brown, M.T., Brandt-Williams, S., 2000. Folio #1: introduction andglobal budget. Handbook of Emergy Evaluation: A Compendium of Data forEmergy Computation Issued in a Series of Folios. Center for EnvironmentalPolicy, University of Florida, Gainesville, FL, USA.

Patten, B.C., 1985. Energy cycling in the ecosystem. Ecological Modelling 28, 1–71.Rapport, D.J., 1989. What constitutes ecosystem health. Perspectives in Biology and

Medicine 33 (2), 120–132.Rapport, D.J., 1992. What is Clinical Ecology? Ecosystem Health-New Goals in

Environmental Management. Island Press, Washington, DC.Rapport, D.J., 1993. Ecosystems not optimized a reply. Journal of Aquatic Ecosystem

Health 2 (1), 57.Rapport, D.J., 1998. Dimensions of ecosystem health. In: Rapport, D.J., Costanza, R.,

Epstein, P.R., Gaudet, C., Levins, R. (Eds.), Ecosystem Health. Blackwell Science,Malden and Oxford, pp. 34–40.

Rapport, D.J., Costanza, R., McMichael, A.J., 1998a. Assessing ecosystem health.Trends in Ecology and Evolution 13, 397–402.

Rapport, D.J., Gaudet, C., Karr, J.R., Baron, J.S., Bohlen, C., Jackson, W., Jones, B.,Naiman, R.J., Norton, B., Pollock, M.M., 1998b. Evaluation landscape health:integrating societal goals and biophysical process. Journal of EnvironmentalManagement 53, 1–15.

Rees, W.E., 1992. Ecological footprints and appropriated carrying capacity: whaturban economics leaves out. Environment and Urbanization 4 (2), 121–130.

Rees, W.E., Wackernagel, M., 1996. Urban ecological footprints: why cities cannot besustainable—and why they are a key to sustainability. Environmental ImpactAssessment Review 16, 223–248.

Romitelli, M.S., 2000. Emergy analysis of the new Bolivia–Brazil gas pipeline. In:Brown, M.T., Brandt-Williams, S., Tilley, D.R., Ulgiati, S. (Eds.), Emergy Synth-esis: Theory and Applications of the Emergy Methodology. Center for Environ-mental Policy. University of Florida, Gainesville, pp. 53–69.

Spiegel, J.M., Bonet, M., Yassi, A., Molina, E., Concepcion, M., Mast, P., 2001.Developing ecosystem health indicators in centro Habana: a community-based approach. Ecosystem Health 7 (1), 15–26.

Su, M.R., Chen, B., Xu, L.Y., Zhao, Y.W., Liu, G.Y., Zhang, Y., Yang, Z.F., 2011a. Anemergy-based analysis of urban ecosystem health characteristics for Beijingcity. International Journal of Exergy 9 (2), 192–209.

Su, M.R., Fath, B.D., Yang, Z.F., 2010. Urban ecosystem health assessment: a review.Science of the Total Environment 408 (12), 2425–2434.

Su, M.R., Yang, Z.F., Chen, B., 2009a. Set pair analysis for urban ecosystem healthassessment. Communications in Non-linear Science and Numerical Simulation14 (4), 1773–1780.

Su, M.R., Yang, Z.F., Chen, B., Ulgiati, S., 2009b. Urban ecosystem health assessmentbased on emergy and set pair analysis―a comparative study of typical Chinesecities. Ecological Modelling 220 (18), 2341–2348.

Su, M.R., Yang, Z.F., Liu, G.Y., Chen, B., 2011b. Ecosystem health assessment andregulation for urban ecosystems: a case study of the Yangtze River Delta urbancluster, China. Journal of Environmental Informatics 18 (2), 65–74.

Tilley, D.R., 2011. Dynamic accounting of emergy cycling. Ecological Modelling 222,3734–3742.

Ulgiati, S., Ascione, M., Zucaro, A., Campanella, L., 2011. Emergy-based complexitymeasures in natural and social systems. Ecological Indicators 11, 1185–1190.

Ulgiati, S., Brown, M.T., 2009. Emergy and ecosystem complexity. Communicationsin Nonlinear Science and Numerical Simulation 14 (1), 310–321.

Walker, B.H., Carpenter, S.R., Rockstrom, J., Crépin, A.S., Peterson, G.D., 2012.Drivers, “Slow” variables, “Fast” variables, shocks, and resilience. Ecology andSociety 17 (3), 30.

Walker, B., Holling, C.S., Carpenter, S.R., Kinzig, A., 2004. Resilience, adaptability andtransformability in social and cological systems. Ecology and Society 9 (2), 5.

http://refhub.elsevier.com/S0301-4215(13)00254-1/sbref1
















































































































































Waltner-Toews, D., 2004. Ecosystem Sustainability and Health-A PracticalApproach. Cambridge University Press, Cambridge.

Whitfield, D.F., 1994. Emergy Basis for Urban Land Use Patterns in Jacksonville,Florida. Department of Landscape Architecture, University of Florida,Gainesville.

Zeng, Y., Shen, G.X., Guang, S.F., Wang, M., 2005. Assessment of urban ecosystemhealth in Shanghai. Resources and Environment in the Yangtze Basin 14 (2),208–212, in Chinese.

Zhang, L.X., 2007. Comparative Analysis of Typical Ecological Economic Systems ofUrban Cluster in China. Doctoral Degree Dissertation, Beijing Normal University,Beijing. (in Chinese).

Zhang, H., Wang, X.R., Ho, H.H., Yong, Y., 2008. Eco-health evaluation for theShanghai metropolitan area during the recent industrial transformation (1990–2003). Journal of Environmental Management 88 (4), 1047–1055.

Zhang, Y., Yang, Z.F., Liu, G.Y., Yu, X.Y., 2011. Emergy analysis of the urbanmetabolism of Beijing. Ecological Modelling 222, 2377–2384.

Zhao, K.Q., 1989. Set pair and set pair analysis-a new concept and systematicanalysis method. In: Proceedings of the State Forum on System Theory andRegional Planning in China. pp. 87–91.

Zou, J., Wu, S.S., 2011. Urban ecosystem health evaluation and analysis based onDPSIR model—a case study in Hengyang. Journal of Hengyang Normal Uni-versity 32 (6), 113–117 , (in Chinese).

















Documents

Ecosystem health pattern analysis of urban clusters based on emergy synthesis: Results and implication for management