Ecological Network Analysis for Carbon Metabolism of Eco-industrial

May 18, 2015 - 2.1Network Model for Carbon Metabolism in Eco-industrial Parks .... of information and communication technology, high-tech commerce, ...
43 downloads 4 Views 4MB Size
Download PDF
Article pubs.acs.org/est

Ecological Network Analysis for Carbon Metabolism of Eco-industrial Parks: A Case Study of a Typical Eco-industrial Park in Beijing Yi Lu,† Bin Chen,*,† Kuishuang Feng,‡ and Klaus Hubacek‡ †

School of Environment, Beijing Normal University, State Key Joint Laboratory of Environmental Simulation and Pollution Control, Beijing 100875, China ‡ Department of Geographical Sciences, University of Maryland, College Park, Maryland 20742, United States S Supporting Information *

ABSTRACT: Energy production and industrial processes are crucial economic sectors accounting for about 62% of greenhouse gas (GHG) emissions globally in 2012. Ecoindustrial parks are practical attempts to mitigate GHG emissions through cooperation among businesses and the local community in order to reduce waste and pollution, efficiently share resources, and help with the pursuit of sustainable development. This work developed a framework based on ecological network analysis to trace carbon metabolic processes in eco-industrial parks and applied it to a typical eco-industrial park in Beijing. Our findings show that the entire metabolic system is dominated by supply of primary goods from the external environment and final demand. The more carbon flows through a sector, the more influence it would exert upon the whole system. External environment and energy providers are the most active and dominating part of the carbon metabolic system, which should be the first target to mitigate emissions by increasing efficiencies. The carbon metabolism of the eco-industrial park can be seen as an evolutionary system with high levels of efficiency, but this may come at the expense of larger levels of resilience. This work may provide a useful modeling framework for low-carbon design and management of industrial parks. neutral” industrial parks have been constructed in Belgium,22 and 55 low-carbon industrial parks are currently under operation in China.23 Recently, there have been a number of studies with a focus on emission mitigation in eco-industrial parks,21,24−29 however, in most cases with a strong emphasis on the input (supply) and output (emission) side of carbon flows to and from the park but omitting important carbon cycling processes within these parks. For example, embodied carbon in materials are imported from the external environment and transferred to firms for producing products and services within the park or consumed by visitors or staff of eco-industrial parks, and which are eventually flowing back to the external environment in terms of products, wastes, or GHG. For this reason, it is necessary to adopt a metabolism-based perspective to show how carbon is supplied, consumed, and cycled within the eco-industrial park and finally released or exported to its surrounding environment. The additional focus on carbon flows within the park allows to identify the major consumers and pathways and thus leverage points for intervention. Metabolism is a metaphor borrowed from the biotic world, describing how material and energy are required and flow in open systems. Eco-industrial parks, as open systems, require import of raw materials and energy from external environments

1. INTRODUCTION Among diverse anthropogenic activities producing greenhouse gas (GHG) emissions, industrial sectors are the major contributor, globally.1 For example, about 62% of global GHG emissions resulted from energy and other industrial production in 2012.2 As a consequence, industrial emitters have become a main focus for tracing and mitigating GHG emissions at source.3−11 In China, which accounted for 25.4% of global GHG emissions in 2011,12 industrial activities contributed about 90% of territorial carbon emissions.13 Thus, managing industrial GHG emissions is playing a main role for China to achieve its climate mitigation target of reducing 40−45% of CO2 intensity by 2020 compared to 2005.14 One important step toward a low-carbon industry is to reduce total energy consumption and emissions through efficiency improvement. However, in China efficiency gains were outpaced by emissions associated with economic growth, structural change, and lifestyle changes.15−17 Decoupling carbon emissions from economic growth through increasing efficiencies and reorganizing industrial production chains will play an increasingly significant role in mitigating industrial emissions. Industrial symbiosis in eco-industrial parks, which consists of physical exchanges among separate collaborative industries striving for synergistic possibilities and collective benefits,18,19 has proven to be an effective tool for the mitigation of carbon emissions.20,21 Subsequently, low-carbon industrial parks have been pursued by managers and scientists due to their great potential for reducing emissions. In practice, so-called “carbon © 2015 American Chemical Society

Received: Revised: Accepted: Published: 7254

November 20, 2014 May 17, 2015 May 18, 2015 May 18, 2015 DOI: 10.1021/es5056758 Environ. Sci. Technol. 2015, 49, 7254−7264

Article

Environmental Science & Technology

Figure 1. Conceptual network model for carbon metabolic processes in eco-industrial parks.

indirect effects, and revealing properties of structure and function through ecological flows (e.g., carbon flows).45,46 This approach can be regarded as a combination of the two streams of metabolic methods, since it is a process-based analytical tool on the basis of a flow inventory. It has been widely applied to natural ecosystems, their metabolic processes and mutual interactions based on specific chemical fluxes.47,48 Applications of ENA in the metabolism of artificial systems have recently become popular,49−51 but there are only very few such studies for carbon metabolism in Chinese industries.39,52 This study aims to investigate the processes and properties of the carbon metabolism for eco-industrial parks based on ENA. In the following, Section 2 introduces an ENA-based methodology to simulate the carbon metabolism in ecoindustrial parks. Section 3 illustrates a case study of a typical park in Beijing, and provides a systematic analysis on its metabolic behavior, structural stability, sectoral interactions, control conditions and systematic properties in the context of ENA. Section 4 discusses practical barriers and possible solutions for applying the ENA framework to analyze the carbon metabolism in eco-industrial parks.

and discharge wastes to support daily operations. After great interest in other flows,30−39 the carbon metabolism is seen as providing additional benefits when being integrated into the metabolic framework at the level of eco-industrial parks. Tracking of carbon in this framework of eco-industrial parks will facilitate our cognition of carbon metabolic processes and properties, and facilitate a systematic implementation of emission mitigation. Two streams of metabolism-orientated approaches have been applied to the investigation of carbon metabolic processes in eco-industrial parks. One is based on the metabolic inventories to quantify and decompose life cycle emissions through inventory accounting,25,26,29 decomposition analysis,24,28 and life cycle assessment.21,40,41 The other one is process-based methods tracking metabolic flows using material flow analysis.42−44 Both of them have been adopted to quantify metabolic fluxes via carbon inventories and flow-charts, where particular emphasis has been put on quantifying inflows and emissions. However, more information on sectoral behavior, especially the cumulative effects due to indirect interactions between nonadjacent sectors, is required to better understand the mechanism of carbon metabolism in eco-industrial parks. Investigation on metabolic hierarchy shaped by the trophic roles of producer, consumer, and decomposer also greatly contributes to predict system stability in the long-run. The complex interactions between parks and supporting environments also deserve more investigations. All of them can be implemented in the framework of ecological network analysis (ENA). ENA is a system-oriented methodology with unique advancements in measuring mutual interactions and system structure. It makes it possible to integrate different components as a linked system, signifying and quantifying both direct and

2. METHODOLOGY 2.1. Network Model for Carbon Metabolism in Ecoindustrial Parks. The concept of metabolism depicts physical and chemical processes in organisms or ecosystems including resource supply from the external environment, nutrition and energy exchange within the system, and waste emission.53 Due to the similarity of natural ecosystems and artificial systems, the metabolic perspective has been widely adopted for the analysis 7255

DOI: 10.1021/es5056758 Environ. Sci. Technol. 2015, 49, 7254−7264

Article

Environmental Science & Technology of stock-flow based systems subsequently, with applications at urban, industrial, or household scales.54−56 From a metabolic perspective, eco-industrial parks host complex processes and need external support, such as flows of raw materials, energy, and labor for their support. Carbon is embodied in the production, transformation, consumption, emission, and decomposition processes involving materials and energy to support the daily operation of eco-industrial parks. It is a natural-anthropogenic dualistic process including both (1) natural processes as carbon absorption by photosynthesis and emission via respiration by plants and soils; and (2) anthropogenic processes as production and consumption of fuels, foods, and services, etc. These carbon fluxes (embodied in the materials and energy accounted for and calculated based on environmental input−output analysis)57,58 not only flow among different metabolic sectors within the administrative boundary of an eco-industrial park, but also flow between the park and the external environment. To account for this, the system boundary shall include both the eco-industrial park and its external environment (i.e., the upstream processes of imports in urban areas). It is an activity based boundary embracing all carbon metabolic processes within the parks and their environments to better distinguish both inner-park and transpark flows of these processes. Inner-park flows refer to carbon fluxes among diverse production/consumption sectors mainly in terms of exchanging embodied carbon in materials and energy. These metabolic sectors can be classified as follows: (1) “producers” such as agriculture (which can be found in some food processing dominated parks), energy and construction sectors provide resources and services to meet demands of industrial production and maintenance (these sectors only play the role as “producers” within the administrative boundary of the eco-industrial parks, whereas they could be regarded as “primary consumers” considering extraction and inflows from surrounding environments); (2) “consumers” such as manufacture and processing, domestic and business sectors contribute to the symbiotic chain by making use of the supplied resources; and (3) “decomposers” mainly undertake the service of sewage treatment and waste management, and natural absorption processes (as carbon sinks, etc.). In addition, eco-industrial parks also require external support. Thus, the external environment is considered as the real “producer” of the carbon metabolic system. The ecological network model is used to represent the carbon metabolism of eco-industrial parks through carbon flows among several network components. Figure 1 shows a conceptual network model illustrating how carbon metabolizes from, to, and within eco-industrial parks. 2.2. Ecological Network Analysis. To better understand the mechanism of carbon metabolism in eco-industrial parks, mutual interactions, control conditions, structural stability, and system-wide performance of the metabolic system were evaluated through network utility analysis (NUA), network control analysis (NCA), and network stability analysis (NSA), utilizing a set of indicators based on the ENA framework.45,46 More details of these tools are provided below. 2.2.1. Network Utility Analysis (NUA). NUA is mainly applied to quantify (also in terms of effects on the whole system) the flows between sectors in a network.59 Mutual relationships among network compartments can be expressed by the direct utility matrix D and dimensionless integral utility matrix U. D = [dij] interprets the direct utility of each metabolic routes in the network system,59

dij =

fij − f ji Ti

(1)

where dij is the direct utility between sector i and j, indicating the net input ratio of sector i; f ij is the carbon fluxes from sector j to i; Ti represents the total input or output flows of i-th sector. In addition to the investigation of direct connections (those associated with carbon flows between adjacently connected sectors in the metabolic system), indirect effects by the reason for those nonadjacent sectors should be emphasized as well.60 Thus, the net integral effect (include direct and indirect effects) of each sector is represented by U, U = D0 + D1 + D2 + ··· + Dm = (I − D)−1

(2)

where I is the identity matrix; U shows both direct and indirect utilities transmitted by pathways in lengths 1, 2, ..., m (m is number of metabolic sectors), reflecting the strength of network organization as well. SignD and SignU are sign matrices introduced to simplify mutual relationships quantified by D and U. Thus, mutual interactions between sectors can be interpreted as conditions of mutualism (+/+), exploitation (±), being exploited (∓), competition (−/−), and neutrality (0/0).61 Variations of signs also indicate the nature of network interactions and organization. On the basis of these signs, network mutualism index (MI) and synergism index (SI) are introduced to measure the performance of the metabolic system,20,62 where MI indicates the ratio of positive and negative utilities, while SI quantifies the total magnitude of them,50 MI = SignU ( +)/SignU ( −) n

SI =

(3)

n

∑ ∑ uij j=1 i=1

(4)

where SignU(+) = ∑ij max(Sign(uij), 0), SignU(−) = ∑ij − min(Sign(uij), 0). System mutualism and synergism occurs when MI > 1 or SI > 0. 2.2.2. Network Control Analysis (NCA). Searching for leverage points (i.e., emission controllers) to influence emissions is of great significance for carbon mitigation. In a complex interconnected metabolic system, mutual interactions among all network components reflect the interdependence and ability to influence, organize, and regulate the whole system. In this way, all sectors can be defined as system controllers with different capacities of dominance. They influence each other and contribute to the network organization, interpreting the control of the entire system.63 So NCA is based on a pairwise integral flows to measure the control and dependence capacities of network components, and indicate the control or level of influence of each sector within a system.49,63,64 In NCA, network flow interactions are also divided into the direct and integral parts. The integral dominances of each component are determined by direct interaction matrices G, G′, and indirect matrices N, N′, N = [nij] = (I − G)−1

(5)

N ′ = [nij′] = (I − G′)−1

(6)

where G = [gij], gij = f ij /Tj; G′= [gij′], gij′= f ij /Ti, quantifying direct influences of metabolic sectors; and N sums the infinite power series of direct interactions.45 Two distributed control 7256

DOI: 10.1021/es5056758 Environ. Sci. Technol. 2015, 49, 7254−7264

Article

Environmental Science & Technology

properties can be regarded as a pair-wised contradiction, only by balancing which can stabilize a metabolic system. So an information stability index (ISI) is proposed to articulate the balancing potential of network organization,

metrics as control allocation matrix (CA) and dependence allocation matrix (DA) are thus developed to reveal the control and dependence condition, respectively: ⎧ nij − n′ ji ⎪ ⎪ nij − n′ ji > 0, caij = m ∑i = 1 nij − n′ ji CA = [caij] ≡ ⎨ ⎪ ⎪ nij − n′ ji ≤ 0, caij = 0 ⎩

m

ISI =

m

1 − ∑ j = 1 ∑i = 1 sij Resilience = m m ∑ j = 1 ∑i = 1 sij Efficiency

(12)

where sij = H/Hmax. The ISI is a comprehensive indicator in revealing both network stability and evolution stage as shown in Figure 2. When (1) ISI = 1, network efficiency is equal to

(7)

⎧ nij − n′ ji ⎪ ⎪ nij − n′ ji > 0, daij = m ∑ j = 1 nij − n′ ji DA = [daij] ≡ ⎨ ⎪ ⎪ nij − n′ ji ≤ 0, daij = 0 ⎩ (8)

where caij indicates the control degree sector j on i, 0 ≤ caij ≤ 1; while daij reflects the degree of dependence of sector j on i, 0 ≤ daij ≤ 1.50 A system control index (CI) is formulated by combining CA and DA to reflect the control condition of the entire network. It interprets the control power and organization capacity of the network, measuring the ability for self-regulation of the system’s metabolism as a consequence,50 m

CI =

m

m

m

∑ j = 1 ∑i = 1 caij + ∑ j = 1 ∑i = 1 daij m2

(9)

Figure 2. Stability balance between efficiency and resilience in a specific system.

2.3. Network Stability Analysis (NSA). 2.3.1. Information Stability Analysis. Efficiency and resilience are both necessary for the stability of eco-industrial parks’ carbon metabolism, since pursuing the former ensures higher benefits with lower costs (time, labor, money, waste, etc.), while the latter guarantees a sustainable status of eco-industrial parks. Nevertheless, balancing these prerequisites is difficult: highly ordered systems can survive when facing risks, but they are always short of processing efficiency that is deemed unacceptable in modern production systems; while those with high efficiencies are too “brittle” to reserve sufficient freedom to reconfigure themselves for resisting disturbance.65 Therefore, evaluation on the deviation of efficiency and resilience promises a comprehensive insight in defining the stability and succession phase of the carbon metabolic network for eco-industrial parks. Measurements of network efficiency and resilience are both referred to in information theory. This issue was initiated by MacAuthor,66 who primarily introduced Shannon’s information measure to the flows in network systems, hij = −

⎛ fij ⎞ ⎟⎟ log⎜⎜ TST ⎝ TST ⎠

resilience, suggesting a sustainable stability of the metabolic structure; (2) ISI > 1, more carbon are utilized to stabilize the structure rather than improve efficiency, being the symbol of mature systems; (3) ISI < 1, higher efficiency suggests an evolutionary network that requires, consumes, and emits more carbon for system yields. The closer the ISI to 1, the more stability a network will achieve. 2.3.2. Trophic Structure Analysis. For further depicting the structure of carbon metabolism, trophic dynamics theory is introduced to observe the transfer of nutrition and energy from one components (as prey) of the system to another (as predator), and emphasize the trophic or energy-availing relationships of several trophic levels.67 Due to the function ecological components undertake, they should be primarily aggregated as some functional units (namely trophic levels). Mimicking the natural ecosystems, artificial network also has its unique morphology with linked functional components and transferring pathways.68 It is feasible to introduce the trophic dynamic view into the investigation of metabolic structure. Nevertheless, challenges are posed to describe trophic structure and identify sectoral roles clearly in artificial networks. The formula is introduced for the classification of sectors’ trophic roles,

fij

(10)

where hij is the connecting diversity of each network flow, whose sum H = ∑ hij signifies the overall systematic diversity. In a specific network, the contemporary pattern reflected by H always struggles to approximate to the most stable structure with the largest connecting diversity Hmax,

Hmax = −log

1 m2

L=1+

∑ Pn·Ln

(13)

where L shows the trophic level of each metabolic sector; Pn is the percentage that sector’s n-th input source to its total input; Ln is the trophic level corresponding to the n-th input. Similar to the biotic world, a pyramid structure shows the metabolic system is well-organized. Thus, a system structure index (STI) is newly proposed to measure the deviations to the ideal form,

(11)

The ratio of the actual connecting diversity H to the maximum pattern Hmax indicates the network efficiency in transmitting carbon fluxes, while its deviation to 1 measures the capacity in resisting external disturbances. These stabilized 7257

DOI: 10.1021/es5056758 Environ. Sci. Technol. 2015, 49, 7254−7264

Article

Environmental Science & Technology Table 1. System-Wide Indicators of the Carbon Metabolism Network Model in BDA Parka results

a

indicators

nomenclature

brief description

with EE

without EE

nodes links link density connectivity TST FCI MI SI CI ISI STI

m l l/m l/m2 ∑mj = 1∑mi = 1f ij ∑mj ((njj − 1)Tj/njj)/TST eq2 eq 3 eq 9 eq 12 eq 14

numbers of metabolic sectors numbers of direct metabolic routes metabolic linking degree metabolic connectivity the total system throughflow to show the metabolic magnitude Finn’s cycling effect to indicate the dominance of the cycle network mutualism index to show the metabolism mutualism degree network synergism index to show the metabolic synergism magnitude system control index to show the metabolic self-regulation capability Information stability index to show the potential of metabolic stability system structure index to show the structural deviation to the pyramid

7 23 3.29 0.47 64 306 1.00 2.06 7.00 0.29 0.64 0.21

6 12 2.00 0.33 29 716 1.00 2.00 6.02 0.26 0.75 1.00

EE stood for the external environment.

Figure 3. Carbon metabolism network model for carbon metabolism in BDA Park (Unit: t/a). n

STI =

2.4. System-Wide Indicators. A series of indicators have been introduced in this paper to represent the overall performance of the carbon metabolism in eco-industrial parks from different angles. Some of them are derived from network utility analysis, network control analysis and network stability analysis as discussed above, while the rest are based on other studies.69,70 Nomenclature, equations and brief descriptions are shown in Table 1. 2.5. Description of Case Study and Data. Business Development Area International Business Park (short as BDA Park) has been chosen as case study, being a typical ecoindustrial park, located in the E-Town of southeastern Beijing (Yizhuang region in Daxing District). It is both a national

∑ rp·Sign(rp − 1 − rp) p=1

(14)

where p represents the number of trophic levels; rp is the trophic constitutions indicating the ratio total carbon fluxes of each level to the total system throughflow; Sign(r0 −r1) is assumed to be positive when p = 1. When (1) STI = 1, the system is sustainable with a perfect pyramid structure; (2) STI = 0, a uniform shape is reflected; (3) 0 < STI < 1, the system is in a moderate condition despite of some structural defects; (4) STI < 0, a brittle structure implies an unsustainable future (Supporting Information, SI, Figure S4). 7258

DOI: 10.1021/es5056758 Environ. Sci. Technol. 2015, 49, 7254−7264

Article

Environmental Science & Technology

emission mitigation of eco-industrial parks. In addition, carbon emissions were mainly associated with the industry and business sector (7341 t) that was corresponding to the characteristic of the industrial park. The waste management (6541 t) and residential living sector (2573 t) also significantly contributed to emission processes. All of these sectors would undoubtedly become the hotspot of emission control in BDA Park. We also noticed that indirect emissions via symbiotic chains played a more important role in the network, where the external environment and energy providers were the major drivers accounting for the increasing emissions of the hotspot sectors (i.e., sectors of industry and business, waste management and residential housing) in terms of embodied carbon import, whereas construction and infrastructure sector acted as the intermediate step in the process chain from drivers to the final emitters. In addition, metabolic components were also investigated in the model. External environment (26.9%) and energy providers (26.0%) were the most significant parts of the carbon metabolism, which are also key sectors in BDA Park. Shares of industry and business (14.0%), construction and infrastructure (13.5%), waste management (10.5%) and residential housing sectors (8.7%), are also main emitters, significantly influencing the carbon balance as well. Thus, all of these metabolic drivers are obvious first targets for introducing cleaner technologies to increase utilization efficiencies. 3.2.2. Metabolic Stability. Network stability analysis (ISI = 0.64) indicates a higher efficiency than resilience in the case study, due to the fact that as a developing system, it needed more carbon to stimulate its productivity rather than to diversify metabolic routes. The insufficient resilience might make the BDA Park vulnerable when facing a reduction of external supply of carbon embodied resources (e.g., insufficiency power supply would be a crushing blow for BDA Park since it sustained almost all activities in situ). Investigations on the metabolic structure also help in clarifying the contribution each metabolic sector makes and prove information for the long-term stability of a specific system. In a next step, we reinterpreted the metabolic structure by the identification of trophic roles as producer (external environment), primary consumers (energy providers and construction and infrastructure sector), secondary consumers (industry and business and residential housing sectors), and decomposers (waste management and landscaping) using eq 13. Figure 4 visualizes an irregular pyramid structure of the case park. The imperfect hierarchy (STI = 0.21) indicates that not all sectors were organized “harmoniously” in the model, especially due to overrepresented secondary consumers. As a high-tech eco-industrial park, BDA Park attaches great importance to the commerce of high-end goods and services, which both require large intakes of carbon embodied in energy and industrial resources. Thus, the energy providers and infrastructure sector should be selected as a prior control target, and also for the fact that they have been identified as major driver and intermediate step of carbon emission. 3.2.3. Mutual Relationships. Figure 5 shows the direct and integral relationships between metabolic sectors of BDA Park based on SignD and SignU. Results indicate that the number of positive relationships were larger than negative ones both in direct and integral utilities, showing that the metabolic system tended to be organized to encourage mutually beneficial exchanges. Alterations of signs and values expressed in different colors were also significant. Higher synergism index (SI = 7.0) and mutualism index (MI = 2.1) in the integral utility matrix

economic technological development zone and a national hightech industrial development zone that has been accredited as a national demonstration eco-industrial park in 2011. About 159 high-tech enterprises embracing pillar industries of information and communication technology, high-tech commerce, bioengineering and medicine, and automobile manufacturing are housed in an area of 0.1735 km2. The carbon metabolic inventory is based on primary data provided by park managers. Metabolic processes are identified as (1) artificial processes including the extraction, transportation, exchange, consumption and emission of carbon embodied in materials and energy; and (2) natural processes mainly based on carbon absorption by plants. Carbon fluxes of all materials and fuels are as the embodied carbon based on the environmental input−output analysis.57

3. RESULTS 3.1. Carbon Metabolic Network Model for BDA Park. Seven metabolic sectors were recognized based on each function within the system boundary of BDA Park: (1) Energy Providers, energy suppliers and distributors providing electricity, natural gas, heating and transportation fuels; (2) Construction and Infrastructure Sector mainly including building maintenance and facility services (e.g., ventilations, elevators, fire equipment), providing physical support for the whole park; (3) Residential Housing Sector providing apartments and houses for the employees; (4) Industry and Business Sector for the production of high-end goods; (5) Waste Management Sector disposing sewage and solid wastes discharged from other sectors; (6) Landscaping providing open spaces according to esthetical considerations but also served as carbon sink; and (7) External Environment. Apart from these functional sectors, twenty-three direct metabolic pathways running through the network model were classified as (1) carbon extraction from the external environment to other sectors; (2) carbon exchange including material and power transfer among multiple sectors along the symbiotic chain, and the carbon absorption by photosynthesis of landscaping; and (3) carbon emission directly discharged into the external environment via industrial production, energy consumption, residential activities, and services. Hence, the carbon metabolism network model for BDA Park was built to allow tracing carbon flows within the specific industrial system as shown in Figure 3. 3.2. Metabolic Performances. 3.2.1. Metabolic Processes. Carbon fluxes within the carbon metabolism network model were illustrated in Figure 3 (and SI Table S2). Results show that the annual total system throughflow reached 64,306 t, which was dominated by processes of carbon extraction (especially from the external environment to energy providers, 26.0%), exchange (from energy providers to construction and infrastructure sector, 12.6%) and emissions (from industry and business and waste management sectors back to the environment, 11.4% and 10.2%, respectively). Specifically, energy providers (16 726 t) were the major source delivering embodied carbon in terms of fossil fuels and electricity from the external environment. It also significantly contributed to exchange processes (56.3%), comprising the main part of the entire metabolic system. It also shows that about 46.2% of TST was derived from exchange processes (as from energy providers to sectors of construction and infrastructure and industry and business),71 providing examples of how restructuring of the intersectoral symbiotic chain might be of great significance in 7259

DOI: 10.1021/es5056758 Environ. Sci. Technol. 2015, 49, 7254−7264

Article

Environmental Science & Technology

utility matrix (see Figure 5), but it changed to slight exploiting or mutualism through indirect connections. Interrelations for each sector were also strengthened as mutualism (+/+), indicating wide connections were the reason for selfpromotions, i.e. positive contribution in terms of providing higher degrees of metabolic mutualism to the overall system. However, negative alterations as competence (−/−) were observed as well, as industry sector and residential or waste management sectors should be sustained by energy providers to supply power, thus whose relationships were changed from (0/ 0) to (−/−). It implied the overlapping functions between these key emitters in grabbing carbon embodied resources. 3.2.4. Control and Dependence Conditions. Average control and dependence conditions of metabolic components are shown in Figure 6. On the dominant side, external environment and energy providers were chief controllers for BDA Park’s carbon metabolism with the highest average control capacity (26.9% and 26.0% respectively). Sectors of industry and business, construction and infrastructure, residential housing and waste management also show high influencing power on the case park, whose average control capacities reached 14.0%, 13.5%, 10.4%, and 8.7%, respectively. It also matches our previous findings on dominant flows of carbon extraction, exchange, and emissions, suggesting that the metabolic system was mainly controlled by the original supply (i.e., external environment), intermediate consumption (energy providers and infrastructure sector), and final emitters (industrial, housing, and waste sectors): the environment restricted the productivity of its downstream sectors by supplying accessible metabolic resources, whereas consumers also adjusted other sectors related to their consumption. An interesting observation is that proportions of control degree and sectors’ total system throughflow were quite similar implying that the more carbon flows through a sector, the more control this sector exerts upon the whole network. On the other hand, sectors relied on others to the same extent (14.3%) in terms of network dependence, reflecting that each sector is equally required by the others. 3.2.5. Overall Performance. The carbon metabolism of BDA Park was assessed by a series of system-wide indicators, as shown in Table 1. Comparisons between the network with and without the consideration of external environment were done to recognize external impacts for the whole system (SI Tables S2 and S3). Results show almost all indicators significantly decreased without considering the environment, verifying the strong influence external environment in our case study. More routes of carbon fluxes appeared in the complete metabolism network model, where sectors were more widely connected showing a higher link density and connectivity of the entire network, and thus shaping a higher FCI. Higher MI, SI, and CI also occurred in the complete model, indicating more positive impacts were organized by the external environment-related indirect connections as similarly found in natural ecosystems.59 However, the ISI and STI increased without the participation of the environment, showing that the participation of environment was significant in facing external risks and shaping a perfect metabolic hierarchy.

Figure 4. Metabolic structure of the carbon metabolism network model in BDA Park. The left side shows the trophic roles of all metabolic sectors in BDA Park. The line at the bottom shows the trophic constitution (rp, see eq 14) of each sector, representing the ratio of each sectors’ carbon fluxes relative to total system throughflow. For example, external environment (EE) is defined as producer, contributing 26.9% of total system carbon throughflow of the park.

Figure 5. Direct and integral utility of the carbon metabolism network model in BDA Park. Mutual relationships between two sectors are shown in different colors, whose direct and integral utilities were also isolated in two boxes. For example, two boxes located in the first line and second row represent the relationship between external environment (EE) and energy providers (EP). The left box in orange indicates that the direct mutual relationship of them was strong exploiting (±). While the right box in yellow showed that their integral relationship altered into slight exploiting (±) due to indirect connections.

also provide evidence for quantitative and qualitative changes of mutual relationships among metabolic sectors. Almost all variations tended to show changes toward positive values to benefit the whole system, confirming that indirect processes contributed more to the metabolic mutualism than direct processes.59 For instance, sectors of construction and infrastructure, industry and business, residential housing and waste management strongly utilized carbon inputs from their upstream sector (i.e., energy provider) as shown in the direct

4. DISCUSSION AND CONCLUSIONS The carbon metabolism of eco-industrial parks can be characterized and evaluated by direct/indirect interactions between connected components and pathways.53 In this context, the ENA-based approach is vital to unveil the 7260

DOI: 10.1021/es5056758 Environ. Sci. Technol. 2015, 49, 7254−7264

Article

Environmental Science & Technology

Figure 6. Average control and dependence degree of the carbon metabolism network model in BDA Park.

composing a linear and irreversible carbon supply chain. It suggests that the contemporary carbon metabolism in case park was not sustainable.71,73 As a result, supplying nonsustainable carbonic resources was seemingly the sole solution to support the daily operation of eco-industrial parks, where external environments played the key role. It was the external subsidy providing almost all original sources of carbon-contained resources and maintained and controlled other sectors. The unstable state with imperfect hierarchy based on NSA also reflects the conflict between environment-induced excessive efficiency and deficient resilience, which is a symbol of the unsustainable carbon metabolism. So it is crucial to balance the pairwise antagonists in metabolized parks to mitigate carbon emissions. In pursuit of emission mitigation, update of technologies and devices are suitable to enhance the efficiency of carbon utilization. As for BDA Park, applications of heat exchange system, solar photovoltaic cells and vertical gardens have been introduced to increase the heating efficiency, carbon absorption capacity and reduce the dependence on fossil fuels in situ. However, efficiency improvement is not the only solution to this issue: although productivity can be stimulated by advanced technologies and higher efficiencies, more dependence of environments’ supply and carbon emissions are foreseeable due to the current production and consumption schemes, and it will eventually cause higher environmental and social costs.73 For this purpose, network diversification (e.g., internalizing enterprises that undertake some functions of the environment, and designing more carbon exchange routes based on recycling and reusing waste) is equally important in enhancing system resilience and weakening the reliance on the external environment at the expense of efficiency loss. But it poses a new set of optimization problems: not only the practical barrier of incorporating newly designed sectors and processes into the metabolic network, but also the feasibility of the trade-off between optimizing costs and benefits and process choices.74 To some extent, this issue depends on whether firms inside the park are more efficient than those outside the park, namely whether the newly achieved stability is acceptable. The answer deserves deeper analysis in probing a better interpretation of metabolic stability in eco-industrial parks, where ascendancy analysis,75,76 another branch of network analysis, may provide a holistic perspective

mechanism and properties of the carbon metabolism in terms of mutual relationships, metabolic hierarchy, and control functions. Regarding the case of BDA Park, the results of the carbon metabolism network model show that indirect interactions were the largest contributors to total carbon emissions. The examination of network utility analysis (NUA) also provides evidence of sectoral interactions and behaviors, especially for those seemingly unrelated (e.g., construction and infrastructure was not directly connected with the waste sector in our case study, but in fact the waste sector always exploited waste from the construction sector via some intermediaries). This can help to make the transition to low-carbon industrial parks by redirecting supply chains, which can be achieved, for example, through connecting flows from the waste sector to industry and business by reusing residuals and so reducing carbon emissions. Creating new pathways from the waste sector and landscaping to energy providers would also contribute to achieve lowcarbon targets by decreasing the dependence of imported carbon. Network control analysis (NCA) illustrates another important aspect of the metabolic performance in terms of control and dependence capacities, which emphasizes the importance of combining top-down and bottom-up control in network systems.72 It would be feasible in determining key emitters and allocating goals of emission mitigation through the carbon-distribution chain in eco-industrial parks. Both sectors and system properties were investigated through the traditional ENA framework, revealing some intrinsic characteristics of the carbon metabolism in eco-industrial parks. In order to better regulate emission trajectories, the newly proposed tool network stability analysis (NSA) was introduced into the ENA framework providing insights into system hierarchy, structural stability, and metabolic dynamics. Unique roles of each sector in symbiotic webs were mimicked with trophic ecosystems,68 useful in defining major emitters, and understanding the metabolic hierarchy constructed upon diverse processes these sectors undertook. The stability analysis also depicts a future perspective of the model based on information theory, which indicates that the BDA Park was in the primary phase of system evolution to extract and store more resources from the environment, since metabolic processes in BDA Park were mainly driven by energy obtained from fossil fuels. These dominated the metabolic magnitude of the park, 7261

DOI: 10.1021/es5056758 Environ. Sci. Technol. 2015, 49, 7254−7264

Article

Environmental Science & Technology

(No. 51421065), Major Research Plan of the National Natural Science Foundation of China (No. 91325302), National Natural Science Foundation of China (No. 41271543), and Specialized Research Fund for the Doctoral Program of Higher Education of China (No. 20130003110027).

in balancing the efficiency-resilience contradictions appropriately. Compared with inventory accounting of carbon emissions in eco-industrial parks by other metabolic approaches,24,29,41,42 more detailed insights into the carbon metabolism are revealed by ENA. The life cycle perspective provides a detailed process evaluation (especially including activities of upstream supply chains) on carbon emissions,41 and recognizes the key emitter and processes in eco-industrial parks, while network utility analysis of ENA further identifies the indirect emitters by considering the cumulative effects due to several indirect interactions between metabolic sectors. For example, emissions in BDA Park can be partly traced back to energy providers and the construction and infrastructure sector, delivering materials for the production of high-end goods and supporting electricity to incinerate waste. The material flow-based inventory analysis shows that carbon flows in eco-industrial parks are dominated by energy-related emissions.29,42 By comparison, network control analysis of ENA reveals more emission controllers (environment and waste sector) by considering mutual interactions between metabolic sectors. Decomposition analysis allows attributing emissions to socioeconomic drivers (e.g., production effect of the park and surrounding regions),24 while ENA also specifies the interactions between all intrapark sectors and the external environment and their respective contributions. The newly proposed analytical tool network structural analysis of ENA also helps to clarify the trophic roles (producers, consumers and decomposers) of each sector and thus to shape the metabolic hierarchy and forecast system stability in the long term. It may contribute to improve the strategy and technology of eco-industrial parks by identifying the emission controller. Nevertheless, there is still more room to explore carbon metabolic measurements in eco-industrial parks. ENA is a static analysis where series of snapshot of carbon metabolism can be tracked,77 but it is insufficient to simulate metabolic dynamics. A metabolic database of different types of eco-industrial parks would enable a systematic evaluation and comparison of metabolic properties using the health assessment-integrated ENA framework.78 This study is a first step toward establishing such a framework based on ENA to unveil the mechanism and properties of the carbon metabolism of eco-industrial parks.

■

■

ASSOCIATED CONTENT

S Supporting Information *

Abbreviations; system description; network utility, control, and stability analysis; environmental input−output analysis; carbon flows in network model; carbon flows between park sectors; and additional references. The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/es5056758.

■

REFERENCES

(1) Moomaw, W. R. Industrial emissions of greenhouse gases. Energy Policy 1996, 24 (10−11), 951−968. (2) IEA (International Energy Agency). CO2 Emissions from Fuel Combustion Highlights, 2014 edition; IEA: Paris, 2014. (3) Chang, Y. F.; Lin, S. J. Structrual decomposition of industrial CO2 emission in Taiwan: an input-output approach. Energy Policy 1998, 26 (1), 5−12. (4) Worrell, E.; Price, L.; Martin, N.; Hendriks, C.; Meida, L. O. Carbon dioxide emissions from the global cement industry. Annu. Rev. Energy Env. 2001, 26, 303−329. (5) Jaber, J. O. Future energy consumption and greenhouse gas emissions in Jordanian industries. Appl. Energy 2002, 71 (1), 15−30. (6) Hondo, H. Life cycle GHG emission analysis of power generation systems: Japanese case. Energy 2005, 30 (11−12), 2042−2056. (7) Rehan, R.; Nehdi, M. Carbon dioxide emissions and climate change: policy implications for the cement industry. Environ. Sci. Policy 2005, 8 (2), 105−114. (8) Liu, L. C.; Fan, Y.; Wu, G.; Wei, Y. M. Using LMDI method to analyzed the change of China’s industrial CO2 emissions from final fuel use: An empirical analysis. Energy Policy 2007, 35 (11), 5892− 5900. (9) Al-Ghandoor, A.; Al-Hinti, I.; Jaber, J. O.; Sawalha, S. A. Electricity consumption and associated GHG emissions of the Jordanian industrial sector: Empirical analysis and future projection. Energy Policy 2008, 36 (1), 258−267. (10) Lee, C. F.; Lin, S. J.; Lewis, C. Analysis of the impacts of combining carbon taxation and emission trading on different industry sectors. Energy Policy 2008, 38, 722−729. (11) Yang, J.; Chen, B. Using LMDI method to analyze the change of industrial CO2 emission from energy use in Chongqing. Front. Earth Sci. 2011, 5 (1), 103−109. (12) IEA (International Energy Agency), CO2 Emissions From Fuel Combustion, 2013 edition; 2013. (13) Xia, X. H.; Huang, G. T.; Chen, G. Q.; Zhang, B.; Chen, Z. M.; Yang, Q. Energy security, efficiency and carbon emission of Chinese industry. Energy Policy 2011, 39 (6), 3520−3528. (14) The State Council Information Office of the People’s Republic of China, National Climate Change Plan (2014−2020), 2014. (15) Peters, G. P.; Weber, C. L.; Guan, D.; Hubacek, K. China’s growing CO2 emissionsA race between lifestyle changes and efficiency gains. Environ. Sci. Technol. 2007, 41, 5939−5944. (16) Guan, D.; Peters, G. P.; Weber, C. L.; Hubacek, K., Journey to world top emitter: An analysis of the driving forces of China’s recent CO2 emissions surge. Geophys. Res. Lett., 2009, 36; doi: 10.1029/ 2008GL036540. (17) Minx, J. C.; Baiocchi, G.; Peters, G. P.; Weber, C. L.; Guan, D.; Hubacek, K. Carbonizing dragon: China’s fast growing CO2 emissions revisited. Environ. Sci. Technol. 2011, 45 (21), 9144−9153. (18) Chertow, M. R. Industrial symbiosis: literature and taxonomy. Annu. Rev. Energy Environ. 2000, 25, 313−337. (19) Chertow, M. R. “Uncovering” industrial symbiosis. J. Ind. Ecol. 2007, 11 (1), 11−30. (20) Lobanova, G.; Fath, B. D.; Rovenskaya, E. Exploring simple structural configurations for optimal network mutualism. Commun. Nonlinear Sci. 2009, 14 (4), 1461−1485. (21) Hashimoto, S.; Fujita, T.; Geng, Y.; Nagasawa, E. Realizing CO2 emission reduction through industrial symbiosis: A cement production case study for Kawasaki. Resour. Conserv. Recy. 2010, 54 (10), 704− 710. (22) Maes, T.; Van Eetvelde, G.; De Ras, E.; Block, C.; Pisman, A.; Verhofstede, B.; Vandendriessche, F.; Vandevelde, L. Energy manage-

AUTHOR INFORMATION

Corresponding Author

*Tel.: +86-10-58807368; fax: +86-10-58807368; e-mail: [email protected] (B.C.). Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This work was supported by the Fund for Innovative Research Group of the National Natural Science Foundation of China 7262

DOI: 10.1021/es5056758 Environ. Sci. Technol. 2015, 49, 7254−7264

Article

Environmental Science & Technology ment on industrial parks in Flanders. Renew. Sust. Energy Re. 2011, 15 (4), 1988−2005. (23) Ministry of Industry and Information Technology of the People’s Republic of China; National Development and Reform Commission. The pilot list of national low-carbon industrial parks (the f irst batch). 2014. (24) Jung, S.; An, K. J.; Dodbiba, G.; Fujita, T. Regional energyrelated carbon emission characteristics and potential mitigation in ecoindustrial parks in South Korea: Logarithmic mean Divisia index analysis based on the Kaya identity. Energy 2012, 46 (1), 231−241. (25) Liu, L. X.; Zhang, B.; Bi, J.; Wei, Q.; Pan, H. The greenhouse gas mitigation of industrial parks in China: A case study of Suzhou industrial park. Energy Policy 2012, 46, 301−307. (26) Midthun, K.; Norstebo, V. S.; Perez-Valdes, G.; Bjorkvoll, T. Investment analysis of an integrated industrial park with carbon capture. J. Nat. Gas. Sci. Eng. 2012, 7, 44−51. (27) Wang, H. S.; Lei, Y.; Wang, H. K.; Liu, M. M.; Yang, J.; Bi, J. Carbon reduction potentials of China’s industrial parks: A case study of Suzhou industrial park. Energy 2013, 55, 668−675. (28) Jung, S.; Wang, L. P.; Dogbiba, G.; Fujita, T. Two-step accelerated mineral carbonation and decomposition analysis for the reduction of CO2 emission in the eco-industrial parks. J. Environ. Sci. 2014, 26, 1411−1422. (29) Liu, W.; Tian, J. P.; Chen, L. J. Greenhouse gas emissions in China’s eco-industrial parks: A case study of the Beijing Economic Technological Development Area. J. Clean. Prod. 2014, 66, 384− 39130. (30) Casavant, T. E.; Cote, R. P. Using chemical process simulation to design industrial ecosystems. J. Clean. Prod. 2004, 12, 901−908. (31) van Beers, D.; Corder, G.; Bossikov, A.; van Berkel, R. Industrial symbiosis in the Australian minerals industry. J. Clean Prod. 2007, 11 (1), 55−72. (32) van Berkel, R.; Fujita, T.; Hashimoto, S.; Fujii, M. Quantitative assessment of urban and industrial symbiosis in Kawasaki, Japan. Environ. Sci. Technol. 2009, 43, 1271−1281. (33) Jiang, W. L.; Yuan, Z. W.; Bi, J.; Sun, L. Conserving water by optimizing production schedules in dyeing industry. J. Clean Prod. 2010, 18, 1696−1702. (34) Geng, Y.; Zhang, P.; Ulgiati, S.; Sarkis, J. Emergy analysis of an industrial park: The case of Dalian, China. Sci. Total Environ. 2010, 408, 5273−5283. (35) Mattila, T. J.; Pakarinen, S.; Sokka, L. Quantifying the total environmental impacts of an industrial symbiosisA comparison of process-, hybrid, and input−output life cycle assessment. Environ. Sci. Technol. 2010, 44, 4309−4314. (36) Tian, J. P.; Shi, H.; Chen, Y.; Chen, L. J. Assessment of industrial metabolisms of sulfur in a Chinese fine chemical industrial park. J. Clean Prod. 2012, 32, 262−272. (37) Montastruc, L.; Boix, M.; Pibouleau, L.; Azzaro-Pentel, C.; Domenech, S. On the flexibility of an eco-industrial park (EIP) for managing industrial water. J. Clean Prod. 2013, 43, 1−11. (38) Tong, L.; Liu, X.; Liu, X. W.; Yuan, Z. W.; Zhang, Q. Life cycle assessment of water reuse systems in an industrial park. J. Environ. Manage. 2013, 129, 471−478. (39) Zhang, Y.; Zheng, H. M.; Yang, Z. F.; Liu, G. Y.; Su, M. R. Analysis of the industrial metabolic processes for sulfur in the Lubei (Shandong Province, China) eco-industrial park. J. Clean Prod. 2014, DOI: 10.1016/j.jclepro.2014.01.096. (40) Kantor, I.; Fowler, M.; Elkamel, A. Optimized production of hydrogen in an eco-park network accounting for life-cycle emissions and profit. Int. J. Hydrogen Energy 2012, 37 (6), 5347−5359. (41) Dong, H. J.; Geng, Y.; Xi, F. M.; Fujita, T. Carbon footprint evaluation at industrial park level: A hybrid life cycle assessment approach. Energy Policy 2013, 57, 298−307. (42) Tian, J.; Guo, Q.; Chen, Y.; Li, X.; Shi, H.; Chen, L. Study on industrial metabolism of carbon in a Chinese fine chemical industrial park. Environ. Sci. Technol. 2013, 47, 1048−1056. (43) Zhang, H.; Dong, L.; Li, H. Q.; Fujita, T.; Ohnishi, S.; Tang, Q. Analysis of low-carbon industrial symbiosis technology for carbon

mitigation in a Chinese iron/steel industrial park: A case study with carbon flow analysis. Energy Policy 2014, 61, 1400−1411. (44) Yu, C.; de Jong, M.; Dijkema, G. P. J. Process analysis of ecoindustrial park development - The case of Tianjin, China. J. Clean Prod. 2014, 64, 464−477. (45) Fath, B. D.; Patten, B. C. Quantifying resource homogenization using network flow analysis. Ecol. Model. 1999, 123 (2−3), 193−205. (46) Fath, B. D.; Scharler, U. M.; Ulanowicz, R. E.; Hannon, B. Ecological network analysis: Network construction. Ecol. Model. 2007, 208 (1), 49−55. (47) Li, Y.; Chen, B.; Yang, Z. F. Ecological network analysis for water use systems - A case study of the Yellow River Basin. Ecol. Model. 2009, 220 (22), 3163−317348. (48) Yang, Z. F.; Mao, X. F.; Zhao, X.; Chen, B. Ecological Network Analysis on Global Virtual Water Trade. Environ. Sci. Technol. 2012, 46 (3), 1796−1803. (49) Chen, S. Q.; Fath, B. D.; Chen, B. Information-based network environ analysis: A system perspective for ecological risk assessment. Ecol. Indic. 2011, 11 (6), 1664−1672. (50) Chen, S. Q.; Chen, B. Network environ perspective for urban metabolism and carbon emissions: A case study of Vienna, Austria. Environ. Sci. Technol. 2012, 46 (8), 4498−4506. (51) Dai, J.; Fath, B.; Chen, B. Constructing a network of the socialeconomic consumption system of China using extended exergy analysis. Renew. Sust. Energy Rev. 2012, 16 (7), 4796−4808. (52) Lu, W.; Su, M.; Zhang, Y.; Yang, Z.; Chen, B.; Liu, G. Assessment of energy security in China based on ecological network analysis: A perspective from the security of crude oil supply. Energy Policy 2014, 74, 406−413. (53) Smart, B. Industry as a metabolic activity. Natl. Acad. Sci. 1992, 89 (3), 804−806. (54) Wolman, A. The metabolism of cities. Sci. Am. 1965, 213, 179− 190. (55) Ayres, R. U.; Simonis, U. E. Industrial Metabolism, Restructuring for Sustainable Development; United Nations University Press: Tokyo, 1994. (56) Noorman, K. J. Green Households? Domestic, Consumers, Envrionment and Sustainability; Earthscan Publishments Ltd.: London, 1998. (57) Feng, K.; Siu, Y. L.; Guan, D.; Hubacek, K. Analyzing drivers of regional carbon dioxide emissions for China: A structural decomposition analysis. J. Ind. Ecol. 2012, 16 (4), 600−611. (58) Lenzen, M. Primary energy and greenhouse gases embodied in Australia final consumption: An input−output analysis. Energy Policy 1998, 26 (6), 495−506. (59) Fath, B. D.; Patten, B. C. Network synergism: Emergence of positive relations in ecological systems. Ecol. Model. 1998, 107 (2−3), 127−143. (60) Fath, B. D.; Patten, B. C. Review of the foundations of network environ analysis. Ecosystems 1999, 2, 167−179. (61) Fath, B. D. Network mutualism: Positive community-level relations in ecosystems. Ecol. Model. 2007, 208, 56−57. (62) Zhang, Y.; Yang, Z. F.; Yu, X. Y. Ecological network and emergy analysis of urban metabolic systems: Model development, and a case study of four Chinese cities. Ecol. Model. 2009, 220 (11), 1431−1442. (63) Fath, B. D. Distributed control in ecological networks. Ecol. Model. 2004, 179 (2), 235−245. (64) Patten, B. C.; Auble, G. T. System-Theory of the Ecological Niche. Am. Naturalist 1981, 117 (6), 893−922. (65) Ulanowicz, R. E., Ascendancy: A Measure of Ecosystem Performance. In Handbook of Ecosystem Theories and Management; Jorgensen, S. E., Muller, F., Eds.; CRC Press: Boca Raton, FL, 2000, 303−315. (66) MacArthur, R. Fluctuations of Animal Populations, and a Measure of Community Stability. Ecology 1955, 36 (3), 533−536. (67) Lindeman, R. L. The trophic-dynamic aspect of ecology. Ecology 1942, 23 (4), 399−418. (68) Hardy, C.; Graedel, T. E. Industrial systems as food webs. J. Ind. Ecol. 2002, 6 (1), 29−38. 7263

DOI: 10.1021/es5056758 Environ. Sci. Technol. 2015, 49, 7254−7264

Article

Environmental Science & Technology (69) Finn, F. T. Flow analysis of models of the Hubbard brook ecosystem. Ecology 1980, 61 (3), 562−571. (70) Fath, B. D.; Halnes, G. Cyclic energy pathways in ecological food webs. Ecol. Model. 2007, 208, 17−24. (71) Jelinski, L. W.; Graedel, T. E.; Laudise, R. A.; McCall, D. W.; Patel, C. K. Industrial ecology: concepts and apporaches. Natl. Acad. Sci. 1992, 89 (3), 793−797. (72) Polis, G. A. Why are parts of the world green? Multiple factors control productivity and the distribution of biomass. Oikos 1999, 86 (1), 3−15. (73) Dangerman, A. T. C. J.; Schellnhuber, H. J. Energy systems transformation. Natl. Acad. Sci. 2013, 110 (7), E549−E558. (74) Frosch, R. A. Industrial ecology: a philosophical introduction. Natl. Acad. Sci. 1992, 89 (3), 800−803. (75) Ulanowicz, R. E.; Kay, J. J. A package for the analysis of ecosystem flow networks. Environ. Soft. 1991, 6 (3), 131−142. (76) Heymans, J. J.; Ulanowicz, R. E.; Bondavalli, C. Network analysis of the South Florida Everglades graminoid marshes and comparison with nearby cypress ecosystems. Ecol. Model. 2002, 149 (1−2), 5−23. (77) Gattie, D. K.; Schramski, J. R.; Bata, S. A. Analysis of microdynamic environ flows in an ecological network. Ecol. Eng. 2006, 28 (3), 187−204. (78) Costanza, R. Toward an operational definition of ecosystem health. In Ecosystem Health: New Goals for Environmental Management; Costanza, R., Norton, B., Haskell, B. J., Eds.; Island Press: Washington, DC, 1992, 239−256, 269.

7264

DOI: 10.1021/es5056758 Environ. Sci. Technol. 2015, 49, 7254−7264