Nonzero-Sum Relationships in Mitigating Urban ... - ACS Publications

Sep 4, 2015 - A carbon metabolic network is constructed by modeling the carbon ... which is that cities have the ability to design advanced technologi...
0 downloads 3 Views 1MB Size
Download PDF
Article pubs.acs.org/est

Nonzero-Sum Relationships in Mitigating Urban Carbon Emissions: A Dynamic Network Simulation Shaoqing Chen, Bin Chen,* and Meirong Su State Key Joint Laboratory of Environmental Simulation and Pollution Control, School of Environment, Beijing Normal University, Beijing 100875, P R China S Supporting Information *

ABSTRACT: The “stove-pipe” way of thinking has been mostly used in mitigating carbon emissions and managing socioeconomics because of its convenience of implementation. However, systems-oriented approaches become imperative in pursuit of an efficient regulation of carbon emissions from systems as complicated as urban systems. The aim of this paper is to establish a dynamic network approach that is capable of assessing the effectiveness of carbon emissions mitigation in a more holistic way. A carbon metabolic network is constructed by modeling the carbon flows between economic sectors and environment. With the network shocked by interventions to the sectoral carbon flows, indirect emissions from the city are accounted for under certain carbon mitigation strategies. The nonzero-sum relationships between sectors and environmental components are identified based on utility analysis, which synthesize the nature of direct and indirect network interactions. The results of the case study of Beijing suggest that the stove-pipe mitigation strategies targeted the economic sectors might be not as efficient as they were expected. A direct cutting in material or energy import to the sectors may result in a rebound in indirect emissions and thus fails to achieve the carbon mitigation goal of the city as a whole. A promising way of foreseeing the dynamic mechanism of emissions is to analyze the nonzero-sum relationships between important urban components. Thinking cities as systems of interactions, the network approach is potentially a strong tool for appraising and filtering mitigation strategies of carbon emissions. energy use scenarios9,10 or how mitigation goals can be allocated to various economic sectors.11,12 Others developed ad-hoc models to simulate the effect of specific sectors (such as energy, industry, and transportation)13,14 or engineering15 on emissions control in a direct way. Systems dynamics models are also used in projecting future urban carbon emissions based on the initial sectoral emissions and dominant valuables through time.16 The stove-pipe manner of mitigation (i.e., an abrupt and simple cutting of certain flows) could fail when carbon emissions are coupled with other economic activities and ecological processes in an interactive way. For example, if the residential demand is assumed to be constant, decreasing the direct carbon that is imported to certain urban sectors will result in a compensatory increase in the inner industry or the growth of carbon import in other related sectors, in either way the carbon footprint of citizens augments. It has been recognized that relationships between socioeconomic compo-

1. INTRODUCTION Cities are deemed to be a central battlefield in the war against global warming. Approximately 70% of the energy consumption and CO2 emissions occurs in urban areas1,2 Moreover, the abundant consumption activities of urban energy facilities and residents drive the growth of carbon emissions globally.3 Great challenges will probably remain with regard to global warming mitigation for a long period of time in this century. There is an optimiztic side of the coin, which is that cities have the ability to design advanced technologies in efficient energy use and green infrastructure, which tend to be more affordable in cities than in rural areas.4 Additionally, the complexity of socioeconomics in cities have greater potential for optimization by adjusting the behaviors of the urban sectors that contribute to emissions growth.5,6 The aim of the study is to establish a new and holistic modeling tool for simulating the dynamic changes in emissions in response to the regulation of sectoral activities from a practical perspective, and therefore evaluating and filtering strategies for mitigating carbon footprints of cities,. There has been a vast body of literature that investigates the effectiveness of certain strategies and methods in cutting carbon emissions from cities. Some studies explored the generic potential of carbon emissions mitigation in urbanization7,8 and © 2015 American Chemical Society

Received: Revised: Accepted: Published: 11594

May 30, 2015 August 15, 2015 September 4, 2015 September 4, 2015 DOI: 10.1021/acs.est.5b02654 Environ. Sci. Technol. 2015, 49, 11594−11603

Article

Environmental Science & Technology

Figure 1. Technical framework and model structure for dynamic simulation of urban carbon emissions. Notes: Agr: Agriculture; Ind: Industry & commerce; Con: Construction; Tra: Transportation; Dom: Domestic; Was: Waste; Emi: Emissions; Loc: Local ecosystem; SC: Stock change; Ext: External ecosystems. The shown structure in the figure is not necessary to be the exact scheme of carbon metabolism for all cities (depending on the existence of their industries).

of socioeconomic behaviors. The network simulation can offer a holistic approach of emissions mitigation apart from stovepipe thinking: scrutinize the alteration of emissions within the carbon metabolism by simulating the dynamics of the urban metabolic system under different policy scenarios. With sufficient methodological development, ENA will possess the built-in feature of evaluating and optimizing options of emissions mitigation among sectors and looking for a multiwin low-carbon profile for cities. In this paper, using Beijing as a case study, we develop an equilibrium carbon metabolic network based on ENA, wherein carbon flows among economic sectors and between the urban economy and environment are encompassed. We then target carbon emissions to examine the changes that occur in response to the interventions on the imports to each sector in the network. The contribution of various economic sectors to indirect emissions under such interventions are simulated and quantified. In the end, the mechanism of the dynamics in urban emission trajectories is identified through the analysis of nonzero-sum relationships between urban components.

nents are key to managing urban systems, and neglecting them will result in suboptimal solutions.17−19 A clear definition and determination of interactions between socioeconomic components is prominent for the assessment and mitigation of carbon footprints.20,21 Modeling indirect emissions has been increasingly important because highemitting sectors often greatly rely on each other, such as energy industry and manufacture. Input-output analysis (IOA) proposed by Leontief22 has been widely used to quantify the indirect emissions that are embodied in the supply chains that cross the cities’ geographical boundaries.23,24 Hannon25 first applied IOA to modeling material flows in natural ecosystems. Patten and his colleagues proposed a form of network-oriented metrics called Ecological Network Analysis (ENA)26,27 to examine the indirect effects based on interactions between ecosystem components, and scholars have developed various ENA toolboxes and software ever since.28−33 Recently ENA has been reintroduced to the modeling of urban energy and carbon flows, given its ability to uncover the socioeconomic structure and functioning underlying nonzero-sum relationships between urban components.6,34 In ENA terms, “non-zero-sum relationship” refers to integral relations, summing them from all of the pathways from one sector to the other, which results in diverse situations for the sectors, for example, increasing both or decreasing both.35 In contrast to zero-sum relations that convey in adjacent interactions (direct effects), nonzero-sum relations explain a system’s dynamics originates from both direct and indirect effects. ENA has the potential of digging into the mechanism of emissions based on the indirect effect that are transmitted as nonzero-sum relationships between urban components, which has been shown useful in various ecological systems.36,37 The study of carbon flows of cities has been mostly concentrated in accounting for emissions from production or consumption perspective and analyzing the driving forces of emissions change through time.34,35 The dynamic mechanism of carbon emissions contingent on the nature of intercomponent relationships has not been given much attention, hindering decision makers from a more efficient and prescient regulation

2. MATERIALS AND METHODS 2.1. Technical Framework and Model Structure. Figure 1 shows the technical framework and model structure for the dynamic modeling of carbon emissions mitigation in cities. The modeling procedure is followed by a five-step analysis: (A) Developing a carbon metabolic network (CMN) based on local energy/material flow data. To acquire a holistic view of how emissions are connected and interact with other carbon flows within the urban system, all of the socioeconomic sectors and the natural environment should be considered in the network model. (B) Modeling the propagations of flows via ENA for both adjacent and nonadjacent interactions. All of the carbon flows between economic sectors and between the urban economy and environment are accounted for. This step quantifies the indirect emissions of sectors while they build their production and consumption activities with other sectors and the environment. (C) Establishing a set of scenarios for each sector based on policy strategies for emissions mitigation, 11595

DOI: 10.1021/acs.est.5b02654 Environ. Sci. Technol. 2015, 49, 11594−11603

Article

Environmental Science & Technology

whereas yi is carbon export from component i. The indirect flows of sectors can be quantified through eqs 3−5.

and simulating the changes in indirect carbon emissions from different sectors while responding to preset interventions to sectoral activities. (D) Analyzing the inherent mechanism for emissions trajectory dynamics. The nonzero-sum relationships between sectors are explored for their impact in emissions changes. (E) The outcome of emission trajectory changes and inherent mechanism will offer feedback to urban decision makers, and reevaluate and regulate the emissions mitigation strategies. Under the framework, a generic CMN is developed for tracing the dynamics of carbon flows in cities. The conceptual structure of CMN is represented by bidirectional interactions between economic sectors and the environment (Figure S1). The CMN consists of five aggregated economic sectors, including Agriculture (Agr), Industry & commerce (Ind), Construction (Con), Transportation (Tra), and Domestic (Dom), two sectors of environmental discharge, that is, solid and liquid waste (Was) and gaseous emissions (Emi), carbon stock change of sectors (SC), and a natural component, the local ecosystem within the city’s boundaries (Loc). In assessing the intersector relationships, the market and environment outside the city is defined as a distinct compartment (external ecosystems, Ext). This setting of sectors refers to the economy’s structure from the physical input-output table in the literature.38 Two environmental discharge components are separate from the system (Emi and Was) and are utilized for targeting the impact of other carbon flows on emissions trajectories. Here, carbon emissions refer to the carbon that is contained in CO2 and CH4. 2.2. Network Flow Modeling. The methodology of accounting for direct carbon flows between economic sectors was described in the work of Chen and Chen.6 The indirect carbon flows between urban components are modeled from direct flows in the CMN. ENA can be used to model how much energy or mass reaches one component from all of the pathways built by other components and itself. Throughflow analysis is the basis of all functional analyses based on ENA, which has been used and described in numerous studies for quantifying indirect flows of ecosystems.32,37,39 The urban carbon emissions are of primary concern in this specific study. Herein, the data of direct emissions is based on the hybrid boundary,6 wherein production-based and electricity-related emissions are considered. Throughflow analysis is utilized for modeling the indirect emissions from all components in the CMN with regard to both the original urban system and the systems that are subject to interventions. Throughflow is defined as the total of all flows in or out of component i in an n-compartment system.32,40 The sum of inflows of component i should equal to the sum of its outflows in a steady-state system eqs 1−3. n

Tiin ≡ zi +

Ti ≡ Tiout = Tiout

j=1

(3)

N(n × n) = G(n × n)0 + G(n × n)1 + G(n × n)2 + ... + G(n × n)m = (I(n × n) − G(n × n))−1 Emi

(4)

Emi F(̃ n × 1) = T(n × 1)·(N(Emi n × 1) − G(n × 1))

(5)

where N(n×n) is a dimensionless n × 1 matrix that sums up all of the direct and indirect interactions from one component to the other. gij is the fraction of output-oriented throughflow at donor Emi component i contributed to the focal component j. F(̃ n × 1) is a n × 1 vector of all the indirect flows to the component Emi (i.e., sectoral indirect emissions) from all of the other components in the CMN. T(n×1) is a n × 1 vector of total throughflows of all the components. NEmi (n × 1) is a dimensionless n × 1 vector that sums up all of the direct and indirect effects on Emi from the remainder of the network. GEmi (n × 1) is a dimensionless n × 1 vector of the proportional direct flows to Emi. 2.3. Network Dynamics Simulation. The network dynamics simulation consists of three stages: (I) Intervention to the carbon network. This stage is to select a set of carbon emissions mitigation strategies for specific economic sectors in the CMN and to translate these strategies into interventions to the carbon flows in the network (imports, exports or interflows). This is a primary process of examining how the urban system as a whole responds to the variations of flows. For a demonstrative investigation, we set the intensity of intervention to the economic sectors at the same level during the simulation (i.e., 10% fluctuation). (II) Rebalancing of the carbon network. The intervention of the sectoral flows will result in an unbalance between the inputs and outputs of sectors as well as the original carbon network. The network should be rebalanced for indirect emissions modeling and utility analysis. The input- and output-based approaches have been used in ENA for balancing unbalanced networks.41 These approaches may cause a major distortion of the results by changing the boundary and intermediate flows to a large extent. Therefore, we chose the RAS method (or called biproportional matrix balancing) for network balancing. It biproportionally scales a matrix with minimum unexpected changes to other flows.42 The RAS method has been widely applied and described in input-output analyses. 43,44 A introduction of the balancing process is provided in SI. (III) Simulation of emissions trajectory. The rebalanced networks that correspond to preset interventions are processed to perform network flow modeling. This method has been described in Section 2.3. The indirect emissions from all of the sectors will be simulated under all types of interventions to network flows, which are then compared to emissions from the original system to determine the effectiveness of certain mitigation strategies. The dynamics of the emissions trajectory of the whole city from a consumption perspective will also be simulated in response to the regulation of system flows.

n

∑ f ji , Tiout ≡ yi + ∑ fih j=1

G(n × n) = [gij], gij = fij /Ti

(1) (2)

where i, j, h = 1,2,..., n. (i, j, and h are different components within the carbon network model, and n is the number of components). Tiin is the input-oriented throughflow of component i, and Tout the output-oriented throughflow of i component i. Ti is the throughflow of component i when the system is at steady state.f ij is the carbon flow from component i to component j. zi is the carbon import to component i, 11596

DOI: 10.1021/acs.est.5b02654 Environ. Sci. Technol. 2015, 49, 11594−11603

Article

Environmental Science & Technology Table 1. Carbon Metabolic Network (CMN) of Beijing in 2010 (Unit: Mt)a Agr f% Ind f% Con f% Tra f% Dom f% Was f% Emi f% Loc f% SC f% Import f% a

Agr

Ind

Con

Tra

Dom

Was

Emi

Loc

SC

export

0.00 0.0% 0.45 0.2% 0.28 0.1% 0.00 0.0% 0.00 0.0% 0.00 0.0% 0.00 0.0% 1.51 0.7% 0.00 0.0% 4.97 2.2%

0.86 0.4% 0.00 0.0% 1.86 0.8% 0.00 0.0% 0.00 0.0% 0.00 0.0% 0.00 0.0% 2.16 0.9% 0.00 0.0% 20.07 8.7%

0.02 0.0% 1.16 0.5% 0.00 0.0% 0.02 0.0% 0.00 0.0% 0.00 0.0% 0.00 0.0% 1.91 0.8% 0.00 0.0% 13.67 5.9%

0.00 0.0% 0.55 0.2% 0.37 0.2% 0.00 0.0% 0.00 0.0% 0.00 0.0% 0.00 0.0% 0.17 0.1% 0.00 0.0% 14.82 6.4%

0.91 0.4% 1.92 0.8% 1.68 0.7% 0.05 0.0% 0.00 0.0% 0.00 0.0% 0.00 0.0% 1.18 0.5% 0.00 0.0% 13.44 5.8%

0.36 0.2% 1.87 0.8% 3.88 1.7% 0.79 0.3% 1.68 0.7% 0.00 0.0% 0.00 0.0% 0.00 0.0% 0.00 0.0% 0.00 0.0%

2.75 1.2% 14.82 6.4% 4.21 1.8% 13.20 5.7% 12.50 5.4% 0.00 0.0% 0.00 0.0% 0.00 0.0% 0.00 0.0% 0.00 0.0%

0.20 0.1% 0.84 0.4% 1.02 0.4% 0.06 0.0% 0.93 0.4% 0.30 0.1% 0.23 0.1% 0.00 0.0% 4.22 1.8% 1.31 0.6%

1.26 0.5% 1.42 0.6% 2.97 1.3% 1.45 0.6% 4.08 1.8% 0.00 0.0% 0.00 0.0% 1.26 0.5% 0.00 0.0% 0.00 0.0%

0.84 0.4% 1.93 0.8% 0.50 0.2% 0.33 0.1% 0.00 0.0% 8.29 3.6% 47.25 20.6% 0.92 0.4% 8.22 3.6% 0.00 0.0%

f% is the proportion of specific carbon flow to the total system throughflow (the sum of the Ti for all of the components).

2.4. Network Utility Analysis. Utility analysis (UA) is introduced for analyzing the nonzero-sum relationships between urban components. Direct interaction is presented with a dimensionless direct utility matrix D = [dij] and with consideration of indirect flow pathways from one to the other (pathway length >1), and a dimensionless integral utility matrix (U) is developed eqs 6-8.32,40,45 The integral utility matrix is used for identifying possible mechanisms of emissions change in cities, wherein the urban components have positive or negative impact with each other and cause the dynamics of urban carbon metabolism. D(n × n) = d ij ≡

mitigation. Beijing is the capital city of China, and the authority has been moving the heavy-polluted industries away from the city and transforming the economy into a low-carbon profile. The effect of the mitigation strategies need to be measured with proper tools. Both the activities of the sectors within the urban geographical boundary and the material exchanges between the city and remainder of the world (such as imports and exports) were considered. The system boundary is consistent with the work in Chen and Chen,6 which covers all of the socioeconomic processes that are driven by carbon metabolism in the city. The base carbon flow model investigated here is based on the situation of Beijing in 2010 (detailed data compilation is provided in Supporting Information). To observe how different methods of regulating sectors will impact indirect emissions of the city, four types of intervention are designated (S1, S2, S3, and S4) according to the adaptive response of flows through the sector (Table S1). In all types of intervention, the external import to the sector is reduced by 10%, but differently, in S1 only the Ti of this sector changes while no any changes occur to other carbon outflows, in S2 the intervention only causes the changes in the sector’s emissions and waste, in S3 it causes the changes in the sector’s all carbon outflows except emissions and waste, while in S4 it causes the changes in all carbon outflows of the intervened sector. The five economic sectors that undergo such interventions are Agr, Ind, Con, Tra, and Dom. The 20 scenarios are labeled as (Agr S1, Agr S2, Agr S3, Agr S4), (Ind S1, Ind S2, Ind S3, Ind S4), (Con S1, Con S2, Con S3, Con S4), (Tra S1, Tra S2, Tra S3, Tra S4) and (Dom S1, Dom S2, Dom S3, Dom S4).

(fij − f ji ) Tj

(6)

U(n × n) ≡ [uij] = D(n × n)0 + D(n × n)1 + D(n × n)2 + ... + D(n × n)m = (I(n × n) − D(n × n))−1

(7)

SigU(n × n) ≡ [sign(uij)]

(8)

where U accounts for interflows over all of the pathways in the system of all accessible lengths (1, 2, ..., m). The pathway length is the number of arrows from an initial sector to a terminal sector. sign(uij) indicates the positive (+) or negative (−) signs of uij, which is used to assess the nature of nonzero-sum (integral) relationships between sectors (Table S1). Nonzerosum relationships have been widely used in political, economic, and cultural contexts to describe a nonmechanical system. There is no simple win−lose (or increase−decrease) situation for the performances among the system components. Often a win−win (increase−increase) or lose−lose (decrease−decrease) situation can occur as a result of the complex pairwise relationships between these components. Herein, the dynamics of the components is identified through the integral utility matrix (U) (see an example in Figure S1). 2.5. Case Study and Scenarios Setting. In this study, Beijing was used as the case study to exemplify the application of a dynamic network simulation approach to carbon emissions

3. RESULTS 3.1. Carbon Metabolism of the City. Table 1 shows the carbon metabolic network of Beijing in the investigated year, including the flows between urban economic and ecological components. The most dominant flows are imports to economic sectors from external markets and ecosystems. The sum of all of the imports into the urban economy is 68.28 Mt (of carbon, the same as below), which accounts for 30% of the 11597

DOI: 10.1021/acs.est.5b02654 Environ. Sci. Technol. 2015, 49, 11594−11603

Article

Environmental Science & Technology

Figure 2. Alterations in the carbon metabolic network during simulation via the RAS method. Notes: Each box is the matrix of the carbon metabolic network under different scenarios, with flows that are represented by quadrate spots. Light-up spots are the altered flows that are different from the original network based on the value of ω% (the larger the change the brighter the spot, see SI for details), whereas the spots that do not have arrows pointed at them are possible noise for the simulation (changes in the RAS balancing process).

The changes of indirect emissions of each sector in various simulation scenarios are compared to the base system of Beijing, as displayed in Figure 3 (The alterations in the imports (preset and RAS-balanced imports) and throughflows (Ti) of the sectors in each scenario are provided in Table S3). The results suggest that there are clear patterns for the total indirect emission changes that emerge from all of the scenarios. The type of intervention to the carbon network has an almost decisive impact on the emissions trajectory. The two simulation types, S1 and S3, always result in increases in the emissions, while the other two (S2 and S4) tend to decrease the emissions regardless of which sector receives the intervention during the simulation. Reducing the import to each sector by 10% while accordingly changing the metabolic activeness (i.e., the sectoral throughflow) actually augments the indirect emissions from the city. For example, reducing the import to the industry and commerce sector by 10% (Ind S1) results in a 1.8% increase (or 1.19 Mt) in indirect emissions for the city. Similarly, for other sectors (Agr S1, Dom S1, Tra S1, and Con S1), the indirect emissions increase to different extents (from 0.4% to 1.4%). One of the plausible reasons is that the imports to the other sectors have to increase to meet the constant final demand for carbon in a short period. Another cause could be that the change in the proportional flows that take part in the pathways end in emissions, making the “stove-pipe” mitigation strategy useless. The other type of intervention (S3) changes the throughflow on both the input and output sides, which means that we assume that other sectors respond by reducing their consumption from the sector that received the intervention. Different from what may be expected, the indirect emissions still augment under this scenario, even to a slightly larger degree for some economic sectors. For example, the increment in the indirect emissions caused by the change in Ind (Ind S3) is approximately 9% (1.22 Mt) compared to the base system, while that in Dom S3 is 1.2% (0.81 Mt). It occurs that the responsive changes in the other carbon flows compromise

total carbon throughflow of the system. The largest part of the imports of the city goes to the industry and commerce sector (Ind), while the import to agriculture (Agr) is the lowest. In comparison with external import, the supply of carbon from local ecosystems is relatively poor ( uext,ems). A compromise between the interactions also occurs in S3, where (Agr, Emi) and (Ext, Emi) are at play because the import to and activities in Agr are adjusted. The indirect emissions from the city increase given that Ext has a stronger relation with Emi (higher utility). All of the three relations are at play with regard to S4, in which both Agr and Emi decrease the indirect emissions while Ext has a negative effect for the emissions mitigation. Eventually, the city’s indirect emissions diminish given the larger utility combinations (Agr, Emi) and (Emi, Emi) compared with the emissions in (Ext, Emi). Though the nature of nonzero-sum relationships between the economic sectors and Emi can be different, the inherent mechanism that underlies the relationships resemble that of Agr. This explains why indirect emissions of all sectors change similarly under the same type of intervention (e.g., Agr S1, Ind S1, Tra S1, Con S1, and Dom S1).

identified through a more holistic view of the interactions that underlie urban carbon metabolism. Figure 4 shows the

Figure 4. Nonzero-sum relationships between the sectors that are represented by integral utilities in the carbon metabolic network (the numbers are the utilities from integral utility matrix U).

intensities of integral utilities between all of the components within the urban system. Direct and integral utilities of CMN are assessed based on utility analysis (UA) to address the relationships between the sectors. The difference of interbidirectional signs between direct and integral utility matrices suggests that the nature of nonzero-sum relationships between the sectors is significantly distinct from zero-sum relationships (provided in Table S4). This is why one should look into the nature of nonzero-sum relationships for uncovering real mechanism of carbon metabolism. The nature of nonzerosum relationships between one another includes exploitation (such as between Ind and Loc), competition (such as between Agr and Dom) and mutualistic relation (such as between Agr and Ext). Figure 5 shows how the indirect emissions from the city are influenced by the nonadjacent interactions of Emi with different sectors in the urban system that is subjected to intervention. Only two types of nonzero-sum relationships exist between urban sectors and Emi: exploitation (for Ind in Figure 5b, Tra in Figure 5e, and Dom in Figure 5f) and mutualistic relation

Figure 5. Inherent mechanism of carbon emissions changes based on the intervention to (a) Agr, (b) Ind, (c) Con, (d) Tra, and (e) Dom, building on intersector nonzero-sum relationships. Note: the numbers are the sums of integral utility values between components in the CMN. 11600

DOI: 10.1021/acs.est.5b02654 Environ. Sci. Technol. 2015, 49, 11594−11603

Article

Environmental Science & Technology

proach.10,21,56 This study provides a technique of tracking the dynamics of emission via the modeling of carbon flows. The nonzero-sum relationships between urban components of the carbon metabolic network can be explicitly revealed by utility analysis (UA). UA has been used for recognizing sectoral linkages and economic structure.34,57 This study extends the application of UA to identifying the inherent mechanism of carbon metabolism. UA is found to be useful in explaining why the emissions will grow or diminish under certain mitigation strategies, though it still needs verification as a common rule for controlling the dynamics of indirect emissions from cities. One of the best features of dynamic network simulation is that it enables the preliminary prediction of whether the emissions of specific sectors decrease or increase even before the quantitative simulation is performed. The identification of nonzero-sum relationships can be applied to searching proper solutions for emissions control regarding various economic sectors constituting the city. The current network approach is limited to one-year modeling of cities, which can be enhanced by a time-series simulation of changes in emission structure driven by the dynamics of carbon metabolism. Another future step could be using more realistic scenarios in compliance with national and local mitigation goals for sectors to better inform the low-carbon policy for cities.

4. DISCUSSION Cities are the biggest contributor to global warming, and also the most promising filed of achieving sustainable and lowcarbon economy.46 Developing efficient tools for accounting for and evaluating the carbon profiles of cities is a primary task for urban carbon management.4 There has been numerous studies that account for the carbon emissions within urban territory7,8 as well as embodied in the whole supply chain.47,48 The modeling of socioeconomic flows associated with carbon emissions will provide important references for the mitigation of carbon footprint of cities and their economic sectors.49 In this study, a dynamic network approach is established for examining the changes of indirect carbon emissions of the city with regards to various mitigation strategies. Afterward, we also attempt to match the changes of emissions with the inherent mechanism underlie nonzero-sum relationships within the network. The simulation results of the case study of Beijing suggest that the “stove-pipe” cutting of the consumptive activities of sectors may not be efficient for mitigating emissions as one may think it is. The sole reduction of production activities and supply abilities of one economic sector will cause a rebound in other sectors and thus raise the total emissions of the city, while cutting the supply chain of one sector from imports to direct emissions will receive less resistance from other sectors. It is clear that focusing on the control of sectoral emissions and waste appears to be more efficient than shrinking the whole sector and its activities in consumption-based emissions mitigation. Though there are two types of nonzerosum relationship at play between the economic sectors and emissions (exploitation and mutualistic relation), the inherent mechanism that underlies the network interactions is quite similar. They provide explanations of why indirect emissions of sectors change in the same direction under the certain type of intervention. The network simulation clearly indicates that it is crucial to encompass the modeling of all the carbon flows in the urban metabolism for better evaluating the effectiveness of carbon mitigation strategies. Scholars have proposed the accounting of energy consumption and carbon emissions under the framework of urban metabolism.50−52 Assessing the metabolism of carbon provides a broader view of how to regulate the detrimental emissions from cities.5,6 Input-output analysis (IOA) is capable of accounting for the indirect emissions from an urban economy and its different sectors, as driven by the final demands of the cities.24,53 Compared to IOA, the dynamic network simulation based on ENA can be more flexible in examining how the alterations of specific carbon flows (and changes in the economic structure) result in changes in indirect emissions trajectories. ENA can even work out which sectors contribute the most to such changes. It has been clear that efficient urban carbon governance is achieved by determining which factors impact urban carbon mitigation and to what extent.7 The network simulation approach offers a holistic tool for evaluating the effectiveness of strategies for carbon emissions mitigation, and it is potentially applicable to addressing various environmental footprints of cities from a systems perspective. Another important aspect of carbon mitigation is the regulation of factors or components that drive the growth of emissions. A variety of methods have been used to locate the main factors that cause the emissions increase of cities, including land-use models,3 systems decomposition analysis,54,55 systems dynamics analysis15 and network ap-

■

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.est.5b02654. Details in methods (Table S1, Table S2 and Figure S1), description of data, and supporting details of model results (Figure S2, Table S3 and Table S4) (PDF)

■

AUTHOR INFORMATION

Corresponding Author

*Phone: +86−10−58807368; fax: +86−10−58807368; e-mail: [email protected]. 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 (No. 51421065), Major Research Plan of the National Natural Science Foundation of China (No. 91325302), the China Sustainable Energy Program of Energy Foundation (G-140721749), National Natural Science Foundation of China (No. 41371482), National Natural Science Foundation of China (No. 41271543), Specialized Research Fund for the Doctoral Program of Higher Education of China (No. 20130003110027), and China Postdoctoral Science Foundation funded project.

■

REFERENCES

(1) IEA. World Energy Outlook 2012; International Energy Agency (IEA): Paris, 2012. (2) GEA. Global Energy Assessmenttoward a Sustainable Future; Cambridge University Press: Cambridge, UK, 2012. (3) Seto, K. C.; Güneralp, B.; Hutyra, L. R. Global forecasts of urban expansion to 2030 and direct impacts on biodiversity and carbon pools. Proc. Natl. Acad. Sci. U. S. A. 2012, 109 (40), 16083−16088.

11601

DOI: 10.1021/acs.est.5b02654 Environ. Sci. Technol. 2015, 49, 11594−11603

Article

Environmental Science & Technology (4) Rosenzweig, C.; Solecki, W.; Hammer, S. A.; et al. Cities lead the way in climate-change action. Nature 2010, 467, 909−911. (5) Kennedy, C.; et al. Greenhouse gas emissions from global cities. Environ. Sci. Technol. 2009, 43, 7297−7302. (6) 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, 4498−4506. (7) Dhakal, S. GHG emissions from urbanization and opportunities for urban carbon mitigation. Current Opinion in Environmental Sustainability 2010, 2 (4), 277−283. (8) Satterthwaite, D. Cities’ contribution to global warming: notes on the allocation of greenhouse gas emissions. Environment and Urbanization 2008, 20 (2), 539−549. (9) Dhakal, S. Urban energy use and carbon emissions from cities in China and policy implications. Energy Policy 2009, 37 (11), 4208− 4219. (10) Chen, S. Q.; Chen, B. Urban energy consumption: different insights from energy flow analysis, input-output analysis and ecological network analysis. Appl. Energy 2015, 138, 99−107. (11) Su, M. R.; Liang, C.; Chen, B.; Chen, S. Q.; Yang, Z. F. Lowcarbon development patterns: observations of typical Chinese cities. Energies 2012, 5 (2), 291−304. (12) Wang, J.; Su, M.; Chen, B.; Chen, S. Q.; Liang, C. A comparative study of Beijing and three global cities: A perspective on urban livability. Front. Earth Sci. 2011, 5 (3), 323−329. (13) He, D.; Liu, H.; He, K.; et al. Energy use of, and CO2 emissions from China’s urban passenger transportation sector−Carbon mitigation scenarios upon the transportation mode choices. Transportation Research Part A: Policy and Practice 2013, 53, 53−67. (14) Lin, T.; Gibson, V.; Cui, S.; et al. Managing urban nutrient biogeochemistry for sustainable urbanization. Environ. Pollut. 2014, 192, 244−250. (15) Chen, B.; Chen, S. Q. Life cycle assessment of coupling household biogas production to agricultural industry: A case study of biogas-linked persimmon cultivation and processing system. Energy Policy 2013, 62, 707−716. (16) Feng, Y. Y.; Chen, S. Q.; Zhang, L. X. System dynamics modeling for urban energy consumption and CO2 emissions: A case study of Beijing, China. Ecol. Modell. 2013, 252, 44−52. (17) Brown, H. Next Generation Infrastructure; Island Press: Washington, DC, 2014. (18) Liu, J.; Mooney, H.; Hull, V.; et al. Systems integration for global sustainability. Science 2015, 347, 1258832. (19) Xu, M.; Weissburg, M.; Newell, J. P.; et al. Developing a science of infrastructure ecology for sustainable urban systems. Environ. Sci. Technol. 2012, 46 (15), 7928−7929. (20) Peters, G. P. Carbon footprints and embodied carbon at multiple scales. Current Opinion in Environmental Sustainability 2010, 2 (4), 245−250. (21) Chen, S. Q.; Chen, B.; Fath, B. D. Urban ecosystem modeling and global change: Potential for rational urban management and emissions mitigation. Environ. Pollut. 2014, 190, 139−149. (22) Leontief, W. W. The Structure of American Economy, 1919−1939: An Empirical Application of Equilibrium Analysis. Oxford University Press, New York, 1951. (23) Feng, K.; Davis, S. J.; Sun, L.; Li, X.; Guan, D.; Liu, W.; Liu, Z.; Hubacek, K. Outsourcing CO2 within China. Proc. Natl. Acad. Sci. U. S. A. 2013, 110 (28), 11654−11659. (24) Minx, J.; Baiocchi, G.; Wiedmann, T.; Barrett, J.; Creutzig, F.; Feng, K.; Förster, M.; Pichler, P.; Weisz, H.; Hubacek, K. Carbon footprints of cities and other human settlements in the UK. Environ. Res. Lett. 2013, 8 (3), 035039. (25) Hannon, B. The structure of ecosystems. J. Theor. Biol. 1973, 41, 535−546. (26) Patten, B. C. Systems approach to the concept of environment. Ohio. J. Sci. 1978, 78, 206−222. (27) Patten, B. C.; Auble, G. T. System theory of the ecological niche. Am. Nat. 1981, 117, 893−922.

(28) Ulanowicz, R. E.; Kay, J. J. A package of the analysis of ecosystem flow networks. Environmental Software 1991, 6 (3), 131− 142. (29) Christensen, V.; Pauly, D. ECOPATH IIda software for balancing steady-state ecosystem models and calculating network characteristics. Ecol. Modell. 1992, 61 (3−4), 169−185. (30) Allesina, S.; Bondavalli, C. WAND: an ecological network analysis user friendly tool. Environmental Modelling & Software 2004, 19 (4), 337−340. (31) Kazanci, C. EcoNet, a new software for ecological model simulation and network analysis. Ecol. Modell. 2007, 208 (1), 3−8. (32) Schramski, J. R.; Kazanci, C.; Tollner, E. W. Network environ theory, simulation, and EcoNet® 2.0. Environmental Modelling & Software 2011, 26 (4), 419−428. (33) Borrett, S. R.; Lau, M. K. enaR: An R package for ecosystem network analysis. Methods in Ecology and Evolution 2014, 5 (11), 1206−1213. (34) Zhang, Y.; et al. Analysis of an urban energy metabolic system: comparison of simple and complex model results. Ecol. Modell. 2011, 223 (1), 14−19. (35) Patten, B. C. Systems ecology and environmentalism: Getting the science right. Ecological Engineering 2013, 61, 446−455. (36) 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. (37) Salas, A. K.; Borrett, S. R. Evidence for the dominance of indirect effects in 50 trophic ecosystem networks. Ecol. Modell. 2011, 222 (5), 1192−1204. (38) Mulalic, I. Material Flows and Physical Input-Output Tables-PIOT for Denmark 2002 based on MFA; Statistics Denmark, 2007. (39) Fath, B. D.; Borrett, S. R. A Matlab® function for network environ analysis. Environmental Modelling & Software 2006, 21 (3), 375−405. (40) Fath, B. D.; Patten, B. C. Review of the foundations of network environ analysis. Ecosystems 1999, 2, 167−179. (41) Allesina, S.; Bondavalli, C. Steady state of ecosystem flow networks: a comparison between balancing procedures. Ecol. Modell. 2003, 165 (2), 221−229. (42) Deming, W. E.; Stephan, F. F. On least-squares adjustments of a sampled frequency table when the expected marginal total are known. Ann. Math. Stat. 1940, 11, 427−444. (43) Lenzen, M.; Wood, R.; Gallego, B. RAS Matrix Balancing under Conflicting Information. Intermediate Input−Output Meetings, 2006. (44) Junius, T.; Oosterhaven, J. The solution of updating or regionalizing a matrix with both positive and negative entries. Economic Systems Research 2003, 15, 87−96. (45) Fath, B. D.; Patten, B. C. Network synergism: emergence of positive relations in ecological systems. Ecol. Modell. 1998, 107, 127− 143. (46) Grimm, N. B.; et al. Global change and the ecology of cities. Science 2008, 319, 756−760. (47) Chen, G. Q.; et al. Three-scale input−output modeling for urban economy: carbon emission by Beijing 2007. Communications in Nonlinear Science and Numerical Simulation 2013, 18 (9), 2493−2506. (48) Lenzen, M. Primary energy and greenhouse gases embodied in Australian final consumption: an input-output analysis. Energy Policy 1998, 26 (6), 495−506. (49) Churkina, G.; Brown, D. G.; Keoleian, G. Carbon stored in human settlements: the conterminous United States. Global Change Biology 2010, 16, 135−143. (50) Kennedy, C.; Cuddihy, J.; Engel-Yan, J. The changing metabolism of cities. J. Ind. Ecol. 2007, 11 (2), 43−59. (51) Conke, L. S.; Ferreira, T. L. Urban metabolism: Measuring the city’s contribution to sustainable development. Environ. Pollut. 2015, 202, 146−152. (52) Newell, J. P.; Cousins, J. J. The boundaries of urban metabolism: Towards a political−industrial ecology. Progress in Human Geography 2014, 0309132514558442. 11602

DOI: 10.1021/acs.est.5b02654 Environ. Sci. Technol. 2015, 49, 11594−11603

Article

Environmental Science & Technology (53) Wang, Z.; Yang, L. Indirect carbon emissions in household consumption: evidence from the urban and rural area in China. J. Cleaner Prod. 2014, 78, 94−103. (54) Peters, G. P.; Weber, C. L.; Guan, D.; et al. China’s growing CO2 emissions a race between increasing consumption and efficiency gains. Environ. Sci. Technol. 2007, 41 (17), 5939−5944. (55) Zhu, Q.; Peng, X.; Wu, K. Calculation and decomposition of indirect carbon emissions from residential consumption in China based on the input−output model. Energy Policy 2012, 48, 618−626. (56) Chen, B.; Chen, S. Q. Eco-indicators for urban metabolism. Ecol. Indic. 2014, 47, 5−6. (57) Liu, G. Y.; Yang, Z. F.; Chen, B.; et al. Ecological network determination of sectoral linkages, utility relations and structural characteristics on urban ecological economic system. Ecol. Modell. 2011, 222 (15), 2825−2834.

11603

DOI: 10.1021/acs.est.5b02654 Environ. Sci. Technol. 2015, 49, 11594−11603