Transnational city carbon footprint networks – Exploring carbon links between Australian and Chinese cities

Applied Energy xxx (2016) xxx–xxx

Contents lists available at ScienceDirect

Applied Energy journal homepage: www.elsevier.com/locate/apenergy

Transnational city carbon footprint networks – Exploring carbon links between Australian and Chinese cities Guangwu Chen a, Thomas Wiedmann a,b, Yafei Wang c,⇑, Michalis Hadjikakou a a

Sustainability Assessment Program (SAP), UNSW Water Research Centre, School of Civil and Environmental Engineering, UNSW Australia, Sydney, NSW 2052, Australia Integrated Sustainability Analysis (ISA), School of Physics A28, The University of Sydney, NSW 2006, Australia c School of Statistics, Beijing Normal University, Beijing 100875, People’s Republic of China b

h i g h l i g h t s
A trans-national, multi-region input-output analysis for cities is presented.
We examine the carbon footprint network of ten cities.
The balance of emissions embodied in trade discloses a hierarchy of responsibility.
We model how emissions reductions spread through the city carbon networks.
Implications on the Chinese and Australian carbon trading schemes are discussed.

a r t i c l e

i n f o

Article history: Received 5 April 2016 Received in revised form 6 August 2016 Accepted 9 August 2016 Available online xxxx Keywords: Cities Carbon footprint Carbon trading City carbon footprint network City carbon map Multi-region input-output modelling

a b s t r a c t Cities are leading actions against climate change through global networks. More than 360 global cities announced during the 2015 Paris Climate Conference that the collective impact of their commitments will deliver over half of the world’s urban greenhouse gas emissions reductions by 2020. Previous studies on multi-city carbon footprint networks using sub-national, multi-region input-output (MRIO) modelling have identified additional opportunities for addressing the negative impacts of climate change through joint actions between cities within a country. However, similar links between city carbon footprints have not yet been studied across countries. In this study we focus on inter-city and inter-country carbon flows between two trading partners in a first attempt to address this gap. We construct a multi-scale, global MRIO model to describe a transnational city carbon footprint network among five Chinese megacities and the five largest Australian capital cities. First, we quantify city carbon footprints by sectors and regions. Based on the carbon map concept we show how local emissions reductions influence other regions’ carbon footprints. We then present a city emissions ’outsourcing hierarchy’ based on the balance of emissions embodied in intercity and international trade. The differences between cities and their position in the hierarchy emphasize the need for a bespoke treatment of their responsibilities towards climate change mitigation. Finally, we evaluate and discuss the potentially significant benefits of harmonising and aligning China’s carbon trading schemes with Australia’s cap and trade policy. Ó 2016 Elsevier Ltd. All rights reserved.

1. Introduction Major cities around the world are concentrating their efforts to tackle climate change through global networks. The Compact of Mayors was launched under the leadership of the world’s global city networks (C40,1 ICLEI,2 UCLG3), with support from ⇑ Corresponding author. 1 2 3

E-mail address: [email protected] (Y. Wang). Cities Climate Leadership Group (C40). ICLEI – Local Governments for Sustainability (ICLEI). United Cities and Local Governments (UCLG).

UN-Habitat, the UN’s lead agency on urban issues. The potential reductions of urban greenhouse gas (GHG or ‘‘carbon”) emissions of the Compact of Mayors initiative are up to 3.7 gigatons (Gt) annually by 2030 [1]. More than 360 global cities announced at the Paris Conference that the collective impact of their commitments will deliver over half of the world’s urban emissions reductions by 2020 [2]. Previous studies of city carbon footprint networks using multiregional input-output (MRIO) consumption-based accounting (CBA) have demonstrated that cities – similar to nations – are linked to significant trans-boundary emissions through trade

http://dx.doi.org/10.1016/j.apenergy.2016.08.053 0306-2619/Ó 2016 Elsevier Ltd. All rights reserved.

Please cite this article in press as: Chen G et al. Transnational city carbon footprint networks – Exploring carbon links between Australian and Chinese cities. Appl Energy (2016), http://dx.doi.org/10.1016/j.apenergy.2016.08.053

2

G. Chen et al. / Applied Energy xxx (2016) xxx–xxx

[3–5]. This opens up new opportunities for designing climate change mitigation policies. A recent study on carbon footprint linkages between five major Australian cities has shown that more than half of the nation’s carbon footprint (CF) is associated with these five cities, and that 43–71% of the cities’ CF is from emissions embodied in imports [6]. In comparison, the consumption of four Chinese provincial municipalities not only causes a large amount of emissions within their own territories, but also imposes many more emissions on their surrounding provinces via interregional supply chains [3]. A recent, comprehensive review on city carbon footprints based on consumption-based accounting (CBA) has been provided by Wiedmann et al. [7], which includes [5,8–19], while a summary on corresponding research based on the communitywide infrastructure footprint (CIF) method can be found in [20–25], with the most recent progress described in [25–29]. Around one fourth of global CO2 emissions is associated with the production of goods and services which are exported and used to satisfy demand from countries other than those where the CO2 is emitted [30]. A recent study suggests that the rapid growth of exports in carbon-intensive goods from Australia to China in the first decade of the 21st century has slowed global GHG emissions growth, owing to the significantly lower carbon intensity of goods produced in Australia compared to those produced in China [31]. International trade, even when transport-related emissions are factored into the calculation, therefore has the potential to lower global emissions if specific goods are produced in the location where environmental impacts are lowest [32,33]. Cities are not only engines of economic growth and innovation, but also hotspots of consumption, and they inevitably have to satisfy local demand though domestic and global trading, making them responsible for significant environmental impacts beyond their boundaries. The reduction of emissions embodied in trade is therefore crucial in designing urban carbon mitigation policies. In Australia, 29–39% of the carbon footprint (CF) of the five Australian capital cities is from emissions embodied into overseas imports [6]. Similarly, in the Chinese city of Xiamen, 40% of the CF comes from overseas [24]. However, no previous study has yet assessed emissions embodied into trade between cities in different countries, even though emissions transferred between countries have attracted significant academic attention in recent years [32,34–37]. The necessity to move towards transnational MRIO models with nested city regions, allowing links between consumption and production across cities, regional areas and nations, has been highlighted as a key requirement for future footprint research [38]. The present study is a comprehensive first step addressing this requirement. It is now well established that considering direct emission reductions alone (such as under the Kyoto Protocol) while neglecting outsourced emissions, renders any carbon mitigation efforts less effective [37,39–42]. In a similar fashion, global cities should also conduct the same analysis if local governments are to adopt holistic, full supply chain mitigation efforts. More and more cities are aiming at achieving carbon neutrality [43]. The Australian Government is planning to extend its Carbon Neutral Program that currently allows to certify organisations, products, services and events as carbon neutral, so that whole cities can be certified as carbon neutral as well [44]. This raises vital questions as to how much cities outsource their emissions to other cities through their global carbon networks and what the net emissions hierarchy between cities is. Addressing these questions outlines responsibility of cities for emissions and promotes collaborative actions for climate change mitigation. In this study, we employ a multi-scale MRIO model to analyse the transnational flow of embodied GHG emissions between four Chinese province-level municipalities (Beijing, Tianjin, Shanghai and Chongqing), a special Chinese administrative region (Hong

Kong) and the five largest Australian capital cities (Sydney, Melbourne, Brisbane, Adelaide and Perth). We tailor the model setup to suit a highly relevant case study, since China is now Australia’s largest two-way trading partner in goods and services. China is the largest destination for Australian exports of materials and goods as well as the largest source of merchandise imports into Australia [45]. The China Australia Free Trade Agreement (ChAFTA), which came into force on 20 December 2015 [45], is expected to further influence the intensity and transfer of embodied environmental impacts between these two countries [46]. The China-Australia case study aims to investigate how joint actions could be designed in order to maximise global GHG mitigation. Calculating the balance of emissions embodied in trade offers a city hierarchy of net carbon transfers. Based on the city carbon map concept [7], we model how carbon reductions in a city influence other regions through carbon networks. In the final section we discuss the implications of Chinese and Australian carbon trading schemes and the common benefits of matching both.

2. Materials and methods We include four Chinese provincial municipalities, Hong Kong and the largest five Australian capital cities in a globally closed multi-regional input-output table (MRIOT) with 25 sectors. The boundaries of Hong Kong and the four Chinese provincial municipalities are defined by their administrative area, which include both urban and rural areas (see population and areas in Table S1 of the Supplementary Material, SM). To make boundaries comparable with Chinese cities, we choose the Australian city’s boundary based on the Greater Capital City Statistical Areas (GCCSAs) published by the Australian Bureau of Statistics [47]. The GCCSAs is designed to include not only the urban area of the city but also the small towns and rural areas surrounding the city, thus covering a population which regularly socialise, shop or work between the cities and its surrounding areas. In the Eora database, the global MRIO tables for 2009 already include China, Hong Kong and Australia (http://worldmrio.com). The national tables of China and Australia are further replaced by detailed city MRIO tables from previous work. Following the approach taken in [48] to link 30 Chinese provinces (including the four provincial municipalities) with 185 countries in a global MRIOT (see more details in the SM). The trade links between Australian cities and the rest of the world is established in the same way (for details see papers [6,7,49]). We allocate the international import to local (city-scale) economies according to the share of local industrial intermediate use and the export to overseas destinations according to the share of local industrial intermediate supply. This implicitly assumes that international inputs are proportional to local inputs and exports are proportional to local supplies (see Fig. S1 in the SM). Precedents of such approaches estimating international trade between regional economies have been described in [50,51]. All city input-output tables are nested in the final MRIOT, with the rest of China (RoC) representing the domestic ‘hinterland’ of Chinese cities and the same applies for the rest of Australia (RoA). The rest of the world (RoW) is aggregated into one region to represent the global hinterland. While the resulting 13-region MRIOT is purpose-built for this study, the foundations of the underlying large-scale MRIO were laid in previous work on global and sub-national MRIOTs that required substantial data mining, compilation and optimisation [48,52–54]. Data for estimating GHG emissions of the four Chinese provincial municipalities are mainly taken from the 2010 Chinese Energy Statistics Yearbook and the 2010 Statistical Yearbooks of Beijing Shanghai, Tianjin and Chongqing [55]. Provincial CO2e emissions

Please cite this article in press as: Chen G et al. Transnational city carbon footprint networks – Exploring carbon links between Australian and Chinese cities. Appl Energy (2016), http://dx.doi.org/10.1016/j.apenergy.2016.08.053

G. Chen et al. / Applied Energy xxx (2016) xxx–xxx

were determined following the method described in [56,57] as well as and the IPCC’s sectoral accounting approach [58]. The input-output data and GHG extensions for Hong Kong and the rest of world (RoW) have been extracted from the Eora global MRIO database [52,53]. Three GHGs are considered for the present analysis – carbon dioxide, methane and nitrous oxide – expressed in CO2 equivalent emissions (CO2e) based on their global warming potential. The base year for all data is 2009. The Australian city and RoA MRIO table for the year 2009 is generated with the Australian Industrial Ecology Virtual Laboratory (IELab, http://ielab.info, [54]) by using Flegg’s adjusted location quotient method [59,60] as a non-survey method for subnational regionalisation of MRIOTs. The data necessary for this method include business turnover, employment and income data from the latest census, available at the Statistical Area 2 (SA2) level which has a population of about 10,000 [54]. A modified RAS method is used for re-balancing the MRIOT [54,61,62]. Our final multi-scale MRIOT with 13 regions has a mixed structure of supply/use and symmetric IO tables, reflecting the nature of the original data: Australian MRIOTs are in industry-bycommodity format, Chinese tables in commodity-by-commodity format and Hong Kong and RoW tables in industry-by-industry format. The exemplified MRIO table structure is shown in the form of a ‘heat map’ in Fig. 1, see also [53]. We adopt the concept and methodology of city carbon maps from [7] and extend it to a transnational city-scale study. In brief, a city carbon map, based on environmental input-output analysis, displays how the sectoral emissions are linked between cities and local, regional, national and global origins and destinations. A certain region’s carbon map can be calculated by allocating the emissions from sectors to final demand of products using the common Leontief demand-pull model [7]:

^i Ci ¼ ^f  L  y

3

ð1Þ

where Ci is region i’s carbon map of dimensions n  n, ^ f is a vector (n  1) of direct industry emission intensities that has been diago^i is nalised (n  n), L is the MRIO Leontief Inverse (n  n), and y region i’s final demand vector (n  1) that has been diagonalised (n  n). Emissions embodied in imports (EEI) to region i can be distinguished as:

^ij EEIji ¼ ^f  L  y

ð2Þ

where EEIji are the emissions embodied in imports from region j to i ^ij is a final demand (n  n), f^ and L remain the same as in Eq. (1). y vector (n  1) that has been diagonalised (n  n) and that refers to products produced in region j and consumed in region i. Emissions embodied in exports (EEE) from region i to j are simply the mirror of EEI:

EEEij ¼ EEIji

ð3Þ

We can further model how regional emissions reductions spread through the carbon networks. A change in a regional sector’s direct emissions will lead to the change in direct industry emission intensities, thus changing all regions’ carbon maps. The change can be expressed as:

cf  L  y ^i DCi ¼ D

ð4Þ

where DCi is change in region i’s carbon map of dimensions n  n, cf is a vector (n  1) of change of direct industry emission intensiD

^i remain the ties that has been diagonalised (n  n), and L and y cf  L represents the change of total same as in Eq. (1). In addition, D carbon intensity.

Fig. 1. Exemplified format and heatmap of the MRIOT structure used in this study. Remaining cities and rest-of-nation regions have been aggregated with the RoW regions for illustrative purposes (Ind = industries, Com = commodities, FD = final demand).

Please cite this article in press as: Chen G et al. Transnational city carbon footprint networks – Exploring carbon links between Australian and Chinese cities. Appl Energy (2016), http://dx.doi.org/10.1016/j.apenergy.2016.08.053

4

G. Chen et al. / Applied Energy xxx (2016) xxx–xxx

3. Results 3.1. City carbon footprints by sectors The city carbon footprint defined in [7] is designed to avoid double counting by deducting emissions embodied in exports from total emissions. In brief, the city carbon footprint (CF) equals the sum of all territorial GHG emissions, plus emissions embodied in imports (EEI), minus emissions embodied in exports (EEE). The remaining territorial emissions (RTE) are defined as territorial emissions minus EEE. The CF thus accounts for all emissions associated with the final demand for products in the city. Australian cities’ CFs range from 11 t CO2e/cap for Brisbane to 31 t CO2 e/cap for Perth (Fig. 2), not including direct emissions from private transport and gas use of Australian households which are calculated separately (and shown in the Supplementary Data, SD). Direct emissions from private transports of households in the four Chinese municipalities are not included in the transport sectors in Fig. 2 due to the lack of data. The energy consumption of private cars began to be taken into accounts in statistics since 2015. However, emissions from private car use can be obtained from the literature [63]. For example, car use in Beijing accounted for 0.4 t/cap in 2009 [64]. The CFs of the four Chinese municipalities are 18, 17, 21 and 6.4 t CO2e/cap for Shanghai, Beijing, Tianjin and Chongqing, respectively. These results are almost twice as high as reported by Feng et al. [3]. The likely reason for this is that we use more detailed 39-sector industrial energy use data for each provincial municipality whereas Feng et al. adopt the production based data from [56,65] which are based on the Chinese Economic Census Yearbook (CECY) [66] and the Chinese Energy Statistics Yearbooks (CESY) [67]. In addition to these two data sets we also collect energy consumption data for detailed industrial sectors

from all Provincial Statistics Yearbooks (PSY) [55]. The CECY with energy data is only updated to 2008 while PSY includes the latest updated time series up to 2015. CECY and PSY show large differences in production-based data, for example, total territorial emissions in Shanghai are 50% higher than the data used in Feng et al. [3] (see comparison in SD). The higher territorial emissions also result in higher emissions for the rest of China (RoC). The total gap between the sum of provincial statistics and the national statistics can be as high as 1.4 Gt [57]. This accounts for the significantly higher EEI from the RoC in this study. We also include EEI from the RoW which further increases the CFs of the four municipalities by up to 2 t CO2e/cap. The construction sector is the largest contributor in all cities. In Australia it represents 21–24% of the total CF, due to sustained growth in the population and the economy [68]. In China, this share is even higher with 23–41%, which reflects a strong growth in the construction and real estate sectors following China’s twoyear 4 trillion Yuan (US$ 586bn) stimulus since 2008 to boost the economy mainly through infrastructure projects [69]. In Chinese cities, 0.7–1.9 t CO2e/cap are allocated to the electricity, gas and water sector, which is smaller than the territorial emissions as the electricity generated in cities is often embodied in services exported from the cities. Compared to Chinese cities, the proportion of emissions allocated to electricity, gas and water is higher in Australia (between 8% and 16%), which can be explained by a combination of high per-capita usage of electricity (10,972 kW h/cap in Australia vs 2633 kW h/cap in China [70]) and carbon-intensive power generation, predominantly based on coal-fired power plants. Four to nine percent of embodied emissions are attributed to the electrical equipment and machinery sector in China compared to 1–2% in Australia. Transport makes up 1.2–2.9 t CO2e/cap,

Fig. 2. Total per-capita carbon footprints of Australian and Chinese main cities by sector (excluding direct emissions from households).

Please cite this article in press as: Chen G et al. Transnational city carbon footprint networks – Exploring carbon links between Australian and Chinese cities. Appl Energy (2016), http://dx.doi.org/10.1016/j.apenergy.2016.08.053

G. Chen et al. / Applied Energy xxx (2016) xxx–xxx

showing that the transport in Australian cities is not as carbonefficient as in China’s densely populated megacities where transport emissions are between 0.3 and 0.9 t CO2e/cap. According to the official data [71], Hong Kong emits 42 Mt CO2e annually from its territory, with 69% and 17% from electricity use and transport, respectively. Presently, the Hong Kong government does not account for the emissions embodied in imported services and goods. In comparison, under consumption-based accounting, the total CF of Hong Kong is 241 Mt CO2e (33 t CO2e/cap). The electrical and machinery sector and construction sector both take up 14% and 15%, respectively. Food and retail trade makes up around 9%, mainly because Hong Kong is a well-known dining hub and popular with travellers. The transport sector only contributes 4% or 8 Mt CO2e which is close to the official data for territorial emissions (7.4 Mt CO2e). The rest of emissions mainly come from the demand for other services and products sectors. 3.2. City carbon footprint networks Territorial emissions in all regions can be mapped to city and regional carbon footprints following the flows of emissions embodied in interregional trade (from bottom to top in Fig. 3; see sectoral details in SD and total emissions in Fig. S1 in the SM). For example, the territorial emissions of Melbourne are 9.7 t CO2e/cap, of which 7.5 t CO2e/cap (the RTE) remain in Melbourne, 1.0 t CO2e/cap are exported to Sydney and 1.2 t CO2e/cap to all other regions that were modelled. The sum of embodied emissions flows to the top represents a regional CF. For Melbourne, this is 16.2 t CO2e/cap coming from the own city territory (RTE = 7.5 t CO2e/cap) and from emissions embodied in imports from RoA (3.0 t CO2e/cap), RoC (1.6 t CO2e/cap), RoW (2.8 t CO2e/cap) and all other regions (1.3 t CO2e/cap). The RTE shows how much territorial emissions are released to satisfy the demand of city residents. It is obvious that the RTEs of all the Chinese cities including Hong Kong only make up a fraction of the city CFs (8–27%) while these proportions are much higher in Australian Eastern capital cities with 30–50% and the highest in Perth with 62% (19.5 t CO2e/cap). Perth relies more on GHG emissions from own industries to satisfy final city demand than from anywhere in Australia and the world. Furthermore, economic output in Perth is high relative to population [72]. For all other cities, and in particular the Chinese megacities, the demand for resources such as labour and materials has long reached beyond city boundaries, resulting in an increase of outsourcing and transferring large amounts of emissions at the same time. In contrast to China, lower population density in Australian cities mean larger distances for transport and a higher (yet decreasing) share of remaining

5

agricultural land. This, together with a carbon-intensive, highconsumption lifestyle means that territorial emissions are larger in Australian cities on a per-capita basis. The EEI from RoC to Australian cities are 1.2–2.8 t CO2e/cap showing that Australian cities outsource some of their carbon footprints by purchasing low-price products and services from the rest of China. For Chinese cities, the EEI varies with 4.4 t CO2e/cap from RoC to Chongqing to 12.3–15.8 t CO2e/cap for Beijing, Tianjin, Shanghai and Hong Kong. This highlights, on the one hand, the high carbon-intensity production in mainland China, leading to high carbon-intensity domestic imports to these cities [73]. On the other hand, the extent of EEI reflects consumption levels and lifestyle patterns of cities. Slower economic development in Chongqing, for example, results in a less modern lifestyle compared to the other three municipalities results, resulting in a significantly lower CF and EEI for Chongqing (see GDP data in Table S1 in the SM). The EEI from RoW to Australian cities are in the range of 2.3–5.0 t CO2e/cap while this is only 0.1–2.3 t CO2e/cap for the four Chinese municipalities. This means that urban residents in Australia outsource more emissions to the RoW than their Chinese counterparts. Or, in other words, Chinese cities are less dependent on foreign products than Australian cities. In contrast, Hong Kong which is a free port and does not levy any duty on imports [74] has much higher EEI with 14.7 t CO2e/cap from RoW. The carbon footprint networks can be further magnified for specific cities with spatial and sectoral information to reveal ‘hotpots’ and possible priority intervention areas. The example from Melbourne is shown in Fig. 4 (for all other cities see the Figs. S3– S11 in SM). From top to bottom, the EEI from RoA, RoC and RoW make up the largest percentage of total EEI. The service sector from these regions contributes the largest amount, followed by manufacturing and construction. This reflects the large amount of imports of goods, materials and services by Melbourne residents. But in the manufacturing sector, Melbourne is exporting an even larger amount of embodied emissions than it imports, confirming Melbourne’s position as a manufacturing hub. 3.3. Balance of emissions embodied in intercity trade The urban environment has its limits in providing all resources for the growing demand of residents, and cities inevitably rely on supplies from their hinterlands [61,75]. In addition, cities outsource the production of goods and services to each other, which leads to a complex web of trade links and associated embodied flows of greenhouse gas emissions. The net carbon transfer between cities can be calculated by subtracting EEE from EEI for

Fig. 3. Inter-regional, per-capita flows of embodied carbon emissions instigated by city final demand, (the largest bilateral flows are labelled).

Please cite this article in press as: Chen G et al. Transnational city carbon footprint networks – Exploring carbon links between Australian and Chinese cities. Appl Energy (2016), http://dx.doi.org/10.1016/j.apenergy.2016.08.053

6

G. Chen et al. / Applied Energy xxx (2016) xxx–xxx

Fig. 4. Flows of embodied GHG emissions related to Melbourne.

each city. We depict net carbon flows in the form of a hierarchy in ascending order of net carbon receipts, i.e. if a city receives positive net embodied carbon emissions from another, the city is higher in the hierarchy (Fig. 5). Interestingly, all Australian cities appear in the middle of the hierarchy graph whereas Chinese cities can be found both at the bottom and at the top of the graph (Fig. 5). Hong Kong ranks at the top of the hierarchy and receives the largest net carbon transfer from both Chinese and Australian cities. Hong Kong is an international financial and commercial centre serving the Asia-Pacific region and the Mainland of China [74]. However, much of Hong Kong’s prosperity is not because of trading its own resources or products, but by taking advantage of its free trade policy to trade commodities mainly between the mainland of China and other global trade centres [76]. The largest flow of embodied net emissions between the ten cities investigated is 8.5 Mt CO2e from Shanghai to Hong Kong. Shanghai is the largest container port of mainland China [77] and a hub for shipping made-in-China products globally. It also serves as a net exporter of embodied emission to Beijing and all Australian cities, only receiving net emissions from Tianjin and Chongqing. Beijing is second highest in the hierarchy. As an inland capital it is not only driving the flows from other Chinese cities but also from the Australian cities. Net flows, however, seem relatively small since both EEE and EEI are of similar magnitude (e.g. EEE from Sydney to Beijing is 174 kt CO2e and EEE from Beijing to Sydney is 120 kt CO2e, resulting in a net flow of 54 kt CO2e). Within Australia, Sydney is leading the net carbon transfer hierarchy (mostly receiving from Melbourne and Perth), although it has lower per-capita emissions than Melbourne. Chongqing as the least developed city, compared with the other nine cities, is ranked at the bottom as it mainly exports net emissions to all other cities. Unlike at the country level where net carbon emissions are exclusively transferred from developing to developed countries [39], the city hierarchy in (Fig. 5) demonstrates a developing country can have rich regions like Beijing and Hong Kong that are receiving net carbon transfers from cities in both developed and developing countries. Therefore cities should be differentiated

from countries in their responsibility for net carbon transfers. This is even more important in large countries with considerable spatial heterogeneities in economic development and lifestyle, such as China. 3.4. Modelling the influence of local mitigation spreading through carbon networks A reduction in territorial emissions in a region will lead to indirect reductions elsewhere, thus transferring the benefits through the carbon network. We demonstrate this effect following the reduction of embodied flows in other regions if emissions in Shanghai are reduced. We choose this city for two reasons. Firstly, Shanghai constitutes a real-world case where emissions reductions are actually happening and secondly, Shanghai has larger flows of embodied emissions to Hong Kong and Australia than other municipalities. During the 2005–2010 periods, Shanghai had developed an emission-reduction project under the Clean Development Mechanism (CDM). The project resulted in 6500 kt CO2e reduction per year by introducing a gas-to-electricity technology in a local steel company (namely Baoshan Iron & Steel), after adjusting for an offset in grid electricity due to the use of blast furnace gas technology [78]. Based on the cities’ carbon map, we hypothetically subtract these emissions from the energy sector to model how these reductions transfer from Shanghai to other regions using Eq. (4) (in Fig. 6, the negative numbers represent reductions). As is expected, the initial direct emissions reduction in energy sector leads to the mitigations in all sectors in Shanghai as they use lower carbon-intensity energy. In order to display results succinctly, we aggregate the 25 model sectors into 6 sectors (see details of 25 sectors in the SD). The manufacturing and services sectors rely heavily on the energy sector and therefore share the most reductions with 2573 and 1692 kt CO2e, while agriculture and food, electricity, gas and water and transport sector register reductions in the order of 271, 593 and 194 kt CO2e. Also, the reductions from the energy sector will lower the carbon intensity in the local production systems, thus reducing the

Please cite this article in press as: Chen G et al. Transnational city carbon footprint networks – Exploring carbon links between Australian and Chinese cities. Appl Energy (2016), http://dx.doi.org/10.1016/j.apenergy.2016.08.053

G. Chen et al. / Applied Energy xxx (2016) xxx–xxx

7

Fig. 5. Hierarchy of net carbon transfers between Australian and Chinese cities (only flows larger than 0.2 Mt CO2e are shown).

Fig. 6. Influence of GHG emissions mitigation in Shanghai on other cities. Numbers show the net emissions reductions (rounded up to the nearest kt) in all cities as a consequence of lowering emissions in Shanghai.

emissions embodied into products and services and benefiting those regions that import these products and services. Most of the reductions from the manufacturing sector are transferred to the RoC and the RoW as a result of Shanghai exporting less carbon-intensive manufacturing products to these two regions while only a small percentage are consumed within the city boundary. In contrast, construction and electricity, gas and water share more reductions within the city boundary as products from these sectors are rarely exported. The total benefits for all regions are shown at the bottom. Only 32% of reductions (2056 kt CO2e) stay within Shanghai itself, whereas 68% of reductions become embodied in exports and are transferred to other regions. The five Australian capital cities reduce their CF by around 4.5–17.2 kt CO2e from EEI coming from

Shanghai while the other Chinese cities achieve a mitigation of 18.8–142 kt CO2e. The RoC and RoW benefit the most, with reductions of 1207 and 2898 kt CO2e, respectively. In other words, from a consumption perspective, reducing Baoshan’s carbon emissions has a higher mitigation benefit in the rest of the world than in the city of Shanghai or even China as a whole. The reductions in Shanghai also lead to the decrease of total carbon intensity thus transforming production systems in other regions (Fig. 7). Hong Kong benefits the most, with manufacturing and services seeing a reduction in the order of 0.01–0.1%. This confirms the hierarchy where Hong Kong was shown to have a strong connection with Shanghai in these sectors. The electricity, gas and water sector is hardly affected though, as Hong Kong seldom imports these products from Shanghai. For Australian regions,

Please cite this article in press as: Chen G et al. Transnational city carbon footprint networks – Exploring carbon links between Australian and Chinese cities. Appl Energy (2016), http://dx.doi.org/10.1016/j.apenergy.2016.08.053

8

G. Chen et al. / Applied Energy xxx (2016) xxx–xxx

Fig. 7. Reduction of sectoral carbon intensity in all regions as a result of GHG emissions reductions in the Baoshan company in Shanghai.

the most significant impact is on financial intermediation and business activities with a reduction of 0.02–0.03% in carbon intensity. This is likely because Shanghai is the top financial centre in the mainland of China [79]. For other Chinese regions, the transport equipment sector carbon intensity is most reduced by 0.01–0.03% since Shanghai accommodates many transport equipment manufacturing companies. Some of this reduction also translates through to Australian cities.

modelling any induced economic (and related emissions) changes as a result of mitigation measures. This is based on the assumption that the mitigation in Shanghai does not significantly influence the structure or size of the economies of Shanghai or other regions. For other uncertainties and limitations associated with the city carbon map concept we refer to [7].

3.5. Uncertainties and limitations of this study

4.1. Implications for China and Australia

Governments usually do not collate or publish detailed data for inter- or intra-regional trade. This is the case for trade between regions within a nation but even more so for trade between regions from different countries. This lack of ‘real’ data based on actual surveys has been a problem for inter-regional IO modelling since its beginnings in the 1950s and ever since modellers have developed a number of non-survey methods that allow for the estimation of inter-regional and inter-sectoral trade flows [54,80]. In this study we used the Adjusted Flegg’s Location Quotient (AFLQ) method for estimating interregional trade matrices within Australia [54,60]. For the trade matrices between Australian and other regions, however, we were limited to a simplified approach as described in Section 2. This means that inter-city trade patterns follow the general trade structure between Australia and other regions whereas trade volumes for individual sectors follow industries’ demand and supply in the cities. This part can be improved once the detailed data is available or by using non-survey models [80]. When modelling the effects of direct emissions reductions in Shanghai, we only consider first-order effects without further

In the analysis of city carbon footprints by sector, the largest percentage of CF of Chinese municipalities comes from the construction sectors (up to 40%) which could be related to the China’s 4 trillion Yuan stimulus to boost the economy. Hong Kong as a financial and commercial hub has the highest per-capita emissions, with the majority coming from goods, services, construction and electricity, gas and water sectors. For Australian cities, the construction, transport, and food and beverage sectors make up about 40% of CF while the rest is allocated to goods and services as well. Even though the government of Hong Kong has announced that it will direct efforts to enhance the energy efficiency in buildings and also plans on purchasing nuclear power from the Chinese mainland [81], the real challenge for the city lies in the emissions embodied in imported goods and services from both mainland China and RoW. The top position in the hierarchy demonstrates Hong Kong’s responsibility for its net transferred emissions. Hong Kong could thus prioritise investing in carbon trading projects with Shanghai as this will not only reduce the emissions embodied in imports but also lead to a production system with lower carbon intensity in Hong Kong. The solution could be joining the carbon

4. Discussion and conclusions

Please cite this article in press as: Chen G et al. Transnational city carbon footprint networks – Exploring carbon links between Australian and Chinese cities. Appl Energy (2016), http://dx.doi.org/10.1016/j.apenergy.2016.08.053

G. Chen et al. / Applied Energy xxx (2016) xxx–xxx

trading scheme with mainland of China where the main sources for Hong Kong’s embodied emissions are found. The study of carbon footprint networks reveals a disparity of regional CFs, RTE and EEI. Chinese cities have lower RTE because of limited land, resources and large populations while Australian cities share higher RTE because of lower population densities. The EEI from RoC, RoA and RoW to Australian cities make up about half of their CFs whereas the EEI from RoC to Chinese cities constitute most of their CFs, indicating the high-carbon intensity production system and high mitigation potentials in RoC. Cities in both countries have already set their own individual targets to reduce urban carbon emissions. Hong Kong has proposed to set a carbon intensity reduction target of 50–60% by 2020 compared to 2005 levels [81]. The four Chinese provincial municipalities are currently evaluating the outcomes of their previous five-year (2011–2015) mitigation plans aimed at reducing carbon intensity by 17–19% [82], with mitigation plans for the running 5-year period (2016–2020) already underway by the central China government. The Chinese government has further released seven pilot carbon trading schemes that cover the four provincial municipalities, the city of Shenzhen and two other provinces, in an attempt to improve the efficiency of emissions reduction [83]. China’s pilot trading scheme is a promising step to create a financial market capable of capping and reducing emissions through economic incentives [84,85]. It has thus laid the foundation for implementing a national emissions trading scheme in 2017 [86]. The national scheme is important for the four municipalities because most of their emissions are coming from the Chinese hinterland (i.e. RoC). Thus cities should purchase credits through this scheme from their hinterlands in order to fulfil their 13-five-year mitigation targets. On the one hand, this benefits hinterlands which have a high potential in reductions and shortage in funding for mitigation [87]. On the other hand, municipalities could reduce their emissions economically through buying cheaper credits from outside the boundary, and cut down the emissions embodied into trade from hinterlands. This may have the added benefit of reducing income inequality across regions within China. In addition, the benefits of mitigation from the RoC will spread worldwide by reducing the emissions embodied in products and services originating from China. Australian cities, on the other hand, have set themselves far more ambitious long-term emissions targets. In joining the Carbon Neutral Cities Alliance [43], Melbourne and Sydney commit to reducing greenhouse gas emissions by 80% or more by 2050 and to finally becoming carbon neutral cities [43]. Adelaide City Council has claimed that the city of Adelaide could become the world’s first carbon neutral city [88]. The carbon hierarchy presented in our study suggests that a responsibility dividend for cities could provide an incentive that drives low-carbon outcomes through carbon trading among cities. To claim carbon neutrality, Australian cities need to offset indirect emissions beyond the city’s sphere of influence [89]. Under the Australian cap and trade policy (the so-called Safeguard Mechanism) [90], cities can prioritise purchasing carbon credits from their domestic sources as EEI makes up a large percentage of their total CF. The option to buy offshore credit units is not part of the current policy but is under consideration by the new government [91]. This option has the advantage that emissions reductions in overseas cities via offsets would not only reduce the territorial emissions of these cities but would also benefit the city who buys the credits as the emissions embodied in its imports are also reduced. As an example, if a company from Sydney pays to reduce the direct emissions from a power station in Shanghai, the emissions embodied in products traded from Shanghai to Sydney will be reduced as well, and so will be Sydney’s total CF (see case study in Section 3.4). Furthermore, as the Baoshan example shows, the

9

consumption-based carbon impact multipliers are reduced across the entire city network in any location dependent on imports from Shanghai. We show that Hong Kong’s carbon mitigation efforts benefit significantly as it is highly dependent on trade from Shanghai. Carbon trading schemes thus play an important role in the joint actions of cities. Both China’s trading scheme and Australia’s cap and trade system will help to cut down the emissions embodied in trade and spread benefits globally through carbon networks. Hence it would be beneficial to streamline the two nation’s trading schemes between each other as well in global carbon trading markets. 4.2. Global policy implications Joint mitigation actions based on carbon trading or CDM type projects on a substantial scale could help cities to achieve a winwin scenario. Even though a global carbon market would arguably be the most efficient solution to reduce emissions worldwide, a multilateral, top-down solution has proven elusive and improbable in the short to medium term [6]. Global city networks, however, could pioneer city-scale links in global carbon trading markets that can help to realize many benefits more efficiently and/or sooner, and help pave the way towards a larger, more efficient global emissions market. The approach and analysis presented in this paper raises interesting prospects with regards to who should be shouldering responsibility for carbon mitigation efforts. Whilst the notion of developed countries outsourcing their emissions to the developing world is clearly valid at the national scale [39], consumption-based accounting at the city scale reveals that large and affluent cities in developing countries have already developed urban economies and lifestyles which also entail considerable outsourcing of emissions to other areas of the country [3]. In the case of Beijing, Tianjin, Shanghai and in particular Hong Kong in our study, the percapita emissions are comparable or even higher than most large cities in developed Australia. Beijing is also placed higher in the net carbon transfer hierarchy, second only to Hong Kong. This certainly supports the notion of considering the role of cities in developing countries not only as producers but also as consumers who require considerable amounts of resources [92] and embedded energy [93,94] from other regions and, increasingly, other nations. Our analysis also highlights the potential of utilising highly disaggregated nested MRIOTs in order to mitigate carbon emissions by locating the most interconnected points in the city supply chains. This is where carbon mitigation efforts should be targeted to ensure an optimum reduction in global emissions and also to devise ways to attract carbon trading investment in these regions. Unlike in the case of other environmental issues [95], when it comes to carbon emissions, reductions anywhere can have a positive global impact irrespective of their location. 4.3. Concluding remarks This study conducts the first investigation into embodied carbon emissions of inter-city trade across countries. The results challenge the stereotypical view that developed countries generally shift the burden of GHG emissions to developing countries, which does not necessarily appear to be the case at the city level. Beijing and Hong Kong outsource more emissions than all the five large Australian cities combined. Per-capita emissions are also not determined by the development status of Australia and China, but by the economic idiosyncrasies of the cities themselves. Our study also reveals new opportunities for joint responsibility and action, including direct investment in cities and regions which represent upstream GHG ‘hotspots’ related to city final demand. The method

Please cite this article in press as: Chen G et al. Transnational city carbon footprint networks – Exploring carbon links between Australian and Chinese cities. Appl Energy (2016), http://dx.doi.org/10.1016/j.apenergy.2016.08.053

10

G. Chen et al. / Applied Energy xxx (2016) xxx–xxx

presented can assess the collective total impact of the commitments that cities made at the Paris Climate Conference. Further improvements in data and methodologies will allow exploring alternative carbon trading scenarios. This could, in turn, lead to the optimisation of mitigation efforts to reduce supply chain emissions in different sectors in different cities at minimum cost. Acknowledgement This research was enabled by the Industrial Ecology Virtual Laboratory (IELab, www.ielab.info), originally funded under the Australian National eResearch Collaboration Tools and Resources project (NeCTAR, code VL201). IELab datafeeds and routines written by Manfred Lenzen, Arne Geschke, Joe Lane and Hazel Rowley were utilised. Yafei Wang acknowledges the support of the Program for New Century Excellent Talents in University (Grant NCET-13-0060) and the National Philosophy and Social Science Foundation of China (Grant No. 12BTJ013). Guangwu Chen acknowledges a PhD scholarship jointly granted by the China Scholarship Council (CSC), UNSW Australia and the Australian Cooperative Research Centre for Low Carbon Living (project code RP2007). Appendix A. Supplementary material Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.apenergy.2016. 08.053. References [1] C40 Cities Climate Leadership Group. Compact partners announce aggregate emissions impact at Climate Summit for Local Leaders. ; 2015 [accessed on 05 August 2016]. [2] Compact of Mayors. Announcing the collective global impact of the compact of Mayors at the climate summit for local leaders. ; 2016 [accessed on 05 August 2016]. [3] Feng K, Hubacek K, Sun L, Liu Z. Consumption-based CO2 accounting of China’s megacities: the case of Beijing, Tianjin, Shanghai and Chongqing. Ecol Indic 2014;47:26–31. [4] Chavez A, Ramaswami A. Articulating a trans-boundary infrastructure supply chain greenhouse gas emission footprint for cities: mathematical relationships and policy relevance. Energy Policy 2013;54:376–83. [5] Minx J, Baiocchi G, Wiedmann T, Barrett J, Creutzig F, Feng K, et al. Carbon footprints of cities and other human settlements in the UK. Environ Res Lett 2013;8:035039. [6] Chen G, Wiedmann T, Hadjikakou M, Rowley H. City carbon footprint networks. Energies 2016;9:602. [7] Wiedmann TO, Chen G, Barrett J. The concept of city carbon maps: a case study of Melbourne, Australia. J Ind Ecol 2015. http://dx.doi.org/10.1111/jiec.12346. [8] Larsen HN, Hertwich EG. The case for consumption-based accounting of greenhouse gas emissions to promote local climate action. Environ Sci Policy 2009;12:791–8. [9] Lenzen M, Peters GM. How city dwellers affect their resource hinterland. J Ind Ecol 2009;14:73–90. [10] Larsen HN, Hertwich EG. Identifying important characteristics of municipal carbon footprints. Ecol Econ 2010;70:60–6. [11] Larsen HN, Hertwich EG. Implementing carbon-footprint-based calculation tools in municipal greenhouse gas inventories. J Ind Ecol 2010;14:965–77. [12] Zhou SY, Chen H, Li SC. Resources use and greenhouse gas emissions in urban economy: ecological input–output modeling for Beijing 2002. Commun Nonlinear Sci Numer Simul 2010;15:3201–31. [13] Chavez AA. Comparing city-scale greenhouse gas (GHG) emission accounting methods: implementation, approximations, and policy relevance. University of Colorado at Denver; 2012. [14] Guo S, Shao L, Chen H, Li Z, Liu JB, Xu FX, et al. Inventory and input–output analysis of CO2 emissions by fossil fuel consumption in Beijing 2007. Ecol Inform 2012;12:93–100. [15] Ala-Mantila S, Heinonen J, Junnila S. Greenhouse gas implications of urban sprawl in the Helsinki metropolitan area. Sustainability 2013;5. [16] Chen GQ, Guo S, Shao L, Li JS, Chen Z-M. Three-scale input–output modeling for urban economy: carbon emission by Beijing 2007. Commun Nonlinear Sci Numer Simul 2013;18:2493–506.

[17] Paloheimo E, Salmi O. Evaluating the carbon emissions of the low carbon city: a novel approach for consumer based allocation. Cities 2013;30:233–9. [18] Dias AC, Lemos D, Gabarrell X, Arroja L. Environmentally extended input– output analysis on a city scale – application to Aveiro (Portugal). J Clean Prod 2014;75:118–29. [19] Hermannsson K, McIntyre SG. Local consumption and territorial based accounting for CO2 emissions. Ecol Econ 2014;104. [20] Kennedy C, Steinberger J, Gasson B, Hansen Y, Hillman T, Havránek M, et al. Greenhouse gas emissions from global cities. Environ Sci Technol 2009;43:7297–302. [21] Kennedy C, Steinberger J, Gasson B, Hansen Y, Hillman T, Havránek M, et al. Greenhouse gas emissions from global cities. Environ Sci Technol 2011;45:3816–7. [22] Ramaswami A, Chavez A. What metrics best reflect the energy and carbon intensity of cities? Insights from theory and modeling of 20 US cities. Environ Res Lett 2013;8:035011. [23] Kennedy CA, Stewart I, Facchini A, Cersosimo I, Mele R, Chen B, et al. Energy and material flows of megacities. Proc Natl Acad Sci 2015;112:5985–90. [24] Lin J, Hu Y, Cui S, Kang J, Ramaswami A. Tracking urban carbon footprints from production and consumption perspectives. Environ Res Lett 2015;10:054001. [25] Tan X, Dong L, Chen D, Gu B, Zeng Y. China’s regional CO2 emissions reduction potential: a study of Chongqing city. Appl Energy 2016;162:1345–54. [26] Mi Z, Zhang Y, Guan D, Shan Y, Liu Z, Cong R, et al. Consumption-based emission accounting for Chinese cities. Appl Energy 2016. http://dx.doi.org/ 10.1016/j. apenergy.2016.06.094. [27] Chen S, Chen B. Urban energy consumption: different insights from energy flow analysis, input–output analysis and ecological network analysis. Appl Energy 2015;138:99–107. [28] Tan S, Yang J, Yan J, Lee C, Hashim H, Chen B. A holistic low carbon city indicator framework for sustainable development. Appl Energy 2016. http:// dx.doi.org/10.1016/j.apenergy.2016.03.041. [29] Wei J, Chen H, Cui X, Long R. Carbon capability of urban residents and its structure: evidence from a survey of Jiangsu Province in China. Appl Energy 2016;173:635–49. [30] Wiedmann T. Impacts embodied in global trade flows. In: Clift R, Druckman A, editors. Taking stock of industrial ecology. Cham: Springer International Publishing; 2016. p. 159–80. [31] Tan H, Sun A, Lau H. CO2 embodiment in China-Australia trade: the drivers and implications. Energy Policy 2013;61:1212–20. [32] Kander A, Jiborn M, Moran DD, Wiedmann TO. National greenhouse-gas accounting for effective climate policy on international trade. Nat Clim Change 2015;5:431–5. [33] Brizga J, Feng K, Hubacek K. Household carbon footprints in the Baltic States: A global multi-regional input–output analysis from 1995 to 2011. Appl Energy 2016. http://dx.doi.org/10.1016/j.apenergy.2016.01.102. [34] Andrew RM, Davis SJ, Peters GP. Climate policy and dependence on traded carbon. Environ Res Lett 2013;8:034011. [35] Peters GP, Marland G, Le Quere C, Boden T, Canadell JG, Raupach MR. Rapid growth in CO2 emissions after the 2008–2009 global financial crisis. Nat Clim Change 2012;2:2–4. [36] Davis SJ, Peters GP, Caldeira K. The supply chain of CO2 emissions. Proc Natl Acad Sci 2011;108:18554–9. [37] Scott K, Barrett J. An integration of net imported emissions into climate change targets. Environ Sci Policy 2015;52:150–7. [38] Hubacek K, Feng K, Minx JC, Pfister S, Zhou N. Teleconnecting consumption to environmental impacts at multiple spatial scales. J Ind Ecol 2014;18:7–9. [39] Peters GP, Minx JC, Weber CL, Edenhofer O. Growth in emission transfers via international trade from 1990 to 2008. Proc Natl Acad Sci 2011;108:8903–8. [40] Barrett J, Peters G, Wiedmann T, Scott K, Lenzen M, Roelich K, et al. Consumption-based GHG emission accounting: a UK case study. Clim Policy 2013;13:451–70. [41] Scott K, Daly H, Barrett J, Strachan N. National climate policy implications of mitigating embodied energy system emissions. Clim Change 2016:1–14. [42] Xia Y, Fan Y, Yang C. Assessing the impact of foreign content in China’s exports on the carbon outsourcing hypothesis. Appl Energy 2015;150:296–307. [43] CNCA, Carbon Neutral Cities Alliance. Framework for long-term deep carbon reduction planning. ; 2016 [accessed on 05 August 2016]. [44] Australian Government. Government paves way for carbon neutral cities. Joint media release of the Minister for the Environment and the Minister for Cities and the Built Environment. ; 2016 [accessed on 05 August 2016]. [45] Australian Government. China-Australia free trade agreement. ; 2016 [accessed on 05 August 2016]. [46] Jayanthakumaran K, Liu Y. Bi-lateral CO2 emissions embodied in AustraliaChina trade. Energy Policy 2016;92:205–13. [47] ABS. Statistical geography factsheet – greater capital city statistical areas. . . Canberra, ACT. Australia: Australian Bureau of Statistics. ; 2012 [accessed on 05 August 2016]. [48] Wang Y, Geschke A, Lenzen M. Constructing a time series of nested multiregion input-output tables. Int Region Sci Rev 2015. 0160017615603596. [49] Chen G, Hadjikakou M, Wiedmann T. Urban carbon transformations: unravelling spatial and inter-sectoral linkages for key city industries based on multi-region input–output analysis. J Clean Prod 2016. http://dx.doi.org/ 10.1016/j.jclepro.2016.04.046.

Please cite this article in press as: Chen G et al. Transnational city carbon footprint networks – Exploring carbon links between Australian and Chinese cities. Appl Energy (2016), http://dx.doi.org/10.1016/j.apenergy.2016.08.053

G. Chen et al. / Applied Energy xxx (2016) xxx–xxx [50] Stilwell FJB, Boatwright BD. A method of estimating interregional trade flows. Region Urban Econ 1971;1:77–87. [51] Testa W, Klier T, Zelenev A. Estimating US metropolitan area export and import competition. Econ Perspect 2003;27:13–27. [52] Lenzen M, Kanemoto K, Moran D, Geschke A. Mapping the structure of the world economy. Environ Sci Technol 2012;46:8374–81. [53] Lenzen M, Moran D, Kanemoto K, Geschke A. Building Eora: a global multiregion input–output database at high country and sector resolution. Econ Syst Res 2013;25:20–49. [54] Lenzen M, Geschke A, Wiedmann T, Lane J, Anderson N, Baynes T, et al. Compiling and using input–output frameworks through collaborative virtual laboratories. Sci Total Environ 2014:485–6. [55] National Bureau of Statistics of China. Provincial statistical yearbooks. Beijing: China Statistics Press; 1998–2012. [56] Liu Z, Geng Y, Lindner S, Guan D. Uncovering China’s greenhouse gas emission from regional and sectoral perspectives. Energy 2012;45:1059–68. [57] Guan D, Liu Z, Geng Y, Lindner S, Hubacek K. The gigatonne gap in China/’s carbon dioxide inventories. Nature Clim Change 2012;2:672–5. [58] Intergovernmental Panel on Climate Change (IPCC). 2006 IPCC guidelines for national green house gas inventories; 2006. [59] Mccann P, Dewhurst JHL. Regional size, industrial location and input-output expenditure coefficients. Region Stud 1998;32:435–44. [60] Flegg AT, Webber CD. Regional size, regional specialization and the FLQ formula. Region Stud 2000;34:563–9. [61] Lenzen M, Peters GM. How city dwellers affect their resource hinterland – a spatial impact study of Australian households. J Ind Ecol 2010;14:73–90. [62] Lenzen M, Gallego B, Wood R. Matrix balancing under conflicting information. Econ Syst Res 2009;21:23–44. [63] Zeng Y, Tan X, Gu B, Wang Y, Xu B. Greenhouse gas emissions of motor vehicles in Chinese cities and the implication for China’s mitigation targets. Appl Energy 2016. http://dx.doi.org/10.1016/j.apenergy.2016.06.130. [64] Wang Z, Liu W. Determinants of CO2 emissions from household daily travel in Beijing, China: individual travel characteristic perspectives. Appl Energy 2015;158:292–9. [65] Liu Z, Liang S, Geng Y, Xue B, Xi F, Pan Y, et al. Features, trajectories and driving forces for energy-related GHG emissions from Chinese mega cites: the case of Beijing, Tianjin, Shanghai and Chongqing. Energy 2012;37:245–54. [66] National Bureau of Statistics. Chinese economic census yearbook; 2008. [67] National Bureau of Statistics. China energy statistical yearbook Beijing: China Statistics Press; 1998–2010. [68] Australian Bureau of Statistics. 4102.0 – Australian social trends, 2014. ; 2014 [accessed on 05 August 2016]. [69] Zheng Y, Chen M. How effective will China’s Four Trillion Yuan Stimulus Plan be. The University of Nottingham China Policy Institute Briefing Series; 2009. [70] World Bank. Electric power consumption (kWh per capita). ; 2009 [accessed on 05 August 2016]. [71] Hong Kong Government. Greenhouse gas emissions in Hong Kong. ; 2016 [accessed on 05 August 2016]. [72] The office of Robin Chapple MLC. Western Australian Greenhouse Gas Estimates 2012. ; 2012 [accessed on 05 August 2016]. [73] Bai H, Qiao S, Liu T, Zhang Y, Xu H. An inquiry into inter-provincial carbon emission difference in China: aiming to differentiated KPIs for provincial low carbon development. Ecol Ind 2016;60:754–65. [74] Hong Kong Government. Hong Kong’s Trade Policy. ; 2016 [accessed on 05 August 2016]. [75] Wackernagel M, Rees W. Our ecological footprint: reducing human impact on the earth. New Society Publishers; 1998.

11

[76] Hong Kong Government. Trade statistics. ; 2014 [accessed on 05 August 2016]. [77] Shanghai International Shipping Institute. China shipping database 2014, ; 2014 [accessed on 05 August 2016]. [78] Shanghai Government. Shanghai 12-five-year plan on energy saving and climate change. ; 2012 [accessed on 05 August 2016]. [79] Florida R. London Beats New York as the World’s leading financial center in the latest ranking. ; 2016 [accessed on 05 August 2016]. [80] Sargento AL, Ramos PN, Hewings GJ. Inter-regional trade flow estimation through non-survey models: an empirical assessment. Econ Syst Res 2012;24:173–93. [81] Government HK. Climate change initiatives in Hong Kong. ; 2016 [accessed on 05 August 2016]. [82] China Government. China’s twelfth five-year plan (FYP). ; 2011 [accessed on 05 August 2016]. [83] Zhang Z. Carbon emissions trading in China: the evolution from pilots to a nationwide scheme. Centre for Climate Economics & Policy, Crawford School of Public Policy, The Australian National University; 2015. [84] Xiong L, Shen B, Qi S, Price L, Ye B. The allowance mechanism of China’s carbon trading pilots: a comparative analysis with schemes in EU and California. Appl Energy 2016. http://dx.doi.org/10.1016/j.apenergy.2016.01.064. [85] Zhou P, Zhang L, Zhou DQ, Xia WJ. Modeling economic performance of interprovincial CO2 emission reduction quota trading in China. Appl Energy 2013;112:1518–28. [86] Mehta A. China launches nationwide emissions trading scheme. Chemistry world. ; 2016 [accessed on 05 August 2016]. [87] Han G, Olsson M, Hallding K, Lunsford D. China’s carbon emission trading. An Over-view of Current Development; 2012. [88] Adelaide City Council. CARBON NEUTRAL ADELAIDE: a shared vision for the world’s first carbon neutral city. ; 2015 [accessed on 05 August 2016]. [89] Australian Government. Carbon neutral program guidelines: National carbon offset standard. Version 4.0. ; 2015 [accessed on 05 August 2016]. [90] Australian Government. The safeguard mechanism. ; 2016 [accessed on 05 August 2016]. [91] Theguardian. Australia signs up for clear carbon trading rules, hinting at policy change. ; 2016 [accessed on 05 August 2016]. [92] Yu Y, Feng K, Hubacek K, Sun L. Global implications of China’s future food consumption. J Ind Ecol 2016;20:593–602. [93] Wang S, Fang C, Guan X, Pang B, Ma H. Urbanisation, energy consumption, and carbon dioxide emissions in China: a panel data analysis of China’s provinces. Appl Energy 2014;136:738–49. [94] Levine MD, Aden NT. Global carbon emissions in the coming decades: the case of China. Ann Rev Environ Resour 2008;33:19–38. [95] Liu J, Diamond J. China’s environment in a globalizing world. Nature 2005;435:1179–86.

Please cite this article in press as: Chen G et al. Transnational city carbon footprint networks – Exploring carbon links between Australian and Chinese cities. Appl Energy (2016), http://dx.doi.org/10.1016/j.apenergy.2016.08.053