Case Studies of Urban Metabolism: What Should be Addressed Next?

This paper analyzes case studies of Urban Metabolism (UM), an interdisciplinary field that studies the flow of materials and energy in cities. It focuses on global cases to help researchers identify research gaps. I have categorized the studies based on location, scale, and urban system. Two findings need to be specified: first, the geographic distribution of UM case studies is uneven. Only limited studies have been developed for emerging African cities despite expected large future populations. Second, neighborhood-scale cases do not use an appropriate local scale, primarily due to the lack of reliable data sources. Upon noticing concerns over (1) the evaluation of optimized metabolisms, (2) the effectiveness of knowledge transfer, and (3) the awareness of timeframe in delivering practical policy, researchers may now focus on developing more applicable planning and design guidelines while paying attention to the early communication of UM assessment results between scientists and practitioners.


Introduction
To capture the process of how a city "attains resources" and "releases wastes"-while providing "economic outputs and social services"-the Urban Metabolism (UM) concept has inspired ideas for designing sustainable cities (Musango, Currie and Robinson 2017) despite years of being overlooked (Barles 2010). A broad definition from Kennedy, Pincetl and Bunje (2011) specifies UM as "the sum total of the technical and socio-economic processes that occur in cities, resulting in growth, the production of energy, and the elimination of waste" (the view adopted in this paper). The main schools of thought see UM as an interdisciplinary field in which cities are studied as organisms or ecosystems (Golubiewski 2012). This trend furthers several quantitative methods that can be used to assess urban resource flows (Musango, Currie and Robinson 2017). To change the "linear processes" into "networked and cyclical processes" aimed at resource-efficacy, researchers from different backgrounds have widely used UM to assess sustainable policies for cities worldwide. UM is now central to the study of sustainable urbanism.
UM involves diverse quantitative approaches to analyzing how environmental, social, and economic factors interact in urban settings. It is also used to develop cross-disciplinary interventions for various urban systems and thus contributes to urban sustainability. Although challenges remain, resource-efficacy initiatives are an inevitable starting point in attaining the improved urban resource management needed to develop more resilient cities (Agudelo-Vera et al. 2012). Applying UM across disciplines is hence a global movement.
The growth of relevant UM case studies has encouraged researchers to develop literature reviews, with six notable reviews documenting numerous cases to clarify our current understanding of UM (Beloin-Saint-Pierre et al. 2017;Holmes and Pincetl 2012;Pincetl, Bunje and Holmes 2012;Zhang 2013;Zhang, Yang and Yu 2015;Wang et al. 2021). While these reviews categorize cases based upon approaches such as Material Flow Analysis (MFA) and Energy Flow Analysis (EFA), they fail to address the context such as location. In other words, these reviews do not reveal the specifics of how UM has been used, how these cases have benefited cities, or which regions or cities need further study. A comprehensive review that includes context is needed to identify gaps in UM research.
This paper looks at the contexts of UM case studies to reveal how, where, and why cases have been developed. It applies a global perspective to reveal whether cases are evenly distributed; when they are not, it asks why this uneven distribution occurs and where studies should be located in the future. Methodologically, this paper develops a comprehensive categorization regarding location, scale, and urban system. After the introduction, the paper is organized as follows: Materials and Methods depicts the materials and methods; Results presents findings of categorization; Discussion reveals significant places that lack UM case studies and considers city dynamics. My conclusion is provided in the last section.

Materials and Methods
Because this paper aims to categorize case studies, it requires a list of cases for further analysis. I assemble the collected cases of Beloin-Saint-Pierre et al. (2017) and Wang et al. (2021) to take advantage of their contributions. The work by Beloin-Saint-Pierre et al. (2017) is critical as it reviews the extensive literature needed to explain what UM methods should be used in various research contexts. This work addresses those methods that have been empirically applied and provides a "map of methodological choices" (Beloin-Saint-Pierre et al. 2017, p. s232). As such, it becomes critical knowledge for UM research development. The review by Wang et al. (2021) is unique because of the rigorous literature search process used. They have assembled the literature from the Web of Science Core Collection and gathered 1,084 publications dated April 13th, 2020, for analysis. Their work offers an approach that researchers may apply to explore the latest UM cases. As a result, these two reviews covered extensive literature, providing critical contributions to UM research development.
I decide to use their work instead of performing my own review for two reasons. First, Wang and his colleagues' work helped resolve the technical challenge of a literature search with a collection date to 2020, meaning I might not retrieve a list much different from what they had collected. Second, I aim to provide different perspectives to understand the progress in UM research and explore the underlying problem of knowledge production in this field. Analyzing their collections could therefore result in additional knowledge for researchers and decision-makers. These reasons lead me to focus on what they have collected rather than retrieving the literature redundantly on my own.
I then combine these two databases and manually review the title, abstract, and content to remove the irrelevant publications. I exclude these publications because they overlap or do not primarily provide empirical and place-based results from UM approaches, including (1) theoretical papers, perspectives, or white papers, (2) review papers, (3) publications providing limited timeframe, location, and scale information, and (4) technical reports that overly emphasize methodological and technological breakthroughs rather than provide new information about specific geographical areas. I finalize with 472 case studies (journals, book chapters, and theses) and examine them using three central aspects: location, scale, and urban system. The rules for categorization are as follows: • Location: The location dimension shows the distribution of cases. The outcome will point out the places that have not been adequately studied. The locations of cases are divided into 23 regions, including the 22 regions 1 suggested by the United Nations (United Nations, Department of Economic and Social Affairs, Population Division 2015) and one "Global" category-if the studies were comparing metabolisms across regions. • Scale: The scale dimension focuses on the scope of the dataset applied.
Because these scopes can vary significantly from nations to neighborhoods, the scale dimension depicts the concentration of cases among scales, the causes of this concentration, and suggestions to develop more case studies in the future. This paper uses nation, metropolis, and neighborhood as its three categories of scale. 2 If a study has compared different cities, it is considered a case of metropolis, such as the study by Su et al. (2009) that assesses the urban ecosystem health of Chinese cities. Similarly, suppose a study applies metropolitan data for its analysis (like states, provinces, and regions) but does not go above national boundaries. In that case, it is also considered as a case of metropolis. Notably, a study will be assigned as a case of neighborhood only when it specifies a clear research area and its boundary clearly is the scale of a village. • Urban System: This dimension discusses systems such as water, transportation, and food. It describes the spectrum of urban systems investigated. Based upon different objectives, approaches, contexts, and the major findings of cases, thirteen urban systems are identified as 3 (1) Water, (2) Substance, 4 (3) Greenhouse Gases (GHGs), 5 (4) Buildings, 6 (5) General Materials, (6) Comprehensive Systems, 7 (7) Overall Energy, 8 (8) Food, (9) Ecosystem, (10) Transportation, (11) Urbanization, Land Use and Urban Planning, (12) Waste, and (13) Heat.

Results
Based upon the rules in Materials and Methods, the findings derived are addressed (see Table A1 and Figure A1 in the Appendix).

Location
The locational distribution of UM cases is greatly uneven. Eastern Asia, Northern America, Northern Europe, Western Europe, and Southern Europe are the most concentrated regions that account for 74 percent of UM cases. In contrast, four regions remain to be investigated, and four regions have less than five cases 9 (mostly in Africa). A global review highlights those regions that should be prioritized in future research. Uneven distribution is also visible within the same region, such as the concentration of cases in specific cities. For example, Beijing, Toronto, Paris have been frequently studied in their respective regions. Upon reviewing the major contributors and institutions within these cities, I have found Prof. Yan Zhang at Beijing Normal University in Beijing, Prof. Christopher Kennedy at University of Toronto in Toronto, and Prof. Sabine Barles at Université Paris 1 Panthéon-Sorbonne in Paris (see Table A2 in the Appendix for more details). These cities have been studied due to the robustness of local or adjacent research institutions. In other words, the presence of a research institution influences the richness of UM cases among regions. In this regard, investing more research resources in pioneering institutions could initiate more UM cases.
Using UM to compare global cities is a growing research concern; furthermore, data availability can determine if a city will be chosen for global comparison. The frequency of compared cities (see Table A3 in the Appendix) shows that London is the most popular city appearing in six cases (Goldstein et al. 2013;Kennedy et al. 2009Kennedy et al. , 2010Sovacool and Brown 2010) while Los Angeles (County) and New York City are the two cities presenting in four cases. To explain London's popularity, I further looked at the data resources of local cases. Three data resources (Chartered Institution of Wastes Management and Environmental Body 2002; Greater London Authority 2007; Mayor of London 2006) were used at least twice in the UM case studies. Of note, these data resources were developed either by a governmental office or research institutions to address carbon issues (see Table A4 in the Appendix). Such a finding suggests that when a city proactively develops official reports or encourages other institutions to conduct thematic studies, the robust data resources could inspire researchers to include such a city in their future research.

Scale
About 82 percent of UM cases are metropolitan-scale, while national and neighborhood cases are relatively rare. This paper argues that the popularity of metropolitan cases is related to the accessibility of data and the applicability of research outcomes. For example, metropolitan data is often accessible for researchers because it is readily derived from local governments. Therefore, metropolitan case studies dominate either due to the availability of data or the influence of research outcomes.
This finding also highlights the scarcity of national and neighborhood cases. Only 9 percent of UM cases are national-scale, and they often involve global comparisons. These cases are categorized as national cases because they use national data to present international or transregional comparisons among urban systems. Examples include the assessment of Folke et al. of the Ecological Footprint (EF) of the twenty-nine largest cities of Baltic Europe (1997), and Yi et al. study (2007), which applies Life Cycle Assessment (LCA) to evaluate environmental impacts of regional activities in Japan. More recently, Ciacci  By contrast, the scarcity of neighborhood cases is often related to data resources. For example, a city's population can be easily accessed if the census has been updated regularly; however, determining these demographics can be difficult when researchers have to complete the census by themselves. It can be even more challenging if they simultaneously hope to access other neighborhood data such as materials and energy. This disadvantage has limited the development of neighborhood cases and resulted in about 8 percent of UM cases in this paper.

Urban System
The categorization of the urban system dimension shows that (1) Overall Energy, (2) General Materials, and (3) Comprehensive Systems account for 46 percent of UM cases. Regarding this concentration, the following section addresses (1) details of the concentrated categories and (2) growing cases in GHGs category.
The Concentrated Categories. In this paper, cases are added to the General Materials category when applying the MFA approach. While these case studies share an identical approach, they differ in the types of materials included. For example, the materials that Schulz includes 10 (2007) (2013), and Swilling addresses Cape Town's infrastructure and highlights how ecological issues have been ignored in the city's policy making (2006). Such an approach is still actively used today (Baabou et al. 2017;Lu and Chen 2017).
I also find cases using approaches other than EF to consider both material and energy flows. For instance, Conke and Ferreira (2015) study material and energy use in Curitiba from 2000 to 2010 and demonstrate how the city has contributed to sustainable development in South America. Other cases include Ngo and Pataki's study (2008) that explores Los Angeles County's energy and mass balance, and Pincetl and her team's effort (2014) that reveals urban flows and sinks in the Los Angeles region. Although approaches in these cases may not be generalizable because they are tailored for different purposes, these cases still share the common goal of increasing environmental sustainability. UM provides a clear foundation in all of them.
The Growing GHGs Category. Case studies of GHGs or Carbon flow are increasing in number; this discourse has transitioned from a focus on a single city (Baldasano, Soriano and Boada 1999;Harvey 1993) to comparisons between cities (Kennedy et al. 2009(Kennedy et al. , 2017Sovacool and Brown 2010). As one of the earliest cases, Harvey's study (1993) responded to Toronto's ambitious plan for CO 2 reduction. His work highlights how the community energy-saving program and improved land-use planning can reduce pollution. While GHGs reduction has become a shared value, numerous other case studies have addressed how cities can mitigate climate change, including Baldasano, Soriano, and Boada's survey (1999) that uncovered the GHG emissions in Barcelona from 1987to 1994, Xia and colleagues' effort (2015 that clarifies the reasons behind the increase of GHG emissions in Beijing, and finally Zheng, Fath, and Zhang's examination (2017) of embodied energy consumption and energy-related carbon footprints in the Jing-Jin-Ji region. Because they focused on one specific city or city-region, these studies are considered a part of the typical GHGs category in this paper.
In addition to the cases focusing on a single city, others look at comparisons between cities. These cases began to emerge around 2010, such as Hillman and Ramaswami's study (2010) that evaluates GHG emission footprints in eight US cities, other professionals' effort (2010, 2017) that assesses GHG emissions of global cities or city-regions, and Sovacool and Brown's work (2010) that presents carbon footprints of twelve metropolitan areas. The growing number of such cases indicates that global comparison is a central direction for future research while it emphasizes the value of collaboration among cases. Such collaborations are expected to be more frequent in the future, which surely will not be restricted to GHGs issues.

Discussion
In Results, this paper presents findings based upon the categorization of UM cases. Along with the above findings, this section further interprets places in need of UM case studies and the importance of including city dynamics in UM cases.

Places in Need of UM Cases
This paper highlights the regions with no UM cases to date, mostly in Africa, Central Asia, and Eastern Europe. Meanwhile, although some regions account for numerous cases, these studies often focus on specific cities such as Beijing or Toronto. This outcome relates to the series of cases developed by iconic research institutions. Namely, an uneven distribution of research resources has led to the uneven distribution of UM cases. Furthermore, when considering why some cities have been frequently selected for global comparison (see Table A3 and A4), this paper finds a positive correlation between governmental investigation and academic output. UM researchers may develop more cases focusing on a city with more official reports and thematic studies.
Therefore, these findings point out opportunities for increasing UM cases, especially for regions with limited cases today and are expected to be highly populated in the future. According to the World Population Prospects from the UN (United Nations, Department of Economic and Social Affairs, Population Division 2017), Africa is expected to be the fastest-growing continent in the following decades. Considering the emerging megacities in Africa, redistributing research resources to this continent is reasonable. Meanwhile, developing these studies is also urgently needed because the knowledge generated is tied to the local context and may not be transferable. As a result, if future research can aggressively assess the metabolisms of these emerging cities, the findings may help these cities grow in a sustainable manner.
In this vein, an opportunity arises to strengthen the alliance between China and African countries, thereby allowing them to shift from their roles as business partners (Lee 2018) to one of knowledge exchange. Of note, a handful of Chinese scholars have developed UM research through the integration of (1) the extensive data resources, including statistical yearbook (Gao et al. 2020), planning documents (Chen, Wen and Wang 2020), or land use types (Xia et al. 2019a), (2) the suitable coefficients for data conversion, such as integrating material flows into energy system models for a more consistent assessment Liu et al. 2019;Wang, Zhang and Yu 2019;Xu et al. 2020), and (3) advanced techniques of simulation, such as design scenario model (Yin 2019), product system model (Qi et al. 2019), 4D-GIS model ), or network model (Xia et al. 2019b). The proposed knowledge exchange could start from these three components and boost different academic alliance modes for establishing African environments that will foster more comprehensive and longitudinal UM research. Indeed, these various mechanisms could be further elaborated. Yet, the disparity in UM case studies that this paper presents should serve as a critical starting point for the proposed new partnership.
Furthermore, increasing UM cases require cities to share more reliable data from an official perspective. London, for example, is one of the most frequently studied cities because of the reports it has released (Chartered Institution of Wastes Management and Environmental Body 2002; Greater London Authority 2007; Mayor of London 2006). Cities looking for UM studies may also draw inspiration from this finding. However, to practice this proposition in cities with limited capacities, I suggest city leaders pursue more cost-efficient techniques to monitor the urban performance of their interests that more specifically speak to emerging challenges like water consumption or energy efficiency. This focus can target, but is not limited to, assembling more reliable information through affordable and accessible approaches. This paper finds that 82 percent of UM cases are of a metropolitan scale based on the scale dimension. Metropolitan cases have prevailed because of the accessibility of data resources and the applicability of research outcomes. Future research is encouraged to focus on innovations in data collection (Browne, O'Regan and Moles 2011;Codoban and Kennedy 2008) as this would facilitate the collection process and increase the reliability of data resources. An emerging trend is integrating Geographic Information System and Remote Sensing to reveal metabolism processes (Miatto et al. 2019;Xia et al. 2018). The breakthroughs in this field should facilitate progress in data collection, while boosting the development and application of UM research, by making the findings spatially explicit.

Dynamic Progress of City Making
The timeframe affects both the approach and applicability of a given UM case study. First, we might gather some insights from an earlier work by Codoban and Kennedy (2008). The case by Codoban and Kennedy considers that neighborhoods may not be in a continual state of construction when compared with cities (2008, p. 21). They evaluate their existing neighborhoods without consideration for the construction materials. Hence, their findings reveal the "operation" instead of the "remaking" of neighborhoods. Their design guidelines for prioritizing the rebuilding of existing building stock over retrofitting (2008, p. 29) are improper because reconstruction costs have not been investigated. Reflecting on such work indicates that certain design guidelines might be reversed had the dynamic process been better captured. For example, while addressing the decision between rebuilding and retrofitting existing buildings, I argue that retrofits should be preferred based on long-run improvement. Through retrofitting, the reuse of existing buildings will reduce the consumption of materials and energy at the start. Neighborhood metabolism can be improved without the additional cost of rebuilding. Moreover, if such reuse can be integrated with new services and functions, the benefits will be even more significant. As a result, without considering the dynamic process of citymaking, the design guidelines remain unconvincing.
Considering such a dynamic will reveal the need to understand the timeframe of implementations. While delivering design guidelines, earlier UM cases might not clarify which should be carried out quickly and which are long-term goals. The timeframe was not discussed. This absence could be related to the fact that research has not yet addressed what an optimized metabolism indicates. As a result, although UM research was set to manage the operations of urban systems, earlier works did not clarify under what conditions a system would be considered sustainable.
The development of recent neighborhood UM studies helps specify how the above concern may be addressed, including the works from Chrysoulakis et al. (2013) and Yin (2019)-one at the planning scale and another at the design scale. At the planning scale, Chrysoulakis and colleagues outline the approach taken in the FP7 (European Union's 7th Framework Program for Research and Technological Development) BRIDGE (sustainaBle uRban plannIng Decision support accountinG for urban mEtabolism) project to develop a Decision Support System (DSS) for sustainable urban planning. The BRIDGE project provides a structured assessment of UM processes in different planning alternatives/scenarios. Building on such a foundation, the developed DSS allows end-users (such as urban planners, architects, and engineers) to modify sustainability objectives and indicator weights regarding a specific future scenario and then generate separate spider diagrams for each scenario. The derived results, in turn, provide information for end-users to determine planning decisions.
At the design scale, the study by Yin (2019) shows how design guidelines can be generated through UM research, based mainly on the Chinese context. Yin quantifies design methods and strategies, with an analysis of urban resource flows, to decide on the design proposal for the China World Trade Center area in Beijing. It involves four phases: urban status analysis, design scenario setting, design alternatives, and design evaluation. Using a three-dimensional urban space model as its working platform, Yin measures the water, energy, organic waste, and food flows of different design alternatives. These alternatives come with seven parameters: urban density, coverage rate, architectural form, open space, street attributes, land use, and underground space. In other words, once UM research further applies the metabolic processes of current urban status and the anticipated outcomes of different design alternatives, we can better choose the design proposal.
Although the studies from Chrysoulakis et al. (2013) and Yin (2019) demonstrate how simulation techniques are being progressively adopted to explore UM processes at planning and design scales, some challenges remain that limit UM knowledge's applicability in planning policy practice. In inquiring about the effectiveness of knowledge transfer between UM scientists and practitioners, Perrotti uses the BRIDGE project to retrospectively evaluate the relevance and impact of UM studies in urban planning (2019). On the one hand, she finds that the practitioners acknowledge the relevance of such retrospective evaluation and the applicability of the BRIDGE project in their practice. On the other hand, the scientists show moderate interest in engaging in her evaluation, feeling that BRIDGE was already "behind them" (Perrotti 2019(Perrotti , p. 1470. She also summarizes some opportunities that will lead to more effective knowledge transfer: (1) clarifying the connection between scientific research and the practical aspects of urban planning; (2) developing a common language across science and practice; and (3) hosting preliminary consultation to make sure that visions and objectives are shared between parties. The early communication of UM assessment results to practitioners is crucial in influencing the applicability of UM research in planning.
In sum, UM research still faces the challenge of exploring optimized metabolisms, the need to enhance the effectiveness of knowledge transfer from science and practice, and the concern over the timeframe of implementation. Suggestions from Perrotti and other scholars (Binder et al. 2015;Lang et al. 2012) will provide a crucial heads up once the field starts to assess the optimized metabolisms of the emerging megacities (especially in the places addressed in Places in Need of UM Cases). Hopefully, in keeping with their suggestions, we can better transfer UM knowledge to create resource-efficient cities one day.

Limitation
This paper has its limitation regarding the categorization of the scale dimension. Finding that 82 percent of UM cases are metropolitan-scale could be a self-evident result given that my emphasis on "urban metabolism" cases states a clear "urban" focus. Nevertheless, it is also worth noting that I assembled my UM literature from reviews by Beloin-Saint-Pierre et al. (2017) and Wang et al. (2021). The cases I have categorized at the national or neighborhood level were included in their reviews as UM literature. In this vein, it seems that "UM" per se is deemed an umbrella term broadly perceived within the concept rather than its spatial interpretation. I recognize such a notion while looking to future research to discuss the issue of scale in the UM review.

Conclusions
Because previous literature reviews on UM research (Beloin-Saint-Pierre et al. 2017;Holmes and Pincetl 2012;Pincetl, Bunje and Holmes 2012;Zhang 2013;Zhang, Yang and Yu 2015;Wang et al. 2021) often categorize studies based on the approaches applied rather than the contexts addressedsuch as location and urban system-I present a systematic review of the context of UM cases. I have developed my categorization based on the latest cases collected by Beloin-Saint-Pierre et al. (2017) and Wang et al. (2021), considering their contributions. The findings highlight that Africa, Central Asia, and Eastern Europe are the primary uninvestigated regions globally. Meanwhile, even within regions with rich academic output, their cases focus on specific cities or city-regions because these places have robust research institutions to support their work. Moreover, some megacities, such as London, are frequently investigated because they have more reliable official reports and thematic studies useful to researchers. The location distribution of UM cases hence is extremely uneven. In addressing this disadvantage, this paper proposes a redistribution of research resources to places needing UM case studies, especially emerging cities in Africa. Such a redistribution should inspire more cases and support sustainable development in cities. On the other hand, city leaders could pursue more cost-efficient techniques to monitor the urban performance of their interests by targeting more reliable information through affordable and accessible approaches.
I also point out that future neighborhood cases should focus on developing more applicable planning and design guidelines while paying attention to the early communication of UM assessment results between scientists and practitioners. To consider the dynamic progress of city-making matters, I have reviewed the works from Codoban and Kennedy (2008), Chrysoulakis et al. (2013), Yin (2019), and Perrotti (2019). I noticed concerns over the evaluation of optimized metabolisms, the effectiveness of knowledge transfer, and the awareness of the timeframe in delivering practical policy. That said, suggestions derived from UM case studies will remain unconvincing if studies do not relate their insights to the dynamics of city growth. Future research should be aware of accommodating visions and objectives between academia and industry while assessing optimized metabolisms of cities; such efforts will contribute to establishing desired benchmarks for sustainable urbanism.
Lastly, I believe that collecting UM case studies and displaying their distribution is critical to providing helpful information for urban planning, design, and policymaking, especially when cities and neighborhoods share similar urban settings, cultures, challenges, and opportunities. An example would be comparing Accra with Johannesburg. Aside from understanding that their metabolic processes should differ based on local contexts, they are, in many ways, more similar to each other when comparing them with an Asian city. As a starting point, the insights derived from this collection allow adjacent places to learn from each other, and more specifically by following these steps. First comes recognizing which of them has a better metabolic process. Second is understanding what urban policies have been adopted for such outcomes, like lifestyle changes. Thirdly, evaluate the results once replicating such policies. Hopefully, this paper imparts a sense of why collecting UM case studies matters and explains how urban planning, design, and policymaking may benefit afterward. Zhang, Yang and Yu (2015, p. 11252) uses (1) household,

A previous review by
(2) neighborhood, (3) urban, (4) regional, (5) social, and (6) anthroposphere to illustrate the multiple scales of metabolism; this paper, however, considers that not all six scales can be commonly adopted in cases. It hence simplifies them into three scales. 3. This dimension looks at contexts and purposes of each case for categorization. Therefore, reviewing UM approaches such as MFA is merely one of the factors. For example, suppose a case applies an energy analysis, an approach of EFA that converts different forms of material and energy into one unified energy unit (seJ), to understand the urbanization process. In that case, this case is considered "Urbanization, Land Use and Urban Planning," not "Overall Energy." 4. This refers to cases that track specific substances, such as Nitrogen (Baker et al. 2001), Copper (Van Beers andGraedel 2003), nutrient (Faerge, Magid and de Vries 2001), and Phosphorus (Nilsson 1995). It is also understood as the "partial macro-MFA" (Ferrão and Fernandez 2013, p. 80). 5. This category tracks greenhouse gases (GHGs) or carbon flow. Carbon tracking was originally categorized as Substance category because it tracks a specific substance. However, since carbon tracking also addresses climate issues, this review regards carbon tracking as part of the GHGs category. 6. This category discusses the energy consumption of households, either the lifecycle of buildings (Balocco et al. 2004) or the energy demand of households (Lenzen, Dey and Foran 2004). 7. This category indicates cases that cover both material and energy sectors for assessment. 8. Including two types of research: (1) addressing energy issues such as generation, transmission, or consumption as its research subject, or (2) applying EFA, such as emergy or exergy, to synthesize both material and energy for analysis. 9. Regions have no cases: Western Africa, Central Asia, Melanesia, and Micronesia; regions have less than five cases: Eastern Africa, Middle Africa, Northern Africa, and Eastern Europe. 10. Including products, industrial minerals, construction minerals, fossil fuels, and biomass. 11. Including food, clothing and textiles, paper and printed material, non-durable household goods and miscellaneous plastic products, household durable goods, total manufactured goods, residential fuel and electricity, and motor fuel.