Accounting and management of city carbon emissions: Trajectories towards advanced data use

Accounting and management of city carbon emissions: Trajectories towards advanced data use Will Brown * , Kristen MacAskill University of Cambridge, Centre for Sustainable Development, Department of Engineering, Trumpington Street, Cambridge CB2 1PZ, United Kingdom ARTICLE INFO Keywords: Urban carbon accounting Carbon emissions Data Literature review ABSTRACT Cities simultaneously play a significant role in the production of global carbon emissions, whilst being essential in facilitating their reduction. It is common practice for cities to monitor, assess and manage their carbon emissions to inform approaches to reducing emissions. Whilst carbon accounting practice is evolving alongside accessibility of data and availability of resources, city authorities remain limited in their ability to accurately assess the carbon emissions produced within their respective city – often being constrained by political, financial and capacity factors. This paper reviews academic literature concerning the utilisation of carbon emission data to establish trajectories for advancing urban carbon accounting practice. The cases reviewed expand more typical contemporary urban carbon accounting practice through at least one of five trajectories: combining different data sources, visualising carbon emissions, developing district level mitigation policies, accounting for project level emissions, or estimating policy emission impacts. These are discussed in relation to broader narratives within urban carbon accounting research. Each trajectory offers potential development routes for city authorities as their urban carbon accounting practices mature, thereby working towards closing the gap between carbon accounting practice and academic research. 1. Introduction Cities are simultaneously responsible for an estimated 70 % of the world’s emissions (IPCC, 2023) whilst being critically important to the reduction of global carbon emissions. City authorities have been considered to be more active in pursuing mitigation policies when compared to national governments (Lo, 2014), supported by their local knowledge, creativity, and resources to reduce them (Kennedy et al., 2009). However, cities are fundamentally complex (Alberti et al., 2018), which produces significant challenges for urban governance and policymaking (McPhearson et al., 2016). This complexity increases the likelihood of undesirable outcomes emerging from well-intentioned interventions (Pak et al., 2017), thereby requiring systemic approaches to consider the connections between different systems (Carr, 1996) and the utilisation of carbon emission data to inform and enhance decision making. Whilst many city authorities are actively working to account for carbon emissions, their actions are limited. City emission baselines are predominately constructed through accounting for scope 1 and scope 2 emissions (GPC, 2021), whilst scope 3 emissions, often the most significant urban emission sources (Goodwin et al., 2023), are seldom accounted for. This is owing to the complexities of accounting for them, with city authorities lacking direct control over these emission sources and reliable data to account for them (Millward-Hopkins et al., 2017), as well as complexities around how to measure a city’s consumption of goods and services; the driver of urban scope 3 emissions (Dorr et al., 2022). Another element to consider within urban carbon emission reduction is the maturity of a city’s ability to conduct carbon accounting and utilise emission data to inform decision making. City authorities often utilise city-wide carbon emission inventories, usually produced through a combination of approaches, including: the ‘downscaling’ of national carbon emission data, produced via the use of input–output methods based on national economic data (Davey, 2025); leveraging connections with local utility providers to estimate household emissions (Baltar de Souza Le˜ ao et al., 2020); and, estimating transportation emissions via the application of an emission factor to city-wide fuel sales data (Kongboon et al., 2022). Whilst city-wide emission data is useful at the city-wide scale, it is limited in application at smaller scales, owing to the complexity of * Corresponding author. E-mail address: [email protected] (W. Brown). Contents lists available at ScienceDirect Sustainable Cities and Society journal homepage: www.elsevier.com/locate/scs https://doi.org/10.1016/j.scs.2025.106677 Received 2 December 2024; Received in revised form 24 July 2025; Accepted 24 July 2025 Sustainable Cities and Society 131 (2025) 106677 Available online 25 July 2025 2210-6707/© 2025 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY license ( http://creativecommons.org/licenses/by/4.0/ ). accurately ‘downscaling’ the data (Gately & Hutyra, 2017). Whilst there are examples of cities pursuing more advanced carbon accounting, the minimal standard of urban carbon accounting (GPC, 2021) requires cities to develop city-wide carbon emission inventories (Davey, 2025), resulting in this level of accounting being the most prevalent globally. To address these limitations, a literature review of academic research on the application of carbon emission data in cities has been conducted. In comparison to the current practices adopted by many cities such research represents the ‘state of the art’, defined as ‘very modern and using the most recent ideas and methods’ (Cambridge Dictionary, 2025). This paper demonstrates to city authorities, practitioners, and academia potential trajectories associated with advancing carbon accounting in the urban realm. These trajectories characterise the application and use of carbon emissions data, marking an important contribution towards bridging the gap between academic research and carbon accounting practice applied by city authorities. This paper has conducted a literature review to surface the different approaches to using urban carbon emission data by reviewing 41 case studies covering different cities from across the globe. During the review process, five distinct trajectories were revealed, each of which can be pursued by cities to develop their carbon accounting practice. Each is discussed in relation to the relevant literature in Section 3, with the article concluding by discussing the key insights derived from this review. The next section provides an overview of the methodological approach used. 2. Methodology This literature review is an ‘extending review’ (Xiao & Watson, 2019) and whilst being structured around a rigorous search process, it is intentionally not a systematic review. This is owing to the purpose of the search. A typical goal of a systematic literature reviews is to quantify the results of the literature search, here the authors sought to identify academic articles which place a detailed focus upon the application of carbon emission data in the urban realm. Key terms from the Global Protocol for Community-Scale Greenhouse Gas Inventories (GPC) (GPC, 2021), the leading standard for structuring urban carbon emission inventories, used by over 13,000 cities worldwide (Global Covenant of Mayors, 2024), were applied as a search term framework (Dixon-Woods, 2011). This standard is premised upon the identification of six sectors: Stationary energy, Transportation, Waste, Industrial processes and product use, Agriculture, forestry, and other land use and other scope 3 (GPC, 2021). These sectors were used as initial search terms (the Web of Science journal database was utilised for all searches) with eight individual searches being conducted by applying either Energy, Transport, Waste, Industrial, Product Use, Agriculture, Forestry or Scope 3 to the below search term: “Urban [GPC sector] Carbon Emissions” A review process focused on long-listing city-specific case studies. This selection criteria was included to reflect the focus on city approaches to carbon accounting. Subsequent refinements (described in Fig. 1). led to 41 articles for review. A paper was selected for review if it featured a city-specific case study where carbon accounting was conducted, with the resulting emission data being applied to produce more nuanced or detailed findings beyond the initial creation of an inventory. The five manifestations of this data application formed the trajectories discussed in Section 3. An example of a leading paper which did not meet this requirement was Lenk et al.’s study comparing territorial and consumption carbon accounting in Berlin, producing an emission inventory which was not subsequently applied to the city to produce more nuanced or detailed findings (Lenk et al., 2021). This paper and others like it (See: Appendix) were removed at the ‘refinement 4 ′ stage. Conversely, Marchi et al. (2023) for example, produced similar insights into the Italian municipality of Grosseto, but used them to map the emission profiles of different districts and proposing specific policies to combat them (Marchi et al., 2023). This development of the five trajectories (detailed in Section 3) was supported by developing a series of codes which focused on the context, use of emission data, methods used, governance, urban environment and type of transition investigated within each paper. This review includes older papers dating back to 2010. Despite the relative vintage of this research, the methods included remain more advanced than the prevailing norm for urban carbon accounting today. Throughout, the term ‘carbon emissions’ is used to describe the gasses being accounted for, owing to all articles reviewed either solely, or in combination with other greenhouse gasses, accounting for carbon dioxide emissions. There is the potential for the selective reporting of cases, where less successful examples or attempts are less likely to be formally reported. The role of this paper is to capture potential trajectories of advancing urban carbon accounting. In doing so it does not claim to explore all the complexities of their implementation or provide a comprehensive review of current practices used within cities. The authors are aware of numerous governance barriers, beyond issues relating to data, which can manifest themselves within the context of advancing urban carbon accounting. The European Union’s Net Zero Cities project has identified three barriers which prevent the advancement of carbon neutrality – institutional, behavioural and infrastructural barriers (Net Zero Cities, 2024). A reflection of urban governance and capacity limitations is provided in the discussion section, but concerns of governance are not a central focus of this review. Another element to raise is that whilst different technological approaches to carbon accounting are discussed in the context of setting the scene for different case-studies, the scope of this paper is to focus upon the potential trajectories of future applications of carbon emission data within cities, rather than on the merits of different technologies in producing emission data. Therefore, the different forms of sensing technologies and techniques are solely discussed in the context of their application to produce carbon emission data. 3. Findings Within the 41 case studies (summarised thematically in Table 2) most of the reviewed articles use the city as a case study, rather than being a representation of the actions of a city authority – although there are examples of collaboration between researcher and authority (Goodwin et al., 2023; Keller et al., 2023; Marchi et al., 2023). There is a diversity of cities represented, demonstrated by Fig. 2, which maps each case study and carbon accounting trajectory onto Loughborough University’s Globalization and World Cities research network’s (GaWC) ranking (GaWC, 2024). This ranking classifies cities into 12 levels of world city network integration, from alpha ++ (the most economically integrated cities) through to sufficiency (cities with their own services that are not reliant on other world cities), based on the prevalence of leading financial firms within the city and the connections between them. Owing to the multiplicity of factors which make a particular city attractive for investment (Murray, 2022) the world city ranking provides a multifaceted means of categorising the city case studies reviewed here. While alpha cities slightly dominate the analysis, there is generally a reasonable spread across the city rankings, suggesting the findings of the paper could be generalized beyond a particular city type. Although, the significant majority of cases are based in high-income countries, a point that will be returned to in wider discussion in Section 4. The majority of case studies included in this review are based upon bottom-up accounting practices as opposed to the top-down approach adopted by many cities today. This distinction is important considering the time and skillset required to conduct a bottom-up analysis in comparison to the relative simplicity of a top-down assessment (Davey, 2025), especially within the constraints of urban governance. The five trajectories are presented within the context of a starting point scenario, with this paper exploring how each trajectory builds on this scenario. W. Brown and K. MacAskill Sustainable Cities and Society 131 (2025) 106677 2 While there is variability in existing practice, for the purpose of this paper this scenario (Table 1) is framed as a city currently conducting the ‘BASIC’ level of carbon accounting contained within the GPC standard (scope 1 and 2 emissions for stationary energy, transportation and waste) using either top-down or bottom-up accounting. Owing to the less labour-intensive nature of ‘top-down’ carbon accounting, this starting point will be based on a city performing top-down accounting. The first two trajectories centre upon better ways to utilise preexisting data, firstly by combining emission data with other data sources to support more enhanced, holistic decision making, and secondly through detailing approaches to the visualisation of carbon emissions. The trajectories subsequently expand from current practice by applying emission data in a more targeted manner to develop neighbourhood/ district level carbon reduction policies (rather than at a city-wide scale), through conducting carbon accounting of project emissions to assess their emission reducing potential and impact, and by estimating the emission impacts of proposed and already existing governmental actions through the carbon accounting of policy initiatives. Each trajectory is discussed through the relevant cases. Initially with a thematic overview involving a brief description of case studies and a reflection upon urban carbon accounting maturity, before delving deeper into a featured case study. This expanded analysis unpicks specific thematic elements, with supporting cases that more widely demonstrate application of the trajectory in question. 3.1. Trajectory 1: Combining data sources A crucial element of enhancing urban carbon accounting is the alignment and combination of different data sources. Be this through: the combination of different scales of top-down emission data (citywide, regional and national), as demonstrated by Wiedmann et al. (2016) development of Melbourne’s ‘carbon map’, a chart-based tool which demonstrates the link between emission sources and consumption within the city; the implementation of bottom-up accounting practices, whereby activity data is utilised to calculate and estimate a city’s carbon emissions via the application of an emission factor (Dorr et al., 2022; McPherson & Kendall, 2014; Sigurðard´ ottir et al., 2023); the combination of socio-economic or census data to further interrogate the production of urban emissions (Goodwin et al., 2023; Millward-Hopkins et al., 2017; Patarasuk et al., 2016); or the use of carbon emission data with economic or financial data to support the prioritisation of carbon neutrality efforts within a city. A New York case study (explored below) (Eicker et al., 2020) not only demonstrates an approach to combining different forms of urban data to support decision making, but also proposes an approach to developing better data integration. 3.1.1. New York Whilst America’s largest city has passed legislation to reduce their carbon emissions by 80 % until 2050, it is essential that the data utilised to inform mitigation policies reflects New York’s dense complexity, entailing a diversity of data forms, sources and applications. Eicker et al. Fig. 1. Diagram for literature search process. W. Brown and K. MacAskill Sustainable Cities and Society 131 (2025) 106677 3 Fig. 2. Reviewed papers’ case study city world city rankings mapped according to key trajectory themes (some of the 41 cases feature more than once). Cities listed as ‘not included’ (purple) are deemed to be non-world cities by GaWC and therefore do not feature in the world city ranking. W. Brown and K. MacAskill Sustainable Cities and Society 131 (2025) 106677 4 (2020) identified these challenges facing New York and proposed the development an open urban data platform, which centres upon urban data collection and analysis from a diverse range of sources, such as sensors, municipal data records, knowledge repositories and social media streams. Yet, these are often siloed, disconnected from each other, rarely exchanging data. Therefore, the role of an open urban data platform is to interconnect data from multiple urban infrastructures – such as built environment, transportation, natural environment and utilities. To overcome data siloisation, the CityGML building modelling standard was utilised within the platform as a common format for data exchange between datasets. The platform analyses and optimises urban infrastructures (energy, water and food or goods consumption) and was applied to investigate the food-water-energy nexus of Borough Hall (Brooklyn). Open-source data related to the energy demand of the building, food and water sector was collected and combined, supporting a range of nuanced policy recommendations, with the authors arguing that districts with a high ratio of home ownership and high-income population could afford to modernise their built environment more easily in comparison with low income or city-owned apartments, where other approaches are needed. Another insight concerns the importance of building age, where informed refurbishment and system modernization decisions can be made. However, as is argued byEicker et al. (2020), the use of multiple data sources can be challenging, because they are not always on the same level and scale and are usually in different data Table 1 Overview of starting point city scenario. Element City Scenario Data Source Top-down national level data downscaled to city Scope GPC scopes 1 and 2 - BASIC accounting Advancements Limited awareness of potential carbon accounting advancements/ next steps +Limited data availability and expertise to conduct bottom-up accounting Data Use No connection between decarbonisation actions and currently available carbon emission data Table 2 Summary of the trajectories and the 41 shortlisted cases. Trajectory Sub-Theme Featured Case Study Other Relevant Cases Combining Data Sources Combining Data Sources Eicker, Weiler, Schumacher and Braun (2020) - New York Demonstrates the combination of different forms of urban data to support decision making, proposes an approach to developing better data integration Dorr et al. (2022) Montreuil, McPherson and Kendall (2014) Los Angeles, Sigurðard´ ottir et al. (2023) Reykjavik, Patarasuk et al. (2016) Salt Lake City, Goodwin et al. (2023) Canberra, Millward-Hopkins et al. (2017) Bristol, Wiedmann et al. (2021) Melbourne Mapping Emissions District/ Neighbourhood Mapping Tan et al. (2021) - Shenzhen Demonstrates the mapping of emissions at district level and across different communities. Marchi et al. (2023) Grosseto, Han & Ge (2021) Suzhou Fine-Scale Mapping Keller et al. (2022) - Auckland Represents the fine-scale carbon emissions of a geographically complex area, demonstrating the need for this type of approach. Cai et al. (2020) Hong Kong, Zhou and Gurney (2010) Indianapolis, Boitor et al. (2019) Cluj-Napoca, Schandl, Marcos-Martinez, Baynes, Yu, Miatto and Tan (2020) Canberra Representing Building Level Emissions Zhang et al. (2022) - Xi’an Models building emissions across a city, identifying themes which expand beyond political boundaries Kellett et al. (2013) Vancouver, Wu, Tao, Cao, Fan and Ramaswami (2019) Shanghai, Lorenzo-S´ aez et al. (2020) Quart de Poblet Mapping Directly Monitored Emissions Doukalianou et al. (2020) - Xanthi Lee et al. (2017) - Vancouver Pugliese et al. (2018) - Toronto Granular Neighbourhood /District Specific Policies Granular Neighbourhood /District Specific Policies Cheng et al. (2022) - Chongqing An informative example of creating granular policies which reflect the specific contexts of the areas they are proposed for. Rivera-Marín, Alfonso-Solar, Vargas-Salgado and Catal` a-Mortes (2023) Valencia Marchi et al. (2023) - Grosseto Creates new forms of understanding carbon reduction through devising new districts to focus policies and actions. Dorr et al. (2022) Montreuil Patarasuk et al. (2016) - Salt Lake City Brings in socio-economic data which proposes creating policies along not only physical or energy consumption lines, but also social factors such as wealth. Tan et al. (2021) Shenzen, Kellett et al. (2013) Vancouver, Huang et al. (2017) Adelaide Assessing Project Emissions Built Environment/ Retrofit Kaandorp et al. (2022) - Amsterdam Example of assessing the impact of installing energy efficiency technologies in retrofit scenarios across different contexts. Sigurðard´ ottir et al. (2023) Reykjavik, Lin et al. (2017) Tainan Nature Based Solutions McPherson and Kendall (2014) - Los Angeles An in-depth analysis of the carbon costs, and benefits, of planting trees within the urban realm. Emphasises the importance of systemic benefits of trees, via the reduction of HVAC associated emissions. G´ omez-Villarino et al. (2021) Madrid, Li et al. (2020) Singapore, Liu et al. (2023) Xiamen New Approaches Dong et al. (2014) - Kawasaki An illustrative example of assessing the emission reductions associated with adopting circular economy practices in the urban realm. Kumdokrub et al. (2023) Ithaca, Hsu et al. (2010) Los Angeles Estimating Policy Emissions Estimating Proposed Policy Impacts Dorr et al. (2022) - Montereuil An example of combining the accounting for urban emissions and neighbourhood level analysis. This is combined with estimating the impact of the policies on emissions across the area. (Millward-Hopkins et al., 2017) Bristol, Goodwin et al. (2023) Canberra, Yang, Wang, Liu and Zhou (2018) Ningbo, Sugar and Kennedy (2013) Toronto Assessing Government Policies Chen, Zhang, Chen, Li and Cui (2023) - Nanjing Estimates the effectiveness of national policy on a city’s efforts to become carbon neutral. Further assesses what would be required to achieve this end goal.  Proposes Policies, but does not assess their impacts Lombardi, Laiola, Tricase and Rana (2018) Foggia, (Lwasa, 2017) Kampala, Gu, Fu, Thriveni, Fujita and Ahn (2019) Shanghai, Zhang et al. (2015) Beijing W. Brown and K. MacAskill Sustainable Cities and Society 131 (2025) 106677 5 formats. Combining these diverse data sources however provides information that can be crucial to making sensible scenarios for a future renewable and CO2-reduced city planning. This case study illustrates the benefits of using different forms of data to create more informed analysis of complex urban environments, marking an important step towards the development of holistically informed decarbonisation decisions, and represents the first trajectory towards advancing carbon accounting practice. 3.2. Trajectory 2: Mapping emissions Another development towards enhanced urban carbon accounting is the visualisation of a city’s carbon emissions. Given the imperceptible nature of carbon emissions, the surfacing of their distribution across a city can support more informed decision making (Marchi et al., 2023; Wiedmann et al., 2016). Four forms of emission visualisation were identified in the review, representing: district/neighbourhood level emissions; city emissions at fine scale; emissions at the building level; and the representation of emission data collected through direct sensing (Fig. 3). 3.2.1. District/Neighbourhood mapping A form of emission mapping concerns those which represent the relative differences across neighbourhoods. Tan et al. (2021) analysis of emissions from passenger transport in Shenzhen is an example of this. Here, the authors analysed five communities within the city, which were split into a total of 34 ‘traffic analysis zones’ (TAZ). Bottom-up data was collected through using local authorities and research institutes and accessing statistical reports, literature and other online resources (ibid). This data collection identified the requisite variables for the study, which were divided into four categories: vehicular energy, land use, socioeconomic factors, and transport accessibility. After ascertaining the travel demand for each TAZ, the authors applied emission factors to the vehicle activity data to estimate the car and bus emissions across the study area. The study provides maps relating to the distribution of vehicular emissions, per capita emissions and emission intensity across the 34 TAZs, revealing the impact of cars and the significant role of ‘spatial inequalities’ of travel demand, transport emissions and land use layouts on passenger transport emissions. This form of mapping is also demonstrated in the work of Marchi et al. in Grosseto and, at a less granular scale, by Han and Ge’s investigation into the role of land-use and urban carbon emissions in Suzhou (Han & Ge, 2023). 3.2.2. Fine-Scale mapping Auckland is working to half its carbon emissions by 2030 and to reach net-zero emissions by 2050 (Keller et al., 2023). Keller et al. (2023) argue that detailed, sector-specific emission monitoring is required. In 2016, the city produced a city-wide emissions inventory, serving as the basis for the development of Mahuika-Auckland, a spatially and temporally resolved fossil fuel CO2 emissions data product for the city. The genus of this tool came from the inadequacy of pre-existing tools to effectively monitor urban carbon emissions at a scale which accurately represents geographic complexity. Using the city’s emission inventory, the scope 1 emissions of various sectors were mapped at a scale of 500 ×500 m across the political boundary of Auckland: operating at a finer scale than those at the district/neighbourhood level. One particular benefit is the ability for this form of mapping to reveal different conceptualisations of urban carbon emissions, moving beyond comparison of districts or census data zones, providing a visual means of categorising areas of a city along more nuanced lines. Other papers which develop a finer resolution of visualisation are Cai et al.’s (2020) study of Hong Kong and Zhou and Gurney’s (2010) analysis of Indianapolis. Both adopted a combination of top-down and bottom-up methods, with the former modelling building and transportation emissions through accessing open data (Cai et al., 2020) and the latter downscaling US national data and combining it with bottom-up datasets and conducting building energy simulations (Zhou & Gurney, 2010). Another example of fine scale carbon emission mapping, but solely from the perspective of those produced through mobility and traffic, is Boitor et al.’s (2019) use of GPS tracking and vehicle speed assessment to estimate vehicle emission hotspots in the Romanian city of Cluj-Napoca, observing that the spaces which feature both congested streets and high-speed roads are those which produce the most carbon emissions. 3.2.3. Representing building level emissions Another scale is at the level of buildings. This form of mapping is often conducted through creating ‘building architypes’, a bottom-up approach to accounting for carbon emissions based on the geometry of Fig. 3. A visual representation of the four different forms of carbon emission visualisation, produced by the authors using Google Earth as a base image: District/ neighbourhood level emissions (Green); City emissions at ‘fine scale’ (Blue); Emissions at the building level (Yellow); Emission data collected through direct sensing (Wifi symbol). W. Brown and K. MacAskill Sustainable Cities and Society 131 (2025) 106677 6 the building stock and building characteristics (building area, height, type) by combining emissions from the manufacture of construction materials, building stock and the construction industry (Zhang et al., 2015). Archetypes are produced via Geographic Information System (GIS) maps (Kaandorp et al., 2022), LIDAR (Kellett et al., 2013), machine learning (Wu et al., 2019) and satellite imagery programmes such as Google Earth (Zhang et al., 2022). Schandl et al. (2020) used building archetypes to assess the carbon impacts of land-use change over time in the Australian suburb of Braddon (Canberra), estimating the embodied carbon emissions produced over a 60-year period. Another example is Zhang et al.’s (2022) case study of Xi’an. Applying a hypothesis that each building is defined by its land use to estimate the city’s building carbon footprint, the authors utilised GIS mapping software to locate buildings within different land-use data zones. Eight building types, with a combined total of 17 archetypes, were identified. These were utilised to identify 15 emission hot spots across the city. Whilst this section commenced with an overview of district or neighbourhood level mapping, Zhang et al. identified a trend for the building emissions of Xi’an to cross such boundaries, observing the influence of building type over that of functional and administrative zones. Whilst this does not invalidate district/neighbourhood level mapping as an approach, for local decision making is often conducted at the district level as is the allocation of city budgets, it does represent the benefits of different approaches to visualising carbon emissions within cities. It is of note that whilst building archetypes are often utilised to calculate building emissions, not all studies which map building emissions use archetypes as an approach. An example of this is the use of preexisting building energy certificate data to calculate building emissions in the Valencian suburb of Quart de Probert (Lorenzo-S´ aez et al., 2020). 3.2.4. Mapping directly monitored emissions Despite the difficulties of directly measuring carbon emissions in the urban realm, four papers included in this review applied sensor-based approaches to directly measure carbon emissions– three of which mapped their findings (the fourth case via Hsu et al. (2010) is covered in Section 3.4). Doukalianou et al. (2020) assessed the potential for carbon mitigation in the peri‑urban hillside forests of Xanthi, Greece, by utilising the static closed chamber technique, assessing gas exchanges between the soil below and above the chamber. This was conducted to assess the impacts of novel forest thinning activities on carbon sequestration. Another paper which directly assessed urban emissions was Pugliese et al. (2018) development of the ‘Southern Ontario CO2 Emissions’ inventory through monitoring the carbon emissions of Toronto from four local sites. Here, the authors used optical techniques for detecting gases to monitor carbon emissions and create a new emission inventory for Canada’s largest city. Lee et al. (2017) developed a mobile sensor network in Vancouver, by modifying five cars and a single bicycle to assess the carbon flux across a range of different neighbourhoods within the city. Each of these papers conducted different forms of assessment: one remote (Pugliese et al., 2018), one mobile (Lee et al., 2017) and one in the ground (Doukalianou et al., 2020). Whilst each paper operates at different scales, the two Canadian city studies produced similar maps based around identifying emission hot spots. Whilst these Canadian studies produced fine scale maps, the analysis of Xanthi’s peri‑urban forest followed an approach akin to district/neighbourhood mapping, by identifying nine plots where different forms of tree thinning would take place and utilising GIS to map their impacts (Doukalianou et al., 2020). Whilst the forms of mapping described in this section possess differing attributes and limitations, they each provide a means of visually demonstrating the intra-city reality of urban carbon emissions. Owing to the adoption of systemic perspectives being supported by visual artefacts and the visualisation of complex phenomena, the visualisation and mapping of a city’s carbon emissions serves an important function in the pursuit of advanced carbon accounting. 3.3. Trajectory 3: The development of more granular, neighbourhood/ district specific emission reduction policies The drive to reduce urban carbon emissions is promoted and informed by the actions of the local or city authority and is often manifested through the development of ‘roadmaps’ for the entire city. Whilst this process is important, they do not necessarily relate to the complex differences present within a city’s boundaries. Policies which reflect such intra-city differences are better positioned to succeed (Chandler, 2014), with several papers utilising local carbon emission data to create granular, local-scale policy recommendations. This marks the first trajectory which advances practice towards the need for more advanced forms of data collection. Many cities reliant upon downscaled national level data may face challenges in finding appropriate methods to downscale further to the neighbourhood level. Rivera-Marín et al. (2023) work in Valencia attempted such downscaling, where the authors utilised a downscaling approach to account for the decarbonisation potential of the scope 1, 2 and 3 emissions of the city’s La Carrasca neighbourhood—focusing upon the built environment, transportation, consumption of goods, waste management and green areas. This work produced an emission inventory for the neighbourhood, enabling the authors to propose policies to reduce the emissions contained within. A similar study in geographical scope, but different in methodology, was conducted by Dorr et al. (2022) in the Parisian suburb of Montreuil, where building retrofit, transportation and food related strategies were assessed—which will be covered in more detail in Section 3.4. In another Valencian neighbourhood, Lorenzo-S´ aez et al. (2020) developed a methodology to map primary energy consumption and carbon emissions in buildings, with the findings utilised to support the implementation of the city authority’s building retrofit policy. In the Australian city of Adelaide, Huang et al. (2017) used LCA to model the emissions of the built environment (both embodied and activity) and transportation in two precincts. The authors observed estimated reductions in emissions if residents increased public transportation use and increased installation of photovoltaics, with a significant difference between the two locations—emphasising the importance of incorporating local features in carbon neutrality policy setting. In Shenzhen, Tan et al. (2021) applied a development model to assess the carbon emissions from passenger transport, identifying that a lack of mixed land use contributes to more intensive private vehicle use. Another paper which proposes neighbourhood specific approaches to reducing emissions is Kellet et al.’s (2013) bottom-up modelling of a residential neighbourhood in Vancouver, produced by focusing on buildings, transport, people (food, waste and the carbon emissions of the human body itself) and plants. The impacts of three scenarios were modelled in comparison to the baseline scenario, with optimising existing approaches, transit-oriented development and maximising low carbon approaches, with the latter producing 31 % of the baseline emissions. Below are three case studies which apply neighbourhood and district level policies through different approaches. 3.3.1. Chongqing Cheng et al. (2022) developed a carbon accounting methodology and an efficiency assessment model designed to be applicable at the neighbourhood level to measure the carbon emission characteristics and reduction potential of each neighbourhood. The authors identified 19 emission sources and sinks across 15 communities. Bottom-up activity data was collected from a variety of sources, through different methods, from statistical yearbooks and data from neighbourhood committees, alongside surveys and semi-structured interviews with residents, merchants and local institutions. This data was converted into emission data by using emission factors and used to assess each community’s carbon footprint and efficiency. The authors categorised communities into four W. Brown and K. MacAskill Sustainable Cities and Society 131 (2025) 106677 7 ‘quadrants’ relating to whether they had high or low emissions and emission efficiency. This resulted in tailored policy recommendations, from promoting highly energy-efficient lifestyles for residents, rehabilitating the building envelope and optimising the transportation network for those with high emissions and emission intensity, to promoting resident awareness of environmental conservation, improving land use efficiency and optimizing community resource allocation for those with low emissions and efficiency. 3.3.2. Grosseto In 2018, the Italian municipality of Grosseto developed a carbon emissions inventory which was integrated with Geographic Information Systems based maps, to visualize the spatial distribution of the ‘greenhouse gas balance’ (Marchi et al., 2023). This inventory was created using the IPCC Guidelines for National Greenhouse Gas inventories, mapping prevalence of emissions across the municipality using primary data covering 46 emission sources provided by local authorities, companies, and sector operators. According to the authors these enable the connection of “human activities, GHG [greenhouse gas] emissions and landscape, providing tools to orient possible decarbonization measures for cities and rural areas” (ibid, p9). It was through these maps that four GHG ‘action zones’ were identified: the City of Grosseto, agricultural areas, coastal areas and agro-forestry surfaces. Each zone was subsequently paired with policy suggestions and the departments and stakeholders able to assist in their delivery. 3.3.3. Salt Lake City Patarasuk et al. (2016) quantified Salt Lake County’s fossil fuel carbon emissions across eight sectors to ascertain the main drivers of the city’s emissions. To this end, the authors applied a bespoke bottom-up dataset and conducted regression analysis concerning population statistics, resident affluence and building age; mapping the results across the Salt Lake City region. Whilst the authors observed that population levels have the greatest proportional influence on carbon emissions, they also argued that resident wealth is an important factor. The policy implications of this finding suggest “that emissions reductions may find greatest efficacy among the high-income census block groups” (p1032). This illustrates another form of granular policy development. Rather than targeting different neighbourhoods or districts within a city, an authority could tailor their emission mitigation policies to different socio-economic groups. Given the proportional significance of consumption emissions within the urban realm (see: Goodwin et al., 2023), and the well observed link between wealth and increased emissions, tailoring policies towards wealthier residents may be a fruitful policy approach to reduce city emissions, as well as supporting more just decarbonisation practices (Fuller, 2017). 3.4. Trajectory 4: Assessing project emissions A further step in the enhancement of a city’s carbon accounting practice lies in assessing and quantifying the emission impacts of actions envisaged to reduce a city’s carbon footprint. It is at this stage that a city expands from monitoring its carbon emissions, towards understanding the impact of their actions to reduce them. The gap between the monitoring of citywide emissions and those from reduction approaches represents a significant hurdle for cities to overcome. However, there are examples of cities working towards the closing of this. For example, some Nordic cities, such as Oslo (City of Oslo, 2024) are applying emission factors to city council procurement data to calculate their organisational scope 3 emissions, an approach which can lay the foundations for calculating bottom-up project level emission data. Within the review, examples of accounting for a project’s carbon emissions include Liu et al.’s (2023) assessment of the impact of wetland renovation for rural wastewater treatment through conducting a LCA in the Chinese city region of Xiamen. By using the data to estimate the long-term emission impacts of the project, the authors estimate the carbon impacts of the project and the corresponding carbon recovery period. Another nature-based project is covered in G´ omez-Villarino et al.’s (2021) assessment of the carbon emission mitigation potential of an agricultural and forestry project in Madrid, identifying the net sequestration benefits of the project outweighing the emissions produced through its implementation. On a different scale, Li et al. (2020) assessed the carbon impacts of indoor farming systems in Singapore, by modelling different energy sources and crop choices, identifying the importance of solar energy and recycling materials in reducing carbon impacts. Within the built environment sector, Sigurðard´ ottir et al. (2023) analysed the embodied emissions of the development of a new neighbourhood in Iceland’s capital city, Reykjavik, by utilising planning documentation to conduct an LCA. Here the authors estimated the emission impacts of using timber instead of reinforced concrete, identifying potential emission reductions of 43 % within the neighbourhood’s construction. Lin et al. (2017) explore building energy consumption by modelling the impacts of solar panel installation, using building archetypes for analysis in Tainan. Another approach to assessing project emissions is to validate new approaches to reducing emissions. One such example is (Kumdokrub et al., 2023) tracking of the energy and material flows in the energy, food waste, and construction materials sectors of Cornell University to assess the metabolism and the feasibility of a circular economy approach to reducing emissions, highlighting the importance of renewable energy generation and composting of food waste as key elements. Beyond assessing the emissions of a project, Hsu et al. (2010) took the step to validate the emission data itself, by conducting a verification study of a carbon and methane inventory which covered the Los Angeles area. The authors used a suite of sensors at the Mt Wilson observatory to produce a top-down inventory and compared it with a prior, bottom-up inventory produced by the California Air Resources Board, observing that the top-down inventory reported a third-greater level of emissions in comparison with the bottom-up approach. Below are three examples which highlight the assessing of emissions bound in projects at different scales—building retrofit, city-wide nature-based solutions, and the adoption of new practices to reduce emissions. 3.4.1. Amsterdam Kaandorp et al.’s (2022) analysis of building insulation and electricity decarbonisation across three sites in Amsterdam exemplifies the assessment of novel approaches across different contexts within a city. The authors investigated the configuration of urban heat systems to identify the lowest cumulative carbon emissions over time. To ascertain the emission impact of the assessed heat systems, a bottom-up heat demand model was developed, comprising of GIS informed, building archetype derived models of heating demand, heating system emission factors and insulation and decarbonisation trajectories. Based on the study’s findings the authors argue that the viabilities of technologies for heat supply depends on the ambitiousness of the policy to decarbonise electricity as well as the rate of building insulation. This finding has a systemic quality, requiring consideration of the available technology is required as well as how it operates within its context. 3.4.2. Los Angeles McPherson and Kendall’s (2014) LCA research into Los Angeles’ ‘Million Trees’ programme assesses fuel use, material inputs, and biogenic carbon flows for each life stage of the programme over 40 years. This analysis covers planting, maintenance, growth, removal and disposal of three types of tree planting, street, park and yard trees. This inventory accounted for the impact of shading from trees on energy demand, revealing that whilst the project overall would be a net carbon emission sink, this was not true for all types of tree planting, with park trees projected to be net carbon emission sources; owing to the emissions associated with their planting, maintenance and removal not being balanced by their sequestration. Nor do they provide secondary benefits, W. Brown and K. MacAskill Sustainable Cities and Society 131 (2025) 106677 8 like in building shading, which could reduce building heating, venting and air conditioning emissions. This finding further illustrates the virtue of accounting for project emissions. The conventional wisdom is the more trees planted within a city, the more carbon will be sequestered, yet, if the processes behind the actual implementation and management of the trees is accounted for, then they could be a net source of carbon emissions. 3.4.3. Kawasaki Dong et al. (2014) analysed carbon emission reduction through industrial and urban symbiosis in the Japanese industrial city of Kawasaki. Urban symbiosis is the use of by-products (waste) from cities as alternative raw materials or energy sources in industrial operations. This form of ‘circular economy’ is of importance within reducing urban carbon emissions, given the significant embodied emissions contained within materials such as steel and concrete. The authors evaluated the carbon footprint from a lifecycle perspective along three lifecycle stages: upstream, onsite and downstream. If emissions occurred within the administrative boundary, they were classified as onsite direct carbon footprint, whilst emissions which occurred outside the administrative boundary were classified as upstream or downstream depending on whether they were produced before or after those onsite. The authors categorised the carbon footprint into six ‘parts’ and collected activity data from technical documents, published literature and an onsite survey, then applied emission factors. Regarding the reduction of emissions from urban symbiosis, its largest impacts concerned the city’s iron and steel industry, where emission reductions were mainly found in material carbon footprint reduction, which contributed to the positive impact of urban symbiosis on reducing overall emissions. In opposition to consuming new materials, the recycling of materials central to urban symbiosis reduced the eco-town’s carbon footprint. 3.5. Trajectory 5: Modelling future policy impacts Beyond assessing project emissions, the development of policy level emission data is a complex, but important task to conduct to understand the possible emission impacts of policies going forward. This is owing to the influence city authorities have on reducing emissions currently being unaccounted for in terms of their carbon impact. Despite city councils often being constrained by national level policies—for example the inability to introduce carbon taxes—the passing of local, city level carbon neutrality policies, especially in relation to reducing consumption and scope 3 emissions, is a key source of leverage for city authorities to reduce their city’s carbon emissions. However, the ability for a city authority to estimate the carbon impacts of a policy or range of policies is a particularly advanced and complex form of carbon accounting. This is in part due to the complexities of setting the boundary of accounting, thereby raising questions of what is determined to be impacted by a policy and what is not, as well as the requirement to do speculative accounting grounded in assumptions which is predominately beyond the capacity for city authorities. However, within academia, several papers have estimated the emission impact of different policies upon a city’s carbon footprint. Adopting a broad perspective on urban emission reduction approaches, Sugar and Kennedy (2013) modelled a range of different interventions within building retrofit, green infrastructure, alternative energy supply and transportation. The estimated impacts of which were assessed in line with current policy and more aggressive policies in 2031 for the city of Toronto (Sugar & Kennedy, 2013). In Bristol, Millward-Hopkins et al. (2017) calculated the production and consumption emissions of the British city, as well as estimating its future emission trajectory up to 2035. The authors proposed mitigation scenarios to reduce production emissions but find that these measures will have a limited impact on the consumption emissions of the city, which are estimated to be three times larger than production emissions. Another paper which targeted consumption emissions was an assessment of Canberra’s emissions, identifying that consumption emissions of the Australian capital were 83 % of its total carbon footprint (Goodwin et al., 2023). The authors estimated that by implementing policies designed to a 1.5 ◦C warming scenario, the city’s emissions could be reduced by 85 % by 2045. Another paper which sought to estimate the potential impacts of policies was Yang et al.’s (2018) assessment of the Chinese city of Ningbo in relation to the national government’s launching of a low-carbon city pilot program. This assessment was conducted through the development of a long-range energy alternatives planning system model which simulated the emission impacts of six energy sectors, with the authors observing the need to both emphasise increasing energy efficiency and reducing consumption rates. The review also identified papers which propose policies to reduce carbon emissions, but do not estimate their impacts. Examples of this include: Lombardi et al. (2018) developing an action plan for the Italian municipality of Foggia, based upon the calculation of the city’s territorial (scopes 1 and 2) carbon footprint; Lwasa (2017) accounting of Kampala’s territorial carbon footprint identified interventions from the building, transportation and nature based solutions sectors; Zhang et al.’s (2015) analysis of the carbon emissions of the different sectors of Beijing’s economy; and Gu et al. (2019) system dynamics modelling of Shanghai’s carbon footprint, identifying that improving energy intensity would be the most effective means of reducing emissions. 3.5.1. Montreuil Owing to the detailed carbon data collected and focus on consumption emissions, Dorr et al. (2022) assessment of the Parisian suburb of Montreuil’s carbon emissions features an insightful estimation of policy induced emission reductions. Through applying environmental modelling across different sectors— combining food, mobility and buildings—the authors evaluated the impact of various climate mitigation interventions, involving data collection from a range of sources. For food, national level data was downscaled for use at a city level, using socio-economic data to facilitate this process. Mobility and residential buildings were modelled by adopting the widely used four-step transportation modelling process (to generate resident trips and associated emissions) and by applying a building energy simulation model (which created 14 building archetypes representing various heating loads and environmental impacts across building lifecycles). These models produced an emissions baseline against which to assess eight policies across the three sectors: reducing food waste and replacing red meat in local diets; increasing electric and hybrid vehicle use as well as implementing a speed limit of 15 km/h and not allowing car journeys below 3 km; and the replacement of windows, renewal of gas boilers and the insulation of building facades. The study predicted a 23 % reduction of Montreuil’s carbon emissions as a combined impact of these policies. The authors found that the building sector emissions reduction had the most impact from local intervention. However, the food sector emissions are mostly created by agricultural sites outside the city’s boundaries and thus are not under the direct influence of elected officials. 3.5.2. Nanjing Whilst the above papers estimated the impacts of hypothetical emission reductions, another form of estimating policy impacts is to model currently enacted policies. Considered to be a vital economic centre in East China, Nanjing is home to heavy industry which accounts for >90 % of city energy consumption—a critical issue given the its reliance on energy derived from coal and oil (Chen et al., 2020). To this end, the Chinese government has proposed emission reduction policies designed to control industrial energy consumption, promote new power system construction, waste classification management, the development of green transportation and industry restructuring. Within their case study, Chen et al. (2020) coupled these policy approaches with three contextual factors: economic development, population growth and the urbanisation rate. They then set four scenarios covering the W. Brown and K. MacAskill Sustainable Cities and Society 131 (2025) 106677 9