A Review of the Impact of Urban Form on Building Carbon Emissions

Abstract

With the intensification of urbanization, resulting in the growing building stock, building operations have become the main contributors to greenhouse gas emissions. However, the relationship between urban form and carbon emissions remains unclear, which limits the sustainable development of cities. This study reviews the definition of carbon sources, data characteristics, and evaluation methods of carbon emissions. In addition, the impact of urban form on building carbon emissions at the macro, meso, and micro scales is reviewed, and low-carbon design strategies for urban form are discussed. Finally, the existing problems in this field are pointed out, and future research directions are proposed. Our review found that small and medium-sized compact cities tend to have less carbon emissions, while large cities and megacities with compact urban forms have more carbon emissions. The carbon reduction design of urban form at the meso scale is often achieved by improving the microclimate. Developing a research framework for the impact mechanism of building carbon emissions in a coordinated manner with multi-scale urban forms can effectively promote the development of low-carbon sustainable cities. This review can assist urban planners and energy policymakers in selecting appropriate methods to formulate and implement low-carbon city analysis and planning projects based on limited available resources.

1. Introduction

The rapid development of urbanization, one of the most prominent global trends in the 21st century, has been accompanied by a steady rise in carbon emissions, stemming from energy needs. Climate change, driven by global warming due to intensified carbon emissions, will cause a series of serious global environmental problems such as rising sea levels, frequent extreme climate disasters, and reduced biodiversity, which are increasingly attracting the attention of the international community [1]. In response to this reality, the global community has proactively and systematically implemented measures to address climate change. The Kyoto Protocol clearly stipulates that during the compliance period from 1990 to 2012, industrialized countries must achieve a reduction target of no less than 5% in their total greenhouse gas emissions [2]. The Paris Agreement initiated global actions to address climate change post-2020 [3]. The European Climate Law officially came into effect in 2021, incorporating the 2050 carbon neutrality target into the EU legal system, and explicitly requiring member states to formulate a phased carbon reduction path and establish a climate adaptation mechanism.

Urban areas serve as hubs of human socioeconomic activities, characterized by high population densities and intensive resource consumption. Although urban areas occupy only 2% of global land use, they host more than half of the world’s population, with the rate expected to reach 68% by 2050 [3,4,5]. Additionally, about 75% of global energy consumption and nearly 70% of carbon emissions come from cities [6]. Furthermore, it is impossible to overlook that the construction industry accounts for approximately 40% of total energy consumption and 36% of global greenhouse gas (GHG) emissions, which has identified it as a key sector for carbon emission reduction. Compared to the material production and construction stages, the operational phase is the largest contributor to carbon emissions, such as heating, cooling, ventilation, and lighting [7,8]. Urban spatial form exerts a locking effect on urban operations [9]. Once established, it is difficult to alter, thereby exerting long-term and profound impacts on urban economic activities and carbon emissions [7]. Consequently, urban spatial form plays a critical role in enhancing urban energy efficiency and reducing carbon emissions, gradually emerging as a research hotspot in urban carbon emission studies [10].

The concept of urban form can be defined in two ways: narrow and broad. The narrow definition of urban form refers to the physical spatial form constituted by urban entities [11]. In contrast, the broad definition of urban form denotes the outcome of the coupled interaction between human activities and natural factors within a specific geographical space and socioeconomic context, representing the overall image perceived and understood by people through various means [12]. The term “urban form” in this study adheres to the narrow definition.

Different scholars have proposed a variety of classic theoretical frameworks related to urban form. The urban form research theory was proposed by March and Martin in the 1950s. Their theory believes that cities are composed of basic spatial elements such as different open and enclosed spaces and various traffic corridors, and analyzes these basic elements and their relationships from different scale levels to qualitatively and quantitatively analyze them [9]. Kevin Lynch’s form believes that people obtain an overall image of the city through five elements: Paths, Edge, District, Node, and Landmark. He believes that a good urban form should have elements such as diversity, permeability, accessibility, recognizability, flexibility, and vitality [13]. Bill Hillier and Julienne Hanson proposed the space syntax theory based on topological ideas. This theory aims to explore the relationship between spatial organization and human society by quantitatively describing the spatial structure of various human settlements such as buildings, villages, and cities. The space syntax theory focuses on the configuration of space, which is the relationship between spaces that constitute the system. Space is abstracted into quantitative symbols and then converted into relationship diagrams to achieve a quantitative description of space [14]. Michael Batty [15] is the pioneer of complexity theory. In his book Cities and Complexity, he proposed that cities present multifaceted complexity, which is reflected in the fact that cities are composed of networks of connections around the world, including trade, social interactions, and even knowledge networks that have become global through the Internet. With the support of complexity theory, he introduced a computational model based on cellular automata to simulate the relationship between cities and the external complex environment [15]. In addition, different scholars may have different views and opinions on the same research theory. For example, supporters of the compact city theory believe that high-density compact cities are crucial to reducing greenhouse gas emissions. However, another group of scholars holds a critical view of the compact city theory, in which they believe that it exaggerates the role of compact cities in reducing greenhouse gas emissions and its implementation cost is extremely expensive [16].

Urban areas are dynamic hierarchical networks of nested structures characterized by diverse spatial scales [17]. Based on spatial scales, Caniggia categorized urban forms from the macro to micro scales as regions, cities, blocks, streets, and buildings [18]. Based on the spatial scale classification commonly used in current research on urban form and carbon emissions, this study categorizes urban form into three scales: the macro, meso, and micro scales. The macro scale focuses on the urban level, where the key factors influencing carbon reduction primarily include urban land use and urban spatial layout. The meso scale targets the street and block levels, with key carbon reduction influencing factors such as the street layout and block structure. The micro scale zeroes in on the characteristics of individual buildings, with its carbon reduction influencing factors primarily focusing on the building shape coefficient, height, shape, and orientation, among others. It is worth noting that existing related studies have slight differences in the classification of their spatial scales, which need to be determined according to the specific research content and objectives.

The impact mechanism of urban spatial form on carbon emissions is intricate, with varying degrees of influence and distinct mechanisms among different urban spatial form elements [19]. This research categorizes the impact mechanisms of urban spatial form on carbon emissions into two types: direct and indirect influence. A direct impact means that urban morphological elements directly affect energy consumption and carbon emissions, and its response is that the characterization index itself is a factor that can directly affect energy consumption and carbon emissions. For example, urban spatial forms with high compactness are characterized by dense populations, clustered production activities, efficient infrastructure utilization, and short commuting distances, which directly lead to a reduction in carbon emissions [20]. Indirect impact means that no direct relationship exists between urban morphological elements and carbon emissions, but an effect is still felt by carbon emissions as corresponding intermediary elements are affected. For example, block network density is positively correlated with surface temperature [21], and surface temperature, as a mediating factor, can further influence energy consumption and carbon emissions. Another example is that a building’s layout influences the surrounding air velocity and solar radiation [22], thereby affecting internal cooling and heating energy consumption and its associated carbon emissions in buildings [23]. It is worth noting that the same urban spatial form element may impose both direct and indirect influences on carbon emissions, with these effects often presenting significant trade-offs in their cumulative impact. For instance, increasing urban blue-green spaces, such as water bodies, green spaces, and vertical greening, can not only directly reduce urban carbon emissions but also indirectly mitigate urban heat island effects by regulating the urban microclimate, thereby reducing carbon emissions [24,25,26,27,28,29]. However, the inappropriate configuration of blue-green spaces may also lead to increased energy consumption and carbon emissions due to higher winter heating demands, a problem that is more pronounced in hot-summer–cold-winter regions.

This study aims to comprehensively review the relationship between urban form and operational building carbon emissions, with the goal of informing science-based low-carbon urban planning theories and methods. This paper is structured into six sections. Section 1 provides an overview of the characteristics of urban form and operational building carbon emissions, as well as their impact relationships. Section 2 employs bibliometric analysis techniques to collect and conduct qualitative and quantitative analyses of the literature on urban form and building carbon emissions over the past decade, presenting the research status and development trends in this field. Section 3 introduces the carbon source definition, data characteristics, and accounting methods. Section 4 discusses the research methods used in the field of urban form and building operation carbon emissions; clarifies urban form indicators and their impact mechanisms on building operation carbon emissions from the macro, meso, and micro levels; and provides corresponding design inspirations. Section 5 summarizes the theoretical framework of urban form and carbon emissions research, urban form factors, and influencing mechanisms and proposes future prospects. Section 6 summarizes the content of this study and presents the conclusions drawn.

2. Bibliometric Methods and Results

2.1. Materials and Methods

Figure 1 illustrates the research framework for this study. Using bibliometric methods, this research undertook quantitative and qualitative analyses of the relevant literature on the impact of urban form on building carbon emissions. Two specific steps were taken: data collection and bibliometric analysis.

2.1.1. Data Collection

This study utilized the Web of Science (WOS) database as its literature data source. The literature retrieved in this study covers the period from 2015 to 2024. After pre-analysis and comparison, the search keywords were determined as follows: “(TS = (urban form) OR TS = (urban morphology) OR TS = (urban spatial structure)) AND (TS = (carbon emission) OR TS = (carbon emission during building operation) OR TS = (operational carbon emission) OR TS = (carbon emission of building) OR TS = (carbon emission during use stage) OR TS = (direct carbon emissions) OR TS = (building energy consumption)).” Additionally, the core collection database was selected, and the literature type was set as articles and reviews. The retrieved 2007 English articles were screened according to the PRISMA flow chart shown in Figure 2. In the identification stage, 100 articles were excluded, including duplicate articles and articles detected as unqualified by automatic detection tools. In the screening stage, 366 articles that did not match the article topic and 488 articles that did not match the article research object were excluded by reading the title and abstract. Then, three reviewers conducted a comprehensive evaluation of the conceptualization, research design, and data quality of the 598 articles that could be obtained based on the quality assessment criteria shown in Figure 3. The scoring criteria are as follows: 0 points: not reported/not considered; 1 point: partially reported/partially considered; 2 points: fully reported/fully considered. Finally, according to the scoring results, the articles were divided into three categories according to the total score. Among them, the articles with a total score of ≥9 points were high-quality articles, totaling 98 articles; the articles with a total score of 6–8 points were medium-quality articles, totaling 151 articles; and the articles with a total score of <6 points were low-quality articles, totaling 349 articles. Low-quality literature will be excluded, and finally, 249 pieces of literature were obtained, see Supplementary Materials.

2.1.2. Bibliometric Approach

Bibliometric analysis is a technical means of conducting research in a quantitative manner. It has advantages such as objectivity, comprehensiveness, efficiency, and accuracy. The commonly analyzed data in bibliometric analysis include keywords, citations, authors, institutions, and their affiliations. Common methods include collaboration network analysis, keyword co-occurrence analysis, and cluster analysis. To explore the development trends of publications, author collaboration networks and research hotspots, researchers often employ bibliometrics as a research method [30].

This study uses bibliometric analysis methods to deeply explore the knowledge system of the impact of urban form on building carbon emissions.

In bibliometric analysis, commonly used software tools include VOSviewer, CiteSpace, and BibExcel, among which CiteSpace is the most widely used. CiteSpace is suitable for literature data analysis in multiple disciplines. Its prominent advantage lies in its ability to efficiently handle massive amounts of literature and present literature information in a visual form. This paper uses CiteSpace (6.3.1) to evaluate and visualize research results.

2.2. Bibliometric Results

Figure 4 indicates a notable upward trend in the annual publication volume of articles examining the impact of urban form on building carbon emissions from 2015 to 2024. Research in this field can be roughly divided into two stages. The period from 2015 to 2021 was one of steady growth, with the number of publications gradually increasing from 2 to 18. The period from 2022 to 2024 was one of rapid growth, with the number of published articles rising to 45–68.

By analyzing the journal distribution characteristics of the literature, the influence of the included literature can be reflected upon. The journals that have published research on the impact of urban form on building carbon emissions are counted, as shown in Figure 5a. The top two journals are Land (25) and Sustainability (25). Followed by Journal of Cleaner Production (19), Ecological indicators (15), Sustainable cities and society (13) and Journal of environmental management (11). The remaining journals have published fewer than 10 articles. One of the important indicators for measuring the significance of a journal article is its journal impact factor. Among them, the journal with the highest impact factor is Sustainable cities and society (12). Next are Journal of cleaner production (10), Energy (9.4), Journal of environmental management (8.4), and The science of the total environment (8). The impact factors of the remaining journals are all less than eight. In addition, according to different research methods, all the literature is divided into three categories: empirical studies, simulation studies and conceptual studies, as shown in Figure 5b. There are 66 empirical studies, 24 simulation studies and 159 conceptual studies.

In recent years, research on the impact of urban form on building carbon emissions has shown obvious regional distribution characteristics. Figure 6 illustrates the research progress on the effects of urban form on building carbon emissions in various countries worldwide. In terms of quantity, countries with the highest number of published articles include China, the United States, the United Kingdom, Japan, and South Korea. In terms of publication time, China, the United States, and the United Kingdom were the earliest to conduct research in this area. From the perspective of research directions, these countries can be mainly divided into three camps: The camp led by China, including Australia, Malaysia, Russia, Thailand, Singapore, and other countries, primarily conducts research on greenhouse gas emissions. The camp led by the United States, including Japan, South Korea, Austria, and other countries, mainly conducts research on sustainable development. The camp led by the UK, including Ethiopia, Germany, India, Italy, and other countries, primarily conducts research on decarbonization. More cooperation occurs within the same camp, but mutual communication also occurs between different camps.

Keyword co-occurrence network analysis is a method of analyzing the frequency and correlation of keywords in the literature, which helps identify hot topics and main research directions within a research field. Figure 7 presents the keyword co-occurrence network developed in this study, which was constructed using the keywords extracted from 254 collected articles. CO₂ and carbon emissions have the highest frequency of occurrence, with 126 and 68 occurrences, respectively. The frequency of “city” is 67, accounting for a relatively large proportion, indicating that research on urban carbon emissions has been given special attention. The frequency of “energy consumption” is 62, indicating a strong correlation between energy consumption and carbon emissions. The frequency of “China” is 52, indicating that China has a high level of concern regarding carbon emissions and numerous related studies. The frequency of “urbanization” and “urban form” is 51, indicating that the process of urbanization and urban form also has an impact on carbon emissions. The frequency of other keywords is relatively low, such as “economic growth”, “land use”, and “climate change”. Still, this suggests that carbon emissions are linked to issues such as economic growth, land use, and climate change.

Figure 8 shows a timeline of keyword co-occurrences to illustrate the evolution trend of keywords. Between 2015 and 2016, keywords such as “city”, “urbanization”, “urban form”, and “energy consumption” emerged, indicating that researchers began to pay attention to the impact of factors such as urbanization, urban form, and energy consumption on building carbon emissions. Between 2017 and 2019, keywords such as “energy use”, “climate change”, “economic growth”, “land use”, and “socioeconomic factors” emerged, indicating that researchers began to pay attention to the impact of energy use, climate change, economic growth, land use, and socioeconomic factors on building carbon emissions. Between 2020 and 2022, keywords such as “population”, “urban agglomeration”, and “social network analysis” emerged, indicating that factors such as population, urban agglomerations, and social network analysis started receiving attention. Between 2023 and 2024, the appearance of keywords such as “spatial heterogeneity”, “urban compactness”, and “country” indicates that, in recent years, research on carbon emissions has tended to be conducted at a macro scale.

3. Critical Research Parts of Building Operational Carbon Emissions

This section introduces the three key components of carbon emissions research during building operations: carbon source definition, data characteristics, and accounting methods. The detailed process framework is illustrated in Figure 9. Considering the entire life cycle, energy consumption and carbon emissions during the operation phase of buildings are significantly higher compared to those from building material production, construction, and building demolition after the scrapping stages [31]. Therefore, in most cases, the building operation stage is used as the primary stage for calculating carbon emissions. Unless otherwise specified, the energy consumption and carbon emissions discussed in this article refer solely to those associated with the building operation phase.

3.1. The Definition of Carbon Sources

The definition of carbon sources is essential for quantifying and evaluating carbon emissions. The sources of operational carbon emissions (OCES) are mainly related to the system boundary under study [32], the greenhouse gases considered, and the operating space.

The system boundary marks the boundary between OCES-related activities and the outside world and determines the number and type of activities [33]. A LCA system boundary mainly includes cradle-to-grave, cradle-to-gate (for building product analysis), or gate-to-gate (for construction process analysis) system boundaries. The cradle-to-grave system boundary is usually adopted, which refers to the building, starting from its pre-use to end-of-life phases [34], including the pre-use, construction, use, and end-of-life phases [35]. Among these phases, the carbon emissions produced due to operations during the building use phase are OCES, and the sources mainly include heating, cooling, ventilation, lighting, cooking, information technology (IT), and equipment [36]. The definition of the system boundary of OCES research mainly differs in the following three aspects: First, the types of operational projects that should be considered in research are not clear and unified, so most studies define the system boundaries subjectively. Second, how far the research should extend upstream and downstream of the operational stage is unclear [37]. Finally, due to different building structures and functions, some studies consider the entire building, while others only consider a subdivision or even a component. The Kyoto Protocol stipulates six greenhouse gases that affect global warming, with the main impacts being from CO₂, CH₄, and N₂O [2]. Among these greenhouse gases, the greenhouse effect produced by CO₂ accounts for the largest proportion, at 55%, and is increasing year by year. Most studies and standards only focus on CO₂, but in some high-precision studies, the contribution of non-CO₂ greenhouse gases is also considered to more comprehensively quantify the carbon footprint during the operational phase [38]. The operating space of OCES is not limited to the internal space of the building but should include all spaces that generate carbon emissions required to support the normal operation of the building. In 2023, about 28% of the carbon emissions generated by global building operations were direct emissions from the internal site of the building and 72% came from indirect emissions from the outside of the building generated by electricity and central heating to ensure the normal operation of the building [39]. Therefore, the OCES operational space includes all spaces where carbon emissions are produced due to the operation of buildings or building components.

3.2. Data Characteristics

Data quality not only determines whether emission reduction targets can be effectively quantified but also directly affects the scientific validity and practicality of the policies and emission reduction measures. OCES data have many characteristics. First, the data types are diverse. Building operation carbon emissions usually include multiple systems and may involve different energy types at the same time, so data collection requires the coordination of multiple departments to ensure the data’s accuracy and completeness [40]. Second, the sources of data are widely distributed. Unlike the construction phase, which is concentrated on the construction site, operational emissions data are distributed across various locations and terminals throughout the city [41]. For large-scale building clusters or building complexes, the statistics of operational carbon emissions are more likely to span multiple administrative regions or multiple energy service providers, making data acquisition more cumbersome. Third, the data changes dynamically. After a building is officially put into use, its energy consumption often fluctuates cyclically and seasonally with the outside temperature, usage time, resident behavior, and energy supply price [42,43]. Therefore, the collection of real-time data and a multi-period weighted analysis can ensure the representativeness of building operation carbon emission accounting. Some studies have proposed that dynamic models and real-time monitoring should be used to grasp the changes in building operation emissions [44]. However, due to the limitation of data acquisition frequency, achieving real-time carbon emission tracking for all buildings remains difficult. Fourth, there is a contradiction between the accuracy of carbon emission data and carbon reduction efficiency. If you want to reduce emissions in the early design process, you are often limited by the lack of actual data or accurate input of usage scenarios. However, when the building is put into operation, although more comprehensive data can be obtained, the opportunity to maximize emission reduction in the design stage is missed. This contradiction of “lack of data support at the stage where intervention is most needed, and difficulty in making fundamental changes at the stage where data is most available” has become a key challenge for building emission reductions. Fifth, data collection is private and sensitive. Since operational data are often closely related to user behavior [45], companies or residents may be unwilling to disclose data for privacy and profit considerations [46]. Therefore, corresponding mechanisms need to be established in terms of data security and privacy protection to ensure that all parties are willing to share data.

3.3. Assessment Methods

Carbon emission measurement is the process of calculating carbon emissions using data from relevant sources such as the economy and energy consumption. Nowadays, common methods for assessing carbon emissions can be divided into direct carbon emission general calculation methods and life cycle-based building carbon emission accounting methods.

3.3.1. General Methodology for Measuring Carbon Emissions

The three most widely used direct carbon emission accounting methods are the emission factor method (EFA), the mass balance method (MBM), and the measured method (AMA) [38].

The EFA was proposed by the Intergovernmental Panel on Climate Change (IPCC) in 1966 in the Guidelines for National Greenhouse Gas Emission Inventories and remains the most widely used technique for carbon emission accounting [47]. This method estimates carbon emissions (E) based on the carbon emission inventory by multiplying the activity data (A) of each emission source by the emission factor (EF):

$$E=A\times EF$$

where E is the greenhouse gas emissions, A is the activity data (a specific amount directly associated with the carbon emissions of a single emission source); and EF is the carbon emission factor, that is, the amount of greenhouse gases emitted per unit usage of the emission source. This method is simple to calculate and highly operational [48] and has a relatively mature emission database [49]. However, emission factors are regional in nature, subject to the level of technology, production level, and process flow. Therefore, many countries have developed a set of correlation coefficient specifications that are suitable for their national conditions [50].

MBM is grounded in the law of conservation of mass, which means the carbon content of the input material equals the carbon content of all output materials. The formula is as follows:

$$E=\left[\sum (M_{in}\times C_{in}\right)-\sum (M_{out}\times C_{out}\left)\right]\times {\displaystyle \frac{44}{12}}\times GWP$$

where M_in is the input material quantity; C_in is the carbon content of the input material; M_out is the output material quantity; C_out is the carbon content of the output material; ${\displaystyle \frac{44}{12}}$ is the conversion factor from carbon to carbon dioxide; and GWP refers to the global warming potential. The MBM offers the benefit of a systematic and comprehensive approach to analyzing carbon emissions, allowing it to capture differences between various equipment types. However, this method demands an in-depth understanding of both the production process and the chemical reactions involved. The workload is large, and the data requirements are also high [51].

AMA refers to the use of professional instruments to measure the gas emission rate, emission, and concentration to calculate carbon emissions [38].

$$E=Q_{air}\times c_{air}\times \alpha$$

where Q_air represents the flow rate of the medium (air), C_air represents the CO₂ concentration within the air, and α is the unit conversion factor. Although actual measurement provides highly accurate and straightforward results, it is both challenging and expensive to obtain the necessary data. Consequently, it is typically employed for limited sample testing, where the representativeness of the sample is critical to ensuring result accuracy.

3.3.2. The Life Cycle Approach for Measuring Carbon Emissions

According to different system boundaries and methodological principles, traditional LCA assessment methods can be divided into the PBM (process-based method), the EIO (economic input–output) method, and the hybrid method (HM) [52]. These traditional methods still have certain limitations in calculation [53]. According to their methods of calculation, these methods can be divided into bottom-up and top-down, which are consistent with PBM and EIO methods, respectively [54].

The PBM divides the life cycle of a building, analyzes its energy consumption and emission data one by one in stages, multiplies them by the corresponding coefficients, and accumulates them to obtain the total emissions [55,56]. The advantages of the PBM are its high calculation accuracy and strong customizability, which are suitable for detailed studies on individual buildings or small-scale projects [57]. However, the PBM has high data requirements for each link, and the analysis process is complex and time-consuming. In addition, the PBM has uncertainty in the division of system boundaries and is easily affected by upstream indirect influences [58]. Therefore, it is best suited for systems with clearly defined boundaries or products with clearly defined manufacturing processes, and it is less effective for evaluating carbon emissions from urban residents based on urban form [59]. The EIO model proposed by Leontief uses an input–output table to assess carbon emissions, which makes it possible to assess the interrelationships and efficiency of carbon emissions within or between specific sectors [60]. This model calculates final carbon emissions at a macro level and connects them to the input–output economic data across sectors. It adopts the “integrated system boundary” to avoid random boundary division [61]. The EIO method not only studies the direct environmental impact of products or services but also considers all indirect consequences involved [62]. Therefore, the EIO model estimates are usually higher than the PBM [33]. While data for the EIO method are relatively easy to obtain, it is a commonly used method in macro-accounting of urban carbon emissions. [63]. However, its accuracy depends on the updating frequency and accuracy of the input–output table and emission factors, and performing subdivided accounting for specific projects or building types is difficult [61].

The HM merges the precision of the PBM with the completeness of the EIO method, which can not only achieve a detailed analysis of key processes and specific buildings but also take into account macro-assessments in a larger regional or socioeconomic framework [64]. However, when merging data from process-based methods and input–output methods, HM can introduce truncation errors, which can lead to omissions or data duplication [65]. Moreover, the HM also faces the problems of complex model architecture and huge data requirements [66].

4. The Relationship Between Urban Form and Building Carbon Emissions

4.1. Research Methods

In terms of spatial statistics and geometrics, existing studies often use methods such as GWR (geographically weighted regression), spatial autocorrelation analysis, and GIS (geographic information system) spatial analysis [67,68] to identify the spatial differences and spillover effects or impact mechanisms of carbon emissions [69]. GWR can introduce geographic location dimensions into regression coefficients to identify the spatial variation characteristics of the impact of a certain urban form indicator on carbon emissions. Additionally, many studies also use cluster analysis methods [70] to classify cities or blocks into types, thereby comparing carbon emission differences at a more detailed urban unit scale. These spatial statistics and clustering methods can more accurately characterize the characteristics and mechanisms of urban form and carbon emissions in different regions and can further provide support for optimizing urban planning layouts. In addition, some studies are integrating machine learning methods such as Lasso regression and random forest [71,72] to process high-dimensional or multicollinear urban form data and to evaluate importance ranking and nonlinear associations. Though still emerging in this research area, such methods can effectively support higher-precision carbon emission prediction and form optimization simulation in the context of big data. Regarding the relationship between urban form variables and carbon emissions, studies generally use multiple linear regression analyses, OLS (ordinary least squares), correlation analyses, STIRPAT (stochastic impacts by regression on population, affluence, and technology) models, and other research methods [73]. Among these methods, the multiple linear regression and OLS methods are often applied to explore the linear impact mechanism of urban form variables on carbon emissions, which can intuitively present the degree and direction of impact [69]. Although the STIRPAT model is popular in comprehensively assessing the factors that affect carbon emissions, it lacks the ability to detail the impact path of individual factors, which limits its application in urban form research [74].

In summary, the current quantitative methods for the study of urban form and building operation carbon emissions have gradually expanded from traditional regression models to GWR, spatial autocorrelation, cluster analyses, and machine learning models. The comprehensive application of multiple methods helps to more accurately analyze the differences in various influencing factors at different regional scales, as well as the potential contribution of morphological optimization to carbon emission reduction.

4.2. The Impact of Urban Form on Building Carbon Emissions and Design Implications

As shown in Figure 10 and

Table 1. The impact direction and mediating mechanism of important indicators of urban form on carbon emissions at different scales.

Spatial Scale	Factors Affecting	Characterization Indicators	Impact Direction	Main Mediating Factors	Main References
Macro-scale	Urban size	Built-up area	Nonlinear	Population density, Traffic behavior, Energy demand	Wang et al. [75], Zhou et al. [76]
	Urban compactness	Compactness index	Bidirectional	Infrastructure utilization, Traffic behavior, Heat island effect	He et al. [77], Huang et al. [78]
	Urban spatial structure	Urban polycentric structure	Negative	Traffic behavior	Wolday [67], Martin et al. [79]
	Urban road system	Road network density	Bidirectional	Traffic behavior, Local microclimate	Sharma and Mathew [80], Su et al. [81]
	Blue-green space structure	Blue-green space coverage	Negative	Carbon sink, Heat island effect, Local microclimate	Hong et al. [82], Shen et al. [83], Koch et al. [84]
Meso-scale	Street canyon ratio	Street height-to-width ratio	Negative	Local microclimate	Strømann-Andersen and Sattrup [85]
	Block functional layout	Land use mix	Negative	Traffic behavior, Space efficiency	Song and Knaap [86], Chen et al. [87]
	Block green space layout	Accessibility to green space, Green area ratio	Negative	Local microclimate, Traffic behavior	Wang et al. [88], Ramyar et al. [89]
	Block development capacity	Floor area ratio, Site coverage	Nonlinear	Heating/cooling needs	Li and Yan [90], Xu et al. [91], Leng et al. [92]
	Block complex layout	Block orientation, Building spacing, Building average height, Building height standard deviations, Building group layout	Bidirectional	Local microclimate	Kamal et al. [93], Bai et al. [94], Shareef et al. [95]
Micro-scale	Building geometric parameters	Building shape coefficient	Positive	Heat transfer area	Muhaisen et al. [96]
	Building geometric parameters	Building orientation	Negative	Solar heat gain, Natural ventilation	Strømann-Andersen and Sattrup [85], Mahmoud [97]
	Building geometric parameters	Window-wall ratio	Bidirectional	Lighting, heat transfer, air conditioning loads	Susorova et al. [98]
	Building energy efficiency design	BIPV, Green roof, Highly reflective material	Negative	Heat island effect, Renewable energy use	Gibbs and O’Neill [99], Taminiau et al. [100], Berardi et al. [101]

Note: Positive, negative, bidirectional and nonlinear, respectively, mean that as the indicator increases, carbon emissions increase, decrease, increase or decrease and can be U-shaped or inverted U-shaped.

, this section will clarify the important urban forms from the macro, meso and micro scales, as well as their impact direction and mediating mechanism on building carbon emissions, and propose design inspirations.

4.2.1. Macro Scale

At the macro scale, the impact of urban form on building carbon emissions is mainly reflected in two major dimensions: urban land use and urban spatial layout, which include multiple specific factors (see Figure 11). These factors have different degrees of impact on building carbon emissions under different urban environments and development levels.

Urban land use can affect urban carbon emissions in many ways by regulating land use pattern and urban functional layout. On the one hand, land use patterns are highly linked to energy performance and are often characterized by regional development intensity [102]. Excessive development intensity often means the concentrated distribution of large-scale building clusters. Although it helps to save building land, it may also lead to the urban heat island effect, thereby increasing cooling carbon emissions [71,103]. On the other hand, the impact of an urban functional layout on carbon emissions is mainly reflected in the job–housing balance ratio and industrial land use layout. Job–housing imbalance will hinder the effective deployment of multifunctional land use and infrastructure [104]. The urban industrial layout structure can also reduce carbon emissions by integrating industrial space and enhancing energy efficiency [105]. Li et al. [106] analyzed 268 cities in China and found that optimized land use structures can reduce energy-related CO₂ emissions by about 12%.

Existing research in the field of carbon emissions focuses on five aspects of urban spatial layout: urban size, urban compactness, urban spatial structure, urban road system, and blue-green space structure [107]. Among the commonly used indicators for assessing the urban size, the built-up area is often used to measure urban size, and the land use expansion and fractal coefficients are often used to measure the scale expansion status of urban areas. Prior research has highlighted that the impact of urbanization on CO₂ emissions is complex [76]. Wang et al. [75] developed a CSI (city size index) to investigate the impact of urban scales on carbon emissions and found different relationship curves between different levels of urban scales and per capita carbon emissions. The city size index and per capita carbon emissions exhibit an inverted U-shaped relationship in small and medium-sized cities and a U-shaped relationship in large and megacities. Common indicators of urban compactness include the shape index and compactness index [108,109]. Compact cities usually have good land continuity, which can improve infrastructure utilization, thereby reducing per capita energy consumption and carbon emissions [110,111,112]. However, this effect varies depending on geographic location, resource availability, and city scale [20]. Similar studies have found that in exploring the potential mechanism between urban compactness index and carbon emissions, in China’s small and medium-sized urban areas, the inflection point value between the quadratic term of urban compactness index and the total number of green patents is 0.7589, and the inflection point value between urban compactness index and the rationality of industrial structure is 0.4441, which is similar to the value in the benchmark regression. This proves that in small and medium-sized cities, carbon emissions can be effectively reduced through potential channels such as technological innovation and industrial upgrading [113]. However, in large and megacities, excessive compactness may lead to issues such as traffic congestion and worsened urban heat environments, thereby driving up emissions, which need to be balanced through planning [20,113,114]. At the same time, infrastructure utilization beyond carrying capacity may also lead to problems such as reduced efficiency and traffic congestion, offsetting the early emission reduction benefits brought by dense forms [78]. An analysis of 260 cities in China also shows that when the urban form changes from dispersed to compact, carbon emissions first decrease and then slightly increase, and there is an optimal density range [104]. In addition, when He et al. [77] studied the relationship between the three-dimensional compactness of buildings and building energy consumption in different types of urban functional areas in hot summer and warm winter regions. They found that carbon emissions of residential and commercial functional areas with different compactness levels were significantly different, but this was not the case for industrial functional areas. Urban spatial structure is often characterized by polycentricity structure and centralization index. For megacities, polycentric spatial structure helps to improve carbon emission efficiency [115]. Studies have shown that the polycentric development model can balance economic vitality and carbon emission reduction benefits in large cities and can reduce carbon emissions per unit GDP compared with the monocentric model [79]. Therefore, introducing a polycentric layout in urban master planning is considered an effective measure to deal with high carbon emissions in large cities. However, in small cities, the monocentric centralized distribution of major facilities has better performance [67]. The urban road system is the basic framework of urban form. A reasonable road hierarchical structure and a reasonable road network density can not only improve urban traffic efficiency [80] but also affect carbon emissions by changing the urban microclimate [81]. The blue-green space structure not only affects the total amount of urban carbon sinks [116], but also significantly reduces the urban heat island effect [117]. Three indicators, namely, blue-green space coverage, blue-green spatial integration, and blue-green land connectivity, are often used to evaluate the balance and integrity of the blue-green space structure. The optimization of the layout of urban green spaces and ventilation corridors is also one of the urban form optimization strategies in reducing energy consumption from building operations. Parks and green spaces not only reduce the surrounding temperature through transpiration but also serve as ventilation corridor nodes to guide wind flow. Studies have shown that establishing connected green corridors in the southern coastal areas of Xiamen that are more affected by the ocean can enhance urban ventilation and heat dissipation, thereby reducing the air conditioning energy demand of the building complex [82,83]. In areas with less marine influence, appropriately breaking up large areas of continuous natural space to enrich the blue-green system can produce greater heat mitigation effects, thereby reducing cooling carbon emissions. Moreover, findings from Ye et al. [118] suggest that compared with green space, water bodies have better carbon reduction capabilities. And when green space and water bodies are combined in a certain way, the carbon reduction effect is more obvious. In planning, it is necessary to balance high-density development and blue-green infrastructure layout to achieve efficient land use and thermal environment regulation at the same time [84].

4.2.2. Meso Scale

At the meso scale, many scholars have focused on the key morphological elements of street and block structure (see Figure 12) and have conducted a large number of quantitative and qualitative studies on urban form and building carbon emissions.

At the street level, the street density and street canyon ratio jointly affect street layout. Street density includes intersection density and road network density. Studies have found that fragmentation of street networks increases carbon emissions, while moderately connected street patterns help reduce carbon emissions [71,119]. Street canyon ratio is often characterized by street height-to-width ratio, which is closely related to street microclimate and thus affects building carbon emissions within the block [120]. Strømann-Andersen and Sattrup [85] found in their simulation that urban canyon morphology can change the annual energy consumption of low-energy buildings in temperate regions by up to 19–30%. Krüger et al. [121] found that increased street depth can reduce cooling requirements in street canyons in summer.

At the block level, block scale division, block functional layout, block green space layout, block development capacity, and block complex layout together constitute the block structure. Among them, block scale division is directly determined by block area and block length. As for the block functional layout of blocks, land use mix and the mixed-use block ratio are commonly used indicators for evaluating the functional layout of blocks. In general, reasonable functional mix can effectively reduce traffic energy consumption [86] and avoid idle buildings due to improper functional zoning, thereby reducing building carbon emissions [87,107]. Block green space layout can affect traffic carbon emissions by changing the accessibility to green spaces and adjust block carbon sinks by adjusting the proportion of green area ratio [88]. In addition, the green space ratio also indirectly affects carbon emissions by regulating the microclimate of the block [89]. In compact high-rise and mid-rise blocks in three cities in the United States, areas without trees and green vegetation have higher daytime emission levels than other areas [122]. Floor area ratio and site coverage are core indicators of block development capacity. However, under different urban climate backgrounds, the effects of floor area ratio and site coverage on carbon emissions are significantly different [123]. In hot summer and cold winter areas, floor area ratio and site coverage together have a nonlinear reinforcing effect on carbon emissions [90,91]. In severely cold areas, a larger floor area ratio and site coverage are accompanied by less conductive and convective heat loss, thus reducing residents’ demand for heating [92]. In addition, differences in local climate zones can also lead to differences in carbon emissions. In Nanjing, where summers are hot and winters are cold, different urban forms lead to differences in local temperatures of more than 2 °C. The cooling energy consumption of buildings in high-density areas in the city center is about 19% higher than that in low-density suburbs, while the heating energy consumption is about 18% lower, reflecting the trade-off effect of dense forms on cooling and heating demands [124]. Block complex layout also plays a decisive role in building carbon emissions by changing building average height, building height standard deviations, block orientation, building spacing, and building group layout. Building average height and building height standard deviations affect air flow within the plot, thereby affecting the block microclimate and building carbon emissions [125]. A study of residential building energy in Seattle, a temperate maritime climate, showed that as the average height of block buildings decreased, the annual carbon emissions of multi-family buildings decreased [126]. Kamal et al. [93] found that building height diversity has a positive effect on reducing building cooling load in tropical climates and can optimize shadow layout in combination with block orientation. Further research shows that in hot and humid areas, adopting a combination of multi-story, high (medium)-story, and multi-story buildings distributed from south to north can reduce air conditioning load by 20–25% [94]. Shareef et al. showed that in the hot desert climate, in the same residential group, a north–south orientation layout can reduce cooling energy consumption by about 6.4% compared with a northwest–southeast orientation [95]. Building spacing affects microclimate through mutual shading of buildings, thereby affecting carbon emissions [127]. Different building group layouts mainly adjust the microclimate by changing the enclosure of blocks, thereby changing carbon emissions. In hot climates, courtyard layouts hinder air flow between buildings, thereby increasing cooling loads [128]. In hot summer and cold winter regions, row-column layouts have been found to notably lower building energy consumption [129]. For the cold Beijing region, the combination of low-rise buildings + towers and low-rise buildings + towers are the best layout patterns for commercial buildings [130].

4.2.3. Micro Scale

As shown in Figure 13, at the micro scale, on the one hand, the contribution rate of building geometric parameters to building energy carbon emissions is about one-third, which is a key factor affecting OCES [131]. For building units, research has primarily focused on the impact of building shape coefficient, building shape and building orientation on OCES [132]. These factors directly determine the heat transfer between the building and the external environment and the indoor light and heat conditions, which in turn significantly affect energy usage [85]. An increase in the shape coefficient has been linked to higher electricity consumption in urban residential buildings and campus dormitories situated in climates with hot summers and cold winters, and the OCES of the building also increases accordingly [133]. Muhaisen [96] proposed setting a set of ratios in each climate condition to achieve high efficiency in summer and winter at each location.

For example, in hot and humid climate conditions, the perimeter to height ratio between 3 and 7 is the most effective. In hot and dry climate conditions, the recommended ratio range is between 4 and 8. Building shape, orientation and window-to-wall ratio affect solar heat gain and ventilation, thereby changing the air conditioning load [134]. In cold climates, residential buildings should have a low aspect ratio and be oriented south to maximize winter sunlight [85]. Simulations for the Middle East climate found that optimizing the aspect ratio of residential buildings to 0.8 and choosing a north–south orientation could reduce annual energy consumption by about 39% [97]. Susorova [98] evaluated the influence of building geometric parameters on the energy performance of commercial office buildings in six climate zones in the United States. The study showed that in tropical climates, the energy savings rate of window geometry parameters can reach 14%. In tropical and temperate climates, the “best energy performance” occurs in shallow rooms with medium windows and deep rooms with large windows. In cold climates, the “best energy performance” occurs in shallow rooms with small windows and deep rooms with medium windows. On the other hand, building energy design is a core way to reduce carbon emissions. Actual cases show that compared with traditional designs, optimized building solutions can reduce annual energy consumption by about 30% while maintaining comfort [135]. With the development of green energy technology, building-integrated renewable energy has become an important way to reduce building carbon emissions [99,136,137]. Integrated photovoltaic (BIPV) technology enables the building surface to become a power generation unit, reducing dependence on fossil energy [138,139,140]. A study in Daejeon, South Korea, found that solar energy can meet more than half of the city’s energy needs through the deployment of rooftop photovoltaic panels throughout the city [100]. In addition, measures such as green roof and highly reflective materials can effectively reduce the heat island effect, thereby reducing cooling needs and carbon emissions in building clusters [101,141].

5. Discussion

5.1. Theoretical Framework for Studying Urban Form and Building Carbon Emissions

To achieve effective low-carbon city planning, we first need to clarify the relationship between urban form and carbon emissions. Existing research in the field of carbon emissions is mostly based on the logical framework of “urban form-intermediary factors–carbon emissions” [19], but problems such as insufficient theoretical depth and disconnection from applications occur. For example, the theoretical framework proposed by Ewing et al. [142] supposes that urban form influences carbon emissions through electricity distribution efficiency, the housing market, and the urban heat island effect (UHI) but ignores key intermediary variables such as traffic behavior and building energy efficiency. At the same time, the optimization method of low-carbon city planning mainly focuses on two aspects. The first involves the management and operational aspects of cities, including energy systems, consumption patterns, industrial growth, operations, and transportation systems, etc.; the second aspect is urban form planning, which emphasizes the physical form and structural features of urban space [59]. In efforts to better understand the impact of different urban forms on carbon emission behavior, recent studies have also introduced refined classification methods such as local climate zones (LCZs) [122], but the theoretical framework is still mainly based on descriptive analysis and lacks an in-depth explanation of the dynamic interaction mechanism. In the future, an interdisciplinary theoretical model will need to be constructed to reveal the nonlinear relationship between urban form and carbon emissions.

5.2. The Impact Mechanism of Urban Form on Building Carbon Emissions

Existing studies have identified the impact of various urban form indicators on carbon emissions, from the initial investigation of a single factor to the comprehensive analysis of the impact of multiple factors. However, the impact intensity and mechanism of different indicators exhibit differences, and the weights set by various studies have not yet reached a consensus [98]. In general, urban form affects carbon emissions through two pathways: direct and indirect effects. Direct effects refer to the direct changes in energy consumption caused by urban form factors. At the macro level, larger city sizes and highly compact urban structures may improve energy efficiency, on the one hand, but may also increase total emissions due to higher population and activity densities, on the other hand, and their impacts need to be weighed. High building densities and a high floor area ratio will change the urban microclimate and residents’ lifestyles, thereby affecting energy consumption and carbon emissions during building operations. These indirect effects at the meso and micro levels are reflected in the fact that urban form indirectly determines the level of carbon emissions by affecting intermediary factors. It is worth noting that urban carbon emissions usually have obvious spatial dependence and aggregation characteristics, and areas with similar forms often show similar emission levels. Guo et al. [143] further divided urban morphological factors into “geometry” and “construction environment.” Their research revealed that geometric characteristics of urban form influence carbon emissions not only through direct mechanisms but also indirectly by modifying construction environmental factors, such as access to public amenities and the efficiency of the transport system. In summary, the impact mechanism of urban form on carbon emissions is complex, and the intensity of each factor is constrained by conditions such as socioeconomic characteristics. This highlights the importance of developing a comprehensive research framework to systematically characterize the path by which urban form affects carbon emissions and to provide a scientific basis for low-carbon spatial planning.

5.3. Policy Impact

In addition to planning and technical means, policies and management are essential in guiding carbon reduction in cities and buildings. Governments of various countries have formulated multi-level policy frameworks to promote the development of low-carbon cities and green buildings [144,145]. At the urban planning level, compact city policies are widely acknowledged as one of the effective ways to reduce carbon emissions. By increasing urban density and improving public transportation, policies such as reducing commuting distances and dependence on private cars can reduce carbon emissions from the transportation sector. Studies have shown that most compact city policies emphasize the strengthening of public transportation, and high density and reasonable urban spatial layout have a significant influence on urban carbon emissions [146]. Although compact city policies have great carbon reduction potential in theory, there are still large regional differences in actual carbon reduction effects [99]. By exploring the relationship between land use regulations, urban form, and greenhouse gas emissions in nearly 40 middle-income countries, it was found that regions with stricter land use regulations have higher density and lower emissions [147]. However, in major metropolitan areas in the United States, the carbon reduction effect brought about by land use regulations is often not as good as directly levying energy use taxes, and too many land management policies will lead to a significant reduction in population growth [148]. For small cities with serious aging populations, such as Toyama City in Japan, by improving energy efficiency in the city center and promoting housing near public transportation, emphasizing accessibility to create compact communities, effectively reducing urban carbon emissions, it has become a model for global compact city development [146].

At the building level, energy efficiency regulations and standards remain key policy tools for countries to control carbon emissions from building operations [149]. Many developed countries have implemented strict building energy efficiency standards since the end of the last century to reduce building energy consumption. In developing countries, building energy efficiency standards are gradually shifting from voluntary to mandatory, but there is still room for improvement in policy enforcement and coverage [150]. In addition to mandatory regulations, governments have also adopted economic incentives to accelerate the adoption of low-carbon technologies and buildings. Fiscal incentive policies have been proven to be very effective in promoting carbon emission reduction in the building sector [52]. Among them, carbon taxes make energy-saving technologies and renewable energy more economically competitive by increasing the price of high-carbon emission energy. Studies have shown that after the introduction of carbon taxes, the energy intensity of the US commercial building sector is significantly lower than that in the absence of carbon taxes, because carbon taxes promote the transition of the power sector to low-carbon resources and increase users’ motivation to save energy [151]. In addition, a reasonable carbon tax price will encourage developers to choose low-carbon buildings. For China, 200 yuan/ton has the best incentive effect [152]. However, the effects of carbon tax policies are not completely uniform. An evaluation of the EU carbon emissions trading system found that new member states suffered greater losses because industries with higher carbon emissions accounted for a larger proportion of the economy in these countries [153]. In addition, user behavior factors are often ignored in policies, but in fact have a significant impact on building energy performance [154,155]. A systematic evaluation by Hu et al. [156] found that current building energy policies often simplify the role of user behavior, which may weaken the actual energy-saving effect of the policy. The study recommends that the complexity of occupant behavior should be fully considered when formulating building energy-saving policies, and residents’ and users’ behavior should be better incorporated into the policy framework by improving energy information feedback, incentives, and policy evaluation mechanisms. At the policy implementation level, researchers also proposed the need to effectively transform macro policy goals into specific actions at the city and community scale. Lei et al. [157] combined China’s carbon peak policy in the construction sector and constructed a low-carbon community planning evaluation tool based on the Analytic Network Process (ANP) method to refine and prioritize the strategic indicators covered by the policy and guide low-carbon planning practices at the city and neighborhood levels.

5.4. Future Studies

Although research on urban form and carbon emissions has made important progress in recent years, there are still many areas to be deepened and room for improvement, which need to be paid attention to and expanded in future research.

First, the current understanding of the mechanism of urban form affecting carbon emissions is still incomplete, and the uncertainty of the intensity of the impact of various morphological indicators is a prominent problem. Different studies have different conclusions on indicators such as density and compactness, which reflect the lack of a unified analytical framework to integrate the effects of multiple factors. In addition, it is essential to standardize indicators and data to ensure the scientific nature of urban carbon emission research [158]. Future research should develop more sophisticated models and analytical methods to quantitatively evaluate the independent contribution and interactive effects of various morphological elements on carbon emissions. In particular, it is important to enhance the research on intermediary mechanisms, such as how the urban thermal environment, residents’ lifestyle, housing type selection, etc., play a bridging role between urban morphology and carbon emissions, so as to clarify the direct and indirect impact pathways. In summary, there is an urgent need to establish a systematic research framework to guide future work, integrate multidisciplinary methods and multi-scale perspectives, and provide a structured roadmap for low-carbon urban morphology research. A possible framework is: with macro policy goals as the background, determine the key direction of morphological optimization; with meso urban design as the bridge, study the mechanism of action and synergy of morphological variables; with micro technical measures as the support, provide specific solutions and quantitative effects for achieving morphological optimization. The framework should include a clear chain representation of morphological variables, impact mechanisms, and carbon emission results, and incorporate feedback loops (such as resident behavior feedback, climate change impact, etc.). Future research should systematically integrate the classical theories of urban morphology and spatial planning as the conceptual support for the multi-scale analysis framework. For example, at the macro level, the urban metabolism theory emphasizes the quantification of urban material and energy flows, and the “metabolic efficiency” under different urban forms directly affects energy use and carbon emission levels [159]. Critical discussions on the compact city model can also be taken into consideration [160]. Qiang et al. [161] explored the correlation between urban compactness and household CO₂ emissions in more than 284 cities in China. The findings showed that, except for heating-related emissions, household CO₂ emissions were positively correlated with two urban compactness indices. At the meso level, Li et al. [162] used space syntax to quantify spatial structure indicators, thereby establishing a carbon emission transformation pathway to analyze the nonlinear effects of multi-scale spatial structures in severe cold regions. At the micro level, the socio-technical transition model emphasizes that morphological transformation requires the coordinated evolution of technological systems, user behaviors, and institutional structures. Urban carbon emission reduction is not only a planning optimization problem, but also a transformation problem of multi-dimensional social systems [159]. Through such a framework, researchers can more systematically identify the contributions and shortcomings of existing research and thus identify the knowledge gaps that need to be filled in the future.

Second, there is a significant knowledge gap in scale and data. Previous studies have focused on macro-scale urban comparisons and lack a full understanding of the relationship between urban form and carbon emissions at the meso- and micro-scales. Due to the challenges associated with obtaining fine-scale carbon emission data, some studies have limitations in their methods, such as unclear calculation boundaries and incomplete data. Therefore, future research should make greater use of big data and new technologies, such as remote sensing, big data platforms, and IT sensors, to obtain multi-scale, high-temporal, and high-resolution carbon emission and urban form data [163]. Xia et al. [102] proposed a machine learning approach based on a BP (back propagation) neural network to make nonlinear predictions of CO₂ emissions in 2035 under different urban land use intensities. This proves that emerging machine learning and artificial intelligence models can also be used to explore the complex nonlinear relationship between forestry and carbon emissions and improve prediction and scenario simulation capabilities. These technologies will help to make up for the shortcomings of existing data and models and more comprehensively reveal the mechanism by which urban form affects carbon emissions [164]. Deep learning models such as LSTM and Transformer can effectively identify the temporal dynamics and spatial coupling patterns of morphological variables [165]. Graph neural network (GNN) models can construct a structural map of urban space and capture the energy consumption diffusion relationship between cross-regional building complexes [166].

Third, research on policy effectiveness and behavioral factors needs to be strengthened. There is a lack of sufficient quantitative evaluation of how urban planning and building energy-saving policies actually affect carbon emissions and which policy combinations are most effective. In the future, cross-city comparative studies and policy experimental studies should be carried out to monitor the long-term carbon emission reduction effects of measures in order to obtain stronger empirical support. At the same time, research on resident behavior and socioeconomic factors needs to be more deeply integrated into the low-carbon city research framework. For example, how to incorporate the randomness and diversity of user behavior into the building carbon emission model and how to guide residents to adopt a low-carbon lifestyle at the community level are important directions for future research. Finally, future research should pay more attention to the integration of interdisciplinary and life cycle perspectives. Research on low-carbon cities and buildings needs to combine multidisciplinary knowledge to achieve a comprehensive approach and perspective. Existing research has explored the correlation between urban block form and carbon emissions based on classification frameworks such as LCZ. In the future, based on this, a full life cycle low-carbon zone classification can be constructed based on the regional carbon emission intensity of different classification frameworks [123]. In addition, Xia et al. [167] performed a spatial analysis of urban carbon transition and showed that urban growth patterns characterized by strong connectivity and a well-integrated relationship between urban layout and transportation structure are favorable for reducing emissions. Therefore, it is also necessary to strengthen collaborative research among departments such as energy, transportation, and land use. In the future, we should formulate cross-departmental common carbon neutrality goals and roadmaps to coordinate the optimization of urban form with the supply of clean energy and the construction of a green transportation system. In short, future research should focus on filling the above knowledge gaps and developing more complete theoretical frameworks and decision-making support tools. This not only includes refined, multi-scale data and model innovation, but also means conducting comprehensive research at the socio-economic and policy levels to guide urban planning practices towards low-carbon transformation and provide solid scientific support for achieving carbon neutrality goals.

6. Conclusions

Progress in research on the impact of urban form on building carbon emissions is crucial to the sustainable development of cities. Therefore, this study reviews such progress from the past decade and discusses the definition of carbon sources, data characteristics, and accounting methods of building operation carbon emissions, as well as the impact of multi-scale urban form on carbon emissions and design inspiration, in detail. The conclusions drawn from this review are as follows:

(1)

The emission factor method is currently the most widely used carbon emission accounting method. It simplifies the calculation of direct carbon emissions from urban energy use and has a relatively mature carbon emission database.

(2)

At the meso scale, the block development capacity index is an important factor affecting the carbon emissions of the mesoscale blocks and is also affected by the urban climate background. In hot summer and cold winter areas, the floor area ratio and building density jointly play a nonlinear role in enhancing carbon emissions, while in cold areas, a larger floor area ratio and building density can reduce residents’ demand for heating.

(3)

At the meso scale, the block development capacity index, as an important factor affecting meso-block carbon emissions, is affected by both the climate background and building functions. In hot summer and cold winter regions, carbon emissions from office blocks are positively correlated with the floor area ratio, while in extremely cold regions, the relationship between carbon emissions from commercial buildings and the floor area ratio is U-shaped. For hot regions, a high building density will increase building cooling carbon emissions, while for cold regions, a high building density will reduce building heating carbon emissions.

(4)

At the micro scale, the characteristics of the building unit directly determine the heat exchange between the building and the external environment, as well as the indoor light and heat conditions, which have significant impacts on the building’s own carbon emission levels. Building orientation and window-to-wall ratio affect solar heat gain and ventilation and thus affect carbon emissions. Buildings integrating renewable energy, rooftop greening, and highly reflective materials are important ways to reduce building carbon emissions.

(5)

Urban form can indirectly affect carbon emission levels through intermediary factors. At the macro scale, policies and management affect residents’ behavior, thereby affecting carbon emissions. At the meso scale, urban form indicators often affect urban carbon emissions by changing the microclimate of the block.

In addition, this study proposes a theoretical framework for future research: with macro-policy goals as the background, the key directions of morphological optimization can be determined; meso urban design can be used as a bridge to study the mechanisms behind and the synergistic effects of morphological variables; and with the support of micro-technical means, a clear chain representation of morphological variables–influence mechanism–carbon emission results can be constructed and combined with feedback loops, specific plans, and the quantitative effects of morphological optimization. The contribution and scientific significance of this study lie in its ability to help scholars better understand the key aspects of this specific field and make more informed decisions on research topics, content, and methods, providing a reference for the development of low-carbon cities.