Study on Carbon Emission Accounting Method System and Its Application in the Iron and Steel Industry

Abstract

Amid global climate change and the pursuit of carbon neutrality, the steel industry, a major source of carbon emissions, urgently requires a robust and scientific carbon emission accounting system to achieve sustainable development. This study systematically examines carbon emission accounting methods in the steel industry, constructing a comprehensive framework that includes method classification, application analysis, and future trend projections. The aim is to provide theoretical support and practical guidance for the industry’s low-carbon transition. Through in-depth analysis of existing accounting methods, this study summarizes their characteristics regarding accounting boundaries, calculation principles, and data requirements, and explores their current applications and limitations in the steel industry. Looking ahead, the research anticipates that with the advancement of new-generation information technologies and increasing global climate governance demands, carbon emission accounting methods in the steel industry will evolve towards digitalization, refinement, and standardization, offering more reliable data support for the industry’s low-carbon transformation. This research offers a theoretical foundation and practical direction for precise carbon emission accounting and management in steel enterprises and provides a scientific basis for policymakers to develop effective emission reduction policies and strategies, thereby promoting the sustainable development of the steel industry.

1. Introduction

Climate change, a defining challenge of the 21st century, exerts widespread impacts, including ecosystem degradation, agricultural decline, and rising sea levels [1,2,3]. In order to deal with greenhouse gas emissions, the international community has established a global emission reduction framework through the Kyoto Protocol and the Paris Agreement. As one of the main sources of carbon emissions, the steel industry is facing tremendous pressure to reduce emissions, especially in the context of the EU Carbon Border Adjustment Mechanism (CBAM) and the carbon neutrality commitments of various countries. This industry now confronts unprecedented challenges in emission reduction, underscoring the urgency for effective carbon accounting systems to assess the efficacy of mitigation strategies and guide targeted policy development. In 2019, the global steel industry’s carbon emissions totaled 2.8 billion tons CO₂eq, with China contributing 1.62 billion tons, representing 11% of the nation’s total emissions [4]. Therefore, it is urgent for the steel industry to develop an effective carbon emission accounting system to scientifically assess the effectiveness of emission reduction measures and formulate accurate environmental impact mitigation strategies, so as to meet the international climate goals [5]. Accurate carbon emission accounting is pivotal for enterprises in the iron and steel industry, enabling the identification of critical emission points, optimization of production processes, enhancement of energy efficiency, and establishing a robust data foundation for participation in global carbon market trading [6].

Carbon emission accounting methodologies in the steel industry have evolved from a single country-level inventory approach, such as the Intergovernmental Panel on Climate Change (IPCC) method, to a multi-dimensional, complex system covering life cycle assessment (LCA), input–output analysis (IOA), and Kaya identity. With the strengthening of emission reduction requirements and the advancement of technology, these methods have improved in accuracy and scope, but at the same time, they have exposed their respective limitations. For example, the IPCC method is the basis for the preparation of international greenhouse gas inventories, covering key emission sources such as energy, industrial processes, agriculture and land use change. However, its broad applicability challenges the precision required to detect microscopic changes at the firm or project level [7,8,9]. LCA provides a comprehensive evaluation of a product’s environmental impacts from cradle to grave, demanding high data integrity and numerous assumptions when data is incomplete, potentially affecting analysis accuracy [10,11]. IOA, a “top-down” approach, captures inter-sectoral economic interactions and carbon emission transfers, yet its lengthy compilation cycle and lack of temporal resolution limit micro-level analysis [12,13,14]. The Kaya identity simplifies the carbon emissions analysis framework by dissecting emissions through key variables such as population, per capita GDP, energy consumption per unit GDP, and carbon emissions per unit energy. However, this approach, while offering a clear lens into macro-level emission dynamics, struggles to elucidate the causal interplay among these factors at the micro-level, such as within individual enterprises or projects [15,16,17]. Current research highlights a notable absence of systematic synthesis and practical guidance for carbon emission accounting methodologies, particularly within the context of the iron and steel industry’s intricate production processes and variegated energy consumption patterns [18]. The applicability of these methods to specific industrial processes is underexplored, and with the advent of policies such as the CBAM, there is an acute need for the iron and steel industry to adopt a unified, scientifically robust, and nuanced carbon emission accounting system.

Within the iron and steel industry’s carbon emission accounting methodologies, the coexistence of “top-down” and “bottom-up” approaches [19] has not yet been systematically dissected to elucidate their individual characteristics and limitations. Moreover, the practical application of these methods is fraught with issues such as inconsistent accounting boundaries and substantial divergences in calculation principles. This paper begins by categorizing the carbon emission accounting systems, providing an in-depth introduction to the “bottom-up” method exemplified by LCA and the “top-down” method founded on IOA. Subsequently, the paper makes an in-depth study of the specific applications of international common methods (IPCC method, International Iron and Steel Institute method, ISO 14404 standard [20]) and regional accounting methods in accounting boundaries, calculation principles, and data requirements. Finally, the paper summarizes the current research progress of carbon emissions accounting in the iron and steel industry, including the application of LCA technology, the quantification of carbon footprints, and the innovative refinements to carbon emission accounting methodologies.

Upon an exhaustive examination of the carbon emission accounting methodologies and their application within the iron and steel sector, this paper establishes a comprehensive research framework. It encompasses methodological categorization, application analysis, and the projection of future trends. This framework offers innovative insights into addressing pivotal challenges in carbon emission accounting within the industry, such as the selection of appropriate methods, the definition of accounting boundaries, and the specification of data requirements. This study furnishes a theoretical foundation and pragmatic direction for iron and steel enterprises in the quest for precise carbon emission accounting and efficacious management. Concurrently, it affords policymakers a scientific underpinning for the development of pertinent emission reduction policies and strategies. As new-generation information technology advances and the imperatives of global climate governance intensify, the iron and steel industry’s carbon emission accounting methodologies are on a trajectory towards digitalization, refinement, and standardization. This shift is expected to yield a more robust and reliable dataset, which is pivotal for the industry’s low-carbon transition. Driven by the ambition to achieve global carbon neutrality and sustainable development principles, the steel industry’s research in this field will continue to deepen, and methods and applications are expected to continue to improve and evolve.

2. Classification of Carbon Emission Accounting System

In accordance with the research ideas, methodologies for carbon footprint accounting are divided into two distinct categories: the “bottom-up” and the “top-down” approaches [21]. The “bottom-up” methodology is exemplified by LCA, a systematic process that collates and scrutinizes activity data across each phase of a product’s lifecycle to appraise its environmental ramifications. Conversely, the “top-down” methodology is predicated on IOA, which tracks carbon emissions by dissecting the economic interdependencies among various sectors. The two aforementioned methods have been extensively implemented in carbon footprint accounting within the iron and steel industry, yet they exhibit distinct characteristics and scopes of applicability.

2.1. LCA

LCA is a rigorous ‘bottom-up’ process analysis methodology, designed to systematically evaluate the potential environmental impacts of products or services across their entire life cycle, from cradle to grave. This approach is aligned with international standards, such as ISO 14040 [22]. This study constructs a technical framework that encapsulates four core steps: goal and scope definition, life cycle inventory analysis, life cycle impact assessment, and result interpretation. The goal and scope definition phase establishes the evaluation object and the research boundary, which lays the foundation for the subsequent evaluation; the life cycle inventory analysis phase establishes the material and energy input and output lists of the product life cycle through systematic data collection; the life cycle impact assessment phase translates the inventory data into quantifiable environmental impact indicators. In the result interpretation phase, the evaluation results are analyzed and discussed to provide a scientific basis for decision-making.

The current life cycle impact assessment methods are mainly divided into two categories: the midpoint method (mid-point) and the endpoint method (end-point) [23]. The midpoint method encompasses established frameworks such as CML2001, ReCiPe midpoint (H), and the IPCC GWP 100a [24], which quantify and characterize environmental impacts based on pollutant equivalents. Notably, the CML2001 method is extensively utilized in the environmental impact assessment of products and processes, while the ReCiPe midpoint rule places a greater emphasis on the characterization of direct environmental impacts. The IPCC GWP 100a, on the other hand, is specifically tailored towards the assessment of greenhouse gases, employing the global warming potential to convert the impacts of various greenhouse gases into CO₂ equivalents. These methodologies are characterized by their straightforward evaluation processes and accessibility of data; however, they often fall short in accurately reflecting the ultimate environmental impacts. The endpoint approach includes the IMPACT2002+ and ReCiPe endpoint methods, among others [25]. These methods prioritize the comprehensive environmental damage incurred by the exposure of receptors encompassing human health, health systems, resources, and more to various emissions. The interdisciplinary approach combines environmental science, environmental meteorology, toxicology, and epidemiology, reflecting the systematic and complex nature of environmental impact assessments. The IMPACT2002+ method facilitates a holistic assessment of environmental damage, integrating diverse environmental impacts into a final damage index through weighting; the ReCiPe endpoint method offers a multidimensional impact assessment system, evaluating three dimensions: human health, ecosystem quality, and resource availability. These methods are compared in Table 1.

Rui Zhang et al. [26] use the ReCiPe2016 method for life cycle assessment to meticulously quantify the environmental performance of electrolytic manganese metal production. The normalization results at the midpoint of the ReCiPe2016 are depicted in Figure 1, revealing that the scarcity of mineral resources and the impact category of terrestrial ecotoxicity exert the most substantial influence on the environment. Notably, the scarcity of mineral resources secures the highest score across all impact categories. In contrast, the impact of global warming and the scarcity of fossil resources is comparatively minor, collectively accounting for 13% of the overall environmental burden.

Lingyan Deng et al. [27] used the ReCiPe2016 method and the IMPACT2002+ method to carry out LCA on the environmental impact of a by-gas utilization system in the iron and steel manufacturing process. The ReCiPe2016 method was deployed to scrutinize the environmental performance, with the normalization results presented in Figure 2. The sum of the weighted factors (that is, human health, ecosystem, and resource damage) will be 100% at any point within the triangle, with each weighted factor ranging from 0% to 100%. The findings indicated that, from an environmental standpoint, regions such as Ontario, Finland, and China are best suited for the implementation of methanol production systems (CBMeOH), while the United States and Mexico would benefit more from adopting combined cycle power plant systems (CCPP). The IMPACT2002+ method was also utilized to corroborate these findings, with the results depicted in Figure 3. The analysis revealed a concordance with the outcomes obtained from the ReCiPe2016 method. In essence, the CBMeOH system emerges as the most environmentally benign option for facilities in Ontario, Finland, and China, whereas the CCPP system is deemed the most environmentally friendly for those in the United States and Mexico.

Di Maria Francesco et al. [28] employed an LCA approach to evaluate the environmental and health impacts of municipal solid waste incineration (MSWI), utilizing the IMPACT 2002+ method for a quantitative assessment of human health endpoint indices. The analysis, as depicted in Figure 4, identifies direct and indirect emissions from slag and fly ash landfills, particularly the inertization of fly ash, as the primary sources of positive average net human toxic non-cancer (HTnc) impacts (CTUh/tonnes MSW). In contrast, contributions from incineration emissions and slag treatment are minimal. The endpoint indicators for human toxic cancer (HTc) and overall human health (HH) suggest that the environmental impacts of MSWI are partly mitigated by the benefits of energy and material recycling. This study highlights the differential impacts of small, medium, and large incineration facilities, emphasizing the potential for reducing environmental and health impacts through optimized incineration processes and enhanced resource recovery. The findings provide critical insights to inform policy and technological advancements in waste management, steering towards more sustainable practices.

Yuan et al. [29] utilized the IMPACT 2002+ methodology to quantitatively evaluate the life cycle impacts of municipal solid waste management (MSWM) systems in China, with a particular focus on three proposed waste classification scenarios. The findings, as depicted in Figure 5, reveal that Scenario S-1 exerts the minimal impact on ecosystem quality, indicating its superior performance in safeguarding biodiversity and ecosystem services. Conversely, Scenario S-2 emerges as the most significant contributor to human health, ecosystem quality, and climate change, potentially attributable to its higher energy and resource consumption during processing. Scenario S-3, on the other hand, is associated with the least adverse effects on human health, climate change, and resource depletion, suggesting that more refined waste classification can substantially mitigate environmental harm. The study also incorporated a Monte Carlo uncertainty analysis to substantiate the robustness of the findings, ensuring the statistical validity of the comparisons across scenarios. This research offers policymakers data and insights, underscoring the potential of waste classification to boost the environmental performance of MSWM systems and highlighting the necessity for a judicious balance when designing MSWM systems to consider diverse environmental impacts. The study enhances the understanding of the integrated effects of waste classification on environmental and human health and provides a scientific foundation for refining MSWM practices.

The endpoint method, with its comprehensive evaluation dimensions, better reflects the evolution of carbon footprint accounting in the iron and steel industry, from single-index to integrated assessments, positioning it as a research priority and future direction. This trend is mainly reflected in the following aspects: first, the endpoint method’s superiority is epitomized in its ability to translate diverse environmental stressors, such as energy consumption, raw material utilization, and emission discharges in iron and steel production, into quantifiable final environmental damages. This transformation renders the evaluation outcomes more interpretable and readily comparable [30]. Secondly, through interdisciplinary research, the endpoint method provides a more scientific and systematic evaluation framework, which can better reflect the complexity of the environmental impact of the upstream and downstream of the iron and steel industry chain [31]. The evaluative outcomes yielded by the endpoint method align more closely with the decision-making imperatives of iron and steel enterprises. They offer actionable insights that can be directly harnessed to optimize production processes and enhance the adoption of cleaner production technologies [32]. However, there are still some limitations in the current methods of life cycle assessment in the iron and steel industry [33,34,35,36,37]: first, there is a notable heterogeneity in the assessment methods and impact categories employed across studies. Some research focuses exclusively on greenhouse gas emissions, while others concurrently evaluate a spectrum of environmental impacts, including acidification and eutrophication, leading to a lack of comparability in results. Second, the endpoint method, due to its interdisciplinary nature, entails a complex evaluation process. This is particularly evident in the assessment of critical processes such as high-temperature operations and chemical reactions in steel production, where extensive data requirements amplify the uncertainty in evaluation. Third, most of the existing evaluation methods and databases are based on developed countries, which may not be fully applicable to the actual situation of China’s iron and steel industry, such as different process routes, energy structure, and pollutant emission characteristics.

Therefore, future research needs to promote the standardization of evaluation methods, data quality control, and cross-industry applicability. To this end, a dual-pronged approach is advocated: firstly, the formulation of a standardized framework that is attuned to the global iron and steel industry, complemented by a robust enhancement of emission factor databases; and secondly, the fortification of synergies with complementary evaluative techniques, such as material flow analysis (MFA) [38] to track raw material flows, and economic assessments to dissect the financial implications of emission mitigation across diverse national contexts. Furthermore, the development of cutting-edge, real-time monitoring technologies, underpinned by the principles of intelligent manufacturing and the pervasive reach of the global industrial internet [39], is posited as a critical third pillar. Such advancements are envisioned to augment the precision and immediacy of data acquisition, thereby mitigating the pervasive issue of uncertainty in current evaluative practices. By marshaling these strategic enhancements, the scholarly community stands poised to furnish a more holistic and robust evaluative foundation for the low-carbon transition and the pursuit of sustainable development within the global steel industry. Concurrently, these efforts are anticipated to foster a climate conducive to international collaboration and the sharing of best practices, in a concerted bid to surmount the formidable challenges posed by global climate change.

Table 1. Comparison of the advantages and disadvantages of the LCA method.

LCA Method		Advantages	Disadvantages	Application Scenario	References
Midpoint method	CML2001	(1) A wide range of environmental impact categories are covered to provide a comprehensive assessment of the potential environmental burden and impact of a product, process or activity. (2) Standardized evaluation criteria are provided to make comparisons between different studies and projects possible.	(1) The assessment demands substantial input data, including material flows, energy use, and emissions. (2) The application may require adjustments based on the environmental and socio-economic conditions of specific regions. (3) The result is affected by the uncertainty of the data input, and the reliability of the result needs to be evaluated through uncertainty analysis.	Applicable for in-depth analysis of specific environmental issues.	[40,41,42]
	Middle point (H) of ReCiPe	(1) The midpoint index can be combined and compared flexibly, which is helpful to analyze specific environmental problems, such as global warming potential, acidification potential, etc. (2) The calculation process of the midpoint method is more transparent and easier to understand and verify.	(1) Directly correlating these to specific environmental damages is challenging. (2) Multiple impact categories may be involved, requiring more data and complex calculations, adding to the complexity of the assessment.	Applicable to contexts necessitating a holistic assessment of the environmental impacts associated with products or processes.	[43,44,45]
	IPCC GWP100a	(1) Provides a clear time frame for the assessment of long-term climate change impacts. (2) Provides a standardized framework for comparison that allows the emission impacts of different greenhouse gases (e.g., carbon dioxide, methane, nitrogen oxides, etc.) to be compared.	(1) A 100-year time horizon may not be enough to capture the full impact of some greenhouse gases, especially those with longer lifespans. (2) A single indicator may not fully reflect the complex impacts of GHG emissions on the environment and socio-economic systems. (3) The focus on global warming potential may overlook the direct impact of greenhouse gas emissions on other environmental aspects, such as the ozone layer and air quality.	Applicable for evaluating the contribution of greenhouse gas emissions from products or activities to global warming.	[24,46,47]
End-point method	IMPACT2002+	(1) Integrates multiple assessment models to provide a comprehensive environmental impact assessment, including multiple dimensions such as human health, ecosystem quality, and climate change. (2) Global and regional environmental impacts are considered, making them applicable to environmental assessment at different regional and global scales. (3) Allowing the standardization and weighting of different environmental impact types helps to compare the importance of different impact types and identify the most critical environmental issues.	(1) Due to the integration of multiple evaluation models, the operation can be complex and requires professional knowledge and skills to apply correctly. (2) As with all LCA methods, the results of IMPACT2002+ are subject to uncertainty in the input data and require uncertainty analysis to assess the reliability of the results.	Applicable for assessing the overarching impacts of products or activities on human health and ecosystem quality from a macroscopic perspective.	[48,49,50]
	ReCiPe end method	(1) Directly links to ultimate damages to human health and ecosystems, offering a more intuitive understanding of environmental impacts. (2) Integrates multiple midpoint effects into a few major endpoint effects, simplifies the evaluation process, and makes the results easier to interpret and communicate. (3) A damage-oriented approach helps to assess and compare the overall impact of different products or activities on human health and ecosystems.	(1) Complex environmental processes may need to be simplified, which may lead to increased uncertainty in the outcome. (2) It may not be possible to provide detailed information about specific environmental issues because it combines multiple impacts into a few endpoint categories.	Applicable for evaluating the environmental impacts of products or processes across multiple dimensions.	[51,52,53,54]

Table 1. Comparison of the advantages and disadvantages of the LCA method.

LCA Method		Advantages	Disadvantages	Application Scenario	References
Midpoint method	CML2001	(1) A wide range of environmental impact categories are covered to provide a comprehensive assessment of the potential environmental burden and impact of a product, process or activity. (2) Standardized evaluation criteria are provided to make comparisons between different studies and projects possible.	(1) The assessment demands substantial input data, including material flows, energy use, and emissions. (2) The application may require adjustments based on the environmental and socio-economic conditions of specific regions. (3) The result is affected by the uncertainty of the data input, and the reliability of the result needs to be evaluated through uncertainty analysis.	Applicable for in-depth analysis of specific environmental issues.	[40,41,42]
	Middle point (H) of ReCiPe	(1) The midpoint index can be combined and compared flexibly, which is helpful to analyze specific environmental problems, such as global warming potential, acidification potential, etc. (2) The calculation process of the midpoint method is more transparent and easier to understand and verify.	(1) Directly correlating these to specific environmental damages is challenging. (2) Multiple impact categories may be involved, requiring more data and complex calculations, adding to the complexity of the assessment.	Applicable to contexts necessitating a holistic assessment of the environmental impacts associated with products or processes.	[43,44,45]
	IPCC GWP100a	(1) Provides a clear time frame for the assessment of long-term climate change impacts. (2) Provides a standardized framework for comparison that allows the emission impacts of different greenhouse gases (e.g., carbon dioxide, methane, nitrogen oxides, etc.) to be compared.	(1) A 100-year time horizon may not be enough to capture the full impact of some greenhouse gases, especially those with longer lifespans. (2) A single indicator may not fully reflect the complex impacts of GHG emissions on the environment and socio-economic systems. (3) The focus on global warming potential may overlook the direct impact of greenhouse gas emissions on other environmental aspects, such as the ozone layer and air quality.	Applicable for evaluating the contribution of greenhouse gas emissions from products or activities to global warming.	[24,46,47]
End-point method	IMPACT2002+	(1) Integrates multiple assessment models to provide a comprehensive environmental impact assessment, including multiple dimensions such as human health, ecosystem quality, and climate change. (2) Global and regional environmental impacts are considered, making them applicable to environmental assessment at different regional and global scales. (3) Allowing the standardization and weighting of different environmental impact types helps to compare the importance of different impact types and identify the most critical environmental issues.	(1) Due to the integration of multiple evaluation models, the operation can be complex and requires professional knowledge and skills to apply correctly. (2) As with all LCA methods, the results of IMPACT2002+ are subject to uncertainty in the input data and require uncertainty analysis to assess the reliability of the results.	Applicable for assessing the overarching impacts of products or activities on human health and ecosystem quality from a macroscopic perspective.	[48,49,50]
	ReCiPe end method	(1) Directly links to ultimate damages to human health and ecosystems, offering a more intuitive understanding of environmental impacts. (2) Integrates multiple midpoint effects into a few major endpoint effects, simplifies the evaluation process, and makes the results easier to interpret and communicate. (3) A damage-oriented approach helps to assess and compare the overall impact of different products or activities on human health and ecosystems.	(1) Complex environmental processes may need to be simplified, which may lead to increased uncertainty in the outcome. (2) It may not be possible to provide detailed information about specific environmental issues because it combines multiple impacts into a few endpoint categories.	Applicable for evaluating the environmental impacts of products or processes across multiple dimensions.	[51,52,53,54]

2.2. IOA

IOA, a “top-down” methodology initially introduced by American economist Wassily Leontief in 1936, is designed to examine the economic interdependencies among sectors [55]. The technical framework of IOA encompasses four interconnected phases: the selection and processing of the input–output table, the construction of a carbon footprint model, the accounting of carbon footprints, and the analysis of the results [56]. In the phase of selecting and processing the input–output table for IOA, it is imperative to choose an input–output table with an appropriate scale and timing that aligns with the accounting objectives, followed by preprocessing such as sector aggregation to mitigate errors stemming from inter-sectoral differences in production technology and scale. The subsequent phase constructs the carbon footprint model by defining three core elements: (1) the carbon emission intensity coefficient matrix, (2) the Leontief inverse matrix, and (3) the final demand matrix [57]. These matrices quantify economic interdependencies and sectoral carbon transfers. The carbon emission intensity coefficient matrix is derived from energy consumption, carbon emission factors, and sectoral output values detailed in statistical yearbooks, while the Leontief inverse matrix and final demand matrix are derived from input–output tables. During the carbon footprint accounting and result analysis phase, the complex economic interdependencies between sectors are translated into the physical relationships of greenhouse gas emissions, clarifying the nexus between direct and indirect carbon emissions.

IOA has specifically evolved into two distinct models for application: the single-region input–output (SRIO) model and the multi-region input–output (MRIO) model, each tailored to address the economic interdependencies within and across different geographical scopes, respectively [58].

The SRIO model, as the pioneering application of IOA, simplifies the data collection and computational demands of traditional IOA by positing a critical assumption [59]: the technology level of imported goods and services is equivalent to that of domestic production, as illustrated in

Table 2. Data form of the Tokyo single-regional input–output table [60].

			Intermediate Consumption			Final Demand				Import Inf
						Consumption Export Outflow and Investment
			T (Tokyo)			T	O
			1	…	n
Intermediate Input	T	1	${ttT}^{11}$		${ttT}^{11}$	${tty}^1$	${toy}^1$	${tE}^1$	${tN}_{out}^1$	${tM}^1$	$tN$
		…
		n	${ttT}^{n1}$		${ttT}^{nn}$	${tty}^n$	${toy}^n$	${tE}^n$	${tN}_{out}^1$	${tM}^n$	$tN$
Added Value			${tV}_1$		${tV}_n$
Gross Input			${tX}_1$		${tX}_n$

. The SRIO model, as one of the early applications of IOA, simplifies the data collection and calculation process by assuming that the technical level of imported products and services is equivalent to domestic production. This assumption, while simplifying the analysis, overlooks the impact of production technology differences between domestic and foreign entities on carbon footprints, leading to significant distortion in the implied carbon accounting outcomes for imported and exported products. In reality, as most countries and regions operate as open economies with intermediate goods and final consumption reliant on imports, the associated greenhouse gas emissions and resource consumption are not accurately reflected within the SRIO model framework.

The MRIO model, evolving from the Inter-Regional Input–output (IRIO) model, is depicted in its fundamental form in

Table 3. Chinese MRIO table [62].

Input			Output
			Intermediate Use												Final Use				Total
			Province 1				…	Province m			Province 1			…	Province m			Exports	Output
			Sector		…	Sector n		Sector	…	Sector n	Category	…	Category		Category	…	Category
			1					1			1		k		1		k
Intermediate	Province 1		Sector 1	$z_{11}^{11}$	…	$z_{1n}^{11}$		$z_{11}^{1m}$	…	$z_{1n}^{1m}$	$y_{11}^{11}$		$y_{1k}^{11}$		$y_{11}^{1m}$	…	$y_{1k}^{1m}$	$e_1^1$	$x_1^1$
input
			⋮	⋮		⋮	…	⋮		⋮	⋮		⋮	…	⋮		⋮	⋮	⋮
			Sector n	$z_{n1}^{11}$	…	$z_{nn}^{11}$		$z_{n1}^{1m}$	…	$z_{nn}^{1m}$	$y_{n1}^{11}$	…	$y_{nk}^{11}$		$y_{n1}^{1m}$	…	$y_{nk}^{1m}$	$e_n^1$	$x_n^1$
		⋮					⋮							⋮				⋮	⋮
			Sector 1	$z_{11}^{m1}$	…	$z_{1n}^{m1}$		$z_{11}^{mm}$	…	$z_{1n}^{mm}$	$y_{11}^{m1}$		$y_{1k}^{m1}$	…	$y_{11}^{mm}$		$y_{1k}^{mm}$	$e_1^m$	$x_1^m$
	Province m		⋮	⋮		⋮	…	⋮		⋮		…	⋮		⋮	…		⋮	⋮
			Sector n	$z_{n1}^{m1}$	…	$z_{nn}^{m1}$		$z_{n1}^{mm}$	…	$z_{nn}^{mm}$	$y_{n1}^{m1}$		$y_{nk}^{m1}$	…	$y_{n1}^{mm}$		$y_{nk}^{mm}$	$e_n^m$	$x_n^m$
		Imports		$I_1^1$	…	$I_n^1$	…	$I_1^m$	…	$I_n^m$	$I_1^{f,1}$	…	$I_k^{f,1}$	…	$I_1^{f,m}$	…	$I_k^{f,m}$
Value-added				$v_1^1$	…	$v_n^1$	…	$v_1^m$	…	$v_n^m$
Total input				$x_1^1$	…	$x_n^1$	…	$x_1^m$	…	$x_n^m$
Direct carbon emissions				$d_1^1$	…	$d_n^1$	…	$d_1^m$	…	$d_n^m$

. The IRIO model was initially developed to capture the technological disparities in production between regions within a nation, addressing the limitations inherent in the SRIO model. The MRIO model diverges from the IRIO model primarily in its approach to calculating regional production technology matrices. To surmount data constraints, the global MRIO model standardizes production technology within nations, making it a streamlined version of the IRIO model [61]. The MRIO model, a successful application of economic principles in environmental research, enables the precise tracking of the geographical and spatial distribution of environmental impacts. This model provides a viable approach for quantifying the cross-regional transfer and distribution of waste emissions or resource consumption, offering detailed insights into the economic relationships that drive environmental pressures across different regions.

The MRIO model has emerged as the preeminent and most widely adopted model within the field of IOA [63]. The disparities in data sources and compilation methods across this model, such as Eora, WIOD, EXIOBASE, and GTAP, particularly in the calculation of carbon emission factors, introduce uncertainty affecting the comparability and credibility of accounting results. Studies have indicated that these variations can lead to discrepancies of up to 50% in carbon footprint accounting for major economies like the United States, China, Russia, and India [64,65]. As the globalized production landscape evolves, the traditional MRIO model confronts new challenges, notably its inability to account for technological disparities between nations and regions. Consequently, scholars have advanced the development of a multi-scale, multi-regional input–output (MSMRIO) model to address this shortcoming more effectively. For instance, applications of the MSMRIO model in examining trade dynamics between China and Japan, ASEAN, East Asia, and the United States have yielded promising outcomes, demonstrating the ongoing evolution and potential for further enhancement of the MRIO model in the future.

IOA, as a “top-down” approach, has significant advantages in macro-scale carbon footprint analysis, as shown in

Table 4. Summary of the carbon accounting emission system.

Carbon Emission Accounting System
Method	LCA	IOA
Scope of application	Suitable for accounting at the micro-level, such as individual products. In steel production, it is suitable to analyze the carbon footprint of individual products or production links.	Applicable to macro-level computations, such as those conducted at the national or corporate sector scale. In steel production, it is suitable to analyze the carbon emissions of the whole industry.
Advantages	(1) It encompasses the entire life cycle of products or services, spanning from raw material procurement to production, usage, and waste management, thereby offering a holistic assessment of environmental impacts. (2) The assessment integrates insights from environmental science, environmental meteorology, toxicology, epidemiology, and other disciplines, underscoring the systematic and intricate nature of environmental impact evaluations. (3) High precision.	(1) This approach is adept at delivering a comprehensive quantification of the carbon footprint and environmental impacts associated with products or services, providing a holistic view of their ecological footprint. (2) It differentiates between direct and indirect greenhouse gas emissions across the life cycle, adeptly addressing the pervasive truncation errors that often arise at system boundaries during life cycle assessments. (3) Demonstrating high operational feasibility, the method, upon establishment, can rapidly produce accounting outcomes, offering a swift and efficient means of data synthesis for environmental impact evaluations.
Disadvantages	(1) There are major obstacles in the acquisition of macro-level data. The heterogeneity in the evaluation methods and impact types across various studies introduces substantial discrepancies, thereby diminishing the comparability of research outcomes. (2) The evaluative process is notably intricate, particularly when it comes to assessing the high-temperature processes, chemical reactions, and other specialized stages within steel production, which increase the uncertainty of the evaluation. (3) The prevailing evaluation methodologies and databases, largely contoured by the contexts of developed countries, may not be entirely germane to the nuanced realities of China’s steel industry.	(1) Micro-level applications, such as those at the organizational and product levels, often incur substantial errors. (2) The input–output table has a lengthy preparation period with slow updates, and there are notable discrepancies between various databases. (3) The annual aggregate data in the input–output table struggles to support more granular time-scale analyses, thereby constraining the utility of IOA in dynamic assessments.
Improvement direction	(1) Establish a localized evaluation system that is tailored to the nuances of China’s steel industry, and enhance the emission factor database to ensure it reflects the specificities of this sector. (2) Integrate this system more robustly with other complementary evaluation methodologies to create a more holistic assessment framework. (3) Leveraging advancements in intelligent manufacturing and industrial Internet, developing real-time monitoring technologies that can significantly bolster the precision and immediacy of data acquisition, thereby informing more accurate environmental impact assessments.	(1) Enhancing the intrinsic limitations of IOA involves increasing data update frequency, standardizing evaluation metrics, and bolstering database infrastructure. (2) Integrating IOA with life cycle assessment strategies can leverage their synergistic strengths, merging macroeconomic statistics with micro-level process data to forge a more holistic and robust environmental impact assessment framework.
Calculation error	Due to the subjectivity of boundary setting and truncation error, it may lead to uncertainty in evaluation results.	It has good system integrity at the macro level, but due to the differences in departmental mergers and the annual summary characteristics of data, the accounting results at the micro level are not accurate enough.
Data dependency	A large amount of high-precision data is needed to support its detailed life cycle analysis. Data is usually mainly based on physical units, which requires high data quality.	Relying on the input–output table, its data update cycle is long, and it is difficult to quickly reflect the impact of technological changes on the environment. Data is usually dominated by monetary units, which makes it difficult to accurately reflect physical flow.

, being able to capture the complexity of cross-sectoral economies through input–output tables, systematically distinguish direct and indirect greenhouse gas emissions, and avoid system boundary truncation errors common in LCA. However, there are significant inaccuracies in the application of IOA at the micro level, such as at the individual organization and product level, in contrast to the precision of LCA at the micro level. In the steel production process, LCA is more suitable for analyzing the carbon footprint of a single product or production link, because it can evaluate the environmental impact at all stages from raw material mining, production, and processing to waste disposal in detail. IOA is more suitable for analyzing the overall carbon emissions of the steel industry, and can reflect the economic ties between various departments within the industry and the transmission relationship of carbon emissions. From the calculation error, LCA may lead to the uncertainty of evaluation results due to the subjectivity of boundary setting and truncation error. However, IOA may have inaccurate accounting results at the micro level due to differences in departmental mergers and annual summary characteristics of data. In terms of data dependence, LCA requires a large amount of high-precision data to support its detailed life cycle analysis [66], while IOA relies on input–output tables, and its data update cycle is long [67], making it difficult to quickly reflect the impact of technological changes on the environment. Furthermore, annual summary data for input–output tables limits the application of IOA to more detailed, dynamic time-scale analyses. Therefore, IOA is more economical and practical in carbon footprint analysis at the macro level, but its accuracy at the micro level is insufficient, and the data timeliness and dynamics are poor. These limitations limit its application in some scenarios.

To address the limitations of IOA, current mainstream research is optimized in two primary dimensions: enhancing the methodology of IOA itself and integrating IOA with LCA. Enhancements to IOA include increasing the frequency of data updates, standardizing evaluation criteria, and bolstering database construction, which are crucial for improving the accuracy and reliability of environmental impact assessments. Conversely, the integration of IOA and LCA leverages the complementary strengths of both methods, merging macro-level statistical data with micro-level process data to establish a more comprehensive and robust environmental impact assessment system. Li et al. [68] used the Economic Input–output Life Cycle Assessment (EIO-LCA) model to decompose environmental impacts into various stages of the production chain, considering not only direct emissions, but also upstream industries (such as raw material supply, energy production) and indirect emissions from downstream industries (such as product use and waste disposal). The results show that coke and coal produce the most direct CO₂ emissions. This comprehensiveness enables EIO-LCA to more accurately assess CO₂ emissions throughout the steel industry’s life cycle, providing support for the development of more effective emission reduction strategies.

3. Analysis on the Application of Carbon Emission Accounting System

Based on the previous analysis of the main carbon emission accounting methodologies, this chapter further studies the carbon emission accounting methodologies currently in use, introducing the international standard methods represented by the IPCC method, the ISO 14404 standard, and the International Iron and Steel Institute method, and analyzes the regional standards such as the CBAM and the Chinese standard. A systematic comparison of these methodologies is conducted, focusing on aspects of accounting boundaries, calculation principles, and data requirements, to provide a robust framework for carbon emission assessments.

3.1. International General Method

3.1.1. IPCC National Greenhouse Gas Inventory

The 2006 IPCC Guidelines for National Greenhouse Gas Inventories, meticulously crafted by the IPCC’s Task Force on National Inventories, underwent a comprehensive revision in 2019 [69]. These guidelines offer a triad of fundamental methodologies for emissions estimation: the emission factor method, predicated on default coefficients; the mass balance method, grounded in the principles of material equilibrium; and the actual measurement method, which relies on direct empirical data. These approaches collectively form the cornerstone of a rigorous framework for assessing national greenhouse gas emissions.

• Emission factor method

The Emission Factor Method employs an estimation approach where default emission factors are multiplied by national production data [70]. Given the substantial variability in emissions per unit of steel output due to the diversity in steelmaking processes, it is imperative to ascertain the proportion of steel produced by different types of metallurgical processes. Thereafter, emissions from each process are calculated, and the estimates are aggregated to derive a comprehensive assessment.

CO₂ emissions from steel production:

$$E_{{CO}_2,Non-energy}=BOF·{EF}_{BOF}+EAF·{EF}_{EAF}+OHF·{EF}_{OHF}$$

(1)

CO₂ emissions from pig iron production from unprocessed steel:

$$E_{{CO}_2, Non-energy}=IP·{EF}_{IP}$$

(2)

CO₂ emissions from direct reduction iron production:

$$E_{{CO}_2, Non-energy}=DRI·{EF}_{DRI}$$

(3)

CO₂ emissions from molten slag production:

$$E_{{CO}_2,Non-energy}=SI·{EF}_{SI}$$

(4)

CO₂ emissions from sand blocks production:

$$E_{{CO}_2,Non-energy}=P·{EF}_P$$

(5)

where:

• E_{CO2, Non-energy} = CO₂ emissions to be reported in the IPPU department, in tons;
  BOF = The amount of crude steel of the alkaline oxygen converter, produced in tons;
  EAF = Crude steel quantity of electric arc furnace produced, in tons;
  OHF = The amount of crude steel produced, in tons;
  IP = Output of pig iron not converted into steel, in tons;
  DRI = The quantity of direct reducing iron produced by the state, in tons;
  SI = The amount of molten slag produced by the state, in tons;
  P = The amount of pellets produced by the state, in tons;
  EF_x = Emission factor, measured in x per ton of CO₂/production.

The emission factor method, noted for its simplicity and low data demand, is marred by high computational uncertainty, a critical factor in emissions inventory accuracy.

2.

Mass balance method

The mass balance method takes into account the inputs and outputs of other processes [71] and is suitable for obtaining national data on the use of process materials in steel production, slag production, pellet production and direct reduction iron production. Notably, by considering the actual input quantities contributing to CO₂ emissions, this method surpasses the emission factor method in accuracy, providing a robust foundation for emissions inventory assessments.

CO₂ emissions from steel production:

$$E_{{CO}_2,Non-energy}=\left[\begin{array}{c}PC·C_{PC}+{\displaystyle \sum _a}\left({COB}_a·C_a\right)+CI·C_{CI}+L·C_L+D·C_D+CE·C_{CE}\\ +{\displaystyle \sum _b}\left(O_b+C_b\right)+COG·C_{COG}-S·C_S-IP·C_{IP}-BG·C_{BG}\end{array}\right]·{\displaystyle \frac{44}{12}}$$

(6)

CO₂ emissions from molten slag production:

$$E_{{CO}_2,Non-energy}=\left[CBR·C_{CBR}+COG·C_{COG}+BG·C_{BG}+{\displaystyle \sum _a}\left({PM}_a+C_a\right)-SOG·C_{SOG}\right]·{\displaystyle \frac{44}{12}}$$

(7)

CO₂ emissions from direct reduction iron production:

$$E_{{CO}_2,Non-energy}=\left({DRI}_{NG}·C_{NG}+{DRI}_{BZ}·C_{BZ}+{DRI}_{CK}·C_{CK}\right)·{\displaystyle \frac{44}{12}}$$

(8)

Among them, for iron and steel production:

E_{CO2, Non-energy} = CO₂ emissions to be reported in the IPPU department, in tons;

PC = The amount of coke consumed in iron and steel production (excluding slag production), in tons;

COB_a = Quantity of on-site coke oven by-product a consumed in blast furnaces, in tons;

CI = The amount of coke injected directly into the blast furnace, in tons;

L = The amount of limestone consumed in steel production, in tons;

D = The amount of dolomite consumed in iron and steel production, in tons;

CE = Number of carbon electrodes consumed in EAF, in tons;

O_b = The amount of other carbon aerosols and process materials b consumed in steel production, such as molten slag or waste plastic, in tons;

COG = The amount of coke oven gas consumed in blast furnaces in iron and steel production, in m^3;

S = The quantity of steel produced, in tons;

IP = The output of iron not converted into steel, in tons;

BG = The quantity of blast furnace gas transfer out of the field, in m³;

C_x = The carbon content of input or output material x, in tons C/(units of material x) [for example, tons C/ton].

For slag production:

E_{CO2, Non-energy} = CO₂ emissions to be reported in the IPPU department, in tons;

CBR = The quantity of coke powder purchased and produced on site for slag production, in tons;

COG = The amount of coke oven gas consumed in the blast furnace in slag production, in m³;

BG = The amount of blast furnace gas consumed in slag production, in m³;

PM_a = The quantity of other process materials a, that is, the amount of coke and slag production consumption in integrated coke production and steel production facilities, in tons;

SOG = The amount of slag smoke transferred from off-site to iron and steel production facilities or other facilities, in m³;

C_x = The carbon content of input or output material x, in tons C/(units of material x) [for example, tons C/ton].

For directly reducing iron production:

E_{CO2, Non-energy} = CO₂ emissions to be reported in the IPPU department, in tons;

DRI_NG = The amount of natural gas used in the production of direct reducing iron, in GJ;

DRI_BZ = The amount of coke powder used in the production of direct reducing iron, in GJ;

DRI_CK = The amount of metallurgical coke used in the production of direct reducing iron, in GJ;

C_NG = Carbon content of natural gas, in tons of C/GJ;

C_BZ = Carbon content of coke powder, in tons of C/GJ;

C_CK = Carbon content of metallurgical coke, in tons of C/GJ.

While the mass balance approach demands a higher level of data precision and technical expertise, its computational intricacies yield more accurate emissions estimates compared to the emission factor method.

3.

Actual measurement method

The actual measurement method refers to the estimation of CO₂ emissions by obtaining specific factory data through actual monitoring, offering higher accuracy than the mass balance method. When empirical CO₂ emission data from iron and steel production facilities are accessible, they can be aggregated to determine national emissions. Where such data are lacking, emissions can be derived from plant-specific activity data concerning individual reductants, exhaust gases, and other process materials and products, culminating in a national total that is the sum of facility-reported emissions.

The actual method, while labor-intensive and costly, and often challenged by the inaccessibility of certain data, offers greater accuracy in emission calculations compared to both the emission factor and mass balance approaches.

Table 5. Comparison of IPCC methods of accounting for carbon emissions.

Method	Method Characteristics	Data Needed to Calculate Emissions from Fossil Fuel Combustion	Data Required for Calculation of Industrial Process Emissions
Emission coefficient method	This approach is relatively simple, user-friendly, and imposes low demands on data quality, albeit with a higher degree of uncertainty in the calculated results.	(1) Fuel volume data. (2) Default emission factor.	(1) The output of steel in different types of steelmaking processes, the amount of pig iron not converted to steel, and the amount of directly reduced iron. (2) Default emission factor.
Mass balance method	While more complex and demanding in terms of data and technical expertise, this method yields comparatively more accurate computational outcomes.	(1) Fuel volume data. (2) Country-specific emission factors and various gas fuels.	(1) The amount of material used in the process of steel production, slag production, and directly reduced iron production. (2) Other process inputs and outputs. (3) Calorific value and carbon content of each material.
Actual measurement method	The method, though labor-intensive and costly, with many data points that are challenging to ascertain, offers a heightened degree of precision in its computational outcomes.	(1) The amount of fuel burned for each technology type. (2) Greenhouse gas emission factors are given by fuel and technology type.	(1) To obtain actual measured CO2 emission data in iron and steel making facilities. (2) Specific factory activity data of each process material and product, calorific value, and carbon content of each material in a specific plant.

contrasts these three accounting methodologies:

3.1.2. International Iron and Steel Association

The life cycle carbon emissions of iron and steel products of the International Iron and Steel Institute are calculated in units of each production process, which extends the sources of greenhouse gas emissions to include all material items in the steel production process on the basis of material flow [72]. This approach is systematically broken down into the following steps:

• Draw a boundary

The system boundary of the International Iron and Steel Association excludes the extraction and transportation of raw materials. Instead, it adopts a “gate-to-gate” approach, encompassing the processes from the entry of raw materials into the facility to the dispatch of finished products, as delineated in Figure 6.

2.

Calculation method

Within the framework of carbon balance, the total carbon emissions of the steel industry are calculated by summing the product of CO₂ emissions per ton of process-specific steel and the corresponding steel ratio factor for each process. This is succinctly represented by the following formula:

$$E_{total}={\sum }_{i=1}^n{E}_i\times p_i$$

(9)

In the form, E_total—Total carbon dioxide emissions per unit product, in tons;

E_i—The amount of carbon dioxide produced in the production process, in tons;

p_i—Steel ratio coefficient of the rolling process.

Carbon emissions within each process of the iron and steel industry are derived from the difference between the carbon content of input materials and energy and the carbon content of output products and by-products. These emissions are categorized into three components: direct emissions, indirect emissions, and carbon emissions deductions. Direct emissions originate from materials with inherent carbon content or gases composed directly of CO₂, encompassing fossil fuels, fluxes, raw materials, and external by-product gases. Indirect emissions arise from the production of materials or energy that do not contain carbon elements but consume carbon-containing materials or energy, such as electricity and steam, along with the upstream emissions of all purchased materials and energy, including blast furnace gas, converter gas, coke, and oxygen. Carbon emissions deductions account for by-products like blast furnace slag and converter slag. The calculation is formulated as follows:

$$E_{\mathrm{w}\mathrm{o}\mathrm{r}\mathrm{k}\mathrm{i}\mathrm{n}\mathrm{g} \mathrm{p}\mathrm{r}\mathrm{o}\mathrm{c}\mathrm{e}\mathrm{d}\mathrm{u}\mathrm{r}\mathrm{e}}={\sum }_i^n{C}_{in,i}\times {EF}_i-{\sum }_j^n{C}_{out, j}\times {EF}_j$$

(10)

In the form, E_{working procedure}—Carbon dioxide emissions from each process, in tons;

C_in,i—Input stream of carbon carrier in each process, in tons;

C_out,j—Output stream of carbon carrier in each process, in tons;

EF—CO₂ emission factor of carbon carrier materials, in t/t.

The life cycle carbon emission calculation method of iron and steel products of the International Iron and Steel Institute provides a comprehensive and standardized framework to evaluate the carbon footprint of iron and steel products. This approach encompasses the entire life cycle, offering a comprehensive view of carbon emissions and focusing on the carbon flow between processes at the per-ton steel product level, thereby highlighting the potential for emission reduction. However, this methodology demands extensive background and prospective data, entailing significant effort. Moreover, the data, often lacking uniform standards across enterprises, presents challenges of high data requirements, pronounced subjectivity, and substantial expertise demands.

3.1.3. ISO 14404 Calculation Method of Carbon Dioxide Emission Intensity in Iron and Steel Production

The ISO standard employs a life cycle approach to quantify the greenhouse gas emissions associated with iron and steel products, factoring in both the direct and indirect emissions of carbon dioxide, as well as the CO₂ inherent in the products themselves [73]. This methodology is bifurcated: the first segment pertains to steel mills utilizing the blast furnace process, while the second addresses those employing electric furnace technology. This is an internationally unified method for calculating carbon dioxide emission intensity, thereby enabling steel mills to standardize and benchmark CO₂ emissions across diverse production junctures. The accounting boundaries are graphically represented in Figure 7 and Figure 8.

The formula for calculating CO₂ emissions from iron and steel products is as follows:

$$E_{CO2, annual}=\sum _{t=1}^N{K}_{t,d,CO2}\times Q_{t,d,CO2}+\sum _{t=1}^N{K}_{t,u,CO2}\times Q_{t,u,CO2}+\sum _{t=1}^N{K}_{t,c,CO2}\times Q_{t,c,CO2}$$

(11)

In the form, $K_{t,d,CO2}$—Emission factors of direct carbon dioxide emission sources, tCO₂/unit;

$Q_{t,d,CO2}$—Number of direct carbon dioxide emission sources;

$K_{t,u,CO2}$—Emission factors of upstream carbon dioxide emission sources, tCO₂/unit;

$Q_{t,u,CO2}$—Number of upstream carbon dioxide emission sources;

$K_{t,c,CO2}$—Emission factors of carbon dioxide emission sources produced, tCO₂/unit;

$Q_{t,c,CO2}$—number of carbon dioxide emission sources produced.

As an international standard, ISO 14040 2013 provides a standardized method for calculating carbon dioxide emissions in iron and steel production, enhancing the consistency and comparability of global data. This framework, which encompasses the entire lifecycle of steel production, facilitates a holistic assessment of carbon emissions and pinpoints avenues for reduction. The standard underscores the importance of transparent and precise accounting principles, thereby enhancing the reliability of emissions reporting. Nonetheless, challenges persist, such as the demand for extensive data, operational complexity, and inherent uncertainties.

3.2. Regional Accounting Method

3.2.1. EU “Carbon Tariff”

In March 2022, the European Union adopted the CBAM, marking the world’s inaugural “carbon tariff” regime [74]. This pioneering mechanism mandates that importers procure carbon emission allowances at prices aligned with the EU’s carbon market for high-emission goods, such as iron and steel, with accounting based on CO₂ emissions attributable to fossil fuel combustion within the production process.

$$\mathrm{D}\mathrm{i}\mathrm{s}\mathrm{c}\mathrm{h}\mathrm{a}\mathrm{r}\mathrm{g}\mathrm{e} \mathrm{a}\mathrm{m}\mathrm{o}\mathrm{u}\mathrm{n}\mathrm{t}\left(\mathrm{t}{\mathrm{C}\mathrm{O}}_2\right)=\mathrm{Q}\mathrm{u}\mathrm{a}\mathrm{l}\mathrm{i}\mathrm{t}\mathrm{y}\times \mathrm{E}\mathrm{m}\mathrm{i}\mathrm{s}\mathrm{s}\mathrm{i}\mathrm{o}\mathrm{n} \mathrm{i}\mathrm{n}\mathrm{t}\mathrm{e}\mathrm{n}\mathrm{s}\mathrm{i}\mathrm{t}\mathrm{y}$$

(12)

In the form, Quality—Fossil fuel combustion, ton, megawatt-hour;

Emission intensity—Consumption of CO₂ emissions per unit of fossil fuels, tons of CO₂/products, tons of CO₂/megawatt-hours.

To ensure parity in carbon emission costs between imported and EU-produced goods, CBAM allows for two types of carbon emission deductions: (1) to prevent double protection for EU enterprises, importers’ taxable carbon emissions may be adjusted based on the amount of free emissions allocated to similar EU products, akin to a tax base adjustment; (2) to avoid double taxation on imported products, the carbon emission credits already paid by imported products in their countries of origin should be deducted from the importers’ taxable carbon emissions, akin to a tax deduction. This approach underscores the evolving nature of carbon accounting and its integration with international trade policies.

The implementation of CBAM has had a profound impact on China’s steel industry. As a major steel producer, China accounted for 57% of the world’s total steel output and approximately 60% of global steel-related carbon emissions in 2020 [75]. CBAM has increased the cost of Chinese steel exports to the EU, thereby weakening the international competitiveness of Chinese steel products [76,77]. However, this mechanism has also brought new opportunities for low-carbon transformation to China’s steel industry, driving the research and application of low-carbon innovative technologies, such as hydrogen-enriched carbon recycling blast furnaces and top-gas recycling oxygen blast furnaces [78]. Moreover, the development of electric arc furnace steelmaking, which has a significant low-carbon advantage compared with the traditional blast furnace–basic oxygen furnace steelmaking process, has also been promoted.

To cope with CBAM, China needs to adjust its domestic carbon emission accounting system to align with international standards. China has already released the “Action Plan for Improving the Carbon Emission Statistics and Accounting System” to accelerate the construction of a sound carbon emission statistics and accounting system. China should also strengthen the quality management of carbon emission data and improve the carbon emission data quality of enterprises in the industry. Through these measures, China’s steel industry can not only meet the challenges brought by CBAM, but also seize the opportunities to achieve green, low-carbon, and high-quality development [77].

3.2.2. Guidelines for Accounting Methods and Reporting of Greenhouse Gas Emissions from China

The guidelines for Accounting Methods and Reports of Greenhouse Gas Emissions from China’s Iron and Steel Production Enterprises, compiled by experts from the National Center for Strategic Research on Climate Change and International Cooperation and commissioned by the National Development and Reform Commission, draws on practical experience and research outcomes in greenhouse gas accounting reports from both domestic and international enterprises [79]. Grounded in input–output theory, this guideline calculates the macro-level greenhouse gas emissions of steel production processes, excluding material flow cycles within the system, with a focus on carbon dioxide emissions. This methodology demonstrates the calculation of greenhouse gas emissions within the iron and steel industry, highlighting the importance of international unity and comparability in carbon accounting standards and their alignment with global trade policies.

Figure 9 illustrates the boundaries of the iron and steel production system, which includes direct production systems, auxiliary production systems such as power supply and utilities, and ancillary production systems that directly support production operations, including management structures and on-site services like catering and healthcare facilities. The scope of greenhouse gas emissions accounting and reporting for iron and steel enterprises encompasses emissions from fuel combustion, industrial production processes, net purchased electricity, thermal energy emissions, and embodied emissions from carbon sequestration products. This comprehensive framework ensures a thorough assessment of the sector’s environmental footprint, aligning with the stringent standards necessary for global emissions reporting.

The total CO₂ emissions from the iron and steel production process are equivalent to the aggregate of all fossil fuel combustion emissions within the accounting boundary, emissions from industrial production processes, and CO₂ emissions associated with the net purchase of electricity and heat. It is imperative to also account for the implied emissions from carbon sequestration products, which should be deducted in the overall calculation.

$$E_{CO2}=E_{com}+E_{pro}+E_{ind}-R_h$$

(13)

In the form, $E_{CO2}$—Total carbon dioxide emissions from enterprises, in tons;

$E_{com}$—CO₂ emissions from all corporate net consumption of fossil fuel combustion activities, in tons;

$E_{pro}$—CO₂ emissions from industrial production processes of enterprises, in tons;

$E_{ind}$—CO₂ emissions from net purchase of electricity and heat by enterprises, in tons;

$R_h$—CO₂ emissions implied by enterprise carbon sequestration products, in tons.

The CO₂ emissions resulting from fuel combustion activities are aggregated emissions from various fuel combustion activities within the enterprise, with the calculation formula as follows:

$$E_{com}={\sum }_{i=1}^n{AD}_i\times {EF}_i$$

(14)

In the form, $A_{Di}$—Activity level of the second fuel, GJ;

${EF}_i$—The carbon dioxide emission factor of the second kind of fuel, in t/GJ.

The formula for calculating the activity level of the first kind of fuel, AD_i, is as follows:

$${AD}_i={NCV}_i\times {FC}_i$$

(15)

In the form, ${NCV}_i$—the average low calorific value of the first kind of fuel, for solid or liquid fuel, in GJ/t; versus gaseous fuel, in GJ/ten thousand Nm³;

${FC}_i$—The net consumption of the first type of fuel, in t for solid or liquid fuel, and ten thousand Nm³ for gaseous fuel.

The formula for calculating the carbon dioxide emission factor of the first kind of fuel EF_i is as follows:

$${EF}_i={CC}_i\times {OF}_i\times {\displaystyle \frac{44}{12}}$$

(16)

In the form, ${CC}_i$—carbon content per unit calorific value of the second kind of fuel, in t/GJ;

${OF}_i$—Carbon oxidation rate of the first kind of fuel, in %.

Greenhouse gas emissions from fuel combustion are calculated based solely on the net consumption of fossil fuels, with the exception of carbon sequestration associated with energy use. Consequently, the emissions from fossil fuel combustion detailed in the Guidelines for Accounting Methods and Reporting of Greenhouse Gas Emissions from China’s Iron and Steel Production Enterprises slightly exceed those calculated within the framework for the compilation of Provincial Greenhouse Gas Inventories.

Greenhouse gas emissions from the industrial production process encompass CO₂ emissions derived from the consumption of solvents such as dolomite and limestone, CO₂ emissions resulting from electrode consumption in electric arc furnaces and refining furnaces, and CO₂ emissions associated with the consumption of carbon-containing raw materials, including purchased pig iron.

$$E_{pro}={\sum }_{j=1}^n{P}_j\times {EF}_j+P_e\times {EF}_e+{\sum }_{k=1}^n{M}_k\times {EF}_k$$

(17)

In the form, $P_j$—Net consumption of the first type of solvent, in tons;

${EF}_j$—CO₂ emission factor of the second kind of solvent, in t/t;

$P_e$—The amount of electrode consumed by electric furnace steelmaking and refining furnace, in tons;

${EF}_e$—CO₂ Emission Factor of Electrodes consumed by Electric furnace steelmaking and Refining Furnace, in t/t;

$M_k$—Consumption of the first kind of carbon-containing raw materials, in tons;

${EF}_k$—CO₂ Emission Factor of the first carbon-containing Raw material, in t/t.

The formula for calculating CO₂ emissions implied by net purchase of electricity and heat is as follows:

$$E_{ind}={AD}_E\times {EF}_E+{AD}_H\times {EF}_H$$

(18)

In the form, ${AD}_E$—Net purchase of electricity, in MWh;

${EF}_E$—CO₂ emission factor of electric power, in t/MWh;

${AD}_H$—Net calorie purchase, in GJ.

${EF}_E$—CO₂ emission factor of heat (such as steam), in t/GJ.

Carbon sequestration products include exported gases (blast furnace gas, converter gas, coke oven gas), carbon by-products (coal tar, crude benzene, other coking by-products), and crude steel. The greenhouse gas emissions are calculated as follows:

$$R_h={\sum }_{h=1}^n{AD}_h\times {EF}_h$$

(19)

In the form, ${AD}_i$—Output of the first kind of carbon sequestration products, in tons;

${EF}_i$—CO₂ emission factor of the first carbon sequestration product, in t/t.

3.3. Comparative Analysis and Summary

3.3.1. Boundary Contrast

In carbon emission accounting, the delineation of system boundaries is crucial for ensuring the accuracy and comparability of accounting results [80,81,82]. Different accounting methods have adopted boundary settings ranging from narrow to comprehensive based on their purposes and scopes of application. The IPCC method primarily focuses on direct emissions, featuring a narrow boundary that facilitates unified accounting and comparison [83]. In contrast, the methods of the International Iron and Steel Institute and ISO employ life cycle analysis, with comprehensive boundaries that can more accurately reflect the carbon emissions of steel products throughout their life cycles [84]. Guidelines such as the “Provincial Greenhouse Gas Inventory Preparation Guide” and those related to Chinese steel production enterprises have established boundaries covering direct emissions, some indirect emissions, and the implicit emissions of carbon sequestration products, considering the actual situations of regions and enterprises to meet the accounting needs at different levels. Boundary selection needs to take into account the purpose, data availability, accounting feasibility, and consistency with international standards in a comprehensive manner [85]. For assessing the environmental footprint of steel products and exploring emission reduction potential, the boundary setting of life cycle analysis is more appropriate. However, for inventory preparation at regional and enterprise levels, boundaries should be flexibly defined according to specific circumstances to ensure the feasibility and accuracy of accounting.

The computational efficiency and scalability of different carbon emission accounting methods vary significantly in large-scale enterprises and cross-regional applications [86,87,88]. The IPCC accounting method, with its narrow boundary and relatively simple calculation process, is primarily suitable for national and regional greenhouse gas inventory compilation, demanding lower computational resources and fitting well for small to medium-scale data processing. However, its limited accounting scope may fail to comprehensively reflect the complex carbon emission situations in large-scale enterprises or cross-regional contexts. In contrast, the methods of the International Iron and Steel Institute and ISO, employing life cycle analysis, feature comprehensive accounting boundaries that can more accurately reflect the carbon emissions of steel products throughout their life cycles. These methods involve complex calculations and require higher computational resources, especially when dealing with large-scale data, necessitating stronger computational support [89]. Guidelines such as the “Provincial Greenhouse Gas Inventory Preparation Guide” and those related to Chinese steel production enterprises have established boundaries covering direct emissions, some indirect emissions, and the implicit emissions of carbon sequestration products, considering the actual situations of regions and enterprises to meet the accounting needs at different levels. But in cross-regional or large-scale enterprise applications, their scalability may be somewhat restricted. To enhance the computational efficiency and scalability of these methods in large-scale enterprises and cross-regional applications, it is recommended to assess their computational resource requirements, particularly in terms of data processing, storage, and computing capabilities [90,91]. Meanwhile, the use of parallel computing technology should be considered to accelerate large-scale data processing through distributed computing resources and efficient load balancing, as well as leveraging the dynamic scaling capabilities of cloud computing platforms to flexibly allocate computing resources according to demand, thereby optimizing computational efficiency and ensuring the feasibility and accuracy of the accounting process [92].

3.3.2. Data Collection and Uncertainty

In the practice of carbon accounting, data collection and accuracy management constitute the basis of the reliability of accounting results, and their technical complexity and management requirements vary significantly with different accounting systems. When the IPCC method adopts the basic framework of “activity level data × emission factors”, it often faces the dual challenges of fuzzy statistical boundaries of activity data and insufficient temporal and spatial representation of emission factors. Especially in developing countries, due to imperfect statistical systems, omission of activity data and insufficient localization of emission factors have become the main sources of systematic errors [93]. In comparison, the ISO 14044 standard [94] extends this challenge to the supply chain dimension through its life cycle data requirements [95]. Its multi-node data collection not only needs to address the accuracy of a company’s direct emission data but must also tackle the coordination challenges of upstream and downstream data, including practical operational obstacles such as inconsistent data formats and reporting frequencies. CBAM further raises the complexity of data collection to the transnational level. The accounting of implied carbon emissions from imported products involves the conversion of accounting methods in different countries, data sovereignty restrictions, and differences in monitoring, reporting, and verification (MRV) systems. Institutional obstacles produce data comparability and verifiability key bottlenecks. China’s carbon accounting system emphasizes the priority application of actual measurement methods and improves data accuracy through direct measurement methods such as the Continuous Emission Monitoring System (CEMS). However, the promotion of this method is limited by the representative verification of monitoring points and the traceability requirements of equipment calibration. High operation and maintenance costs, especially among small and medium-sized enterprises, create greater implementation resistance.

Data accuracy is not only affected by systematic errors in the collection stage, but also faces multiple uncertainties in subsequent processing. The selection of emission factors often introduces significant deviations. For example, the difference between regional power grid factors and national average factors may cause accounting results to deviate from actual emission levels [96]. Measurement errors include instrument accuracy errors and sampling errors. Especially in discontinuous monitoring scenarios, the choice of data interpolation method directly affects the credibility of the final accounting results. In addition, the problem of data update lag is common, especially after process improvement or energy structure optimization. If historical emission factors or activity data are still used, the accounting results will not be able to reflect the true emission reduction effect. Inconsistencies in statistical calibers, such as differences in the definition of boundaries of scope 1, 2, and 3, further aggravate the problem of data comparability and challenge the reliability of cross-enterprise or cross-industry Benchmarking analysis (Benchmarking).

In order to improve data quality, a comprehensive control system covering technical verification and system management can be built. In the data collection stage, a standardized responsibility matrix (RACI model) is established to clarify the rights and responsibilities of data owners, collectors, and auditors, and differentiated collection frequencies (such as real-time, day, month, and year) are formulated based on data criticality [97]. At the level of technical verification, Monte Carlo simulation can be used to quantify data uncertainty, statistical methods (such as the T test) can be used to cross-validate multi-source data, and potential systematic deviations can be identified by combining the comparative analysis of the material balance method and the emission factor method [98]. In terms of institutional guarantees, establish a data quality scoring system (DQ scoring) and a third-party audit mechanism, and explore the application of blockchain and other technologies in data traceability and certificate storage [99]. In addition, dynamically updating the emission factor database, implementing data quality KPI assessment, and industry-level benchmarking management can form a closed loop of continuous improvement, ultimately ensuring the integrity, accuracy, and verifiability of carbon accounting data and meeting increasingly stringent compliance requirements.

3.3.3. Summary

This chapter provides a systematic analysis of the characteristics and applications of the principal carbon emission accounting methodologies within the iron and steel industry. The IPCC method establishes the foundational framework for compiling international greenhouse gas inventories, yet it lacks the granularity of specific processes. The International Iron and Steel Institute’s method, based on the carbon balance principle, enables comprehensive accounting across the entire process. However, this method’s capacity to account for upstream emissions is somewhat constrained, necessitating advancements in measurement technologies and the refinement of emission factor databases to enhance the inclusivity of carbon accounting practices. From a lifecycle perspective, the ISO 14404 standard establishes detailed protocols for both long and short iron and steel production processes, enhancing the comparability of accounting results. With the advent of policies such as CBAM, the international harmonization and comparability of accounting have become increasingly critical. Moreover, regional standards have been continuously refined in practice, refining accounting boundaries and offering more localized options to support more informed decision-making in emission reduction strategies.

The divergence in carbon emission accounting methodologies is evident in their release timelines, objectives, applicability, calculation principles, accounting techniques, boundaries, and selection of emission factors, as detailed in

Table 6. Summary of accounting methods for carbon emissions in the steel industry.

Serial Number	Accounting Method	Release Time	Main Purpose	Scope of Application	Compute Principle	Accounting Type	Accounting Boundary	Let out Divisor
1	IPCC accounting methods 1	2006	Report on national greenhouse gas emissions and removals.	Guidelines for national and regional greenhouse gas inventories.	Input–output	Emission factor method	(1) Emissions from fossil fuel combustion. (2) Emissions from steel production.	Default value
2	International iron and steel association	2016	Investigate the carbon footprint throughout the steel life cycle and unearth the potential for carbon reduction.	Accounting and reporting of gas emissions of iron and steel enterprises.	Life cycle	Carbon balance method	Carbon emissions from all production processes: (1) Direct discharge. (2) Indirect emissions (consider upstream emissions). (3) Carbon credits.	Default value
3	International Organization for Standards (ISO)	2013	Investigate the carbon footprint throughout the steel life cycle and unearth the potential for carbon reduction.	Accounting and reporting of gas emissions of iron and steel enterprises.	Life cycle	Emission factor method	Carbon emissions from all production processes: (1) Direct discharge. (2) Upstream emissions. (3) Carbon credits.	Default value
4	Guidelines for the Preparation of Provincial Greenhouse Gas Inventories	2011	Preparation of provincial greenhouse gas inventories.	Accounting and reporting of gas emissions of iron and steel enterprises in each province.	Input–output	Emission factor method	(1) Fossil fuel burning (only purchased fuels are considered). (2) Emissions from the industrial production process: high temperature decomposition of ironmaking solvents (limestone and dolomite, etc.) and carbon reduction process of steelmaking (purchased pig iron). (3) Indirect emissions from purchased electricity.	Default value or measured value
5	Guidelines for Accounting Methods and Reporting on Greenhouse Gas Emissions of Chinese Steel Production Enterprises	2013	Establish the enterprise greenhouse gas emission reporting system, improve the steel industry greenhouse gas emission statistical accounting system, and other related work references.	Accounting and reporting of greenhouse gas emissions of Chinese steel production enterprises.	Input–output	Emission factor method	(1) Fossil fuel burning (net fossil fuel use, excluding carbon sequestration for energy use). (2) Industrial production process emissions (solvent consumption, electrode consumption, purchased pig iron, and other carbon-containing raw materials consumption). (3) The net purchase of electricity and heat. (4) Implicit emissions of carbon sequestration products (sold gas, carbon by-products, crude steel, methanol, etc.)	Default value or measured value
6	Requirements of the greenhouse gas emission accounting and reporting—Part 5: Iron and steel production enterprise	2015	It can be used as a reference for the development of carbon emission trading, the establishment of an enterprise greenhouse gas emission reporting system, and the improvement of the greenhouse gas emission statistical accounting system in the steel industry.	Accounting and reporting greenhouse gas emissions of Chinese steel production enterprises.	Input–output	Emission factor method	(1) Fossil fuel burning (all fossil fuel burning within accounting boundaries). (2) Industrial production process (solvent consumption, electrode consumption, purchased pig iron, and other carbon-containing raw materials consumption). (3) Purchase electricity and heat (4) Implicit emissions of carbon sequestration products (such as crude steel, methanol, etc.). (5) The power and heat output of the enterprise.	Default value or measured value
7	Tianjin accounting method	2013	For the use of carbon emission trading and carbon verification in the steel industry in Tianjin City.	Accounting and reporting of carbon emissions of steel production enterprises in Tianjin.	Input–output	Emission factor method	(1) Fossil fuel combustion. (2) Discharge in the production process (carbonate decomposition in flux such as limestone and dolomite; change of carbon content in steelmaking; lime production process discharge; emissions from carbonate desulfurization processes, etc.) (3) The net purchase of electricity and heat.	Default value or measured value
8	Shanghai accounting method	2012	For the use of carbon emission trading and carbon verification in the steel industry in Shanghai.	Carbon dioxide emission accounting and reporting for major emitters in the Steel Industry in Shanghai.	Input–output	Emission factor method (in which some process emissions use the balance method)	(1) Fossil fuel combustion. (2) Production process emissions (carbonate decomposition in limestone, dolomite and other fluxes, and changes in the carbon content of iron- and steelmaking) (3) The net purchase of electricity and heat.	Default value or measured value

. International methodologies prioritize the versatility and comparability of approaches, whereas regional standards focus on practical implementation and local adaptability. Notably, the evolution from the IPCC method to regional standards illustrates the ongoing refinement and enhancement of carbon emission accounting systems. This methodological diversity informs approach selection for carbon emission accounting at various scales and for diverse objectives.

4. Conclusions and Outlook

This study examines carbon emission accounting methodologies in the iron and steel industry, focusing on foundational techniques such as “bottom-up” LCA and “top-down” IOA. It evaluates industry-specific methods, including the IPCC method, ISO 14044, and regional standards like the EU’s CBAM and Chinese norms. The study highlights differences in accounting boundaries, principles, and data requirements, and outlines emerging trends for sustainable development. Anticipating future trajectories, the study outlines the emerging trends in carbon emission accounting for the iron and steel sector:

(1) With the rapid development of technologies such as the Internet of Things (IoT), big data, and artificial intelligence (AI), carbon emission accounting methods are innovating in the direction of real-time, precision, and customization [100]. By deploying sensor networks in production facilities, enterprises can monitor energy consumption, production parameters, and emission levels in real time, and build a dynamic database [48]. Combined with big data analysis technology, these massive datasets can be efficiently integrated and processed to identify abnormal emission patterns, predict equipment maintenance needs, and optimize production processes, thereby significantly improving the timeliness and transparency of carbon accounting [6]. In this process, artificial intelligence shows multi-dimensional core values. Machine learning algorithms (such as neural networks and decision trees) can automatically integrate multi-source data (including energy consumption records, production activity data, etc.), and quickly extract key information through natural language processing (NLP) and time series analysis techniques, reducing traditional accounting manual errors in [101]. At the level of prediction and optimization, AI further deepens the intelligence of carbon emission management. Based on models such as BP neural network and long short-term memory network (LSTM), the system can capture the nonlinear characteristics of carbon emissions and accurately predict emission trends in different scenarios. At the same time, by analyzing enterprise-specific factors (such as production scale, technical capabilities, and product structure), AI can generate customized emission reduction strategies and simulate the cost-effectiveness of different measures.

(2) The ongoing standardization of carbon accounting practices is gaining momentum globally [102]. With the expansion of the global carbon market, the iron and steel sector’s carbon accounting methods are expected to align more closely with international standards, thereby enhancing the global consistency and comparability of carbon accounting outcomes. The application of advanced deep learning algorithms, such as Long Short-Term Memory networks (LSTM) and Generative Adversarial Networks (GAN), will facilitate the identification of emission patterns from extensive datasets and predict emission trends [103]. Compliance with internationally recognized frameworks, such as ISO 14064 [104], enables companies to ensure the precision and transparency of their carbon emissions data, enhancing their engagement in global carbon markets and bolstering their international competitiveness. Moreover, as carbon footprint and LCA become increasingly critical in supply chain management, the steel industry may increasingly employ these approaches to evaluate and mitigate the carbon impact of their products [105]. Additionally, as renewable energy technologies advance and costs decrease, the steel industry may shift more towards cleaner energy sources, which will directly influence carbon accounting methods and outcomes. Policy incentives like carbon taxes and trading systems will also encourage the adoption of more sophisticated accounting methods to meet stringent environmental regulations and market demands [106].

(3) Carbon emission accounting methodologies are increasingly converging toward integration. Future research in carbon emission accounting will emphasize the synergistic application of multiple methodologies, particularly the integration of LCA and IOA [107]. This integration necessitates the establishment of standardized protocols for data collection and processing, as well as the development of comprehensive models that can manage both environmental and economic datasets [108]. Artificial intelligence and machine learning technologies will enable the analysis of complex data, develop decision support systems, and provide real-time carbon emission data and reduction strategies for enterprises. These systems will also evaluate the impact of supply chain risks on carbon emissions and develop corresponding risk management strategies [109].

(4) The scope of carbon emission accounting is set to expand, encompassing not only carbon dioxide but also a range of other greenhouse gases such as methane and nitrous oxide. This expansion necessitates a technical upgrade of current methods, including the development of new measurement technologies, the enhancement of emission factor databases, and the updating of life cycle assessment models [110,111]. Additionally, it requires the regionalization of emission factors and the refinement of relevant policies and standards, thereby providing a more comprehensive foundation for decision-making in emission reduction [112].

The expansion of accounting content within the iron and steel industry’s carbon emission accounting methods will promote more precise, comprehensive, and standardized practices, providing robust support for achieving carbon neutrality goals. Concurrently, the continuous methodological innovations of these methods will enhance the competitive edge of China’s iron and steel industry in the global market, driving the sector’s transformation towards a low-carbon and green development trajectory, and contributing to the achievement of global sustainable development goals.