Equitable Allocation of Interprovincial Industrial Carbon Footprints in China Based on Economic and Energy Flow Principles

Abstract

The equitable allocation of carbon emission responsibility is fundamental to advancing China’s industrial decarbonization, achieving its dual-carbon goals, and realizing regional sustainable development. However, prevailing interprovincial carbon accounting frameworks often neglect the coupled dynamics of economic benefits, energy flows, and ecological capacity, leading to systematic misattribution of industrial carbon footprint transfers. Here, we develop an integrated analytical framework combining multi-regional input–output (MRIO) modeling and net primary productivity (NPP) assessment to comprehensively quantify industrial carbon footprints and their transfers across 30 Chinese provinces. By embedding both the benefit principle (aligning responsibility with trade-generated economic gains) and the energy flow principle (accounting for interprovincial energy trade), we construct a dual-adjustment mechanism that rectifies spatial and sectoral imbalances in traditional accounting. Our results reveal pronounced east-to-west industrial carbon footprint transfers, with resource-rich provinces (e.g., Inner Mongolia, Xinjiang) disproportionately burdened by external consumption, impacting the balance of sustainable development in these regions. Implementing benefit and energy flow adjustments redistributes responsibility more fairly: high-benefit, energy-importing provinces (e.g., Shanghai, Jiangsu, Beijing) assume greater carbon obligations, while energy-exporting, resource-dependent regions see reduced responsibilities. This approach narrows the gap between production- and consumption-based accounting, offering a scientifically robust, policy-relevant pathway to balance regional development and environmental accountability. The proposed framework provides actionable insights for designing carbon compensation mechanisms and formulating equitable decarbonization policies in China and other economies facing similar regional disparities.

1. Introduction

Under the global challenge of climate warming, the conflict between industrial production and consumption and ecological sustainability has intensified. As the primary source of carbon emissions [1], industrial activities—through fossil fuel consumption and the depletion of natural carbon sinks such as forests and grasslands—severely undermine the carbon cycle function of ecosystems, thereby exacerbating climate change and ecological crises [2]. In response, the Chinese government has emphasized the need to “coordinate efforts to reduce carbon emissions, cut pollution, expand green development, and promote growth, precisely control fossil fuel consumption, improve ecological and environmental quality, and enhance carbon sink capacity to break the trade-off between environmental governance and economic development” [3]. However, interprovincial transfers of industrial carbon emissions exacerbate regional ecological imbalances. Energy flows, driven by supply-demand relationships in industrial production and consumption, constitute critical pathways for carbon footprint transfers. Neglecting these flows can result in a mismatch between “economic benefits” and “environmental costs” for energy-exporting and importing regions, making it difficult to impose effective constraints on carbon emissions at the source. In this context, the ecological carbon footprint is adopted as a key indicator for measuring carbon emissions. By integrating trade benefits and energy flows, this study scientifically defines and rationally optimizes the principles for allocating responsibilities for transferring provincial industrial carbon footprints in China, thereby providing an important impetus for regional resource conservation and the sustainable development of eco-friendly green industries [4].

Research on carbon footprints primarily focuses on two aspects: measurement methods and the allocation of carbon footprint transfers responsibility. Major measurement approaches include the IPCC method [5], energy conversion [6], input–output models [7], and net primary productivity (NPP) models [8]. Among these, the multi-regional input–output (MRIO) model enables simultaneous analysis of carbon emissions across multiple provinces, accurately identifying transfer paths in high-energy-consuming industries and facilitating comparisons of regional sustainable development states [7]. The NPP model, on the other hand, links carbon emissions with ecosystems, providing an ecological perspective on regional pollution differences [8].

In the allocation of carbon footprint transfer responsibility, three main principles prevail: producer responsibility, consumer responsibility, and shared responsibility. The producer responsibility principle operates on “whoever produces is responsible,” assigning producers the responsibility for the entire lifecycle emissions of products and services [9]. Conversely, the consumer responsibility principle asserts “whoever consumes is responsible” [10], arguing that regional consumption drives production and therefore should bear the corresponding carbon emission responsibility. Dual-side responsibility research reveals that interprovincial trade in China leads to higher consumption-side carbon footprints in eastern and southern provinces compared to western regions, with carbon footprint transfers flowing spatially from east to central and western areas [11,12,13]. However, assigning transfer responsibility solely to producers or consumers has raised fairness concerns [14]. Consequently, many studies adopt the shared responsibility principle, focusing on the scientific design of allocation factors such as trade value-added [14], equal distribution [15], and bilateral trade carbon pricing [16]. Nevertheless, most existing allocation models remain outcome-based at the national or provincial level, and often overlook the unique characteristics of industrial sectors. Owing to China’s dual challenges of uneven energy supply and indispensable demand, energy flows directly induce carbon footprint transfers from energy-exporting regions to energy-importing ones via industrial production–consumption pathways, exacerbating imbalances in regional resource and environmental pressures [17,18]. In particular, several northwestern provinces have seen their ecological carrying capacity turn negative due to coal exports [19].

Most studies focus on industrial carbon footprint transfers at the national or provincial level, lacking integrated, dual-dimensional (provincial and industrial) collaborative analyses that are essential for revealing cross-provincial transfer paths among various industrial sectors. Furthermore, the design of allocation factors for industrial carbon footprint transfers often fails to fully account for the intrinsic linkages between industrial activities and energy flows, introducing systemic biases in the assessment of actual carbon emission responsibilities and diminishing the accuracy of shared responsibility frameworks.

Therefore, this study measures the characteristics of production-side and consumption-side industrial carbon footprint transfers from both provincial and sectoral perspectives and analyzes the transfer pathways of key industries. This will extend the effectiveness of traditional accounting systems to new dimensions: “ecological carbon sink pressures” and “industrial linkage mechanisms.” Furthermore, it develops a dual-principle carbon responsibility allocation framework based on “benefits and energy flows,” designing a benefit allocation factor derived from trade value added (reflecting the principle that “greater benefit entails greater responsibility”) and an allocation coefficient based on energy flow rates (capturing the principle that “greater energy use entails heavier responsibility”). Together, these form a two-dimensional mechanism that balances economic interests with environmental costs.

On the basis of identifying transfer characteristics, the study further couples the MRIO–NPP model to quantify the ecological impacts of carbon footprint transfer. In addition, it highlights the potential changes to transfer pathways arising from post-2017 renewable energy transitions. Finally, it proposes a set of policy recommendations—namely “differentiated dual-side management,” “ecological carbon compensation mechanisms,” “regional collaborative governance,” “shared benefit–cost responsibility,” and “dynamic expansion of the Carbon Responsibility Allocation Framework”—to form a complete cycle from problem identification to policy implementation. These contributions provide both theoretical foundations and practical pathways for the equitable allocation of provincial industrial carbon footprint responsibilities in China, thereby supporting the implementation of national sustainable development strategies under the “dual carbon” goals.

2. Data and Methodology

2.1. Allocation of Carbon Footprint Transfer Responsibility

The producer responsibility principle and consumer responsibility principle are the fundamental criteria for measuring industrial carbon footprint transfers. Clearly defining inter-provincial industrial carbon footprint transfer responsibilities is essential for implementing differentiated carbon reduction governance across regions, thereby forming an equitable and collaborative shared responsibility mechanism. Under these principles, responsibility for carbon emissions generated throughout the lifecycle of traded products—from raw material extraction to final consumption—is assigned either to industrial producers or consumers. The difference between these two accounting methods constitutes the implicit carbon footprint transfer in industrial trade, representing the gap between production-side and consumption-side industrial carbon footprints [20].

A province is classified as a net carbon footprint inflow region if it undertakes more production tasks due to external consumption demand while maintaining lower domestic demand for products from other regions. Conversely, it is considered a net outflow region if its own consumption drives production elsewhere. Trade benefits serve as the core link between producers and consumers. Allocating carbon responsibility based on the proportion of trade value-added reflects the principle of aligning economic gains with emission reduction responsibilities [14].

In addition, industrial carbon emissions are fundamentally driven by energy consumption. China’s energy flows follow the “West-to-East Power Transmission, North-to-South Coal Transport” pattern. Energy extraction directly damages ecological service functions, while transportation losses and industrial chain extensions further increase the environmental burden on exporting regions [19], resulting in polarized carbon footprints across provinces. According to environmental equity theory, while both producers and consumers share energy usage rights, their responsibility burdens differ significantly. The direct consumption coefficient within the MRIO model quantifies carbon footprint transfers induced by energy flows. Specifically, the direct consumption coefficient of coal mining and other extractive industries reflects the direct energy input required per unit of output, indicating carbon footprint linkages. Moreover, the intensity of industrial linkages—where stronger dependencies and connections exist—correlates with larger-scale carbon footprint transfers.

As shown in Figure 1, by integrating the benefit principle and the energy flow principle, this study constructs an industrial carbon footprint transfer responsibility allocation framework, clarifying the “economic benefit–environmental cost” matching mechanism to achieve inter-provincial shared responsibility. This framework provides systematic theoretical support for the coordinated advancement of industrial green transformation and ecological governance.

2.2. Provincial Carbon Responsibility Accounting Framework

2.2.1. Accounting Framework Based on Benefit–Cost Sharing

The MRIO model estimates inter-regional trade coefficients to construct a multi-regional input-output table [21], linking regional industrial data and clarifying inter-provincial flows of goods and services [22]. Provincial carbon emissions are calculated as Equation (1).

$$\begin{array}{c}\left[\begin{array}{cccc}{\mathit{C}}^{\mathbf{11}}& {\mathit{C}}^{\mathbf{12}}& \dots & {\mathit{C}}^{\mathbf{1}\mathit{n}}\\ {\mathit{C}}^{\mathbf{21}}& {\mathit{C}}^{\mathbf{22}}& \dots & {\mathit{C}}^{\mathbf{2}\mathit{n}}\\ ⋮& ⋮& ⋮& ⋮\\ {\mathit{C}}^{\mathit{n}\mathbf{1}}& {\mathit{C}}^{\mathit{n}\mathbf{2}}& \dots & {\mathit{C}}^{\mathit{n}\mathit{n}}\end{array}\right]=\left[\begin{array}{cccc}{\widehat{\mathit{F}}}^{\mathbf{1}}& \mathbf{0}& \dots & \mathbf{0}\\ \mathbf{0}& {\widehat{\mathit{F}}}^{\mathbf{2}}& \dots & \mathbf{0}\\ ⋮& ⋮& ⋮& ⋮\\ \mathbf{0}& \mathbf{0}& \dots & {\widehat{\mathit{F}}}^{\mathit{n}}\end{array}\right]\times \left[\begin{array}{cccc}{\mathit{B}}^{\mathbf{11}}& {\mathit{B}}^{\mathbf{12}}& \dots & {\mathit{B}}^{\mathbf{1}\mathit{n}}\\ {\mathit{B}}^{\mathbf{21}}& {\mathit{B}}^{\mathbf{22}}& \dots & {\mathit{B}}^{\mathbf{2}\mathit{n}}\\ ⋮& ⋮& ⋮& ⋮\\ {\mathit{B}}^{\mathit{n}\mathbf{1}}& {\mathit{B}}^{\mathit{n}\mathbf{2}}& \dots & {\mathit{B}}^{\mathit{n}\mathit{n}}\end{array}\right]\times \\ \left\{\left[\begin{array}{cccc}{\mathit{Y}}^{\mathbf{11}}& {\mathit{Y}}^{\mathbf{12}}& \dots & {\mathit{Y}}^{\mathbf{1}\mathit{n}}\\ {\mathit{Y}}^{\mathbf{21}}& {\mathit{Y}}^{\mathbf{22}}& \dots & {\mathit{Y}}^{\mathbf{2}\mathit{n}}\\ ⋮& ⋮& ⋮& ⋮\\ {\mathit{Y}}^{\mathit{n}\mathbf{1}}& {\mathit{Y}}^{\mathit{n}\mathbf{2}}& \dots & {\mathit{Y}}^{\mathit{n}\mathit{n}}\end{array}\right]+\left[\begin{array}{cccc}\mathit{E}{\mathit{X}}^{\mathbf{1}}& \mathbf{0}& \dots & \mathbf{0}\\ \mathbf{0}& \mathit{E}{\mathit{X}}^{\mathbf{2}}& \dots & \mathbf{0}\\ ⋮& ⋮& ⋮& ⋮\\ \mathbf{0}& \mathbf{0}& \dots & \mathit{E}{\mathit{X}}^{\mathit{n}}\end{array}\right]+\left[\begin{array}{cccc}{\mathit{\epsilon }}^{\mathbf{1}}& \mathbf{0}& \dots & \mathbf{0}\\ \mathbf{0}& {\mathit{\epsilon }}^{\mathbf{2}}& \dots & \mathbf{0}\\ ⋮& ⋮& ⋮& ⋮\\ \mathbf{0}& \mathbf{0}& \dots & {\mathit{\epsilon }}^{\mathit{n}}\end{array}\right]\right\}\end{array}$$

(1)

In the equation, $\mathit{C}$ represents the carbon emission matrix, $\widehat{\mathit{F}}$ denotes the provincial carbon emission coefficient matrix, $\mathit{Y}$ is the final consumption matrix, where ${\mathit{Y}}^{\mathit{i}\mathit{j}}$ indicates the quantity of final goods flowing from province $i$ to province $j$, $\mathit{B}={\left(\mathit{I}-\mathit{A}\right)}^{-\mathbf{1}}$ stands for the total input coefficient matrix, $\mathit{E}\mathit{X}$ represents the export matrix by sector and province, $\mathit{\epsilon }$ denotes the error-term matrix. This formula represents a matrix-within-matrix operation, in which the elements of the nested matrix ${\mathit{C}}^{\mathit{i}\mathit{j}}, {\widehat{\mathit{F}}}^{\mathit{n}},{\mathit{B}}^{\mathit{i}\mathit{j}},{\mathit{Y}}^{\mathit{i}\mathit{j}}, \mathit{E}{\mathit{X}}^{\mathit{n}}$ and ${\mathit{\epsilon }}^{\mathit{n}}$ are both expressed as matrices. In the construction of MRIO tables, the theoretical identity (total output = intermediate use + final use) often does not hold, primarily due to data heterogeneity and unobserved economic activities (e.g., regional informal trade, gray cross-border markets, household self-sufficiency). Although such activities are excluded from conventional statistical systems, they nevertheless participate in interregional economic circulation. To reconcile this inconsistency, an error term is introduced as a balancing item. Its core functions are to quantify data gaps and statistical errors, proxy unobserved economic activities, and ensure that the MRIO satisfies basic accounting constraints, thereby providing the foundation for analyzing industrial linkages and carbon transfers across regions. This mechanism is especially critical for provincial-level carbon accounting, where data constraints are more severe (e.g., lack of interprovincial trade data and greater influence of unobserved activities). By adjusting for unobserved emissions, the error term improves accounting accuracy and supports cross-regional mitigation policy design. However, data on unobserved economic activities cannot be separated from the error term, resulting in the actual contribution of unobserved economic activities to carbon emissions being broadly masked by the error term. This failure to accurately isolate their independent impact on carbon footprint transfer volumes thereby exacerbates data uncertainty and resulting biases in carbon responsibility allocation.

Furthermore, the reliability of the MRIO model depends on the accuracy of the complete consumption coefficient matrix and the carbon emission coefficient matrix; however, its practical application often introduces biases due to simplified parameterization or data processing. Firstly, the adoption of the assumption that carbon emission coefficients for the same industry are uniform across provinces may underestimate the carbon emission intensity of energy-intensive provinces. Secondly, the insufficient inclusion of implicit carbon emissions from industrial chain extension tends to underestimate the actual environmental costs of energy-exporting regions. Consequently, the subjectivity in MRIO data processing and error-term handling, along with data heterogeneity, may lead to overestimation or underestimation of carbon footprint transfer volumes in key industries, resulting in overestimated responsibilities for resource-rich provinces and underestimated responsibilities for economically developed provinces, ultimately compromising the fairness and scientific validity of inter-provincial carbon responsibility allocation. To address this, it is necessary to reduce uncertainty and enhance the reliability of responsibility allocation through multi-source data fusion and optimization of error term allocation methods. Based on the production responsibility principle, the production-side carbon emissions for province $i$ are calculated as Equation (2).

$$\begin{array}{c}{\mathit{C}}_{\mathit{M}\mathit{P}}^{\mathit{i}}={\displaystyle \sum _{\mathit{j}=\mathbf{1}}^{\mathit{n}}}{\mathit{C}}^{\mathit{i}\mathit{j}}={\widehat{\mathit{F}}}^{\mathit{i}}{\displaystyle \sum _{\mathit{j}=\mathbf{1}}^{\mathit{n}}}{\mathit{B}}^{\mathit{i}\mathit{j}}{\mathit{Y}}^{\mathit{j}\mathit{i}}+{\widehat{\mathit{F}}}^{\mathit{i}}{\mathit{B}}^{\mathit{i}\mathit{i}}\mathit{E}{\mathit{X}}^{\mathit{i}}+{\widehat{\mathit{F}}}^{\mathit{i}}{\mathit{B}}^{\mathit{i}\mathit{i}}{\mathit{\epsilon }}^{\mathit{i}}\\ +{\widehat{\mathit{F}}}^{\mathit{i}}{\displaystyle \sum _{\mathit{j}\ne \mathit{i}}^{\mathit{n}}}{\displaystyle \sum _{\mathit{s}=\mathbf{1}}^{\mathit{n}}}{\mathit{B}}^{\mathit{i}\mathit{s}}{\mathit{Y}}^{\mathit{s}\mathit{j}}+{\widehat{\mathit{F}}}^{\mathit{i}}{\displaystyle \sum _{\mathit{j}\ne \mathit{i}}^{\mathit{n}}}{\mathit{B}}^{\mathit{i}\mathit{j}}\mathit{E}{\mathit{X}}^{\mathit{j}}+{\widehat{\mathit{F}}}^{\mathit{i}}{\displaystyle \sum _{\mathit{j}\ne \mathit{i}}^{\mathit{n}}}{\mathit{B}}^{\mathit{i}\mathit{j}}{\mathit{\epsilon }}^{\mathit{j}}\end{array}$$

(2)

Based on the consumption responsibility principle, the consumption-side carbon emissions for province $i$ are calculated as Equation (3).

$$\begin{array}{c}{\mathit{C}}_{\mathit{M}\mathit{Q}}^{\mathit{i}}={\displaystyle \sum _{\mathit{j}=\mathbf{1}}^{\mathit{n}}}{\mathit{C}}^{\mathit{j}\mathit{i}}={\widehat{\mathit{F}}}^{\mathit{i}}{\displaystyle \sum _{\mathit{j}=\mathbf{1}}^{\mathit{n}}}{\mathit{B}}^{\mathit{i}\mathit{j}}{\mathit{Y}}^{\mathit{j}\mathit{i}}+{\widehat{\mathit{F}}}^{\mathit{i}}{\mathit{B}}^{\mathit{i}\mathit{i}}\mathit{E}{\mathit{X}}^{\mathit{i}}+{\widehat{\mathit{F}}}^{\mathit{i}}{\mathit{B}}^{\mathit{i}\mathit{i}}{\mathit{\epsilon }}^{\mathit{i}}\\ +{\widehat{\mathit{F}}}^{\mathit{j}}{\displaystyle \sum _{\mathit{j}\ne \mathit{i}}^{\mathit{n}}}{\displaystyle \sum _{\mathit{s}=\mathbf{1}}^{\mathit{n}}}{\mathit{B}}^{\mathit{j}\mathit{s}}{\mathit{Y}}^{\mathit{s}\mathit{i}}+{\widehat{\mathit{F}}}^{\mathit{j}}{\displaystyle \sum _{\mathit{j}\ne \mathit{i}}^{\mathit{n}}}{\mathit{B}}^{\mathit{j}\mathit{i}}\mathit{E}{\mathit{X}}^{\mathit{i}}+{\widehat{\mathit{F}}}^{\mathit{j}}{\displaystyle \sum _{\mathit{j}\ne \mathit{i}}^{\mathit{n}}}{\mathit{B}}^{\mathit{j}\mathit{i}}{\mathit{\epsilon }}^{\mathit{i}}\end{array}$$

(3)

In the equation, ${\mathit{C}}_{\mathit{M}\mathit{P}}^{\mathit{i}}$ denotes the production-side carbon emission matrix of province $i$, which is used to quantify the carbon emissions generated by all production activities within province $i$ (encompassing various industrial sectors and production links) during the production process; and ${\mathit{C}}_{\mathit{M}\mathit{Q}}^{\mathit{i}}$ denotes the consumption-side carbon emission matrix of province $i$, which is used to quantify the carbon emissions induced by final consumption activities within province $i$ (including household consumption, government consumption, capital formation, etc.). Where ${\widehat{\mathit{F}}}^{\mathit{i}}{\sum }_{\mathit{j}=\mathbf{1}}^{\mathit{n}}{\mathit{B}}^{\mathit{i}\mathit{j}}{\mathit{Y}}^{\mathit{j}\mathit{i}}+{\widehat{\mathit{F}}}^{\mathit{i}}{\mathit{B}}^{\mathit{i}\mathit{i}}\mathit{E}{\mathit{X}}^{\mathit{i}}+{\widehat{\mathit{F}}}^{\mathit{i}}{\mathit{B}}^{\mathit{i}\mathit{i}}{\mathit{\epsilon }}^{\mathit{i}}$ represents the carbon emissions of province $i$ caused by final product consumption, exports, and the error term. ${\widehat{\mathit{F}}}^{\mathit{i}}{\sum }_{\mathit{j}\ne \mathit{i}}^{\mathit{n}}{\sum }_{\mathit{s}=\mathbf{1}}^{\mathit{n}}{\mathit{B}}^{\mathit{i}\mathit{s}}{\mathit{Y}}^{\mathit{s}\mathit{j}}+{\widehat{\mathit{F}}}^{\mathit{i}}{\sum }_{\mathit{j}\ne \mathit{i}}^{\mathit{n}}{\mathit{B}}^{\mathit{i}\mathit{j}}\mathit{E}{\mathit{X}}^{\mathit{j}}+{\widehat{\mathit{F}}}^{\mathit{i}}{\sum }_{\mathit{j}\ne \mathit{i}}^{\mathit{n}}{\mathit{B}}^{\mathit{i}\mathit{j}}{\mathit{\epsilon }}^{\mathit{j}}$ represents the carbon emissions of province $i$ caused by final product consumption, exports, and the error term in other provinces. ${\widehat{\mathit{F}}}^{\mathit{j}}{\sum }_{\mathit{j}\ne \mathit{i}}^{\mathit{n}}{\sum }_{\mathit{s}=\mathbf{1}}^{\mathit{n}}{\mathit{B}}^{\mathit{j}\mathit{s}}{\mathit{Y}}^{\mathit{s}\mathit{i}}+{\widehat{\mathit{F}}}^{\mathit{j}}{\sum }_{\mathit{j}\ne \mathit{i}}^{\mathit{n}}{\mathit{B}}^{\mathit{j}\mathit{i}}\mathit{E}{\mathit{X}}^{\mathit{i}}+{\widehat{\mathit{F}}}^{\mathit{j}}{\sum }_{\mathit{j}\ne \mathit{i}}^{\mathit{n}}{\mathit{B}}^{\mathit{j}\mathit{i}}{\mathit{\epsilon }}^{\mathit{i}}$ represents the carbon emissions in other provinces caused by final product consumption, exports, and the error term in province $i$. In the calculation of bilateral net carbon transfers between provinces, carbon emission flow values were extracted and aggregated from both the production-side and consumption-side emission matrices. Taking provinces $i$ and $j$ as an example, the bilateral net carbon emission transfer between the two provinces is calculated as Equation (4).

$$T^{ij}=C^{ij}-C^{ji}$$

(4)

$T^{ij}>0$ represents the trade flow between province $i$ and $j$ caused by final product consumption, exports, and the error term, making province $i$ a net carbon inflow region and province $j$ a net carbon outflow region [20].

2.2.2. NPP-Based Carbon Footprint Measurement

The NPP model characterizes the carbon footprint by the ratio of carbon emissions within a province to its average NPP. The regional average NPP represents the ecological carbon absorption capacity allocated per unit of land area. The calculation formula is as follow.

$$EE{F}_{MP}^i={\displaystyle \frac{C_{MP}^i}{\overline{NP{P}^i}}},EE{F}_{MQ}^i={\displaystyle \frac{C_{MQ}^i}{\overline{NP{P}^i}}}$$

(5)

In the equation, $\overline{NP{P}^i}$ is the annual average NPP of province $i$ extracted from MODIS data, $EE{F}_{MP}^i$ is the production-side carbon footprint of province $i$, $EE{F}_{MQ}^i$ is the consumption-side carbon footprint of province. $C_{MP}^i$ is $i$ the aggregated carbon emissions extracted from the provincial production-side emission matrix, $C_{MQ}^i$ is $i$ the aggregated carbon emissions extracted from the provincial consumption-side emission matrix.

Regions with net carbon footprint inflows passively bear higher “carbon absorption” pressure. To estimate the land area required to achieve a “carbon-neutral” state, the net carbon footprint transfer is calculated as:

$$EE{F}_{TRA}^i={\displaystyle \frac{T^{ij}}{\overline{NP{P}^i}}}$$

(6)

In the equation, $T^{ij}>0$, province $i$ is a net carbon footprint inflow region, and $EE{F}_{TRA}^i$ is the net carbon footprint transferred from province $j$ to $i$.

2.2.3. Carbon Footprint Allocation by Benefit Principle

A reasonable design of responsibility allocation coefficients for production-side and consumption-side carbon footprints is key to dividing carbon transfer responsibilities [20]. This study uses the MRIO model to calculate the value-added induced by final product consumption, exports, and the error term between provinces. The formula is as follows:

$$\left[\begin{array}{cccc}{\mathit{V}}^{\mathbf{11}}& {\mathit{V}}^{\mathbf{12}}& \dots & {\mathit{V}}^{\mathbf{1}\mathit{n}}\\ {\mathit{V}}^{\mathbf{21}}& {\mathit{V}}^{\mathbf{22}}& \dots & ⋮\\ ⋮& ⋮& ⋮& ⋮\\ {\mathit{V}}^{\mathit{n}\mathbf{1}}& {\mathit{V}}^{\mathit{n}\mathbf{2}}& \dots & {\mathit{V}}^{\mathit{n}\mathit{n}}\end{array}\right]=\left[\begin{array}{cccc}{\widehat{\mathit{V}}}^{\mathbf{1}}& \mathbf{0}& \dots & \mathbf{0}\\ \mathbf{0}& {\widehat{\mathit{V}}}^{\mathbf{2}}& \dots & \mathbf{0}\\ ⋮& ⋮& ⋮& ⋮\\ \mathbf{0}& \mathbf{0}& \dots & {\widehat{\mathit{V}}}^{\mathit{n}}\end{array}\right]\times \left[\begin{array}{cccc}{\mathit{B}}^{\mathbf{11}}& {\mathit{B}}^{\mathbf{12}}& \dots & {\mathit{B}}^{\mathbf{1}\mathit{n}}\\ {\mathit{B}}^{\mathbf{21}}& {\mathit{B}}^{\mathbf{22}}& \dots & {\mathit{B}}^{\mathbf{2}\mathit{n}}\\ ⋮& ⋮& ⋮& ⋮\\ {\mathit{B}}^{\mathit{n}\mathbf{1}}& {\mathit{B}}^{\mathit{n}\mathbf{2}}& \dots & {\mathit{B}}^{\mathit{n}\mathit{n}}\end{array}\right]\times \left[\begin{array}{cccc}{\mathit{W}}^{\mathbf{11}}& {\mathit{W}}^{\mathbf{12}}& \dots & {\mathit{W}}^{\mathbf{1}\mathit{n}}\\ {\mathit{W}}^{\mathbf{21}}& {\mathit{W}}^{\mathbf{22}}& \dots & {\mathit{W}}^{\mathbf{2}\mathit{n}}\\ ⋮& ⋮& ⋮& ⋮\\ {\mathit{W}}^{\mathit{n}\mathbf{1}}& {\mathit{W}}^{\mathit{n}\mathbf{2}}& \dots & {\mathit{W}}^{\mathit{n}\mathit{n}}\end{array}\right]$$

(7)

In the equation, Matrix $\mathit{W}$ is the sum of the diagonal matrices of final product consumption, exports, and the error term for each province, diagonal matrix $\widehat{\mathit{V}}$ is the value-added matrix, ${\mathit{V}}^{\mathit{i}\mathit{j}}$ denotes the value-added of province $i$ contained in goods outflow from province $i$, where the outflow is induced by final product consumption, exports, and the error term in province $j$.Based on the benefit principle, the carbon footprint transfer allocation factor ${\alpha }^{ij}={\displaystyle \frac{V^{ij}}{V^{ij}+V^{ji}}}$ is designed, with values ranging between [0, 1]. $V^{ij}$ and $V^{ji}$ represent the provincial value-added extracted and aggregated from matrix $\mathit{V}$. A higher ${\alpha }^{ij}$ indicates that province $i$ benefits more from trade flows between province $i$ and $j$, and province $i$ should bear greater carbon footprint transfer responsibility [23]. Therefore, the adjusted industrial carbon footprint for province $i$ under the benefit principle (Equation (8)).

$$EE{F}_{MR}^i=EE{F}_{MP}^i+\sum _{j\ne i}^n{\alpha }^{ij}\times EE{F}_{TRA}^i$$

(8)

In the equation, ${\sum }_{j\ne i}^n{\alpha }^{ij}\times EE{F}_{TRA}^i$ represents the adjustment amount of province $i$ production-side industrial carbon footprint. When $EE{F}_{TRA}^i>0$ province, $i$ is a net carbon inflow region, and the larger the allocation factor ${\alpha }^{ij}$ the greater the economic benefit obtained by province $i$ the more transferred carbon footprint it bears, and the greater the carbon emission responsibility it should assume. When $EE{F}_{TRA}^i<0$ province, $i$ is a net carbon outflow region. If the adjustment is allocated according to the value of ${\alpha }^{ij}$, the larger ${\alpha }^{ij}$ the smaller the transferred carbon footprint assigned to province $i$ which violates environmental equity theory and the principle of matching environmental costs with economic benefits.

To resolve this inconsistency, this paper proposes that when province $i$ is a net carbon outflow region, the allocation factor is switched from ${\alpha }^{ij}$ to $(1-{\alpha }^{ij})$ [20]. In this way, the allocation factor is reversed from a “benefit share” to a “cost share,” so that the degree of responsibility reduction granted to net outflow provinces is positively related to their environmental costs. This corrects the unfairness of the single-benefit principle for resource-based provinces, and the switching mechanism provides a quantitative embodiment of shared responsibility. At this point, $(1-{\alpha }^{ij})$ can be understood as “the proportion of responsibility reduction that a net outflow province should receive for bearing excess production-related carbon emissions.” Through dynamic adjustment of the benefit allocation factor, this approach alleviates the problem of “inflated carbon footprints” in net outflow provinces (energy exporters), strengthens the “consumption-side responsibility” of net inflow provinces (energy importers), and enhances the fairness of carbon responsibility allocation.

2.2.4. Carbon Footprint Allocation by Energy Flow Principle

Building on provincial energy extraction and end-use data, this study calculates interprovincial flow coefficients for four major energy categories: coal, petroleum products, natural gas, and electricity. This approach enables the direct measurement of carbon emissions from four corresponding energy-related industries: coal mining and processing, oil and gas extraction, electricity and heat production/supply, and gas production/supply. The calculated flow coefficients are then applied to adjust carbon footprint responsibility through compensation mechanisms for energy-exporting regions. The relevant calculation formulas are provided below (Equations (9)–(12)).

$$\left[\begin{array}{cccc}{\mathit{C}}_{\mathit{Z}}^{\mathbf{11}}& {\mathit{C}}_{\mathit{Z}}^{\mathbf{12}}& \dots & {\mathit{C}}_{\mathit{Z}}^{\mathbf{1}\mathit{n}}\\ {\mathit{C}}_{\mathit{Z}}^{\mathbf{21}}& {\mathit{C}}_{\mathit{Z}}^{\mathbf{22}}& \dots & {\mathit{C}}_{\mathit{Z}}^{\mathbf{2}\mathit{n}}\\ ⋮& ⋮& ⋮& ⋮\\ {\mathit{C}}_{\mathit{Z}}^{\mathit{n}\mathbf{1}}& {\mathit{C}}_{\mathit{Z}}^{\mathit{n}\mathbf{2}}& \dots & {\mathit{C}}_{\mathit{Z}}^{\mathit{n}\mathit{n}}\end{array}\right]=\left[\begin{array}{cccc}{\widehat{\mathit{F}}}^{\mathbf{1}}& \mathbf{0}& \dots & \mathbf{0}\\ \mathbf{0}& {\widehat{\mathit{F}}}^{\mathbf{2}}& \dots & \mathbf{0}\\ ⋮& ⋮& ⋮& ⋮\\ \mathbf{0}& \mathbf{0}& \dots & {\widehat{\mathit{F}}}^{\mathit{n}}\end{array}\right]\times \left[\begin{array}{cccc}{\mathit{A}}^{\mathbf{11}}& {\mathit{A}}^{\mathbf{12}}& \dots & {\mathit{A}}^{\mathbf{1}\mathit{n}}\\ {\mathit{A}}^{\mathbf{21}}& {\mathit{A}}^{\mathbf{22}}& \dots & {\mathit{A}}^{\mathbf{2}\mathit{n}}\\ ⋮& ⋮& ⋮& ⋮\\ {\mathit{A}}^{\mathit{n}\mathbf{1}}& {\mathit{A}}^{\mathit{n}\mathbf{2}}& \dots & {\mathit{A}}^{\mathit{n}\mathit{n}}\end{array}\right]\times \left[\begin{array}{cccc}{\mathit{W}}^{\mathbf{11}}& {\mathit{W}}^{\mathbf{12}}& \dots & {\mathit{W}}^{\mathbf{1}\mathit{n}}\\ {\mathit{W}}^{\mathbf{21}}& {\mathit{W}}^{\mathbf{22}}& \dots & {\mathit{W}}^{\mathbf{2}\mathit{n}}\\ ⋮& ⋮& ⋮& ⋮\\ {\mathit{W}}^{\mathit{n}\mathbf{1}}& {\mathit{W}}^{\mathit{n}\mathbf{2}}& \dots & {\mathit{W}}^{\mathit{n}\mathit{n}}\end{array}\right]$$

(9)

In the equation, ${\mathit{C}}_{\mathit{Z}}$ represents the direct carbon emission matrix for the four industries (coal mining, etc.). ${\mathit{C}}_{\mathit{Z}}^{\mathit{i}\mathit{j}}$ indicates the direct carbon emissions in the province $j$ caused by the four industries in province $i$. $\mathit{A}$ is the direct consumption coefficient matrix, $\widehat{\mathit{F}}$ is the carbon emission coefficient matrix, and $\mathit{W}$ represents the matrix of final product consumption, exports, and error terms.

$$H^i=(I^i-D^i-P^i)/U^i$$

(10)

In the equation, $H^i$ is the energy flow rate of province $i$, $I^i$ is the energy inflow volume of province $i$, $D^i$ is the energy outflow volume of province $i$, $P^i$ is the energy production volume of province $i$, and $U^i$ represents the end-use energy consumption of province $i$. When $H^i>0$ province $i$ is classified as an energy-importing province, otherwise, it is an energy-exporting province. Assuming there are $n$ provinces, with $m$ being energy importers and $(n-m)$ energy exporters: Energy-importing provinces benefit from energy consumption and thus should assume a greater share of transferred carbon responsibility. This “incremental carbon responsibility” is allocated in proportion to the share of each province’s energy input rate $H^i$ in total input, i.e., higher consumption entails greater responsibility. Conversely, energy-exporting provinces, while gaining economically from energy supply, bear disproportionately high ecological and environmental costs and should thus receive “carbon reduction compensation.” This compensation is allocated based on the share of each province’s energy output rate $H^i$ in total energy output, whereby greater supply volume results in more compensation. Such a differentiated allocation mechanism balances regional economic benefits with environmental costs, providing a quantitative basis for shared carbon responsibility. Therefore, the carbon increase allocation rate for importing provinces is $K^i={\displaystyle \frac{H^i}{{\sum }_{i=1}^m{H}^i}}$, while the carbon reduction allocation rate for exporting provinces is $K^j={\displaystyle \frac{H^j}{{\sum }_{j=1}^{n-m}{H}^j}}$. To address quality differences across coal, natural gas, electricity, and other energy types, the model applies a simplified approach of “physical quantity–based accounting with post-adjustment of environmental attributes.” The key assumptions are as follows: (1) within each energy type, quality is assumed to be homogeneous, with internal property differences ignored; all flows are calculated using standardized physical units (e.g., tons, 100 million m³); (2) flow coefficients are calculated independently for each energy type without cross-interactions, while environmental impacts are incorporated later through an “energy-specific carbon emission coefficient matrix” in the emissions accounting stage; and (3) as a secondary energy, electricity is treated as homogenous across production sources (e.g., thermal, hydro, wind), with physical quantities considered equivalent and environmental differences not reflected in energy flow rates. These assumptions are designed to accommodate the practical limitations of provincial energy statistics, which are primarily reported in physical units, and to enable simplified but comparable allocation of loss responsibilities in quantifying interregional energy flows.

$$EE{F}_{DR}^i=EE{F}_{MR}^i+C_Z^i\times K^i/\overline{NP{P}^i}.$$

(11)

$$EE{F}_{DR}^j=EE{F}_{MR}^j-C_Z^j\times K^j/\overline{NP{P}^j}$$

(12)

In the equation, $EE{F}_{DR}$ represents the adjusted industrial carbon footprint under the energy flow principle. $C_Z^i\times K^i/\overline{NP{P}^i}$ denotes the increased carbon footprint caused by energy imports. $C_Z^j\times K^j/\overline{NP{P}^j}$ indicates the reduced carbon footprint resulting from energy exports.

2.3. Data Sources and Preprocessing

This study utilizes the 2017 China MRIO Table and sectoral carbon emission data published by the China Emission Accounts and Datasets (CEADs). Due to missing data for Tibet, it was excluded. The MRIO categorizes industries into 42 sectors, while the carbon emission data is classified into 45 sectors. To ensure consistency, the two datasets were matched based on industry classifications, with a focus on extractive industries, manufacturing, and utilities (electricity, gas, and water supply sectors), following the sectoral classification framework established by Wang [20]. Data on energy transfers (inbound, outbound), production, and final consumption were sourced from China’s Statistical Yearbooks. NPP data were obtained from NASA’s MODIS MOD17A3HGF satellite product, which provides annual NPP estimates at a 500-m resolution. The data were processed using the MODIS Reprojection Tool (MRT) and ArcGIS 10.8 for batch stitching, cropping, and reprojection. Provincial NPP values were extracted based on China’s administrative boundaries, with results scaled by 0.1 to remove outliers, yielding annual NPP data for each province.

Vegetation NPP was obtained from NASA’s MODIS MOD17A3HGF product (500 m spatial resolution), which reflects the interannual carbon sequestration capacity of provincial ecosystems. The raw datasets were batch-mosaicked and reprojected using the MODIS Reprojection Tool (MRT), then clipped to the administrative boundaries of 30 provinces in ArcGIS 10.8. Since NPP values are stored as integers scaled by a factor of ten (“actual value × 10”), provincial rasters were uniformly corrected by multiplying by 0.1 to avoid systematic overestimation. This correction removed pseudo-abnormal highs and ensured consistency for provincial aggregation and subsequent carbon footprint modeling. While preprocessing improved data reliability, satellite-derived NPP remains subject to sensor error, algorithmic uncertainty, and spatial heterogeneity. NPP, as a core parameter for quantifying ecosystem carbon sequestration capacity, exhibits valuation biases that exert systematic impacts: overestimation may easily lead to the erroneous determination of strong regional carbon sink capacity, potentially resulting in insufficient investment in the protection of carbon pools such as forests and grasslands; while underestimation may exaggerate carbon source risks and trigger excessive intervention. This issue is particularly prominent in specific regions—for resource-rich provinces with inherently low NPP that receive carbon footprint transfers from eastern regions (e.g., northwest arid regions), NPP overestimation caused by spatial heterogeneity will trigger multiple cascading effects: first, it will mask their ecological overload status after receiving carbon footprint transfers, potentially causing biases in carbon footprint quantification and ecological cost assessment; furthermore, leading to deviations of carbon responsibility allocation from reality and undermining the scientificity of the “dual principles of benefit-energy flow” framework, ultimately exacerbating the regional imbalance of “eastern benefits versus western costs”. To reduce these uncertainties, future studies should incorporate ground-based observations or multi-source NPP data fusion.

3. Results and Analysis

3.1. Provincial Carbon Footprint Analysis: Production vs. Consumption

3.1.1. Analysis by Province and Industry Sector

Table 1. Industrial carbon footprint on the production side and consumption side in 2017.

Industrial Carbon Footprint (104 hm2)	Production-side				Consumption-side
	Extractive	Manufacturing	Utilities	Total	Extractive	Manufacturing	Utilities	Total
	7639.05	460,883.59	404,586.00	873,108.63	6982.56	322,255.29	297,833.29	627,071.14

presents the production- and consumption-side industrial carbon footprints for 2017. Against the backdrop of global trade expansion, China’s energy production and consumption have continued to rise, reaching historical peaks. The production-side industrial carbon footprint significantly exceeds the consumption-side, with a gap of 2.46 billion hm². China’s industrial sectors are primarily divided into extractive industries, manufacturing, and utilities (electricity, gas, and water supply).

Manufacturing, as a cornerstone of the national economy, exerts a dominant influence over both trade flows and domestic production and consumption patterns [24]. It represents the largest contributor to industrial carbon footprints, accounting for 52.8% on the production side and 51.4% on the consumption side. Spatially, energy-intensive manufacturing activities are concentrated in northern provinces such as Inner Mongolia and Shanxi, while export-oriented industries are primarily clustered in coastal areas like Guangdong and Jiangsu. Resource-processing industries, meanwhile, tend to be distributed around resource-abundant provinces such as Zhejiang and Henan. This uneven spatial distribution results in significant regional agglomeration of production-side carbon footprints, which are 1.4 times higher than those on the consumption side—constituting the main driver of the observed “dual-side” disparity.

Utilities, which are essential for both industrial operations and residential needs, contribute 46.3% and 47.5% to the industrial carbon footprint on the production and consumption sides, respectively. China’s electricity and gas sectors remain heavily dependent on fossil fuels—including coal, oil, and natural gas—the uneven distribution of which has given rise to interprovincial energy transfer patterns such as “West-to-East Power Transmission” and “North-to-South Coal Transport.” Stringent extraction policies in resource-rich provinces increase energy costs for non-resource provinces, thereby indirectly constraining high-energy demand. Nevertheless, large-scale energy redistribution projects directed toward eastern (e.g., Shanghai, Jiangsu) and southern coastal regions (e.g., Guangdong) unintentionally shift environmental costs, transferring pollution loads from consumers in the east to resource-exporting northwestern and central provinces such as Inner Mongolia, Shaanxi, and Xinjiang.

Thus, cross-provincial energy flows permeate all industrial subsectors, giving rise to a distinct “Western production–Eastern consumption” supply chain. Western resource-exporting provinces are confronted with dual environmental pressures: first, direct ecological degradation resulting from energy extraction activities, which diminishes local carbon sinks; and second, additional carbon pollution arising from the relocation of energy-intensive industries from eastern provinces. These combined pressures significantly inflate production-side carbon footprints in resource-dependent regions. The energy-flow-adjusted accounting framework adopted in this study directly addresses the inequity of “Western emissions versus Eastern consumption benefits,” providing a more balanced approach to responsibility allocation.

3.1.2. Net Carbon Footprint Transfer Analysis

Figure 2a illustrates the net flows of industrial carbon footprints across provinces, with arrows denoting regions of net inflow. At the national scale, carbon footprints are predominantly transferred from the Beijing-Tianjin-Hebei region, the southern coastal areas, and other eastern provinces toward northwestern and northeastern regions. The principal outflow provinces—Beijing, Shandong, Jiangsu, Zhejiang, and Guangdong—together account for 74.21% of the national total outflows. Notably, Guangdong exhibits the largest net outflow, amounting to (−2.74 million hm²). The major inflow provinces include Shanxi, Inner Mongolia, Jilin, Gansu, Qinghai, Ningxia, and Xinjiang, which together account for 86.90% of national inflows. Among these, Xinjiang receives the largest net inflow, totaling (+7.40 million hm²).

The economic hubs of Beijing, Guangdong, Shanghai, and Jiangsu—core areas within the Beijing-Tianjin-Hebei, Pearl River Delta, and Yangtze River Delta economic zones—share common features such as dense populations, high consumption demands, and rapid circulation of goods. These provinces not only transfer substantial carbon footprints to energy-rich northwestern regions, but also generate spillover effects to adjacent areas through their energy-intensive production and distribution activities. This dynamic creates a distinctive “siphon effect” of carbon footprint transfers [16]. Such patterns highlight the urgent need for coordinated interprovincial governance of industrial carbon footprints.

The ecological impacts of industrial carbon footprint transfers on net inflow provinces also exhibit pronounced regional heterogeneity. In the arid and semi-arid northwest (e.g., Inner Mongolia and Xinjiang), large-scale coal mining has damaged grassland vegetation and disrupted soil structure, with warming and drying trends and desertification creating positive feedback loops that further reduce NPP. Transfers from Guangdong alone account for 4.5% of Xinjiang’s annual carbon neutrality target, and similar carbon transfers between eastern provinces and other northwestern resource-based regions such as Inner Mongolia and Gansu highlight the significant transfer pressures faced across the northwest as a whole. In the ecologically fragile southwest (e.g., Qinghai and Gansu), biodiversity loss has been severe; in areas such as the Sanjiangyuan Reserve and the fringes of the Qilian Mountains, rising average temperatures have accelerated grassland degradation and weakened carbon sink capacity. Collectively, these impacts illustrate an imbalance between “eastern economic benefits and western ecological costs,” underscoring the urgent need to establish a carbon responsibility framework that integrates resources, environment, and ecosystem services.

Interprovincial flows of industrial carbon footprints display distinct sectoral patterns, as shown in Figure 2b–e. The electricity supply sector exhibits the most significant transfers, with Guangdong and Shandong together responsible for 53.1% of total outflows directed toward energy-producing regions such as Inner Mongolia and Xinjiang. Notably, this includes a substantial transfer of 1.26 million hm² from Guangdong to Xinjiang, underscoring the pivotal role of basic energy industries in driving carbon footprint redistribution.

Within the manufacturing sector, several subsectors reveal particularly pronounced transfer characteristics. For instance, non-metallic mineral products flow predominantly to Inner Mongolia (2.52 million hm²), with 47.8% originating from Guangdong and Fujian. Transfers of metal products and smelting from Jiangsu, Guangdong, and Shandong to Inner Mongolia total 2.02 million hm². Petroleum, coking, and nuclear fuel products are mainly transferred from Guangdong, Beijing, and Jiangsu to Shanxi and Inner Mongolia, peaking at 0.96 million hm² to Shanxi. Meanwhile, general and special purpose equipment flows are concentrated from Henan to Shanxi (accounting for 86.6% of Henan’s outflow), with Shanxi also receiving 83.97 million hm² from Jiangsu.

These patterns fundamentally reflect regional economic disparities and uneven resource distribution within China. As major suppliers of energy and mineral resources, provinces such as Inner Mongolia and Shanxi absorb substantial carbon footprints in the process of manufacturing basic resource products for more developed regions. The high resource demand of economically advanced provinces drives large-scale transfers of primary products from resource-exporting areas, causing carbon footprints to track resource flow trajectories and greatly intensifying ecological pressures on energy-exporting regions. As China’s leading coal producer, Inner Mongolia has absorbed high energy-intensive industries such as metal smelting and non-metallic mineral products from provinces like Jiangsu and Guangdong. This has driven continuous expansion of industrial land, while open-pit coal mining has reduced grassland vegetation cover. The resulting carbon transfers account for 7.9% of the province’s annual carbon neutrality target. The situation is even more pronounced in Shanxi, where similar industries generate carbon transfers equivalent to 15.1% of its annual neutrality target, compounded by large volumes of acidic wastewater from coal washing that degrade surrounding soils and reduce cropland productivity. In Xinjiang, large-scale thermal power plant construction around the Tarim Basin to meet national electricity demand has overdrawn groundwater, intensified surface evaporation, and accelerated land salinization, further undermining ecosystem stability. Tracing such industry-specific carbon footprint flows clarifies transfer pathways within product supply chains and provides a basis for more equitable allocation of industrial carbon footprint responsibilities.

Accordingly, detailed analysis of sector-specific transfer pathways is essential for clarifying carbon footprint transmission mechanisms along supply chains and for establishing scientifically grounded frameworks for responsibility allocation in industrial carbon transfers.

3.2. Carbon Footprint Responsibility Adjustment

3.2.1. Benefit Principle in Responsibility Allocation

The adjustment of industrial carbon footprint responsibility allocation using the benefit principle reveals distinct provincial patterns, as illustrated in Figure 3. Provinces can be stratified into three tiers according to their trade benefits: high-benefit regions such as Shanghai and Chongqing experience upward adjustments to their production-side footprints (ranging from 1.7% to 5.7%), reflecting their greater capacity to assume emission responsibilities. Medium-benefit provinces, exemplified by Zhejiang, display intermediate allocation factors, often as a result of manufacturing reforms. In contrast, resource-dependent regions such as Shanxi and Xinjiang receive downward adjustments of 0.5% to 0.7%, attributable to limited trade gains and technological constraints. Beijing represents a special case: with allocation factors stabilizing around 0.5, the city transfers a significant proportion of its carbon footprint (53.8% of the total) to five northwestern provinces while maintaining balanced responsibility quotas.

Following adjustment, net carbon-exporting provinces consistently exhibit increased carbon footprints, whereas importing provinces show corresponding reductions. For most provinces, the adjusted values fall between the original production-based and consumption-based measurements—a distribution that strongly validates the principle of “benefit–responsibility proportionality” [20]. This refined allocation framework not only preserves the integrity of national emissions accounting but also addresses inherent spatial inequities, establishing an operational model for equitable interprovincial carbon governance that aligns economic benefits with environmental accountability.

3.2.2. Energy Flow Principle in Carbon Footprint Adjustment

The analysis of industrial carbon footprint adjustments under the energy flow principle reveals pronounced spatial patterns in China’s interprovincial energy transfers (Figure 4). Energy-exporting provinces experience significant reductions in carbon responsibility: Shaanxi, for example, receives a 31.3% deduction for coal exports (equivalent to 25.9 times its local consumption) and a 33.1% reduction for oil transfers, while Inner Mongolia’s extensive natural gas exports (11.1 times its local demand) warrant a 29.0% adjustment. In contrast, energy-importing regions see substantial increases in their assigned responsibility. For instance, Shanghai’s carbon responsibility for coal imports rises by 19.7% (9.2 times its local usage), and Beijing’s responsibility for natural gas dependence increases by 12.3%.

The electricity sector presents unique dynamics: western provinces such as Inner Mongolia and Xinjiang collectively supply 74.8% of power to eastern megacities, yet receive minimal footprint adjustments due to standardized transmission protocols. However, recipient regions like Shanghai face a 35.8% increase in responsibility. These adjustments systematically address the inherent imbalance in China’s “North-South Coal, West-East Power” energy infrastructure, where resource-rich western provinces have historically borne disproportionate environmental costs in supplying developed eastern regions.

By implementing differentiated responsibility allocations—reducing burdens for exporters by 24.1% (Inner Mongolia), while increasing obligations for importers by up to 52.5% (Shanghai)—the framework not only corrects accounting distortions but also establishes an equitable mechanism that more accurately attributes environmental costs to final energy consumers, all while preserving the integrity of national emission totals. This approach offers a replicable model for managing transboundary environmental impacts within federated energy systems and is particularly relevant for other economies experiencing similar regional energy rebalancing.

3.3. Effectiveness of Holistic Carbon Footprint Adjustment

As shown in

Table 2. Carbon footprint and its difference under different measurement principles in 30 provinces in 2017.

Province	Production-Side Industrial Carbon Footprint (104 hm2)	Consumption-Side Industrial Carbon Footprint (104 hm2)	First Adjusted Industrial Carbon Footprint (104 hm2)	Second Adjusted Industrial Carbon Footprint (104 hm2)	First Adjustment Amount (104 hm2)	Second Adjustment Amount (104 hm2)	NPP (tC·hm−2·a−1)
Beijing	2933.75	20,430.99	3011.34	3108.64	77.59	97.30	3.86
Tianjin	22,505.85	13,706.33	22,458.94	22,511.88	−46.92	52.95	2.96
Hebei	35,660.00	38,522.14	35,696.76	35,717.79	36.76	21.03	3.59
Shanxi	62,813.01	20,656.12	62,523.21	62,520.56	−289.80	−2.65	3.76
Inner Mongolia	117,339.41	18,560.69	116,762.58	116,738.45	−576.82	−24.14	2.78
Liaoning	33,100.84	34,279.45	33,095.90	33,095.65	−4.94	−0.25	4.16
Jilin	28,627.90	5670.23	28,551.06	28,545.52	−76.84	−5.54	4.94
Heilongjiang	46,356.43	13,604.89	46,131.72	46,127.60	−224.71	−4.13	4.45
Shanghai	11,058.98	16,677.96	11,101.94	11,154.42	42.96	52.48	4.93
Jiangsu	17,370.51	71,953.29	17,695.66	17,706.54	325.15	10.88	4.89
Zhejiang	7433.64	28,351.15	7560.40	7563.45	126.76	3.05	6.66
Anhui	14,054.79	18,877.43	14,099.70	14,093.83	44.91	−5.87	5.05
Fujian	5579.43	17,196.23	5650.80	5645.93	71.37	−4.86	8.58
Jiangxi	8377.02	12,334.43	8396.90	8396.35	19.88	−0.55	6.82
Shandong	27,005.78	74,776.48	27,286.65	27,295.05	280.88	8.39	3.95
Henan	16,377.99	32,523.64	16,489.44	16,504.61	111.45	15.17	4.28
Hubei	13,039.80	16,491.25	13,066.08	13,052.13	26.28	−13.95	5.83
Hunan	11,255.31	15,974.00	11,303.39	11,299.78	48.08	−3.61	6.79
Guangdong	5021.13	55,059.19	5306.57	5308.96	285.44	2.39	9.71
Guangxi	10,426.92	2835.09	10,387.06	10,382.85	−39.87	−4.21	9.22
Hainan	10,079.29	1891.32	10,023.37	10,020.16	−55.92	−3.21	8.74
Chongqing	8795.10	4648.67	8777.84	8778.48	−17.26	0.64	6.69
Sichuan	13,200.25	17,690.94	13,236.35	13,203.89	36.10	−32.46	5.24
Guizhou	11,572.69	6932.41	11,551.41	11,541.17	−21.28	−10.25	8.15
Yunnan	8817.41	5893.24	8801.33	8780.60	−16.08	−20.73	10.47
Shaanxi	19,724.83	18,505.50	19,730.06	19,723.03	5.24	−7.04	4.71
Gansu	27,462.69	7895.32	27,346.03	27,330.96	−116.65	−15.07	3.69
Qinghai	72,489.71	5183.70	72,096.43	72,066.74	−393.28	−29.69	1.56
Ningxia	79,656.25	9724.97	79,142.88	79,118.53	−513.37	−24.35	2.15
Xinjiang	124,971.91	20,224.10	124,233.10	124,214.36	−738.81	−18.73	1.91

, the initial gap of 2.46 billion hm² between production-side and consumption-side carbon footprints in 2017—highlighting the limitations of single-perspective accounting—was progressively reduced through our two-stage adjustment process. The benefit-based primary adjustment accounted for 46.71 million hm² of changes, narrowing the disparity by 15.94 million hm². The subsequent energy flow-based secondary adjustment contributed an additional 4.96 million hm² of modifications, further reducing the gap by 0.33 million hm². This methodology produced particularly notable results in rebalancing responsibility across key regions: major carbon-exporting provinces such as Guangdong, Jiangsu, and Zhejiang saw their net outflows decrease by 40.4–43.0%, while importing regions including Inner Mongolia, Qinghai, and Xinjiang experienced reductions in net inflows of 58.4–70.5%.

The ecological dimension of the framework adds further validity, as demonstrated by the stark contrast between Guangdong’s high carbon absorption capacity (9.71 tC·hm⁻²·a⁻¹ NPP) and the more vulnerable ecosystem of Inner Mongolia (2.78 tC·hm⁻²·a⁻¹ NPP), justifying the redistribution of environmental burdens. By simultaneously incorporating economic benefits, energy flows, and ecological capacities—while rigorously maintaining the integrity of national emission totals—this methodology achieves a scientifically robust and equitable distribution of carbon responsibilities, in ways that single-factor approaches cannot. The successful application of this balanced framework offers valuable insights for other large economies grappling with similar interregional carbon accounting challenges.

4. Discussion

Current research on carbon emissions has predominantly adopted dual production- and consumption-based perspectives to examine interprovincial trade-embodied carbon transfers and responsibility allocation mechanisms [25,26]. Our analysis of net carbon footprint transfers confirms the spatial pattern of “east-to-west” carbon movement, with footprints shifting from coastal to central and western regions. In this context, energy-supplying provinces become key emission reduction areas, while eastern coastal regions should assume greater transfer responsibilities—findings that align with previous studies [27,28,29]. Departing from prior approaches, our work introduces substantial improvements across three critical dimensions: analytical perspective, responsibility allocation, and research methodology.

By innovatively incorporating carbon absorption pressure into our analytical framework through ecological footprint modeling, we quantitatively assess how carbon transfers impact recipient ecosystems, overcoming the limitation of traditional flow-only analyses [5,30]. This approach explicitly accounts for spatial heterogeneity in carbon sink capacity across provinces. Our results demonstrate that carbon inflows from southern coastal regions significantly increase absorption pressure in western provinces, leading to ecological overload. These findings highlight the necessity for responsibility allocation mechanisms that incorporate “regional carbon sink capacity” and “ecological vulnerability” as key variables, with differentiated weighting to prevent the unfair cost-shifting to ecologically fragile areas [26,31].

Regarding responsibility allocation, while existing studies have largely focused on macro-level producer/consumer debates [26,31], we systematically examine energy–carbon linkages across industrial supply chains. Northern industrial provinces—serving as China’s primary energy production and processing hubs—exhibit carbon intensities 32% higher than non-energy regions [32], with extraction and processing activities inflating their apparent carbon footprints. Our sector-specific approach traces energy–carbon pathways, quantifies transfer coefficients, and uncovers hidden responsibilities across 42 industrial sectors, thereby developing a comprehensive allocation model. This model integrates both interprovincial trade-embodied transfers and direct industrial energy flows, while also considering regional environmental carrying capacities. Interprovincial carbon compensation faces persistent challenges. The core difficulty lies in differing perceptions of value: net outflow provinces (economically developed) favor cost estimates based on carbon market prices, while net inflow provinces (economically underdeveloped) emphasize ecological restoration standards, making consensus elusive. Behaviorally, developed provinces often shift energy-intensive industries to evade responsibility, while resource-based provinces expand coal extraction to boost short-term revenue, creating a disconnect between compensation and mitigation goals. Institutionally, the framework remains underdeveloped, lacking efficiency metrics for fund use, dedicated legal provisions, clear responsibility and penalty mechanisms, and independent third-party oversight—factors that collectively hinder effective implementation.

From a methodological perspective, our study advances beyond single-method approaches such as MRIO [25,33] or IPCC [5] techniques by integrating multi-regional input–output analysis with net primary productivity modeling. Incorporating key parameters such as the carbon footprint transfer coefficient, environmental carbon carrying capacity, and NPP values into the compensation accounting framework. This novel combination enables a comprehensive assessment of both economic linkages and ecological impacts, establishing a systematic theoretical framework for industrial carbon accounting in China.

This study is constrained by the timeliness of data, as the provincial MRIO tables are only available up to 2017. Since then, however, China’s industrial carbon footprint dynamics have shifted significantly. Eastern provinces such as Jiangsu and Guangdong have advanced industrial upgrading by reducing manufacturing dependence and expanding low-carbon industries, thereby lowering their net carbon footprint outflows. In contrast, central and western resource-based provinces like Inner Mongolia and Shaanxi have introduced green manufacturing standards that reduce the intensity of net carbon footprint inflows. Simultaneously, renewable energy transitions—such as “green electricity transmission to the east,” ultra-high-voltage hydropower delivery, and distributed solar in coastal provinces—have reshaped traditional coal-based transfer pathways and reduced overall carbon intensity.

Looking ahead, the deep substitution of fossil fuels by renewables is driving a structural transformation in China’s industrial system. Future research should incorporate renewable energy flows into MRIO frameworks, with attention to interregional allocation mechanisms and substitution effects. Such analysis can strengthen the theoretical and empirical foundation for provincial cooperative mitigation policies, regional carbon equity, and industrial green transitions.

More broadly, the proposed framework—based on the dual principles of benefit allocation and energy flows—offers advantages over producer- or consumer-only approaches and aligns with the CBDR principle in global climate governance. Its applicability spans scales (national, provincial, urban) and sectors (transportation, construction, agriculture via water-footprint parameters). Nationally, it adapts to developed (high-energy-consuming, resource-importing), resource-exporting (high-energy-exporting, ecologically vulnerable), and developing (industrial-expanding) countries, incorporating analysis of globalized energy supply chain expansion and transnational value chain benefit distribution. Industrially, service sectors (transportation, construction) directly apply “benefit allocation factor” and “energy flow rate”, while agriculture requires innovative water-footprint-based parameters (linking virtual water trade to carbon responsibility). By integrating energy, economic, and ecological dimensions, this approach provides a “Chinese experience” that enriches international carbon responsibility allocation and contributes to achieving UN Sustainable Development Goals 12 (Responsible Consumption and Production) and 13 (Climate Action).

5. Conclusions and Policy Implications

5.1. Conclusions

This study integrates MRIO modeling with NPP analysis to quantify industrial carbon footprints and develop an equitable allocation framework across 30 Chinese provinces. The findings reveal substantial east-to-west transfers of industrial carbon, with energy-rich provinces like Shanxi, Inner Mongolia, and Xinjiang receiving the largest inflows, while coastal provinces such as Jiangsu, Zhejiang, Shanghai, and Shandong serve as principal sources. By linking carbon responsibility to both trade benefits and interprovincial energy flows, the framework ensures that high-benefit and energy-importing regions assume a greater share of responsibility, while burdens for resource-exporting provinces are proportionally reduced. Most provinces’ adjusted carbon footprints fall between production- and consumption-based values, highlighting the framework’s effectiveness in addressing regional disparities and providing a more balanced approach to carbon governance. Collectively, these results offer a robust and adaptable model for advancing fair and effective carbon responsibility allocation in China and other large, regionally diverse economies.

5.2. Policy Implications

Based on our findings, we recommend four key policy measures to address interprovincial carbon footprint imbalances and promote equitable decarbonization in China. First, differentiated dual-track carbon footprint management. Provinces should adopt tailored strategies based on their resource endowments, environmental conditions, and industrial structures. For provinces with net carbon footprint inflows (energy-supplying regions), policy priorities include improving industrial energy-use efficiency, accelerating low-carbon energy transitions, and setting fossil-fuel extraction quotas to prevent overexploitation. Greater investment should also be directed toward the research and deployment of renewable energy such as solar and wind. For provinces with net carbon footprint outflows (economically advanced regions), emphasis should be placed on reducing demand for energy-intensive products to curb upstream emissions, while expanding green infrastructure to offset ecological impacts. In ecologically fragile western areas where energy extraction is concentrated, ecological protection zones should be established. Sustained ecological restoration would enhance ecosystem carbon sequestration and mitigation capacity, forming a coordinated system of “production-side mitigation, consumption-side regulation, and ecosystem restoration” to support interregional cooperation on carbon governance [34].

Second, establishing an ecological carbon compensation mechanism. Guided by the principle of shared responsibility for industrial carbon footprint transfers, a compensation framework should be developed that combines dynamic adjustment of carbon footprints with carbon market prices as the central accounting benchmark. By quantifying the amount of carbon footprint transferred across provinces, quota trading in the carbon market can transmit responsibility and cost through market-based mechanisms. For instance, in Guangdong (NPP = 9.71 tC·hm⁻²·a⁻¹), the combined adjustment volume of primary and secondary industrial carbon footprints amounts to 2.88 million hm². Using the national average carbon price of 60 RMB/tC, the corresponding compensation is approximately 45.7 million RMB. This approach minimizes disputes over “monetizing ecological value.” Compared with cost-based ecological compensation, which is hindered by regional heterogeneity (e.g., NPP values of 2.78 tC·hm⁻²·a⁻¹ in Inner Mongolia vs. 9.71 in Guangdong) and substantial cost variation (3–5 times difference), a carbon-price-based mechanism provides a more objective benchmark. To stabilize expectations when disputes cause market volatility, the National Climate Strategy Center could publish a “regional carbon price index” that incorporates NPP values and ecological governance cost coefficients, enabling carbon price adjustments that reduce uncertainty and overcome the heterogeneity of ecological cost accounting [35].

Third, advancing regional ecological strategies through coordinated action. Linking ecological carbon compensation funds with the China Certified Emission Reduction (CCER) mechanism, and integrating water footprints, air pollutants, and other environmental indicators, a synergistic “carbon–water–air” governance framework can be established. Energy-exporting provinces could deploy compensation funds in forestry carbon sink projects, with additional CCER certification standards incorporating water-use accounting (e.g., water consumption from coal mining) and air-pollution mitigation (e.g., desulfurization and denitrification retrofits). The CCERs generated could then be traded to energy-importing provinces such as Guangdong to offset quota obligations. In the Yellow River Basin, carbon footprint transfer responsibilities could be linked to soil–water conservation and joint air-pollution control, with compensation funds earmarked for “mine reclamation + carbon sink forest construction.” Improvements in ecological performance (e.g., higher NPP values) would provide the basis for quota adjustments in subsequent years. In the Beijing–Tianjin–Hebei region, a “Beijing–Inner Mongolia Carbon Compensation Fund” could be established, whereby Beijing purchases Inner Mongolia’s wind-power CCERs proportionally to its transferred carbon footprint, fostering a sustainable pattern of “regional collaboration, emission reduction, and shared benefits [30].

Fourth, sharing the benefits and costs of carbon emissions. Anchored in sustainable development, responsibility for industrial carbon footprints should be allocated to balance economic benefits, environmental costs, and social equity. Building upon existing ecological compensation policy frameworks (e.g., the Ecological Protection Compensation Regulations), carbon footprint transfer compensation is operationalized, and latent responsibilities are explicitized through carbon market quota adjustments, aiming to address the “eastern benefit-western cost” imbalance. Stronger central–local coordination is required, along with improved legislation such as the Ecological Compensation Regulation and the Interim Regulation on Carbon Emission Trading Management. NPP ecological thresholds should be incorporated into compensation rules, and a performance system established linking “financial compensation, emission reduction outcomes, and ecological improvements,” with third-party oversight. Energy-exporting provinces (e.g., Inner Mongolia) could issue “carbon compensation bonds,” channeling raised funds into ecological projects such as photovoltaic-based desert control, with bond yields tied to carbon sink outcomes. Energy-importing provinces (e.g., Guangdong) could purchase these bonds, generating a cycle of “ecological protection, economic benefits, and regional cooperation.” With central policy as guidance, market mechanisms as the core, and legal frameworks as the safeguard, sustainable operations can be achieved through equity returns and bond interest, with compensation payments prioritized from proceeds. Ultimately, this framework enables the scientific allocation of carbon responsibilities, advancing regional green transitions and sustainable development goals in tandem.

Fifth, we propose the dynamic expansion of the carbon transfer responsibility allocation framework. Policy impact prediction is achieved by setting multi-dimensional scenario parameters: Under the energy structure adjustment scenario, we assume a 25% increase in the share of non-fossil energy by 2030, an average annual growth of 6% in carbon prices, and a decrease in power transmission losses. Simulations indicate that the reduction in Xinjiang’s carbon responsibility for traditional energy exports is reduced by over 20%, while Guangdong’s consumption-side responsibility increases by 15% due to indirect carbon transfers from energy-intensive product demand, with compensation funds rising from 45.7 million yuan in 2022 to 68 million yuan in 2030. Under the industrial transfer restriction scenario, the implementation of annual 10% quota management for cross-provincial transfers of energy-intensive industries predicts a 10–15% reduction in the carbon responsibility share of Jiangsu, Shanghai, and other provinces by 2030. Under the dynamic carbon compensation adjustment scenario, compensation standards are linked to NPP values (e.g., a 0.1 increase in the compensation coefficient for every 1 tC·hm⁻²·a⁻¹ increase in Inner Mongolia’s NPP), and when combined with 5-year time slices aligned with the “dual carbon” policy, Shanxi, Shaanxi, and other provinces experience a 5–8% narrowing of traditional energy responsibility reductions due to new energy substitution. This framework integrates industrial transformation, energy substitution, and ecological cost accounting, achieving the transition from static accounting to dynamic prediction and providing a quantitative tool for the state to formulate phased plans and evaluate policy effectiveness.