Spatiotemporal Characteristic and Driving Factors of Synergy on Carbon Dioxide Emission and Pollutants Reductions in the Guangdong–Hong Kong–Macao Greater Bay Area, China

Abstract

As an economically active region, the Guangdong–Hong Kong–Macao Greater Bay Area (GBA) faces dual challenges of carbon and air pollution reduction. Existing studies predominantly focus on single pollutants or engineering pathways, lacking systematic analyses of multi-scale synergistic effects. This study investigates the spatiotemporal distributions, driving factors, and synergistic effects of CO₂ and volatile organic compounds (VOCs) at the multi-scale of urban agglomerations, cities, and industries, using global Moran’s index, standard deviational ellipse, logarithmic mean divisa index decomposition model, and Tapio decoupling model. The results show that the average annual growth rate of CO₂ (7.4%) was significantly higher than that of VOCs (4.5%) from 2000 to 2020, and the industrial sector contributed more than 70% of CO₂ and VOC emissions, with the center of gravity of emissions migrating to Dongguan. Industrial energy intensity improvement emerged as the primary mitigation driver, with Guangzhou and Shenzhen demonstrating the highest contribution rates. Additionally, CO₂ and VOC reduction show two-way positive synergy, and the path of “energy intensity enhancement–carbon and pollution reduction” in the industrial sector is effective. Notably, the number of strong decouplings of the economy from CO₂ (11 times) is much higher than the number of strong decouplings of VOCs (3 times), suggesting that the synergy between VOC management and economic transformation needs to be strengthened. This study provides scientific foundations for phased co-reduction targets and energy–industrial structure optimization, proposing regional joint prevention and control policy frameworks.

1. Introduction

As the world’s largest developing country, China’s carbon emission research provides key support and demonstration for the achievement of global carbon reduction targets. Nonetheless, atmospheric environmental governance confronts more complex and urgent challenges [1,2]. The energy-intensive development model has not only escalated greenhouse gas emissions but also triggered compound air pollution [3]. Notably, the emissions of volatile organic compounds (VOCs) increased by 18.74 Tg during 1990–2017 [4], forming complex secondary reaction mechanisms with particulate matter (PM) and nitrogen oxides (NOₓ), thereby exacerbating regional air quality challenges. China released the Action Plan for Continuous Improvement of Air Quality [5], which for the first time included synergistic control of VOCs, PM, and NOx in the core strategy and strengthened spatially differentiated measures. Within this context, synergistic emission reduction of air pollutants (VOCs, PM, NOx, etc.) and CO₂ constitutes a key pathway for achieving green development and air quality improvement. Research on VOCs as precursors of O₃ [6] is still in its beginning stages (Figure 1).

Established research has formed a theoretical system for carbon and pollution (CAP) synergy. Early research identified significant co-benefits from greenhouse gas mitigation technologies, with European studies confirming that the reduction of carbon emissions caused by energy structure adjustment can simultaneously reduce air-pollutant emissions [7]. Subsequent work revealed that air pollution control technologies also significantly reduce carbon emissions [8]. Post-Paris Agreement (2016), research shifted toward systematic synergy evaluation, integrating theoretical frameworks, measurement techniques, policy instruments, and institutional mechanisms [9,10]. Recent advancements focus on quantitative synergy metrics under climate change pressures [11,12,13,14,15]. Existing evaluation methodologies primarily encompass the coupling coordination model and indicator evaluation method. The coupling coordination model, as an analytical tool for assessing interaction intensity and mutual influences between systems, quantifies the degree of benign coupling among two or more interconnected systems [16]. The indicator evaluation method establishes a multidimensional index system to conduct quantitative assessments of alternative solutions or research objects, ultimately generating composite scores to facilitate optimal decision-making [17]. However, its application of standardization and normalization techniques across heterogeneous domains and indicators may compromise the intuitiveness and accuracy of comprehensive synergistic evaluations. Current research on CAP synergy necessitates paradigm shifts toward quantitative characterization and precision-oriented assessment frameworks.

Current studies on synergistic control of CAP primarily investigate three distinct geographical scales. Urban-scale research is suitable for in-depth analyses of specific issues within cities. Guan et al. (2023) employed the bivariate synergy index (BSI) and integrated synergy index (ISI) to quantify the synergistic relationship between urban-level CO₂ and air-pollutant (PM₂.₅ and O₃) emissions [18]. Xue et al. (2023) developed an index system comprising ISEC-AC (synergy evaluation composite index) with two sub-indices—IHI (industrial harmonization index) and ICE (carbon efficiency index)—to assess CAP synergy from energy structure, industrial composition, and spatial distribution perspectives [19]. Wang et al. (2024) constructed the degree of urban CAP reduction synergy (DUSR) from the three-part connotation of target synergy (T), path synergy (P), and management synergy (M) [20]. For provincial scales, Yi et al. (2022) applied a synergistic assessment model to calculate the coordinated air quality–carbon mitigation degree (CASD) across 30 Chinese provinces during 2005–2018 [21]. Moreover, cities within the urban agglomeration are geographically close to each other and have similar natural geographic environments, and there is a close industrial division of labor and collaboration among them, resulting in the formation of complete industrial chains and clusters. Urban agglomeration contains cities of different sizes and grades, forming a reasonable city system and realizing the complementary functions of large, medium, and small cities. EEach urban assumes different functional roles in the urban agglomeration, such as political, economic, cultural, scientific and technological, financial, and other functions, collaborating and supporting each other. The studies focused on the urban agglomerations have systematically examined metropolitan regions, including Beijing–Tianjin–Hebei, the Yangtze River Delta, and the Pearl River Delta, elucidating their distinct decarbonization pathways and spatial differentiation characteristics [22,23,24]. These multidimensional investigations have significantly advanced the scientific understanding of coordinated emission reductions for CAP.

To analyze the driving factors of synergistic emission reduction, scholars have explored urban agglomerations and urban-scale factors. Sectoral and regional analyses consistently demonstrate significant heterogeneity in emission patterns across electricity generation, industrial manufacturing, transportation, and residential sectors [25,26]. Key determinants encompass economic-development level, population density, GDP growth, urbanization rate, industrial output proportion, energy-consumption structure, and energy intensity [27,28,29,30,31,32,33]. At present, the research methods on the driving factors of carbon emissions and air pollutants are mainly divided into five categories: environmental Kuznets curve model, econometric and statistical model, Kaya equation, decomposition analysis model, and decoupling model [34,35,36,37,38,39]. Impacts by population, affluence, and technology model (IPAT) or stochastic impacts by regression on population, affluence, and technology model (STIRPAT) belong to the econometric statistical models. Structural decomposition analysis (SDA) and log mean dichotomy index (LMDI) are decomposition analysis models. These models are currently widely used. However, the quantitative regression analysis method based on IPAT model or STIRPAT model has some limitations in selecting driving factors. The SDA method is based on input–output analysis, and the analysis process is complex. The LMDI method is easier to analyze in detail. Under the premise of satisfying the reversibility of the factor, the residual and zero values are eliminated, and the accuracy is guaranteed. It is widely used in energy demand or to identify the driving factors of emission changes. If other methods are used instead of LMDI, residuals or improper residual decomposition may occur after the decomposition of explanatory variables, which cannot achieve the similar objective LMDI method. Therefore, in order to obtain reliable results, it is of great significance to analyze the influencing factors of carbon emissions via the LMDI method. The LMDI model has been widely used on a global scale [40], national scale [41], regional scale, and urban scale [42]. The research also involves decomposition analysis of carbon emissions at the industry or sector level, such as electricity, manufacturing, construction, transportation, residential life, etc. [43,44]. Despite systematic revelations about conventional pollutants (SO₂, NOx, PM, and O₃) through spatiotemporal heterogeneity analyses, critical knowledge gaps persist regarding carbon–-VOC co-evolution dynamics. Crucially, achieving synergistic mitigation targets requires transcending the conventional environment–economy dichotomy paradigm through innovative policy-integration frameworks. While existing studies systematically revealed spatiotemporal variations in and driving factors of conventional pollutants like SO₂, NOx, PM, and O₃, systematic insights into the evolution of CAP remain scarce.

The Guangdong–Hong Kong–Macao Greater Bay Area (GBA), a pivotal region under China’s national development strategy, exemplifies high-intensity urbanization and industrial agglomeration. It generates 12.57% of national GDP on merely 0.6% of China’s land area [45], with an economic output density 20.9 times the national average. The traditional atmospheric pollutants (PM₂.₅ and SO₂) in the GBA have crossed the turning point of Environmental Kuznets Curve (EKC), whereas O₃ and VOCs are still in the rising phase of the EKC curve. This transitional state indicates that the region has not fully entered the post-inflection era characterized by sustained environmental-quality improvement, thereby remaining constrained by dual structural pressures from carbon emission-peaking commitments and compound atmospheric pollution challenges [46,47]. While achieving PM₂.₅ reductions (34% decline from 2015 to 2022), secondary pollution dominated by O₃ has intensified, with ground-level concentrations rising from 44 μg/m³ (2015) to 61 μg/m³ (2022), exhibiting 3.8% annual growth [48]. The Implementation Plan for Building a Pioneer Zone of Beautiful China (2025) indicates a shift from single-target governance to systemic governance. However, trans-regional challenges persist in PM₂.₅ reduction, O₃ control, and carbon peaking, underscoring the urgency of GBA-specific CO₂-VOC co-governance research.

Building on the previous research, in order to promote the refined urban management of CAP emission reduction in the GBA, exploration of co-abatement driving mechanisms and decoupling analysis is urgently needed. The interaction between the driving factors determines the level and trend of CAP emissions. Characterizing the spatial–temporal differentiation of CAP and revealing evolution patterns in key cities/sectors provide a foundation for subsequent coordinated governance of CAP reduction in the country. This study investigates CAP synergy in the GBA across three scales: urban agglomeration, 9 + 2 cities, and key sectors (power, industry, transportation, and residential).

The main objectives of this study were to (1) assess the spatiotemporal evolution of CAP to identify emission hotspots across cities and sectors; (2) quantify the driving factors of CAP emissions; (3) analyze the difference between the driving effect of carbon reduction on pollution reduction and the driving effect of pollution reduction on carbon reduction; and (4) evaluate economy–emission disentanglement state. It provides theoretical reference and decision-making guidance for the coordinated emission reduction of carbon emissions and VOCs in urban agglomerations, cities, and industries.

2. Methods and Data

2.1. Research Methods of Spatiotemporal Evolution Characteristics of CAP

2.1.1. Spatial Autocorrelation Analysis

The first law of geography states that “Everything is related to everything else, but near things are more related than distant ones”, which has been proved to be applicable to CAP emissions in previous studies [49]. This indicates that CAP emissions in different cities will be synergistically affected by the surrounding areas [50]. To quantify the degree of spatial autocorrelation, the Global Moran’s I—characterized by its positive/negative values and significance levels—is commonly employed to evaluate the global agglomeration effects of CAP emissions across cities in the GBA. The statistical significance of the global Moran’s I is usually assessed through Monte Carlo simulations or approximate tests based on the normal distribution. A series of simulated values of the Moran’s I are calculated by randomly replacing the spatial locations of the observations several times, and then the actual calculated Moran’s I is compared with the simulated values to determine if it is significantly different from what would be expected in the case of a random distribution. If the actual Moran’s I falls in the extreme tails of the simulated values (usually at the 5 per cent or 1 per cent significance level), the original hypothesis can be rejected, and the observations can be considered to be spatially significantly autocorrelated. These can be expressed as follows:

$$I={\displaystyle \frac{{\sum }_{i=1}^n{\sum }_{j=1}^n{W}_{ij}\left(Y_i-\overline{Y}\right)}{S^2{\sum }_{i=1}^n{\sum }_{j=1}^n{W}_{ij}}}$$

(1)

$$S^2={\displaystyle \frac{1}{n}}{\sum }_{i=1}^n{\left(Y_i-\overline{Y}\right)}^2$$

(2)

$$\overline{Y}={\displaystyle \frac{1}{n}}{\sum }_{i=1}^n{Y}_i$$

(3)

In the formula, $n$ represents the total number of samples of a variable; $W_{ij}$ is the weight matrix, which describes the spatial relationship between geographical units. In this paper, the binary adjacency matrix is used. $Y_i$ and $Y_j$ represent the observed values of region i and region j, respectively. The Moran’s I index is in a range of [−1, 1]. A positive value indicates positive spatial autocorrelation of the studied attribute across regions, with values approaching 1 signifying stronger spatial clustering and more significant agglomeration effects. Conversely, values approaching −1 indicate increasingly divergent spatial distributions. A value of 0 suggests no spatial correlation, meaning regions are independent of one another [51].

2.1.2. Standard Deviation Ellipse (SDE)

As a fundamental tool in exploratory spatial data analysis, the SDE enables systematic identification of spatial clustering patterns, directional heterogeneity, and non-uniform distribution characteristics of geographical elements [52].

SDE is centered on the center of gravity of the spatial distribution of the study object; the long axis of the ellipse indicates the direction of maximum dispersion of spatial elements, and the short axis indicates the direction of minimum dispersion of spatial elements. That is, the changes of the center of gravity, and the long and short axes of the standard deviation ellipse are used to reflect the spatial distribution and movement characteristics of carbon and VOCs. The area of the ellipse reflects the characteristics of emission distribution, and the smaller the area, the more aggregated the emission distribution [53]. The formula is as follows:

$$\overline{X_w}={\displaystyle \frac{{\sum }_{i=1}^n{w}_i{x}_i}{{\sum }_{i=1}^n{w}_i}}; \overline{Y_w}={\displaystyle \frac{{\sum }_{i=1}^n{w}_i{y}_i}{{\sum }_{i=1}^n{w}_i}}$$

(4)

$$tan\theta ={\displaystyle \frac{\left({\sum }_{i=1}^n{\overline{x_i}}^2-{\sum }_{i=1}^n{\overline{y_i}}^2\right)\sqrt{{\left({\sum }_{i=1}^n{\overline{x_i}}^2-{\sum }_{i=1}^n{\overline{y_i}}^2\right)}^2+4{\left({\sum }_{i=1}^n\overline{x_i} \overline{y_i}\right)}^2}}{2{\sum }_{i=1}^n\overline{x_i} \overline{y_i}}}$$

(5)

$${\sigma }_x={\displaystyle \frac{\sqrt{{\sum }_{i=1}^n\left(w_i\overline{x_i}\cos \theta -w_i\overline{y_i}\sin \theta \right)}}{{\sum }_{i=1}^n{w_i}^2}}$$

(6)

$${\sigma }_y={\displaystyle \frac{\sqrt{{\sum }_{i=1}^n\left(w_i\overline{x_i}\sin \theta -w_i\overline{y_i}\cos \theta \right)}}{{\sum }_{i=1}^n{w_i}^2}}$$

(7)

In the formula, ($x_i$,$y_i$) is the spatial location coordinates of the ith sampling point; $w_i$ is the weight; $n$ is the number of cities; $\theta$ is the turning angle, representing the main trend of the spatial distribution of the research object; ${\sigma }_x$ represents the length of the long axis of the ellipse; ${\sigma }_y$ represents the length of the short axis of the ellipse; and the area of the standard deviation ellipse reflects the aggregation status of the distribution of carbon and VOC emissions. The standard deviation ellipse area reflects the aggregation of carbon and VOC distribution: the smaller area indicates the more aggregated emission distribution.

2.2. Decomposition of CAP Synergistic Emission Reduction Effect

2.2.1. CAP Emission Reduction Drive Decomposition Based on Kaya–LMDI Model

Combined with the spatial and temporal characteristics of CAP emissions in the GBA and the feasibility of factor quantification, the main factors affecting CAP emissions can be summarized as intensity factors (emission intensity, energy intensity, etc.), structural factors (energy structure, economic structure, etc.), and scale factors (economic development, population size, etc.) [54] (Figure 2).

In 1989, Japanese scholar Yoichi Kaya proposed Kaya’s constant equation to factorize carbon emissions in response to changes in carbon emissions [55]. The expression is:

$${CO}_2={\displaystyle \frac{{CO}_2}{PE}}·{\displaystyle \frac{PE}{GDP}}·{\displaystyle \frac{GDP}{POP}}·POP$$

(8)

where ${CO}_2$, $PE$, $GDP$, and $POP$ denote carbon emissions, energy consumption, gross domestic product, and population, respectively.

An improved Kaya identity combined with the LMDI method was used to construct a CAP-driven decomposition model for the GBA. This model decomposes total CAP emissions into sectoral contributions across the GBA’s 9 + 2 cities [56]:

$$C={\sum }_{i=1}^{11}{\sum }_{j=1}^4{C}_{ij}={\sum }_{i=1}^{11}{\sum }_{j=1}{\displaystyle \frac{C_{ij}}{F_{ij}}}{\displaystyle \frac{F_{ij}}{Y_{ij}}}{\displaystyle \frac{Y_{ij}}{Y_i}}{Y}_i+{\sum }_{i=1}^{11}{\sum }_{j=2}^4{\displaystyle \frac{C_{ij}}{F_{ij}}}{\displaystyle \frac{F_{ij}}{F_i}}{\displaystyle \frac{F_i}{Y_i}}{\displaystyle \frac{Y_i}{Y}}{\displaystyle \frac{Y}{P}}P$$

(9)

In the formula, $i$ represents 9 + 2 cities in the GBA, where $i$ = 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11; and $j$ represents sectors, with $j$ = 1 Industry, 2 Electricity, 3 Transportation, and 4 Residential.

Equation (9) can be further rearranged as follows:

$$C={\sum }_{i=1}^{11}{\sum }_{j=1}{EMI}_{ij}\times {IEI}_{ij}\times {IES}_{ij}\times Y_i+{\sum }_{i=1}^{11}{\sum }_{j=2}^4{EMI}_{ij}\times {ES}_{ij}\times {EI}_i\times {GS}_i\times PGDP\times P$$

(10)

The definitions of parameters involved in the model calculation process are provided in

Table 1. Symbols and meanings of variables in the model.

Symbols	Meanings	Symbols	Meanings
$C$	CO2/VOC total emissions	${EMI}_{ij}$	CO2/VOC emission intensity of the j sector in City i
$C_{ij}$	CO2/VOC emissions of the j sector in City i	${IEI}_{ij}$	Energy consumption intensity of industrial sectors
$F_{ij}$	Energy consumption	${IES}_{ij}$	Level of industrialization of City i
$Y_{ij}$	Output value of the j sector in City i	${ES}_{ij}$	Energy consumption structure of the j sector in City i
$Y_i$	Economic output of City i	${EI}_i$	Energy intensity of City i
$Y$	Gross regional product (GDP)	${GS}_i$	Proportion of the economic output of City i
$P$	Number of permanent residents	$PGDP$	Per capita GDP

:

This study employs time-series data from 2005 to 2020 for the GBA, with statistical analysis of model variables provided in

Table 2. Summary statistics for the model variables.

Variables	Unit	Mean	Standard Deviation	Min	Max
Total energy consumption	Tons of standard coal	19,972.07	3322.73	14,287.75	24,933.67
Industrial added value	Hundred million yuan	25,350.29	9016.09	10,622.19	38,031.49
Gross regional product	Hundred million yuan	79,592.97	29,957.67	32,711.41	124,690.17
Gross domestic product per capita	Ten thousand yuan/person	11.14	2.65	6.20	14.66
Population	Ten thousand person	7025.69	1093.08	5279.35	8634.52

.

2.2.2. CAP Synergistic Emission Reduction Effect Model Based on Kaya–LMDI Model

The Kaya–LMDI model was also used for the decomposition of the drivers for CAP synergistic emission reduction effect, incorporating the synergistic carbon-pollutant reduction coefficient (Csyn) into the decomposition equation. This study investigated the contributions of the Csyn and CO₂ reduction intensity (CY) to VOC emissions, as well as the reciprocal contributions of the Csyn and VOC reduction intensity (VY) to CO₂ emissions.

$${VOC}_s={\sum }_{i=1}^{11}{\sum }_{j=1}^4{C}_{ij}={\sum }_{i=1}^{11}{\sum }_{j=1}^4{\displaystyle \frac{{VOC}_{sij}}{{CO}_{2ij}}}{\displaystyle \frac{{CO}_{2ij}}{Y_i}}{\displaystyle \frac{Y_i}{P_i}}{P}_i$$

(11)

$${CO}_2={\sum }_{i=1}^{11}{\sum }_{j=1}^4{C}_{ij}={\sum }_{i=1}^{11}{\sum }_{j=1}^4{\displaystyle \frac{{CO}_{2ij}}{{VOC}_{sij}}}{\displaystyle \frac{{VOC}_{sij}}{Y_i}}{\displaystyle \frac{Y_i}{P_i}}{P}_i$$

(12)

$$C_{syn}={\displaystyle \frac{{ER}_{{CO}_2}}{{ER}_{{VOC}_s}}}\mathrm{or}{\displaystyle \frac{{ER}_{{VOC}_s}}{{ER}_{{CO}_2}}}$$

(13)

In the formula, $i$ and $j$ denote the same indices as defined in previous equations; ${ER}_{{CO}_2}$ represents CO₂ emission reductions; ${ER}_P$ denotes synergistic air pollutant reductions; and $C_{syn}$ is the synergistic CAP reduction coefficient, defined as the CO₂ emission reductions per unit of pollutant reduction.

Equations (12) and (13) can be further rearranged as follows:

$${VOC}_s={\sum }_{i=1}^{11}{C}_i={\sum }_{i=1}^{11}{C}_{{syn}_i}\times {CY}_i\times {PGDP}_i\times P_i$$

(14)

$${CO}_2={\sum }_{i=1}^{11}{C}_i={\sum }_{i=1}^{11}{C}_{{syn}_i}\times {VY}_i\times {PGDP}_i\times P_i$$

(15)

The definitions of each variable within the model can be found in

Table 3. Symbols and meanings of variables in collaborative model.

Symbols	Meanings	Symbols	Meanings
$VOC$	Total VOC emissions	$C_{{syn}_i}$	Cooperative emission reduction coefficient of CO2 and VOCs
$CO$	Total CO2 emissions	${CY}_i$	CO2 emission reduction intensity of City i
$C_{ij}$	CO2/VOC emissions of the j sector in City i	${VY}_i$	VOC emission reduction intensity of City i
$Y_i$	Economic output of City i	$PGDP$	Per capita GDP
$P$	Number of permanent residents

.

2.3. Decoupling Analysis Method

Decoupling analysis has been extensively employed to quantify dynamic dissociation mechanisms between economic development trajectories and environmental pollution burdens. Early decoupling studies focused on linear relationships between single pollutants (e.g., SO₂ and NOx) and economic growth, neglecting multi-pollutant interactions and regional heterogeneities [57,58]. Since the early 21st century, climate change agendas have driven research toward co-abatement pathways for carbon and air pollution, incorporating methods such as Tapio decoupling and LMDI decomposition to quantify dynamic decoupling mechanisms [59,60,61,62].

Decoupling analysis is a key tool for assessing whether economic growth can be achieved without exacerbating environmental pressures, such as carbon emissions or air pollution. The method quantifies the dynamic relationship between economic output (e.g., GDP) and environmental indicators (e.g., CO₂ and VOCs). We applied the Tapio decoupling model ($D$) to analyze the relationships between economic growth and carbon emissions, as well as air pollution, in the GBA from 2005 to 2020 [63]. The formula is as follows:

$$D={\displaystyle \frac{∆E/E}{∆GDP/GDP}}={\displaystyle \frac{\left(\frac{E_i}{E_{i-1}}-1\right)}{\left(\frac{{GDP}_i}{{GDP}_{i-1}}-1\right)}}$$

(16)

In the formula, $i$ and $i-1$ denote the final period and base period, respectively; $∆E$ represents the CAP emission increment; $∆GDP$ denotes the GDP increment; and $D$ is the decoupling elasticity coefficient, reflecting the change trend of CAP emissions relative to economic growth between periods (i.e., decoupling state). Based on different degrees of decoupling, decoupling states can be classified into four categories: strong decoupling, weak decoupling, recessionary decoupling, and negative decoupling. Among these, strong decoupling represents the most desirable state, where economic growth occurs concurrently with declining carbon emissions. Conversely, negative decoupling is the most undesirable state, indicating a maladjusted relationship between economic development and carbon emissions [64].

Table 4. Decoupling state and decoupling index.

Decoupling State	Decoupling Type	Environmental Pressure Growth Rate ($∆\mathit{E}/\mathit{E}$)	Economy Driven Growth Rate ($∆\mathit{G}\mathit{D}\mathit{P}/\mathit{G}\mathit{D}\mathit{P}$)	Decoupling Index, $\mathit{D}$
Decoupling	Strong decoupling (SD)	(−$\infty ,$ 0)	(0, $+\infty$)	$D<0$
	Weak decoupling (WD)	(0, $+\infty$)	(0, $+\infty$)	$0\le D<0.8$
	Recession decoupling (RD)	(−$\infty$, 0)	(−$\infty$, 0)	$D>1.2$
Coupling	Expansion coupling (EC)	(0, $+\infty$)	(0, $+\infty$)	$0.8\le D\le 1.2$
	Recession coupling (RC)	(−$\infty$, 0)	(−$\infty$, 0)	$0.8\le D\le 1.2$
Negative decoupling	Strong negative decoupling (SND)	(0, $+\infty$)	(−$\infty$, 0)	$D<0$
	Weak negative decoupling (WND)	(−$\infty$, 0)	(−$\infty$, 0)	$0\le D<0.8$
	Expansion negative decoupling (END)	(0, $+\infty$)	(0, $+\infty$)	$D>1.2$

presents eight decoupling states for carbon-pollutant emissions.

2.4. Data Sources

2.4.1. CAP Emission Data Sources

The data for this study were derived from the Multi-Resolution Emission Inventory Model for Climate and Air Pollution Research (MEIC, http://meicmodel.org.cn). The MEIC dataset is compiled using provincial energy statistics, unit-based power plant emissions, population density, transportation networks, and updated emission factors. This dataset provides annual CO₂ and VOC emission data for electricity, industrial, residential, and transportation sectors, covering chemical mechanisms such as SAPRC07, CB05, and SAPRC99 [65]. It offers comprehensive and continuously updated data support to ensure timeliness and accuracy.

Version 1.4 of the MEIC database (developed in 2021) was employed, which systematically simulates the impacts of structural adjustments and pollution control measures implemented under China’s Three-Year Action Plan for Winning the Blue Sky Defense War on emission changes. The carbon emission inventory and air pollutant data were updated through 2020. Monthly CO₂ and VOC raster data for electricity, industrial, residential, and transportation sectors from 2000 to 2020 were extracted at a spatial resolution of 0.25°.

To minimize zoning statistical errors caused by raster resolution, resampling operations were performed to reduce grid cell size by 100 times. Using ArcGIS 10.7 software’s zonal statistics function, average emissions for each prefecture-level city were batch-extracted.

2.4.2. Socioeconomic Population Data Sources

The data employed in this study primarily consist of energy, economic, and population datasets. Energy data encompass city-specific energy production, consumption, and structural information across different periods, which are used to evaluate energy utilization efficiency and explore energy transition pathways. Economic data detail the economic development trajectories of cities over the past decade, providing a foundation for studying the synergistic development between the economy and environment. Population data are critical for analyzing energy demand, environmental pollution, and social development, as analyzing population-change trends allows for evaluating the pressures exerted by population growth on energy and the environment.

Energy, economic, and population data were sourced from city statistical yearbooks (2005–2020), the National Bureau of Statistics of the People’s Republic of China (https://data.stats.gov.cn), the Census and Statistics Department of the Hong Kong Special Administrative Region (https://www.censtatd.gov.hk/sc/), and the Statistics and Census Service of the Macao Special Administrative Region (https://www.dsec.gov.mo/zh-MO/).

3. Result

3.1. Analysis of Spatial and Temporal Evolution Characteristics of CAP in the GBA

3.1.1. Spatiotemporal Aggregation Trend of CAP Emissions

(1)

Urban agglomeration scale

Figure 3 reveals the spatiotemporal variations in CAP characteristics in the GBA.

The total carbon emissions (TCE) of the GBA reached 84.433 million tons in 2000, growing to 216.155 million tons by 2020, representing an average annual growth rate of 7.4%. Per capita carbon emissions (PCCE) peaked at 3.8 tons in 2011 before declining gradually to 2.5 tons in 2020. Although the GBA has made progress in reducing per capita emissions, this metric remains higher than that of advanced international bay areas (e.g., California and New York), indicating significant room for improvement in low-carbon development.

Total VOC emissions (TVOC) in the GBA increased from 541,000 tons in 2000 to 1,048,000 tons in 2020, with an average annual growth rate of 4.5%. Despite high aggregate emissions, per capita VOC (PCVOC) levels remained relatively low, primarily attributed to the GBA’s efforts in industrial restructuring and pollution control.

The temporal trends of carbon emissions per unit area (CEUA) and VOC emissions per unit area (VEUA) generally mirrored those of total emissions.

(2)

Urban scale

Figure 4 illustrates the spatiotemporal characteristics of carbon emissions and VOC pollution across the GBA’s 9 + 2 cities. From 2000 to 2020, significant disparities were observed in the urban distributions of total emissions, per capita emissions, and emissions per unit area for both carbon and VOCs.

In the GBA, the nine cities in the Pearl River Delta (PRD) contributed 89% of the total carbon emissions, with Guangzhou, Foshan, and Dongguan having consistently high emissions throughout the study period (Figure 4a). Similarly, VOC emissions in the region are mainly concentrated in the core cities of the PRD, and their spatial distribution bears a remarkable resemblance to carbon emissions (Figure 4d). In addition, there were differences in the trends of carbon emission changes among the cities in the GBA: the total carbon emissions of Guangzhou, Foshan, and Zhongshan gradually stabilized after initial growth; those of Shenzhen and Dongguan rose first and then showed a downward trend; and those of Huizhou and Jiangmen maintained a steady growth throughout the period. In contrast, Shenzhen, Hong Kong, and Macao accounted for a relatively low share of total carbon emissions and VOC emissions, and their emission characteristics were significantly different from those of the other nine cities in the PRD (see Figure 4b,e). The per capita carbon emissions and VOC emissions of Hong Kong and Macao were also at a relatively low level. Comparatively, Dongguan, Foshan, and Zhongshan have higher carbon emissions and VOC emissions per unit area (Figure 4c,f). Future studies should explore in depth the impacts of economic structure and emission reduction policies on CO₂ and VOC emissions in these cities, so as to understand more comprehensively the differences in regional emission characteristics and their evolutionary trends.

3.1.2. Analysis of CAP Emission Key Cities

Regional disparities in urban energy resource endowments and development policies have led to spatiotemporal agglomeration trends in carbon and VOC emissions. Using ArcGIS 10.7, GeoDa 1.14, and R 4.4.3 software, this study computed Moran’s I indices to explore the distribution characteristics of carbon and VOC emissions. The results of Moran’s I index show that the regional aggregation characteristics of carbon and VOC emissions in the GBA can be divided into two distinct phases (Figure 5).

The first stage is before 2014. During this period, the Moran’s I of carbon emissions in the GBA increased rapidly from 0.28 in 2005 to 0.39 in 2014, indicating a significant increase in the spatial aggregation of carbon emissions among cities. Meanwhile, the Moran’s I of VOC emissions increased from 0.31 in 2005 to 0.36 in 2014. The second stage is 2014–2020, during which the Moran’s I of carbon and VOC emissions in the GBA shows a small fluctuating upward trend, reaching 0.41 and 0.36 by 2020, respectively. The spatial correlation of carbon emissions increases slightly in this stage.

To further visualize the spatial agglomeration characteristics and dynamic evolution trends of carbon and VOC emissions, this study conducted a standard deviational ellipse analysis (Figure 6). The left part of each sub-picture shows the standard deviation ellipse and the focus-shift trajectory, the top right image is a zoomed-in view of the standard deviation ellipse, and the bottom right image is a zoomed-in view of the focus-shift trajectory.

Analysis results show that the spatial distribution of carbon and VOC emissions in the GBA exhibited specific directional patterns in 2005, 2010, 2015, and 2020. Specifically, the spatial distribution of total emissions and per capita emissions of CO₂ and VOCs mainly presents an “east–west” pattern, indicating that the high-value areas of emissions have strong extensibility in the east-west direction (Figure 6a,b,d,e). The spatial distribution of emissions per unit area shows a pattern of “southeast–northwest”, reflecting the significant difference in emission intensity per unit area in this direction (Figure 6c,f). In addition, the spatial polarization effect of carbon emissions is stronger than that of VOC emissions; that is, the inter-city difference of carbon emissions is greater than that of VOC emissions. In the GBA, the spatial distribution of carbon emissions is more uneven, and the contribution of some cities to the total carbon emissions is more significant, whereas VOC distributions were relatively more homogeneous.

The area of the standard deviational ellipse expanded continuously from 2005 to 2020, indicating increasing spatial dispersion of carbon and VOC emissions. This phenomenon suggests that high-emission zones are no longer confined to specific regions but are gradually diffusing to surrounding cities. Concurrently, the ellipse’s centroid shifted northwestward overall, indicating a gradual concentration of the GBA’s carbon and VOC emission activities in this direction.

3.1.3. Analysis of CAP Emission Key Sectors

From 2000 to 2020, significant disparities in carbon emissions were observed across major sectors in the GBA (Figure 7).

From a sectoral decomposition perspective over time, the industrial and electricity sectors were primary sources of CO₂ emissions, with their emissions peaking in 2011 before showing a downward trend. This shift was likely closely associated with industrial restructuring, energy efficiency improvements, and clean energy replacement policies implemented in the GBA after 2010 [66]. For instance, the industrial sector reduced energy consumption through technological upgrades and capacity optimization, while the electricity sector decreased coal dependence by increasing natural gas and renewable energy usage.

In contrast, CO₂ emissions from transportation and residential sectors exhibited moderate growth trends, collectively accounting for approximately 20% of the GBA’s total CO₂ emissions (Figure 8). The growth in transportation emissions may be attributed to sustained increases in motor-vehicle ownership, while residential-sector growth reflects rising energy demand driven by population growth and improving living standards.

Figure 9a–c illustrate the dynamic changes in city-level and sectoral carbon and VOC emissions across the GBA in 2000, 2010, and 2020.

Figure 9 shows that Guangzhou, Foshan, and Dongguan consistently ranked among the top three cities in carbon emissions, with their distributions highly aligned with key cities of VOCs. As economic cores of the GBA, their high emissions primarily stemmed from intensive industrial activities, heavy traffic volume, and high energy consumption. From a sectoral perspective, carbon and VOC emissions from the electricity sector declined annually, closely linked to the GBA’s recent energy structure optimization policies—such as increasing natural gas and renewable energy adoption while phasing out high-pollution, high-energy-consuming coal-fired power plants. However, reduced electricity sector emissions did not significantly lower total emissions, instead coinciding with a shift in emission contributions to the industrial sector. The industrial sector has emerged as the primary source of both carbon and VOC emissions in recent years. Rising emissions in this sector may be attributed to rapid manufacturing expansion, increased chemical production activities, and escalating industrial energy consumption.

3.2. Analysis of Driving Factors of CAP Emissions

In order to reveal the regional and sectoral differences of CO₂ and VOC pollution in urban agglomerations and cities in the GBA, this study employed the LMDI model to conduct a driving factor analysis of CO₂ and VOC emissions. The investigation focused on the current situation of industrial energy consumption, industrial energy intensity, industrialization level, proportion of non-industrial energy consumption, regional economic development balance, and population size in 9 + 2 cities and their impact on carbon emission reduction.

3.2.1. Analysis of Driving Factors of CAP Emission in Urban Agglomeration Scale

To analyze the impacts of driving factors on carbon and VOC reductions over 16 years in a more intuitive manner, this study divided the research period into three phases (2005–2010, 2010–2015, and 2015–2020).

First, the LMDI method was employed to quantitatively analyze the contribution levels of the driving factors of CO₂ emissions in the GBA (Figure 10). The results show that the growth rate of CO₂ emissions declined significantly over the period 2005–2020. Among them, it was 35.4% in 2005–2010, and it decreased to 2.2% in 2010–2015, with the first negative growth (−0.2%) in 2020. This trend indicates that the GBA has achieved initial successes in carbon reduction.

Positive driving factors for carbon reduction (promoting emission reductions) include industrial energy intensity, energy intensity, and regional economic development balance. Among these, reduced industrial energy intensity was one of the primary driving factors for CO₂ reductions, accounting for more than 20% of the emission reduction. This indicates that technological upgrades and energy efficiency improvements in the industrial sector significantly reduced energy consumption per unit of output. Declining energy intensity also played a positive role in emissions reductions, with a contribution rate comparable to industrial energy intensity, reflecting overall energy efficiency improvements in the GBA—likely due to clean energy replacement and optimized energy management. Additionally, improvements in regional economic development balance positively impacted carbon reductions during 2010–2015, suggesting that narrowing regional economic gaps and promoting coordinated development helped lower aggregate carbon emissions.

Negative driving factors for carbon reduction (increasing emissions) include industrialization level, urban economic development, regional per capita GDP, and population size. During 2005–2010, the industrialization level did not facilitate emission reductions; instead, it caused a 1.03% increase in carbon emissions. This may be attributed to rapid industrial-scale expansion and a high proportion of energy-intensive industries during this phase. Rising urban economic development levels also increased emissions, indicating that economic growth remains reliant on energy-intensive industries and high-carbon development models. Furthermore, growth in per capita GDP exerted a positive driving effect on emissions, reflecting increased energy demand from improving living standards. Expanding population size further exacerbated emissions, highlighting sustained pressures from population growth on energy consumption and carbon emissions.

Next, the study analyzed the driving factors and contribution levels of VOC emissions (Figure 11). Our results showed that industrial energy intensity, energy intensity, VOC emission intensity, and industrialization level made positive contributions to VOC reductions, whereas economic development emerged as the primary factor driving increased VOC emissions.

Positive driving factors for VOC reduction (promoting emission reductions) include industrial energy intensity, energy intensity, VOC emission intensity, and industrialization level. Among them, industrial energy intensity’s contribution to VOC reductions decreased from 23.58% in 2010 to 6.46% in 2020, indicating that early technological upgrades and energy efficiency improvements in the industrial sector played a significant role, though their impact diminished over time. Energy intensity’s contribution declined from 24.55% in 2005 to 5.40% in 2020, reflecting the positive effect of overall energy efficiency improvements in the GBA on VOCs’ reductions, albeit with decreasing contribution rates. VOC emission intensity’s contribution rose from 13.28% in 2005 to 19.20% in 2020, demonstrating that technological advancements and enhanced end-of-pipe pollution control measures significantly reduced VOC emissions per unit of economic output. The industrialization level’s control effect on VOC reductions gradually strengthened, suggesting that industrial structural optimization and phasing out high-pollution industries positively impacted emissions.

The negative driving factor for VOCs’ reduction (increasing emissions) was economic development, though its contribution rate trended downward. This shift indicates that under the context of low-carbon economic transformation, the GBA’s economic growth has become progressively less dependent on VOC emissions, leading to improved environmental quality.

3.2.2. Analysis of the Absolute Contribution of Driving Factors for CO₂ and VOC Emission Reduction at Urban Scale

Based on data analysis of the GBA’s 9 + 2 cities from 2005 to 2020, this study further quantified the absolute contributions of various factors to CO₂ reductions (Figure 12).

Our results indicate that urban economic development, regional per capita GDP, and regional population size were negative driving factors for CO₂ reduction, significantly increasing carbon emissions. Conversely, industrial energy consumption intensity and energy consumption intensity served as positive driving factors for CO₂ reduction. Among these, energy consumption intensity demonstrated the most significant carbon reduction contribution, nearly offsetting the emission increase caused by economic growth.

Figure 13 illustrates the influence effects of various driving factors on VOC emissions in the GBA’s 9 + 2 cities from 2005 to 2020.

The results of the driving-force decomposition analysis for VOC emissions mirrored those for CO₂ emissions, including urban economic development, regional per capita GDP, and regional population size, which exerted negative effects in VOC reductions in most areas, while industrial energy intensity, energy intensity, and industrialization level had a positive effect on VOC reductions in most cities. Among these, declining energy intensity had the most significant impact on VOC reductions.

In summary, Guangzhou, Foshan, Dongguan, Huizhou, and Zhongshan showed larger negative effect of economic development on both the reduction of both CO₂ and VOCs, indicating their continued reliance on high-energy-consuming and high-emission industrial models, which drive sustained emission increases. Optimizing industrial structures and reducing the proportion of energy-intensive industries represent critical pathways for achieving emission reduction targets in these cities. Moreover, Guangzhou and Shenzhen demonstrated significant positive effects of energy intensity and industrial energy intensity on CO₂ and VOC reductions. These cities effectively reduced emissions while maintaining economic growth through improved energy efficiency and optimized industrial energy consumption structures. Thus, lowering energy intensity serves as an effective entry point for achieving urban pollution and carbon reduction goals, enabling reduced energy use while ensuring economic benefits and advancing green and low-carbon urban development.

3.3. Analysis of CAP Collaborative Driving Factors

3.3.1. Analysis of the Contribution Level of CR to PR

Based on the calculation results of the CAP synergistic emission reduction effect decomposition model (Figure 14), this study quantified the contribution levels of CO₂ reductions to VOC emission reductions.

Our results indicate that the carbon reduction intensity effect exerted a significant positive impact on VOC’ reductions, indicating that reducing carbon emissions can effectively drive down VOC pollution. However, economic development and population size remained the primary driving factors contributing to increased VOC pollution. Regional heterogeneity analysis revealed that carbon reduction intensity had more pronounced pollution reduction contributions in Guangzhou, Foshan, Dongguan, and Shenzhen. From an annual variation perspective, the contribution of carbon reduction to VOC reductions fluctuated between 2005 and 2020. The most significant pollution reduction outcomes occurred during 2012–2013 and 2015–2016, while the pollution reduction effects of carbon reduction were relatively weak during 2010–2011 and 2016–2017.

3.3.2. Analysis of the Contribution Level of PR to CR

Based on the calculation results of the CAP synergistic emission reduction effect decomposition model (Figure 15), this study quantified the contribution levels of VOC reductions to CO₂ reductions.

Our results show that the pollution reduction intensity effect exerted a significant positive impact on CO₂ reductions, indicating that reducing VOC pollution can effectively drive down CO₂ emissions. However, economic development and population size remained the primary driving factors contributing to increased CO₂ emissions. Regional heterogeneity analysis revealed that pollution reduction intensity had more pronounced carbon reduction contributions in Guangzhou, Foshan, Dongguan, and Jiangmen. From an annual variation perspective, the contribution of pollution reduction to CO₂ reductions remained relatively stable between 2005 and 2020. The most significant carbon reduction outcomes occurred during 2010–2011, coinciding with a phase when carbon reduction’s pollution reduction effects were relatively weak.

3.4. Decoupling Analysis

To clarify the decoupling status between economic development and CAP emissions during the economic growth process, this study applied the Tapio decoupling model to analyze the decoupling relationships between economic development and CO₂ and VOC emissions in the GBA’s 9 + 2 cities from 2005 to 2020 (Figure 16 and Figure 17).

Our results show that over the 15-year period, strong decoupling and weak decoupling between economic development and carbon emissions occurred 11 and 1 times, respectively, in the GBA’s 9 + 2 cities, while only 3 instances of strong decoupling were observed for VOC emissions. This indicates that VOC emissions currently face stronger constraints in decoupling from economic growth compared to carbon emissions.

Figure 18 illustrates the decoupling analysis results between CO₂ and VOC emissions and economic growth in the GBA’s 9 + 2 cities during 2005–2010, 2010–2015, and 2015–2020.

The decoupling status of the 9 + 2 cities in the GBA is relatively favorable, with weak decoupling and strong decoupling accounting for a higher proportion—with strong decoupling being the optimal state. Specifically, except for Macao, which experienced a decrease in GDP during 2015–2020, the rest of the regions showed growth in GDP during the study period. During 2005–2010, VOC emissions generally increased, with only Hong Kong showing a negative growth rate in CO₂ emission. Guangzhou and Zhuhai showed a negative growth rate in CO₂ emissions during 2010–2015. Between 2015 and 2020, VOC emissions generally decreased, while CO₂ emissions increased slightly in Guangzhou, Zhaoqing, and Huizhou.

4. Discussion

4.1. The Spatiotemporal Characteristics of CAP in the GBA

Similar to the experiences of most developed countries during early industrialization and urbanization, China’s rapid economic growth has been accompanied by substantial pollution emissions and environmental degradation [2]. Prior to 2011, CO₂ emissions and VOC pollution in the GBA showed annual increases. In 2009, Guangdong, Hong Kong, and Macao jointly formulated the Coordinated Development Plan for the Greater Pearl River Delta Urban Agglomeration, aiming to build a world-class city cluster with global competitiveness. In 2010, the Notice on Launching Low-Carbon Province and City Pilot Projects was released, designating Guangdong Province and eight cities, including Shenzhen, as low-carbon pilot areas. During China’s 11th Five-Year Plan (2006–2010), a target of reducing energy consumption per unit of GDP by approximately 20% was set, laying the groundwork for carbon reduction through measures such as restructuring energy-intensive and high-pollution industries and phasing out backward production capacity.

Under this backdrop, the GBA actively responded to emission reduction initiatives, achieving significant progress in CO₂ and VOC reductions between 2011 and 2020. Specifically, CO₂ emissions peaked in 2011, followed by fluctuating declines from 2012 to 2020. The electricity and industrial sectors demonstrated particularly notable carbon reduction achievements. While transportation and residential CO₂ emissions rose overall, they declined in 2020. VOC emissions fluctuated but also decreased in 2020. Industrial VOC emissions, though increasing in earlier years, also dropped in 2020—a change likely influenced by the COVID-19 pandemic [67,68]. China’s strict epidemic control measures, including self-quarantine and transportation restrictions [69], temporarily suppressed industrial production and transportation activities, thereby reducing atmospheric pollutant emissions.

4.2. Sectoral Emission Variance of CAP in the GBA

This study analyzed changes in CO₂ and VOC emissions across sectors in the GBA from 2000 to 2020. Results indicate significant disparities in emission trends among industries, likely closely associated with sectoral energy structures, production processes, and policy interventions [65].

Industry is the primary source of both CO₂ and VOC emissions in the GBA, making it a key focus for pollution reduction and carbon mitigation, which aligned with those reported in previous studies [70]. Synergistic emission reductions can be achieved through technological innovation and industrial upgrading to build efficient, low-carbon, and circular industrial systems that reduce energy and resource consumption at the source [71]. While the electricity sector contributes minimally to VOC pollution, its CO₂ emissions account for over 35% of total emissions [72]. Expanded development and utilization of clean energy, combined with improved energy conversion efficiency, are critical for this sector. Additionally, transportation and residential sectors are significant sources of carbon and air pollutants, underscoring the need for comprehensive governance measures.

Overall, the GBA has achieved notable progress in CO₂ and VOC reductions, but it still faces many challenges. There is a need to further strengthen policy guidance and technology research and development, and to formulate more targeted emission reduction strategies for different industries. For the industrial and electric power industries, we should continue to promote the optimization of energy-structure and production-process upgrading, and strengthen the monitoring and treatment of VOC emissions. For the transportation and civil sectors, it is necessary to accelerate the popularization of new energy vehicles, raise residents’ awareness of environmental protection, and promote green travel and lifestyles, so as to achieve the goal of continuous improvement of air quality and sustainable development in the GBA.

4.3. Synergistic Driving Factors and Decoupling Status of CAP in the GBA

Industrial energy intensity, energy intensity, VOC emission intensity, and industrialization level made positive contributions to CO₂ and VOC reductions in the GBA, while urban economic development, regional per capita GDP, and population size acted as negative driving factors for emission reductions across the GBA and its cities. Among these, Guangzhou, Foshan, Dongguan, and Huizhou—key carbon-pollution-emitting cities—exhibited larger contributions of economic development to CO₂ and VOC emissions. Rapid economic growth and population expansion, such as through large-scale infrastructure construction, industrial production, and household consumption, led to substantial energy consumption and subsequent increases in carbon emissions and air pollution [73].

The significant contribution of VOC reductions to CO₂ reductions during 2010–2011 indicates that stringent VOC emission reduction measures have played an important role in driving CO₂ reduction. Moving forward, strengthened implementation of VOC reduction policies is essential to ensure their alignment with carbon reduction goals. The relatively stable contributions of VOC reductions to CO₂ reductions in other years demonstrate the sustained effectiveness of pollution control measures in promoting CO₂ reductions. Future efforts should focus on optimizing pollution reduction measures—such as promoting clean production technologies and low-VOC raw material substitutions—to enhance their leveraging effect on CO₂ reductions. Overall, a synergistic relationship exists between CAP reduction in the GBA. Notably, Guangzhou, Foshan, and Dongguan, as key CO₂ and VOC emitters in the GBA, achieved the most significant synergistic emission reduction outcomes, though economic development and population size remain primary barriers to integrated emission reductions.

Additionally, the GBA has generally achieved relative or absolute decoupling between economic growth and CAP emissions, reflecting the effectiveness of energy efficiency improvements, clean energy adoption, and end-of-pipe pollution control measures. However, some cities still face expansive coupling or expansive negative decoupling between economic growth and emissions, indicating that their economic growth models require further optimization to achieve more efficient green and low-carbon development.

4.4. Limitation

This study has several limitations. In terms of data timeliness, the CO₂ emission data in the MEIC database had been updated up to 2021, whereas the VOC emission data only cover up to 2020. Given the synergistic effect of CO₂ and VOCs, it is essential to maintain consistency in the statistical caliber of the data. Consequently, the data source for this study spans from 2005 to 2020, and we will continue to monitor and update the data in a timely manner. Although the contribution of PR to CR was quantified based on the decomposition model of CAP synergistic emission reduction effects, some factors (intensity of industrial pollution control, etc.) could not be included in the analysis due to the lack of data on pollutant charging and sewage charges for specific industries in Hong Kong and Macao. Future research could improve analytical precision through field monitoring and multi-model comparisons to address these uncertainties. Additionally, this study did not fully account for the specific implementation effects of policies. Future research could expand the framework of drivers of synergistic emission reduction to comprehensively examine the influencing factors and their synergistic effects of synergistic emission reduction of CAP in urban agglomerations. Additionally, considering policies as a carrier of energy socioeconomic factors, the impacts of policies on energy production, consumption, and socioeconomic development can be assessed in depth by quantifying the effects of policies and their implementation.

5. Conclusions and Policy Proposal

This study analyzed the spatiotemporal distribution characteristics, urban and sectoral evolution patterns, emission reduction driving factors, synergistic effects, and decoupling status of CO₂ and VOCs in the GBA’s 9 + 2 cities from 2000 to 2020. CO₂ and VOC emissions in the GBA exhibited significant spatiotemporal disparities. The nine Pearl River Delta cities served as the primary emission sources within the GBA, with Guangzhou, Foshan, and Dongguan consistently recording higher total emissions than other cities. The spatial distribution of carbon-pollutant emissions demonstrated a clustered pattern, with the CO₂ emission centroid shifting northeastward from Southern Guangzhou, while the VOC emission centroid migrated westward from the Guangzhou–Dongguan border into the interior of Guangzhou. Second, the industrial sector constituted the dominant driving factor in both CO₂ and VOC emissions in the GBA. Reducing both energy intensity and industrial energy intensity emerged as critical a pathway for achieving emission reductions. Notably, Guangzhou and Shenzhen require particular attention to the impacts of energy intensity and industrial energy intensity on carbon emissions, though their effects on VOC emissions were relatively minor. Against the backdrop of rapid urban agglomeration development, economic growth exerted a significant driving effect on CO₂ and VOC emissions, particularly in Guangzhou, Foshan, Dongguan, Huizhou, and Zhongshan, where urban economic expansion is more prone to emission increases. Finally, the GBA’s 9 + 2 cities generally achieved a favorable decoupling status between economic development and emissions, reflecting sustainable development trends in the region. However, some cities still face expansive coupling or expansive negative decoupling between economic growth and emissions, underscoring the necessity for further optimization of their economic growth models.

The main policy proposals for achieving synergistic CO₂ and VOC reductions are as follows:

(1)

Address the negative driving effects of economic development and population size on emissions by promoting decoupling between economic growth and CO₂/VOC emissions. Specific measures include optimizing industrial structure, developing green industries, promoting low-carbon technologies, and strengthening end-management measures.

(2)

Leverage successful pollution and carbon reduction experiences from Guangzhou, Foshan, Dongguan, and Jiangmen for replication in other cities. Enhance regional collaborative governance to achieve holistic improvements in environmental quality.

(3)

Expand clean energy adoption, optimize industrial structures, and improve energy efficiency to realize win-win outcomes between economic growth and environmental quality improvement.