How does green transformation of industrial sector affect carbon emissions? Evidence from China

Abstract

Industry is the core sector with the most challenging task of carbon emission reduction in China. This paper explores how industrial transformation in China’s industrial sector affects carbon emissions. We shed light on the effect of industrial sector transformation on carbon emissions, and divide industrial sector transformation into two aspects: industrial structure optimization and industrial spatial layout.We construct socioeconomic panel data for 30 cities in the urban agglomeration in the middle reaches of the Yangtze River from 2000 to 2020, and explore the influence of industrial transformation on carbon emissions in 33 industrial sectors in China from three dimensions: spatial spillover effect, spatiotemporal heterogeneity, and threshold effect. The results show that: (1) In the industrial structure optimization dimension, industrial rationalization has effect on decreasing carbon emissions, while industrial upgrading increases carbon emissions. In the industrial spatial layout dimension, industrial specialization agglomeration can decline carbon emissions, while industrial diversification agglomeration does not reduce carbon emissions but instead increase carbon emissions.(2)Both industrial structure optimization and industrial spatial layout have spatiotemporal heterogeneous impacts on carbon emissions during the study period.(3) Under different levels of economic development, industrial transformation will have different impacts on carbon emissions. In the future, the Chinese government should increase efforts to promote technological progress and innovation, taking technological progress and independent innovation capabilities as the central link in promoting industrial transformation. Those findings not only provide certain environmental research reference for China’s industrial economic development and transformation, but also provide a practical reference for carbon reduction in other developing countries that are undergoing economic transformation.

Introduction

Economic development has come at the cost of environmental pollution in many developing countries [27, 30, 39]. Like many developing countries that rely on resource inputs to develop their economies, China has experienced a rapid industrialization process and successfully developed from a large agricultural country to a world industrial power, and created a miracle of economic development [16, 24, 25]. Since 2000, the added value of China’s manufacturing industry has grown at a yearly rate of 21.1%, and its proportion of GDP has remained at around 27.8% [30]. However, the rapid development of industrialization is largely based on resource consumption and environmental pollution. In particular, the development of industry has increased the energy demand for high-carbon sectors such as coal and oil, which has resulted in a great deal of carbon emissions (CE), leading to severe environmental problems. China Carbon Accounting Database (CEADS) statistics that from 1997 to 2019, China’s total CE increased every year, with an annual rate growth of 2.99%. At the end of 2019, China’s total CE was nearly 9.8 billion tons [30]. It presents a significant challenge for China to deal with the contradiction between economic development and energy consumption.

For developing economy without polluting the environment, the Chinese government has proposed to promote industrial sector transformation(IST), aiming to reduce CE through industrial transformation(IT), and reduce environmental pollution while ensuring sustainable economic development [17, 29]. Can China’s IST develop the economy while protecting the environment?

In recent years,many studies have focused on CE effects on land use transformation derived from industrial transformation (IT) through internal transformation of farmland and change of land use structure caused by expansion of construction land [3, 15, 19, 32, 35, 38]. These studies analyzed CE from land transformation through changes in carbon sources and sinks,changes in land use structure, IT induced internal farmland and change to construction land. Qu et al. [31] argued that land use change derived from IT is the second biggest root of greenhouse gas emissions which is only second to energy. A lot of CE will be generated in the process of converting farmland to construction land, and in the process of converting forest to farmland and grassland. They analyzed impact of land conversion within farmland, farmland to non-agricultural land, and land conversion within non-farm land on CE. Some studies findings show that farmland change to non-agricultural land will significantly increase CE, and the conversion of secondary industrial land to tertiary industrial land will reduce CE. Arneth et al. [2] believes that the historical CE caused by land transformation and IT may be greater than assumed, and afforestation and avoiding further deforestation are important approaches to mitigate CE. Noojipady et al. [26] employed satellite data on farmland expansion, woodland cover and vegetation carbon storage to examine the forest CE generated by farmland expansion in the Cerrado biome in Brazil. They found that land use change and forestry during the IT process accounted for two-thirds of Brazil’s CE. Many studies have been undertaken to examine IT and the CE of land transformation. However, there are limitations associated with these studies.First, in terms of theoretical logic, existing studies focus on exploring the relationship between IT and CE. Few studies have attempted to use land as vehicle to build a fully logical system to understand the influence of land transformation due to IT on CE from the integrated economic-environmental system. Land is the carrier for industrial transformation and CE which carries various socioeconomic activities of human beings. IT has triggered transformation and spatial reconstruction of land use. Incorporating IT into CE analysis builds a logical analysis framework for CE derived from land transformation due to IT. This enables a more comprehensive and in-depth analysis of how IT achieves goal of protecting environment. Second, existing studies predominately focus on CE from agricultural land transformation under natural ecosystems. By contrast, CE of construction land use under socioeconomic systems have been largely overlooked. In fact, CE process and mechanism of land use under socioeconomic system are much more complicated than those under natural ecosystem. The impact of land use on CE needs to be conveyed through multiple factors, and there are differences between various factors that affect CE. As a crucial carbon source, construction land needs be analyzed in conjunction with all aspects of socioeconomic development when analyzing its CE effects. However, the current studies on CE from construction land mainly focus on carbon accounting of energy consumption, while CE effects of land use changes are largely overlooked. Third, existing studies mostly analyze influence of IT from aspect of industrial structural optimization (ISO) or industrial spatial layout(ISI). Few studies have explored the influence of IT from two dimensions of ISO and ISI. Investigating effect of IT from the two dimensions of ISO and ISI can provide a more in-depth analysis of the impact of IT, which made the research results more valuable.

To fill these gaps, this paper chooses 30 prefecture-level cities in the urban agglomeration of the middle reaches of the Yangtze River in China (UYR) as the research object, and examines the specific impact of China’s IST on CE by establishing panel data from 2000 to 2020.

This paper makes contributions in two dimensions. First, we develop the theoretical mechanism of how IST affects CE by changing land transformation. We constructed a theoretical framework for the impact of IST on CE and analyze CE effects derived from land use changes under socioeconomic system with IST, which can more comprehensively explore CE effects derived from IT. Secondly, IST is analyzed from two dimensions of ISO and ISI, and the impact of IST on CE is analyzed from three dimensions: spatial spillover effect, patiotemporal heterogeneity and threshold effect. As a result, the process of IST’s carbon reduction effect can be examined from a dynamic perspective, especially from different stages of economic development.

Theoretical analysis and research hypotheses

The natural CE effect is the carbon cycle of soil and forest vegetation in the natural ecosystem assuming there is no human socio-economic activity. In this process, the entire ecosystem can automatically maintain the carbon cycle balance on the land.The direct CE effect of land use refers to the impact on carbon storage of soil and vegetation on the land, which corresponds to the natural ecosystem, while the indirect CE effect of land use corresponds to the social and economic system. This kind of land use CE is near to economic development, targeting CE generated from construction land utilization [6, 8, 9, 13].

In the IST process, land is not only an important spatial carrier of economic activities, but also a carrier that directly and indirectly drives CE.The economic and social CE effect refers to the CE produced by various activities carried out by humans on the land. It is the CO2 produced on the land by other factors using the land as the carrier. This type of CE is defined as indirect CE effects.The CE effect caused by IST analyzed in this paper mainly refers to the economic and social CE effect caused by land transformation in the IST.

Proximity distance, convenient transportation and regional cooperation promote the correlation of economic activities between cities, resulting in the phenomenon of spatial dependence and generating spatial spillover effects [1, 4, 19].

On the one hand, IST promotes the reorganization of input factors in land-based industries and flows from sectors from low to high productivity, promotes the flow and distribution of factors between enterprises and regions, and produces industrial correlation effects [45]. Each region has increased technology sharing and regional cooperation in the IST process, strengthened the sharing of advanced technologies among regions, and promoted the carbon reduction effect of IST among neighboring regions. On the other hand, similar location conditions and development levels between neighboring regions can easily lead to cooperation mechanisms. Regional enterprises form integrated groups through mutual cooperation, promote the interoperability and integration of production factors on the land, and promote technology sharing between regions. In the context of the in-depth development and widespread application of advanced technology, media just like the internet,makes the spatial connection between cities closer and closer, and promotes the fuller flow of factors just like advanced technology and scientific researchers with advanced productivity across regions. Each region takes its local comparative advantages and forms an economy of scale in which one plus one is greater than two [37]. This process promotes the diffusion effect of IST between adjacent regions. This diffusion effect further optimizes the cross area allocation efficiency of capital, labor and other resources within region, reduces the sharing costs between enterprises, and realizes inter-regional integration. Resource recycling not only affects the CE effect of the local region, but also affects the CE effect of neighboring regions. The IST process promotes the mobility of capital, labor, land and other factors on land and the spillover of knowledge and technology to form cross-regional flows and allocations. Neighboring areas with a higher level of development provide advanced public infrastructure and service sharing for surrounding areas and form a "demonstration effect" of diffusion.Developed regions promote circulation of production factors within the region through demonstration and in underdeveloped regions through imitation, thereby improving the imbalance of production factors among regions and promoting the sharing of capital, labor and technology by reallocating internal resources in the area [43]. With the continuous circulation of factors and regional cooperation, a “demonstration effect” is produced on neighboring cities and larger cities, which not only affects the CE effect of the region, but also affects the CE effect of neighboring regions (Fig. 1).

Fig. 1.

Theoretical analysis framework of spatial spillover effects of IST on CE

From the above analysis,we put forward research hypothesis 1: IST drives factor circulation among areas, causing circulation of factor resources such as labor, capital, technology and data between regions, and has a spatial spillover effect on CE reduction.

Lin Yifu believes that the prime industrial structure is one that can maximize the “competitive advantage” of the economy in domestic and foreign markets [20]. The optimal industrial structure is endogenous to factor endowments and evolves dynamically over time[21]. In different historical periods, due to differences in resource endowments and economic, the impact of IST on CE varies over time and space.

From the perspective of temporal heterogeneity, IST is a continuous and cyclical evolutionary process, and different evolutionary stages will show different characteristics. The theory of industrial life cycle believes that development of an industry has different cycles such as emerging, developing, mature and declining. Therefore, formation and evolution of an industry have temporal and spatial specificity and differences. The initial state of an industry will be continuously strengthened through the accumulation of causal processes and show different characteristics at different development stages. Affected by external environment and internal growth capacity, IST will show obvious differentiation characteristics in different periods [44]. In the early period of industry life cycle, the effect of industrial transformation on CE is relatively limited. As the industry gradually grows and enters the mature stage, IST gradually improves, and its effect on CE is increasing. When industrial development enters a recession, effect of IST on CE may gradually weaken.

Specifically, different stages of industrial development will lead to different land use structures. In the pre-industrial stage, primary industry has the largest proportion in the industrial structure, agricultural land accounts for the largest ratio in land use structure, and non-agricultural land in the city accounts for a smaller proportion.After entering the industrialization stage, rapid growth of urbanization and industrialization has promoted the development of secondary and tertiary industries. Many people have moved from rural areas to cities, and ratio of first industry has progressively declined. The ratio of urban non-agricultural land in the land use structure has continued to increase, and agricultural land has sharply decreased.

Under rapid economic growth, after entering the post-industrial stage, the ratio of industrial land in the urban land has gradually decreased, and tertiary industry has progressively replaced industry as primary driver for urban economic development. The number of non-agricultural land such as transportation land, residential land, and commercial land that are matched with it has increased, and these land types occupy the core area of the city center.

Therefore, at different stages of industrial development, land use will present different structures, and different land structures will bring different inputs and outputs, which will have different impacts on CE.

From the view of spatial heterogeneity, CE reduction effect brought about by IST in cities in the same administrative unit and different administrative units is different.

During the IST process, due to regional homogeneity, homogeneous industries in prefecture-level cities in the same administrative unit are prone to resource sharing, resulting in industrial linkages and forming advantageous industrial chains [7, 16, 24, 30]. The advent of industrial chains is helpful to the formation of reasonable industrial distribution and regional division of labor. Each region gives play to its own comparative advantages, obtains specialized advantages and innovation, and then promotes technological progress and reduces CE. In addition, it is easier to form a mimicry effect between the same regions, resulting in a strong “demonstration effect”. Local governments in economically underdeveloped regions gradually abandon low-profit industries with comparative advantages, and blindly “imitate” industrial structure and development model of neighboring developed regions, resulting in repeated investment and construction of “high-profit industries”, which will cause overcapacity and industrial isomorphism. At the same time, the protectionism awareness of local governments in the same administrative region is stronger [11, 12, 17]. In order to compete for limited resources, local governments hinder the outflow of various factors, deliberately raise the entry conditions of similar enterprises or products from other places, and even completely prohibit foreign products from entering local market, taking active protective actions against local enterprises or products, resulting in regional market segmentation.This forced segmentation behavior hinders flow of factors and commodities between regions, forms trade barriers, leads to local resource surplus, which is not helpful to further adjustment of local ISI, which causes extensive land use and increases CE.

Neighboring prefecture-level cities that are not in the same administrative unit are conducive to supplementing local industrial advantages and breaking local administrative boundaries. The production factors brought about by industrial transformation can flow and combine across regions in a larger range, promoting mutual penetration and gradual integration between regions [11, 30]. Horizontal cross-regional resource flow can boost the prime fusion of production factors, promote labor mobility, capital mobility, and technology mobility, facilitate the sharing of production resources just like technical equipment, and form a regional specialization division of labor based on effective resource allocation and overall interest maximization. In addition, for sake of economic interests and factor resources, local governments often transfer industries that are not suitable for their own utility preferences to other regions [39].This process not only accelerates transformation of regional individual economic development methods, but also promotes industrial categorization and cooperation between areas, improves intensive land use, and reduces CE.

On the contrary, the competition for factor resources between neighboring cities that are not in the same administrative region will affect the rational cross-regional flow of production factors, resulting in the spatial dispersion and uneven distribution of industrial resource allocation and unreasonable regional division of labor, leading to inefficient distribution of industrial land resources, extensive land use, and increased CE (Fig. 2).

Fig. 2.

Theoretical analysis framework of temporal and spatial heterogeneity effects of IST on CE

In light of the above, this paper presents hypothesis 2: The effect of IST on CE is different in periods and regions, and there is temporal and spatial heterogeneity.

Evolutionary economic geography theory believes that economic activities are always in dynamic change, and the development of things not only depends on the current state of things, but also shows dependence on historical paths. Different levels of economic development will bring different resource consumption, market vitality and technological innovation levels [30, 34], and this will cause IST to have different effects on CE.

From the perspective of resource consumption, under different levels of economic development, enterprises on different land types will consume different resources during the production process.During economic development, population size, economic aggregate, economic investment amount and urbanization level will cause changes in resource consumption. With the rapid development of economy, promotion of clean and sustainable energy technologies has encouraged new energy to gradually replace traditional energy, prompting a greater improvement in the energy structure with large part of fossil energy, and application of resources has gradually decreased, clean energy-saving production resources gradually replace high energy-consuming resources. Different resource consumption will directly impact the input–output structure of land, and then indirectly impact IST and CE reduction.

From the perspective of technological innovation, technological innovation is the core boosting force for enterprises to save energy,improve production efficiency and reduce CE. When technological level is low, carbon reduction effect of IT driven by technology level is smaller. As economy develops, it promotes the progress in technological innovation and the popularization of the transformation and application. Technological innovation has strengthened the impact of IT on CE by reducing the research costs of enterprises, lowering development risks and accelerating technology diffusion and inhibitory effect. Economic development continues to promote scientific and technological progress, and enterprises’ consumption of resources such as fossil energy and production materials is gradually reduced. At the same time, advanced technology will increase the expected output of enterprise production, which will promote CE reduction effect of IST. On the contrary, the CE reduction effect of IST becomes weaker.

From the perspective of market vitality, as market entities and the main body of CE reduction, enterprises often focus on carrying out production activities on land with greater market vitality, and continue to conduct innovation based on their own situation in market competition, thereby enhancing their competitive advantages. The improvement in economic development has led to an increase in market vitality, competition among enterprises has gradually intensified, forcing inefficient enterprises to withdraw from the market. Only by continuously innovating technology and improving productivity can enterprises have a place in the market competition, and continuous innovation in technology and improvement in productivity provide a breakthrough for cities to reduce CE. With the increase of market entities and market vitality, strong demand and smooth flow of goods will generate more supply and sales of goods, thus bringing higher profits to enterprises, and enterprises will have more funds to achieve low-carbon technology research and green production to reduce CE. At the same time, with the improvement of market vitality, intercommunication of production factors just like labor, technology, capital and data on land has further accelerated, especially the “spillover effect” caused by intercommunication of production factors just like knowledge, technology. The potential of technological innovation in the industry changes the input and output scale of factors, reduce factor inputs and increase energy conservation and desired output, which will have a significant effect on reducing CE. Market vitality also creates a good market environment for optimizing the efficiency of enterprise production factor allocation and promoting green economic transformation. When the market vitality is low, the impact of IT on CE will also be smaller. As the economy develops and improvement of market vitality, the important influence of market in resource allocation has been fully exerted, and IT can reduce CE more effectively (Fig. 3).

Fig. 3.

Theoretical analysis framework of the nonlinear relationship between IST and CE

Based on the above analysis, we bring up hypothesis 3: Different levels of economic development bring distinct resource consumption, market vitality and technological innovation levels, which makes the impact of IST on CE a non-linear relationship.

Research region, data and method

Research region and data

Research region

The research region is China’s urban agglomeration in the middle reaches of the Yangtze River (UYR), containing Hunan, Hubei, and Jiangxi three provinces. The study region within 30 prefecture-level cities. UYR is an important growth pole for China’s high-quality economic development, and a pioneer area for new urbanization in China’s central and western regions. Due to lack of some data in Tianmen City after 2010, considering the completeness of data and rigor of this study, Tianmen City was not included in the study area.

Data sources

The CE data get from China Carbon Accounting Database (https://www.ceads.net.cn), industrial transformation data come from the statistical yearbooks, and control variable data get from “China Cities Statistical Yearbook” and bulletins on socioeconomic development of every city[30].

Methods

Measurement of CE

Because most of China’s current CE come from fossil energy CE, drawing on existing research [30], the CE accounting in this paper refers to CE generated by fossil energy in production, CE data are from China Carbon Accounting Database (CEADs) https://www.ceads.net.cn/ collated by the team of Tsinghua University’s Professor Guan Dabo [30].

1

In Formula 1, stands for CE of each city; on behalf of CE of different varieties of energy consumption, such as coal and coke; represents consumption of distinct varieties of energy, just like coal and coke; stand for standards of different types of energy,like conversion coefficient of coking coal; stands for CE factors of different types of energy [30].

Calculate of industrial rationalization in industrial sector (ITL)

This paper divides ISO into industrial rationalization(ITL) and industrial upgrading(IW), while ISI is divided into industrial specialization agglomeration(IKSL) and industrial diversification agglomeration(IDIV).Four industrial transformation measurement index systems were constructed. Citing[10, 30], this paper uses Theil’s coefficient to measure ITL:

2

stands for industrial increased output of i industrial field; stands for gross production worth of thirty-three industrial sectors (Table 1); stands for quantity people in industrial field; is total of labor in thirty-three industrial field; and stands for the quantity of industrial fields. The litter the value, the better the structure of industrial sector is.

Table 1.

Basic information from 33 industrial sectors selected in this paper

Classification
Ferrous metal mining and beneficiation	Paper and Paper Products Industry	Specialized equipment manufacturing industry
Non-ferrous Metal Mining and Dressing Industry	Printing and Recorded Media Reproduction Industry	Transportation equipment manufacturing industry
Non-metallic Mineral Mining and Dressing Industry	Cultural, Educational, Arts and Crafts, Sports and Entertainment Products	Electrical machinery and equipment manufacturing
Agricultural and Sideline Food Processing Industry	Petroleum, Coal and Other Fuel Processing Industry	Computer, communication and other electronic equipment manufacturing industry
Food Manufacturing Industry	Pharmaceutical manufacturing	Instrumentation Manufacturing
Wine, Beverage and Refined Tea Manufacturing Industry	Rubber and Plastic Products Manufacturing Industry	Other manufacturing industries
Tobacco Products Industry	Non-metallic Mineral Products Industry	Waste resource comprehensive utilization industry
Textile Industry	Ferrous Metal Smelting and Rolling Processing Industry	Electricity and heat production and supply industry
Textile, Clothing and Apparel Industry	Non-ferrous Metal Smelting and Rolling Processing Industry	Gas production and supply industry
Leather, Fur, Feathers and Their Products	Metal Products Industry	Water production and supply industry
Wood Processing and Wood, Bamboo, Rattan, Palm and Grass	General equipment manufacturing	Furniture manufacturing

Calculate of industrial sector upgrading in industrial sector (IW)

We used high-tech sectors to explain output of industrial to behalf IW[30].High-tech sector contained instrumentation, medicine, electronics, communication equipment, computers and office equipment, aerospace equipment and machinery.

3

IO stands for yield of industrial sector, IH stands for yield of high-tech industry. The bigger IW, the larger tproportion of high-tech industry in industrial and the bigger level of IW.

Quantification of industrial specialization agglomeration in industrial sector (IKSL)

4

stands for industrial sector, stands for city, and stands for quantity person. The larger , the bigger level of .

Quantification of industrial diversification agglomeration in industrial sector (IDIV)

This paper uses the Herfindahl index to judge IDIV [30]:

5

In the equation, items are same as . The bigger , the more dispersed industrial [30].

Spatial durbin model of IST on CE

Effect of ISO on CE

6

We use the spatial Durbin model to study the influence of ITI on CE. is spatial autocorrelation index,and means the spatial weight matrix. What’s more, , RKMD,, CZHL, FDI are control variables.、 are vectors of their respective regression coefficients; is the error term.

Effect of the ISI on CE

7

Variables meaning are same as Formula 6.

IT Effect on the spatio-temporal variability of CE

This paper uses geographically time-weighted regression (GTWR) model to explore the spatio-temporal diversity of CE effect by IT at distinct time and space.

Threshold effect model of IST on CE

8

9

In the formula, t and s represent time and area; represents CE, are industrial rationalization and industrial upgrading, (*)is the threshold of the function; are the coefficients of the explanatory variables under different threshold intervals, is the threshold variable; represents other control variables except the explanatory variables; is the threshold value, is the sample individual effect; is the disturbance term.

10

11

In the formula, and represent time and area; represents CE, are industrial specialization and diversified agglomeration, I(*) is the threshold representing function; are the coefficients of the explanatory variables under different thresshold intervals, is the threshold variable; represents other control variables besides the explanatory variables; is the threshold value, is the sample individual effect; is the disturbance term. Descriptive statistics for all variables are reported in Table 2.

Table 2.

Descriptive statistics of all variables used

Variable	meaning	Mean	Standard deviation	Minimum	Maximum
lnCE	Carbon emissions	2.674	0.954	0.41	4.849
ITL	Industrial sector rationalization	0.097	0.083	0.004	0.549
IW	Industrial upgrading	0.147	0.073	0.007	0.522
IKSL	Specialized industrial clusters	0.767	0.215	0.091	1.441
InIDIV	Industrial sector diversification and clustering	2.219	0.435	0.747	2.924
InGDP	Economic development level	15.888	1.072	13.198	18.905
InRKMD	Population size and distribution	5.952	0.388	5.176	6.893
FDI	Foreign investment utilization level	0.023	0.016	0.0001	0.089
KJB	Investment in science and technology	0.005	0.007	0.00001	0.057
CZHL	Urbanization rate	0.474	0.131	0.081	0.843

Result

Spatial spillover impacts of IST on CE

Table 3 reports effects of ISO on CE under three spatial weight matrices. The three model’s rho are 0.042, 0.151, and 0.091, and all have passed at least the 10% significance and have spatial spillover impacts. *, **, and *** denote statistical significance at the p<0.10, p<0.05, and p<0.01 levels, respectively. The same applies to all subsequent tables.

Table 3.

Spatial spillover effects of ISO on CE

Explained variable CE	Adjacent distance				Economic distance		Economic geographical nesting
spatial autoregressive coefficient rho	0.042** (2.660)				0.151* (1.930)		0.091*** (-3.030)
	Direct Effect	Indirect Effect	Total Effect	Direct Effect	Indirect Effect	Total Effect	Direct Effect	Indirect Effect	Total Effect
ITL	0.561** (-2.320)	0.757** (2.020)	1.198*** (-2.750)	0.733*** (-2.980)	0.082* (-1.750)	2.398*** (-2.610)	0.445* (-1.690)	1.065** (-2.260)	0.681*** (-3.710)
IW	0.585** (2.440)	1.813** (2.180)	0.559*** (2.780)	-0.367 (1.460)	1.048*** (-3.130)	2.398 (-1.610)	0.327 (1.280)	1.757* (-1.720)	1.048 (-1.390)
FDI	-1.775 (-1.320)	10.521*** (2.950)	8.746** (2.350)	0.735 (0.530)	4.572*** (2.760)	5.308*** (2.680)	-1.166 (-0.810)	6.164 (1.450)	4.998 (1.260)
InGDP	0.521*** (7.290)	-0.226 (-1.360)	0.295* (1.900)	0.434*** (7.140)	-0.039 (-0.540)	0.395*** (6.050)	0.541*** (7.390)	-0.341** (-2.370)	0.200 (1.410)
InRKMD	0.609* (1.960)	5.925*** (5.640)	6.534*** (6.360)	0.896*** (2.900)	0.784* (1.930)	1.680*** (3.830)	1.086*** (3.510)	1.667 (0.970)	2.753* (1.650)
KJB	-4.628** (-2.190)	-0.534 (-0.200)	-5.162** (-2.470)	-5.168*** (-2.630)	0.094 (0.050)	-5.074*** (-2.710)	-3.523 (-1.590)	-1.249 (-0.490)	-4.772*** (-3.170)
CZHL	0.280 (0.940)	-1.934* (-1.900)	-1.654 (-1.590)	-0.227 (-0.710)	0.923** (2.310)	0.696 (1.360)	-0.036 (-0.110)	1.587 (1.360)	1.552 (1.320)
R2	0.690			0.711			0.713

Effect of ITL on CE

Overall, from 2000 to 2020, there was a positive correlation between ITL and CE, indicating that ITL can reduce CE. Under the adjacency distance weight matrix (ADW), the ITL optimized by 1%, and CE increase by 1.198%, that is ITL decreased by 1%, and CE decreased by 1.198%.Under the economic distance weight matrix(EDW), the rationality of ITL optimized by 1%, and CE decreased by 2.398%. Under economic geography nesting weight matrix(EGW), the rationality of ITL is optimized by 1%, and CE decreased by 0.681%. The litter of ITL, the more reasonable the utilization of production factors such as talents and technology, the more conducive it is to adopt advanced production technology in the production field and exploit more clean energy, thereby reducing ratio of energy-intensive sectors like oil, natural gas, and coal within the industry and reducing the CE.

At the same time, regression results show that ITL has played a role in CE reduction with direct, indirect and total effects, it indicates that ITL has reduced CE in the local region while also reducing CE in the neighborhoods. The CE reduction effect of ITL has a spatial spillover impact [40, 44].

Effect of IW on CE

IW has no effect in reducing CE. Under ADW, EDW and EGW, IW increased CE.Under ADW, CE increased by 0.559% for 1% increase in IW. Under EDW, CE increased by 2.398% for 1% increase in IW. Under EGW, CE increased by 1.048% for 1% increase in IW.Under three weights, IW increased CE in both local and neighboring areas.

Under the influence of ISO, three control variables of FDI, GDP and RKMD all increase CE, and indirect effect of FDI on CE is bigger than direct impact. According to pollution haven theory and race to the bottom line of environmental standards hypothesis, influx of foreign investment will aggravate environmental pollution and cause environmental pollution. GDP requires the consumption of large amounts of fossil fuels, which increases CE. And it increases CE of local region, but reduces CE of surrounding regions to a certain extent. This may be because economic development of local area is driven by transfer of polluting enterprises in adjacent area. The transfer of polluting firms in surrounding regions to local area increases CE of local area and reduces CE of neighboring region. The advanced production factors brought by urbanization can reduce CE.Excessive RKMD will drive up CE through surging resource demand and environmental overload. KJB reduce CE by improving energy efficiency and promoting clean energy alternatives.CZHL drives the transformation of industries towards low-carbon service industries and high-end manufacturing, replacing energy-intensive industries and reducing CE,this is consistent with the findings of Huang et al.,[14].

Effect of ISI on CE

Table 4 showed effects of ISI on CE under ADW, EDW and EGW. Coefficients of models are 0.151, 0.091 and 0.268, which also show spatial spillover impact. From 2000 to 2020, the impacts of ISI on CE in UYR are opposite.

Table 4.

The spatial spillover effect of ISI on CE

Explained variable CE	Adjacent distance				Economic distance		Economic geographical nesting
spatial autoregressive coefficient rho	0.151* (1.850)				0.091** (2.510)		0.268* (-1.830)
	Direct Effect	Indirect Effect	Total Effect	Direct Effect	Indirect Effect	Total Effect	Direct Effect	Indirect Effect	Total Effect
IKSL	-0.104 (1.070)	-0.063 (0.240)	-0.181 (0.610)	0.072 (0.720)	-0.243** (-2.020)	-0.167 (-1.150)	0.112 (1.080)	-0.350 (-1.280)	-0.171 (-0.930)
InIDIV	0.253*** (4.390)	0.113 (0.610)	0.016* (1.910)	0.235*** (3.980)	-0.206*** (-2.860)	0.366 (0.310)	0.289*** (4.830)	-0.127*** (-3.640)	0.029** (-2.230)
FDI	-0.465 (-0.340)	19.089*** (4.610)	18.624*** (4.230)	0.883 (0.650)	5.740*** (3.450)	6.623*** (3.250)	-0.608 (-0.430)	8.386* (1.670)	7.778 (1.610)
InGDP	0.593*** (8.450)	-0.035 (-0.220)	0.558*** (3.630)	0.523*** (8.890)	-0.053 (-0.750)	0.470*** (6.920)	0.625*** (8.890)	-0.191 (-1.260)	0.435*** (2.860)
InRKMD	0.584* (1.940)	5.710*** (5.170)	6.294*** (5.730)	0.908*** (3.000)	0.894** (2.220)	1.802*** (3.970)	1.273*** (4.290)	5.233*** (2.890)	6.506*** (3.620)
KJB	-4.769** (-2.290)	-1.171 (-0.390)	-5.940** (-2.450)	-5.574*** (-2.860)	-0.056 (-0.030)	-5.629*** (-2.990)	-3.372 (-1.550)	-1.537 (-0.560)	-4.909** (-2.570)
CZHL	0.218 (0.710)	-2.915*** (-2.720)	-2.698** (-2.410)	-0.381 (-1.170)	0.657 (1.650)	0.276 (0.520)	-0.296 (-0.930)	-0.826 (-0.740)	-1.122 (-1.000)
R2	0.557			0.589			0.566

IKSL under three weight matrices has CE reduction effect. Under ADW, IKSL increases by 1%, and CE decrease by 0.181%.Under EDW, IKSL increases by 1%, and CE decrease by 0.167%. Under EGW, IKSL increases by 1%, and CE decrease by 0.171%, and it is consistent with Dong et al.(2025) and [33].This is because IKSL affects CE by affecting the production activities of the industrial sector. Based on geographical proximity and industrial complementarity and related advantages, the same industry can communicate and cooperate with each other during IKSL process. On the one hand, through learning and sharing, significant technology spillover effects are generated, especially clean technology brought by high-tech industries, which can significantly reduce CE. IKSL forms an environmental protection industry through specialized division of labor, which improves CE [5].

The process of IDIV did not reduce CE, but increased regional CE. Under ADW, IDIV increased by 1%, and CE increased by 0.016%.Under EDW, IDIV increased by 1%, and CE increased by 0.366%.Under EGW, IDIV increased by 1%, and CE increased by 0.029%.This may be because current development of productive service sector is mostly low-end development, service support provided to industry is far from meeting the needs of industrial technology-intensive industries. Therefore, the current collaborative agglomeration of productive service sector and industry is still at a low-quality agglomeration, which have no role in facilitating CE reduction effect of the city, but increases total CE of region.

Regional heterogeneity of IST on CE

Since the three provinces have distinct urban economic levels, distinct resource endowments and geographical locations, marketization level, enterprise technology level, infrastructure level, and transportation level of different regions will affect the effect of IST on CE. The direction and effect of IST on CE in different regions may be distinct. Considering regional heterogeneity, this paper divides total region into Hunan, Hubei, and Jiangxi provinces to explore impact of IST on CE in three sub-regions.

Regional heterogeneity of ISO on CE

Table 5 reports the impact of ISO on CE under ADW, EDW and EGW. Rho is 0.522, 0.007 and 0.127.At the level of ITL, the ITL of Hunan, Hubei and Jiangxi has reduced CE. ITL of Hunan Province increased by 1%, and CE decreased by 0.862%.ITL of Hubei Province increased 1%, and CE decreased by 1.672%.ITL of Jiangxi Province increased by 1%, and CE decreased by 1.860%.

Table 5.

Impact of spatial spillover effect of ISO on CE by region

Explained variable CE		Hunan			Hubei			Jiangxi
spatial autoregressive coefficient rho		0.522*** (-3.820)			0.007** (2.560)			0.127* (1.670)
	Direct Effect	Indirect Effect	Total Effect	Direct Effect	Indirect Effect	Total Effect	Direct Effect	Indirect Effect	Total Effect
ITL	-0.181(0.370)	1.043* (-1.850)	0.862** (-2.250)	1.505*** (-5.210)	0.167 (-0.370)	1.672*** (-3.460)	0.562 (-1.160)	1.299 (-1.630)	1.860** (-2.150)
IW	-1.675*** (-3.340)	-0.233 (-0.310)	-1.908*** (-2.890)	0.860** (2.290)	-1.242* (-1.750)	-0.382 (-0.450)	0.248 (0.600)	0.281 (0.220)	0.530 (0.380)
FDI	-2.979 (-0.600)	8.363 (1.390)	5.384 (1.310)	-2.490 (-1.330)	7.913*** (3.280)	5.423** (2.320)	1.362 (0.580)	2.285 (0.380)	3.647 (0.540)
InGDP	0.573*** (4.580)	-0.011 (-0.070)	0.562*** (6.200)	0.241** (2.500)	0.034 (0.260)	0.274*** (3.050)	1.063*** (5.800)	-0.632* (-1.770)	0.431 (1.270)
InRKMD	2.546*** (3.570)	-3.884** (-2.460)	-1.338 (-0.720)	-0.439 (-1.280)	4.021*** (6.850)	3.582*** (6.930)	-3.319** (-2.460)	2.080 (0.650)	-1.239 (-0.330)
KJB	-2.321 (-0.400)	-4.106 (-0.650)	-6.427** (-2.400)	-6.565*** (-2.930)	0.509 (0.200)	-6.057*** (-3.400)	5.401 (1.130)	-7.519 (-1.330)	-2.118 (-0.490)
CZHL	-5.017*** (-4.700)	5.421*** (4.490)	0.404 (0.460)	1.206*** (4.550)	-1.345** (-2.130)	-0.139 (-0.190)	-0.119 (-0.120)	1.283 (0.660)	1.163 (0.560)
R2	0.582			0.396			0.423

Looking at IW, IW of Hunan and Hubei provinces reduced CE. IW of Hunan Province increased by 1%, and CE decreased by 1.908%. IW of Hubei Province increased by 1%, and CE decreased by 0.382%. IW of Jiangxi Province increased CE, and for every 1% increase, CE increased by 0.530%.

Regional heterogeneity of ISI on CE

Table 6 reports the impact of ISI on CE under ADW, EDW and EGW. The coefficient rho is 0.254, 0.036 and 0.122, respectively. In terms of ISI, ISI in Hunan and Hubei provinces increased CE. IKSI in Hunan increased by 1%, CE increased by 0.335%.IKSI in Hubei increased by 1%, CE increased by 0.040%.Among the three regions, only ISI in Jiangxi Province played a role in reducing CE. IKSI increased by 1%, and CE decreased by 0.597%. In terms of IDIV, IDIV of Hunan and Jiangxi provinces increased CE. For every 1% increase in IDIV in Hunan Province, CE increased by 0.140%. For every 1% increase in IDIV in Jiangxi Province, CE increased by 0.959%.IDIV of Hubei Province played a role in reducing CE,for every 1% increase, CE decreased by 0.079%.

Table 6.

Influence of ISI on CE by region

Explained variable CE		Hunan			Hubei			Jiangxi
Spatial autoregressive coefficient rho		0.254** (-1.970)			0.036*** (2.950)			0.122** (2.320)
	Direct Effect	Indirect Effect	Total Effect	Direct Effect	Indirect Effect	Total Effect	Direct Effect	Indirect Effect	Total Effect
IKSL	0.230 (1.560)	0.105 (0.440)	0.335 (1.370)	0.212** (1.970)	-0.173 (-0.960)	0.040 (0.210)	-0.663*** (-2.770)	0.067 (0.150)	-0.597 (-1.270)
InIDIV	0.510*** (4.740)	-0.371** (-2.030)	0.140 (0.830)	0.185** (2.570)	-0.263* (-1.870)	-0.079 (-0.510)	0.300** (2.250)	0.659** (2.210)	0.959*** (2.820)
FDI	-4.311 (-0.910)	5.745 (0.930)	1.433 (0.280)	-2.536 (-1.280)	8.889*** (3.210)	6.353** (1.990)	2.013 (0.890)	2.503 (0.450)	4.515 (0.720)
InGDP	0.699*** (5.860)	-0.267 (-1.590)	0.432***(3.860)	0.335*** (3.470)	0.049 (0.360)	0.384*** (4.000)	1.089*** (6.190)	-0.361 (-1.170)	0.728** (2.560)
InRKMD	2.961*** (3.850)	-3.221* (-1.750)	-0.261 (-0.110)	-0.441 (-1.330)	4.379*** (7.420)	3.938*** (7.010)	-2.747** (-2.130)	4.651 (1.420)	1.904 (0.490)
KJB	-2.735 (-0.510)	0.688 (0.110)	-2.047 (-0.560)	-7.463*** (-3.230)	1.585 (0.610)	-5.878*** (-2.950)	3.779 (0.840)	-7.822 (-1.450)	-4.042 (-0.990)
CZHL	-3.819*** (-3.600)	4.983*** (3.950)	1.165 (1.070)	1.169*** (4.140)	-1.390** (-2.020)	-0.221 (-0.280)	-0.994 (-1.040)	-2.453 (-1.290)	-3.447* (-1.770)
R2	0.687			0.719			0.523

Robustness test

For further to test the robustness,we test the robustness of model by adjustment sample period, lowering study time by 4 years,and inverse distance weight matrix was using to test robustness of result in Table 7 and Table 8. The effects of ISO and ISI on CE are basically in line with result of original model, which proves estimation results of original model have strong stability.

Table 7.

Robustness test of the impact of ISO on CE

Explained variable CE
Explanatory variable	Direct Effect	Indirect Effect	Total Effect
ITL	0.228 (-1.030)	0.995** (-2.290)	1.224*** (-3.830)
IW	1.600*** (3.600)	-1.538 (-1.590)	0.062* (-1.650)
FDI	0.533 (0.410)	6.056*** (2.640)	6.589*** (3.070)
InGDP	0.427*** (6.990)	0.022 (0.250)	0.450*** (6.470)
InRKMD	0.869** (2.460)	1.933*** (2.680)	2.802*** (4.560)
KJB	-1.183 (-0.660)	-4.220** (-2.280)	-5.404*** (-3.730)
CZHL	-0.094 (-0.330)	-0.849 (-1.610)	-0.943 (-1.630)
Rho	0.025** (-2.200)
R2	0.770

Table 8.

Robustness test of the impact of ISI on CE

Explained variable CE
Explanatory variable	Direct effect	Indirect effect	Total effect
IKSL	-0.147* (-1.760)	-0.341** (-2.230)	-0.195** (-2.460)
InIDIV	0.221** (-2.460)	-0.296*** (2.830)	0.075 (-0.730)
FDI	1.944 (1.510)	7.321*** (3.040)	9.266*** (4.020)
InGDP	0.440*** (7.270)	0.130 (1.490)	0.570*** (7.700)
InRKMD	0.924*** (2.630)	2.440*** (3.310)	3.364*** (5.180)
KJB	-1.960 (-1.100)	-4.943*** (-2.660)	-6.904*** (-4.470)
CZHL	-0.073 (-0.250)	-1.191** (-2.130)	-1.263* (-1.970)
Rho	0.042*** (4.630)
R2	0.774

GTWR Results of IST on CE

Effect of ISO on CE

As shown in Fig. 4, spatial variability of influence of ISO on CE has increased significantly. The box length of ITL has increased every year, which means spacetime heterogeneity of effect of ISO on CE has been increasing every year. This is because economic level was limited in the early stage of the research time, ISO in the industrial sector was low, production capacity was backward, and economic development relied on labor and capital intensive enterprises.Labor-intensive and capital-intensive companies are high-energy-consuming companies, which require a lot of energy input and have little effect on CE reduction. With the advancement of industrialization,gap between economic development level and industrialization processes in different region has widened, which result in different degrees of rationalization of industrial structures in industrial sectors, and differences in CE reduction effects have gradually increased. Spatiotemporal variability affect of IW on CE has been in a relatively stable trend since 2016. In IW, as economy develops, sophisticated technology have become the major impetus for economic. Advanced science and technology have reduced energy consumption by improving energy utilization efficiency,and CE reduction effect of IW has been controlled in a relatively stable state.

Fig. 4.

Spatiotemporal variability of CE reduction effect of IST

Effect of ISI on CE

The spacetime heterogeneity of IKSL on CE shows a u-shaped trend. From 2000 to 2011, spacetime heterogeneity of effect on CE decreased year by year, and after 2012, spatial and temporal heterogeneity of the effect on CE increased again. This is due to industrial sector is primary contributor to energy CE, in initial stage of IKSL, industrial sectors with similar development scales are concentrated in the same region. The difficulties in cooperation between various sectors increase production costs, affect production efficiency, and lead to more energy consumption [18, 36]. In addition, various sectors have invested a lot of infrastructure construction in industrial agglomeration, so CE increased in preliminary period of IKSL. However, as economy develops, IKSL has continued to improve, and industry has continued to develop. In order to reduce production costs and increase economic benefits, industrial sectors have continuously carried out technological innovation. Advanced technologies flow and diffuse between the same and different types of industrial sectors, thereby reducing the CE. What’s more, rapidly developing high-end productive service industry can provide technical support for industrial sector and promote technological progress of industrial sector, which increase the CE reduction effect. Spatiotemporal variability of IDIV on the CE shows a stable trend of change, with little change during the study period.

Results by region

As shown in Fig. 5, from 2000 to 2020, the coefficient of the CE in Hunan and Jiangxi provinces was negative and relatively stable. The regression curve shows that CE reduction on ITL in Hunan and Jiangxi provinces has not changed too much. In contrast, coefficient of CE mitigate effect on ITL in Hubei Province has changed greatly. From 2000 to 2010, CE reduction effect of ITL in Hubei Province increased every year, CE alleviative effect decreased from 2010 to 2015, and increased every year after 2015.

Fig. 5.

Spatio-temporal variability of CE Due to IST

From 2000 to 2010, the IW regression coefficient curve of CE alleviative effect in Hunan and Jiangxi provinces was similar, and emission reduction effect showed a trend of decreasing year by year, and with some differences after 2010. After 2010, CE alleviative effect in Jiangxi Province began to show an increasing trend every year, while CE alleviative effect in Hunan Province on CE stabilized after 2010 and declined after 2018. Compared with Hunan and Hubei provinces, regression curve of CE reduction effect of IW in Jiangxi Province changed greatly, showing a more obvious stage-by-stage change trend. Emission reduction effect of IW increased every year from 2000 to 2005, and decreased every year from 2005 to 2010. CE reduction effect of IW generally showed a decreasing trend from 2010 to 2015, and decreased again after 2015. This shows that technology spillover effect generated by exchanges and cooperation between different sectors of industrial sector has a relatively stable inhibitory effect on CE.

The changes in regression coefficients of IKSL on CE in Hunan and Jiangxi provinces are quite interesting. From 2000 to 2010, the emission reduction effects of IKSL in the two provinces increased year by year and converged in 2010, CE reduction effects of IKSL in the two provinces were the same in 2010. After 2010, CE reduction on IKSL in Jiangxi Province increased year by year, while CE reduction effect of IKSL in Hunan Province weakened. CE reduction effect of IKSL on CE in Hubei Province gradually decreased during 2000 to 2005, increased every year from 2005 to 2015, and then decreased again after 2015.

The CE reduction effect of IDIV on CE in Hunan and Jiangxi provinces increased from 2000 to 2005, and decreased from 2005 to 2010. After 2010, the emission reduction effect of IDIV in Jiangxi Province increased every year, while CE effect of IDIV in Hunan Province was relatively stable and decreased after 2015. CE reduction effect of IDIV in Hubei Province decreased slightly year by year from 2000 to 2010, and CE reduction effect showed a clear downward trend after 2010.

Threshold effect results of IST on CE

Threshold effect of ISO on CE

As shown in Tables 9, 10 and  11, the single threshold P value of 0.000 and double threshold P value of 0.073 for ITL on CE are significant, and third threshold P value is not significant at 10% confidence level, indicating that ITL has a single and double threshold alleviative effect on CE. When economic development level is less than 14.224, coefficient of ITL on CE is -2.981. Economic development level is between 14.224 to 15.112, coefficient of ITL on CE is -1.718,Economic development level is greater than 15.112 and coefficient of ITL on CE is -0.561.

Table 9.

Threshold results of ITL on CE

Model	Threshold	Lower	Upper
Th-1	15.112	15.047	15.146
Th-21	15.410	15.301	15.443
Th-22	14.224	14.004	14.879
Th-3	15.112	14.831	15.146

Table 10.

Threshold test results of CE due to ITL

Threshold	RSS	MSE	Fstat	Prob	Crit10	Crit5	Crit1
Single	50.502	0.083	87.930	0.000	26.576	34.757	46.455
Double	48.325	0.079	27.430	0.073	24.521	30.363	57.290
Triple	47.422	0.078	11.590	0.447	22.233	24.862	39.051

Table 11.

Threshold model parameter results of ITL on CE

lnCE	Coef	Std.Err	T	P > t	[95%Conf	Interval]
ITL
0	-2.981	0.616	-4.840	0.000	-4.241	-1.720
1	-1.718	0.438	-3.920	0.000	-2.614	-0.822
2	-0.561	0.506	4.400	0.001	0.353	1.076
3	0.992	0.636	1.560	0.130	-0.310	2.293

As shown in Tables 12, 13 and  14, single threshold P value of IW on CE is significant at 0.010, and double threshold P value and third threshold P value are not significant, indicating that IW has a single alleviative threshold effect on CE. When the economic development level is less than 14.762, the impact coefficient of IW on CE is -5.976.

Table 12.

Threshold results on CE caused by IW

Model	Threshold	Lower	Upper
Th-1	14.762	14.700	14.794
Th-21	14.509	14.389	14.566
Th-22	15.112	15.016	15.146
Th-3	13.752	13.589	15.648

Table 13.

Threshold test results of CE caused by IW

Threshold	RSS	MSE	Fstat	Prob	Crit10	Crit5	Crit1
Single	55.963	0.092	76.720	0.010	42.749	54.738	70.555
Double	53.038	0.087	33.590	0.170	41.279	50.518	80.277
Triple	51.812	0.085	14.400	0.717	41.801	46.979	72.060

Table 14.

Threshold model parameter results of IW on CE

lnCE	Coef	Std.Err	T	P > t	[95%Conf	Interval]
IW
0	-5.976	1.691	-3.530	0.001	-9.435	-2.518
1	-2.639	1.008	-2.620	0.014	-4.700	-0.578
2	-0.523	0.741	-0.700	0.486	-2.039	0.994
3	1.181	0.566	2.090	0.046	0.023	2.340

Impact coefficient of ITL on CE changed from -2.981 to -0.561, and even reached 0.992, the possible reason is that when the UYR region was at a medium level of economic development, combined effect of structural transformation,technology application and regional collaboration was maximized, which formed a golden period for CE reduction [16, 23, 24]. When economy entered high-income stage, the structural dividend was exhausted, carbon lock-in intensified, demand and scale effects offset each other, and the regional industrial path dependence was superimposed, which ultimately led to a weakening or even reversal of CE reduction effect [28, 41].

Threshold effect of ISI on CE

In Tables 15, 16, and 17, single threshold value P of IKSL on CE is 0.000 and double threshold value P is 0.007. Third threshold value P is not significant at 10% confidence level, indicating that IKSL has a single and double threshold alleviative effect on CE. When economic development level is less than 13.752, IKSL on CE is -1.004. Economic development level is between 13.752 to 14.509 and IKSL on CE is -0.662. Economic development level is greater than 14.509 and coefficient of IKSL on CE is -0.253.

Table 15.

Threshold results on CE caused by IKSL

Model	Threshold	Lower	Upper
Th-1	15.112	15.047	15.146
Th-21	15.443	15.342	15.472
Th-22	14.509	14.389	14.566
Th-3	13.752	13.589	13.848

Table 16.

Threshold value test results of CE caused by IKSL

Threshold	RSS	MSE	Fstat	Prob	Crit10	Crit5	Crit1
Single	52.791	0.087	123.460	0.000	39.029	52.411	83.026
Double	48.480	0.080	54.150	0.007	30.464	35.501	45.821
Triple	47.349	0.078	14.550	0.727	41.813	49.331	66.865

Table 17.

Parameter results of the CE threshold model of IKSL

lnCE	Coef	Std.Err	T	P > t	[95%Conf	Interval]
IKSL
0	-1.004	0.148	-6.760	0.000	-1.307	-0.700
1	-0.662	0.135	-4.900	0.000	-0.938	-0.386
2	-0.253	0.145	-1.750	0.091	-0.549	0.043
3	0.090	0.130	0.690	0.496	-0.176	0.356

In Tables 18, 19 and 20, single threshold value P of IDIV on CE is 0.000 and double threshold value P is 0.013, and third threshold value P is not significant, indicating that IDIV has a single and double threshold alleviative effect on CE. When economic development is less than 14.376,impact coefficient of IDIV on CE is -0.262. Economic development level is between 14.376 and 14.509, IDIV on CE is 0.032. When economic development level is greater than 14.509 and IDIV on CE is 0.154.

Table 18.

Results of IDIV on CE thresholds

Model	Threshold	Lower	Upper
Th-1	14.509	14.389	14.566
Th-21	14.376	14.316	14.462
Th-22	15.112	15.047	15.146
Th-3	15.643	15.542	15.663

Table 19.

Test results of IDIV on CE thresholds

Threshold	RSS	MSE	Fstat	Prob	Crit10	Crit5	Crit1
Single	51.470	0.085	117.490	0.000	34.104	42.544	56.199
Double	45.759	0.075	76.010	0.013	33.559	46.714	85.457
Triple	43.764	0.072	27.750	0.480	54.913	68.001	88.843

Table 20.

Parameter results of the CE threshold model of IDIV

lnCE	Coef	Std.Err	T	P > t	[95%Conf	Interval]
lnIDIV
0	-0.262	0.107	-2.460	0.020	-0.481	-0.044
1	0.032	0.107	3.310	0.008	0.010	0.050
2	0.154	0.097	2.530	0.017	0.047	0.444
3	0.246	0.090	-0.155	0.166	-0.352	0.074

Conclusion and discussion

Conclusion

This paper constructs socioeconomic panel data for 30 cities in the urban agglomeration in the middle reaches of the Yangtze River from 2000 to 2020, and explores the influence of IT on CE in 33 industrial sectors in China from spatial spillover effect, spatiotemporal heterogeneity, and threshold effect. The results show that:

First,under three weights, ITL plays a role in reducing CE in ISO, while IW increases CE.In ISI, IKSL has played a role in reducing CE while IDIV increases CE.

Second,the CE reduction effect of IST exhibits spatiotemporal heterogeneity.The impact of ISO and ISI on CE varies across different regions and time.The spatial heterogeneity level of ISO on CE in IST has increased during research period. Spatial heterogeneity of ITL on CE has increased every year, spatiotemporal heterogeneity of IW on CE shows a propensity of first reducing and then boosting, and then maintaining in a relatively stable trend since 2016.

Third, due to regional heterogeneity, differences in economic development level and technological progress, the impact of IST on CE varies slightly among the three sub-regions.ISO and ISI in the three regions have obvious spatial heterogeneity in their CE reduction effects.

Fourth, under different levels of economic development, the IST has different effects on CE. Both IKSL and IDIV have a single threshold effect and a double threshold effect on CE. From a regional perspective, the threshold effect of IST on CE in different areas is distinct.

Discussion

Given the unique role of the industrial sector in China’s economic and environmental development, the above findings further highlight the research value of this paper. Industry is a core pillar of China’s economic development, playing an irreplaceable and crucial role in driving national economic growth and promoting industrialization and modernization. However, its development has also generated a series of environmental problems, with CE being particularly prominent. Simply shutting down industrial sectors to address environmental issues will inevitably have a negative impact on China’s economic development, making it difficult to achieve the coordinated development of the economy and the environment. Therefore, promoting the green transformation and high-quality development of the industrial sector has become a key path to achieving a win–win situation for stable economic growth and ecological environment improvement for CE reduction.

The findings in this study show that there is a difference between regional and UYR overall results. For example, in UYR, IW increases CE, while Hunan and Hubei Provinces’ IW decreases CE. The core reason for this is that IW in these two provinces is primarily driven by low-carbon, high-value-added industries, coupled with a stronger capacity to absorb green technologies, ultimately achieve a synergy between upgrading and CE reduction.The divergence in regional industrial bases and technological transformation capabilities among different provinces leads to distinct industrial upgrading paths, some leaning towards high-end manufacturing and green transformation, which results in some regions’ results contradicting the overall findings.

This paper makes contributions in two ways. First, we develop the theoretical mechanism of how IST affects CE by changing land transformation. We constructed a theoretical framework for the impact of IST on CE and analyze CE effects derived from land use changes under socioeconomic system with IST, which can more comprehensively explore CE effects derived from IST. Secondly, IST is analyzed from two dimensions of ISO and ISI, and the impact of IST on CE is analyzed from three dimensions: spatial spillover effect, patiotemporal heterogeneity and threshold effect. As a result, the process of IST’s carbon reduction effect can be examined from a dynamic perspective, especially from different stages of economic development.

However,this paper also has research limitations.First, the ISI measures constructed in this paper, such as ITL, may not be comprehensive enough, resulting in some empirical results being insignificant or even negative. More refined measurement methods need to be used in subsequent studies.Second, the differences in the impact of ISI and ISO dimensions on CE have not been thoroughly studied,future research will examine the differences in the impact of ISI and ISO on CE in greater detail.Third, due to limitations in the updating schedule of carbon emission-related databases, data is only available up to 2020; therefore, this study’s timeframe only extends to 2020. Future studies will incorporate updated data to supplement and validate the research findings.

Policy implications

Based on the research findings, this paper proposes the following policy recommendations.

First of all, in the IST process, ISI has not played a good role in CE reduction. In the future, efforts need to be made to develop leading industries, improve the industrial chain, and strengthen the role of ISI in CE reduction. When selecting leading industries, government should fully consider the resource endowment, location factors and economic development level of each region, reasonably select and build large-scale, high-efficiency, technologically advanced and widely diffused leading industries, developed the leading role of advantageous industries and cultivate other related additional industries to increase competitiveness of core industries.What’s more, greater emphasis should be placed on the application of green technologies and the establishment of green innovation indicator system.

Secondly,exert spatial spillover effect between adjacent regions to the fullest, build a reasonable industrial division of labor, focus on cultivating and developing emerging low-carbon industries, accelerate the growth of high-tech industries, build a modern service industry with high-tech content and high economic effects, promote industrialization with service industry, and facilitate the green transformation of regional internal industries. In addition, accelerate the integration of similar and different industries to give full play to Jacobs externalities and MAR externalities. Furthermore, promote the upgrading of producer services towards high-end and green development, and enhance their deep integration with advanced manufacturing, thereby promoting IKSL and IDIV to reduce CE.

Author contributions

W.P.:writing—original draft preparation, review and editing; A.Z.: conceptualization, review, editing and supervision. All authors have read and agreed to the published version of the manuscript.

Funding

This work was financially supported by Major Project of National Social Science Foundation of China (25&ZD171),Ministry of Agriculture and Rural Affairs in 2025(04250040), 2025 Zhejiang Provincial Department of Education Research Project(Y202559389),2026 Research Project of Zhejiang Federation of Social Sciences(2026N041),and the 2025 Zhejiang Province Soft Science Research Project (2025C25027).

Data availability

Data will be made available on request.

Declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.