Achieving the Paris Agreement’s temperature target of 1.5–2.0°C requires reducing carbon emissions in the atmosphere (Terhaar et al., 2022). At the same time, the removal of carbon from the atmosphere should also be increased. Global warming and sustainable development have forced the transformation of the energy system, gradually replacing fossil fuels with clean, efficient, and sustainable energy sources (Bogdanov et al., 2021). Over the past few decades, fossil fuels have contributed to the increase in China’s carbon emissions. Consumption of renewable resources is critical to reducing carbon emissions, and natural resource consumption is stimulating the continuous increase of carbon emissions (Li et al., 2022). Forests are a potentially scalable method for reducing atmospheric carbon and sequestering it in biomass and soil (Busch et al., 2019). During the period 2020–2050, China’s forest vegetation will be able to absorb about 22% of the carbon dioxide emitted by fossil fuel combustion (Qiu et al., 2020).

Among the most effective carbon sinks, forests are effective at storing, conserving, and substituting carbon (Nunes et al., 2020). As carbon sinks, forests often limit climate change by offsetting carbon emissions (Pugh et al., 2019). Carbon capture, storage, and utilization is a key technology for mitigating climate change. Large-scale carbon storage requires the establishment of a carbon transportation network to achieve a certain emission reduction target. Pipelines are more suitable to reduce carbon emissions in northeastern, north, and northwestern parts of China (Wei et al., 2022). During carbon sequestration, trees absorb and store carbon dioxide from the atmosphere, reducing emissions and costs. Sequestering carbon through afforestation is considered the most cost-effective method. Consequently, forests contribute significantly to reducing carbon emissions (Richards and Stokes, 2004; Raihan et al., 2019). Forest biomass or fuel can save 9–38% of annual carbon dioxide emissions by replacing fossil fuels. In comparison to gasoline, bio-methanol reduces carbon emissions by 8–35% (Suntana et al., 2009).

Forest carbon sinks help China balance its economic development and alleviate the effects of climate change (Zhang et al., 2023). The economic value of forest carbon sinks is primarily derived from reducing emissions and costs. When the Chinese economy maintains a medium-low growth rate, carbon sinks in the national carbon trading market will reduce carbon emissions more efficiently (Lin and Ge, 2019b). Global warming must be kept below 2°C by setting carbon prices and carbon caps at levels that allow reductions (Santos, 2022). Due to the high cost of carbon sequestration, landowners primarily sequester large amounts of carbon in forests by increasing forest land and extending rotation periods (Sohngen and Mendelsohn, 2003). Forest resource harvest rotation in China has a significant impact on carbon storage. Short rotation lengths resulted in the lowest simulated carbon stocks, while longer rotation lengths resulted in relatively higher carbon stocks (Jiang et al., 2002). Various combinations of price incentives for carbon sequestration and forest bioenergy expansion, which can reduce marginal abatement costs and increase mitigation potential in the short term, have been explored in a dynamic structural model. However, long-term carbon sequestration costs increase (Baker et al., 2019). The amount of carbon removed from reforestation would increase by about 5.6% under a fixed carbon price. Reforestation can reduce emissions 7.2–9.6 times more than avoided deforestation. It offers more low-cost emission reduction opportunities than preventing deforestation (Busch et al., 2019). In addition, risks associated with economics and climate are heavily influencing carbon sequestration’s social cost (Cai and Lontzek, 2019).

Based on the analysis of China’s Three-North Shelterbelt Project (TNSP), forests are able to sequester carbon well, but the amount of carbon accumulated fluctuates greatly, showing an overall downward trend. Sequestering carbon in forests is economically beneficial since a small amount of monetary investment yields a large amount of benefits (Chu et al., 2019). Due to phenological events occurring in early spring and late autumn, forests consume more carbon through photosynthesis than they release through respiration, and the timing of these events influences an ecosystem’s function. As a result of climate changes, carbon sequestration will likely be enhanced in the present and the future, thereby slowing global warming (Keenan et al., 2014).

Based on a regional tree species analysis of China’s commercial forests under the constraints of logging and afforestation, China’s annual carbon sequestration is estimated to be able to offset 5% of anthropogenic greenhouse gas emissions by 2019. Additionally, social and economic factors such as carbon prices and deforestation have a greater impact on carbon sinks (Liu F. et al., 2023). By adopting carbon pricing policies, CO₂ emissions can be reduced and regional income gaps can be narrowed (Zhang W.-W. et al., 2022). However, the willingness of cities to engage in carbon offset trading depends on a variety of factors, including urbanization level, understanding of carbon sequestration, educational level, and familiarity with market mechanisms (Poudyal et al., 2010).

The goal of reducing carbon emissions by utilizing forest resources has to be properly supervised and managed, even though forests can absorb carbon in the atmosphere through carbon sinks (Jin et al., 2020). Carbon sequestration by forests is declining. However, with proper management and monitoring, 100–200 Tg C per year of forest carbon can still be sequestered (Birdsey et al., 2006). Nevertheless, the impact of tropical forests on carbon storage is unknown. In scenario simulations, species composition has an important impact on carbon storage capacity (Bunker et al., 2005). Forestry stores the most carbon, and low harvesting will increase biodiversity indicators, such as dead wood and old forest areas (Heinonen et al., 2017). Therefore, through biodiversity conservation, it is possible to maximize the co-benefits of greenhouse gas mitigation (Rahman et al., 2021).

Due to climate policies that encourage the enhancement of forest carbon, forest ecosystems have been remodeled into carbon forestry. There is a serious threat to the potential of forests to benefit both nature and people (Ojha et al., 2019). The loss of biodiversity and climate change are causing an environmental crisis. Forests in tropical regions are most vulnerable to these crises. Climate change mitigation policies that reduce carbon dioxide caused by deforestation and degradation are often considered a win-win strategy (Phelps et al., 2012). However, some scholars have found that the coverage of forest resources has no significant impact on carbon emissions reduction (Xuan et al., 2020).

In direct proportion to global carbon emissions, deforestation contributes 6–17% of carbon emissions (Friel et al., 2009). As forests are cleared, carbon stored in leaves, branches, and roots is released into the atmosphere. Carbon dioxide in the atmosphere is primarily emitted by deforestation in the tropics. However, forest carbon storage varies spatio-temporally primarily due to natural and human disturbances (Baccini et al., 2012). The main cause of forest carbon emissions is deforestation. Climate change may cause the forest’s carbon sink function to cease. Consequently, reducing wood harvesting and ending deforestation are key strategies for mitigating climate change (Le Noë et al., 2021). Moreover, forest degradation results in three times more carbon loss than deforestation (Qin et al., 2021).

Forests are important sources and sinks of carbon, contributing to climate change (Köhl et al., 2015). There is a potential for tropical forests to become carbon sources in the future, mainly because of forest loss and the impact of climate change on their ability to absorb excess atmospheric carbon (Mitchard, 2018). Tree biomass sequesters carbon through climate change, which results in a faster increase in growing stocks due to increased tree growth. The effects of large forests can only be achieved with large areas of forest. Therefore, forest carbon balance highly depends on the rational use of forest resources (Eggers et al., 2008).

CO₂ concentrations in the atmosphere can affect the global climate system. Carbon dioxide is sequestered from the atmosphere by forests as part of the carbon cycle. Higher carbon concentrations promote plant growth through a “fertilization effect” (Drake et al., 2016). There are, however, other factors that can limit plant growth, such as pollution or nutrient deficiencies. Consequently, existing forests will not increase carbon storage in response to increased atmospheric carbon concentrations (Beedlow et al., 2004). Forest carbon stocks are the most commonly overlooked factor in assessing forest carbon emissions (Ter-Mikaelian et al., 2015). Plants absorb a great deal of CO₂ from the atmosphere. Due to a lag in tree mortality, forest carbon stocks may increase only temporarily. Thus, the current stimulation of tree growth will eventually result in a lagged increase in mortality, which will eventually lead to carbon gains (Brienen et al., 2020).

Reforestation has potential and value, but it should be carefully considered as one of the best ways to reduce carbon emissions from humans (Taylor and Marconi, 2020). However, there is the problem of carbon leakage that cannot be ignored. When the amount of carbon released elsewhere destroys the direct carbon benefits, forest carbon sequestration leakage happens. The carbon leakage effect should not be ignored when using forests to mitigate carbon emissions. Forest carbon sink leakage is heavily influenced by specific activities and regions (Murray et al., 2004).

China has been focusing on low-carbon policies to reduce city-level emissions (Liu and Qin, 2016). Using the method of Propensity Score Matching-Difference in Differences (PSM-DID), it was verified that the low-carbon city pilots significantly increased the actual performances of local enterprises with data from 285 Chinese prefectures from 2005 to 2015. By implementing a low-carbon pilot policy, enterprises can achieve emissions reductions and high-quality growth (Chen et al., 2021). According to the Difference in Differences (DID) model and the spatial DID model, low-carbon city pilots have enhanced carbon emission efficiency by 1.7%. According to the scenario analysis, low-carbon city pilots increased carbon emission efficiency by 1%, resulting in an 8.37 million tons reduction in carbon emissions (Yu and Zhang, 2021).

In a study of 282 Chinese cities from 2004 to 2018, it was verified that the pilot program can significantly reduce carbon emissions, and urban heterogeneity is evident. Additionally, low-carbon pilot policies help reduce carbon dioxide emissions through resource allocation, energy conservation, and technological innovation (Du et al., 2022). Data at the enterprise level in China shows that low-carbon pilot projects are effective at reducing coal consumption and intensity (Zhou et al., 2022). However, it is found that urban low-carbon pilots with high levels of innovation perform better in terms of urban heterogeneity (Zhang H. et al., 2022). Additionally, resource-based cities have a greater effect on reducing emissions (Zeng et al., 2023). According to the cost–benefit analysis of the pilot policy, the annual dioxide emissions of low-carbon pilot cities are reduced by 2.72 percent. In contrast, the related economic losses from 2013 to 2017 amounted to approximately 1.19 trillion yuan. By upgrading industrial structures and promoting technological innovation, the pilot cities can improve total factor productivity (Huo et al., 2022).

The setting of carbon reduction targets can facilitate the diffusion and learning of low-carbon city policies between different cities (Song et al., 2021). The promotion of low-carbon urban development is essential to reducing dioxide emissions. At different stages of urban development, different factors influence carbon emissions. Taking the capital of China, Beijing, as an example, the driving forces of such an increase in carbon emissions between 1991 and 2004 were economic aggregate and population size, while the real driving force behind reductions was industrial structure. From 2004 to 2022, economic performance contributed the most to the increase in carbon emissions, followed by total population and accessibility to energy (Shen et al., 2018). Government, enterprises, non-governmental organizations, and citizens are the main participants in the governance framework of low-carbon cities. The main framework for target-based low-carbon urban governance can also be used for other cities (Liu et al., 2018).

Despite some reductions in carbon emissions as a result of the pilot policy, insufficient market-oriented measures and supportive policies will impede the creation of truly low-carbon cities (Khanna et al., 2014). Even though the low-carbon pilot policy has the potential to reduce carbon emissions, improving carbon emission efficiency will take longer. Compared to China’s western regions, eastern cities are more affected by the pilot policy, thereby adding to the urban division between the eastern and western regions (Fu et al., 2021). Green total factor productivity increased significantly after the low-carbon pilot policy was carried out in the first year, but its effects gradually waned in subsequent years (Liu et al., 2020). According to panel data collected from 49 cities between 2005 and 2018, it has been quantitatively tested that the low-carbon pilot policy has failed to meet the goals and expectations, with the policy itself increasing carbon intensity by 15–20%. This unintended outcome is likely caused by the construction of initial infrastructure and the development of the economic circle, and the decarbonization process will subsequently increase carbon intensity (Feng et al., 2021).

The continuous advancement of digitalization has provided the basis for the development of smart cities and provided new opportunities for economic development (Krishnan et al., 2020). Carbon dioxide emissions are concentrated in cities and must be controlled in cities (Huang et al., 2022). By combining low-carbon pilots with smart cities, the emission reduction effect is better, and it far exceeds that of single pilot cities. In the long run, low-carbon pilot policies will mainly rely on big data, high technology, and industrial innovation to reduce carbon emissions (Zhang and Fan, 2023). As scarce resources and high-tech innovation become more prevalent, smart city construction will increasingly become a technology-based solution for efficient, affordable, and sustainable urban living. As part of the agenda of environmental sustainability, smart cities will also reduce carbon dioxide (Cavada et al., 2016). Essentially, smart cities are a new kind of urban development model that supports green urban development. The construction of smart cities often promotes sustainable development through urban planning, smart infrastructure construction, and smart governance (Cheng et al., 2022).

Investing in smart cities can indeed improve urban carbon emissions efficiency, and this promotion effect on the low-carbon economy will gradually increase over time (Qiu et al., 2021). Smart cities help build low-carbon economies by optimizing government functions, improving governance efficiency, and implementing openness, especially for large cities (Fan et al., 2021). A smart city’s primary purpose is to reduce per capita carbon emissions through improved energy efficiency, and this is more apparent in cities where the administrative level is higher and green innovation is more prevalent (Guo et al., 2022). When working with big data to reduce emissions in a smart ecosystem, the government may face various challenges (Giest, 2017).

Through the development of smart technologies, smart cities reduce emissions related to transportation and education, thereby reducing household indirect carbon emissions (Wu, 2022). When residents combine energy saving, money saving, changing consumption behavior, and reducing carbon emissions, they will consciously develop a low-carbon lifestyle and reduce the family’s carbon footprint (Hurst et al., 2020). Besides, smart transportation can not only curb transport-related carbon emissions but also greatly curb dioxide emissions in non-traffic sectors. Furthermore, smart transportation affects carbon dioxide emissions indirectly through traffic scale, structure, and technical factors (Zhao et al., 2022). Through the analysis of 279 cities in China from 2008 to 2018, it is verified that smart cities can significantly reduce pollutant emissions and energy intensity, thereby promoting the development of an urban green economy (Liu K. et al., 2023).

Existing literature provides good theoretical and practical support for this research. However, few studies are conducted from the double dual perspective of natural resources and man-made policy. There are several possible marginal contributions of this research. First of all, this study clarified the controversy of forest natural resources as carbon source and carbon sink through empirical analysis, and scientifically analyzed the impact of forest resource abundance on carbon emissions based on fully considering the dual attributes of forest as carbon sink and carbon source. In addition, based on the panel data of 41 cities in the YRD region from 2007 to 2019, this paper empirically tests whether natural resource endowment and low-carbon pilot policies can help reduce carbon emissions from the dual perspectives of resources and policies by using econometric models. Third, to strengthen the urban ecological spatial layout and conform to the trend of urban digital and low-carbon development, this paper uses the Difference in Difference in Differences (DDD) method to test whether the carbon emission reduction effect of low-carbon city pilot policies has significant heterogeneity in resource endowment and smart city construction. Also, this paper provides some experience and reference for the YRD region to realize green transformation development in the digital age.

Forest resources reduce emissions primarily by absorbing carbon dioxide from the atmosphere. In addition to being carbon sinks and sources, forest resources need to be managed properly to reduce carbon emissions effectively. It is not enough to reduce carbon emissions using nature-based solutions. Minimizing carbon emissions effectively requires a synergistic approach. Cities are responsible for the majority of carbon emissions during urban development. Low-carbon city pilots examine heterogeneity in three ways: whether the city is resource-based, smart, or forest-rich. In this paper, according to the practice of previous scholars (Ding et al., 2006), as long as the forest coverage rate exceeds 30%, it is considered a city with abundant forest resources. The paper thus proposes three hypotheses:

In order to examine the impact of forest resources abundance in YRD on carbon emissions, using 41 cities in this region from 2005 to 2019, a double fixed effect model was established as follows:

Where formula (1) represents a linear model, and formula (2) represents a nonlinear model. Carbon in those functions represents the emissions of carbon dioxide per GDP, FC represents the forest coverage rate, FC² represents the square of the forest coverage rate, i represents the city in YRD, t represents the year, μ i represents the individual effect of the city, ξ t represents the period effect, and ε it represents random error term.

Low-carbon development in social development requires more than just reducing carbon emissions. Taking economic development into account is also important. Thus, carbon emissions per GDP serve as the explained variable, and forest cover rate is used as the main explanatory variable. At the same time, economic indicators, environmental indicators, and institutional indicators are selected as control variables to ensure the rationality and credibility of the model results as much as possible. Table 1 lists the data source and variables.

Moreover, this paper presents estimates of net carbon emissions per GDP (Ncarbon) excluding forest carbon sequestration. In the calculation of forest carbon storage, many studies rely on forest coverage or forest area (Sun and Liu, 2019). With the forest area data in YRD, this paper analyzes the forest carbon sink capacity from 2005 to 2019. The net ecosystem productivity of vegetation (i.e., carbon sink) is a measure of how much carbon dioxide enters the ecosystem from the atmosphere. The calculation does not include carbon emissions from vegetation respiration. Here is the formula for calculating vegetation net ecosystem productivity (Yi et al., 2015).

Where C_nep refers to the net ecosystem productivity of the forest ecosystem, which reflects the total carbon sink absorbed by the forest ecosystem every year. C_veg is the net production per unit area of vegetation, that is, the annual carbon sink absorbed by the forest per unit area, which is taken as 3.81thm⁻²a⁻¹ (Xie et al., 2008); A_veg is the vegetation area. Forest area is calculated from forest coverage rate and land area. The net carbon emission per GDP is mainly the carbon emission per GDP minus forest carbon sinks.

This paper uses data from 41 cities in YRD from 2005 to 2019 to examine the abundance of forest resources and how low-carbon pilot policies reduce carbon emissions. Moreover, it proposes strategies that take natural resources and policies into account for sustainable development. Each variable is described in Table 2. In order to remove the influence of each variable’s dimension, all variables are standardized.

Table 3 is the correlation analysis of each variable. In the table, almost all the correlation coefficients are less than 0.6, implying that there is little correlation between variables.

Table 4 shows each variable’s multicollinearity analysis. In order to determine if variables are multicollinear, variance inflation factors (VIF) and tolerance (TOL) are commonly used. Multicollinearity is generally considered to be more serious if the VIF is greater than 10. The VIF and TOL results revealed no multicollinearity between variables, as the VIF and TOL averages are only 2.09 and 0.52, respectively.

Based on panel data from 41 cities in the YRD from 2005 to 2019, from the dual perspectives of resources and policies, this paper first analyzes the impact of forest resources on carbon emissions by using a double fixed effect model and a spatial econometric model. Then, with the DID Model, this paper analyzes the policy effects of low-carbon city pilots in the YRD region on carbon emission reduction. Later, the DDD model is employed to assess the heterogeneity of low-carbon city pilots in smart cities, resource-based cities, and forest-rich cities. The empirical results are proven to be robust. Here are the findings of the hypotheses proposed in the paper:

Although forest resources inhibit carbon emissions, they also act as carbon sinks and sources, resulting in a nonlinear relationship between forest resources and net carbon emissions. There is a profound effect of forest resources on carbon emissions (Pugh et al., 2019). However, taking into account the impact of forest carbon sequestration, forest resources have a greater impact on net emissions. Forest resources are nonlinearly related to net carbon emissions per GDP, and forest coverage suppresses net emissions at first and then promotes them. The U-shaped trend shows that forest resources will promote carbon emissions when they reach a certain level. This also fully demonstrates the carbon-neutral nature of forests (Köhl et al., 2015), and carbon sequestration has a social cost (Taylor and Marconi, 2020). The early contribution of forests to carbon emissions is worth looking forward to. However, in the long run, we should pay attention to the biodiversity ecosystem services of forests, which is the essence of sustainable green and low-carbon development.

Through the analysis of forest resources and low-carbon pilot policies in YRD, the article tries to prove that from the perspective of resources and policies, man and nature can coexist harmoniously, and enable cities to develop sustainable low-carbon. Moreover, this paper shows that forest resources have dual attributes as carbon sinks and carbon sources, and relying only on forest resources for carbon emission reduction will present certain challenges. Therefore, achieving low-carbon development that co-exists with nature will require a combination of policies to promote the low-carbon transition in the region’s industries, transportation, foreign trade and other sectors. In addition, the findings of this study show that forest resources do have the effect of carbon reduction, but the characteristics of forest resources themselves are carbon neutral so forest resources can absorb and emit carbon. In addition, low-carbon pilot policies are indeed effective, and the policy works better in smart cities, resource-based cities, and cities with rich forest resources. Therefore, this paper proposes the following policy implications to help the low-carbon transformation development of the YRD region.

Forest resources should be fully protected and used effectively, while land use should be rationally planned. Because forest resources can reduce carbon emissions if they are used rationally and efficiently, and their carbon-neutral properties must be taken into consideration. Governments should properly protect forest resources and improve their condition through sustainable resource use to combat air pollution, mitigate climate change, and reduce carbon emissions. Forests can bring a lot of biomass energy, and the use of biomass energy can effectively control fossil energy consumption, to achieve the goal of pollution and emission control. Protecting forests contributes to sustainable ecosystem diversity and stability. At the same time, land use types and spatial structures should be optimized, and classified management should be carried out for low-carbon development and ecological diversity (Liu et al., 2022). However, the carbon sequestration function of forests can be relied on to assist in carbon emission reduction temporarily. In the long run, a forest ecosystem with biodiversity is needed to truly achieve low-carbon sustainable development.

The government should actively promote the low-carbon city pilot policy, and make the industrial layout low-carbon. Low-carbon pilot policies should be developed in coordination with ecological development, energy conservation and emission reduction, and environmental protection. Policymakers should timely introduce and improve industrial, fiscal and taxation, financial support, and technology promotion policies to promote low-carbon development. In order to control carbon emissions, the first step should be to develop a target responsibility system and refine the evaluation mechanism. Second, establish a system for monitoring and managing carbon emissions, compile an inventory of corporate carbon emissions, and develop a platform for carbon trading. The third step is to create a low-carbon industrial system and optimize the industrial structure, such as developing a low-carbon intelligent transportation system and a green tourism industry focusing on forest health. The fourth step is to advocate low-carbon life and consumption patterns and to deepen awareness of low-carbon living.

In terms of urban governance, the two major trends of digitalization and low carbonization must be accurately grasped, to empower the construction of smart cities with digitalization and improve the capacity and level of urban governance. Digital development can help the green and low-carbon transformation of cities and achieve low-carbon development through digital platforms, data elements, and technology. As cities become more low-carbon, digitalization and greenization will inevitably go hand-in-hand. The construction of smart cities can reduce emissions and conserve energy in production, consumption, supply, and all-round life scenarios to a greater extent. Meanwhile, the construction of a city’s smart management platform can optimize the city’s carbon emission activities and force the healthy development of the economy, thereby improving the management efficiency of carbon emissions, reducing unnecessary energy waste, and maximizing resource efficiency. What needs to be done now should be to decrease carbon emissions, rationally plan land use, and improve energy efficiency to truly achieve carbon emission reduction (Ho, 2023). Therefore, specific carbon emission reduction measures must be combined with the actual situation and adapted to local conditions to finally achieve environmentally friendly development.