How Digital Intelligence Integration Boosts Forestry Ecological Productivity: Evidence from China

Abstract

In the context of the “Dual Carbon” goals and ecological civilization development, enhancing forestry ecological total factor productivity (FETFP) has become vital for advancing green development and environmental governance. Confronted with tightening resource constraints and pressure to transform traditional growth models, whether digital intelligence integration can effectively empower improvements in FETFP requires in-depth empirical validation. Based on publicly available panel data from 30 Chinese provinces spanning 2012 to 2022, this study constructs an index system for measuring digital intelligence integration and FETFP. Using the Double Machine Learning (DML) framework, the study empirically identifies the impact of digital intelligence development on FETFP and explores its internal mechanisms. The key results show that (1) digital intelligence integration significantly enhances FETFP. For every unit increase in digital and intelligent integration, FETFP rises by an average of 19.97%; (2) mechanism analysis reveals that digital intelligence improves FETFP by optimizing the forestry industrial structure, promoting green technological innovation, and amplifying the synergistic effects of fiscal support; (3) and heterogeneity analysis suggests that the positive impact of digital intelligence integration is more pronounced in regions with higher environmental expenditures and stronger green finance support. Accordingly, this study proposes several policy recommendations, including accelerating digital infrastructure development, strengthening foundational digital intelligence capabilities, enhancing support for green innovation, leveraging the ecological multiplier effects of digital transformation, tailoring digital strategies to local conditions, and improving the precision of regional environmental governance. The findings provide robust empirical evidence for improving FETFP in developing and developed economies.

1. Introduction

Since the onset of the Industrial Revolution, global warming has intensified, giving rise to more frequent and severe extreme weather events [1]. These climatic disruptions have emerged as a formidable challenge to sustainable development, drawing heightened international concern [2]. As the world’s second-largest economy, China has urgently responded to environmental degradation [3]. In this context, the nation’s ecological civilization strategy has progressively emphasized the strategic role of forestry in enhancing ecosystem services, regulating climate systems, and safeguarding biodiversity [4]. Forestry has become a cornerstone sector in advancing green development [5].

Given its pronounced ecological externalities, forestry aligns well with the assumptions of the Cobb–Douglas production function, particularly regarding the estimation of multi-dimensional outputs. This theoretical foundation has facilitated the formulation of forestry ecological total factor productivity (FETFP) [5], a concept that extends conventional green total factor productivity by integrating ecological outputs. FETFP serves as a holistic indicator, capturing resource use efficiency within the forestry sector and the environmental–economic synergy it generates [5]. However, persistent reliance on extensive production models and outdated governance structures constrains resource allocation efficiency in China’s forestry sector [6]. Against this backdrop, a systematic inquiry into the formation mechanisms and enhancement pathways of FETFP holds substantial theoretical and practical relevance for advancing coordinated ecological and economic development [7].

The Decision of the Central Committee of the Communist Party of China on Further Deepening Reform Comprehensively and Advancing Chinese Modernization, adopted at the Third Plenary Session of the 20th Central Committee in 2024, explicitly advocates leveraging national standards to modernize traditional industries and accelerate their transformation through the dual integration of digital intelligence and green technologies. Amidst the broader tide of global technological innovation and industrial upgrading, the convergence of digital and intelligent technologies, commonly termed “digital intelligence integration”, has demonstrated considerable potential in optimizing resource allocation and reshaping industrial paradigms [8]. As a pivotal force driving the transformation of traditional industries, the digital economy has found wide application across agriculture, manufacturing, and services, fostering structural adjustment and boosting productivity [9,10,11].

Despite these advancements, empirical studies addressing the degree of digital intelligence integration remain sparse, with even fewer addressing its implications for forestry productivity. Digital intelligence integration is a key driving force behind digital transformation and Industry 4.0, representing a new business model that promotes the comprehensive intelligent upgrade of various industries by deeply integrating digital technologies with intelligent technologies [12]. In the context of forestry, a sector heavily reliant on natural resources and marked by long production cycles, high ecological sensitivity, and significant externalities, achieving sustained improvements in FETFP poses unique challenges. Research has shown that FETFP, which is an important indicator of resource use efficiency in the forestry sector, has been influenced by a variety of factors, including technological innovation, financial supply, and government policies such as fiscal incentives and ecological compensation mechanisms [5,13,14,15]. These studies provide valuable insights into the multi-dimensional factors affecting forestry ecological efficiency, but there is still a lack of focused analysis on the specific role that digital intelligence integration plays in this context.

Existing research suggests that digital intelligence integration has already demonstrated significant economic benefits in other sectors by improving resource utilization, reducing environmental burdens, and promoting green innovation [16,17]. However, there is limited empirical evidence exploring its impact on forestry. This gap is even more pronounced when considering that much of the literature on digital intelligence and the digital economy has focused on traditional production inputs, such as labor and capital, without adequately addressing the transformative potential of emerging factors like data assets and intelligent algorithms. As such, the relationship between digital intelligence integration and FETFP remains underexplored [15]. Furthermore, while studies on the digital economy have examined its role in improving green total factor productivity in forestry [18], the concept of digital intelligence integration, characterized by a fusion of digital and intelligent technologies, offers a more precise and technical model that could more effectively drive improvements in forestry productivity [19].

This study aims to explore the relationship between digital intelligence integration and forestry ecological total factor productivity. Using provincial panel data from China’s forestry sector, it investigates the underlying mechanisms and operational pathways through which digital intelligence integration influences the evolution of off-site forestry production efficiency. In doing so, it further explores the spatial heterogeneity of digital intelligence-enabled forestry development across regions. The study also suggests how digital strategies can be tailored for precise environmental governance by considering regional differences in digital readiness, local capacities, and institutional contexts. This research enriches the scholarly discourse on forestry production efficiency and offers empirical evidence to inform the high-quality, sustainable transformation of the forestry sector, both in China and other global contexts.

The structure of the paper is as follows: Section 2 presents the literature review and research hypotheses, Section 3 outlines the materials and methods, Section 4 discusses the results, Section 5 provides the discussion, and Section 6 concludes the study.

2. Theoretical Analysis and Research Hypotheses

The Cobb–Douglas production function posits that total factor productivity (TFP) growth is primarily driven by the optimal allocation of production inputs and resource use efficiency and innovation improvements, thereby serving as a fundamental engine of sustained industrial advancement [20,21,22]. Building upon this framework, FETFP extends the classical model by incorporating ecological dimensions into the assessment of output efficiency derived from traditional inputs, namely labor, land, and capital, within forestry production systems. However, China’s forestry sector has long been hindered by information asymmetries, misallocated resources, and a reliance on extensive management practices, all of which undermine the efficiency of resource allocation. For example, without reliable and real-time data, core forestry activities, such as resource inventorying, harvesting planning, and ecological restoration, are often based on subjective judgment or experiential knowledge, resulting in inefficiencies and environmental degradation. Moreover, the inherent complexity and dynamism of forestry ecosystems further challenge the capabilities of traditional management tools, which are ill-equipped to support real-time monitoring and adaptive decision-making, thus limiting the realization of ecological–economic synergies.

In recent years, integrating advanced technologies, including 5G communication, cloud computing, and artificial intelligence, has driven a paradigm shift in forestry practices through the rise of digital intelligence integration. Characterized by the dual processes of digitalization and intelligentization, this transformation has injected new dynamism into forestry production systems. First, leveraging big data platforms, digital intelligence systems enable the construction of dynamic monitoring frameworks that track key ecological indicators, such as forest stock volume, carbon sequestration potential, and biodiversity, in real time. Intelligent algorithms can then inform optimized decision-making in harvesting, tending, and restoration, thereby minimizing ecological losses and enhancing input-output efficiency [23]. Second, applying drones, remote sensing, and intelligent sensor networks helps overcome traditional forestry supervision’s spatial-temporal limitations, facilitating early detection and rapid response to ecological threats such as pest outbreaks or wildfires [24]. These innovations mitigate information asymmetries and promote a shift from experiential to data-driven forest governance, significantly enhancing management precision and responsiveness. Third, artificial intelligence and machine learning enable the identification of latent patterns in forest growth and ecological dynamics, supporting evidence-based decision-making. For instance, predictive models can optimize thinning regimes and harvest timing. At the same time, blockchain technologies can ensure complete traceability in forest product supply chains, thereby improving transparency, market value, and operational efficiency. Based on the above, we advance the following hypothesis:

H1: 

Digital intelligence integration significantly enhances FETFP.

According to industrial structure theory, upgrading and optimizing the industrial structure is a key pathway to improving resource allocation and alleviating inefficiencies caused by structural mismatches [25]. In the forestry context, challenges such as weak value chain integration, limited diversification, and low value-added output constrain long-term TFP growth. The adoption of digital intelligence technologies provides new impetus to overcome these barriers. On the one hand, integrating data-sharing platforms, resource coordination mechanisms, and interconnected information systems can improve forestry resources’ spatial and temporal alignment, thereby enhancing value chain coordination. Technologies such as remote sensing and big data analytics facilitate real-time monitoring and dynamic management, improving ecological traceability and increasing the market premium of forestry products, effectively transforming ecological assets into industrial advantages.

On the other hand, digital intelligence technologies empower forestry actors to more accurately perceive natural resource endowments, market trends, and policy shifts. For instance, satellite imagery combined with AI can help identify forest degradation patterns and predict future resource availability. This capability fosters a transition from intuitive to analytical decision-making and from extensive to precision-based management. By using data-driven tools, forestry managers can optimize harvest cycles and improve conservation efforts, ensuring sustainability while maximizing economic returns. Consequently, the forestry industry is steered toward higher-value, service-oriented, and diversified development trajectories [26], contributing to FETFP enhancement. We thus propose:

H2: 

Digital intelligence integration improves FETFP by optimizing the forestry industrial structure.

The Porter Hypothesis posits that well-designed environmental regulation can catalyze green innovation, thereby enabling simultaneous improvements in ecological and economic performance [27]. However, traditional forestry innovation often faces formidable obstacles, including constrained funding, high R&D costs, and significant market risks. Moreover, the long gestation periods and delayed payoffs associated with eco-technological development tend to suppress innovation incentives. Digital intelligence integration can mitigate these challenges by creating a more favorable informational and technological environment. For example, the fusion of remote sensing with machine learning algorithms has dramatically reduced the cost and enhanced the accuracy of carbon sink assessments, thereby providing technical infrastructure for carbon markets.

Digital technologies also lower the barriers to innovation by reducing costs and risks while enhancing the precision of innovation resource allocation. Intelligent platforms match R&D efforts with innovation needs in real time, optimizing input structures. For example, cloud-based platforms can enable forestry companies to collaborate globally, sharing knowledge and research on sustainable practices, thereby speeding up innovation cycles. With these tools, forestry enterprises are better positioned to identify green innovation opportunities and expedite the commercialization of low-carbon technologies, ultimately improving their adaptive capacity and competitiveness [28]. In addition, digital intelligence integration facilitates the diffusion of green knowledge and accelerates collaborative innovation processes, collectively boosting the sector’s innovation ecosystem [29]. In this regard, we hypothesize:

H3: 

Digital intelligence integration improves FETFP by promoting green technological innovation in forestry.

Fiscal support represents a pivotal policy instrument in advancing the transformation and modernization of agriculture and forestry [30]. It is vital in enabling technological upgrading, structural optimization, and productivity enhancement. In the context of this study, fiscal resources—particularly subsidies, special project funding, and infrastructure investments—can alleviate the financial constraints forestry enterprises face in adopting digital and intelligent technologies. This reduces the marginal cost of the green transition and amplifies the ecological empowerment effects of digital intelligence integration.

Furthermore, fiscal support fosters the development of collaborative innovation platforms that bring together government agencies, enterprises, and research institutions. These platforms accelerate the R&D and application of digital technologies in forestry, thereby supporting the improvement of FETFP. In addition to its material impact, fiscal support also sends a strong policy signal, demonstrating governmental commitment to green development and catalyzing private sector investment in forestry’s digital transformation. This crowding-in effect further amplifies the technological and ecological dividends of digital intelligence integration. Accordingly, we posit the following hypothesis:

H4: 

Fiscal support positively moderates the impact of digital intelligence integration on FETFP. In essence, stronger fiscal support amplifies the influence of digital intelligence integration on FETFP, while weaker fiscal support diminishes its effectiveness.

3. Materials and Methods

3.1. Overview of the Study Area

This study focuses on 30 provinces (autonomous regions and municipalities, excluding Hong Kong, Macau, Taiwan, and Tibet) in China, spanning a diverse range of regions including the eastern, southern, northern, and western parts of the country. These provinces exhibit significant regional heterogeneity in terms of population size, economic development, and geographical environment. The variations in forestry resources, ecological conditions, and industrial development across these provinces provide a rich contextual backdrop and diverse comparative data to explore the impact of digital intelligence integration on FETFP.

Many of these provinces play a pivotal role in forestry development, with some possessing abundant forest resources and high forest coverage rates. They have made substantial progress in ecological protection, green development, and the transformation of the forestry industry. Additionally, as digital technologies continue to proliferate, local governments have actively implemented strategies to promote digital transformation and facilitate the adoption of intelligent technologies. Against this backdrop, this study aims to examine how digital intelligence integration influences forestry ecological efficiency, while uncovering regional variations and the underlying mechanisms at play. The overview map of the study area is shown in Figure 1.

3.2. Double Machine Learning Model

This study centers on investigating the influence of digital intelligence integration on FETFP, with a particular focus on identifying its underlying mechanisms. This paper adopts the Double Machine Learning (DML) framework to rigorously estimate causal effects within a high-dimensional setting. The DML approach offers distinct methodological advantages, particularly in its ability to automatically select relevant control variables from a large pool of potential covariates, thereby mitigating multicollinearity, a common challenge in traditional panel data analyses, and enhancing the robustness and precision of variable selection [31].

In addition, the DML framework is well-suited to modeling complex, nonlinear interactions among variables, making it an ideal tool for analyzing systems such as FETFP, where ecological and economic factors are interdependent and often interact through intricate pathways.

Accordingly, this study employs a partially linear DML model, following the methodological framework proposed by Zhang and Li [32], to empirically assess the causal impact of digital intelligence integration on FETFP. This approach allows for a more accurate and reliable treatment effect estimation while controlling for confounding factors in a high-dimensional data environment.

$$Y_{it}={\theta }_0{D}_{it}+\widehat{g}\left(X_{it}\right)+U_{it}$$

(1)

$$E\left(U_{it}|D_{it},X_{it}\right)=0$$

(2)

$$D_{it}=m\left(X_{it}\right)+V_{it}$$

(3)

$$E\left(V_{it}|X_{it}\right)=0$$

(4)

In Equations (1)–(4), i denotes the province and t denotes the year. $Y_{it}$ represents the dependent variable (or the mediating variable in the mechanism analysis); $D_{it}$ represents the main explanatory variable; ${\theta }_0$ is the coefficient of interest to be estimated; $X_{it}$ denotes the high-dimensional set of control variables, whose functional form is estimated using a support vector machine (SVM) algorithm; $U_{it}$ is the error term with a conditional mean of zero; $m\left(X_{it}\right)$ represents the regression function of the treatment variable on the high-dimensional controls; and $V_{it}$ is its associated error term, also with a conditional mean of zero. In addition, both province and year fixed effects are included in the model to control for unobserved heterogeneity. The analysis idea of DDML is shown in Figure 2.

The data used here are from publicly available databases, extracted and compiled by the author team (for specific sources, see Section 3.3 Data Sources), and regression analysis was performed using the data processing software Stata 16.0.

3.3. Data Sources

To ensure data continuity and availability, this study selects 30 provinces (autonomous regions, municipalities, excluding Hong Kong, Macau, Taiwan, and Tibet) in China as the research subjects, using panel data from 2012 to 2022.

The ecological output data is calculated based on the annual land data coverage from Wuhan University (spatial resolution of 30 m) and the ERA5-Land dataset (spatial resolution of 100 m) provided by organizations such as the European Centre for Medium-Range Weather Forecasts. The Peking University Digital Inclusive Finance Index is published by the Peking University Digital Finance Research Center and Ant Group Research Institute. The number of smart patent applications comes from the Yi Patent Platform. The industrial robot installation density is calculated based on the industrial robot installation data by the IRF Alliance, which corresponds to the national economic industry classification code, with adjustments made based on the proportion of employment in each province. The number of artificial intelligence (AI) enterprises comes from Tianyancha. The degree of AI adoption is measured by the ratio of the disclosed market value of machines to the number of employees in listed companies within the region, following existing studies as a proxy for AI adoption in firms. The remaining data are sourced from the “China Statistical Yearbook,” “China Rural Statistical Yearbook,” “China Forestry and Grassland Statistical Yearbook,” and the Wind database [5]. The aforementioned variables are aggregated at the provincial level.

3.4. Definition of Variables

3.4.1. Forestry Ecological Total Factor Productivity

The dependent variable in this study is forestry ecological total factor productivity (FETFP), a comprehensive metric designed to assess the economic performance, ecological value, and environmental impact derived from the utilization of forestry production inputs such as capital, land, labor, and energy [33]. Unlike the conventional green total factor productivity indicator that only considers economic benefits, FETFP also incorporates ecological output into the production results system, which is more in line with the strategic vision of the high-quality development of forestry under the framework of China’s environmental civilization.

Following the methodologies proposed by Dong et al. [5], the input variables selected for this study include capital investment, land area under forestry operations, energy consumption, and labor input. The output system encompasses three categories: (1) economic output, proxied by the gross output value of the forestry sector; (2) ecological output, including regulatory, support, and cultural ecosystem services provided by forests; and (3) environmental pollutants, specifically emissions of sulfur dioxide (SO₂), wastewater, and solid waste resulting from forestry activities.

To estimate ecological service values, this study draws on the frameworks Xie et al. [34] and Yin [35] developed, employing an improved equivalent factor method. This approach integrates corrections based on region-specific factors such as net primary productivity (NPP), annual precipitation, agricultural net profit, and social adjustment coefficients, thus enhancing the ecological valuation’s accuracy and regional applicability. Environmental emissions associated with forestry production are derived through an indirect estimation method, considering the limitations of direct pollutant data at the provincial level.

To quantify FETFP, the study adopts the Epsilon-Based Measure (EBM) model in combination with the Global Malmquist–Luenberger (GML) index, as proposed by Tone [36]. This methodological approach allows for considering desirable and undesirable outputs simultaneously within a unified production efficiency framework, offering a robust and dynamic assessment of FETFP across provinces and over time. The complete indicator system and logic are shown in

Table 1. Calculation Indicators of Forestry Ecological Total Factor Productivity.

Category	Subcategory	Basic Indicator	Specific Evaluation Metric	Unit
Inputs	Input Indicators	Capital	Completed investment in forestry fixed assets (adjusted using perpetual inventory method)	100 million CNY
		Land	Area of forestry land use	10,000 hectares
		Energy	Total energy consumption (standard coal) × forestry output share in regional GDP	10,000 tons
		Labor	Number of on-the-job employees in the forestry sector	100 persons
Outputs	Desirable Outputs	Economic Output	Output value of the forestry industry	100 million CNY
		Ecological Output	Value of forest regulation services	100 million CNY
			Value of forest support services	100 million CNY
			Value of forest cultural services	100 million CNY
	Undesirable Outputs	Forestry SO2 Emissions	Industrial SO2 emissions × forestry secondary industry output share in total industrial output	10,000 tons
		Forestry Wastewater Discharge	Industrial COD wastewater emissions × forestry secondary output share in industrial output	10,000 tons
		Forestry Solid Waste	Industrial solid waste generation × forestry secondary output share in industrial output	10,000 tons

and Figure 3.

3.4.2. Digital Intelligence Level

The key explanatory variable in this study is the Digital Intelligence Level, which reflects the degree of integration between digital and intelligent technologies across provinces. At present, there is no unified standard for measuring digital intelligence integration. Existing literature has approached this challenge through various methodologies. For example, some studies utilize the establishment of “New Generation Artificial Intelligence Innovation Development Pilot Zones” as quasi-natural experiments to implement difference-in-differences (DID) models and assess the economic impacts of digital intelligence development [37,38]. Others construct multi-dimensional evaluation frameworks based on indicators of regional digitalization and intelligentization [39].

Given the provincial-level focus of this study, relying solely on pilot zone designations may not sufficiently capture inter-provincial heterogeneity. In light of this, and drawing on the framework proposed by prior research, this study conceptualizes digital intelligence integration as a three-dimensional construct comprising (1) digital infrastructure development, representing foundational support; (2) digital application penetration, reflecting core functional capabilities; and (3) intelligent innovation deployment, signifying advanced technological progress. Compared to the single digital economy, digital intelligence integration places more emphasis on the depth of technological innovation application and the driving role of intelligent technologies in the development of society and industries [19].

Twelve representative indicators are selected to capture these three dimensions comprehensively. To enhance the composite index’s scientific rigor and accuracy, the entropy weight method is employed to endogenously determine the weights of individual indicators and calculate a comprehensive Digital Intelligence Level score for each province. The complete evaluation index system is presented in

Table 2. Evaluation Index System for Digital Intelligence Level.

Primary Indicator	Secondary Indicator	Tertiary Indicator	Indicator Nature	Unit	Weight
Digitalization	Digital Infrastructure	Total Postal and Telecommunication Services	Positive	100 million CNY	0.0750
		Mobile Phone Users	Positive	10,000 households	0.0419
		Broadband Internet Subscribers	Positive	10,000 users	0.0525
		Length of Long-Distance Cable Lines	Positive	10,000 km	0.0316
	Digital Application Penetration	Number of Internet Domain Names	Positive	10,000 domains	0.1101
		E-commerce Sales Volume	Positive	100 million CNY	0.1157
		Urban Employment in Information Transmission, Software, and IT Services	Positive	10,000 persons	0.0951
		Peking University Digital Financial Inclusion Index	Positive	/	0.0192
Intelligentization	Intelligent Innovation Deployment	Number of Smart Patent Applications	Positive	Items	0.1659
		Industrial Robot Installation Density	Positive	/	0.1203
		Number of AI Enterprises	Positive	Units	0.1520
		AI Adoption Level in Enterprises	Positive	/	0.0207

.

Among them, the degree of artificial intelligence adoption is measured by using the ratio of the disclosed market value of machines to the number of employees in listed companies within the region, as a proxy for the adoption level of AI technology in firms, following existing studies [40].

3.4.3. Control Variables

By prior research [5], this study incorporates a set of control variables that may influence FETFP to account for potential confounding effects. The selected control variables encompass key dimensions of regional socioeconomic development, ecological conditions, human capital, and technological advancement. Specifically, the following variables are included:

Economic development (ED), measured by the natural logarithm of provincial per capita GDP, serves as a proxy for overall economic capacity;

Soil erosion (SE), captured by the natural logarithm of the total area affected by water and soil loss in each province, reflects ecological degradation pressures;

Forest coverage (FC), represented by the officially reported forest coverage rate, indicates the extent of forest ecosystems within each region;

Education expenditure (EE), measured as the ratio of education spending to total fiscal expenditure, serves as a proxy for regional investment in human capital;

Labor force level (LF), proxied by the natural logarithm of the number of employed persons, reflects the scale of human resource input;

The level of informationization (IL), measured by the ratio of the total volume of postal and telecommunication services to regional GDP, captures the level of digital infrastructure development.

3.4.4. Mechanism Variables

Based on the theoretical analysis above, this study explores how digital intelligence integration affects FETFP, focusing on two key pathways: forestry industrial structure and green technological innovation.

First, the forestry industry structure follows the approach of Dong et al. [41], using the industrial structure hierarchy coefficient method. This method assigns values of 1, 2, and 3 to the primary, secondary, and tertiary industries, respectively, and calculates the forestry industry structure index based on the proportion of output value from each industry. Specifically, as the share of the secondary and tertiary industries in forestry increases, the industry structure level improves, indicating that the forestry sector is gradually shifting from a resource extraction-based model to a higher value-added processing and service-oriented industry. This transformation reflects the upgrade from a resource-based economy to a green, intelligent, and innovation-driven economic model [42].

Second, green technological progress is considered a crucial factor in enhancing forestry ecological efficiency. To measure this capability, this study uses the logarithm of the number of green patents granted in a region as an indicator, reflecting the region’s capacity for green innovation. The stronger the green innovation capacity, the more conducive it is to promoting regional green development and technological accumulation in improving ecological productivity [43], which in turn supports the continuous improvement of forestry ecological efficiency.

3.4.5. Moderating Variable

As suggested in the previous analysis, fiscal support may enhance the effect of digital intelligence integration on FETFP and is, therefore, treated as a moderating variable. Drawing on the approach of Deng et al. [44], the logarithm of the total amount of provincial government expenditure on agriculture and forestry measures fiscal support.

Descriptive statistics are presented in

Table 3. Descriptive statistics results.

Variable	Obs	Mean	S.t	Min	Max
FETFP	330	1.0454	0.0906	0.8301	1.3838
DIL	330	0.1235	0.0991	0.0169	0.5497
ED	330	9.3322	0.4631	8.6605	10.7525
SE	330	7.9859	1.1657	4.0738	9.5905
FC	330	33.7929	17.9881	4.2400	66.8000
EE	330	0.0411	0.0418	0.0137	0.2678
LF	330	7.6017	0.7642	5.7394	8.8479
IL	330	0.0629	0.0557	0.0170	0.2785

, providing a comprehensive overview of the key variables in the analysis. The mean value of FETFP is 1.0454, with a standard deviation of 0.0906. The values range from a minimum of 0.8301 to a maximum of 1.3838, indicating substantial regional variation in forestry productivity across different provinces. The relatively high standard deviation suggests that there are notable disparities in the forestry sector’s efficiency and development across regions, likely influenced by factors such as local resource availability, infrastructure, and policy support.

The mean of DIL is 0.1235, with a standard deviation of 0.0991, and it ranges from 0.0169 to 0.5497. This variation underscores the pronounced regional disparities in the adoption and integration of digital intelligence technologies within the forestry sector. The low minimum and high maximum values reflect a spatial non-equilibrium in digital intelligence development. Some provinces have made significant strides in adopting digital tools, while others remain far behind. This discrepancy aligns with the broader trend in China’s digital transformation, where more developed regions have access to advanced technologies and skilled labor, while underdeveloped areas struggle with access to the necessary infrastructure and investment.

The control variables’ distributional characteristics are consistent with previous literature’s findings and align well with China’s regional development realities [14].

4. Results

4.1. DML Estimation of the Baseline Results

Based on the theoretical analysis and research design presented earlier, and considering that the SVM can effectively handle nonlinear relationships [45], and that the number of sample splits in small sample studies is recommended to be no less than 5 [46], the SVM algorithm is chosen for estimation. The K-fold ratio is set to 5, and empirical tests are conducted to examine the impact of DIL on FETFP.

The empirical findings are summarized in

Table 4. Baseline regression results.

	(1)	(2)
Variable	FETFP	FETFP
DIL	0.1807 **	0.1997 ***
	(0.0713)	(0.0665)
First-order control variables	YES	YES
Second-order control variables	NO	YES
Time Fixed	YES	YES
Province Fixed	YES	YES
Observations	330	330

Note: ** and *** indicate significance at the 5%, and 1% levels, respectively; robust standard errors are reported in parentheses.

. Column (1) presents the baseline estimation results, including first-order control variables and year and province fixed effects. The coefficient for DIL is 0.1807 and is statistically significant at the 5% level. This positive relationship suggests that higher levels of digital intelligence integration are associated with an increase in FETFP, providing initial empirical support for the hypothesis. Column (2) extends the model by incorporating the squared terms of control variables to capture potential nonlinearities. The estimated coefficient for the digital intelligence level increases to 0.1997, with significance at the 1% level, indicating a more pronounced and robust effect. This reinforces the idea that the integration of digital technologies in forestry not only improves productivity but does so in a nonlinear fashion, where the impact intensifies as digital intelligence levels rise.

These results provide strong empirical support for H1, confirming that DIL significantly improves FETFP. In practice, integrating advanced digital intelligence technologies, such as blockchain for ecological traceability, artificial intelligence for environmental risk prediction, and innovative platforms for precision forest governance, has markedly strengthened the resilience and sustainability of forestry ecosystems. These innovations contribute directly to enhanced ecological outputs and more efficient resource utilization, thereby elevating overall FETFP. This suggests that digital intelligence is not just a tool for economic growth but also plays a key role in improving the ecological sustainability of the forestry sector.

To further disentangle the specific contributions of individual components within the broader framework of digital intelligence integration, this study conducts a disaggregated analysis using three secondary indicators: digital infrastructure development, digital application penetration, and intelligent innovation deployment. The corresponding estimation results are reported in

Table 5. Decomposition term regression results.

	(1)	(2)	(3)
Variable	FETFP	FETFP	FETFP
Digital Infrastructure Development	0.9205 ***
	(0.1968)
Digital Application Penetration		0.1642
		(0.1162)
Intelligent Innovation Deployment			0.4694 ***
			(0.1673)
First-order control variables	YES	YES	YES
Second-order control variables	YES	YES	YES
Time Fixed	YES	YES	YES
Province Fixed	YES	YES	YES
Observations	330	330	330

Note: *** indicate significance at the 1% levels; robust standard errors are reported in parentheses.

.

The results reveal that digital infrastructure development and intelligent innovation deployment positively affect FETFP at the 1% significance level. Notably, the estimated coefficients for these two components exceed those of the aggregate digital intelligence index, suggesting that current improvements in FETFP are primarily driven by advancements in foundational digital capacity and the deep integration of intelligent technologies into forestry systems. This highlights the crucial role of robust digital infrastructure, such as high-speed internet networks, cloud computing resources, and data centers, in enabling more efficient forest management. Similarly, the deployment of cutting-edge technologies like AI and machine learning for risk prediction and forest monitoring appears to be more impactful than general digital intelligence measures. These findings underscore the strategic importance of investing in digital infrastructure and accelerating high-level innovation to support sustainable forestry development. In practice, strong digital infrastructure not only facilitates data collection and analysis but also enables the seamless operation of more sophisticated digital applications across large, spatially dispersed forestry areas.

In contrast, the effect of digital application penetration on FETFP is statistically insignificant. This result may be attributed to the structural characteristics of the forestry sector—namely, long production cycles, spatial dispersion, and the inherent complexity of ecological processes—which pose challenges to the seamless integration of digital applications across all stages of production. Consequently, digital tools remain underutilized or are applied in fragmented ways, limiting their transformative impact on forestry productivity [47]. These findings highlight the necessity of optimizing digital application pathways and tailoring technological solutions to the unique operational demands of the forestry industry.

4.2. Robustness Tests of DML

To ensure the reliability of the baseline findings, a series of robustness tests is conducted, including adjustments to cross-validation ratios, outlier treatment, algorithm substitution, and the inclusion of interaction fixed effects. The results are summarized in

Table 6. Robustness test.

Variable	(1)	(2)	(3)	(4)	(5)	(6)
	Sample Split		Winsorization		Algorithm Change	Interaction FE
	1:2	1:6	3%	5%	Gradboost
DIL	0.2076 ***	0.2035 ***	0.2386 ***	0.2217 ***	0.2363 *	0.2006 ***
	(0.0673)	(0.0676)	(0.0639)	(0.0535)	(0.1306)	(0.0665)
First-order control variables	YES	YES	YES	YES	YES	YES
Second-order control variables	YES	YES	YES	YES	YES	YES
Time Fixed	YES	YES	YES	YES	YES	YES
Province Fixed	YES	YES	YES	YES	YES	YES
Observations	330	330	330	330	330	330

Note: * and *** indicate significance at the 10%, and 1% levels, respectively; robust standard errors are reported in parentheses.

.

4.2.1. Robustness Test of Adjusting Sample-Splitting Ratios

The baseline model uses 5-fold cross-validation for estimation. To test the model’s sensitivity to sample partitioning methods, this study adjusts the cross-validation settings by using two different training-to-testing set ratios of 1:3 and 1:6 to re-estimate the model, in order to observe the extent of interference from sample splitting on the regression results. This adjustment is crucial as varying the ratio can sometimes lead to discrepancies in model performance, particularly in terms of overfitting or underfitting. The re-estimated results demonstrate that the effect of digital intelligence integration on FETFP remains statistically significant under both partitioning schemes. Notably, the consistency in coefficient sign and significance levels with the baseline model confirms that the main results are robust to changes in sample-splitting strategies. This reinforces the reliability of the findings, suggesting that the observed relationship between digital intelligence and FETFP is not sensitive to changes in the sample partitioning methodology.

4.2.2. Robustness Test of Removing Outliers

A two-sided winsorization is applied to all continuous variables to mitigate the potential influence of extreme values on parameter estimation. Extreme values or outliers can disproportionately affect model outcomes, especially in regression analysis, where they can distort the relationships between variables. Two truncation thresholds—(3%, 97%) and (5%, 95%)—are adopted to test the sensitivity of results to outlier handling. The adoption of these thresholds ensures that extreme values do not unduly influence the regression estimates. The regression outcomes indicate that the estimated positive effect of digital intelligence integration on FETFP remains robust and statistically significant following outlier adjustment, thereby validating the integrity of the baseline estimates. This suggests that the main findings are not driven by a few extreme observations, further supporting the validity of the results.

4.2.3. Robustness Test of Resetting the Machine Learning Algorithm

Given that different machine learning algorithms vary in function approximation capabilities, parameter tuning procedures, and complexity control, reliance on a single algorithm may lead to estimation bias. The SVM used in the baseline model is replaced with a gradient-boosting algorithm to test the stability of results across learning algorithms. Gradient boosting, being an ensemble method, is known for its strong predictive power and ability to capture complex patterns in the data. The results continue to show a significantly positive effect of digital intelligence integration on FETFP at the 10% level, reinforcing the robustness of the findings under alternative algorithmic settings. This provides further confidence that the observed relationship is not dependent on the specific choice of machine learning model, indicating that the positive effect of digital intelligence is consistently observed across different methods.

4.2.4. Robustness Test of Introducing Province–Year Interaction Fixed Effects

To account for unobservable heterogeneity that evolves simultaneously across provinces and over time, such as differences in ecological conditions, digital infrastructure quality, and policy responsiveness, province–year interaction fixed effects are introduced into the baseline specification. This step is particularly important as it controls for factors that vary both across regions and over time, which could otherwise bias the results. The introduction of province–year interaction fixed effects ensures that the model accounts for both temporal and spatial variations in unobserved factors, such as regional policy changes or shifts in local digital infrastructure development. The estimated coefficient for digital intelligence integration remains significantly positive, indicating that the observed effects are not driven by time-varying structural characteristics specific to individual provinces. This further strengthens confidence in the causal interpretation of the results, as it suggests that the relationship between digital intelligence integration and FETFP is robust even after controlling for region- and time-specific factors.

4.3. Endogeneity Test of the Baseline Results

To avoid the potential endogeneity issues arising from omitted variable bias, reverse causality, and other factors that may affect the impact of digital intelligence integration on forestry ecological total factor productivity, this study uses instrumental variables for the endogeneity test.

First, following the approach of Dong et al. [41], the number of post offices per million people in 1984 is selected as an instrumental variable. On the one hand, the number of post offices reflects the early information flow and communication capabilities of the region, and the accumulation of information infrastructure represented by it may have a path-dependent effect on the current level, satisfying the relevance requirement. On the other hand, as a historical variable, it has no direct correlation with the current error term of forestry ecological total factor productivity, ensuring strong exogeneity, and can effectively serve as an instrumental variable for digital intelligence integration.

Second, the distance from each provincial capital to Hangzhou is used as another instrumental variable. The reasoning is that, on one hand, digital intelligence integration is closely related to geographic location, satisfying the relevance requirement. On the other hand, the distance from the provincial capital to Hangzhou is an objective spatial distance, thus meeting the exogeneity requirement.

Based on this, we interact with the historical number of post offices and city distances with the previous period’s digital intelligence level and conduct an instrumental variable test. The results (

Table 7. Endogeneity analysis: Instrumental Variable Method.

Variable	1984 Post Offices per Million Interacted with the Previous Period’s DIL	Distance from the Provincial Capital to Hangzhou Interacted with the Previous Period’s DIL
DIL	0.2126 ***	0.1445 *
	(0.0800)	(0.0744)
First-order control variables	YES	YES
Second-order control variables	YES	YES
Time Fixed	YES	YES
Province Fixed	YES	YES
Wald F	74.605	940.158
Observations	300	290

Note: * and *** indicate significance at the 10%, and 1% levels, respectively; robust standard errors are reported in parentheses.

) show that the coefficients of the key estimated parameters remain significant, and the Wald F-values proposed using the 2SLS method are 74.605 and 940.158, respectively. This indicates that the instrumental indicator specification is not as problematic as the instrumental indicator, demonstrating that the conclusions remain reliable even after considering the endogeneity issue.

4.4. Mechanism Analysis

This section examines the internal mechanisms through which digital intelligence integration enhances FETFP to empirically validate the theoretical hypotheses proposed in the conceptual framework. Specifically, it investigates the mediating roles of forestry industrial development and green technological innovation and the moderating effect of fiscal support.

4.4.1. Forestry Industrial Development

As reported in Column (1) of

Table 8. Mechanism analysis: Forestry Industrial Development, Green Technological Innovation, and Fiscal Support.

	(1)	(2)	(3)
Variable	Forestry Industrial Development	Green Technological Innovation	FETFP
DIL	0.2056 ***	10.3403 ***
	(0.0197)	(0.6232)
Fiscal Support × DIL			0.2948 ***
			(0.0098)
First-order control variables	YES	YES	YES
Second-order control variables	YES	YES	YES
Time Fixed	YES	YES	YES
Province Fixed	YES	YES	YES
Observations	330	330	330

Note: *** indicate significance at the 1% levels; robust standard errors are reported in parentheses.

, the estimated coefficient of digital intelligence level on forestry industrial development is 0.2056, statistically significant at the 1% level. This implies that a one-unit increase in digital intelligence level is associated with an average increase of 0.2056 units in the level of forestry industrial development. From an economic perspective, this finding provides strong empirical support for H2, which hypothesizes that digital intelligence integration enhances forestry industrial development. The significant positive relationship suggests that digital technologies, through their ability to improve informatization, standardization, and specialization in forestry operations, drive industrial upgrading. This is crucial for improving the modernization and competitiveness of the forestry sector. By enhancing real-time data acquisition, intelligent coordination, and factor allocation, digital intelligence helps reduce inefficiencies and promotes more productive and sustainable forestry practices. As a result, forestry industries become better equipped to respond to both market demands and environmental challenges, ultimately enhancing FETFP.

4.4.2. Green Technological Innovation

Column (2) of Table 8 presents the estimation results for green technological innovation. The coefficient of digital intelligence level is 10.3403, significant at the 1% level, indicating that digital intelligence integration substantially boosts regional capacity for green innovation. This strongly supports H3, which posits that digital intelligence enhances green technological innovation. The results suggest that by enabling technologies like AI, big data, and cloud computing, digital intelligence provides the tools necessary for more effective research, development, and application of green technologies in forestry. The significant effect of digital intelligence on green innovation reflects its role in overcoming traditional barriers to green innovation diffusion, such as slow adoption rates and high costs [48]. By streamlining deployment processes and improving cost-effectiveness, digital technologies accelerate the transition toward a more sustainable and innovation-driven forestry sector, which directly contributes to higher FETFP.

4.4.3. Fiscal Support

Column (3) of Table 8 reports the interaction effect between fiscal support and digital intelligence level. The estimated coefficient is 0.2948, significant at the 1% level, indicating that fiscal support positively moderates the relationship between digital intelligence integration and FETFP, thereby confirming H4. From an economic perspective, this highlights the critical role of fiscal support in accelerating the digital transformation of the forestry sector. Fiscal incentives help alleviate financial barriers for forestry enterprises by funding digital infrastructure, technology adoption, and skilled human capital development. By reinforcing the policy-market synergy, fiscal support ensures that the integration of digital technologies becomes more effective and widespread. The interaction between fiscal support and digital intelligence integration helps create a virtuous cycle, where government incentives encourage innovation, which in turn leads to more sustainable and efficient forestry practices. This directly boosts FETFP, validating the importance of fiscal support in fostering digital and ecological progress in the sector.

4.5. Heterogeneity Analysis

To investigate whether the impact of digital intelligence integration on FETFP varies across institutional and policy environments, this study conducts a heterogeneity analysis based on environmental expenditure levels and green finance development. These dimensions capture the variation in public fiscal investment and financial system support, both critical in shaping the effectiveness of digital intelligence deployment in forestry.

4.5.1. Heterogeneity in Environmental Expenditure

The proportion of environmental protection expenditure in total fiscal expenditure is employed as a proxy for a region’s ecological spending level. As shown in

Table 9. Heterogeneity analysis: Environmental Expenditure and Green Finance Development.

	(1)	(2)	(3)	(4)
Variable	Low Environmental Expenditure	High Environmental Expenditure	Low Level of Green Finance	High Level of Green Finance
DIL	−0.0022	0.4277 ***	0.0531	0.1518 *
	(0.0762)	(0.1164)	(0.1217)	(0.0782)
First-order control variables	YES	YES	YES	YES
Second-order control variables	YES	YES	YES	YES
Time Fixed	YES	YES	YES	YES
Province Fixed	YES	YES	YES	YES
Observations	162	168	155	175

Note: * and *** indicate significance at the 10%, and 1% levels, respectively; robust standard errors are reported in parentheses.

, the estimated coefficient for digital intelligence integration in areas with high environmental expenditure is 0.4277, statistically significant at the 1% level, indicating a strong positive effect on FETFP. In contrast, in low environmental expenditure regions, the coefficient is −0.0022 and fails to reach statistical significance, suggesting an absence of measurable impact.

Higher levels of environmental spending provide essential fiscal and institutional support for ecological governance, facilitating the embedding digital intelligence of digital intelligence tools into ecosystem monitoring, resource allocation, and environmental risk management systems. Furthermore, substantial fiscal commitment enhances the priority of ecological transformation in local policy agendas, encouraging governments and enterprises to adopt and integrate digital solutions for green development [49]. These dynamics collectively contribute to more effective digital intelligence implementation and productivity-enhancing effects in the forestry sector.

Conversely, regions with limited environmental expenditure, underdeveloped infrastructure, and constrained ecological investment hinder the deployment and application of digital technologies. The lack of supportive governance capacity and ecological resources undermines the potential of digital intelligence integration to drive improvements in ecological efficiency, thereby diminishing its contribution to FETFP.

4.5.2. Heterogeneity in Green Finance Development

Drawing on the methodology of Zhang et al. [50], a provincial green finance development index is constructed to reflect the comprehensiveness of green financial systems. The index incorporates five dimensions: green credit, green securities, green insurance, green investment, and carbon finance. The regression results in Table 9 reveal that digital intelligence integration in provinces with well-developed green finance systems exerts a significantly positive effect on FETFP at the 10% level. However, the relationship is statistically insignificant in regions with low levels of green finance.

This discrepancy may be attributed to the role of green finance in mobilizing capital toward environmentally sustainable and technologically advanced projects. By reducing the cost and risk of financing, green finance facilitates the adoption of digital intelligence technologies in forestry enterprises. Moreover, robust green finance systems are often accompanied by strict disclosure standards and environmental risk management mechanisms [51], further incentivizing firms to invest in innovation, digitalization, and ecological modernization. These conditions amplify the productivity-enhancing effects of digital intelligence integration in forestry.

In contrast, underdeveloped green financial ecosystems often lack adequate institutional frameworks and suffer from inefficient capital allocation, resulting in insufficient support for digital and ecological upgrading. Consequently, the enabling role of digital intelligence technologies in improving FETFP is significantly weakened in these regions.

5. Discussion

This study provides empirical evidence that the integration of digital intelligence has become a key driver for enhancing FETFP. This finding aligns with international research on the role of the digital economy, smart agriculture, and digital transformation in improving agricultural development and system resilience [3,5]. The construction of digital infrastructure not only provides efficient information acquisition, data transmission, and intelligent decision-making platforms for forestry, but also lays a solid foundation for collaboration and risk-sharing across the entire forestry supply chain. Additionally, the promotion of digital applications, such as smart forestry machines, precision forestry, and intelligent forest management, has significantly improved forestry’s ability to respond to external shocks like climate change and ecological degradation. This supports existing literature that emphasizes the importance of digital technologies as a core driver for enhancing the resilience of traditional industries [52]. However, the study further reveals that relying solely on smart innovation applications is insufficient for achieving systemic improvements. This resonates with research that cautions that “digital innovation disconnected from practical application can lead to technological silos,” highlighting the importance of not only adopting digital technologies but also deeply integrating them into forestry operations [9]. As such, it is essential that the introduction of digital technologies is closely aligned with practical, on-the-ground needs to achieve true modernization and ensure the sustainable development of the forestry sector.

The analysis also uncovers that digital intelligence integration boosts forestry ecological productivity through three main channels. First, by promoting green technologies, it enhances innovation and self-restoration within forestry systems. Second, it fosters industrial diversification, which in turn improves the risk dispersion and adaptive capacity of the forestry sector. Lastly, it leverages the synergistic effects of fiscal support, which improves the ecological efficiency of forestry. These findings are consistent with innovation diffusion theory and industrial diversification theory, and they help address gaps in the current research on the sustainability of forestry industries [53].

Moreover, the regional heterogeneity analysis reveals that the effect of digital intelligence integration on forestry ecological productivity is more pronounced in regions with lower fiscal support for agriculture and more developed green finance systems. In regions with lower fiscal support, the integration of digital intelligence can compensate for insufficient policy investment. By enhancing the forestry system’s resilience through information empowerment, technological advancement, and industrial collaboration, it strengthens the system’s intrinsic resilience. This aligns with the principle of diminishing marginal returns and the reverse compensation mechanism theory [54]. On the other hand, in regions with more robust green finance systems, well-established institutional guarantees and risk prevention mechanisms support the widespread adoption of digital technologies and green finance innovation, thus amplifying the enabling effects of digital intelligence integration. This supports findings in the literature that emphasize the role of the institutional environment in moderating the effects of digital transformation, while also highlighting the risks of misuse and resource misallocation in areas with weak green finance support [55].

In conclusion, this study not only provides new empirical evidence for enhancing FETFP but also offers deeper insights into how digital intelligence integration functions across different regional contexts. It underscores the pivotal role of digital technologies in promoting sustainable development and facilitating the intelligent transformation of the forestry sector. These findings offer valuable guidance for policymakers, emphasizing the need for region-specific digital strategies and advancing forestry in alignment with green development and intelligent transformation.

6. Conclusions and Implications

This study uses panel data from 30 Chinese provinces between 2012 and 2022 to empirically examine the effects of digital intelligence integration on FETFP. The findings reveal that digital intelligence significantly enhances FETFP, and this relationship remains robust under various robustness tests and instrumental variable estimations addressing potential endogeneity. Mechanism analysis further shows that the positive effects of digital intelligence integration are primarily transmitted through upgrading the forestry industrial structure, stimulating green technological innovation, and leveraging fiscal support. Heterogeneity analysis suggests that the impact of digital intelligence is more pronounced in provinces with higher levels of environmental expenditure and advanced green finance, highlighting the importance of institutional and financial environments in shaping digital transformation outcomes.

However, there are some limitations to this study. First, the provincial-level analysis may overlook intra-provincial variations. Second, the constructed digital intelligence index may not fully capture institutional or behavioral factors that influence technology adoption. Future research could benefit from using enterprise-level or spatially disaggregated data to explore micro-level mechanisms. Additionally, with the rise of carbon markets and digital carbon accounting tools, future work could investigate how digital intelligence supports forestry’s role in achieving carbon neutrality goals.

To effectively integrate digital intelligence, accelerating digital infrastructure development is essential. However, many regions face challenges such as fragmented systems, poor integration, and data silos, which hinder coordinated digital transformation in forestry production, monitoring, and resource management. National and provincial governments should develop clear digital forestry plans, setting infrastructure targets, implementation milestones, and regional priorities. These plans should promote the creation of integrated regional information platforms and forestry data-sharing systems. Investment should also be expanded in core enabling technologies, such as ecological sensor networks, edge computing terminals, satellite-based forest monitoring, and positioning systems, to enable real-time, intelligent tracking, allocation, and response within forestry ecosystems. Moreover, efforts should be made to strengthen local capacity, ensuring that forestry personnel are trained to operate and maintain digital systems effectively. A multi-tiered and regionally adaptive infrastructure strategy will provide a strong foundation for enhancing FETFP through digital transformation.

The study also emphasizes that green technological innovation is a critical pathway through which digital intelligence boosts FETFP, generating multiplier effects that improve both ecological value and systemic efficiency. To fully harness this synergy, coordinated efforts across sectors, such as funding, policy frameworks, and institutional mechanisms, are crucial. First, dedicated green innovation funds should be established to increase R&D investment in key areas like eco-friendly processes, low-carbon equipment, and smart forestry technologies. Collaborative research among universities, research institutes, and forestry enterprises should be promoted to foster a multi-actor innovation ecosystem. Additionally, the commercialization of green technologies should be accelerated through incubation platforms, innovation vouchers, and refined evaluation and incentive systems. Strengthening the alignment between digital intelligence development and green innovation should also be a priority. Targeted support for high-priority areas, such as digital monitoring systems and innovative forestry equipment, will enhance the scalability, stability, and sustainability of green technologies, supporting the transition to a low-carbon, smart, and efficient forestry system.

To optimize ecological governance, regional strategies must be tailored to local contexts. The impact of digital intelligence varies across regions, depending on environmental expenditure and the development of green finance systems. In regions with high environmental expenditure, fiscal resources should be better coordinated to support ecological governance and digital infrastructure integration. In resource-constrained regions, central government transfers and ecological compensation mechanisms should be expanded to strengthen local fiscal capacity, with a focus on developing essential ecological infrastructure and piloting digital forestry initiatives. In provinces with advanced green finance systems, efforts should focus on integrating green finance and digital intelligence through mechanisms like carbon finance, green credit, and green bonds. For regions with less-developed green financial ecosystems, the focus should be on building foundational green finance services and promoting green credit innovation. A differentiated, regionally adaptive policy framework will ensure that digital transformation in China’s forestry sector is tailored to varying ecological and institutional contexts, ultimately improving sustainability, efficiency, and resilience across regions.