Promotion or Hindrance? Assessing Urbanization’s Impact on Forest Ecological Security Through the Lenses of Population, Economy, and Space: Evidence from China

Abstract

Forests serve as “water reservoirs, bank vaults, grain depots, and carbon pools,” and their ecological security plays a critical role in national ecological security. Urbanization, as a long-term development strategy in China, exerts complex and profound impacts on the safety and stability of forest ecosystems. In the context of simultaneously pursuing urbanization and forest ecological security (FES), a systematic analysis of the impact patterns of urbanization on FES from the perspectives of population urbanization (PUB), economic urbanization (EUB), and spatial urbanization (SUB) can effectively uncover the “black box” underlying their complex interrelationship. This study develops a comprehensive FES evaluation system, using panel data from 31 provinces in mainland China over the period from 2004 to 2022. The research employs two-way fixed effects models to examine the actual impacts and heterogeneous characteristics of multidimensional urbanization on FES, while also applying the Environmental Kuznets Curve (EKC) test model to verify potential nonlinear relationships. The main findings are as follows: (1) baseline regression results indicate that during the study period, SUB exerted the strongest negative impact on FES, followed by PUB, while EUB significantly promoted FES improvement. (2) Heterogeneity analysis reveals that the impact of PUB on FES demonstrates both regional and temporal variations, EUB’s effect shows significant temporal differences, and SUB exhibits distinct regional heterogeneity. (3) EKC tests confirm an inverted U-shaped relationship between PUB and FES with an inflection point at 0.481, and a U-shaped relationship between EUB and FES with an inflection point at 0.866. No significant nonlinear relationship is found between SUB and FES. These findings enhance the systematic understanding of how urbanization influences FES in China while offering valuable references for other rapidly urbanizing nations to better coordinate urban development and forest conservation.

1. Introduction

Urbanization, driven by advances in civilization and productivity, is essential for economic growth, industrial upgrading, and modernization [1]. It reflects a nation’s development level and supports urban–rural integration and improved living standards [2]. Unlike developed Western countries that took nearly two centuries to mature their urban systems, China has rapidly urbanized [3], increasing its urban population from 18% in 1978 to 66% by 2024 (https://data.worldbank.org/indicator/SP.URB.TOTL.IN.ZS?locations=CN, accessed on 8 November 2025). This transformation has moved hundreds of millions from rural to urban areas, making urbanization a key driver of economic and social progress with significant benefits [4]. However, China’s rapid, urbanization-driven growth has relied on excessive resource use, causing severe environmental degradation [5,6,7]. Urban expansion pressures—reducing arable land, worsening water scarcity, and increasing air, soil, and groundwater pollution—threaten sustainability and ecological well-being, while declining ecosystem services and biodiversity escalate [8,9,10]. Such trends directly contradict the core objectives of the United Nations Sustainable Development Goals (SDGs) regarding environmental protection, ecological balance, and social equity [11]. Simultaneously, these challenges have spurred extensive research and discussion on ecological security issues among scholars both domestically and internationally [12,13,14,15].

Ecological security refers to the healthy and stable state of natural ecosystems—including wetlands, forests, land, grasslands, and oceans—and includes key aspects such as biodiversity conservation, soil and water conservation, and climate regulation [16,17,18,19,20]. As the dominant component of terrestrial ecosystems, forests provide habitats for numerous plant and animal species and are a key factor in regional and global ecological security [21,22,23]. Described as “water reservoirs, bank vaults, grain depots, and carbon pools,” forests embody a concept of “Four Repositories” that vividly captures their multifaceted value [24]. Facing global challenges like worsening climate change and rapid biodiversity loss [25], strengthening forest ecological development and ensuring forest ecological security have become urgent priorities for humanity, forming a fundamental basis for sustainable socioeconomic development [26].

Faced with the dual challenge of advancing regional urbanization while ensuring the long-term sustainability of forest ecosystems, China appears to have achieved a notably encouraging outcome: the country has maintained over four decades of continuous “dual growth” in both forest area and timber volume despite rapid urbanization (https://www.forestry.gov.cn/c/www/zyxx/589827.jhtml, accessed on 8 November 2025). To date, China’s forest coverage has exceeded 25%, and its timber volume has surpassed 20 billion cubic meters (https://www.forestry.gov.cn/lyj/1/lcdt/20250312/614367.html, accessed on 8 November 2025). However, these quantitative gains in land greening do not fully reflect the overall state of high-level forest ecosystem conservation. Assessing the relationship between urbanization and forest ecosystem development based solely on such metrics risks yielding conclusions that are incomplete and potentially misleading.

The concept of forest ecological security (FES) integrates the natural environment and human activities within a unified framework. It describes an optimal state in which forest ecosystems maintain their structural and functional integrity and effectively support human production and livelihood within a specific spatiotemporal context [27]. This concept provides a viable approach for holistically assessing the harmonious coexistence between humans and forest ecosystems and accurately evaluating the actual effects of urbanization on the forest ecological environment. Academic research on FES focuses on assessment, monitoring, and influencing factors, forming a consensus: (1) FES is a human–nature interaction outcome requiring holistic evaluation of ecosystem pressures, states, and human responses [28,29,30]; (2) prospective monitoring is vital to identify antecedent factors for sustainability [31,32]; (3) human activities (especially urbanization and deforestation) critically impact FES, evidenced by cases like Amazon deforestation and Pakistan’s forest loss due to urban expansion and farming [33,34,35,36].

In practice, China’s rapid urbanization has been a double-edged sword for FES [37]. On one hand, driven by a government-led model, urbanization has advanced rapidly, accompanied by intensive human activities—such as infrastructure construction, real estate development, and industrial expansion—and unsustainable land use practices like cropland over-expansion, lake filling, and forest encroachment. These have led to widespread ecosystem degradation and declining biodiversity. As a result, forest water and soil conservation capacities have weakened, ecological corridors have fragmented, and landscape connectivity has deteriorated, posing significant threats to FES [38]. On the other hand, urban greening initiatives have preserved mature, high-canopy forests around cities, which provide vital services such as climate regulation, air purification, and noise reduction. Moreover, large-scale rural-to-urban migration has accelerated farmland abandonment and supported reforestation programs like “Grain for Green,” reducing agricultural pressure on forests and enabling natural vegetation recovery. This trend is enhancing ecosystem resilience and improving FES in many areas [39].

While previous studies have provided valuable insights into the relationship between urbanization and FES, the precise nature of their interaction remains debated. Much of the literature has focused on assessing coordinated development between the two [40,41,42], but rigorous empirical research on the causal links between urbanization processes and FES is still scarce. Addressing this gap requires deeper quantitative analysis and comprehensive statistical validation to produce more robust conclusions. A key limitation has been the overreliance on a single indicator—the “urbanization rate”—to represent the complex impacts of urbanization on FES. However, urbanization is a multifaceted phenomenon that encompasses far more than demographic shifts, such as the large-scale migration from rural to urban areas. It also involves economic transformation, characterized by the evolution from lower- to higher-productivity industrial structures, and spatial restructuring, manifested in the conversion of agricultural and forest land to urban built-up land. Addressing the identified research gaps, this study establishes a comprehensive analytical framework to advance the global understanding of the urbanization-FES nexus. The investigation proceeds through three systematic phases: first, urbanization is deconstructed into three dimensions: population (PUB), economic (EUB), and spatial (SUB). Second, two-way fixed effects models are employed to quantify both the distinct impacts and heterogeneous effects of each dimension on FES. Third, the Environmental Kuznets Curve (EKC) framework is incorporated to examine potential nonlinearities and identify ecological thresholds. This integrated approach not only clarifies the pathways through which urbanization can be advanced in a manner that enhances FES sustainability but also offers actionable insights for rapidly urbanizing developing countries to better align urban development with forest conservation, thereby fostering long-term human–nature harmony.

2. Materials and Methods

2.1. Variable Selection and Measurement

2.1.1. Dependent Variable

(1)

Variable selection

The accurate measurement of the dependent variable, forest ecological security (FES), is fundamental to empirically testing the impact of urbanization. To comprehensively assess the level of FES, an evaluation index system was constructed based on the pressure–state–response (PSR) model [43,44]. Following established methodologies [45,46], indicators were selected from the three dimensions of pressure (P), state (S), and response (R) to form a comprehensive framework for analyzing the dynamics of FES. The specific types of indicators and their calculation methods are summarized in

Table 1. The indicator system for forest ecological security.

Primary Index	Secondary Index (Unit)	Index Calculation Method	Attributes
Pressure (P)	Per capita wastewater discharge (tons/person)	Total regional wastewater discharge/total regional population at the end of the year	−
	Per capita sulfur dioxide emissions (tons/person)	Regional sulfur dioxide emissions/regional total population at the end of the year	−
	Human Disturbance Index (%)	(Construction land area + cultivated land area)/jurisdiction area × 100%	−
State (S)	Forest coverage rate (%)	Obtain it directly	+
	Forest volume per unit area (cubic meters/hectare)	Regional forest stock volume/regional forest area	+
	Proportion of regional forest land area (%)	Forest land area/jurisdiction area × 100%	+
	Proportion of natural forests (%)	Regional natural forest area/regional forest area × 100%	+
Response (R)	Proportion of newly added soil erosion control (%)	Newly added soil erosion control area/jurisdiction area × 100%	+
	Intensity of government investment in forestry	Regional government forestry investment completion amount/GDP	+
	Green coverage rate of built-up area (%)	Obtain it directly	+

.

(2)

Variable measurement

To mitigate subjective bias and ensure objective assessment, the entropy weight method was employed to evaluate the comprehensive development level of FES [47]. The specific measurement approach is as follows:

• Standardize the positive and negative indicator data:

$$\mathrm{Positive} \mathrm{indicators}: x_{ij}^{\prime }={\displaystyle \frac{x_{ij}-min\left(x_j\right)}{max\left(x_j\right)-min\left(x_j\right)}}$$

(1)

$$\mathrm{Negative} \mathrm{indicators}: x_{ij}^{\prime }={\displaystyle \frac{max\left(x_j\right)-x_{ij}}{max\left(x_j\right)-min\left(x_j\right)}}$$

(2)

2.

Perform normalization processing on the standardized indicator data:

$$P_{ij}={\displaystyle \frac{{x^{\prime }}_{ij}}{{\sum }_{i=1}^m{x^{\prime }}_{ij}}}$$

(3)

where m represents the number of research subjects.

3.

Calculate the information entropy of the jth indicator:

$$e_j=-k\cdot {\sum }_{i=1}^m{P}_{ij}\cdot ln{P}_{ij}$$

(4)

where $k=1/lnm$.

4.

Calculate the entropy weight of the jth indicator:

$$W_j=(1-e_j)/{\sum }_{j=1}^n(1-e_j)$$

(5)

where n represents the number of indicators.

5.

Calculate the comprehensive index of forest ecological security:

$$FES={\sum }_{j=1}^n{w}_j{x^{\prime }}_{ij}$$

(6)

Based on the above indicator system and evaluation method, a comprehensive assessment of the FES level of the 31 provinces in the Chinese mainland was conducted. The results of the FES level of each province in China in 2022 are shown in Figure 1. As shown in Figure 1, in 2022, the FES levels in most Chinese provinces were concentrated within the range of 0.326 to 0.440. Among these regions, Heilongjiang and Jilin in the northeast, Yunnan, Tibet, Jiangxi, and Guangxi in the central and western areas, and Fujian in the east exhibit relatively high FES levels. Most of these provinces are situated in parts of China characterized by abundant forest resources. However, in the eastern region, FES levels in provinces such as Shanghai, Tianjin, Jiangsu, and Shandong are generally low, indicating that the system is under considerable pressure.

2.1.2. Core Independent Variable

To effectively unpack the complex mechanisms through which urbanization affects FES, it is crucial to move beyond a unidimensional perspective. Accordingly, this study employs a multidimensional framework, using population urbanization (PUB), economic urbanization (EUB), and spatial urbanization (SUB) as core independent variables for a more comprehensive assessment. Specifically, PUB captures the process of rural-to-urban migration and the consequent rise in urban population share, reflecting demographic shifts central to urbanization. EUB reflects the structural transformation of the economy from agriculture to industry and services, indicating a fundamental reorganization of economic activities. SUB denotes the physical expansion of urban land, primarily driven by the conversion of rural areas into built-up zones, thereby altering land use patterns and spatial configurations.

Drawing on previous research [48], this study measures the level of PUB by the regional proportion of the urban population, capturing the growth in urban populace driven by rural-to-urban migration. The level of EUB is assessed by the share of regional output from the secondary and tertiary industries, reflecting the degree of economic structural transformation from primary to non-agricultural sectors. The level of SUB is quantified by the proportion of built-up area of the total regional area, indicating the scale of urban land expansion. The specific calculation methods for each indicator are detailed in

Table 2. Core independent variable indicators and their calculation methods.

Primary Indicators	Secondary Indicators (Unit)	Calculation Method of Indicators
Population urbanization (PUB)	The regional proportion of the urban population (%)	(Number of urban population in the region/total permanent population in the region) × 100%
Economic urbanization (EUB)	The share of regional output from the secondary and tertiary industries (%)	(Regional secondary and tertiary industry added value/regional gross domestic product) × 100%
Spatial urbanization (SUB)	The proportion of built-up area of the total regional area (%)	(Area of built-up area in the region/total area of the jurisdiction) × 100%

.

2.1.3. Control Variables

To isolate the net effects of the multidimensional urbanization process on FES, this study systematically controls for a range of confounding factors, thus ensuring the robustness of our estimates by mitigating omitted variable bias. The selection of these control variables is informed by prior literature [49] and includes the following:

(1)

Regional Economic Development (PGDP), measured by per capita GDP (deflated to 2004 constant prices using the Consumer Price Index) to account for inflation.

(2)

Regional Innovation Level (INNO), proxied by the natural logarithm of the number of domestic invention patent applications accepted, reflecting the region’s capacity for technological innovation.

(3)

Environmental Regulation Intensity (ENVI), quantified as the ratio of investment in industrial pollution control to industrial value added, representing the intensity of local governmental efforts in environmental governance.

(4)

Government Intervention (GOVE), represented by per capita government fiscal expenditure, indicating the degree of government involvement in economic and social affairs.

2.2. Empirical Models and Methods

2.2.1. Baseline Regression Model

To address potential confounding bias, the baseline analysis relies on a two-way fixed effects specification [50]. This approach provides a stringent counterfactual framework by netting out all time-invariant and period-specific unobserved factors, forming a credible baseline against which subsequent robustness checks are compared. The specific model formulation is as follows:

$$Y_{it}={\alpha }_0+{\beta }_1{X}_{it}+\sum {\beta }_x{Z}_{it}+{\mu }_t+{\eta }_i+{\epsilon }_{it}$$

(7)

In this model, the dependent variable is Y_it, representing the forest ecological security level of province i in year t; the core independent variable is X_it, indicating the urbanization development level of province i in year t (including population urbanization development level, economic urbanization development level, and spatial urbanization development level); Z_it represents a series of control variables; μ_t denotes the time fixed effect; η_i represents the individual fixed effect; ε_it is the random disturbance term; and α₀ represents the constant term. The statistical significance of the regression coefficients was assessed using two-tailed t-tests based on robust standard errors, with significance levels at 10%, 5%, and 1% denoted by *, **, and ***, respectively.

2.2.2. Robustness Test Methods

To ensure the reliability of the baseline regression results, this study employs four robustness checks to further validate the empirical findings. These tests aim to address potential estimation biases, measurement errors, and model specification uncertainties, thereby enhancing the robustness and credibility of the research conclusions. The specific methods include the following:

(1)

Replacing the explained variable. To reduce potential measurement error bias, this study reassesses regional FES level using the entropy weight-TOPSIS method [51,52] and re-runs the regression with the new index as the dependent variable.

(2)

Data winsorization at the 1% level. To mitigate the influence of extreme outliers on regression estimates, all continuous variables were subjected to 1% winsorization at both tails. This procedure replaces extreme values in the top and bottom 1% of each variable’s distribution with the nearest non-extreme values, thereby reducing distortion while preserving sample size.

(3)

Excluding samples from municipalities directly under the Central Government. Beijing, Tianjin, Shanghai, and Chongqing differ from other regions in administrative status and economic development, potentially introducing structural bias. The regression is re-estimated without these cities to ensure model robustness.

(4)

Alternative model specification. To evaluate the robustness of the estimation results to model specification, the two-way fixed effects model employed in the baseline regression is replaced with an ordinary least squares (OLS) regression model.

2.2.3. EKC Test Model

The Environmental Kuznets Curve (EKC) is a key hypothesis in environmental economics, introduced by Grossman and Krueger (1991) to describe the relationship between economic growth and pollution [53]. It posits that pollution rises with early-stage development but declines after income reaches a threshold, resulting in an inverted U-shaped pattern. In examining the dynamic interaction between urbanization and the environment, both domestic and international scholars frequently adopt the EKC as a theoretical framework to analyze environmental outcomes such as resource consumption and pollutant emissions during the urbanization process [54,55,56].

To further examine whether more complex nonlinear relationships exist between multidimensional urbanization (PUB, EUB, SUB) and FES, this study innovatively introduces the EKC framework into the analysis. Building on the classical EKC model [57], the EKC test model specification maintains the same variables, data definitions, and measurement methods as the baseline regression, ensuring consistency and comparability of results. The model is specified as follows:

$$Y_{it}={\alpha }_0+{\beta }_1{X}_{it}+{\beta }_2{X}_{it}^2+\sum {\beta }_x{Z}_{it}+{\mu }_t+{\eta }_i+{\epsilon }_{it}$$

(8)

2.3. Data Sources and Processing

To enhance the reliability and generalizability of the research findings, this study employs balanced panel data encompassing all 31 mainland Chinese provinces over the 19-year period from 2004 to 2022 for empirical investigation. The datasets are obtained from the China Statistical Yearbook, China Environmental Statistics Yearbook, China Land and Resources Statistical Yearbook, China Forestry and Grassland Statistical Yearbook (https://data.cnki.net/yearBook, accessed on 8 November 2025); the dataset incorporates information from the National Forest Resource Inventory (http://lygc.lknet.ac.cn/sd/si/zgslzy.html, accessed on 8 November 2025), the National Bureau of Statistics of China (https://data.stats.gov.cn, accessed on 8 November 2025), and the EPS database (https://www.epsnet.com.cn/index.html, accessed on 8 November 2025).

For the limited instances of missing values observed in specific years, this study adopts linear interpolation as a scientifically rigorous approach to systematically address data gaps. This method effectively preserves the temporal continuity of the dataset. Regarding all indicator variables pertaining to price fluctuations, Consumer Price Index (CPI) deflation is implemented with 2004 serving as the base year, thereby eliminating inflationary effects and ensuring intertemporal comparability. The descriptive statistics for all variables are presented in

Table 3. Descriptive statistics of each variable.

Variable	Observations	Mean	SD	Min	Median	Max
FES	589	0.343	0.119	0.365	0.057	0.744
PUB	589	0.560	0.147	0.555	0.209	0.896
EUB	589	0.891	0.058	0.895	0.661	0.998
SUB	589	0.019	0.033	0.008	0.000	0.197
PGDP	589	9.240	0.493	9.131	8.012	10.806
INNO	589	8.776	1.849	8.906	2.485	12.399
ENVI	589	0.004	0.004	0.003	0.000	0.031
GOVE	589	1.088	0.883	0.963	0.091	7.124

, which serves as a foundational basis for the subsequent econometric analysis.

3. Results and Discussion

3.1. Baseline Regression Results

To ensure accurate baseline model selection, the Hausman test was conducted before empirical analysis. The results rejected the random effects model, supporting the use of the two-way fixed effects model for baseline regression.

Table 4. Baseline regression results, respectively, for the effect of PUB, EUB, and SUB on FES.

Variable	FES
	(1)	(2)	(3)	(4)	(5)	(6)
PUB	−0.121 ***	−0.151 ***
	(−2.59)	(−2.61)
EUB			0.466 ***	0.222 ***
			(9.30)	(3.49)
SUB					−1.833 ***	−1.795 ***
					(−14.06)	(−12.60)
Constant	0.411 ***	0.247 **	−0.072	−0.761 ***	0.378 ***	0.473 ***
	(15.73)	(2.14)	(−1.62)	(−7.29)	(76.73)	(37.52)
Controls	NO	YES	NO	YES	NO	YES
Individual fixed	YES	YES	YES	YES	YES	YES
Time fixed	YES	YES	YES	YES	YES	YES
R-squared	0.952	0.954	0.941	0.949	0.848	0.980
Observations	589	589	589	589	589	589

Note: *** and ** respectively indicate the significance at 1% and 5% levels, t statistics in parentheses.

shows the baseline regression results of PUB, EUB, and SUB on FES. Columns (1), (3), and (5) report results without control variables; columns (2), (4), and (6) include all controls. All variables are significant at the 1% level, and model fit improves with controls, indicating greater explanatory power. The results show that PUB and SUB had significant negative effects on FES in China, while EUB had a significant positive effect.

3.2. Robustness Test Results

To strengthen the credibility of the baseline results, this study applies the four robustness tests described in Section 2.2.3. The results are reported in

Table 5. Robustness test results for the effect of PUB on FES.

Variable	FES
	(1)	(2)	(3)	(4)
PUB	−0.644 ***	−0.189 ***	−0.188 *	−0.386 *
	(−3.63)	(−4.26)	(−1.94)	(−2.04)
Constant	0.490	0.161 *	0.092	0.503 ***
	(1.31)	(1.66)	(0.45)	(6.78)
Controls	YES	YES	YES	YES
Individual fixed	YES	YES	YES	NO
Time fixed	YES	YES	YES	NO
R-squared	0.940	0.968	0.821	0.363
Observations	589	589	513	589

Note: *** and * respectively indicate the significance at 1% and 10% levels, t statistics in parentheses.

,

Table 6. Robustness test results for the effect of EUB on FES.

Variable	FES
	(1)	(2)	(3)	(4)
EUB	0.227 ***	0.223 ***	0.243 ***	0.186 ***
	(3.54)	(4.11)	(3.50)	(2.87)
Constant	−0.758 ***	−0.712 ***	−0.623 ***	−0.774 ***
	(−7.26)	(−8.04)	(−4.84)	(−6.69)
Controls	YES	YES	YES	YES
Individual fixed	YES	YES	YES	NO
Time fixed	YES	YES	YES	NO
R-squared	0.949	0.964	0.935	0.349
Observations	589	589	513	589

Note: *** indicate the significance at 1% levels, t statistics in parentheses.

and

Table 7. Robustness test results for the effect of SUB on FES.

Variable	FES
	(1)	(2)	(3)	(4)
SUB	−1.270 ***	−1.899 ***	−3.459 ***	−1.825 ***
	(−10.50)	(−12.60)	(−4.11)	(−8.18)
Constant	0.435 ***	0.478 ***	0.475 ***	0.461 ***
	(40.62)	(36.68)	(35.51)	(11.65)
Controls	YES	YES	YES	YES
Individual fixed	YES	YES	YES	NO
Time fixed	YES	YES	YES	NO
R-squared	0.966	0.891	0.875	0.387
Observations	589	589	513	589

Note: *** indicate the significance at 1% levels, t statistics in parentheses.

. Columns (1) to (4) show regressions from tests that replace the dependent variable, apply 1% winsorization, exclude central municipalities, and use an alternative measurement model. The coefficient signs and significance levels remain highly consistent with the baseline regression results, confirming their robustness.

3.3. Endogeneity Test Results

During the research process, endogeneity may arise from sample self-selection, omitted variables, and reverse causality. Following Lian and Jiang [58,59], first-order lagged values of each core independent variable were used as instrumental variables in two-stage least squares (2SLS) estimation. This approach addresses endogeneity and further tests the robustness of the regression results. The three columns in

Table 8. Results of the endogeneity test.

Variable	FES
	(1)	(2)	(3)
PUB	−0.193 ***
	(−3.45)
EUB		0.936 ***
		(4.52)
SUB			−1.501 ***
			(−7.54)
Controls	YES	YES	YES
Individual fixed	YES	YES	YES
Time fixed	YES	YES	YES
Anderson-LM statistic	529.408 ***	72.339 ***	282.149 ***
	[0.000]	[0.000]	[0.000]
Cragg-Donald Wald F statistic	936.928 ***	77.752 ***	547.216 ***
Stock–Yogo 10% critical value	16.380	16.380	16.380
R-squared	0.957	0.943	0.936
Observations	558	558	558

Note: *** indicate the significance at 1% levels, t statistics in parentheses. The values in square brackets represent the p-values for the Anderson-LM statistic.

report the second-stage 2SLS results for PUB, EUB, and SUB, respectively.

The results in Table 8 show that all Anderson-LM statistics have p-values below 0.1, indicating no under-identification of the instrumental variables. Each Cragg-Donald Wald F statistic exceeds the 10% critical value of 16.38 from the Stock–Yogo test, confirming that the instruments are not weak. With instrumental variables, the regression results remain largely consistent with the benchmark model, demonstrating that the baseline findings are robust after accounting for endogeneity.

3.4. Heterogeneity Analysis Results

China’s vast territory leads to significant provincial differences in geography, urbanization, economic development, industrial transformation, and land use. Urbanization has also evolved through distinct phases: before 2012, it focused on quantitative expansion measured by the “urbanization rate,” while after 2012 it emphasizes qualitative development, balancing urban growth with ecological protection [60]. To better understand urbanization’s impact on FES, this study analyzes heterogeneity across regions and time periods. Accordingly, the study area is first divided into eastern and central-western regions following China’s official regional classification for spatial heterogeneity analysis. Meanwhile, with 2012 taken as the temporal threshold—when the “new-type urbanization” concept was formally introduced at the 18th National Congress of the Communist Party of China—the study period is partitioned into traditional and new-type urbanization phases for temporal heterogeneity analysis.

3.4.1. Regional Heterogeneity Analysis

Table 9. Results of the regional heterogeneity analysis.

Variable	FES
	(1)	(2)	(3)	(4)	(5)	(6)
	Eastern	Central-Western	Eastern	Central-Western	Eastern	Central-Western
PUB	−0.185 ***	0.189 ***
	(−2.85)	(2.13)
EUB			0.330 ***	0.206 ***
			(2.43)	(2.77)
SUB					−0.530 ***	2.483 ***
					(−2.03)	(2.06)
Constant	0.414 ***	0.279 ***	−1.092 ***	−0.660 ***	0.299 ***	0.319 ***
	(9.46)	(6.33)	(−6.82)	(−4.23)	(10.85)	(23.61)
Controls	YES	YES	YES	YES	YES	YES
Individual fixed	YES	YES	YES	YES	YES	YES
Time fixed	YES	YES	YES	YES	YES	YES
R-squared	0.968	0.933	0.969	0.925	0.974	0.942
Observations	209	380	209	380	209	380
Intergroup difference test	−0.374 ***		0.125		−3.013 ***

Note: *** indicate the significance at 1% levels, t statistics in parentheses.

shows the grouped regression results for PUB, EUB, and SUB on FES across regions. Columns (1) and (2) reveal significant regional differences in PUB’s impact: it has a significantly negative effect in eastern China but a significantly positive one in central-western China. This divergence reflects differing urbanization stages. In eastern China, high PUB intensifies population concentration, driving urban sprawl and increasing pressure on forest ecosystems, thus reducing FES. In contrast, moderate PUB in central-western China enables scale benefits, improving resource efficiency and easing rural pressure on forests, thereby enhancing forest quantity and quality.

Regarding EUB, columns (3) and (4) indicate no significant regional differences in its impact on FES. Consistent with the baseline regression results, EUB demonstrates significantly positive coefficients at the 1% level in both eastern and central-western China, suggesting that economic urbanization contributes to FES improvement across all regions. For SUB, columns (5) and (6) show a pattern similar to that of PUB, with significant regional variations in its effect on FES. SUB exhibits a significantly negative impact in eastern China but a significantly positive impact in central-western China. This contrast suggests that the threat to FES from SUB in eastern China stems primarily from irrational land use and disorderly expansion of built-up areas. When implemented through scientifically planned and well-paced development strategies, SUB can actually promote ecological improvement and enhance FES through optimized land resource integration and efficient utilization.

3.4.2. Temporal Heterogeneity Analysis

Table 10. Results of the temporal heterogeneity analysis.

Variable	FES
	(1)	(2)	(3)	(4)	(5)	(6)
	Traditional	New-Type	Traditional	New-Type	Traditional	New-Type
PUB	0.276 ***	−0.287 ***
	(2.28)	(−2.75)
EUB			0.245 ***	0.571 ***
			(2.84)	(3.13)
SUB					−1.783 ***	−1.719 ***
					(−6.57)	(−10.38)
Constant	2.096 ***	0.894 ***	0.114 ***	−0.161 ***	0.489 ***	0.478 ***
	(7.78)	(4.95)	(1.51)	(−0.98)	(23.50)	(28.19)
Controls	YES	YES	YES	YES	YES	YES
Individual fixed	YES	YES	YES	YES	YES	YES
Time fixed	YES	YES	YES	YES	YES	YES
R-squared	0.943	0.912	0.959	0.955	0.938	0.944
Observations	279	310	279	310	279	310
Intergroup difference test	0.563 ***		−0.326 ***		−0.064

Note: *** indicate the significance at 1% levels, t statistics in parentheses.

presents the grouped regression results examining the effects of PUB, EUB, and SUB on FES across different time periods. Columns (1) and (2) reveal significant temporal heterogeneity in PUB’s impact on FES. During the traditional urbanization phase, PUB showed a significantly positive effect on FES at the 1% level, while it turned significantly negative at the 1% level in the new-type urbanization phase. This indicates that urban population growth initially contributed to FES improvement but became detrimental in the later phase. Although baseline results suggest PUB generally pressures FES, this heterogeneity implies its impact is not consistently negative but may involve more complex nonlinear relationships, which will be further verified in the following subsection.

Columns (3) and (4) show that EUB significantly promoted FES at the 1% level in both phases, though with notable nonlinear differences in effect magnitude. The positive impact was substantially stronger during new-type urbanization, reflecting China’s economic transition from resource-intensive growth to green, innovation-driven development. This shift has enabled more balanced urban–rural green transformation, providing greater impetus for ecological improvement. For SUB, columns (5) and (6) demonstrate consistently significant negative effects at the 1% level across both periods, with no substantial difference in impact. This suggests SUB’s adverse effect on FES remained largely unchanged in direction throughout the study period, showing no complex nonlinear characteristics.

3.5. EKC Test Results

The temporal heterogeneity analysis shows that the effects of PUB, EUB, and SUB on FES are not constant but vary across urbanization stages. This suggests that beyond linear relationships, nonlinear mechanisms should be examined to fully understand their interactions. Given the widespread use of the Environmental Kuznets Curve (EKC) framework in studying nonlinear urbanization–ecological quality relationships, this study adopts it by estimating quadratic models to test for nonlinearities between each urbanization dimension and FES.

Table 11. Results of the EKC test model.

Variable	FES
	(1)	(2)	(3)
PUB	0.434 *
	(1.93)
PUB 2	−0.451 **
	(−2.39)
EUB		−2.804 ***
		(−3.20)
EUB 2		1.618 ***
		(3.04)
SUB			−0.163
			(−0.35)
SUB 2			2.029
			(1.16)
Controls	YES	YES	YES
Individual fixed	YES	YES	YES
Time fixed	YES	YES	YES
Curve type	Inverted U-shaped	U-shaped	Non-significant
R-squared	0.953	0.949	0.957
Observations	589	589	589

Note: ***, **, and * respectively indicate the significance at 1%, 5%, and 10%, t statistics in parentheses. Variables with a superscript ² represent the squared term of the original variable.

presents the EKC results, including quadratic term coefficients and significance levels. These findings offer key insights into the complex urbanization–FES relationship.

Column (1) of Table 11 shows that the linear term of PUB has a significantly positive coefficient, while its quadratic term is significantly negative, indicating a statistically significant inverted U-shaped relationship between PUB and FES. Figure 2 displays the EKC for PUB’s impact on FES, including the inflection point. Further calculation places the inflection point at 0.481, meaning that PUB enhances FES when the urban population proportion is below this threshold. Beyond it, further PUB growth increasingly pressures FES.

Column (2) of Table 11 shows that the linear term of EUB has a significantly negative coefficient, while its quadratic term is significantly positive, indicating a significant U-shaped relationship between EUB and FES. Figure 3 displays the EKC for EUB’s impact on FES, including the inflection point. The calculated inflection point is 0.866. When the proportion of secondary and tertiary industries in the regional economy remains below this point, the advancement of EUB exerts pressure on FES. However, once this threshold is exceeded, further development of EUB contributes to the improvement and enhancement of FES.

Column (3) of Table 11 shows that both the linear and quadratic terms of SUB are statistically insignificant, indicating no significant nonlinear relationship with FES during the study period. This finding aligns with baseline regression and temporal heterogeneity analyses, demonstrating consistent directionality of SUB’s impact across different periods without U-shaped or inverted U-shaped nonlinear characteristics.

3.6. Further Discussion on the Causes of the EKC Test Results

3.6.1. Discussion on the Causes of the EKC Test Results for the Impact of PUB on FES

The EKC test results reveal a significant inverted U-shaped relationship between PUB and FES over the research period. This research finding is largely consistent with the research findings of scholars such as Feng [61], yet the key contribution of this study lies in the precise identification and quantification of the specific inflection point in the nonlinear influence curve. The underlying mechanism behind this phenomenon can be reasonably explained by the phased characteristics of urbanization development: in the early phase of PUB, moderate and orderly rural-to-urban migration reduces reliance on traditional forest-dependent livelihoods—such as clearing forests for agriculture, excessive fuelwood harvesting, and degradative overgrazing. As a result, fewer people directly consume forest resources, significantly lowering anthropogenic pressure on forest ecosystems. This reduced disturbance not only limits direct resource depletion but also allows time for ecosystem recovery, promoting the restoration of FES. In contrast, during rapid urbanization, a large and sudden influx of rural migrants triggers widespread deforestation for construction, increased vehicular emissions, strained infrastructure, and rising demand for energy and timber. The combined impact of these pressures can quickly erode earlier ecological gains, disrupt ecosystem services, and undermine forest ecosystem stability.

Combining the baseline regression and heterogeneity analysis results, it is clear that during the study period, most regional urban population ratios exceeded the inflection point, leading to a significantly negative overall impact of PUB on FES. In the traditional urbanization phase, urban population ratios were generally lower, with most values below the inflection point, resulting in a significantly positive effect of PUB on FES. In contrast, during the new-type urbanization phase, most regions surpassed the EKC’s inflection point, causing PUB to have a significant negative impact on FES. The identified EKC pattern between PUB and FES aligns closely with the baseline and heterogeneity findings, confirming their validity while further revealing the nonlinear, phased nature of PUB’s impact mechanism. This provides key empirical evidence for understanding forest ecological conservation challenges and opportunities amid population urbanization.

3.6.2. Discussion on the Causes of the EKC Test Results for the Impact of EUB on FES

The EKC test results show a significant U-shaped relationship between EUB and FES over the study period, indicating a phased pattern in their interaction. Previous studies, such as Hou and Deng [62], have also found a U-shaped relationship between regional economic growth and forest quality from a spatial perspective. This convergent evidence from independent research strengthens the credibility and robustness of the findings presented in this study. The underlying mechanism behind this phenomenon can be explained by the different economic development models and goals at different stages of the urbanization process: during the initial stage, regional development undergoes rapid industrial restructuring as economies transition from agriculture and forestry toward manufacturing and construction. This phase maintains frequent direct forest resource exploitation to support industrial expansion and infrastructure development. Concurrently, heavy fossil fuel consumption generates sharply increased industrial pollutant emissions, creating substantial pressure on forest ecosystems. Once these pressures surpass ecological thresholds, forest degradation emerges, producing an overall negative effect of EUB on FES. As economic urbanization advances, industrial structure becomes progressively optimized. Traditional energy-intensive and polluting industries are gradually replaced by knowledge-intensive sectors, high-tech industries, producer services, and eco-tourism. These emerging industries demonstrate reduced direct resource dependence while creating institutional demands for environmental quality. This transformation motivates governments and enterprises to enhance investments in pollution control, ecological restoration, and forest management. The resulting improvements in regional environmental quality and forest ecosystem stability ultimately convert the influence of economic urbanization on FES from negative to positive during later development stages.

The baseline regression and heterogeneity analysis consistently show a significant positive impact of EUB on FES. This is because most regions’ secondary and tertiary industry output shares remained above the inflection point during the study period. The temporal analysis confirms this effect in both traditional and new-type urbanization phases, with stronger benefits in the latter due to intensified industrial optimization. The EKC test further validates these findings. Together, the results clarify the relationship between EUB and FES and provide practical guidance for policies that balance economic growth and environmental protection.

3.6.3. Discussion on the Causes of the EKC Test Results for the Impact of SUB on FES

The EKC test results show no significant nonlinear relationship between SUB and FES during the study period. This is primarily because rapid SUB continuously converts high-quality land—especially agricultural and forest land—into built-up areas, compressing and fragmenting natural ecosystems such as forests and wetlands. As artificial ecosystems expand in megacities and urban agglomerations, structural conflicts arise between natural and human-dominated systems. These manifest as disrupted microclimates, imbalanced local water–heat cycles, and impaired soil–vegetation–atmosphere material exchange. In regions like Beijing–Tianjin–Hebei and the Yangtze River Delta, built-up area expansion far exceeds ecological recovery capacity, causing immediate issues such as habitat loss [63], biodiversity decline [64], and reduced water conservation [65]. Over time, altered growing conditions and disturbed stand structure lead to persistent negative impacts on forest ecosystem health and stability, creating long-term environmental pressures that degrade FES.

4. Conclusions and Implications

4.1. Conclusions

This study represents the first effort to reveal the distinct effects and underlying mechanisms of China’s multidimensional urbanization on FES through three perspectives: PUB, EUB, and SUB. The main findings are summarized as follows: during the study period, both PUB and SUB exerted significantly negative effects on FES, with SUB posing a greater potential threat. In contrast, EUB enhanced FES by driving industrial transformation through green technological innovation and low-carbon sustainability. In terms of regional heterogeneity, EUB’s impact on FES showed no regional variation, positively affecting both eastern and central-western China. PUB and SUB exhibited regional heterogeneity: PUB negatively impacted eastern FES but positively affected central-western areas, while SUB threatened eastern FES yet enhanced it in central-western regions with proper management. Regarding temporal heterogeneity, SUB consistently pressured FES across both periods, showing no temporal heterogeneity. In contrast, PUB shifted from significantly boosting FES in the traditional urbanization phase to significantly harming it in the new-type phase. EUB had consistently positive effects, with stronger benefits during the new-type phase. With respect to nonlinear effects, both PUB and EUB demonstrate significant nonlinear relationships with FES. Specifically, PUB displays a significant inverted U-shaped relationship with FES, with an inflection point at 0.481. EUB exhibits a significant U-shaped relationship with FES, with an inflection point at 0.866. No significant nonlinear relationship is found between SUB and FES during the study period. Throughout urbanization, irrational land use remains a persistent threat to FES.

4.2. Implications

Based on the above findings, three targeted policy implications are proposed to support the coordinated governance of urbanization and FES globally: first, urbanization policies should prioritize green transformation by advancing industrial upgrading, expanding the circular economy, and accelerating green technology innovation, with a particular focus on strengthening the role of EUB in driving high-quality development. Second, the regional heterogeneity in the impacts of PUB and SUB necessitates tailored policy responses: eastern regions must address ecological pressures arising from excessive urban expansion, whereas central and western regions can strategically harness moderate agglomeration to promote ecological restoration. Finally, continuous monitoring of key indicators—such as urban population share and built-up area growth—is critical for identifying ecological thresholds, preventing irreversible environmental degradation, and ensuring that urbanization proceeds within sustainable boundaries.

In summary, this study demonstrates that urbanization and FES are not inherently antagonistic but can achieve synergistic development through scientific planning, green transformation, and differentiated governance. Countries worldwide should strengthen ecological red-line thinking throughout the urbanization process, enhance cross-sectoral policy integration, and promote a human-centered, low-carbon, and ecologically resilient model of urbanization. Such coordinated efforts will yield valuable insights and actionable pathways for achieving global sustainable development goals.

4.3. Research Limitations and Future Directions

The limitations and future research directions of this study are as follows: firstly, constrained by data availability, this study focuses on China as a representative case to examine how multidimensional urbanization affects FES. However, systematic international comparisons and cross-regional analyses have not yet been incorporated. Future research could collect multi-source data from countries at different levels of development to enable large-scale comparative studies of urbanization patterns and their universal as well as context-specific impacts on FES. Additionally, this study only examines the direct effects of multidimensional urbanization on FES at the provincial level in China. Nevertheless, according to Tobler’s first law of geography, environmental and socioeconomic phenomena often exhibit spatial dependence. Therefore, the relationship between urbanization and FES may involve spatial interactions. Future research could thus employ spatial econometric models to further investigate these potential multilevel spatial effects.