Research on the Driving Factors and Trade-Offs/Synergies of Woodland Ecosystem Services in Zhangjiajie City, China

Abstract

Research on woodland ecosystem services is the premise of the formulation of regional woodland policies and ecological protection measures in a new round of woodland protection utilization planning. Based on remote-sensing images and socioeconomic data, this study used the InVEST model, geographical detectors, Spearman correlation analysis, and a coupling coordination degree model to evaluate the spatiotemporal changes, driving factors, and trade-off/synergies relationship of habitat quality (HQ), soil conservation (SC), water conservation (WC), and carbon storage (CS) in the woodland of Zhangjiajie City in 1995, 2005, 2015, and 2022. The results show the following: (1) HQ significantly decreased, SC and WC fluctuated and decreased, and CS continued to increase. (2) Natural factors were predominant, and other factors and natural factors interact to increase the driving effect. (3) The four services were dominated by weak synergy, but SC and CS turned into weak trade-offs in 2022. These research results can provide theoretical support for the optimization of the tourism development model, the formulation of ecological compensation policies, and the high-quality sustainable development of woodland in Zhangjiajie City, and provide a case study of the ecological management of world natural heritage sites.

1. Introduction

The increasingly serious environmental pollution, resource crisis, and loss of biodiversity have promoted research on global ecosystem services [1,2,3]. Ecosystem services are bridges and ties between nature and humans, and they include four types: provisioning, regulating, supporting, and cultural services [4,5,6,7]. Due to rapid economic development [8,9,10], different services often show complex interaction relationships; that is, there are trade-offs/synergistic effects of mutual enhancement and restriction [11,12]. As the main body of terrestrial ecosystems, woodland ecosystems provide multiple ecosystem services, such as timber supply, soil and water conservation, and carbon storage, which play a crucial and irreplaceable role in global and human economic and social development [13,14,15]. Over the last several decades, due to high-intensity economic activities, extensive forest management, and a lack of understanding of ecosystem services, a series of ecological issues, such as biodiversity loss, soil erosion, and pollutant transmission, have occurred [16,17,18]. Since the concept, classification, and evaluation methods of ecosystem services were proposed by Daily [4] and Costanza et al. [5] in the 1990s, forest ecosystem services have also received more and more attention [19,20]. For example, Garmaeepour et al. [21] used the structural equation model and the index of social progress to analyze the complex relationship between the ecosystem services of the mangrove forests of Iran and the welfare of the local community, highlighting the importance of this nexus in achieving sustainability indicators. Kassun et al. [22] argued that conflicts between forest conservation and forest use will lead to trade-offs/synergies between services, exacerbated by land-use changes. However, in studies on value assessments of ecosystem services and their responses to land-use change [23,24] and landscape pattern change [25,26], woodland were only explored as a type of use, and the value of woodland ecosystem services, their driving factors, and the trade-offs/synergies among them still need in-depth research. Currently, research on the ecosystem services of woodland mostly focuses on value assessment, which includes in-depth analysis of single service types, such as habitat quality [27,28], soil conservation [29,30], water conservation [31,32], and carbon storage [33,34], as well as comprehensive discussions of service types [30,35,36], covering multi-scale studies from global [37,38], national [15,39], and regional perspectives [40,41,42]. The study of driving forces is mainly combined with the analysis of ecosystem services [43,44,45]. However, there is a lack of quantitative research on the driving factors or trade-offs/synergies of woodland ecosystem services in different study areas, especially world natural heritage sites, and the ability of woodland ecosystems to provide services is jointly influenced by the structural functions of woodland and external environmental factors, so the driving factors and trade-offs/synergies of woodland ecosystem services in world natural heritage sites need to be strengthened.

Zhangjiajie City is the location of a world natural heritage site and is the ecological barrier of the middle and lower reaches of the Yangtze River, with a prominent ecological status. However, Zhangjiajie City is facing problems due to its fragile ecological background and the low stability of the ecosystem. In addition, due to the impacts of the tourism industry, urbanization, and other factors, the woodland ecosystem is facing problems of insufficient carrying capacity, declining woodland quality, and weakening service capacity.

In this context, it is crucial to reveal the characteristics of the spatiotemporal dynamics of the woodland ecosystem services in Zhangjiajie City and to analyze the driving factors of the woodland ecosystem services in Zhangjiajie City and the trade-offs/synergies among them, as well as their spatial coupling and relationships, in order to deeply explore the impacts of ecological policies and tourism development on the woodland ecosystem services in Zhangjiajie City.

This study analyzes the spatiotemporal evolution of woodland in four periods: 1995, 2005, 2015, and 2022. The InVEST 3.14.2 (Integrated Valuation of Ecosystem Services and Tradeoffs, InVEST) model is used to evaluate the ecosystem services of the woodland in Zhangjiajie City, namely, habitat quality (HQ), soil conservation (SC), water conservation (WC), and carbon storage (CS). The driving factors of these ecosystem services and their trade-offs/synergies, as well as the spatial coupling and coordination relationships between them, are explored using a geographical detector model, correlation analysis, and a coupling coordination model. The purpose of this study is to clarify the spatiotemporal evolution process and influencing factors of woodland ecosystem services in Zhangjiajie City, so as to provide a theoretical basis and technical support for the ecological protection and high-quality sustainable development of woodland in Zhangjiajie City, a new round of woodland protection and utilization planning, and the special planning of woodland resources in the national territorial space.

2. Materials and Methods

2.1. Study Area

Zhangjiajie City is located in the northwest of Hunan Province, between 28°2′ and 29°48′ north and 109°40′ and 111°20′ east. The total area is 9533 km² (Figure 1).

The landforms in the area are diverse, with the most characteristic being the quartz sandstone peak woodland form. It belongs to the mid-sub-mountain-type humid monsoon climate. The area is crisscrossed by streams and rivers, with the Li River and the Lou River as the main water systems. Zhangjiajie City was established for tourism. The regional GDP in 2022 was CNY 59.25 billion. The number of domestic and foreign tourists received was 20.454 million person-times, and the total tourism income was CNY 27.5 billion, accounting for 46.45% of the city’s total GDP.

2.2. Data Sources

The data for this study included the following (

Table 1. Main data sources.

No.	Data Name	Format	Resolution	Data Source
1	Remote-Sensing Imagery (Landsat TM/ETM+/OLI)	Raster	30 m	Geographic Data Spatial Cloud
2	Elevation	Raster	30 m	Geographic Data Spatial Cloud
3	Soil	Raster	30 m	HWSD Soil Database
4	Leaf Area Index (LAI)	Raster	500 m	Institute of Geographic Sciences and Natural Resources Research, CAS
5	Temperature and Precipitation	Raster	1 km	National Earth System Science Data Center
6	Nighttime Light Data	Raster	500 m	National Earth System Science Data Center
7	GDP	Raster	1 km	Geographic Data Spatial Cloud
8	Population Density	Raster	1 km	Geographic Data Spatial Cloud
9	Socioeconomic Data (e.g., Total Tourism Revenue)	Table	-	Regional Yearbooks, Statistical Yearbooks

): remote-sensing images (Landsat TM/ETM/OLI) from 1995, 2005, 2015, and 2022 with a resolution of 30 × 30 m; secondary survey data, fixed plot data, and long-term field positioning monitoring data on forest resources in Zhangjiajie City. The former were used as basic data, and the latter were used as control and auxiliary data; by interpreting remote-sensing images, the woodland data for the four periods were obtained, and the second-class survey data and plot data were used to assist in the modification. In this study, the precision was evaluated using the confusion matrix in ENVI 5.3. Using a random sampling method, the test samples were selected evenly from Google images, secondary survey data, and fixed plot data, and the accuracy was evaluated. After precision evaluation, the overall classification accuracy in 1995, 2005, 2015, and 2022 was 82.37%, 85.63%, 85.78%, and 87.97%, respectively. The reference elements of remote interpretation data in 1995 were relatively few, and the accuracy was relatively low, but overall, it met the requirements of the study. The woodland data were divided into six categories: broadleaved forest, coniferous forest, mixed coniferous and broadleaved forest, bamboo-dominated woodland, shrub forest, and sparse woodland. In the follow-up research, all spatial data coordinates and resolutions were unified to WGS84 and 30 m × 30 m.

2.3. Research Methods

2.3.1. Methods for Evaluating Ecosystem Services

This article follows the principle of combining scientificity with practicality to select the indicators of ecosystem services, and it combines the regional ecological characteristics and human needs to ensure that the selected indicators can effectively reflect the core functions of forest ecosystems and their contributions to sustainable development. Firstly, as the location of a world natural heritage site, the protection of rare species such as giant salamanders and dove trees needs to prioritize the assessment HQ. And improving HQ is in line with the requirements of the Convention on Biological Diversity and the construction of China’s national park system. Secondly, according to the results of the third remote-sensing survey of soil erosion in Hunan Province, the area of soil erosion in Zhangjiajie City accounts for 24.02% of the total land area of the city, which is much higher than the average level of 17.63% in Hunan Province, and it is the most serious area of soil erosion in Hunan Province. The soil formed by the weathering of shale and sandstone in Zhangjiajie City has poor erosion resistance and is prone to soil erosion caused by rainfall and human activities (such as scenic area construction). SC evaluation can assess erosion risk by combining topography and vegetation data and ensure regional ecological security. Thirdly, Zhangjiajie City is the source of the Li River, and the WC capacity directly affects the water resources security in northwest Hun and the middle and lower reaches of the Yangtze River, and the assessment of WC can monitor the impact of deforestation and tourism facilities’ construction on the function of the water source. Fourthly, forest ecosystems are major carbon pools, and the assessment of CS in forest ecosystems in Zhangjiajie City is a response to the “double carbon” strategy, which can help in the realization of the value of ecological products in Hunan Province. Based on the above considerations, this article selects four types of ecosystem services, namely HQ, SC, WC, and CS, for the study of ecosystem services in Zhangjiajie City. These four services are well matched with the special topography, ecological sensitivity, and policy needs of Zhangjiajie City and can provide a scientific decision-making basis for regional sustainable development.

(1) Model of HQ

This study uses the biodiversity module in the InVEST 3.14.2 model to calculate the HQ of woodland in Zhangjiajie City. The calculation formula is as follows:

$$Q_{xj}=H_j(1-{\displaystyle \frac{D_{xj}^z}{D_{xj}^z+K^z}})$$

(1)

where Q_xj is the HQ index of grid cell x in land-use type j; D_xj is the level of disturbance in a grid cell of land-use type j; H_j is the suitability of land-use type j; K is the half-saturation constant, and it is usually taken as half of the maximum of D_xj; z is the normalization constant.

In the InVEST 3.14.2 model, the biodiversity module takes the sensitivity of threat factors and the strength of external threats into account to calculate and select the threat sources, set the weight and maximum impact distance, determine the habitat suitability and sensitivity, calculate the habitat degradation degree through the threat sources, and then integrate habitat suitability to obtain HQ.

The more sensitive a habitat is to threatening factors, the greater the degree of habitat degradation. In combination with the relevant research results and the situation of Zhangjiajie City, this study determines the threatening factors of the woodland HQ in Zhangjiajie City and their maximum influencing distance and weight (

Table 2. Threatening factors and their maximum impact distance and weight.

Threatening Factor	Maximum Influence Distance (km)	Weight	Attenuation Type
Paddy field	8	0.7	Exponential
Dryland	8	0.5	Linear
Urban land	10	1	Exponential
Rural residential land	5	0.7	Exponential
Other construction land	10	1	Exponential

), as well as the habitat suitability of different woodland types and their sensitivity to the threatening factors (

Table 3. Habitat suitability of different woodland types and their sensitivity to threatening factors in Zhangjiajie City.

Woodland Type		Habitat Suitability	Paddy Field	Dryland	Urban Land	Rural Residential Land	Other Construction Land
Tree-dominated woodland	Broadleaf forest	1	0.6	0.6	0.95	0.85	0.8
	Coniferous forest	1	0.6	0.6	0.95	0.85	0.8
	Mixed coniferous–broadleaf forest	1	0.6	0.6	0.95	0.85	0.8
Bamboo-dominated woodland		1	0.6	0.6	0.95	0.85	0.8
Shrubland		0.95	0.7	0.7	1	0.85	0.8
Sparse woodland		0.9	0.8	0.8	1	0.9	0.65

) [46,47]. In order to analyze the spatiotemporal evolution process of the woodland HQ in Zhangjiajie City, the woodland HQ in Zhangjiajie City is divided into five grades according to the natural breakpoint method, where [0, 0.35) is determined as the lowest, [0.35, 0.45) as lower, [0.45, 0.55) as medium, [0.55, 0.65) as higher, and [0.65, 1] as the highest.

(2) Model of CS

This study utilizes the CS module of the InVEST 3.14.2 model to calculate the CS in the woodland of Zhangjiajie City. The calculation formula is as follows:

$$C_{total}=C_{above}+C_{below}+C_{dead}+C_{soil}$$

(2)

Here, ${\mathrm{C}}_{\mathrm{t}\mathrm{o}\mathrm{t}\mathrm{a}\mathrm{l}}$ is the total carbon density (tons/hm²), ${\mathrm{C}}_{\mathrm{a}\mathrm{b}\mathrm{o}\mathrm{v}\mathrm{e}}$ is the aboveground carbon density (tons/hm²), ${\mathrm{C}}_{\mathrm{b}\mathrm{e}\mathrm{l}\mathrm{o}\mathrm{w}}$ is the belowground carbon density (tons/hm²), ${\mathrm{C}}_{\mathrm{d}\mathrm{e}\mathrm{a}\mathrm{d}}$ is the dead organic carbon density (tons/hm²), and ${\mathrm{C}}_{\mathrm{s}\mathrm{o}\mathrm{i}\mathrm{l}}$ is the soil carbon density (tons/hm²) (

Table 4. Carbon density of different woodland types in Zhangjiajie City. Unit: tons/hm².

Woodland Type		Cabove	Cbelow	Csoil	Cdead
Tree-dominated woodland	Broadleaf forest	63.92	132.1	254.8	7.8
	Coniferous forest	83.21	136.25	240.25	7.8
	Mixed coniferous–broadleaf forest	102.5	140.4	225.7	7.8
Bamboo-dominated woodland		20.35	67.5	170	7.8
Shrubland		26.6	67.5	150	6
Sparse woodland		22.35	67.5	120	5

) [33].

(3) Model of SC

The SC capacity of woodland in Zhangjiajie City is calculated using the SDR module in the InVEST 3.14.2 model. This module is based on the pixel-scale USLE method and integrates data such as land-use types, soil characteristics, a DEM, vegetation-cover factors, rainfall, and SC measures [48]. Using a raster as the computational unit, the model performs calculations. When there is no vegetation cover on the ground, the formula is as follows:

$$RKLS=R\times K\times LS$$

(3)

When there is vegetation cover on the ground, the calculation formula for soil erosion using the USLE is as follows:

$$USLE=R\times K\times LS\times P\times C$$

(4)

Here, LS represents the slope length and steepness factor (dimensionless), K is the soil erodibility factor (tons·hm²·h/(hm²·MJ·mm)), C is the vegetation cover and management factor, R is the rainfall erosivity factor, and P is the SC practice factor. Finally, the SC capacity (tons/hm²·a) is obtained by subtracting USLE from RKLS.

The rainfall erosivity factor (R) is calculated using established research methods, which primarily characterize the potential soil erosion that may be caused by rainfall. The calculation formula is as follows:

$$R_j={\alpha }_3\times P_j^{{\beta }_3}$$

(5)

In the formula, P_j represents the rainfall in the j-th year (mm), R_j represents the rainfall erosivity in the j-th year (MJ·mm·hm²·h⁻¹), and α₃ and β₃ are model parameters.

Generally, the steeper the slope and the longer the slope length, the more susceptible the soil is to erosion. During model calculations, the slope threshold set by the user is taken into account, and different formulas are applied to areas with low and high slopes. Ultimately, the LS for the entirety of Zhangjiajie City is obtained.

The K factor reflects differences in erosion rates due to variations in soil properties, including soil texture, organic matter content, soil structure, and permeability, which determine soil erodibility. In this study, the formula proposed by Williams et al. in the EPIC model is used to calculate the soil erodibility value K by utilizing soil particle composition data and soil organic carbon data. The resulting raster layer of the K value is generated using the following formula:

$$\begin{array}{l}K=0.1317\times \left\{0.2\times 0.3exp\left[-0.0265SAN\left(1-{\displaystyle \frac{SIL}{100}}\right)\right]\right\}\times {\left({\displaystyle \frac{SIL}{CLA+SIL}}\right)}^{0.3}\times \{1-\\ {\displaystyle \frac{0.25C}{C+exp\left(3.72-2.95C\right)}}\}\times \{1-{\displaystyle \frac{0.75N}{SN+exp\left(22.95SN-5.51\right)}}\}\end{array}$$

(6)

$$SN=1-{\displaystyle \frac{SAN}{100}}$$

(7)

In the formula, SAN represents the sand content of the soil (%), SIL represents the silt content of the soil (%), and CLA represents the clay content of the soil (%).

The vegetation-cover factor C ranges between 0 and 1 and characterizes the influence of land-use type, vegetation type, and vegetation cover on soil erosion. It primarily refers to the ratio of soil erosion under vegetation cover or field management to soil erosion under continuous fallow with clean tillage. Therefore, the land-use type and vegetation coverage are the main factors affecting the vegetation-cover factor C. In this study, the research results of Cai Chongfa et al. [49] are primarily adopted. The NDVI is used to calculate vegetation coverage, and then the vegetation-cover factor C is calculated based on vegetation coverage. The specific calculation formula is as follows:

$$c={\displaystyle \frac{NDVI-{NDVI}_{min}}{{NDVI}_{max}-{NDVI}_{min}}}$$

(8)

$$C=\left\{\begin{array}{cc}1& c=0\\ 0.6508-0.3436lnc& 0<c\le 78.3\%\\ 0& c>78.3\%\end{array}\right.$$

(9)

P refers to the ratio of soil erosion under specific conservation measures to soil erosion without any measures, typically ranging between 0 and 1, and it is a dimensionless constant. Based on the existing research results, C and P in this study are shown in

Table 5. Vegetation-cover factor C and soil and water conservation measures P of different woodland types in Zhangjiajie City.

Woodland Type		Vegetation-Cover Factor (C)	Soil and Water Conservation Measures Factor (P)
Tree-dominated woodland	Broadleaf forest	0.005	0.99
	Coniferous forest	0.005	0.99
	Mixed coniferous–broadleaf forest	0.005	0.99
Bamboo-dominated woodland		0.01	0.99
Shrubland		0.03	0.99
Sparse woodland		0.05	0.99

.

(4) Model of WC

The WC is derived from water yield data and adjusted by factors such as the topographic index, soil-saturated hydraulic conductivity, and flow velocity coefficient [50]. The calculation formula is as follows:

$$WC=Min\left({\displaystyle \frac{249}{V}},1\right)\times Min\left(\mathrm{1,0.9}\times TI/3\right)\times Min\left(1,K/300\right)\times Y$$

(10)

In the formula, WC represents the water conservation capacity of the study area (mm); V is the flow velocity coefficient, which varies depending on vegetation type, and its value is determined based on existing references [51]; K is the soil saturated hydraulic conductivity; Y is the water yield; and TI is the topographic index.

K is derived from the National Soil Database and calculated using the Cosby soil transfer function. The formula is as follows [51]:

$$K=1.148\times {10}^{(-0.6+1.26\times {10}^{-2}\mathrm{c}2-6.4\times {10}^{-3}c1)}$$

(11)

In the formula, K represents the soil-saturated hydraulic conductivity (m/d); c₁ and c₂ are the clay and sand content of the soil (%), respectively.

The water yield of woodland in Zhangjiajie City is calculated using the water yield module in the InVEST 3.14.2 model. The water yield is primarily based on the water balance, which is the rainfall minus the actual evapotranspiration. It mainly includes canopy interception, soil water content, surface runoff, and litter water retention. Since the actual evapotranspiration cannot be directly obtained, it is calculated based on the ratio of potential evapotranspiration to rainfall. Therefore, the calculation formula for the water yield in Zhangjiajie City is as follows:

$$Y_{xj}=\left(1-{\displaystyle \frac{{AET}_{xj}}{P_x}}\right)\times P_x$$

(12)

In the formula, Y_xj represents the water supply of grid cell x in land-use type j (mm), AET_xj represents the actual evapotranspiration of grid cell x in land-use type j (mm), and P_x represents the annual precipitation of grid cell x (mm).

$${\displaystyle \frac{{AET}_{xj}}{P_x}}={\displaystyle \frac{1+w_x\times R_{xj}}{1+w_x\times R_{xj}+\frac{1}{R_{xj}}}}$$

(13)

In the formula, ${\displaystyle \frac{{AET}_{xj}}{P_x}}$ represents the ratio of actual evapotranspiration to rainfall, R_xj represents the aridity index of grid cell x in land-use type j, and w represents the ratio of vegetation water demand to precipitation, which is a non-physical parameter (dimensionless).

$$w_x=Z\times {\displaystyle \frac{{AWC}_x}{P_x}}+1.25$$

(14)

In the formula, Z is the Zhang coefficient [52], an empirical constant determined based on the seasonal rainfall characteristics of the study area, typically ranging between 1 and 30. After multiple model runs, when the Zhang coefficient is set to 1.1, the water yield is closer to the actual situation. ${\mathrm{A}\mathrm{W}\mathrm{C}}_{\mathrm{x}}$ represents the plant-available water content of grid cell x (mm), which is determined using the soil texture and depth, and this is used to quantify the amount of water provided by the soil for plant growth.

$$R_{xj}={\displaystyle \frac{k_{xj}\times {ETO}_x}{P_x}}$$

(15)

In the formula, ETO represents the potential evapotranspiration of grid cell x (mm), and k is the plant evapotranspiration coefficient. The plant evapotranspiration coefficient varies for different land-use types. This study refers to previous research to determine the plant evapotranspiration coefficients for different land-use types.

$${AWC}_x=Min\left({MSD}_x,{RD}_x\right)\times {PWC}_x$$

(16)

In the formula, MSD_x refers to the maximum soil depth, RD_x refers to the rooting depth of the plant, and PWC_x refers to the available water content.

In the InVEST 3.14.2 model, biophysical parameters are required, including the maximum root depth (mm), available water content (AWC), annual average precipitation (AAP), land use and land cover (LULC), potential evapotranspiration (PET), and plant evapotranspiration coefficient (Kc) (

Table 6. Biophysical parameters for each woodland type of the InVEST model.

Woodland Type		Maximum Root Depth (mm)	Plant Transpiration Coefficient	Vegetation Cover
Tree-dominated woodland	Broadleaf forest	6000	0.95	1
	Coniferous forest	5800	0.95	1
	Mixed coniferous–broadleaf forest	5600	0.95	1
Bamboo-dominated woodland		5400	0.95	1
Shrubland		5200	0.93	1
Sparse woodland		5200	0.90	1

) [50,51].

2.3.2. Analysis of Driving Factors

This study uses the Geodetector model to reveal the driving factors of ecosystem services. Its calculation formula is as follows [53]:

$$q=1-{\displaystyle \frac{\sum _{h=1}^k{N}_h{\delta }_h^2}{N{\delta }^2}}$$

(17)

where q is the influence of a single factor on the change in ecosystem services, q ∈ [0, 1]; h is the stratification of the driving factor (h = 1, …, L); N and N_h are the number of samples in a certain stratum and the entire area, respectively; ${\mathsf{\delta }}_{\mathrm{h}}^2 \mathrm{a}\mathrm{n}\mathrm{d} {\mathsf{\delta }}^2$ are the variance of a certain stratum and the area, respectively.

2.3.3. Trade-Off and Synergy Analysis

This study employs the Spearman correlation analysis method [54] to calculate the correlation coefficients among woodland ecosystem services in Zhangjiajie City, analyzing the trade-offs, synergies, and evolutionary trends of these services. Spearman’s rank correlation coefficient is a nonparametric statistical method used to measure the strength and direction of the monotonic relationship between two variables. Unlike the Pearson correlation coefficient, which measures linear relationships, Spearman analysis is based on the ranks of the variables rather than the original values and is therefore insensitive to outliers and the shape of the data distribution, making it suitable for analyzing nonlinear but monotonic relationships. The specific formula is as follows:

$${\mathrm{r}}_{\mathrm{s}}=1-{\displaystyle \frac{6{\displaystyle \sum _{i=1}^n{(P_i-Q_{\mathrm{i}})}^2}}{n(n^2-1)}}$$

(18)

where r_s is the Spearman rank correlation coefficient; P_i is the serial number of period i to period n arranged from small to large in numerical; Q_i is the serial number arranged in chronological order; N is the time period. A positive coefficient indicates a synergistic relationship between the two ecosystem services, while a negative coefficient indicates a trade-off relationship between the two ecosystem services (

Table 7. Classification of the trade-offs/synergies of ecosystem services in woodland in Zhangjiajie City.

Correlation Coefficient	Trade-Off/Synergy Classification
(−1–−0.5)	Strong trade-off
(−0.5–0)	Weak trade-off
0	No significant interaction
(0–0.5)	Weak synergy
(0.5–1)	Strong synergy

). This study uses the spearmanr function in the scipy library of Python 3.12.7 to directly calculate the correlation coefficient and p-value.

2.3.4. Coupling Coordination Degree Model

To further reveal the spatial and regional differences in the interactions among woodland ecosystem services, this study draws on the coupling coordination degree model [55] from physics to measure the strength of interactions and the degree of coupling coordination among the four ecosystem services. The coupling coordination degree model is usually used to measure the interdependent relationships between different subsystems in a composite system. The concept of coupling originates from physics, indicating the phenomenon where two or more systems interact and influence each other. The coupling degree is used to represent the intensity of interaction between systems. The coordination degree, on the other hand, refers to the measurement of coordination on the basis of the coupling degree. The final coupling–coordination model can reflect the coordination development level among systems. The calculation formula is as follows:

$$C=4\sqrt[4]{U_1\times U_2\times U_3\times U_4}/(U_1+U_2+U_3+U_4)$$

(19)

$$T=a{U}_1+b{U}_2+c{U}_3+d{U}_4$$

(20)

$$D=\sqrt{C\times T}$$

(21)

In the formula, C represents the coupling degree; U1, U2, U3, and U4 represent the four different ecosystem services; T is the comprehensive coordination index of the four ecosystem services; D is the coupling coordination degree, with values ranging from 0 to 1. a, b, c, and d are the weights of the different ecosystem services. Considering that the four ecosystem services participate equivalently, a = b = c = d = 0.25 is assigned. The natural breaks method is used to classify the coupling coordination index into high [0.5, 1], medium [0.2, 0.5), and low [0, 0.2) categories. This study standardized four ecosystem services and then used the grid calculator tool in ARCGIS 10.8 to calculate the coupling–coordination degree.

3. Results

3.1. Analysis of Woodland-Use Changes

The spatial distribution of different woodland types in Zhangjiajie City in different years does not vary much, but the distribution of different woodland types in each year varies significantly (Figure 2). Tree-dominated woodland is widely distributed and is the main woodland type, with larger areas in each district and county. The bamboo woodland is distributed in patches, mainly in the two branches of the Wuling Mountains in Cili County and the southern part of Wulingyuan District. Shrublands and sparse woodland are intermingled in a complicated and sporadic distribution, mainly in the suburbs of cities and towns and around villages, the southern part of the urban area of Yongding District, in the eastern part of Cili County, and in Badagongshan of Sangzhi County.

From 1995 to 2022, the annual change rate of woodland in Zhangjiajie City was 0.01%, indicating a relatively slow growth rate (Figure 2 and

Table 8. Woodland-use structure in Zhangjiajie City from 1995 to 2022. Unit: km².

Year	Tree-Dominated Woodland				Bamboo- Dominated Woodland	Shrubland	Sparse Woodland	Total
	Broadleaf Forest	Coniferous Forest	Mixed Coniferous–Broadleaf Forest	Subtotal
1995	2040.62	2353.96	893.78	5288.36	96.24	1366.58	575.35	7326.52
2005	2170.03	2602.98	802.01	5575.02	102.66	1334.98	422.72	7435.38
2015	2191.81	2956.14	641.91	5789.85	102.52	1365.02	156.73	7414.12
2022	2169.99	2606.45	1135.81	5912.24	109.48	1332.51	142.83	7497.06

). This suggests that the woodland in Zhangjiajie City is primarily managed through conservation measures, with minimal human interference, and it is well protected. However, in terms of woodland types, there were significant changes in various woodland types in Zhangjiajie City from 1995 to 2022. In terms of area, tree-dominated woodland and bamboo woodland generally increased, shrublands slightly decreased, and sparse woodland significantly decreased. The area of tree-dominated woodland showed a continuously increasing trend, with a cumulative increase of 623.89 km². Tree-dominated woodland, which occupied the largest area, dominated absolutely, increasing its proportion of total woodland from 72.18% in 1995 to 78.86% in 2022, with an average annual growth rate of 0.44%. This is attributed to the effective implementation of measures such as mountain closure for afforestation and forest quality improvement in recent years, leading to a steady increase in forest coverage in Zhangjiajie City. Bamboo woodland, which accounted for the smallest proportion at about 1.4% of the total woodland, increased by 13.24 km² during the period, showing a trend of an initial increase, followed by a slight decrease, and then a significant increase, with an average annual growth rate of 0.51%. The growth rates during the periods of 1995–2005 and 2015–2022 were notably higher than the average. From 1995 to 2005, the economic value of bamboo was relatively high, driving farmers’ enthusiasm for planting bamboo and resulting in a significant increase in bamboo woodland. From 2015 to 2022, the primary reason was the provincial policy during the period of the “14th Five-Year Plan” to develop a billion-yuan bamboo industry, which promoted the increase in bamboo woodland. Shrubland, which ranked second in area, accounted for about 18% of the total woodland and decreased by 34.07 km² during the period, showing a fluctuating trend of an initial decrease, followed by an increase, and then another decrease. This was mainly due to early mountain closure for afforestation, where natural growth favored shrubs, and later improvements transformed most shrubland into tree-dominated woodland, leading to a reduction in shrubland area. Sparse woodland, ranking third in terms of area, showed a continuously decreasing trend, with a cumulative reduction of 432.52 km². Its proportion of total woodland decreased from 7.85% in 1995 to 1.91% in 2022, with an average annual growth rate of −2.78%, which was significantly higher than the change rates of the other types. The main reason was that sparse woodland is also a primary source for forest quality improvement and is the preferred land type for the reservation of cultivated land and construction use. Therefore, during the study period, most of the sparse woodland was converted into other land types, resulting in a substantial reduction of 75.18% in area.

From the perspective of the dominant tree species in tree-dominated woodland, coniferous forests occupied the largest area, followed by broadleaf forests, while mixed coniferous–broadleaf forests had the smallest area. In terms of changes during the study period, both broadleaf forest and coniferous forest areas showed an overall increasing trend, with a slight decrease after continuous growth, increasing by 128.38 km² and 252.49 km², respectively. The area of mixed coniferous–broadleaf forests initially decreased continuously and then increased significantly, with a net increase of 242.03 km². This was mainly due to factors such as geographical conditions and tourism stimulation, leading to the selection of tree species such as Masson pine and Chinese fir, which are coniferous species, for afforestation and forest improvement in Zhangjiajie City. Consequently, the areas of coniferous forests and mixed coniferous–broadleaf forests increased significantly. Additionally, the development of medicinal plants and forest fruit economies drove the increase in broadleaf forest areas.

3.2. Analysis of Changes in Woodland Ecosystem Services

3.2.1. HQ

The regional differences in woodland HQ in Zhangjiajie City in different years are significant (Figure 3). Near-town areas have fewer high-grade areas, and far-from-town areas have more high-grade areas. The higher-grade areas were more widely distributed in 1995, mainly in the southern part of Yongding District, the western part of Wulingyuan District, the northern and southeastern parts of Cili County, and the northern part of Sangzhi County. The distribution of high-grade areas has significantly decreased since 2005, mainly scattered in small patches in the north of Sangzhi County and the south part of Yongding District. The low-grade area shows contrary results. Moreover, the low-grade area is obviously in the form of independent blocks before 1995, and gradually connects into pieces after 2005. The HQ of the woodland in Zhangjiajie City formed three relatively distinct stages: the HQ in 1995 was relatively good; HQ in 2005 and 2015 was general; the HQ in 2022 was relatively poor.

During the study period, the overall HQ in Zhangjiajie City declined significantly, with the area of medium-or-above HQ decreasing from 98.30% to 15.88% (

Table 9. Distribution of the HQ of woodland in Zhangjiajie City. Unit: km².

Year	Lowest	Lower	Medium	Higher	Highest
1995	26.86	114.76	377.48	830.60	5976.82
2005	392.95	1174.79	2371.30	2052.13	1444.21
2015	507.37	1470.68	2444.74	1874.69	1116.65
2022	3613.42	2693.38	844.64	264.77	80.84

). By grade, the areas of higher HQ and the highest HQ both decreased by 5994.96 km² and 463.27 km², respectively. In contrast, the areas of medium, lower, and lowest HQ all increased by 543.34 km², 2599.90 km², and 3583.80 km², respectively. The HQ in each district and county also showed a notable decline. By 2022, Sangzhi County still had 24.71% of its area with medium-or-above HQ, which was slightly better than other districts and counties. In contrast, Cili County had only 7.71% of its area with medium-or-above HQ, the lowest among all districts and counties.

Among the various woodland types, the area of high-grade HQ generally showed a continuous decline, while the area of low-grade HQ mostly increased. Broadleaf forests had the highest proportion of medium-or-above HQ at 25.95%, making it the best among all woodland types. In contrast, sparse woodland had the lowest proportion at only 0.54%.

The primary reason is that HQ is significantly influenced by urbanization and transportation infrastructure. Before 1995, urbanization levels were low, and transportation networks were underdeveloped, resulting in relatively high overall HQ in Zhangjiajie City’s woodland. However, with the rapid development of urbanization and the tourism industry, the implementation of the “village-to-village road connectivity” project, and the expansion and hardening of rural roads, the HQ of woodland in Zhangjiajie City continued to decline. By 2022, the rate of hardened road access in Zhangjiajie City’s administrative villages had reached 100%, and natural villages with 25 households or more were connected by cement (asphalt) roads, forming a rural road network that reaches every group and every household, with most of the road networks extending deep into the forests. The density of the road network in Zhangjiajie City has increased nearly eight times from 1995 to 2022. The urbanization level increased from 13.82% in 1995 to 53% in 2022. A-level scenic spots increased from 0 in 1995 to 30 in 2022. The large-scale construction of urban facilities, tourism infrastructure, and transportation roads has promoted economic growth, but it has also exerted tremendous pressure on the woodland habitat quality, especially in those surrounding urban areas and villages. The above analysis shows that the degradation of HQ in Zhangjiajie City presents a composite feature of “road-oriented spread and tourism-driven intensification”. To improve the HQ of Zhangjiajie City’s woodland, pilot habitat restoration can be carried out preferentially in places with good woodland quality such as Zhangjiajie National Forest, Jiangya Provincial Forest Park, Tianmen Mountain National Forest Park, and the line of Lishui and Loushui, etc., by setting up ecological protection points and ecological corridors, reconstructing the pattern of ecological security, and implementing the ecological transformation of some infrastructure, etc., and exploring the coordinated development of man and nature in the world heritage sites.

3.2.2. SC

The SC of woodland in Zhangjiajie City varied across different years (Figure 4). The maximum high-value area of SC was 1635.21 tons/pixel in 2015, while the minimum was 1108.23 tons/pixel in 2005, showing a relatively large gap. The low-value areas all had values of 0. The highest average value was 173.90 tons/pixel in 2015, and the lowest was 114.95 tons/pixel in 2022. The size and disparity of high-value areas, as well as the average values, indicate the strength of SC. The total SC in Zhangjiajie City decreased from 1.26 × 10⁹ tons in 1995 to 9.54 × 10⁸ tons in 2005, then increased to 1.43 × 10⁹ tons in 2015, and dropped again to 9.57 × 10⁸ tons in 2022, showing a noticeable overall fluctuating decline. Therefore, relatively speaking, the SC of woodland in Zhangjiajie City was stronger in 1995 and 2015, while it was weaker in 2005 and 2022. This indicates a clear temporal variation that was primarily related to the precipitation levels in different years. The high-value areas of SC in woodland are located in relatively high-altitude regions, which are mainly distributed in the northwest of Sangzhi County, the southwest of Yongding District, the border area between Cili County’s Yakou Reservoir and Sangzhi County, and the southwest of Cili County. The SC is lower around urban areas. Sangzhi County and Yongding District have relatively strong SC values, while Wulingyuan District and Cili County are relatively weaker. In tree-dominated woodland, broadleaf forests and mixed coniferous–broadleaf forests have a relatively higher SC per unit area, with four-period averages of 1.70 × 10⁵ tons/km² and 1.65 × 10⁵ tons/km², respectively, while sparse woodland has lower values (1.10 × 10⁵ tons/km²).

The process of spatiotemporal evolution of SC in Zhangjiajie City’s woodland is the result of comprehensive action of natural factors and human factors. Firstly, there are topographical factors. Zhangjiajie City has many mountains and hills, but few plains, which provides topographical conditions for accelerated soil erosion. The average slope of the ground surface in Zhangjiajie City is 16.51°, second only to Xiangxi in the province. Moreover, the area with a slope of more than 15° accounts for 57.13% of the urban area, and the erosion is more serious as the slope increases. Secondly, there is the factor of precipitation, which is affected by the monsoon and the topography, with more rainfall, greater rainfall intensity, and a wider coverage of torrential rain in Zhangjiajie City. The maximum annual rainfall in Zhangjiajie City reaches 2171 mm, and the distribution of precipitation in time and space is uneven, with the period from April to July as the rainy season. Zhangjiajie City’s northwest Badagongshan (Li River source area) is the largest rainstorm center in Hunan Province and Zhangjiajie City, with an average annual precipitation of 2300 mm; the southern Xiejiayan Township and Siduping Township are secondary rainstorm centers, with an average of 1664 to 1872 mm. The higher the altitude, the richer the precipitation and the greater the potential soil erosion, but because the vegetation coverage is better in these areas, the actual erosion is not large, resulting in a larger SC, forming a high-value area, but there are also cases where the precipitation is large, the potential soil erosion is large, the actual erosion is large, but the SC is also large. The maximum annual average precipitation in 1995 (1469.05 mm) and 2015 (1548.34 mm) was higher than that in 2005 (1213.97 mm) and 2022 (1212.78 mm), resulting in significant inter-annual differences in SC. Thirdly, there is the vegetation factor. Due to the combined impact of various factors, the primary vegetation in Zhangjiajie City has been severely damaged. The role of artificial vegetation in controlling soil erosion, conserving water, and regulating runoff is very limited, and the phenomenon of woodland erosion is very common; moreover, some forest fruit lands, such as Camellia sinensis, Camellia oleifera, Eucommia ulmoides, etc., due to long-term cultivation and reclamation, also have serious soil erosion, and the task of soil conservation is heavy. Fourthly, it is caused by human factors. The main human factors affecting SC include agricultural and forestry production, resource development, urban infrastructure construction, tourism development, and so on. With the rapid development of tourism in recent years, the economic level of Zhangjiajie City has been continuously improved, but the situation of soil erosion caused by unreasonable use still exists. According to the “Overall Planning of Mineral Resources of Zhangjiajie City” (2021–2025), the destruction of water and land resources by mineral development in the city is relatively serious, but the treatment rate and greening rate are relatively low. Moreover, with the continuous development of the urban economy in Zhangjiajie City, the scale of urban infrastructure construction and tourism development activities in the city has been expanding, resulting in more serious soil erosion hazards.

Measures can be taken in the following aspects to improve the ability of woodland SC. Firstly, implement spatial zoning and precise governance. The high-value areas around Badagongshan, Tianmen Mountain, and Yaokou Reservoir are subject to natural forest closure; for the low-value areas such as the valley plains in Cili County and the central urban radiation belt in Yongding District, a 30 m wide buffer forest belt is established along the roads to reduce runoff scouring; limit the density of tourism facilities. Secondly, the vegetation structure can be optimized, mainly by adopting pure forest to mixed forest, optimizing the proportion of broadleaf forest and coniferous forest in mixed forest, and transforming shrubland and sparse woodland into tree-dominated woodland. Select tree species combinations with both economic and ecological values to balance the needs of local economic development. For example, coniferous forests are preferred to Pinus massoniana (timber), Cunninghamia lanceolata (fast-growing timber); broadleaved forests are preferred to Cyclobalanopsis glauca (strong CS), Magnolia officinalis (medicinal value), Camellia oleifera (economic forest fruit); At the same time, actively develop the underforest economy, planting shade-tolerant medicinal materials (such as Polygonatum sibiricum, Paris polyphylla) or edible fungi (such as Lentinula edodes) to promote local economic development. Thirdly, implement engineering facilities to control gully erosion, such as the construction of biological stone dams in Sangzhi County, planting Vitex negundo var. heterophylla with developed roots to intercept silt, popularizing terraced fields transformation in Cili County agriculture area, and excavating fish scale pits on bare slopes, etc. Fourthly, carry out institutional innovation from the aspects of ecological compensation system and community co-governance, such as the pilot horizontal compensation system, in which Yongding District and Wulingyuan District pay 1% of tourism revenue to Sangzhi County and Cili County for ecological restoration services, which are used exclusively for ecological restoration and forest management transformation. Attempt to establish an “Ecological Service Credit” trading platform, where enterprises need to purchase credit quotas for occupying woodland (1 hectare of development = 2 hectares of restoration), and incentivize social capital to participate in ecological restoration. Actively build a community co-governance network and establish an ecological cooperative. Villagers invest woodland as shares and participate in afforestation and management, and the benefits are distributed according to the shares (such as management wages, dividends). Experiment with an “ecological points” system, where farmers can earn points for implementing returning farmland to forestry, which can be exchanged for agricultural input subsidies or CS benefits, to boost participation.

3.2.3. CS

The regional differences in CS in Zhangjiajie City’s woodland were also significant across different years (Figure 5). The maximum value for each year in the chart was 42.88 tons/pixel, and the minimum value was 19.34 tons/pixel. The highest average value was in 2022 at 37.79 tons/pixel, followed by 2015 at 37.57 tons/pixel and then 2005 at 36.87 tons/pixel; the lowest was in 1995 at 36.28 tons/pixel. During the study period, the CS in Zhangjiajie City’s woodland continuously increased from 2.95 × 10⁸ tons in 1995 to 3.15 × 10⁸ tons in 2022, indicating that the efforts to improve the woodland CS in Zhangjiajie City were effective. The spatial distribution of CS shows fewer high-value areas near urban regions and more high-value areas in regions farther from urban centers. The distribution of high-value and low-value areas is generally similar across different time points, except for the northwest part of Cili County, which shows noticeable differences between 1995 and other years. High-value areas are mainly distributed in the southern part of Yongding District, the central part of Wulingyuan District, the northern and southeastern parts of Cili County, and the northern part of Sangzhi County, forming relatively contiguous patches. Low-value areas are primarily located around urban roads and appear as scattered blocks or strips within high-value areas. Among the districts and counties, Wulingyuan District has the highest CS per unit area, ranging from 4.21 × 10⁴ to 4.27 × 10⁴ tons/km². Among the different woodland types, mixed coniferous–broadleaf forests in tree-dominated woodland have the highest CS per unit area at 4.76 × 10⁴ tons/km², while sparse woodland has the lowest at 2.15 × 10⁴ tons/km². The distribution of high-value and low-value CS areas in Zhangjiajie City’s woodland is highly correlated with the distribution of different woodland types. Relevant studies indicate that the transfer between high-carbon-density and low-carbon-density land types is a key factor driving changes in CS. Tree-dominated woodland types such as broadleaf forests, coniferous forests, and mixed coniferous-broadleaf forests have high carbon densities and stronger CS capabilities. Therefore, regions with larger areas of tree-dominated woodland tend to have higher CS values.

The reasons for the above-mentioned spatiotemporal evolution characteristics of the CS of Zhangjiajie City’s woodland mainly include two aspects. The first is the positive driving force, including the impact of natural factors, and high-value areas are mostly located in mountainous areas with higher altitude and abundant precipitation, which is conducive to the growth of tree-dominated woodland; the optimization of woodland structure, which increases the area of tree-dominated woodland with high carbon density such as mixed coniferous–broadleaf forests; and the implementation of ecological projects such as closing hillsides to facilitate afforestation and returning farmland to forests, which improves the area and quality of woodland. The second is the negative driving force, including the development of tourism; the expansion of construction land occupying woodland or making woodland fragmented; the conversion of inefficient woodland, mainly involving the conversion of high-carbon-density woodlands into low-carbon-density woodlands, resulting in the loss of CS; and the pressure of abnormal climate change.

Although the CS of the woodland in Zhangjiajie City has continued to increase, there are signs of a decrease in the CS per unit area in Yongding District and Wulingyuan District, which need to be paid attention to. In order to promote the sustainable development of Zhangjiajie City and achieve the goal of “double carbon”, and improve the CS of woodland, we can start from the following aspects. Firstly, the internal structure of woodland should continue to be optimized, and measures such as pure forest to mixed forest, mixed forest quality improvement, and low-efficiency forest quality improvement should be implemented to optimize the woodland structure. Secondly, it is necessary to strengthen the measures of spatial control. We can try to divide the woodland into protected areas, buffer zones, and improvement zones, strictly manage the control system, and manage the areas separately. Thirdly, we should carry out institutional innovation, establish a carbon sink trading mechanism, and use remote sensing, big data, and large AI models to build an airspace–ground integrated monitoring network; we can also pilot community participation models.

3.2.4. WC

The distribution of WC in woodland also shows significant spatial and temporal differences (Figure 6). The maximum value of high-value areas was in 2015 at 676.04 mm, while the minimum was in 2022 at 423.59 mm. The highest average value was also in 2015 at 214.85 mm, and the lowest was in 2022 at 129.34 mm. This similarly indicates that the WC of Zhangjiajie City’s woodland was strongest in 2015 and weakest in 2022. The WC in Zhangjiajie City increased continuously from 144.37 mm in 1995 to 214.85 mm in 2015 and then decreased to 129.35 mm in 2022, showing an overall fluctuating decline with noticeable interannual variations. This is primarily related to significant differences in precipitation across different years. Additionally, in 2022, the annual average temperature in Zhangjiajie City was higher than usual, enhancing evapotranspiration and exacerbating WC. Other factors include the negative effects of tourism development and the lagged effects of returning farmland to forest.

High-value areas were mainly distributed in Wulingyuan District, the Badagongshan area in northern Sangzhi County, some strip areas in southern Yongding District, Gaogiao Town in southern Cili County, and the western part of the county’s urban area. Low-value areas were concentrated in Cili County and the central part of Yongding District. This study uses the average WC per unit area over four years to characterize WC capacity. Wulingyuan District has the strongest WC capacity (2.58 × 10⁵ m³/km²), while Cili County has the weakest (1.28 × 10⁵ m³/km²). Among tree-dominated woodland types, mixed coniferous–broadleaf forests have the strongest WC capacity (1.89 × 10⁵ m³/km²), while sparse woodland has the weakest (1.04 × 10⁵ m³/km²). The main reason is that mixed coniferous–broadleaf forests, through their multi-layered structure of a main canopy, sub-canopy, shrub/grass layer, and litter layer, form a multi-level interception system, with higher canopy interception than that in pure broadleaf forests. Additionally, the root structure of mixed coniferous–broadleaf forests enhances soil erosion resistance, and their soil erosion modulus is significantly lower than that of mixed broadleaf forests and pure coniferous forests. The above analysis indicates that mixed coniferous–broadleaf forests, through the synergy of multi-level ecological functions, are the optimal forest type for WC.

Because the multi-layer structure of the tree canopy, the litter layer, and the soil layer jointly achieve the WC function of the woodland. The influencing factors of the function of WC can also be divided into natural factors and anthropogenic factors. Natural factors include precipitation, topography, and soil characteristics. The more precipitation, the greater the general WC capacity; the topography will also vary with the difference in altitude, and the impact on WC capacity will also vary. Soil water content and soil water potential affect the WC capacity through the forest stand evapotranspiration. And the soil particle composition is the dominant factor affecting the soil water content. Anthropogenic factors include vegetation structure degradation and tourism development, etc. Vegetation structure degradation mainly refers to the conversion of woodland types with strong WC function into woodland types with low WC function due to human activities, such as the expansion of economic forests (camellia oleifera, citrus) leading to a decrease in soil organic matter content, etc. Tourism development, however, leads to an excessive density of tourism infrastructure, exerting pressure on the WC function or even infringing upon key areas of WC. To improve the WC capacity of woodland, based on its influencing factors and spatial–temporal evolution characteristics, it is also possible to start from spatial zoning control, optimization of the vegetation structure, and the management of litter and implement engineering measures.

3.3. Analysis of Driving Factors for Woodland Ecosystem Services

In recent years, with the intensification of human activities in woodland ecosystems, changes in woodland ecosystem services and their driving factors have garnered widespread attention. The analysis of driving factors for woodland ecosystem services in Zhangjiajie City is crucial for the sustainable development of its woodland. By assessing the driving factors of different ecosystem services in woodland, the effectiveness of woodland resource management and environmental protection can be improved, promoting regional ecological balance. Clarifying the driving factors of woodland ecosystem services helps to advance the synergy between woodland ecological protection and economic development. The Geodetector, a spatial analysis method capable of detecting spatial heterogeneity and revealing causal relationships between variables, provides new perspectives and tools for analyzing the driving factors of woodland ecosystem services. This study selects 10 influencing factors from five dimensions: climate change, topography, vegetation, soil, and socioeconomics (

Table 10. Influencing factors of the ecosystem services of the woodland in Zhangjiajie City.

Type	Specific Driving Factors
Climate	MAT (Mean Annual Temperature)
	AAP (Annual Average Precipitation)
Topography	Elev. (Elevation)
	Slope
Vegetation	LAI (Leaf Area Index)
Soil	SOM (Soil Organic Matter)
Socioeconomics	GDP (Gross Domestic Product)
	PD (Population Density)
	NLI (Nighttime Light Index)
	TDI (Tourism Dependency Index)

). And, the test between these influencing factors by using the Ordinary Least Squares method shows that there is no collinearity between them. The factor detector and interaction detector of the Geodetector are used to analyze the contributions of single-factor and dual-factor effects of these influencing factors on woodland ecosystem services in Zhangjiajie City.

The driving force of each factor in different woodland ecosystem services is shown in Figure 7. The driving factors influencing HQ, SC, WC, and CS in woodland exhibit both similarities and differences. The commonality lies in the fact that single factors—primarily natural factors such as climate and topography—have a dominant influence, while other single factors have relatively smaller impacts. From the perspective of interaction effects, the q-values after factor interactions are greater than those of individual factors, indicating a dual-factor enhancement and suggesting synergistic effects among factors that collectively influence HQ, SC, WC, and CS in woodland. The influence of single factors and interaction factors generally shows fluctuating changes across different years.

The differences lie in the various single factors and interaction factors with larger q-values. Among the single factors affecting HQ, the factor with the largest q-value is MAT, followed by Elev. Both show a trend of an initial increase, followed by a decrease and then another increase, reflecting the phased characteristics of climate change and ecological restoration. Among the interaction factors, the strongest interactions vary by year. MAT and Elev., when interacting with other factors, significantly enhance the driving force of these factors. The strongest interactions in 1995 and 2022 were between MAT∩LAI, while in 2005 and 2015, they were between MAT∩NLI. This indicates that urban development, under the influence of the synergy of natural factors, amplifies its impact on HQ. This requires us to pay attention to the combined effects of climate warming and urban expansion and to strengthen the construction of ecological corridors to mitigate the negative effects of tourism activities.

Among the single factors affecting SC in woodland, the factor with the largest q-value is the slope. The driving effect of the slope varied little across the four years. Similarly, it significantly enhances the driving force of other factors through interactions, with the strongest being slope ∩ AAP and slope∩Elev., showing little interannual variation. This suggests that the slope directly influences soil erosion by altering surface runoff paths and soil erosion resistance. We can enhance SC through slope management projects (e.g., terracing) and the application of organic fertilizers to improve soil erosion resistance while reducing unreasonable land-use conversions.

Among the single factors affecting the WC in woodland, the factor with the largest q-value is SOM. The strongest interactions are for SOM∩TDI and SOM∩AAP, indicating that improving soil pore structure enhances WC capacity. At the same time, TDI has a significant impact on WC. Tourism activities indirectly affect WC through changes in surface cover, necessitating the optimization of forest vegetation configuration (e.g., increasing the proportion of mixed coniferous–broadleaf forests) to enhance SOM accumulation and WC capacity.

The driving effects of various factors on the CS in woodland are relatively weak. Factors with larger q-values include Elev., MAT, and SOM. The strongest interaction in 2022 was that of MAT∩TDI, while in other years, it was that of SOM∩MAT, indicating that the impact of tourism development on SC is gradually increasing. Tourism may affect carbon balance by altering surface-cover types (e.g., underforest economic development) or increasing carbon emission sources.

The above analysis shows that we need to give priority to climate-adaptive design in woodland ecology protection and management. It is necessary to regulate the development of urbanization and tourism, strictly implement the management of three zones and three lines, optimize the layout of tourism facilities, and promote the model of eco-tourism.

3.4. Trade-Off and Synergy Analysis of Woodland Ecosystem Services

In 1995, 2005, and 2015, the correlation coefficients of the four ecosystem services were all positive but less than 0.5, indicating weak synergistic relationships (Figure 8). However, by 2022, the relationship between SC and CS shifted from a weak synergy to a weak trade-off, which was primarily due to a significant reduction in precipitation in 2022. The other relationships remained synergistic. The synergy between HQ and the other three services was relatively strong. The synergistic relationship between HQ and SC fluctuated but was generally strengthened, while the correlation coefficient between HQ and WC initially increased and then decreased, remaining largely unchanged overall. This was mainly because HQ indirectly enhanced the synergistic effects of SC and WC by maintaining vegetation cover and biodiversity. The synergy between HQ and CS continued to weaken, which was possible because CS relies more on long-term stable vegetation structures, and short-term ecological restoration struggles to rapidly enhance carbon sequestration capacity. The correlation coefficients between SC and WC and between SC and CS fluctuated and decreased, indicating weakening synergy and even trade-offs. The correlation coefficient between WC and CS also continued to decrease, showing a weakening synergistic relationship. The dynamic changes in the correlation coefficients of the four ecosystem services in Zhangjiajie City are influenced by climate change, human activities, and functional competition among ecosystem services. Therefore, in woodland management, differentiated restoration strategies, climate-adaptive management, and enhanced policy synergy and community participation are imperative to promote the synergy of woodland ecosystem services. In the context of a decreasing trend in precipitation, prioritize deep-rooted tree species (such as Cyclobalanopsis glauca) and mycorrhizal plants (like Pinus) in drought-sensitive areas (like the valley in Cili County) to enhance drought resistance and carbon stability, while also considering surface soil fixation and deep-layer CS; Expand the enclosure of natural forests in the core area of Wulingyuan, prohibit tourism development, and maintain synergistic gains driven by biodiversity. Establish a new system of ecological compensation, and incorporate the synergistic benefits of carbon sinks and soil and water conservation into the compensation standards. Establish a community co-governance platform to involve villagers in the management of ecosystem services, allocate benefits based on contributions, and enhance ecological conservation awareness.

3.5. Coupling Coordination Analysis of Woodlan Ecosystem Services

The spatial distribution of the coupling coordination degree of HQ, SC, WC, and CS in Zhangjiajie City’s woodland during the study period is shown in Figure 9. High-value areas are widely distributed, and they are generally located far from urban centers and concentrated in higher-altitude regions. These areas are notably clustered in the central part of Wulingyuan District, the southern and southeastern parts of Yongding District, the north–central, central–eastern, and northwestern parts of Sangzhi County, and the northwestern and southwestern parts of Cili County. Within these regions, the various functions of woodland develop in a coordinated manner. Medium-value areas are scattered, with notable concentrations in the central–southern part of Yongding District, Wudaoshui and Renchaoxi Towns in Sangzhi County, and Guangfuqiao Town in Cili County. Low-value areas are relatively small and mainly concentrated around the urban areas of each district and county. The area ratio of low-value area continues to decrease from 10.99% to 5.17%, and the area ratio of medium value continues to increase from 34.24% to 70.17% (

Table 11. The area ratio of the coupling coordination degree of the four ecosystem services in woodland of Zhangjiajie City from 1995 to 2022.

Year	High	Medium	Low
1995	54.77%	34.24%	10.99%
2005	54.91%	36.35%	8.74%
2015	57.04%	37.61%	5.35%
2022	24.16%	70.17%	5.17%

). The area ratio of high-value areas reached a peak in 2015 (57.04%), but significantly declined in 2022 (24.16%), while the area ratio of medium-value areas increased significantly in 2022. This suggests that the overall coupling coordination of the four ecosystem services was relatively good, but there was a risk of decline in 2022. The main reason is that the other three types of the four types of ecosystem services, except for CS, have declined to some extent in 2022 due to natural, man-made, and other reasons. The spatial heterogeneity of the coupling coordination of woodland ecosystem services in Zhangjiajie City is significant, and it is necessary to take “zoning control and dynamic monitoring” as the core, combined with ecological restoration, woodland optimization, and policy innovation, to achieve a balance between ecological security and sustainable development. The focus should be placed on the risk tracing of the high-value area decline in 2022, strengthening the cross-sector governance and public participation, and ensuring the long-term stability of the ecological conservation goals.

4. Discussion

4.1. Driving Mechanisms of Dynamic Changes in Ecosystem Services

This study reveals that the ecosystem services of woodland in Zhangjiajie City exhibit significant spatial and temporal heterogeneity, with their evolution being closely related to the interaction of natural and anthropogenic factors. Natural factors (climate, topography) are the fundamental drivers of service changes, while the influence of socioeconomic factors (TDI, NLI) is increasing. The continuous decline in HQ is directly related to urbanization, transportation expansion, and tourism development. This is consistent with the findings of Wang et al. [56] in Huangshan City, indicating that the stress effects of human activities on habitats are regionally universal. The fluctuation in SC functions is closely tied to precipitation patterns (e.g., the peak during the wet season in 2015), echoing the “soil–vegetation–hydrology” coupling mechanism proposed by Li et al. [57]. Notably, the continuous increase in CS (from 2.95 × 10⁸ tons to 3.15 × 10⁸ tons) supports the conclusion of Liu et al. [58] on the enhancement of CS through forest structure optimization. However, the weakening synergy between HQ and CS suggests that short-term ecological restoration struggles to balance biodiversity and long-term CS.

4.2. Temporal and Spatial Characteristics of Trade-Offs in Ecosystem Services

During the study period, the four services were primarily weakly synergistic, but the trade-off between SC and CS in 2022 highlights the conflict between climate fluctuations and human intervention. This may stem from reduced extreme precipitation (18% less in 2022 compared with normal years), which weakened soil erosion but simultaneously suppressed the vegetation CS under drought stress. This aligns with the finding of Cui et al. [59] that the impact of drought on ecosystems exhibits nonlinear thresholds, and abrupt changes in vegetation function may occur when climatic factors (e.g., temperature, precipitation) exceed these thresholds. The continuous weakening of the synergy between WC and CS reflects the lagged effects of forest structure changes—while mixed coniferous–broadleaf forests improved WC efficiency (1.89 × 10⁵ m³/km²), their CS potential requires longer accumulation periods. This is consistent with the findings of Zhang et al. [60] in a multi-scale study of the Funiu Mountain region.

4.3. Limitations and Future Directions

This study has the following limitations: (1) The InVEST 3.14.2 model’s assessment of HQ relies on parameters of the sensitivity to land-use type, and future studies should optimize threat factor weights through field surveys. (2) The CS calculations did not account for differences in stand age structure, potentially underestimating the CS potential of mature forests. (3) The socioeconomic data resolution is low (1 km), making it difficult to capture the fine-scale impacts of tourism activities on local ecosystem services.

Future research could expand in the following directions: (1) incorporating more service types to build a comprehensive assessment framework; (2) combining drone remote sensing and ground monitoring to improve model accuracy; (3) exploring the applicability of “ecological compensation + carbon trading” mechanisms in Zhangjiajie City to provide empirical support for policy formulation.

4.4. Implications for Regional Sustainable Development

This study provides scientific evidence for Zhangjiajie City’s “ecological priority, green development” strategy: (1) strengthening ecological corridor construction during tourism development to reduce the impact of landscape fragmentation on HQ; (2) promoting mixed coniferous–broadleaf forest models to balance WC and CS; (3) establishing a collaborative “climate–vegetation–soil” monitoring system to enhance the adaptive management of ecosystem services. These findings not only offer reference value for the Wuling Mountain area but also provide a Chinese case study for the coordinated development of ecological protection and tourism in global heritage sites.

5. Conclusions

This study analyzed the spatiotemporal dynamics of woodland ecosystem services (HQ, SC, WC, and CS) in Zhangjiajie City from 1995 to 2022. Using the Geodetector model, a correlation coefficient method, and a coupling coordination model, the driving factors, trade-offs, synergies, and spatial coupling coordination relationships were explored. The main conclusions are as follows.

• During the study period, the area of tree-dominated woodland and bamboo-dominated woodland in Zhangjiajie City continuously increased, with cumulative growth of 623.89 km² and 13.24 km², respectively, primarily due to ecological protection policies such as mountain closure for afforestation and forest quality improvement. The reduction in shrubland and sparse woodland areas reflects human disturbances such as tourism development and urbanization.
• HQ declined significantly, with the area of medium-or-above HQ decreasing from 98.30% to 15.88%, mainly due to the influences of urbanization, transportation expansion, and tourism activities. Pilot habitat restoration can be carried out in places with better woodland quality, and key ecological protection points, ecological corridors, and other measures can be set up to improve HQ. SC showed fluctuating changes, with high-value areas being concentrated in mountainous regions such as Sangzhi County and Yongding District. Measures such as the management of spatial zoning, optimizing vegetation structure, implementing engineering facilities control, and innovating the system of ecological compensation and community co-governance can enhance the SC capacity. WC peaked in 2015 and then declined, with 2022 being the worst; this is linked to reduced precipitation, increased evapotranspiration, and changes in surface cover due to tourism development. To improve the WC capacity of woodland, it is also possible to start from spatial zoning control, optimization of the vegetation structure, and the management of litter and implement engineering measures. CS increased continuously, with high-value and low-value areas being highly correlated with different woodland types. Mixed coniferous–broadleaf forests had the highest CS per unit area (4.76 × 10⁴ tons/km²). The enhancement approach is similar to other ecosystem services.
• HQ showed weak synergy with WC, SC, and CS, indirectly enhancing ecological functions through vegetation cover and biodiversity. In 2022, SC and CS shifted to a weak trade-off, reflecting the conflict between climate change (reduced precipitation) and short-term ecological restoration. The synergistic effect of WC and CS is weakened. The synergistic relationship of woodland ecosystem services can be promoted by implementing differentiated remediation strategies, climate-adaptive management, and strengthening policy synergy and community participation.
• Natural factors (MAT, AAP, Elev., slope) are the fundamental drivers, while socioeconomic factors (TDI, NLI) have weaker influences. The strengthening of interactions indicates that the combined effects of natural and anthropogenic factors require focused attention. Climate-adaptive design should be prioritized in woodland ecological conservation and management. We strictly implement the management of Three Zones and Three Lines and promote the model of eco-tourism.
• The overall coupling coordination of the four ecosystem services was good, but in 2022, high-value areas decreased significantly, and medium-value areas increased substantially, indicating a higher risk of declining coupling coordination. There is a need to strengthen cross-sectoral collaborative governance and public participation to achieve a balance between ecological security and sustainable development.

This study reveals the dynamic evolution patterns and driving mechanisms of woodland ecosystem services in Zhangjiajie City, as well as the trade-offs, synergies, and spatial coupling coordination relationships among different ecosystem services. It provides a scientific basis for the ecological protection and high-quality sustainable development of woodland in Zhangjiajie City. Future research should further explore the long-term impacts of climate change and the compound effects of new tourism models on ecosystem services, deepening multi-scale studies of the driving mechanisms.