Quantitative Evaluation of the Integrity of Natural Ecosystems and Anthropogenic Impacts in Shennongjia National Park, China

Abstract

China launched its pilot national park system in 2015, and Shennongjia National Park has attracted much attention as one of the first batch of pilot national parks. The primary goal of national park construction in China is to effectively protect the integrity and authenticity of nationally important natural ecosystems. Based on the theory of landscape ecology, this study interpreted data from high-resolution remote sensing images and used landscape pattern analysis, spatial correlation analysis, and geographic weighted regression analysis to quantitatively evaluate the integrity of natural ecosystems at the landscape scale. A more accurate and operable calculation method was proposed to analyze the spatial variation in natural ecosystem integrity and to explore the scope and intensity of the impact of different anthropogenic activities on natural ecosystem integrity. The results showed that: (1) from the perspective of the spatial distribution patterns of natural ecosystem integrity, the degree of natural ecosystem integrity of Shennongjia National Park was generally high, with an integrity index value of 96.06, and the proportion of high-integrity areas accounted for 72.38%. The integrity index value of the strict protection zone was 98.83, and the proportion of the low-integrity index area only accounted for 0.01% of the strict protection zone, which was mainly distributed in the main urban areas of the nearby townships and along the highways in the national park, as well as in other areas with intensive anthropogenic activities. (2) From the perspective of the degree of impact of anthropogenic activities on natural ecosystem integrity, population density (0.3344), traffic accessibility (0.2389), traditional utilization activities (0.1101), and industrial and mining activities (0.0095) were, in descending order, the most impactful, and there was no significant correlation between ecotourism activities and natural ecosystem integrity. (3) From the perspective of the impact range of anthropogenic activities on natural ecosystem integrity, traditional utilization activities had the largest impact, accounting for 19.71% of the total area of the national park. The area affected by population density accounted for 1.52%. Industrial and mining activities had an influence of 4.75%, and the area affected by traffic accessibility accounted for 9.28%. Through conducting quantitative research into the integrity of natural ecosystems in Shennongjia National Park, this study provides a new research paradigm for the conservation of natural ecosystems and for the sustainable development of resources in protected areas, which is of great significance for the sensible development of national park conservation and management.

1. Introduction

National parks in China play a dominant role in the system of natural protected areas, and their primary function is to ensure that the authenticity and integrity of important natural ecosystems are effectively protected. The purpose of evaluating the integrity of natural ecosystems in national parks is to protect the ecological environment, biodiversity, and natural resources, and to provide a basis for the rational planning of protected areas and the formulation of corresponding conservation policies by management authorities. At present, the study of natural ecosystem integrity is mainly understood from two perspectives. On the one hand, integrity is considered to be the optimal state of an ecosystem in a given geographic area in terms of the components of that natural ecosystem; that is, the ecosystem has all the native biodiversity and ecological processes that the given region’s natural habitats should contain, the ecosystem’s structure, functions, and processes are not threatened or compromised by humans, the ecosystem is within the range of natural variability and maintains appropriate natural cycles, and the native species are at a population level that can sustain reproduction [1,2,3]. On the other hand, the systemic characteristics of natural ecosystems are described in terms of their stability, resistance and resilience, and self-organization; these characteristics include ecosystem health, biodiversity maintenance, sustainability, and naturalness, which together constitute physical, chemical, and biological integrity [4,5,6,7].

The concept and evaluation of the integrity of natural ecosystems has been discussed and studied extensively by domestic and international scholars. The Panel on the Ecological Integrity of Canada’s National Parks first noted that the natural ecosystem integrity of a national park is defined as the abundance, rate of change, and supporting processes of species composition, native species, and biological communities that are typical of the natural area in which the park is located; moreover, its structure and function should not be stressed by anthropogenic disturbances, and ecosystem biodiversity and supporting processes should remain intact [8]. Some experts have used the vegetation-based Index of Biotic Integrity (vIBI) to assess the ecological integrity and health of thousands of hectares of natural, degraded, or reclaimed wetlands in Alberta, Canada [9]. The U.S. National Park Service (NPS) defines ecosystem integrity as the ability to maintain ecosystems with the same species composition, abundance, and structure, as compared to ecosystems under natural conditions, in a given region [10]. The Ecological Integrity Assessment Framework (EIAF), proposed by the NPS, is considered the most highly developed method for assessing ecosystem integrity. The EIAF requires managers to carefully select the most representative monitoring indicators based on the typical characteristics of various ecosystems in national parks, such as forest lands, wetlands, and grasslands, and to independently construct assessment criteria for different ecosystems at different scales. At the landscape level, ecosystem integrity can be assessed on a large scale by remote assessment using GIS and remote sensing technology. For example, land-use data and vegetation net primary productivity data are used [11]. At the small and medium scales, indicators such as vegetation composition, animal population density, and species habitat connectivity can be selected and corrected for large scale remote sensing data, or by intensive assessment.

In recent years, many Chinese research teams have conducted studies on the quantitative evaluation of the natural ecosystem integrity of national parks. A team from Beijing Forestry University evaluated the pilot area of Qianjiangyuan National Park in relation to five aspects (the protected area, protected landscape, protected vegetation, flagship suitable habitat, and main food chain) by constructing a quantitative assessment index system [12]. A team from the Beijing Municipal Research Institute of Environment Protection developed a framework for assessing the integrity and authenticity of the ecosystems in the potential construction areas of the Tibetan Plateau National Park Complex; the framework considered habitat quality, biological communities, ecosystem services, natural conditions and hazards, human activities, and human landscapes [13]. A team from the Forestry and Grassland Survey and Planning Institute of the National Forestry and Grassland Administration established qualitative indicators from three dimensions of ecosystem structure and process integrity, spatial integrity, and spatial pattern integrity, and evaluated the natural ecosystem integrity of the Northeast Tiger and Leopard National Park using the expert scoring method [14]. Moreover, some experts and scholars have also evaluated the natural ecosystem integrity of the natural ecosystem of the Sanjiangyuan National Park using the index construction method [15]. Based on these studies, a series of standards and specifications have been issued, such as the Specification for National Park Establishment (GB/T 39737-2021) [16], the Specification for Assessment of National Park (GB/T 39739-2020) [17], and the Regulation of Resources surveying and Evaluating in National Park (LY/T 3189-2020) [18]; together, these provide many research bases for the evaluation of natural ecosystem integrity in national parks.

However, the index systems or methods used in these studies are too complex to quickly evaluate the integrity of natural ecosystems. Furthermore, most of these studies evaluate the integrity of natural ecosystems in national parks or other nature reserves as a whole, and are unable to identify spatial differences in the integrity of natural ecosystems, making it difficult to provide targeted guidance for ecological restoration, the control of anthropogenic activities, and other conservation and management work. To address this issue, this study proposes a rapid method for assessing the integrity of natural ecosystems in national parks based on landscape ecology by conducting a case study that can accurately identify the spatial distribution patterns of natural ecosystem integrity.

In China, there are a total of 11,800 natural protected areas of different types, covering an area of more than 1.8 million km² [19]. Anthropogenic activities are widespread in these areas and have a significant impact on the integrity of natural ecosystems. Related studies have shown different types and degrees of anthropogenic activities in the 10 existing national parks (including 5 pilot areas of the national park) and 446 national nature reserves in China [20], and the intensity of anthropogenic activities shows an increasing trend [21]. Researchers have evaluated the impact of anthropogenic activities in various types of nature reserves from different perspectives, using methods such as the anthropogenic activity interference degree [22], the anthropogenic activity intensity index [23,24], and the anthropogenic footprint index [25,26], providing different perspectives for analysis. However, the correlation analysis and quantitative evaluation of the degree of impact of these anthropogenic activities on the integrity of natural ecosystems have rarely been studied. The method proposed in this study for assessing the integrity of natural ecosystems and identifying their spatial distribution patterns makes it possible to quantitatively evaluate the impact of anthropogenic activities on the integrity of natural ecosystems from a spatial perspective.

Shennongjia is one of the most important ecological functional areas and ecological vulnerability zones in the world; it has attracted much attention for its superior natural ecological endowments [27]. Shennongjia National Park was one of the first pilot areas of the national park system in China and will soon become part of the next batch of officially established national parks. Based on the theory of landscape ecology, this study selects Shennongjia National Park as the research object and quantitatively evaluates the integrity of its natural ecosystems and their spatial differences through landscape pattern analysis, GIS spatial analysis, spatial correlation analysis, and geographically weighted regression analysis. The scope and intensity of the impact of anthropogenic activities on the integrity of natural ecosystems are quantitatively discussed. This study provides a new research paradigm for the conservation of natural ecosystems and the sustainable development of resources in natural protected areas; it also provides a theoretical basis for the scientific implementation of national park conservation and management.

2. Materials

2.1. Overview of the Study Area

Shennongjia National Park (SNJNP) (Figure 1) is located in the western part of Hubei Province, Central China (109°56′3″ E–110°36′26″ E, 31°21′24″ N–31°36′27″ N); it has a total area of 116,988 hm², accounting for 35.97% of the Shennongjia Forest Region. The National Park covers as many as seven protected sites at various levels, including the Shennongjia World Natural Heritage Site, World Geopark, the World Man and Biosphere Reserve, the National Nature Reserve, the National Geological Park, the National Wetland Park, and the National Forest Park.

Shennongjia is located in the transition zone between the second and third terraces in China and is endowed with extremely rich resources. It has the only well-preserved subtropical forest ecosystem in the global mid-latitudes and the richest biodiversity in the world, and is a gene bank of global significance, known as the “Green Miracle”. The area is rich in ancient and rare endemic species, represented by the Shennongjia golden snub-nosed monkey, and has some of the best-preserved mixed evergreen deciduous broad-leaved forests in the northern hemisphere, with extremely rich wildlife resources. It has been listed as a “Priority area for biodiversity protection in the World” and has been selected by the United Nations “Man and Biosphere Program”. It is one of the fourteen key areas of international significance for biodiversity conservation and research and one of the sixteen hotspots for biodiversity conservation and research in China [28]. In November 2016, the Shennongjia National Park Administration was established and entered the pilot implementation phase of the national park system, which is the first batch of pilot areas of the national park system in China.

Shennongjia National Park is divided into four functional zones: the strict protection zone, the ecological conservation zone, the scientific and educational recreation zone, and the traditional use zone (Figure 1), with the areas accounting for 53.5%, 39.6%, 3.5%, and 3.4%, respectively. There are a total of 25 administrative villages in 5 townships, including Dajiuhu Township, Xiagu Township, Muyu Township, Hongping Township, and Songluo Township. There are 21,072 indigenous inhabitants in the national park, and anthropogenic activities are frequent, mainly comprising agricultural cultivation, mining, and industry, as well as ecotourism activities. These anthropogenic activities have a potential impact on the fragmentation of the natural ecosystem and the survival of the main protected objects in the area.

2.2. Data Sources and Processing

The remote sensing images of Shennongjia National Park used in this study were selected from a high-resolution cloud-free image (spatial resolution of 2 m) taken in 2020, which was captured by the China Gaofen-1 (GF-1) satellite, successfully launched in 2013. The remote sensing image was pre-processed with ENVI 5.1 for geometric correction, multi-band fusion, panchromatic sharpening, and cropping. According to Chinese standard of Current Land-Use Classification (GB/T 21010-2017) [29] and the actual situation of the study area, the landscape of the protected area landscape was classified into 6 primary landscape types and 18 secondary landscape types (

Table 1. Landscape types in Shennongjia National Park.

Primary Landscape Type	Secondary Landscape Type	Landscape Attribute	Primary Landscape Type	Secondary Landscape Type	Landscape Attribute
Woodland	High-canopy-density woodland	Protected landscape	Water and Wetland	River wetland	Protected landscape
	Low-canopy-density woodland	Protected landscape		Lake wetland	Protected landscape
	Shrubland	Protected landscape		Artificial wetland	Protected landscape
	Artificial woodland	Protected landscape		Mud flat	Protected landscape
Grassland	High-coverage grassland	Protected landscape	Artificial land	Industrial land	Interfered landscape
	Medium-coverage grassland	Protected landscape		Transportation land	Interfered landscape
	Low-coverage grassland	Protected landscape		Residential land	Interfered landscape
Cropland	Dry land	Interfered landscape		Other artificial land	Interfered landscape
	Greenhouse	Interfered landscape	Bare land	Bare land	Protected landscape

). After supervised classification, to test the accuracy of the interpretation results, the attribute accuracy and boundary accuracy of the interpretation results were found to be greater than 97%, thus meeting the requirements of this study in terms of accuracy.

3. Methods

3.1. Ecosystem Integrity Index

In this study, the Ecosystem Integrity Index (EII) was constructed to characterize the habitat integrity of the main protected objects in the national park, which include a mosaic of protected landscapes consisting of woodlands, grasslands, and water [30] (Table 1). The Landscape Fragmentation Index and the Edge Effect Index (the Fractal Dimension Index) were selected to evaluate the protected landscape and to provide a comprehensive characterization of the integrity of the natural ecosystem at the landscape level. The calculation formula is as follows [31]:

$$EII=\sqrt[3]{\frac{{{\displaystyle \sum }}_{j=1}^n{S}_j}{S_T}·\left(1-I_F\right)·\left(2-I_{FD}\right)}\times 100$$

(1)

$$I_F=1-{{\displaystyle \sum }}_{j=1}^n{\left(\frac{S_j}{{{\displaystyle \sum }}_{j=1}^n{S}_j}\right)}^2$$

(2)

$$I_{FD}={{\displaystyle \sum }}_{j=1}^n\left[\frac{S_j}{{{\displaystyle \sum }}_{j=1}^n{S}_j}·\frac{2\lg \left(0.25P_j\right)}{\lg S_j}\right]$$

(3)

where EII is the Ecosystem Integrity Index, with values ranges from 0 to 100; the higher the value, the lower the degree of fragmentation of natural ecosystem in protected areas, the weaker the edge effect, and the higher the degree of integrity. S_j is the area of the jth protected landscape mosaic patch; n is the number of all protected landscape mosaic patches; I_FD is the protected landscape edge effect index; P_j is the perimeter of the jth protected landscape mosaic patch; and S_T is the total area of the national park.

In order to explore the spatial variation in natural ecosystem integrity in Shennongjia National Park and to quantitatively evaluate the impact of anthropogenic activities on it, grids of 100 m, 300 m, 500 m, 1000 m, and 2000 m were used to divide the study area. The integrity index was calculated for each grid cell at different scales based on the formula of natural ecosystem integrity, and the grid of 500 m × 500 m was finally determined to be the best scale. Using this scale, the Shennongjia National Park was divided into 4992 grid cells, and the center of mass of each grid cell and its integrity index value were extracted. The spatial distribution of natural ecosystem integrity of Shennongjia National Park was obtained using the Kriging interpolation and resampling tools in ArcGIS (Esri, RedLands, CA, USA, 10.2).

3.2. Anthropogenic Activity Indicator System

In this study, five anthropogenic activity indicators were selected: population density, traditional utilization activities, industrial and mining activities, ecotourism activities, and traffic accessibility (

Table 2. Anthropogenic activity indicators.

Index	Index Meaning	Selection Basis
Population Density	The degree of population concentration.	The higher the population density, the greater the demand for natural ecosystem resources and the greater the impact on the ecosystem [32].
Traditional Utilization Activities	Agricultural cultivation, farming, fishing and other activities; traditional utilization in Shennongjia National Park is dominated by cultivation.	The traditional use of the national park by its indigenous inhabitants is dominated by agricultural cultivation, which is a potential threat to the natural ecosystem [33].
Industrial and Mining Activities	Activities such as industrial production and mineral exploitation.	Industrial and mining activities cause serious surface damage and threaten the balance of the natural ecosystems within a certain area [34].
Ecotourism Activities	Activities such as ecotourism, forest recreation, research, and education.	Ecotourism activities pose a potential threat to the natural ecosystem of the national park, while providing recreational services and shared access for all [35].
Traffic Accessibility	The degree of ease with which humans can move from one location to another using a particular transport system.	Traffic accessibility increases as the level of the road and the degree of impact on the natural ecosystems along the road increases [30,36].

). The degree and range of their impact on the integrity of natural ecosystems were calculated separately.

As there are differences in the orders of magnitude and positive and negative effects of each evaluation indicator, in order to eliminate the possible influences of these differences on the data analysis, this study adopts the range method to standardize the raw data of each indicator; it then extracts the values of each anthropogenic activity indicator to a 500 m × 500 m grid cell to obtain the spatial distribution of various anthropogenic activities (Figure 2). The calculation formulae for the standardization of the indicators are as follows:

$$\mathrm{Positive} \mathrm{evaluation} \mathrm{index}: {\mathit{Y}}_{\mathit{i}}=\frac{{\mathit{X}}_{\mathit{i}}{ - \mathit{X}}_{\mathit{min}}}{{\mathit{X}}_{\mathit{max}}{ - \mathit{X}}_{\mathit{min}}}\times 10$$

(4)

$$\mathrm{Negative} \mathrm{evaluation} \mathrm{index}: {\mathit{Y}}_{\mathit{i}}=\frac{{\mathit{X}}_{\mathit{i}} {- \mathit{X}}_{\mathit{min}}}{{\mathit{X}}_{\mathit{max}}{ - \mathit{X}}_{\mathit{min}}}\times 10$$

(5)

where Y_i is the value of the ith indicator after standardized calculation, and its range is 0–10. The larger Y_i is, the higher the ecological vulnerability of the region is, and the natural ecosystem is vulnerable to external disturbance and damage. X_i is the original value of the ith index, X_max is the maximum value of the original value of the ith index, and Xmin is the minimum value of the original value of the ith indicator.

3.2.1. Population Density

Referring to the method of Venter et al. [36] for assessing population density, the raster with a population density greater than 1000 persons per square kilometer was assigned a value of 100; meanwhile, for the raster with fewer than 1000 persons per square kilometer, the assignment was calculated according to the following formula. Reclassification was performed in ArcGIS 10.2 and the spatial distribution map was derived using spatial interpolation with the following formula:

P = 33.33 × log (P_d + 1)

(6)

where P is the population density assignment for the raster data and P_d is the population density value for each raster image element.

3.2.2. Traditional Utilization Activities

The main traditional utilization activities in Shennongjia National Park comprise agricultural cultivation, so this study selected cropland as the main area of traditional utilization activities to be analyzed. The land-use type of cropland was extracted in ArcGIS 10.2, and the Euclidean distance between each raster and cropland were calculated. According to the relevant research results, the maximum distance of the impact of cropland on the natural ecosystem was set to 500 m [37,38], and the raster with a distance greater than 500 m was assigned a value of 0. For the raster with a distance smaller than 500 m, the impact score of traditional utilization activities was calculated using the distance decay function in the habitat quality module of the InVEST model [39]. The formula is as follows:

T_u = exp [−(2.99/d_max)d] × 100

(7)

where T_u indicates the traditional utilization activity impact score, d is the raster data value calculated by Euclidean distance, and d_max is the maximum impact distance of traditional utilization activities, i.e., 500 m.

3.2.3. Industrial and Mining Activities

The exploitation of mineral resources will threaten the regional ecological environment to a certain extent, causing the atmospheric dispersion of pollutants and an increase in the surface temperature. The land cover of the mined area changes drastically during this process, generating a significant impact on the surface landscape pattern and the integrity of the natural ecosystems. The land-use type of industrial and mining land in the study area was extracted in ArcGIS, and the Euclidean distance between each grid and the industrial and mining land was calculated. According to the relevant research results, the maximum impact distance of industrial and mining land on the natural ecosystem was set to 1000 m [40,41], so the distance of the raster greater than 1000 m was set to 0. For the raster with a distance smaller than 1000 m, the distance decay function in the habitat quality module of the InVEST model was used to represent the impact score of industrial and mining activities. The formula is as follows:

I_a = exp [−(2.99/d_max)d] × 100

(8)

where I_a represents the impact score of the industrial and mining activity, d is the raster data value calculated by the Euclidean distance, and d_max is the maximum impact distance of industrial and mining activities, i.e., 1000 m.

3.2.4. Ecotourism Activities

Numerous monopolistic world-class landscape resources create a unique style of ecotourism in Shennongjia National Park. With the rapid development of tourism, anthropogenic activities are bound to disturb the natural ecosystem. The main ecotourist activity sites in the study area were extracted in ArcGIS 10.2, and the Euclidean distance between each raster and the ecotourism activity sites was calculated. The maximum distance of ecotourism activities from the natural ecosystem was set to 500 m [37], so that the distance of a raster greater than 500 m was assigned a value of 0. For a raster less than 500 m, the distance decay function in the habitat quality module of the InVEST model was used to represent the impact score of ecotourism activities. The formula is as follows:

E_t = exp [−(2.99/d_max)d] × 100

(9)

where E_t indicates the impact score of ecotourism activities, d is the raster data value calculated by the Euclidean distance, and d_max is the maximum impact distance of ecotourism activities, i.e., 500 m.

3.2.5. Traffic Accessibility

With the deepening of urbanization and the development of the tourism industry in the Shennongjia area, the road network and infrastructure construction will cause the division of the landscape to a certain extent. In this study, referring to the assignment method of Woolmer et al. [42] for roads, the roads in Shennongjia National Park were reclassified, and the accessibility assignments of the distances on both sides of different types of roads are shown in

Table 3. Assignment table of traffic accessibility.

Road Type	0–100 m	100–500 m	500–1000 m	1000–3000 m
National Highway	80	60	40	20
County Road	60	40	20	0
Township Road	40	20	10	0

. Three data layers for road type were established, each data layer is superimposed with equal weights, and the superimposed raster results are standardized using the min–max standardization method. The formula is as follows:

T_a = (d − d_min)/(d_max − d_min) × 100

(10)

where T_a represents the score of traffic accessibility after standardized processing, d is the value of the superimposed raster data, d_max is the maximum value in the superimposed raster data, and d_min is the minimum value in the superimposed raster data.

3.3. Analysis of the Impact of Anthropogenic Activities on Natural Ecosystem Integrity

3.3.1. Ordinary Least Squares Model

The ordinary least squares (OLS) model is a linear regression model commonly used to explore the relationship between explanatory and dependent variables [43,44]. In this study, the OLS model was used to establish regression equations using natural ecosystem integrity indicators as the dependent variable and five types of anthropogenic activities as the explanatory variables to quantify the degree of influence of anthropogenic activities on natural ecosystem integrity. The OLS was calculated as follows:

$$y_i=\beta +{{\displaystyle \sum }}_{k=1}^p{\beta }_k{x}_{ik}+{\epsilon }_i$$

(11)

where y_i is the natural ecosystem integrity index at position i in space; β is the squares spatial intercept; β_k is the regression coefficient of the kth anthropogenic activity indicator; x_ik is the value of the kth anthropogenic activity indicator at position i in space; and ε_i is the error correction term.

3.3.2. Geographically Weighted Regression Model

The geographically weighted regression (GWR) model is used to embed the geographical location of the sample data into the regression equation, which provides a visual representation of the degree and extent of influence of each explanatory variable on the dependent variable [45]. The formula is as follows:

$$y_i={\beta }_0\left(u_i,v_i\right)+{{\displaystyle \sum }}_{k=1}^p{\beta }_k\left(u_i,v_i\right)x_{ik}+{\epsilon }_i$$

(12)

where y_i is the natural ecosystem integrity index at position i in space; (ui, vi) is the spatial coordinate of position i; β₀ and β_k are the intercept and local regression coefficient of position i, respectively; x_ik is the value of the kth anthropogenic activity indicator at position i in space; and ε_i is the error correction term.

Before conducting the GWR, a spatial autocorrelation test needs to be conducted for each variable in the study area to determine whether the spatial autocorrelation of each variable can meet the necessary requirements for constructing a GWR model [46]. In this study, Moran’s I index was used to test the spatial correlation of the variables within the region. The formula is as follows:

$$Mora{n}^{\prime }s I=\frac{n}{S_0}\frac{{{\displaystyle \sum }}_{i=1}^n{{\displaystyle \sum }}_{j=1}^n{w}_{ij}\left(x_i-\overline{x}\right)\left(x_j-\overline{x}\right)}{{{\displaystyle \sum }}_{i=1}^n\left(x_i-\overline{x}\right)}$$

(13)

where n is the total number of spatial units in the national park; x_i and x_j represent the values of the variables on the ith and jth spatial units, respectively, w_ij is the spatial weight matrix between unit i and unit j; and S_o is the sum of all spatial weight matrices.

4. Results

4.1. Spatial Distribution Pattern of the Landscape

There are four main protected landscape types in Shennongjia National Park: woodland, grassland, water and wetland, and bare land. Additionally, there are two main interfered landscape types: cropland and artificial land [47] (Table 1).

Table 4. Statistics of landscape-type areas.

Landscape Type		Whole National Park		Strict Protection Zone
Primary Landscape Type	Secondary Landscape Type	Area (hm2)	Proportion (%)	Area (hm2)	Proportion (%)
Woodland	High-canopy-density woodland	111,087.95	93.55	60,071.09	94.23
	Low-canopy-density woodland	1065.83	0.90	982.19	1.54
	Shrubland	1474.36	1.24	901.41	1.41
	Artificial woodland	285.26	0.24	3.57	0.01
Grassland	High-coverage grassland	1677.81	1.41	905.17	1.42
	Medium-coverage grassland	86.53	0.07	45.96	0.07
	Low-coverage grassland	115.92	0.10	94.13	0.15
Water and Wetland	River wetland	636.07	0.54	98.53	0.15
	Lake wetland	1.90	0.00	1.72	0.00
	Artificial wetland	97.39	0.08	91.78	0.14
	Mud flat	71.19	0.06	10.22	0.02
Cropland	Dry land	663.83	0.56	16.04	0.03
	Greenhouse	8.08	0.01	0.12	0.00
Artificial land	Industrial land	322.41	0.27	8.35	0.01
	Transportation land	2.08	0.00	0.12	0.00
	Residential land	387.03	0.33	33.09	0.05
	Other artificial land	61.35	0.05	2.17	0.00
Bare land	Bare land	700.36	0.59	482.17	0.76

and Figure 3 show that the largest area of woodland in the study area is 113,913.39 hm², accounting for 96.07% of the total area, which is the landscape matrix of Shennongjia National Park, while the types of grassland, artificial land, and water are embedded in the form of patches or corridors. The area of the grassland is 1880.26 hm², accounting for 1.59% of the total area; it is mainly distributed in the higher altitude area of Dajiuhu Village in the western part and the central part of the national park. The area of water and wetland is 636.07 hm², accounting for 0.54% of the total area, consisting of the largest lake in the national park, Dajiuhu Lake, and three watersheds, namely the Du River, Xiangxi River, and Yandu River, which are largely distributed in the Jiuhu Township. The area of cropland is 671.91 hm², accounting for 0.57% of the total area; it is scattered in Dongxi Village in the northwestern part of the national park, the Jiuhu Township in the western part, the Xiaguping Township in the south-central part, Muyu Township in the central part, and Jiuchong Township in the east. The area of artificial land is 772.87 hm², accounting for 0.65% of the total area; it is mainly distributed in the core urban areas of each township. The area of bare land is 700.36 hm², accounting for 0.59% of the total area; it is mainly distributed in the higher altitude areas in the western strict protection zone in the national park. The total area of protected landscape in the strict protection zone is 99.9%, while the area of interfered-with landscape accounts for only 0.1%, indicating that the concentrated distribution of the main protection objects is relatively light in terms of anthropogenic activities, and that the regional ecology has been effectively protected from the perspective of the landscape pattern.

4.2. Spatial Distribution Pattern of Natural Ecosystem Integrity

In order to more clearly identify the spatial differences in the integrity of natural ecosystems in national parks, and to ensure the maximum difference in integrity between each level and the minimum difference in integrity within each level, the natural breakpoint method was selected for classification [48].

Using the reclassification tool in ArcGIS 10.2, the natural breakpoint method was used to classify the Ecosystem Integrity Index values into five levels: high (95–100), relatively high (90–95), medium (80–90), relatively low (70–80), and low (0–70) (

Table 5. Area and proportion of different integrity levels.

Level		Whole National Park	Strict Protection Zone
High	Area (hm2)	85,821.04	54,553.73
	Proportion (%)	72.38	85.66
Relatively High	Area (hm2)	13,081.05	4827.94
	Proportion (%)	11.03	7.58
Medium	Area (hm2)	15,378.09	3775.23
	Proportion (%)	12.97	5.93
Relatively Low	Area (hm2)	3994.8	528.13
	Proportion (%)	3.37	0.83
Low	Area (hm2)	288.02	4.22
	Proportion (%)	0.24	0.01
EII Value		96.06	98.83

).

From the perspective of the whole area, the Ecosystem Integrity Index value for the national parks was 96.06, indicating a high overall level of natural ecosystem integrity. Among the areas, the strict protection zone has an integrity index value of 98.83, and its high-integrity area covers 54,553.73 hm² (85.66%), indicating a low degree of fragmentation of the landscape in the strict protection zone. However, the degree of integrity in local areas is low, showing high integrity on the whole. The risk of ecological fragmentation is gradually emerging. Combined with the spatial distribution map of natural ecosystem integrity (Figure 4), it can be seen that these areas are mainly distributed in the main urban areas of townships and along highways within the national park, as well as in other areas with intensive anthropogenic activities.

4.3. Quantitative Analysis of the Impact of Anthropogenic Activities on the Natural Ecosystem Integrity

Degree of Impact of Anthropogenic Activity Indicators on Natural Ecosystem Integrity

The calculation results of the OLS model are shown in

Table 6. OLS model regression results.

Dependent Variable	Explanatory Variables	Regression Coefficient	t	p	VIF
EII	Population Density	−0.3344	−11.7910	0.000 *	1.225
	Traditional Utilization Activities	−0.1101	−19.1009	0.000 *	1.3016
	Industrial and Mining Activities	−0.0095	−2.3769	0.000 *	1.6171
	Ecotourism Activities	−0.0248	−2.0058	0.081	1.4935
	Traffic Accessibility	−0.2389	−33.3603	0.000 *	1.8092

OLS: ordinary least squares; EII: Ecosystem Integrity Index; VIF: variance inflation factor; * correlation is significant at the 0.01 level.

, where the variance inflation factors (VIF) of the 5 categories of anthropogenic activity indicators are all less than 10. It is generally agreed that, when the VIF is less than 10, there is no obvious problem of multicollinearity among the variables [49]. As for the significance test level, ecotourism activities did not pass the significance test at the 0.05 level, which indicates that neither the natural ecosystem integrity index nor ecotourism activities are statistically significant in the study area. The area characterized by ecotourism activities in Shennongjia National Park is relatively small, 63.43 hm², accounting for only 0.054% of the total area. Moreover, the ecotourism activities in Shennongjia National Park are mostly carried out in the form of ecotourism and scientific research, which do not rely on the construction of large-scale tourism infrastructure, so their impact on the integrity of the natural ecosystem is weak. Therefore, the explanatory ability of ecotourism activities for natural ecosystem integrity is not significant.

The regression coefficients in Table 6 show that EII is significantly and negatively correlated with population density, traditional utilization activities, industrial and mining activities, and traffic accessibility, and the absolute value of the regression coefficient between each anthropogenic activity indicator and population density is the largest at 0.3344. For every 1% increase in the influence value of population density, the natural ecosystem integrity index decreases by 33.44%; this indicates that population density has the most significant effect on natural ecosystem integrity in Shennongjia National Park. The other three indicators of anthropogenic activity (absolute values of the regression coefficients) are, in descending order, traffic accessibility (0.2389), traditional utilization activities (0.1101), and industrial and mining activities (0.0095).

As the OLS model calculations showed that the regression relationship between ecotourism activities and the Ecosystem Integrity Index was not significant, the anthropogenic activity indicator of ecotourism activity was removed to ensure the accuracy of the GWR calculation results. As seen in Figure 5, the scatter plots of the Moran’s I index for both the dependent variable (the Ecosystem Integrity Index) and the explanatory variables (population density, traditional utilization activity, industrial and mining activity, traffic accessibility) are distributed along the angular parallels of quadrants one and three, with more pronounced aggregation characteristics (high–high or low–low). The results pass the test at the highly significant level (p ≤ 0.01) (Z > 2.58), indicating that there is a strong spatial autocorrelation in the spatial distribution of each variable; therefore, the impact of each anthropogenic activity indicator on natural ecosystem integrity can be analyzed by further constructing a GWR model.

The spatial pattern of the correlation between the Ecosystem Integrity Index and each anthropogenic activity indicator was obtained using the natural breakpoint method to grade the GWR model regression coefficients (Figure 6). In general, the spatial pattern of the regression coefficients of the Ecosystem Integrity Index and each anthropogenic activity indicator showed a clustered pattern, mainly in the negative direction, indicating that each anthropogenic activity indicator had a strong negative correlation with the Ecosystem Integrity Index. The higher the impact of each type of anthropogenic activity, the lower the value of the Ecosystem Integrity Index.

Therefore, in this study, the low-value area of the regression coefficient of each indicator was taken as the high-impact area for analysis; the area of the high-impact area of the anthropogenic activity indicators (low-value regression coefficient area) is shown in

Table 7. Proportion of high-impact areas of anthropogenic activity indicators.

Anthropogenic Activity Indicators	Impact Area (hm2)	Proportion in the Whole Area (%)	Impact Area of Strict Protection Zone (hm2)	Proportion in Total Impact Area (%)
Population Density	1776.47	1.52	1424.99	80.21
Traditional Utilization Activities	23,063.14	19.71	12,489.28	54.15
Industrial and Mining Activities	5561.5	4.75	916.66	16.48
Traffic Accessibility	10,852.3	9.28	2976.28	27.43

.

As shown in Figure 6 and Table 7, the regression coefficient of the low-value area of traditional utilization activities is significantly lower than that of other anthropogenic activity indicators, with a regression coefficient ranging from −6.82 to −0.1, indicating that the impact degree of this index is significantly higher than that of other indexes. The impact area is 23,063.14 hm², accounting for 19.71% of the total area of Shennongjia National Park and 54.15% of the strict protection zone; it is mainly located in Yinyuhe Village, Dajiuhu Village, and Xiaodangyang Village.

The regression coefficient of the high-impact area of population density ranges from −5.08 to −2.62, and the impact area is 1776.47 hm², accounting for 1.52% of the total area of the national park; this area is mainly located in the strict protection zone south of Yinyuhe Village, 80.21% of which is located in the strict protection zone. The regression coefficient of the high impact area of industrial and mining activities ranges from −0.44 to −0.12, with an impact area of 5561.5 hm², accounting for 4.75% of the total area; this area is mainly located in Changling Village in the northeast and Muyu Village in the central part of the Shennongjia National Park, with 16.48% of the area located in the strict protection zone.

The regression coefficient of the high-impact area of traffic accessibility ranges from −0.71 to −0.45, with an impact area of 10,852.3 hm², accounting for 9.28% of the total area. This area is mainly located in Dongxi Village in the northwest and Heping Village in the south-central part of the National Park, and 27.43% of the area falls within the strict protection zone.

5. Discussion

5.1. Quantitative Analysis of the Integrity of Natural Ecosystems

The results of this study show that the natural ecosystem integrity index of Shennongjia National Park is 96.06. In recent years, Chinese researchers have evaluated the integrity of natural ecosystems in several national parks by constructing an integrity evaluation index system using a scoring method. For example, the natural ecosystem integrity evaluation of Qianjiangyuan National Park scored 52.83, which takes into account the natural conservation geographical units [12]. Due to the small area of Qianjiangyuan National Park, it is not possible to cover the complete geographical unit of nature conservation, resulting in its low overall score. The evaluation score for the natural ecosystem integrity of the Northeast Tiger and Leopard National Park is 81. The method used in that evaluation was based on qualitative indicators of natural ecosystem integrity constructed from ecosystem structure, processes, functions, and patterns, and it is evaluated using an expert scoring method [14]. The EII proposed in this study is a comprehensive index that takes into account the fragmentation status of protected patches and the strength of edge effects, focusing on evaluating the integrity of protected landscapes within the delineated national park. It constitutes a quick, accurate, and operational evaluation method. In addition, by constructing grid cells at different scales and using geospatial interpolation, this study determined the differences in the spatial distribution of natural ecosystem integrity.

Overall, the area of protected landscape in Shennongjia National Park accounts for 98.2% of the total area. This is due to the importance attached to ecological protection in the Shennongjia area at the national level and the effective implementation of park protection and management over the years. In recent years, various ecological protection projects, including the Natural Forest Protection Project, the ‘Returning Farmland to Forests’ Project, and the ‘Three Green’ Projects, have been implemented to effectively protect and restore the ecological environment in most areas. However, the integrity of the natural ecosystem is low in some local areas of Shennongjia National Park, fragmentation is still significant, the interference of anthropogenic activities is growing stronger, and the impact range is larger.

5.2. Quantitative Analysis of the Impact of Anthropogenic Activities on the Integrity of Natural Ecosystems

In terms of the impact of population density on natural ecosystem integrity, there are still many indigenous inhabitants in Shennongjia National Park; they come from 25 villages in 5 townships, with a total population of 21,072. The expansion of these artificial patches has changed the land-use types and led to the fragmentation of the landscape in the national park. The results of this study show that population density is the most important factor influencing the Ecosystem Integrity Index (where the absolute value of the regression coefficient is 0.3344).

In terms of the impact of traditional utilization activities on natural ecosystem integrity, the traditional utilization activities of the indigenous residents, which mainly comprise agricultural cultivation, are scattered throughout the national park and cover a large area; they have a significant impact on the integrity of natural ecosystems, covering an area of 23,063.14 hm². Taking farming as an example, on the one hand, chemical fertilizers, pesticides, and other chemical agents applied during the cultivation process may cause soil consolidation and the deterioration of physical and chemical properties; meanwhile, when harmful substances flow into lakes, rivers, and other bodies of water through various channels, they may cause the nutrient enrichment of water bodies and the pollution of groundwater. On the other hand, the cultivation of cropland divides the woodland matrix, reduces the degree of aggregation and connectivity of natural landscape patches, and leads to local fragmentation. The fragmentation of the woodland has reduced the cohesiveness and connectivity of natural landscape patches, resulting in increased local fragmentation and reducing the natural ecosystem integrity. All of these factors pose direct or potential threats to the main conservation objects in the national park, and a transformation of the green industry development model is urgently needed.

In terms of the impact of industrial and mining activities on natural ecosystem integrity, industrial and mining activities still exist in the national park. The existing enterprises employ outdated mining methods and low utilization of mineral resources, and the mining activities and the abandoned land left behind have caused serious ecological damage to the natural ecosystem. The results of this study show that the area of the national park whose natural ecosystem integrity is significantly impacted by industrial and mining activities is 5561.5 hm²; although most of the high-impact areas are located outside the strict protection zone, the over-exploitation of industrial and mining land still poses a significant threat to the main conservation objects.

In terms of the impact of traffic accessibility on natural ecosystem integrity, the Shennongjia National Park has a complex road network, with National Highway G208 running through the middle of the park from south to north, as well as provincial and rural roads of different types totaling 369.5 km, connecting more than 20 villages in the national park. The natural ecosystem integrity of an area of 10,852.3 hm² is characterized as being significantly impacted. To a certain extent, these roads divide the habitats of the main conservation objects, causing local fragmentation of the landscape, resulting in ecological gaps in the originally continuous ecological network, and posing a serious threat to biological migration within the national park.

5.3. Suggested Strategies for Enhancing the Integrity of Natural Ecosystems

The Shennongjia National Park still faces major challenges in maintaining the integrity of the natural ecosystem. While emphasizing the maintenance of the integrity of the natural ecosystem, it is also necessary to continuously explore new models for the sustainable development of natural protected areas. Through ecological restoration methods such as constructing ecological corridors and the near-naturalization of damaged habitats, implementing control measures on anthropogenic activities, such as logging, hunting, and fishing bans, advocating industrial development models such as replacing traditional industries with green industries and creating ecological products, and developing educational channels shared by all people such as ecotourism, study and education, and forest recreation can play a vital role in maintaining the integrity and sustainability of the natural ecosystem of national parks.

5.4. Innovations

At present, most evaluation methods for natural ecosystem integrity are based on the construction of an evaluation indicator system, which can be too large and involve complicated data collection and quantitative calculation; moreover, some of the indicators rely on a highly subjective scoring method, making them difficult for managers to use. Some of the indicators need to be monitored over a long period of time in practice, making it difficult to obtain data and to apply the evaluation indicators. The existing evaluation methods focus on the overall evaluation of a certain protected area or conservation area and cannot reflect the differences in natural ecosystem integrity at the spatial level.

This study attempts to construct a method for calculating natural ecosystem integrity from the perspective of natural ecosystems’ structure and function, which is highly representative and has wide coverage and easy access to data. The method of using the natural Ecosystem Integrity Index to evaluate the integrity of natural ecosystems is proposed, solving the problem of the quantitative evaluation of natural ecosystem integrity at the spatial level. By discussing the distribution patterns and differentiation characteristics, the strict protection zone of national parks can be better identified, and the key control areas and requirements can be determined, which can be easily understood and used by managers.

5.5. Limitations and Perspectives

This study has the following limitations: firstly, the optimal scale of 500 m × 500 m, set according to the landscape distribution pattern of the study area, is not necessarily suitable for all study scales or study areas. We believe that the scale setting is not only related to the spatial resolution of the landscape interpretation results, but also has a certain relationship with the fractal dimension of the landscape, and the threshold value of the division scale setting needs further analysis. In addition, the anthropogenic activity indicator system constructed in this study is based on the quantification of land-use types, and the indicator system is relatively rough; for example, traditional utilization activities generally include planting activities, grazing activities, and fishing activities, etc., and industrial and mining activities generally include mineral extraction and water development and so on. Future studies need to further refine the types of anthropogenic activities under consideration and combine them with town or village statistical yearbooks, with the aim of providing a more in-depth analysis of the types and extent of anthropogenic activities that affect the integrity of natural ecosystems in national parks, in order to provide clearer guidance for the conservation and management of national parks.

6. Conclusions

The degree of natural ecosystem integrity in Shennongjia National Park is generally high, and the concentrated distribution areas of the main conservation objects are effectively protected. Meanwhile, the integrity of local areas is low, mainly in the main urban areas of the townships and along the roads in the national park, as well as in other areas characterized by intensive anthropogenic activities. From highest to lowest, the degree of impact of the following anthropogenic activities on natural ecosystem integrity is population density (0.3344), traffic accessibility (0.2389), traditional utilization activities (0.1101), and industrial and mining activities (0.0095). The correlation between ecotourism activities and natural ecosystem integrity is not significant; meanwhile, in terms of the impact range of anthropogenic activities on natural ecosystem integrity, traditional utilization activities have the largest area of impact, followed by the traffic accessibility. Meanwhile, population density causes a higher degree of impact at the small scale, with the risk of fragmentation gradually emerging in local areas.

This study evaluates the integrity of the natural ecosystem of Shennongjia National Park and its response to anthropogenic activities, providing a theoretical basis and data support for management agencies to carry out ecological restoration and the control of anthropogenic activities. We provide a new research paradigm for the conservation of natural ecosystems and for the sustainable development of resources in natural protected areas, which is important for the scientific construction of the future national park system.