Ecological space management and control zoning of Giant Panda National Park from the perspective of ecosystem services and land use

Abstract

Since China proposed building a national park system in 2017, the establishment of a planning system for nature reserves, with national parks as the main body, is being actively promoted around the country. Among them, scientific ecological space management and control zoning (ESMCZ) is an important link in maintaining the ecological stability of national parks. How to zone national parks and how to improve the precision of zoning has become a new task for national parks. Therefore, this study takes the Giant Panda National Park as the study area, takes ecosystem services and land use/cover change as the research perspective, integrates the InVEST model, PLUS model and bayes belief network (BBN) model, and builds a set of ecological space management and control zoning (ESMCZ) spatial zoning framework based on raster scale, dividing the study area into strictly protected zone, ecological buffer zone, ecological control zone and controlled development zone. The results showed that: (1) The study area showed an increasing trend in water conservation, soil conservation and carbon storage from 2005 to 2020, and the habitat quality index was generally high. The spatial heterogeneity of ecosystem services in the study area was significant, and the effect of a single factor on ecosystem services was most pronounced. (2) Large variation in area for different land uses under natural development scenarios and ecological protection scenarios. In both scenarios, the area of cultivated land, the area of grassland and the area of unused land decrease relative to 2020, and the area of forested land, the area of water and the area of constructed land increase relative to 2020. (3) The Giant Panda National Park is divided into strictly protected zone, ecological buffer zone, ecological control zone and control development zone, of which the strictly protected zone have the largest area and the best ecosystem background condition, and the control development zone have the smallest area and the worst ecosystem background condition. (4) The ecological space management and control zoning (ESMCZ) framework provides a more refined method for the secondary zoning of nature reserves such as the Giant Panda National Park, which is valuable for the implementation of zoning and categorization management for ecological conservation in the Giant Panda National Park.

Introduction

Nature Conservation International defines a national park as an area of large size and proximity to nature whose function is to maintain the integrity of the species and natural environment within the area while providing educational, scientific, cultural, and recreational places adapted to its environment¹. In October 2020, China formally established five first national parks, including the Giant Panda National Park, and explicitly proposed to improve the main functional area system and optimize the spatial development pattern of the national territory². As the first batch of national parks established with a single species as the core, the Giant Panda National Park is of great significance in reasonably constructing and optimizing the ecological spatial pattern for its protection and restoration of biodiversity, maintaining the integrity of ecosystem structures and processes, and safeguarding the ecological functions of regional systems³.

Ecosystem services are the various types of benefits that humans derive directly or indirectly from ecosystems⁴, which play an important role in shaping and maintaining the environmental conditions and the material basis for human survival and development, and are commonly used in the spatial optimization of a wide range of ecosystem services and the exploration of their spatial differentiation⁵. Land use is the process by which human beings obtain material products and services by combining with land, and it is commonly used in regional land use planning and spatial zoning, and our common zoning is mainly based on the spatial division of land use, and the combination of ecosystem services and land use can identify the dominant ecosystem services of different types of land use, and then carry out functional zoning based on its functional structure and ecological characteristics⁶. In order to ensure that the ecological functions of the Giant Panda National Park are fully realized, ecological zoning is the key to balancing conservation and development in the area⁷. Among them, ecological space management and control zoning (ESMCZ) is an important measure to optimize the spatial structure of national parks. It is also the basis for formulating differentiated spatial control policies⁸. Ecological space management and control zoning adjusts and optimizes space as an object, aiming to coordinate, balance and distribute the spatial layout from various aspects such as quantity, structure, location and time, etc. It is to optimize the spatial pattern of ecosystem services and land use for spatial zoning based on clarifying the current spatial distribution pattern and problems to achieve reasonable ecological space planning⁹.

With the continuous development of eco-spatial zoning and corresponding measures, how to effectively reduce the negative impact of eco-spatial zoning on future ecosystem services and land use planning to enhance the ecological adaptability and foresight of zoning schemes remains one of the important challenges scholars face. It is mainly in two aspects: on the one hand, how to scientifically delineate ecological control zones and optimize the ecological effects of different zones based on the existence of various combinations of different ecosystem services while proposing differentiated ecological management measures; on the other hand, how to face the future spatial evolution of land use trends and predict the impact of land use patterns on ecological space management and control zoning under different scenarios from a future perspective.

In response to the above challenges, UNESCO Man and Biosphere Reserve adopt a single zoning method to divide the reserve into core area, buffer area and transition area. The advantage of this zoning is that its zoning boundaries are easy to identify¹⁰. Some countries and organizations have also proposed a composite zoning model. The International Union for Conservation of Nature summarized the zoning patterns of national parks into protected natural areas¹¹, covered human districts¹², and eight types of zoning in three categories, such as protected historical or archaeological zones. The development intensity of each type of zoning is specified. The composite zoning approach is more refined, but the zoning identification and management are difficult. Both single and composite zoning ideas provide theoretical references for the zoning of ecological protection elements in China. Based on advanced foreign experiences, some domestic scholars, to further explore a new set of ecological spatial control zoning schemes, used the spatial superposition method and cluster analysis method, to incorporate ecological red lines into ecological space management and control zoning, the research scale from national to local, and the research method^13,14,15. The scale of research ranges from national to local, and the research methods range from semi-quantitative to big data^17,18. National parks have social, economic, and ecological functions. Therefore, the multi-functionality of the area should be fully considered when zoning and the maximization and sustainable use of composite functions should be pursued⁸. In the zoning process, national parks have focused more on conservation and utilization, and the zoning model has gradually evolved from a two-zone and three-ring model to a variety of models such as four-zone, five-zone, and ORRRC, and often simplified or increased zoning according to their own needs¹⁹. With the development of biophysical models and the application of high-resolution datasets, spatial zoning of regional ecological functions based on the spatial divergence patterns of ecosystem services and land use is of great practical importance to enhance human well-being²⁰. At the present stage, ecological space management and control zoning mainly includes ecological function zoning²¹, ecological pattern zoning, and ecological control boundary delineation, usually using GIS, RS, and other technical means, with ecological sensitivity, the importance of ecosystem services, biodiversity, and landscape pattern as the direction²², the zoning focuses on the protection and utilization of ecological elements carrier space.

In general, the study of Ecological space management and control zoning has gradually emphasized the integrity of ecosystems, but there are still some limitations. China’s national parks have implemented the “two-zone system”, i.e., a two-level zoning system of control zones and functional zones, but only two types of zoning ignore the spatial heterogeneity, and lack a clear boundary of community life and production, and the zoning of national parks should not be limited to control zones only. The zoning of national parks should not be limited to control zones, and the second level of functional zoning should be added to realize the fine management of the parks. In terms of the selection of ecosystem service indicators, most national park studies lacked the selection of typical ecosystem services, as well as the consideration of correlation factors for each ecosystem service. In terms of zoning methods, traditional zoning mostly adopts qualitative methods such as expert integration, which is more subjective and influenced by a priori knowledge, and it is not easy to form unified zoning standards and technical specifications; most of the existing studies are at the macro level, ignoring the key factors of ecosystem services and the role of land use, and there is a lack of scientific dynamic scenario simulation and benefit assessment as to whether future ecological spatial control can be practically improved in the study area. How to rationalize and optimize the allocation of factors such as natural factors and land use within the study area and how to quantify the degree of influence of key variable factors on zoning to achieve ecological space management and control zoning are precisely the problems to be solved in this study, while there are few systematic frameworks for national park zoning in existing national park zoning studies.

This study follows the main line of “ building the ecological space management and control zoning (ESMCZ) framework”-“spatial and temporal variation of ecosystem services”-“land use scenario simulation”-“ Ecological space management and control zoning (ESMCZ)”. It combines the bayes belief network (BBN) for ecological spatial management and control zoning. The main objectives of this study are as follows: (1) Developing a set of ideas and solutions for spatial zoning of national parks based on ecosystem services and land use perspective. (2) Dividing spatial areas conducive to hierarchical management and zonal protection. (3) Provide suggestions and measures for the zoning management and protection of the Giant Panda National Park. Firstly, this ecological space management and control zoning (ESMCZ) framework introduces land use uncertainty based on ecosystem service assessment, which improves the accuracy and visualization of zoning. Secondly, the framework uses conditional probability to fully identify key ecosystem service factors and quantify zoning indicators from a more micro perspective, while incorporating different scenarios of future, which improves the objectivity of zoning for fine management of parks. Finally, the framework follows national park planning policies and development goals, it has reference value for other national parks and nature reserves in terms of zonal protection and hierarchical management. In conclusion, the main purpose of this study is to construct a zoning framework for national parks, which is able to assess the ecosystem service status of the region, and also zoning from the perspectives of ecosystem service and land use, which can provide a reference for the protection and management of national parks and other nature reserves. To our knowledge, this is the first known attempt to construct a zoning framework in a national park study.

Methods

Study area

The Giant Panda National Park is located in the western region of China and consists of the Minshan Area in Sichuan Province, the Qionglai Mountain-Daxiangling Area in Sichuan Province, the Qinling Area in Shaanxi Province and the Baishuijiang Area in Gansu Province, with a planned area of 27,134 km². The geographical coordinates lie between 102°11′10"-108°30′52"E and 28°51′03"-34°10′07"N. The overall topography is high in the north-west and low in the south-east, with a maximum altitude of 5588 m. 72.07% of the area is covered by forest, with 3446 species of seed plants in 197 families and 1007 genera, and 641 species of vertebrate animals²³. Giant Panda National Park is the key area of the "two screens and three belts" of the national ecological security strategy pattern, and is the last "shelter" to protect the wild pandas. Although it only accounts for 0.3% of the national land area, it is the most important resource reserve in China and maintains the national ecological balance²³. As shown in Fig. 1.

Figure 1

Location map of Giant Panda National Park. (The map was generated by ArcGIS 10.8 http://www.esri.com/sofware/arcgis and does not require any permission from anywhere).

Data

Data sources

The data involved in this paper mainly include DEM (Digital Elevation Model) data, land use data, rainfall data, normalized difference vegetation index (NDVI), and evapotranspiration (ETO) data, soil data, and basic geographic data (administrative district boundary data, traffic data). All data are selected from four years: 2005, 2010, 2015, and 2020. The specific data sources are shown in Table 1.

Table 1 The data set used for the study area.

Data processing

In this paper, all the data are in the same projected coordinate system, the raster pixel size is the same, and the raster resampling resolution is 30 m. In this paper, after multidimensional analysis, cropping, interpolation, computation, and resampling of the rainfall data based on ArcGIS, the raster data of the average annual rainfall in the study area for the period of 2005–2020 are obtained. In this paper, the Modified-Hargreaves method was calculated using the ArcGIS processing to obtain the annual mean potential evapotranspiration raster data of the study area. In this paper, the soil data of the study area were obtained by mask extraction after calculation by connecting the EXCEL table with the ArcGIS attribute table using the Harmonized World Soil Database (HSWD). The remaining data were mosaicked using ArcGIS.

Building ESMCZ framework and zoning

Based on the results of ecosystem service assessment and land use simulation, this paper uses the BBN model to build the ecological space management and control zoning (ESMCZ) framework. The ecological space management and control zoning (ESMCZ) framework diagram is shown in Fig. 2. The main ideas of the framework are as follows:

• (1) To assess the spatio-temporal changes of the four typical ecosystem services in the study area from 2005 to 2020, using 2020 as the time baseline, the bayes belief network (BBN) was used to calculate the variance reduction (VR) value of each ecosystem service node, and the key factors leading to changes in ecosystem services were identified as the key variables of the zoning, and land use was identified as the common key variable of the four typical ecosystem services.

• (2) Using the conditional probabilities derived in the Bayes belief network (BBN) model, different ecosystem service states corresponding to the key variables with the largest conditional probability as the key state of the variables. When each ecosystem service node is in the low, medium, high, and highest levels, respectively, the set of key states corresponding to the factor nodes is determined—the “subset of key states”.

• (3) The intersection of ecosystem services and key state subsets (KSS) under different land use scenarios in 2020 and 2035 was taken as the intersection of the raster layers of elements that were satisfied simultaneously. The intersection of the raster layers of elements that satisfy the key status subsets in different land use scenarios in 2020 and 2035 was taken, and the subsets of ecosystem services and key statuses were superimposed according to the same status. The subsets of ecosystem services and key states were superimposed by the same state to get the comprehensive key state subsets (CKSS) of ecosystem services under different land use scenarios in 2020 and 2035.

• (4) According to the ecological space management and control zoning (ESMCZ) based on the ecosystem service level and the state corresponding to the comprehensive key state subsets (CKSS), the Giant Panda National Park is divided into strictly protected zone, ecological buffer zone, ecological control zone and control development zone.

Figure 2

ESMCZ framework.

Ecosystem service assessment

Based on the actual situation of the Giant Panda National Park, this article selects four ecosystem services: water conservation (WC), habitat quality (HQ), carbon storage (CS), and soil conservation (SC), and uses the InVEST (Integrated Valuation of Ecosystem Services and Tradeoffs) model to assess the spatiotemporal changes of various ecosystem services in 2000, 2005, 2010, and 2020²⁴.

Water conservation refers to the ability of ecosystems to retain, infiltrate and store precipitation in a given time and space. The capacity of water conservation is estimated from the “Water Yield” module of the InVEST model and is calculated as follows²⁵:

$$Y_{x} = \left( {1 - \frac{{AET_{x} }}{{P_{x} }}} \right)P_{x }$$

(1)

where $Y_{x}$ is the annual water yield depth, $AET_{x}$ is the actual evapotranspiration, $P_{x }$ is the annual rainfall and $\frac{{AET_{x} }}{{P_{x} }}$ is the evapotranspiration as a percentage of the water balance, calculated as follows:

$$\frac{{AET_{x} }}{{P_{\left( x \right)} }} = 1 + \frac{{PET_{x} }}{{P_{\left( x \right)} }} - \left[ {1 + \left( {\frac{{AET_{x} }}{{P_{\left( x \right)} }}} \right)^{{W_{\left( x \right)} }} } \right]^{{\frac{1}{{W_{\left( x \right)} }}}}$$

(2)

$$PET_{x} = K_{{C\left( {lx} \right)}} \times ET_{O\left( x \right)}$$

(3)

$$\omega_{\left( x \right)} = Z\frac{{AWC_{\left( x \right)} }}{{P_{\left( x \right)} }} + 1.25{ }$$

(4)

where $PET_{x}$ is the potential evapotranspiration, $\omega_{\left( x \right) }$ is the soil characteristic. $ET_{O\left( x \right)}$ is the reference evapotranspiration, $K_{{C\left( {lx} \right)}}$ is the vegetation evapotranspiration coefficient corresponding to the vegetation type ($lx$), Z is a constant characterizing seasonal factors, and $AWC_{\left( x \right)}$ is the effective moisture content of the vegetation.

Habitat quality refers to the availability of subsistence resources, and high habitat quality can be seen as being structurally and functionally intact and within historical dynamics. The index of habitat quality is estimated from the “Habitat Quality” module of the InVEST model and is calculated as follows⁶:

$$Q_{xj} = H_{j} \left[ {1 - \left( {\frac{{D_{xj}^{2} }}{{D_{xj}^{2} + K^{z} }}} \right)} \right]$$

(5)

where $Q_{xj}$ is the habitat quality index; $H_{j}$ is habitat fitness; $D_{xj}$ is the habitat degradation index.

Carbon storage usually refers to the amount of carbon stored in a carbon pool (forest, ocean, land, etc.), which is estimated based on the storage of four types of substances: aboveground biomass, underground biomass, soil, and dead organic matter, to determine the current amount of carbon stored in the landscape or the amount of carbon sequestered over time. The capacity of carbon storage is estimated from the “Carbon” module of the InVEST model and is calculated as follows²⁶:

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

(6)

where $C$ is the total carbon storage (t/hm2); $C_{above}$ is the carbon storage of aboveground vegetation; $C_{below}$ is the underground biological carbon storage; $C_{dead}$ is the carbon storage of dead organic matter; $C_{soil}$ is soil carbon storage.

Soil conservation services refer to the ability of ecosystems to regulate erosion to prevent soil loss and to store and retain sediment. The capacity of soil conservation is estimated from the “SDR” module of the InVEST model and is calculated as follows²⁷:

$$T = R \times K \times LS \times S \times \left( {1 - C \times P} \right)$$

(7)

where $T$ is the soil conservation amount (t/hm²), $R$ is the rainfall erosivity, $K$ is the soil erodibility, $LS$ is the slope length factor, $C$ is the vegetation management factor, and $P$ is the soil and water conservation measure factor.

Geographical detectors can either test the spatial heterogeneity of a single variable or detect possible causal relationships between two variables by examining the coupling of their spatial distribution. In this paper, spatial and temporal variation in ecosystem services in Giant Panda National Park was analyzed using a geographic detector, selecting rainfall, evapotranspiration, population density, land use, elevation, slope, and normalized difference vegetation index as influencing factors. The calculation formula is as follows:

$$q = 1 - \frac{{\mathop \sum \nolimits_{h = 1}^{L} N_{h} \sigma_{h}^{2} }}{{N\sigma^{2} }}$$

(8)

where $L$ is the stratification variable; $N$ is the number of cells; ${\upsigma }_{{\text{h}}}^{2}$ and ${\upsigma }^{2}$ are the variances of the regional $Y$ values.

Land use scenario simulation

Land use transfer matrix

The land use transfer matrix indicates the change process between land use types over time²⁸. The calculation formula is as follows:

$$S_{{{\text{ij}}}} = \left( {\begin{array}{*{20}c} {S_{11} } & \cdots & {S_{{1{\text{m}}}} } \\ \vdots & \ddots & \vdots \\ {S_{{{\text{m}}1}} } & \cdots & {S_{{{\text{mm}}}} } \\ \end{array} } \right)$$

(9)

where $S_{ij}$ is the converted area of land use type i, j and i, j (I, j = 1,2,3,…m) is the land use type at the beginning and end of each time period; m is the number of land classes.

Land use scenario simulation

In this paper, we use land use data from 2005, 2010, 2015, and 2020, and simulate the land use demand in 2035 using the Markov model based on the analysis of land use changes in different periods, using 2020 as the time baseline, and simulating the spatial pattern of land use in 2035 using the PLUS(patch-generating land use simulation) model on this basis²⁹.

This study sets up the natural development scenarios (NDS) and the ecological protection scenarios (EPS) to simulate the future spatial pattern of land use in the region more comprehensively. The natural development scenarios (NDS) means that the region is not affected by natural disasters and national policies³⁰, and based on the expansion characteristics and change patterns of different land use types in Giant Panda National Park during the historical process, without considering the influence of policy regulation, the natural and socio-economic development indicators are mainly considered and change with time changes. The ecological protection scenarios (EPS) refers to responding to ecological protection policies, following the law of natural development to allocate resources rationally, and prohibiting large-scale development and utilization³¹.

ESMCZ framework

Zoning indicator selection

In this study, slope, soil erodibility(K), soil type, population density (PD), Evapotranspiration (ETO), lucc, rainfall, normalized difference vegetation index (NDVI), rainfall erosion force (R), and topographical relief (TR) are selected as zoning sub-nodes, four ecosystem services as parent nodes. And the data of each node was discretized using ArcGIS, and the natural breakpoint method was used to classify each node into four levels: highest, high, medium and low, and exported as training data, as shown in Supplementary Table S1.

Building a bayes belief network (BBN) of ecosystem services

The nodes in the bayes belief network (BBN) represent discrete variables, and these nodes can be represented by a conditional probability table (CPT), which can reflect the relationship between individual nodes³². The probability distribution of each node is calculated as follows:

$$P\left( {X_{1} ,X_{2} , \ldots ,X_{n} } \right) = \mathop \prod \limits_{i = 1}^{n} P\left( {X_{i} {|}parents\left( {X_{i} } \right)} \right)$$

(10)

where, $\left( {X_{1} ,X_{2} , \ldots ,X_{n} } \right)$ are random variables, this paper uses Netica(Produced by NORSYS software corp) to construct a bayes belief network(BBN) for parameter learning, the conditional probability calculation formula is shown below³³:

$$P\left( {Y{|}X} \right) = \frac{{P\left( {XY} \right)}}{P\left( Y \right)}$$

(11)

where: $P\left( {Y{|}X} \right)$ is the joint probability of event Y and X; $P\left( Y \right)$ is the prior probability of Y; $P\left( X \right)$ is the probability of event X occurring.

Identifying key variables

Sensitivity analysis (SA) in bayes belief network (BBN) uses a quantitative analysis method to compare the influence degree of each indicator or a group of indicators by changing the value of relevant variables one by one³⁴. This paper uses variance reduction formula to determine the sensitivity of nodes:

$$VR = V\left( {ES} \right) - V\left( {ES{|}I} \right) = \mathop \sum \limits_{s} p\left( s \right) \times \left( {s - E\left[ {ES} \right]} \right)^{2} - \mathop \sum \limits_{s} p\left( {s|I} \right) \times \left( {s - E\left[ {ES|I} \right]} \right)^{2}$$

(12)

where $VR$ is the variance reduction value, $V\left( {ES} \right)$ is the variance value of the ecosystem service node; $V\left( {ES{|}I} \right)$ is the variance value of the ecosystem service node in the state of the influencing factor node I; s is the state of the output variable, $E\left[ {ES} \right]$ is the expected value; and $E\left[ {ES|I} \right]$ is the expected value in the state of the known influencing factor node I. This study selects the top three factors in variance reduction and impairment as key variables.

Identifying a subset of key states

When the conditional probability value of different levels of the parent node is maximum, the level of the corresponding key variable is the key state at this time, and different key variables constitute a subset of key states.

Ecological space management and control zoning

This study uses ArcGIS software to spatially overlay different ecosystem services with their corresponding key state subsets (KSS) to obtain a comprehensive key state subset (CKSS) of different ecosystem services. The integrated key state subsets (KSS) for the two scenarios of 2020 and 2035 are overlaid again and partitioned by the level of the integrated key state subsets (KSS), divided into a strictly protected zone, ecological buffer zone, ecological control zone, and control development zone. As shown in Fig. 3.

Figure 3

Diagram of the zoning process.

Results

Assessment results of ES

Spatial and temporal variation in ES

Carbon storage

Regarding spatial distribution, the overall carbon storage in Giant Panda National Park shows the spatial distribution characteristics of high in the north and south and low in the middle. Regarding time, the variation of carbon storage in Giant Panda National Park from 2005 to 2020 was not obvious, as shown in Fig. 4. So the carbon storage in different years was calculated by using ArcGIS, the carbon storage in Giant Panda National Park in 2005, 2010, 2015, 2020 were 78.28 × 10⁷t, 78.77 × 10⁷t, 79.03 × 10⁷t, 79.45 × 10⁷t, with an overall increase of 1.17 × 10⁷t from 2005 to 2020, including an increase of 0.49 × 10⁷t from 2005 to 2010, 0.26 × 10⁷t from 2010 to 2015, and 0.42 × 10⁷t from 2015 to 2020, as shown in Table 2. This is mainly due to the increasing area of forest and grassland in the region, increasing forest cover, and increased carbon storage services.

Figure 4

Spatial distribution of carbon storage from 2005 to 2020. (The map was generated by ArcGIS 10.8 http://www.esri.com/sofware/arcgis and does not require any permission from anywhere).

Table 2 Changes in the quality of ES from 2005–2020.

Habitat quality

Regarding spatial distribution, the habitat quality of the Giant Panda National Park showed a spatial distribution pattern of low in the middle and high around, with little spatial interannual variation, as shown in Fig. 5. The average habitat quality indices of different years were calculated using ArcGIS, as shown in Table 2, the average habitat quality indices of giant panda national park in 2005, 2010, 2015 and 2020 were 0.89, 0.90, 0.90 and 0.84, respectively, and the average habitat quality index was generally high, but decreased in 2020. Human activities and disturbance of the various site types are the root causes of changes in habitat quality.

Figure 5

Spatial distribution of habitat quality from 2005 to 2020. (The map was generated by ArcGIS 10.8 http://www.esri.com/sofware/arcgis and does not require any permission from anywhere).

Water conservation

Regarding spatial distribution, the water conservation content of Giant Panda National Park spatially showed the spatial distribution characteristics of high in the west and low in the east. Regarding time, the spatial variation of water content in Giant Panda National Park from 2005 to 2020 is not apparent, as shown in Fig. 6. So the water conservation content of different years was calculated by using ArcGIS, as shown in Table 2, the water conservation content of Giant Panda National Park in 2005, 2010, 2015 and 2020 were 52.157 billion cubic meters (bcm), 62.093 bcm, 57.315 billion cubic meters and 69.617 bcm, with an overall increase of 17.46 bcm from 2005 to 2020. Among them, 9.936 bcm increased from 2005 to 2010, 4.778 bcm decreased from 2010 to 2015, precipitation fluctuations are the main cause of water conservation phase changes, and 12.302 bcm increased from 2015 to 2020.

Figure 6

Spatial distribution of water conservation from 2005 to 2020. (The map was generated by ArcGIS 10.8 http://www.esri.com/sofware/arcgis and does not require any permission from anywhere).

Soil conservation

In terms of spatial distribution, the soil conservation high value area of Giant Panda National Park mainly appeared in the center of the study area. Regarding time, the spatial variation of soil conservation content in Giant Panda National Park from 2005 to 2020 is not obvious, as shown in Fig. 7. So the soil conservation content of different years was calculated by using ArcGIS, as shown in Table 2. The overall trend of soil conservation in Giant Panda National Park from 2005 to 2020 is increasing, with a total increase of 6.22 × 10⁹ t. The soil conservation in Giant Panda National Park in 2005, 2010, 2015, and 2020 is 37.49 × 10⁹ t, 40.82 × 109 t, 41.75 × 10⁹ t, and 43.71 × 10⁹ t, respectively, of which 3.33 × 10⁹ t from 2005 to 2010, 0.93 × 10⁹ t from 2010 to 2015, and 1.96 × 10⁹ t from 2015 to 2020. The small increase in soil conservation from 2010 to 2015 over 2005–2010 is primarily related to the rate of increase in forest cover, which increased faster in 2010–2015 than in 2005–2010.

Figure 7

Spatial distribution of soil conservation from 2005 to 2020. (The map was generated by ArcGIS 10.8 http://www.esri.com/sofware/arcgis and does not require any permission from anywhere).

Analysis of ecosystem services influence factors

Determinants of the spatial heterogeneity of ecosystem services q values for each influence factor for 2005–2020 were obtained by geographic detector. Lucc was the main factor influencing the spatial heterogeneity of carbon storage from 2005 to 2020. The q values of lucc from 2005 to 2020 were 0.98, 0.99, 0.98, and 0.97, respectively. Lucc was the main factor influencing the spatial heterogeneity of habitat quality from 2005 to 2020. The q values of lucc from 2005 to 2020 were 0.48, 0.42, 0.42, and 0.44, respectively. Evapotranspiration (ETO) was the main factor influencing the spatial heterogeneity of water conservation from 2005 to 2020. The q values of Evapotranspiration (ETO) from 2005 to 2020 were 0.68, 0.66, 0.65, and 0.64, respectively. Rainfall was the main factor influencing the spatial heterogeneity of soil conservation from 2005 to 2020. The q values of Evapotranspiration (ETO) from 2005 to 2020 were 0.47, 0.51, 0.55, and 0.60, respectively. As shown in Fig. 8.

Figure 8

Determinants of force q for different ecosystem service influences. (The map was generated by Microsoft Office PowerPoint and does not require any permission from anywhere).

Land use scenario simulation

Analysis of land use change

From 2005 to 2010, cultivated land and grassland in the Giant Panda National Park decreased by 57.3 km² and 477.58 km², respectively, while forest land, water area, construction land and unused land increased by 441.26 km², 20.14 km², 50.64 km² and 22.84 km², respectively. From 2010 to 2015, cultivated land, forest land, grassland and unused land decreased by 9.58 km², 7.51 km², 7.32 km² and 0.36 km², respectively, while water area and construction land increased by 10.34 km² and 14.43 km², respectively. From 2015 to 2020, cultivated land, forest land, construction land and unused land decreased by 1.6 km², 50.25 km2, 0.01 km² and 1.98 km², respectively, while grassland and water area increased by 53.44 km² and 0.4 km², respectively. The trend in land-use type change is mainly from cultivated land to forest land and grassland. This is mainly because after the national park area has been demarcated, the area is mainly for the protection of giant panda habitats, and it is necessary to return cultivated land to forest land and grassland in order to restore the local ecological functions, so part of the cultivated land will be converted into forest land and grassland. As shown in Fig. 9.

Figure 9

Different land use change maps.

Land use scenario simulation results

Figure 10 and Table 3 show that the land use structure of Giant Panda National Park is dominated by forest land and grassland. Comparing the area of land use types in 2035 with that in 2020, under the natural development scenarios (NDS), the area of cultivated land decreases by 5.58 km², the area of forest land increases by 9.23 km², the area of grassland decreases by 30.21 km², the area of water increases by 0.91 km², the area of construction land increases by 0.69 km², and the area of unused land decreases by 0.14 km². Under the ecological protection scenarios (EPS), the cultivated land area decreased by 5.58 km², the forest land area increased by 326.06 km², the grassland area decreased by 335.89 km², the water area increased by 0.91 km², the construction land area decreased by 10.47 km², and the unused land area decreased by 0.14 km².

Figure 10

Spatial distribution of land use under different scenarios. (The map was generated by ArcGIS 10.8 http://www.esri.com/sofware/arcgis and does not require any permission from anywhere).

Table 3 Map of land use change under different scenarios (km²).

Ecological space management and control zoning results

Conditional probability values for each node

The nodes extracted after the discretization process were fishnetted to get 38,595 sample data, and the “Incorp Case File” tool of Netica was used for parameter learning to get the bayes belief network (BBN) of Giant Panda National Park in 2020. The arrow pointing between the nodes represents the causal link, the node led by the arrow line is the cause, and the node introduced by the arrow line is the effect, the results are shown in Fig. 11.

Figure 11

Ecosystem services BBN.

In this study, the discrete data of land use under the natural development scenarios (NDS) and ecological protection scenarios (EPS) were input into the BBN-based ecosystem services model, and the probability values of habitat quality (HQ), water conservation (WC), soil conservation SC), and carbon storage (CS) under the natural development scenarios (NDS) and ecological protection scenarios (EPS) of Giant Panda National Park in 2035 were obtained, as shown in Table 4.

Table 4 Probability of distribution of ecosystem service classes under different scenarios.

Identifying key variables

In this study, the top three influence factors for each ecosystem services variance reduction value were selected as key variables, as shown in Table 5, where the key variables for habitat quality were lucc, rainfall erosion force (R), and population density (PD), the key variables for water conservation were evapotranspiration (ETO), lucc, and rainfall, the key variables for soil conservation were lucc, normalized difference vegetation index (NDVI), and rainfall erosion force (R), and the key variables for carbon storage were lucc, population density (PD), and normalized difference vegetation index (NDVI).

Table 5 Ecosystem services key variable sensitivity.

Spatial distribution of key state subsets

In this study, Netica software was used to obtain the key variable key state subsets (KSS) corresponding to different levels of each ecosystem services in Giant Panda National Park, as shown in Table 6.

Table 6 Ecosystem services key state subset.

Based on the identified key variables and key state subsets (KSS), the raster layers of the key variables and the key state subsets (KSS) were intersected using the Con function of ArcGIS software raster calculator. The spatial distribution of key state subsets (KSS) under the natural development scenarios (NDS) and ecological protection scenarios (EPS) in 2020 and 2035 was obtained for Giant Panda National Park. As shown in Fig. 12.

Figure 12

Subset of key states of ecosystem services under different scenarios. (The map was generated by ArcGIS 10.8 http://www.esri.com/sofware/arcgis and does not require any permission from anywhere).

The key state subsets I (KSS I) of carbon storage is the set of high probability states when the carbon storage is low, corresponding to the lucc is unused land, high population density (PD), and low normalized difference vegetation index (NDVI), and the area is distributed in strips in each area. The key state subsets II (KSS II) is the set of high probability states when the carbon storage is medium, which corresponds to the lucc is cultivated land, and the population density (PD) and the normalized difference vegetation index (NDVI) are high. The key state subsets III (KSS III) is the set of high probability states when the carbon storage is high, which corresponds to the lucc is grassland, low population density (PD) and high normalized difference vegetation index (NDVI). The key state subsets IV (KSS IV) is the set of high probability states when carbon storage is the highest, corresponding to the lucc is forest land, with low population density (PD) and the highest normalized difference vegetation index (NDVI), and the region has a wide spatial distribution and the largest area.

The key state subsets I (KSS I) of habitat quality is the set of high probability states when the habitat quality is low, corresponding to the lucc is forest land, rainfall erosion force(R) is medium and the population density (PD) is high. The key state subsets II (KSS II) is the set of high probability states when the habitat quality is medium, corresponding to the lucc is forest land, R is low and the population density (PD) is low. The key state subsets III (KSS III) is the set of high probability states when the habitat quality is high, corresponding to the lucc is grassland, rainfall erosion force(R) is medium and population density (PD) is low. The key state subsets IV (KSS IV) is the set of high probability states when the habitat quality is the highest, corresponding to the lucc is forest land, rainfall erosion force (R) is medium and population density (PD) is low, the ecological condition of this area is better, less affected by human activities, and the spatial distribution area is wider.

The key state subsets I (KSS I) of water conservation is the set of high probability states when the water conservation is low, which corresponds to the area where the lucc is forest land, the Evapotranspiration (ETO) is high and the rainfall is low. This area is mainly distributed in the Qinling area and Baishuijiang area, while other areas are scattered. Key state subsets II (KSS II) is the set of high probability states when the water conservation is medium, which corresponds to the lucc is grassland and the area with medium Evapotranspiration (ETO) and rainfall. The spatial distribution of this area is fragmented and has no obvious spatial characteristics. The key state subsets III (KSS III) is the set of high probability states when water conservation is high, corresponding to the lucc is grassland, low Evapotranspiration (ETO) and medium rainfall, which is mainly distributed in the northern part of Qinling Mountains, the western part of Minshan Area and the central part of Qionglai Mountain-Daxiangling Area, and also distributed in the edge of subset IV. The key state subsets IV (KSS IV) is the set of high probability states when water conservation is the highest, corresponding to the lucc is grassland, with low Evapotranspiration (ETO) and high rainfall, which is mainly distributed in the northern part of Minshan Area and Qionglai Mountain-Daxiangling Area, with more concentrated distribution.

The key state subsets I (KSS I) of soil conservation is the set of high probability states when soil conservation is low, corresponding to the lucc is cultivated land, and the normalized difference vegetation index (NDVI) and rainfall erosion force (R) are the highest. This area is mainly located in the Qinling Mountains and the southern part of the Minshan Area, with a relatively large area. The key state subsets II (KSS II) is the set of high probability states when soil conservation is medium, corresponding to the lucc is cultivated land, the highest normalized difference vegetation index (NDVI) and high rainfall erosion force (R), and the distribution of this area is more scattered, and the area is most widely distributed in the Minshan area and Qionglai Mountain-Daxiangling area. The key state subsets III (KSS III) is the set of high probability states when the soil conservation is high, corresponding to the lucc is forest land, with the highest normalized difference vegetation index (NDVI) and medium rainfall erosion force(R). The key state subsets IV (KSS IV) is the set of high probability states when soil conservation is the highest, corresponding to the lucc is forest land, with the highest normalized difference vegetation index (NDVI) and low rainfall erosion force (R).

Ecological space management and control zoning

This paper uses the ecological space management and control zoning (ESMCZ) framework to spatially overlay different key state subsets of ecosystem services in Giant Panda National Park under two scenarios of natural development and ecological protection in 2020 and 2035, and obtains the comprehensive key state subsets (CKSS) of ecosystem services in Giant Panda National Park under different scenarios in 2020 and 2035, as shown in Fig. 13, and overlaying the comprehensive key state subsets (CKSS) for different scenarios. Meanwhile, with reference to the General Plan of Giant Panda National Park and the Interim Measures for National Park Management, the Giant Panda National Park is divided into strictly protected zone, ecological buffer zone, ecological control zone and control development zone, as shown in Fig. 14.

Figure 13

Comprehensive key state subset spatial distribution. (The map was generated by ArcGIS 10.8 http://www.esri.com/sofware/arcgis and does not require any permission from anywhere).

Figure 14

Ecological spatial management and control zoning of Giant Panda National Park. (The map was generated by ArcGIS 10.8 http://www.esri.com/sofware/arcgis and does not require any permission from anywhere).

Strictly protected zone

The strictly protected zone has a good ecological and natural background condition and a large potential demand, mainly concentrated in the southern part of Qinling Area and Baishuijiang Area, the western and northern part of Minshan Area, and the north and south ends of Qionglai Mountain-Daxiangling Area, which is dominated by forest land and grassland, a natural ecological barrier and the area with the highest ecological value. The strictly protected zone is 19,370.9km², accounting for 50% of the total area of the study area, with the widest distribution area.

Ecological buffer zone

The ecological buffer zone has a medium ecological background and low potential demand, mainly concentrated in the southern and central part of the Qinling Area, the northern part of the Baishuijiang Area, and the central part of the Qionglai Mountain-Daxiangling Area, which is dominated by grassland and forest land. The control objective is to improve the productivity and conservation of ecosystem services based on the optimization of ecosystem services in the area. The ecological buffer zone covers an area of 13,247.52km², accounting for 34% of the total area of the study area.

Ecological control zone

The ecological control zone is mainly forest land and cultivated land, with a small portion of land for construction, distributed primarily on the northern part of the Qinling Area, the southern fringe of the Baishuijiang Area, the eastern part of the Qionglai Mountain-Daxiangling Area, and the southeastern part of the Minshan Area. Most of the existing developed areas of the Giant Panda National Park sightseeing spots are in the ecological control zone, which covers an area of 5646.86 km², accounting for 15% of the total area of the study area.

Control development zone

The control development zone is mainly cultivated land and construction land. The ecological environment carrying capacity is high, and this area is mainly set up to meet the needs of ecotourism services. The spatial distribution of this area is relatively scattered, concentrated primarily on the intersection zone between the ecological control zone and other areas. Reasonable development can be carried out without damaging the original ecological environment. The controlled development zone is 352 km², accounting for 1% of the total area of the research area.

Discussion

Feasibility of ESMCZ framework

Zoning is the basic way of geographic synthesis and the basis and prerequisite for cognitive geographic laws. It is a basic requirement for the effectiveness of zoning and zoning that zoning elements vary as much as possible between regions and as little as possible within regions⁸. Ecosystem services and land use have wide applications in ecological environment assessment and ecological function zoning³⁵, this study constructs the ecological space management and control zoning (ESMCZ) framework from the perspective of ecosystem services and land use using bayes belief network (BBN).

First, the framework clarifies current ecological conditions and land use patterns and reflects future conservation priorities. In this process, we added scenario simulations considering the current conservation policies and future planning goals of Giant Panda National Park²³, while the area is rarely affected by urbanization and industry. We only consider population density in terms of human activities, while adding ecosystem services as the third category of indicators, which enriches the existing zoning indicator system and also provides a set of feasible systematic solutions for subsequent zoning studies.

Secondly, the ecological space management and control zoning (ESMCZ) framework mainly utilizes the probabilistic inference mechanism of bayes belief network (BBN), which is used to reduce subjectivity in the partitioning process. From the existing studies, the application of bayes belief network (BBN) in ecosystem services is becoming more and more widespread, and the study mainly used bayes belief network (BBN) for trade-offs and optimal allocation of regional ecosystem services predicted ecosystem service capacity through bayes belief network (BBN) scenario analysis, and the results showed that key decisions can optimize a variety of ecosystem services^16,36,37,38. And this study continues to explore zoning based on ecosystem service optimization to achieve the best optimization in spatial regions.

Unlike most previous zoning studies, the core of this framework is to further identify the key variables of zoning by quantifying the degree of influence of key factors on the ecological spatial control zoning of national parks and also to determine the subset of key states of zoning by using the CPT values of nodes. This zoning method can make full use of the degree of influence of index data on zoning and identify areas with common ecological conditions based on the spatial location relationship of different categories, which can effectively avoid the defects such as strong subjectivity and lack of precision in the zoning process of qualitative methods. It can be an important supplement to quantitative zoning methods. To study the feasibility of the framework, we brought each zoning index into each node for error calculation, in which the correct rate of habitat quality was 88.16%. The correct rate of water conservation is 90.21%; the correct rate of soil conservation is 88.34%; the correct rate of carbon storage is 89.01%, which proves that the framework has some scientific reliability and can be used for zoning.

Finally, based on the ecological space management and control zoning (ESMCZ) framework, the ecological space management and control zoning (ESMCZ) of Giant Panda National Park not only shows the spatial distribution characteristics of ecosystem services and land use in different periods of Giant Panda National Park but also combines the natural geographical features with ecological processes and disturbances of human activities, reflecting the dynamic interaction characteristics of ecological environment in the contradictory relationship between human and land. The ecological space management and control zoning (ESMCZ) framework integrates multi-source data and multiple modeling methods to reveal the spatial differences of each zoning area from the perspective of ecosystem services and land use, determine the dominant service functions of each zoning area, and clarify the ecological conservation protection areas and directions of different areas, which is of great significance to realize the refined management of land resources in regional Giant Panda National Park. Other researchers can utilize the process of the entire framework by first selecting ecosystem service types and indicators that fit the study area, identifying the key variables of the area according to the process of the zoning framework, and then zoning the area, and finally the zoning results can be compared to the dichotomous zoning results of the national parks that have been issued by the state.

Generalizability of zoning results

First, the strict protection zone, ecological buffer zone, ecological control zone, and control development zone proposed in this study are quadratic divisions concerning the core protection zone and general control zone in the scheme of the Interim Measures for National Park Management, in which the core protection zone is divided into strict protection zone and ecological buffer zone, and the general control zone is divided into ecological control zone and control development zone⁹. The results of ecological spatial management and control zoning reflect the overall condition of the ecosystem of Giant Panda National Park and have a greater fit with the regional functional positioning of Giant Panda National Park. Secondly, from the perspective of the need for national parks for conservation and development, any type of national park needs to put ecological protection in the first place⁹. In this study, two types of areas were selected in the process of determining strict protection zones, namely the highest ecosystem service level and the integrated KSS IV, to cover the needs for ecological protection as much as possible, and their establishment is more dependent on the natural landscape and biological categories they contain, so it is necessary to establish separate strict protection zones to highlight ecological protection³⁹. The ecological buffer zone has an excellent natural background condition and a high potential demand, including some transition areas for converting cultivated land to forest land. The ecological control zone is a zoning feature of national parks, which is set up mainly for establishing ecological corridors. Finally, national parks are closest to human beings, who have a strong demand for natural resource utilization, and therefore need to set up control development zones that can be appropriately utilized³⁷.

This study discusses the relationship between ecological conservation, human disturbance, and resource use according to the possible activities in national parks, mainly to scientifically delineate different adaptive areas to implement more fine-grained differentiated control measures so that each area can be used most appropriately and effectively⁸. When the utilization of ecological control zones and control development zones is enhanced, the impact on strictly protected zones and ecological buffer zones will be reduced, which has a certain effect on national parks' balanced protection and utilization. We found that a part of the existing national park zoning studies mainly study the adaptive areas of species, and these studies are more focused on species protection^9,37, but most of the national parks are still aiming at maintaining the natural ecology^40,41, and at the same time, more attention should be paid to the natural ecological environment of the region itself on the premise of protecting the species' habitats. Compared with the results of these studies, this paper mainly focuses on the evaluation of the overall ecological status of the region, and the selection of four typical ecosystem services not only reflects the ecosystem status of the region, but also is similar to the typical ecosystem services of other national parks^42,43, which is applicable to a certain extent.

Setting up strictly protected zones and ecological buffer zones in national parks is the same as the conservation objectives of nature reserves and the primary choice for setting up protected areas. Still, because of the relatively large number of species protected in national parks and the sufficient richness of ecosystem types, their protection will be strong⁴⁴. The national park is large and focuses on protecting the integrity of the ecosystem. In terms of the zoning that currently exists, the establishment of more areas can be an aspect to explore, and for ecological control zones and controlled development zones, although other nature reserves also need utilization, their utilization is relatively less intense, while the national park has a greater need for utilization⁴⁵. This zoning clarifies different ecological protection areas and directions, providing a realistic basis and spatial reference for the current and future conservation and management of Giant Panda National Park. The zoning proposed in this study is reasonable and has a certain degree of general applicability. In this study, due to the use of data with different resolutions throughout the study, the resultant map of the zoning has a part of the more fragmented area, and there may be limitations in the view of the whole map, and at the same time, our study period is from 2005 to 2020, and we have given less consideration to the condition of the study area before 2005, and the condition of the region over a long period of time should be explored as much as possible in the future study.

The relevance of the research findings for management and conservation

From the correlation between ecosystem services and ecological spatial control zoning, ecosystem service assessment is the basis and prerequisite for ecological space management and control zoning (ESMCZ)⁴⁶. We compared the results of our analysis with those of existing studies on the assessment of ecosystem services in Giant Panda National Park, which showed a high degree of similarity in spatial distribution characteristics. In analyzing the drivers of ecosystem services, this study considers the geospatial differences in the study area. It can clarify the influence of drivers on ecosystem services in different regions¹⁶, which can better provide a decision-making basis for constructing ecological space management and control zoning (ESMCZ) in Giant Panda National Park and lay a better foundation for subsequent studies.

From the correlation between land use and ecological spatial control zoning, the National Park Spatial Layout Program proposes that by 2035, the construction of the spatial layout of national parks will be completed, and the world's largest national park system will be built¹⁷. To make our zoning reflect the objectives of this program, land use scenario simulation is critical, and we can simulate the spatial pattern of land use under different policy-oriented and control purposes by combining the current actual situation in the study area⁴⁵. Land use scenario simulation provides both spatial reference and policy support in spatial terms, so it is necessary to set two scenarios of ecological protection and natural development in this study. In both scenarios, forest land and grassland with higher ecological value were targeted for transfer⁴⁵. Under the natural development scenarios (NDS), the change of each land use type is consistent with 2005–2020, and we can obtain the land use change without policy and human intervention. Under the ecological protection scenarios (EPS), the change of each land use type is more different from 2005 to 2020 because we set this scenario mainly influenced by the policy and use forest land as the restricted conversion type. The ecological protection effect is more obvious. This paper simulates the land use pattern under two scenarios of giant panda national park in 2035, which can clarify the land use change characteristics in the region firsthand and provide a basis for zoning on the other hand.

The purpose of functional zoning of national parks is to manage and protect resources in space effectively. Because national park spatial zoning maintains and restores the integrity of ecosystem services through land use, it is necessary to conduct ecological spatial control zoning in local areas and specific periods from the perspective of ecosystem services and land use. The zoning of the National Park facilitates the effective management and protection of the ecological resources and spaces of the National Park.

Suggestions for countermeasures

• (1) For strictly protected zone, the ecological protection mechanism of Giant Panda National Park should be improved, the ecological protection red line should be strictly delineated, development activities should be strictly prohibited from entering strictly protected areas, the objectives and requirements of the management of the national park should be strictly observed, and construction facilities that damage the ecological environment, etc. should not be built except for necessary field surveys and environmental monitoring.

• (2) For the ecological buffer zone, which is the second largest area in the study area, the area should limit regional development, follow local management objectives, conduct appropriate research and ecotourism activities without damaging the ecological environment, rehabilitate areas that have been damaged, and focus on enhancing the ecological environment to prevent ecological damage.

• (3) For the ecological control area, the focus should be on protecting the natural landscape of the Giant Panda National Park, and the focus of ecological tourism development in the area can be gradually shifted to the area by strictly following appropriate development and construction policies, appropriately building a giant panda conservation and learning base and related tourism support facilities, moderately using vacant land, and strengthening forest construction while establishing the national park.

• (4) For the control development zone, the area can be used as a priority selection area for the development of scenic spots in Giant Panda National Park, where larger development can be carried out without destroying the ecological environment, making reasonable use of the unused land in the area to carry out agrotourism, planning and building hiking trails and other ecological infrastructure, and promoting the development of ecological tourism so that the development of scenic spots can be coordinated with ecological, environmental protection and the sustainable.

Conclusion

Based on the perspective of ecosystem services and land use, this study analyzed the spatial and temporal variation of ecosystem services in Giant Panda National Park and the influencing factors, simulated the spatial pattern of land use under different scenarios, constructed the ecological space management and control zoning (ESMCZ) framework, and conducted the spatial zoning, and the main conclusions are as follows:

• (1) Strictly protected zones are mainly located in the southern part of the Qinling Area and the Baishuijiang Area, the western and northern part of the Minshan Area, the north and south ends of the Qionglai Mountain-Daxiangling Area, covering an area of 19,370.9 km². Ecological buffer zones are mainly located in the southern and central part of the Qinling Area, the northern part of the Baishuijiang Area, and the central part of the Qionglai Mountain-Daxiangling Area, covering an area of 13,247.52 km². Ecological control zones are mainly distributed in the northern part of the Qinling area, the southern fringe of the Baishuijiang area, the eastern part of the Qionglai-Daxiangling area, and the southeastern part of the Minshan area. Control development zones are mainly located at the interface between the ecological control areas and other areas, covering an area of 352 km².

• (2) This study assessed the different ecosystem services in the study area, based on which a zoning framework was constructed, different zones were divided, and corresponding measures were proposed for the different zones to achieve the research objectives of this study. The ecological space management and control zoning (ESMCZ) framework couples multiple models and methods with high partitioning accuracy, which is highly feasible and practical. The zoning results can also provide a basis for zonal ecological protection and hierarchical management of the Giant Panda National Park.

• (3) This study not only reflects the ecosystem status of the region, but also is similar to typical ecosystem services in other national parks, which is applicable to a certain extent. Based on the framework construction process and results, we found that the final results are highly consistent with existing research and national park zoning schemes. Every step of the framework construction has undergone accuracy checks, and it also has high applicability and accuracy for new datasets, so it is possible to use the framework in other national parks. When applying this framework to different national parks, specific consideration should be given to the future development direction of the region, and the selection of variables should be more in line with the actual situation.