Introduction

The rapid economic development of countries around the world since the 1950s contributed to resource shortages and ecological deterioration. In 1972, Meadows pointed out in his report on the plight of mankind that the world would face a catastrophe if the pattern of economic growth did not change1. In the same year, representatives of 113 countries and more than 400 non-governmental organisations adopted the Declaration on the Human Environment at the United Nations Conference on the Human Environment in Sweden, which for the first time proposed sustainable development2. In 1983, responded to the theory of sustainable development by introducing the concept of ecotourism to address dilemmas in Latin America. He posited that ecotourism could generate economic benefits while fostering environmental stewardship through education3.

The development of eco-tourism is now widely recognized and used in national park protection and utilization4. The concept of ecotourism in national parks originated from the zoning practice in national parks, which includes the functional division of core protection areas, recreation buffer zones, and intensive recreation areas, which is regarded as the initial stage of ecotourism5. Strategies for effective establishment of ecological function zoning are a requirement for improving the main function area and realising the construction of beautiful China.

The mechanism of ecotourism needs to be further explored6. China is learning from the experience of the United States7, Canada8, Australia9, Japan10 and other countries11. Lessons to date are that relevant laws and regulations should be introduced in the process of national park construction in China, ecological tourism infrastructure should be improved, and the relocation should be carried out carefully. Zhang studied the eco-tourism model of national parks in China and laid a solid theoretical foundation for further development of eco-tourism in national parks12. Important cases of ecotourism research in China 's national parks include the Sanjiangyuan, Wuyishan, Qilian Mountains, and Shennongjia13. The suitability of ecotourism activities14, the carrying capacity15, stakeholders16, and the impact of ecotourism on wildlife17 have become frontier hotspots in the field of national park research.

Suitability evaluation is a method and process for evaluating and classifying the most suitable spatial layout for a specific application scenario based on a variety of factors such as demand, preference, or prediction18. Ecotourism suitability evaluation belongs to the category of land use suitability evaluation19. Scholars have mainly analysed and evaluated the climate suitability of ecotourism20, the comfort of the tourism environment21, and the suitability of land use22, forming a system of quantitative methodology based on the evaluation of natural environmental resources. Most of the studies have used multi-criteria hierarchical decision analysis (AHP) to spatially classify the suitability of ecotourism in specific regions23,24.

AHP is based on the process of weighted evaluation of two-by-two comparisons of each parameter, resulting in an optimal solution to a given problem. This method is widely used in the context of land use suitability evaluation, especially when combined with GIS and remote sensing25. The combination of GIS and AHP provides a reliable, efficient, and accurate methodology that is applicable to different regions of the globe26. Nonetheless, the AHP method relies on expert opinion to determine indicator weight values, which is subjective and cognitively limited27. The inclusion of the Promethee method in this study based on GIS and AHP can ensure the objectivity of the results, reduce the influence of subjectivity, and improve the reliability of suitability assessment. The suitability of ecotourism activities requires the integration of different data, including hydrological data such as topographic features extracted from DEMs and geomorphological information28. These important datasets are usually obtained through remote sensing techniques and subsequently analysed using Geographic Information Systems (GIS) to construct comprehensive ecological mapping29. The construction of indicators for evaluating the suitability of ecotourism from the perspective of sustainable development requires the consideration of environmental constraints such as soil erosion, geological hazards, and other indicators30, 31.

The Qilian Mountains National Park has played a crucial role in the implementation of China’s "One Belt, One Road" policy and the construction of an ecological civilisation32. The park maintains the health of the ecological security barrier in western China, and provides valuable experience in the restoration and protection of the ecological environment33. In recent years, with the rise in popularity of tourism in Gansu and Qinghai provinces, some areas within the Qilian Mountains National Park have become popular tourist destinations; however, there is a lack of consensus on the recreational use of the national park at the local level, restricting tourism and other development activities within the park and contradicting tourism development and ecological protection.

We aimed to establish a spatial layout of tourism activities for the ecology of Qilian Mountains National Park. The main objective of the study was to assess and analyse the suitability of the area for ecotourism by integrating various data sources including topographic maps, hydrological data, geomorphological information, and geohazards using Geographic Information System (GIS) techniques and a multi-criteria decision-making methodology. Specifically, this study was designed for the following research purposes:

  1. (1)

    to construct an evaluation index system for the suitability of ecotourism activities;

  2. (2)

    divide the landscape units in Qilian Mountain National Park by combining the topography and vegetation distribution types;

  3. (3)

    sort the suitability of ecotourism activities using the Promethee method combined with ArcGIS software.

The results provide important insights for policy makers and national park authorities for effective ecotourism development and implementation. Additionally, the proposed spatial layout of ecotourism provides a reference for ecotourism development in other national parks.

Materials and methods

Study area

Qilian Mountain National Park is located in the northeastern Qinghai-Tibet Plateau, at the junction of Gansu and Qinghai provinces, with a total area of 5.02 × 104 km2, Gansu area of 3.44 × 104 km2, accounting for 68.5% of the total area of the national park. The administrative division includes 11 counties (districts) in five prefecture-level cities of Jiuquan, Zhangye, Wuwei, Jinchang, and Lanzhou (Fig. 1). Forests, grasslands, deserts and wetlands are distributed in the area; there are numerous glaciers, rivers, and abundant water resources, mainly in the basins of the Heihe, Shule, and Shiyang Rivers, and Sugang Lake.

Fig. 1
figure 1

Study area. The base map was obtained from the Standard Map Service System of the Ministry of Natural Resources of China, Map Review Number: GS(2022)1873. The maps were generated using ArcGIS software version 10.3 (https://www.arcgis.com/index.html).

Full size image

Fourteen species are listed in the study area as national first-class protected wildlife, 39 species as national second-class protected wildlife, and 135 species of animals and plants have important economic and scientific research value. The main tourist destinations in the Qilian Mountains National Park Gansu Province Area are Tianzhu Three Gorges National Forest Park, Horseshoe Temple Forest Park, Binggou River Forest Park, Binggou River Ecological and Cultural Tourism Scenic Spot, and Qilian Heiheyuan National Wetland Park. The surrounding main tourist areas are Sunan Danxia Landscape Scenic Spot, China Yugu Style Corridor Scenic Spot, Bars Snow Mountain Scenic Spot, Changgou Temple Scenic Spot, Hangtang Temple Danxia and Xiazhitala Scenic Spot, Qilian Heihe Canyon, and Bayi Glacier.

Collection and processing of data

The steps of data collection in this study are divided into five main parts:

  1. (1)

    Initial stage involving data collection;

  2. (2)

    Identification of ecotourism activities, construction of an evaluation index system for ecotourism activities and determination of weights;

  3. (3)

    Delineation of landscape units using Geographic Information System (GIS) for data processing to obtain a zoning map of influencing factors;

  4. (4)

    Ranking the suitability of ecotourism activities using the Promethee method combined with GIS to generate ecological cartography;

  5. (5)

    Spatial layout of ecotourism activities.

The evaluation process of ecotourism suitability is comprehensive, and the evaluation process must account many aspects such as geographic environment and human activities. The data required for the research mainly include DEM data, vegetation distribution data, NDVI data, soil erosion severity, geological hazard sites, and animal activity ranges. The elevation data was obtained from the Geospatial Data Cloud platform, with a spatial resolution of 30 m × 30 m. The sources of other data are listed in Table 1, and the resolutions used are suitable for both the study area and the research objectives31,34.

Table 1 Data sources.
Full size table

Ecotourism activity selection

Ecotourism activity projects were divided into five categories, according to the standard of "China Ecological Experience Base Classification and Recognition": ecological scientific research, natural wonders immersion experience, natural adventure experience, outdoor recreation and sports, and aboriginal life and culture experience. Ecotourism activities are usually classified based on resources or ecosystem types35, and ecological resources such as geomorphology, hydrology, biology, and forest substrate of the study area; finally, project and activities were combined with the current status quo of tourism development, and ecotourism activities in the Gansu area of the Qilian Mountains National Park were divided into three categories and nine subcategories, namely, educational examination, sport adventure, and leisure tourism36 (Table 2).

Table 2 Main types of ecotourism activities.
Full size table

Promethee

The Promethee is a type of multi-criteria decision-making system37. We used an objective assignment followed by the preference order structure assessment method for the suitability evaluation of ecotourism activities; the core steps were based on the two-by-two comparison of the evaluation schemes, using the preference difference method to measure and organize the suitability ranking levels of the individual evaluation indicators, and then summing up these partial scores into global scores to obtain a partially pre-ranked level (Promethee I) or a fully pre-ranked level (Promethee II)38.The final evaluation criterion of Promethee II is determined by the generated net flow values in the range of [-1, 1], with larger values being valued more highly in the rankings. The Promethee methodology for the metrics and rankings is based on Brans’ proposals of six different types of preference functions and criteria, including the usual, U, V, level, linear, linear, U-shaped, and the V-shaped criterion; criterion, level criterion, linear criterion and Gaussian criteria, which cover most attribute decision problems. Considering that the latter five preference functions need to be configured with many different types of preference parameters and that it is usually difficult for decision makers to make a clear decision, we chose the preference function corresponding to the most basic usual criterion for an in-depth analysis in this study. Let Bx and By be the specific evaluation programmes in the set of evaluation programmes \(B\left( {B_{x} ,B_{y} \in B} \right)\), and j be the specific evaluation index of each evaluation programme (j = 1,2,⋯,n), then the specific steps of the Promethee evaluation method are as follows39:

Calculation of the difference in evaluation programme values based on the two-by-two comparison method Dj(Bx,By):

$${\text{D}}_{\text{j}}\left({\text{B}}_{\text{x}}{,\text{B}}_{\text{y}}\right)={\text{D}}_{\text{j}}\left({\text{B}}_{\text{x}}\right){-\text{D}}_{\text{j}}\left({\text{B}}_{\text{y}}\right)$$
(1)

where: Dj(Bx) and Dj(By) are the values of the evaluation programmes Bx and By at indicator j, respectively; Dj(Bx,By) is the difference between the two.

B. Calculation of the preference of evaluation programmes based on the difference in their indicator values Pj(Bx,By):

$${\text{P}}_{\text{j}}\left({\text{B}}_{\text{x}}{,\text{B}}_{\text{y}}\right)={\text{F}}_{\text{j}}\left[{\text{D}}_{\text{j}}\left({\text{B}}_{\text{x}}{,\text{B}}_{\text{y}}\right)\right]$$
(2)

where: Pj(Bx,By) is the degree of preference of the evaluation programme Bx and By with respect to the evaluation indicator j; Fj is the preference function of PrometheeII, capable of transforming the range of values of Dj(Bx,By) between 0 and 1.

C. Calculation of an overall preference index based on indicator weights and preferences ∏(Bx,By) and ∏(By,Bx):

$$\prod \left({\text{B}}_{\text{x}}{,\text{B}}_{\text{y}}\right)={\sum }_{\text{j}=1}^{\text{n}}{\text{W}}_{\text{j}}{\text{P}}_{\text{j}}\left({\text{B}}_{\text{x}}{,\text{B}}_{\text{y}}\right)$$
(3)
$$\prod \left({\text{B}}_{\text{y}}{,\text{B}}_{\text{x}}\right)={\sum }_{\text{j}=1}^{\text{n}}{\text{W}}_{\text{j}}{\text{P}}_{\text{j}}\left({\text{B}}_{\text{y}}{,\text{B}}_{\text{x}}\right)$$
(4)

where: ∏(Bx,By) and ∏(By,Bx) denote the preference of an evaluating programme x over y and the preference of an evaluating programme y over x when all evaluation indicators are considered; Wj is the weight of the evaluation indicators.

D. The positive flow rate φ+(Bx) and the negative flow rate φ-(Bx) are calculated as follows:

$$\varphi^{ + } \left( {{\text{B}}_{{\text{x}}} } \right) = \frac{1}{{{\text{n}} - 1}}{ }\mathop \sum \limits_{{{\text{x}} \in {\text{B}}}} \prod \left( {{\text{B}}_{{\text{x}}} ,{\text{B}}_{{\text{y}}} } \right)$$
(5)
$$\varphi^{ - } \left( {{\text{B}}_{{\text{x}}} } \right) = \frac{1}{{{\text{n}} - 1}}\mathop \sum \limits_{{{\text{x}} \in {\text{B}}}} \prod \left( {{\text{B}}_{{\text{y}}} ,{\text{B}}_{{\text{x}}} } \right)$$
(6)

The criteria for decision-making are as follows:

when: φ+(Bx) ≥ φ+(By) and φ-(Bx) ≤ φ-(By), the evaluation scheme x is preferred (equiprobable) to y;the evaluation scheme x, etc. is better than y if and only if φ+(Bx) = φ-(Bx) and φ+(By) = φ-(By);the rest of the situation cannot be determined.

E. The calculation of net flows φ(Bx) is as follows:

$$\varphi \left( {{\text{B}}_{{\text{x}}} } \right) = \varphi^{ + } \left( {{\text{B}}_{{\text{x}}} } \right) - \varphi^{ - } \left( {{\text{B}}_{{\text{x}}} } \right)$$
(7)

The net flow is the Promethee II measure, where φ(Bx) is as large as possible.

Evaluation of the suitability of ecotourism activities

Division of landscape units

An evaluation unit is the smallest landscape unit in the evaluation of tourism suitability for which information is obtained, processed, and displayed. As the basic object of ecotourism activity suitability research, a landscape unit can express the spatial range of surrounding polygons of landscape feature types, and effectively alleviate the problem of no clear boundary in traditional tourist point description methods. The basic geomorphological types in China are determined by elevation and altitude which reflect the difference between internal and external forces.40. The human perception of landscape units has a holistic nature, mainly based on the differences in the topographic and geomorphological layers, and considering the characteristics of the vegetation types, therefore, landscape units can well meet the needs of environmental protection and management41,42. Thus, we selected altitude, , and vegetation types of Qilian Mountain National Park, and used overlay analysis in GIS to divide the study area to obtain 1641 landscape units in which we analyzed suitability of ecotourism activities.

Constructing an evaluation indicator system

The premise of ecotourism suitability evaluation is to clarify the influencing factors of suitability and construct a reasonable index system43. The study was conducted by combing domestic and international literature related to ecotourism, tourism suitability assessment, ecotourism development potential, and resort development planning. The available studies used elevation, slope, FVC, distance to rivers, and distance to settlements34,44. National parks carry out ecotourism activity projects, which inevitably impact the environment; impact of ecotourism behaviors on a fragile environmental system cannot be considered comprehensively by analyzing activity suitability alone. It is necessary to quantify impacts of a multitude of factors to avoid damage to the local ecosystems45. For example, environmentally restrictive indicators such as soil erosion and geological hazards need to be considered for goals related to sustainable development46. Based on the current status of the environment and ecotourism development in the Gansu area of Qilian Mountain National Park, factors affecting the suitability of ecotourism in Qilian Mountain National Park are divided into three major aspects: geological environment, ecological environment, and human activities.

The results are based on the weights of the indicators and hierarchical analysis following a questionnaire survey by a 16-member group composed of experts and scholars who have long been engaged in tourism and geography research, as well as managers who work in the Qilian Mountain National Park. The process of the survey was carried out in accordance with Delphi Method guidelines and regulations. The constructed indicators of each level were scored according to the degree of importance in a two-by-two comparison, and the results were summarised to construct a decision matrix. A maximum characteristic root (λmax), weight value, and consistency index CI value for each decision matrix were obtained, consistency test results were derived, and the weight of each index was obtained (Table 3).

Table 3 Index weights of different ecotourism activities.
Full size table

The weights of the indicators varied with ecotourism activities, with vegetation cover was assigned higher weights in ecological assessments, while settlement distance and elevation had the lowest weights suggesting that accessibility and altitude were key in trekking. Elevation and road traffic distance received a high weight for professional visits, indicating that altitude and traffic accessibility played a key role in hiking. Animal activity range and road traffic distance received a high weight for animal viewing. For hiking, elevation and road traffic distance received a high weight, indicating that altitude and traffic accessibility played a key role in hiking. The highest weights for ice and snow sports were given to elevation and slope, and basic service facilities were also important influencing factors. Vegetation coverage showed the greatest weight for forest recreation. Human activities such as field camping and farming-pastoral experience, residential distance, and road traffic distance received a high weight, indicating a key role in those activities.

Distribution of physical parameters

The study referenced the Interim Regulations on Planning Techniques for Ecological Functional Areas to establish elevation classifications. Slope grades were determined based on the National Basic Geomorphological Types of China 1:1,000,000 Geomorphological Map Mapping Specification (for Trial Implementation) prepared by the Institute of Geography of the Chinese Academy of Sciences. The degree of soil erosion, the degree of susceptibility to geological hazards, and the percent of vegetation cover were determined using national standards for grading criteria. Distances to water bodies, road traffic, service facilities, and settlements, and animal activity range were reclassified using the natural discontinuity grading method (Jenks) of ArcGIS software to determine the grade classification standard for each index value in the study area.

Then, the zoning statistics tool in ArcGIS was used to assign different scores to the suitability of all landscape units within the area. Scores 1–5 represented five levels of suitability: low, low average, high, and high, respectively (Table 4). The MAJORITY method (to determine the most common values among pixels in the same area as the output pixel in the value grid) was used for partition statistics according to the landscape unit. Finally, the suitability of ecotourism in each landscape unit was determined, and the landscape unit zoning map was produced with ArcGIS (Figs. 2 and 3).

Table 4 Ecotourism suitability index grading standard.
Full size table
Fig. 2
figure 2

Distribution of physical parameters. (A) Elevation map; (B) Slope map (C) Soil erosion levels; (D) Geological hazard levels. Drawn by the author using ArcGIS version 10.3 (https://www.arcgis.com/index.html).

Full size image
Fig. 3
figure 3

Distribution of physical parameters. (A) Distance to rivers; (B) Animal Activity Range (C) Percent vegetation coverage; (D) Distance to roads; (E) Distance to infrastructure; (F) Distance to settlements. Drawn by the author using ArcGIS version 10.3 (https://www.arcgis.com/index.html).

Full size image

Evaluation steps

We used the Promethee method with the Python software to evaluate the suitability of eco-tourism activities in 1641 landscape units. To do that, we calculated positive flow, negative flow, and net flow value of the landscape unit. The suitability of eco-tourism activity types of a landscape unit is comprehensively sorted to lay the foundation for a further spatial layout.The Promethee calculations were implemented in Python 3.9 using custom scripts to handle pairwise comparisons and flow calculations.

The assigned values of slope, soil erosion, and geological disaster susceptibility for ecotourism activities in unit 1 are all highly suitable; the distance from the range of animal activities and from the settlements are great, which indicates low suitability for animal viewing (Table 5).

Table 5 Scorecard for indicator "1" of the landscape unit.
Full size table

The net flow results Fig. 4 show that snow, ice sports, and cross-country adventure. Outdoor hiking and forest recreation was the next large one, and the smallest one was animal viewing and agricultural and animal husbandry experience. This indicates that landscape unit "1" was the most suitable for snow and ice sports ecotourism activities, followed by cross-country adventure, outdoor hiking, and forest recreation. Among them, the net flow values and negative flow values of snow and ice and off-road adventure were similar, indicating that landscape unit "1" has similar suitability for snow and ice and off-road adventure. Animal watching ecotourism exhibited the weakest suitability, and the small difference in positive flow values between animal watching and agro-pastoral experiences suggests that the two types of tourism activities have similar competitive advantages. The results of the suitability evaluation can be analysed to be consistent with the reality.

Fig. 4
figure 4

Net flow of ecotourism activities in landscape unit "1".

Full size image

The above evaluation steps were repeated for 1641 landscape units. As a result, a superiority rating was obtained for each landscape unit for ecotourism activities. The results of the ranking of ecotourism activities for all landscape units were summarised and counted. The activity suitability rankings (1–9) for all landscape units within the study area were obtained, and it was generally agreed that the top 4 ranked tourism activities were more important for the spatial layout of the study area.

The first four levels of suitability ranking of landscape units visualized with ArcGIS showed that the area suitable for cross-country exploration accounted for the largest proportion (69.87%) of the first-level suitable activities (Fig. 5-A), followed by forest recreation, and professional investigation. Animal viewing was not included in the first-level suitable activity map but in the second- and third-level. In the secondary suitable activities (Fig. 5-B), the largest area of the most suitable ecotourism activities was that of forest recreation, accounting for 45.02%. Other secondary suitable activities included professional investigation, cross-country adventure, skiing, and others. Among the third-level suitable activities (Fig. 5-C), the largest area of the most suitable ecotourism activities was that of professional investigation (35.67%) and forest recreation (18.88%). Among the fourth-level suitable activities (Fig. 5-D), the largest area of the most suitable ecotourism activities was that of outdoor camping (33.83%). The remaining detailed ranking results and activity suitability distribution are shown in Table 6. In general, the Gansu area of Qilian Mountain National Park should focus on the protection and sustainable use of semi-arid ecosystems, and design eco-tourism products around sports such as cross-country adventures, specialist visits, and forest recreation.

Fig. 5
figure 5

Ranking of suitability for ecotourism activities 1–4. Drawn by the author using ArcGIS version 10.3 (https://www.arcgis.com/index.html).

Full size image
Table 6 Area ranked for suitability for ecotourism activities (1–4).
Full size table

Discussion

The study’s findings align with China’s ‘One Belt, One Road’ initiative by promoting sustainable development in ecologically fragile regions like the Qilian Mountains. This study advances the field of ecotourism suitability assessment by integrating the AHP-Promethee-GIS into a unified framework. This hybrid approach leverages the strengths of both AHP and Promethee methods to provide a more robust and objective evaluation, addressing limitations of single-method approaches such as standalone AHP22. Furthermore, by incorporating qualitative assessments22,44, our multimethod integration addresses subjective biases in weight assignments and compensatory scoring limitations30,47. For instance, AHP’s pairwise comparison ensures transparent prioritization of criteria, while Promethee’s outranking method eliminates compensatory effects, enhancing objectivity37,38. Comparative studies in ecotourism zoning22,30 typically focus on single criteria like soil erosion or accessibility. In contrast, our framework incorporates dynamic factors such as animal activity ranges and geological hazards, reflecting the complexity of fragile alpine ecosystems50,52. This holistic approach aligns with global sustainability goals (SDGs) by balancing ecological protection and socioeconomic benefits1,53.

Based on landscape-unit-based ecotourism suitability assessments (Fig. 6) , this study proposes differentiated spatial management strategies and a policy framework with significant practical implications. This study’s framework offers practical tools for tourism and forestry departments to plan and manage eco-tourism activities effectively, ensuring they protect the environment while boosting local economies. In highly suitable zones (Level 1)—encompassing western glaciated regions and alpine meadows—cross-country adventure tourism and scientific expeditions are prioritized. Trail networks are optimized using slope stability analyses (Fig. 6) to ensure visitor safety while avoiding permafrost-sensitive areas. For instance, low-impact adventure tourism like cross-country expeditions can be developed in these high suitability zones, appealing to high-end markets and creating local employment opportunities. Strict visitor quotas and permit systems regulate human impact, balancing recreation with ecological preservation. In moderately suitable zones (Level 2–3), such as eastern forested areas, forest recreation and agro-pastoral tourism are developed through elevated boardwalks that mitigate vegetation disturbance and solar-powered campsites reducing fossil fuel reliance. These activities in moderate suitability zones can increase residents’ income and enhance their livelihoods. Seasonal closures during wildfire peaks and flood seasons further align tourism with environmental resilience. Lower suitable zones (Level 4)—marked by steep erosion-prone slopes and geological hazard hotspots—are designated as ecological reserves, restricting tourism to AI-monitored scientific research. Virtual reality experiences simulate glacier retreat (Fig. 6) and soil erosion dynamics, offering alternative educational pathways for visitors.These initiatives should actively involve local communities to ensure they directly benefit from eco-tourism.

Fig. 6
figure 6

Spatial layout of ecotourism activities. Drawn by the author using ArcGIS version 10.3 (https://www.arcgis.com/index.html).

Full size image

Policy frameworks reinforce these spatial distinctions through regulatory zoning, satellite monitoring, and AI-driven anomaly detection, imposing penalties for unauthorized access in Levels 1–2 zones. Eco-infrastructure standards mandate geohazard-resilient designs—such as avalanche barriers in high-risk zones—and seasonal firebreaks in grasslands. Community partnerships incentivize agro-tourism revenue-sharing mechanisms, including allocations of camping profits to herder cooperatives, and support cultural program development. Dynamic monitoring networks integrate Internet of Things sensors (IoT sensors) with machine-learning algorithms to predict risks such as landslides and water quality degradation, triggering real-time alerts. Transboundary collaboration establishes cross-border wildlife corridors and joint enforcement protocols with adjacent protected areas, exemplified by partnerships like Sanjiangyuan National Park.

The AHP-Promethee-GIS framework demonstrated in this study exhibits broad applicability across diverse national park contexts through adaptable parameterization and modular design. While calibrated to the geomorphological and ecological characteristics of the Qilian Mountains, the framework can be tailored to other regions by adjusting indicator weights to prioritize metrics such as water scarcity in arid zones or biodiversity indices in tropical rainforests. Contextual criteria—including cultural heritage values in socially sensitive parks—can be systematically integrated into the human activity layer to address region-specific needs. Its technical scalability further enhances transferability, as the Python-based Promethee implementation readily incorporates emerging technologies such as real-time IoT sensors or satellite imagery updates, enabling dynamic monitoring in ecologically fragile or rapidly changing environments.

Despite its methodological rigor, the study is constrained by data resolution limitations:including soil erosion datasets with 1 km resolution that may not capture micro-scale geomorphic variations critical for trail planning. Additionally, static modeling assumptions—such as linear preference functions in Promethee that oversimplify cross-criteria trade-offs—limit the dynamic adaptability of results. It is important to acknowledge that the current indicator system, grounded in landscape-unit analysis, prioritizes objective physical metrics to ensure consistency and reproducibility across the study area. However, this approach inherently lacks the rich qualitative insights that stakeholders—such as local communities, tourists, and conservationists—could provide. Their perspectives are crucial for a holistic evaluation, as they can offer valuable context regarding cultural values, community needs, and visitor preferences. Future research should aim to integrate these social dimensions into the evaluation framework, potentially through participatory GIS or mixed-methods approaches. This would not only enrich the assessment but also enhance the relevance and acceptability of the proposed ecotourism strategies. We emphasize that the current framework is not intended as a fully holistic evaluation but rather as a physically grounded foundation that can be expanded to accommodate social and cultural indicators. By explicitly recognizing these constraints, we hope to lay the groundwork for more inclusive and comprehensive research in the future. Expert-driven AHP weighting processes, which are susceptible to subjective biases, particularly risk misrepresenting socio-cultural values when local stakeholder input is absent. Implementation challenges further include potential resistance from tourism operators due to restrictive zoning recommendations allocating substantial high-altitude areas to specialized activities, alongside significant infrastructure costs for hazard mitigation systems like avalanche barriers. Addressing these limitations requires iterative refinements, such as integrating seasonal dynamic data streams and adopting participatory GIS tools to align conservation goals with community livelihood needs.

In conclusion, while our study provides a robust framework for assessing ecotourism suitability based on physical and geospatial indicators, it is important to recognize the limitations arising from the omission of stakeholder perspectives and the constraints of data resolution and static modeling assumptions. Future research should aim to integrate social dimensions and address these limitations to enhance the comprehensiveness and applicability of the framework.

Conclusions

This study develops an AHP-Promethee-GIS framework to resolve critical gaps in ecotourism suitability assessment for fragile alpine ecosystems. By integrating multi-criteria decision-making with spatial analysis, we developed a landscape-unit-based system that balances ecological constraints and socioeconomic demands. The framework identifies optimal spatial distributions for cross-country adventure, scientific research, and forest recreation in the Qilian Mountains National Park.

Promethee’s outranking method eliminates compensatory biases in weight assignments, while AHP pairwise comparisons ensure transparent criteria prioritization. These innovations achieve high-resolution spatial planning by dividing the study area into 1641 homogeneous landscape units, overcoming ambiguous zoning boundaries in conventional models.

Policy implications are immediate: Highly suitable zones (69.87% of western glaciers) prioritize low-impact activities like cross-country adventure, supported by slope stability analyses (Fig. 6). Moderate zones adopt agro-pastoral tourism via elevated boardwalks and solar infrastructure. Low suitability zones are designated as AI-monitored reserves for research.

The AHP-Promethee-GIS framework not only aids in protecting the ecological environment of Qilian Mountain National Park but also provides sustainable livelihood options for local residents, thereby improving their economic well-being. For example, in high suitability zones, adventure tourism and scientific research can generate employment and income for communities. In moderate suitability zones, forest and agro-pastoral tourism can further diversify income sources.

The framework’s design allows for adaptability in different regions and integration of new technologies for dynamic monitoring, ensuring long-term benefits for both ecosystems and communities. This study advances data-driven, spatially explicit planning for protected areas, offering a replicable blueprint to reconcile ecological protection with socioeconomic development in fragile ecosystems. This framework advances Goal 15(sustainable use of terrestrial ecosystems) and Goal 8 (sustainable economic growth) by balancing tourism revenue generation with ecological preservation.