Ecological vulnerability assessment of the Ya’an-Qamdo section along the southern route of the Sichuan-Tibet transportation corridor

Identifying the ecological vulnerability of the sensitive and fragile ecosystem of the Ya’an-Qamdo section along the southern route of the Sichuan-Tibet transport corridor is of paramount importance to reduce environmental damage resulting from infrastructure construction. This paper divided the Ya’an-Qamdo transport section into 22 subzones according to their ecological environment characteristics. Based on the vulnerability evaluation model established by the fuzzy matter-element analysis method, the eight main assessment indicators of ecological vulnerability were windstorm, rainstorm, snowstorm, extreme temperature, extreme air pressure, geological hazard, natural conditions, and social resources. The rating and ranking of vulnerability in each subzone were based on the weight of the judgment indicators. Scientific processes were used to verify the rationality of both the indicators themselves and their weights. The results of this study show that subzone 9, located in the subalpine cold and humid forest and scrubland zone, is the most vulnerable, and subzone 1, located in the low- to mid-land warm and humid forest zone, is the least vulnerable. The conclusion of the study suggests that targeted measures of ecological protection should be formulated before development and construction of major transportation infrastructure. Construction should evade the most vulnerable areas, and in-depth research on ecological restoration should be carried out in low- to mid-vulnerability areas so that the ecological environment along the route can be protected effectively for sustainable economic and social development.


Introduction
Since the Qing Dynasty, the Sichuan-Tibet transportation corridor has been comprised of the northern "commercial route" and the southern "official route" (Qiao 2013). Both routes start from Ya'an, Sichuan and pass through Tianquan, Luding, and Kangding (originally Dajianlu). The northern route goes northwest to Ganzi and Derge, then west through Qamdo into Lhasa, Tibet. The southern route runs westbound through Litang, Batang, and Nyingchi into Lhasa. The beautiful and longestablished Sichuan-Tibet transport corridor has rich and diverse ecosystems and natural landscapes, including forests, bushlands, meadows, and wetlands Ecological vulnerability assessment of the Ya'an-Qamdo section along the southern route of the Sichuan-Tibet transportation corridor (Ni et al. 2013;Luo et al. 2005). Many towns such as Ganzi and Dege on the northern route, and Litang and Batang on the southern route, are important economic and cultural centers of the region. As a typical region with Han and Tibetan cultures, it is important to build and develop the region reasonably so that its glorious natural beauty and unique culture can be showcased to the world. However, with diverse geological and harsh climatic conditions, the Sichuan-Tibet transport corridor is one of the most ecologically vulnerable areas worldwide. Due to global warming, increasing human activity, decreasing overall recovery capacity of the natural ecosystem, and a seriously threatened biodiversity, the region has presented the features of extreme sensitivity and extreme vulnerability (Li et al. 2003;Luo et al. 2004;Lai 1996).
Ecological vulnerability refers to the sensitivity and ecological resilience of an ecosystem to external disturbances under specific regional conditions and is the result of the internal succession of ecosystems and the interaction between natural factors and human activity (Wei et al. 2015). For an ecological vulnerability assessment, indicators are usually assigned to build measurable models for evaluation both at home and abroad, including the 'pressurestate-response' (PSR) model (Men and Liu 2018), the 'vulnerability scoping diagram' (VSD) model (Polsky et al. 2007;Li, Fan 2014;Chen et al. 2018), and the 'spatially explicit resilient vulnerability' (SERV) model (Frazier et al. 2014), with VSD being the most widely used. Established research involves a variety of quantitative assessment methods, which fall into two main categories: a) data assessment methods based on sampling (or historical data). This category is represented by principal component analysis (Abson et al. 2012;Kan et al. 2018), the TOPSIS method Yang et al. 2018), and the data envelopment method (Huang et al. 2019;Zhang and Liu 2013); b) methods for related problem-solving in uncertain environments under the domain of fuzzy mathematics. This category is mainly based on three theories: fuzzy mathematical theory, mutation theory, and matter-element extension theory (Zhang et al. 2009;Wang et al. 2019;Forootan, 2022). The main methods for determining the weights of judgment indicators are expert scoring, hierarchical analysis, entropy weighting, and evidence weighting (Chen et al. 2022;Li et al. 2011;Zou et al. 2021;Dagnino et al. 2008;Xia et al. 2016;Suter et al. 2020).
The Ya'an-Qamdo section has a complex topographic and geomorphological environment, which has attracted great attention from scholars for its ecological vulnerability. Literature review shows that existing studies have focused on sensitivity to external natural disturbances such as climate change, climatic hazards, and geological hazards (Zou et al. 2013;Gao et al. 2016;Guo 2017;Ma et al. 2019;Zhong et al. 2010;Zhang et al. 2004;Wang et al. 2021), which has laid a foundation for understanding the ecological context and environmental conditions along the route. However, with rapid economic and social development, the impact of human activity has become increasingly prominent. Existing research results have been unable to meet the practical needs of engineering and construction. Moreover, as the Sichuan-Tibet railway line has been announced to pass through this section, revealing its ecological vulnerability based on the impact of human activity and proposing quantitative evaluation methods will provide scientific and technological support for ecological environmental protection and sustainable economic and social development of the region along the line.

Study Area
In this paper, based on the scope of environmental assessment defined by the railway project construction (Jie et al. 2015), we identify the 10 km range on both sides of the centerline of the newly built railway line from Ya'an to Qamdo as the specific study object (Fig. 1). The main landscape formations of this section are the Sichuan Basin, the Hengduan Mountains, and the Qinghai-Tibet Plateau (Xu 1991). The Hengduan Mountain Range, which runs north-south in parallel rows, includes Qionglai Mountain, the Dadu River, Snowy Mountain, the Yalong River, Shaluli Mountain, the Jinsha River, Mangkang Mountain, the Lancang River, Nu Mountain, the Nujiang River, and Gaoligong Mountain. It is a vast and treacherous terrain with deep river valleys and fluctuations in elevation of up to 1,000-2,500 m, as well as diverse climate conditions including subtropical, plateau marine, and continental. With an altitude of more than 2,500 meters above sea level, the regions have remarkable climatic characteristics such as large daily fluctuations in temperature, uneven precipitation distribution, and seasonal frosts. The river valleys are arid and have prominent hydrothermal conflicts, with annual evaporation reaching 3-6 times the amount of rainfall.
The geography of this section includes lowland, midland, subalpine, and alpine elevation levels, as well as river valleys. The corresponding climatic environment and ecosystem of each elevation level is complex and variable, including warm and humid forests, cold and humid forests, dry scrublands, and alpine meadows. Satellite film and GIS raster data were combined and used to assess ecological vulnerability more accurately. The section is divided into 22 ecological subzones falling under five categories: low-to mid-land warm and humid forest zone; subalpine cold and humid forest and scrubland zone; subalpine river valley arid scrubland and grassland zone; alpine hill and high plain scrubland and grassland zone; and alpine hill and high plain river valley arid scrubland and grassland zone ( Fig. 2 and Fig. 3).

Evaluation model
A theoretical model for evaluating ecological vulnerability was constructed based on previous research results and the fuzzy matter-element analysis method (Bao and Qiu 2021). The model summarizes 20 influencing factors of ecological vulnerability, including temperature, precipitation, snowfall, herbaceous cover, number of plant species, population density, and GDP per capita. Under the VSD vulnerability analysis framework, the 20 influencing factors are generalized into eight judgment indicators, and an indicator evaluation system is established accordingly. In addition, Python programming software was used to construct a multi-factor and multi-attribute ecological vulnerability assessment and measurement model by combining fuzzy mathematics and physical element analysis (Xu et al. 2021). After obtaining the base scores of each judgment indicator through cluster analysis and determining each one's weight based on fuzzy matter-element analysis, we finally ranked the comprehensive vulnerability scores of each subzone. Among them, the judgment indicator base the weighted sum of the k th element of the vector of decision indicator weights k R for the k th column of L with a weight of w R . Finally, the correlation values are ranked to obtain the ranking results P of each scheme.
This paper conducts an empirical study of a theoretical model of ecological vulnerability assessment by calculating, scoring, and ranking the vulnerability of 22 ecological subzones of the Ya'an-Qamdo section along the southern route of the Sichuan-Tibet transport corridor.

Data collection and analysis
Mathematical calculation began with data collection and analysis of specific indicators for each of the 22 ecological subzones. Field sample research was used for the judgment indicators of herbaceous cover and number of plant species. A combination of field research and aerial photogrammetry was used for the judgment indicator of geological hazards, and GIS, remote sensing technology, and various statistical methods were used for data analysis. Specifically, we processed a large amount of data regarding meteorological stations and geological conditions such as elevation, slope rate, and slope direction.

Validation of the evaluation model
After calculating the scores and ranking, we investigated a validation method to address two problems that are rarely mentioned in existing research regarding the theoretical framework of subjective analysis, the rationality of the judgment indicators, and the validity of their weight assignments.

Rationality of judgment indicators
In order to verify the rationality of the judgment indicator scores, this paper proposes a new idea of partition clustering: Firstly, cluster analysis is done on 20 influencing factors (tertiary indicators) in each of the 22 subzones, with the number of clusters being If the indicator scores are rational, the clustering results    The physical meaning of separation degree: Select the largest th cluster in B as the corresponding cluster of the i th cluster in A . The higher the degree of separation, the lower the probability that each cluster in A will have the same corresponding cluster in B . A larger separation means that A and B are more similar. When A and B are exactly the same, K and is a diagonal matrix and the separation degree is at its maximum.
In summary, from the perspective of the physical meaning of aggregation degree and separation degree, if clustering results A and B are the same, then the aggregation and separation degrees reach the maximum; the greater the degrees of aggregation and separation, the more similar the clustering results of A and B are.

Validity of indicator weights
In order to verify whether the indicator weights are valid, the eight indicators of the 22 subzones are subjected to cluster analysis. The vulnerability scores and vulnerability ranking of the five clustering centers can be obtained based on the weights calculated by the fuzzy matter-element analysis method. The same method is applied again in each class to obtain the ranking of each subzone, thereby the vulnerability ranking vector Q of each subzone can be obtained. Then compare the similarity between Q and the vulnerability sorting result P obtained by the above algorithm. To illustrate the similarity of the sorting vectors obtained by the two methods, this paper calculates the cosine of the vector angle. Suppose the vector  is , , vector  is , , and the angle of α and β is  .
Then the similarity of  and  can be expressed by the cosine value of  , which is calculated as: where   cos 0,1   , the closer cos is to 1, the smaller the angle between  and  , and the more similar  and  are.
In summary, the similarity of the two ranking results can be obtained by calculating the cosine of the vector angle for the two vectors Q and P. The rationality of the indicator weights can then be verified.

4
Results of Ecological Vulnerability Assessment

Judgment indicator scores
In this paper, based on the influencing factors of eight judgment indicators, including windstorm, rainstorm, snowstorm, extreme temperature, extreme air pressure, geological hazard, natural conditions, and social resources, the 22 subzones are clustered into three classes according to k=3, which are shown in Fig. 4. Taking windstorm as an example, its influencing factors include annual maximum wind speed and annual average wind speed. Data for both factors are clustered into three classes according to k=3, which are shown in Fig. 3(A), where subzones 1, 2, 3 , 19, 20, 21, and 22, subzones 6, 10, 11, 12, 13, 14, 15, 16, 17, and 18, and subzones 4, 5, 7, 8, and 9 form the three classes.
According to the clustering results of Fig. 4, the rating of natural hazards in the 22 subzones can be determined. The subzone with the highest hazard rating has a score of s1=1; the subzone with the second-highest hazard rating has a score of s2=0.667; and the subzone with the lowest hazard rating has a score of s3=0.333. Taking windstorm as an example, the blue subzone has the highest hazard rating, with a score of s1=1; the yellow subzone has the secondhighest hazard rating, with a score of s2=0.667; and the green subzone has the lowest hazard rating, with a score of s3=0.333. The score of each judgment indicator can be further determined for each of the 22 subzones. The results are shown in Table 1.
To illustrate the eight judgment indicators in each of the 22 subzones more accurately, a visual aid is created of the numerical results of Table 1 for mapping thermodynamic diagrams. The darker the color, the more vulnerable the subzone is to that particular hazard. Taking the windstorm hazard as an example, you can see clearly that subzone 8 is the most vulnerable, and subzones 21 and 22 are the least vulnerable (Fig. 5).

Weighting of judgment indicators
The data of the eight judgment indicators are brought into the weight calculation model. Among the eight indicators, windstorm, rainstorm, snowstorm, extreme temperature, extreme air pressure, geological hazard, and sensitivity are of the smaller the better type, and adaptability is of the larger the better type. The correlation coefficient values corresponding to each indicator in each of the 22 subzones can be obtained, and thus the compound fuzzy matter elements are:  The vulnerability of each of the 22 subzones was rated by the degree of their correlation, and the results were ρ1 <ρ2 <ρ3 <ρ6 <ρ20 <ρ21 <ρ19 <ρ22 <ρ4 <ρ15 <ρ8 <ρ13 <ρ12 <ρ5 <ρ7 <ρ10 <ρ14 <ρ16 <ρ11 <ρ17 <ρ18 <ρ9 where ρi indicates the vulnerability of ith subzone. It can be seen that the most vulnerable is subzone 9, and the least vulnerable is subzone 1. The vulnerability rating of each of the 22 subzones is shown in Fig. 6.

Validation of Vulnerability Assessment Results
According to formulas 1 and 2 in section 3.3, we carried out rationality verification of the judgment indicators and the indicator weights in the ecological vulnerability assessment of the Ya'an-Qamdo section along the southern route of the Sichuan-Tibet transportation corridor. The results were as follows.
2 0 0 0 0 1 0 0 0 0 0 5 1 9 0 0 0 1 0 1 Then the separation and aggregation were calculated, and the results were Therefore, 0.74   is relatively close to 1. The higher the similarity of A and B , the more rational the scoring of the eight indicators is.

Conclusion and Suggestions
The ranking of the indicator weights shows that extreme climate, frequent earthquakes, and lack of social resources are the driving factors of ecological vulnerability on the Ya'an-Qamdo section along the southern route of the Sichuan-Tibet transportation corridor. The vulnerability score and ranking shows that subzone 9 is the most vulnerable and subzone 1 is the least vulnerable. With diverse vegetation types and an ecosystem that is resistant to external disturbance, subzones 1 and 2, both located in the low-to mid-land warm and humid forest zone, covering Tianquan County, Dayuxi Township, and the main urban area of Ya'an City, Sichuan Province, are sthe least vulnerable. Subzone 9, located in the subalpine cold and humid forest and scrubland zone, covers Kangding City and Yajiang County, Ganzi Prefecture, Sichuan Province, where a large number of ethnic minorities live. Due to its large topographic relief, various natural hazards, and underdeveloped economy, subzone 9 lacks the ability to effectively resist and recover from external disturbances, making it very vulnerable. The results of the vulnerability ranking show that forest areas and gently sloping areas are less vulnerable. This paper evaluated original data collected for each judgment indicator, used the fuzzy matterelement analysis method to rank the degree of ecological vulnerability, and verified the rationality of the evaluation model and assignment. Although some of the ranking results were a little different from what was expected, the overall results are consistent with professional empirical perceptions and are in line with common sense. Therefore, the research findings provide significant theoretical and data support for the ecological protection of the Sichuan-Tibet transportation corridor. Suggestions for follow-up studies are as follows: (1) There is still room for adjusting and improving the parameters and the verification method of the mathematical model. (2) Since the ecological vulnerability of Ya'an-Qamdo section involves other influencing factors, research methods from new perspectives need to be supplemented for cross-validation and improvement. Suggestions for planning and construction projects are as follows: (1) More attention should be paid to the dynamic patterns of local ecological vulnerability characteristics in regular forestry and natural resource management through leveraging scientific methods.
(2) The principle of avoidance in the permanent site selection of planning and construction projects is the priority in making ecological protection strategies. For example, we should avoid the most vulnerable subzones 9, 18, and 17 in surface engineering during the planning and construction, especially of permanent roads and large-scale land extraction and disposal sites. To reduce serious ecological damage that may hardly be recovered, we should analyze the key influencing factors of ecological restoration and then determine the restoration mode and measures for each subzone. (3) Scientific and technological research and field trial studies on solutions to ecological restoration and technology integration should be advanced actively in highly vulnerable subzones.

Acknowledgments
The paper is sponsored by the National Natural Science We would like to express our gratitude to Professor Li Xiaoping of the University of Electronic Science and Technology of China for his guidance and assistance in model construction and data processing for this study.

Open Access
This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.