Assessing the Sustainable Development Level of the Tourism Eco-Security System in the Chengdu-Chongqing Urban Agglomeration: A Comprehensive Analysis of Dynamic Evolution Characteristics and Driving Factors

Abstract

Based on the DPSIR framework, this study constructed an evaluation index system to assess the sustainable development levels of the tourism eco-security system (TESS-SDL) present in the Chengdu-–Chongqing urban agglomeration and synthesized multi-dimensional analysis methods to explore its spatiotemporal evolution characteristics and driving factors to provide an important theoretical and practical basis for promoting the sustainable development of the regional tourism eco-security system. The results showed the following: (1) From 2011 to 2021, the regional TESS-SDL was generally at a medium level and showed a trend of steady growth. Although the gap between cities was widening year by year, the speed of the TESS-SDL exceeded the speed of the widening gap, promoting the coordinated development of the regional TESS-SDL. (2) The spatial spillover effect of the regional TESS-SDL was obvious; however, the siphoning effects of Chongqing and Chengdu were strong, and the demonstration effect was insufficient. (3) The dynamic evolution process of the TESS-SDL shows a strong self-locking effect. The risk of downward development (lower sustainability) is greater than the potential for upward development (higher sustainability) and is significantly influenced by neighboring cities. (4) In terms for assessing the driving factors, open-door and green-development policies show positive facilitating effects, while the positive influencing capacities of information technology, economic development, and tourism are moving toward having negative effects; the influencing effect of technological innovation has transitioned from positive to negative.

1. Introduction

The recovery of the global economy and increasing tourism demand in the post-pandemic era have stimulated the tourism economy, which is now experiencing a growth spurt. The massive flow of tourists and the development of tourism and related industries have brought a number of challenges and problems to regional tourism ecosystems. Because of China’s rapid economic development and accelerated urbanization, tourism has become an important pillar of the national economy; however, environmental problems relating to the rapid development of the tourism industry cannot be viewed optimistically. In April 2024, at the Symposium on Promoting the Development of the Western Region in the New Era, President Xi put forward a major requirement concerning “six adherences”, which includes “adhering to a high level of ecological protection to support high-quality development and building a firm national eco-security barrier”. He emphasized the importance for balancing economic growth and environmental protection and advocated transforming the rich natural and cultural resources of the western region to economic assets while promoting the formation of green production methods and consumption patterns. The western region is the main source of China’s great rivers and is a concentrated area of forests, grasslands, wetlands, and lakes. It is both an important ecological barrier and a treasure trove of precious ecological resources, with rich architectural and natural landscapes, providing abundant resources for the development of the regional tourism.

The Chengdu–Chongqing urban agglomeration (Hereinafter referred to as the “Chengyu region”) is located in the western region of China; it includes Chongqing, Chengdu, and the neighboring cities (as shown in Figure 1), and it is an important platform for China’s Western Development Policy. As the most populous economic segment in inland China, the Chengyu region plays an important role in China’s economic development.

The regional economy has remarkably improved in the second round of the Western Development Policy, resulting in a gross regional product (GRP) share of 6.5% of the country’s GDP, an increase of 6.1% compared to the previous year, which is 0.9 percentage points higher than the average of the country (China National Bureau of Statistics, 2023). Tourism is of strategic importance for the economic development of the region. The Chengyu region, with its long history and culture and rich natural landscapes, as well as its increasingly improved infrastructure, public services, business environment, and visibility and reputation, which have been enhanced by large events and festivals, continues to play the role of the “locomotive and ballast” of the tourism economy, together with the Yangtze River Delta and other places. Located in the upper reaches of the Yangtze River Basin and connecting the east and west of China, the Chengyu region is an important economic hub and transportation node. Its special geographical location promotes regional economic integration and plays an important role in protecting the ecological environment. In recent years, the Chengyu region has undergone significant urbanization and industrialization and is facing severe ecological challenges alongside its rapid economic development. Frequent extreme natural disasters, such as floods and earthquakes, have exacerbated the vulnerability of regional ecosystems, further highlighting the importance of ecological security. In addition, the Chengyu region faces increasingly serious problems relating to resource depletion and environmental pollution, and there is an urgent need to explore sustainable development pathways.

In this context, this study takes the Chengyu region as its research object, integrates tourism development and ecological security issues, and constructs a comprehensive model of ecological security for tourism. By analyzing the spatiotemporal evolution and key driving factors relating to the Chengyu region, this study aims to reveal a sustainable development pathway for an ecological security system for tourism in the region. As the Chengyu region is an important center of economic growth in western China, this study not only helps to improve the regional eco-security level of the tourism there but also provides a reference for studying the ecological security systems present in other regions. Furthermore, it contributes to the realization of an ecological civilization and the fulfillment of the sustainable development goals in China.

Tourism eco-security (TES) is not only related to the sustainable development of tourism but also has a profound impact on the global ecological environment and human wellbeing. The study of TES originated from the concern about environmental problems involved in the processes of tourism development [1]. With the booming development of tourism, its impacts on the natural environments, cultural heritage, and human communities have become increasingly significant. Therefore, scholars have begun to explore and research the definitions, evaluation methods, and influencing factors relating to TES from multiple perspectives. These studies help to deepen our understanding of TES and provide a scientific basis for the sustainable development of regional tourism. However, current TES research still has many shortcomings. First, the construction of an evaluation index system needs to be further improved to ensure the accuracy and comprehensiveness of evaluations. Second, empirical research methods need to be constantly innovated and optimized to adapt to the complex and changing ecological environments. Finally, the methods used to transform research results to practical applications to promote the green development of tourism need to be further explored.

In order to systematically analyze the complexity of the ecological security systems relating to tourism, this study introduces the DPSIR framework. The DPSIR framework (driver–pressure–state–impact–response) is an integrated approach that can be used to analyze environmental problems to effectively reveal the causal relationships between various elements, and it is now widely used in research focusing on ecological security for tourism. Huong et al. (2020) used the DPSIR model to describe the logical interactions of coastal ecosystems and identify the causes and consequences of environmental and resource impacts caused by socioeconomic development [2]. Additionally, Kristiadi et al. (2022) used the DPSIR framework to assess the sustainability of urban ecosystems in order to study the drivers of urban climate change and propose countermeasures [3]. Compared with similar research models, such as PSR (pressure–state–response) and DSR (driver–state–response), the DPSIR model has significant advantages. First, the DPSIR model contains five components—the driver, pressure, state, impact, and response—providing a more comprehensive analytical framework when conducting systematic research. This comprehensiveness allows for a study to capture the complex interactions between the elements in greater detail. Second, the DPSIR model has a clear causal logic. The driving force triggers pressure, pressure affects the state of the environment, a change in state has an impact, and the impact prompts society to respond. Such a logical chain helps to understand the root causes of environmental problems and potential corresponding countermeasures. By clearly demonstrating the causal relationships among the elements, the DPSIR model can effectively explain the operating mechanisms of complex environmental systems. Finally, the DPSIR model has strong policy guidance capabilities. As it covers the complete process from drivers to responses, it can provide greater operational guidance when policymaking. By identifying and analyzing the key factors in each link, the DPSIR model can help policymakers to find the most effective intervention points and management strategies, thus improving the effectiveness and implementation of policies.

This study examines the Chengyu region in 2011–2021, utilizing the DPSIR model, coupled coordination analysis, spatiotemporal geographically weighted regression (GTWR), spatial autocorrelation, and Markov chain to assess the TESS-SDL in the region and to determine the driving factors and characteristics of the spatial evolution of its development and eventual developmental situation.

Previous research on ecological security for tourism has achieved rich results; however, there are still places worth exploring. First of all, the existing results are mostly focused on countries [4], provinces [5], ecological functional areas [6], ecologically fragile areas [7], and other typical areas. There is a lack of research on China’s largest inland area, and the important strategic area of the Chengyu region has been insufficiently considered. Second, this study expands the application of the Markov chain by applying it within the scope of a regional tourism ecological security system (TESS), combining it with spatial factors to analyze the characteristics of the evolution of the study area under the influence of adjacent spatial units from a geospatial perspective. Introducing a spatial Markov chain enriches the methodology used to conduct TESS research, provides a new analytical tool for understanding the dynamic changes and steady-state characteristics of a system, and deepens our understanding of the evolutionary law relating to the sustainability of the TESS. Finally, most existing studies use econometric models to analyze the drivers at the aggregate level, which allows for the key drivers to be identified but neglects the local characteristics and influences of each driver at temporal and spatial scales. This study uses spatiotemporal geographically weighted regression to reveal the spatial heterogeneity and dynamic trends of the drivers to reflect the driving mechanisms of the TESS-SDL more precisely in the region. At the practical level, this helps to realize the TESS-SDL in the Chengyu region by balancing the relationship between economic benefits and ecological protection; secondly, it provides effective policy guidance, and through the analysis of the current situation of the TESS-SDL, it provides a scientific basis for the government and related organizations to formulate policies concerning tourism development and the protection of the ecological environment.

The remainder of this article is structured as follows: Section 2 provides a literature review, Section 3 provides the research methods and data sources, Section 4 provides the research results, and Section 5 provides a discussion.

2. Literature Review

The current state of the literature is summarized in the following sections.

2.1. Definition and Importance of Tourism Eco-Security

Tourism eco-security (TES) refers to the preservation of the natural environment, cultural heritage, human wellbeing, and the safety of tourism destinations in relation to tourism development processes; it is the basic tenet of sustainable tourism development [8]. In the early stage, the main focus was on interpreting concepts and definitions. Because of different research perspectives, the definitions of TES also vary. Dong et al. (2003) introduced eco-security theories and methods to sustainable tourism development and proposed that the natural resources and ecological environments on which sustainable TES is based should exist in a healthy, balanced state that is free from threats or risks. With such conditions and development trends, the ecosystem can continue to exist while also meeting the needs of sustainable tourism development [9]. Zhang et al. (2008) studied the eco-security of tourist areas from the perspective of human “consumer” characteristics, focusing on ecology and mobility in relation to tourism consumption [10]. TES refers to the ability of a regional ecological environment to meet the needs of sustainable tourism development in that region, support the cyclical development of living systems, ensure that the productivity and wellbeing of residents and the social economy are not harmed, provide tourists with better tourism experiences, and promote the unified and coordinated development of the natural, economic, and social systems in a tourist area [11].

2.2. Evaluation Methods for Tourism Eco-Security

In recent years, scholars have made significant progress in evaluating TES, employing a variety of methods to assess its status. Evaluation methods can be categorized into single-indicator evaluations and indicator system evaluations, both of which can be categorized as comprehensive evaluations. Single-indicator evaluations focus on measuring a specific ecological indicator of ecological security. For instance, the tourism ecological footprint can be quantified in terms of the pressure that tourism places on the ecosystem [12,13]; the SBM model can be used to assess the eco-efficiency of TESS in terms of economic and ecological performance [8], and habitat quality can be used to evaluate the harmonious relationship between humans and nature through land use changes [14]. In contrast, indicator system evaluations construct a systematic framework by integrating ecological, tourism, and socioeconomic indicators. Common models include the DPSIR model [15,16], the PSR model [17], and the DSR model [18]. Using these frameworks, scholars have mostly applied methods like improved TOPSIS [19,20], GM (1, 1) [21,22], and principal component analysis [23] for multi-attribute decision-making and comprehensive evaluation. To enhance the comprehensiveness, diversity, and reliability of these methodological systems, scholars have also employed advanced models, like the entropy matter-element model [24], cloud model [25], fuzzy comprehensive evaluation method [26,27], and rough sets [28]. These methods compensate for data uncertainty by bridging fuzziness and randomness, allowing for the evaluation of complex systems. Additionally, remote sensing and geographic information system (GIS) technologies can provide data on the ecological environment with extensive spatial and temporal resolutions, supporting the dynamic analysis of TES [29]. With advancements in machine learning and artificial intelligence, these technologies are increasingly being applied to TES evaluation to deal with complex nonlinear relationships and large datasets to significantly improve evaluation accuracy and efficiency [30]. Furthermore, TES evaluations are gradually moving toward multidisciplinary integration, combining theories and methods from fields such as ecology, economics, and sociology to conduct comprehensive evaluations and provide a more holistic analysis of the ecological security status and the influencing factors.

2.3. Driving Factors of Tourism Eco-Security

Driving factors are the central forces driving systemic change, and driver analysis is crucial in terms for obtaining a comprehensive understanding of TES and related research. Driving factors can come from within or outside of the system, and this study attempts to construct a system of drivers that combines the internal and external factors of the system. Ruan et al. used the geo-detector to explore the key impact indicators and their driving mechanisms by detecting systemic factors [31]. Ma et al. employed spatial econometric models to study the spatial effects of the regional TES and the driving effects of its variables [32]. Shi et al. conducted a barrier degree analysis of a TES evaluation system, finding that industrial sulfur dioxide emissions and urban population density are the main obstacles to TES, and, thus, pressure system is the main pathway to improving TES [33]. Related studies have often selected core indicators for driver analysis, which are typically classified into the following categories: social and environmental factors (e.g., the population density, urbanization rate, and infrastructure development); economic factors (e.g., GDP per capita), education and environmental protection (e.g., education expenditure and environmental protection support), tourism infrastructure and capacity (e.g., the tourism reception capacity, including the number of star-rated hotels and travel agencies), and environmental factors (e.g., the green area of built-up areas, number of tourists, and industrial wastewater discharge) [15,34,35]. In addition to research on internal drivers, some studies have also examined external influences on the system, highlighting the complex relationship between TES development and change. Common external influences include the population density, economic levels, environmental regulations, and scientific and technological innovations [36,37]. Furthermore, studies combining internal and external factors are essential for understanding TES. For example, Tian et al. selected the regional economic strength, industrial structure, market scale, policy regulation, and technological innovation ability to explore the coupled role of tourism, urbanization, and the ecological environment when conducting driving mechanism research [38]. Yang et al. introduced a tobit model to investigate the effects of patent authorization, the tourism industry’s structure, the total energy consumption, environmental governance, financial expenditures, and environmental construction on tourism eco-efficiency [39]. Lai et al. chose prior variables to explore the combination of factors for high TES levels based on the three dimensions of the TOE framework, choosing technological innovation and R&D investment for the technological dimension, policy attention and financial support for the organizational dimension, and economic development and industrial structure for the environmental dimension [40]. When studying the driving effects of these factors, scholars have often used methods such as geo-detectors [41,42], barrier factor analysis [43], the tobit model [39], stepwise linear regression [44], the vector autoregression (VAR) model [45], and spatial econometric models. In summary, combining internal and external drivers in research methods can comprehensively cover the various factors required by the system, ensuring its adaptability and sustainability in different contexts. This approach provides solid theoretical and practical foundations for the sustainable development of the tourism industry.

2.4. Review of Tourism Eco-Security Research

TES, a research field that has recently attracted much attention, not only concerns the sustainable development of tourism but also has a far-reaching impact on our understanding of the global ecological environment and human wellbeing. According to the literature, scholars have previously evaluated TES as a research concept; however, this study treats TES as a system, clarifies the scope of such research, and determines the system’s boundaries from a systemic perspective. When evaluating the sustainable development levels of TESS, researchers have mostly adopted comprehensive evaluation methods; however, this study introduces a system-coupling coordination index, which has greater research significance. Studies on the spatial distribution characteristics of the regional TES often adopt static distribution methods. However, this study adopts a dynamic approach and uses spatial autocorrelation and Markov chain models to investigate the probability transition matrix of the spatial dependence relating to the sustainable development levels of a regional TESS. Finally, most previous studies have explored the driving factors of TES development using models such as econometric models and geo-detectors, and they used idealized means to construct the regression model, which led to biased conclusions. This study uses the GTWR model to analyze the differences in the effects of the TESS-SDL driving factors in time and space and to provide a reference for the coordinated development of the regional TESS-SDL.

3. Materials and Methods

3.1. Construction of an Indicator System

This study constructs an evaluation indicator system for the study of the TESS-SDL in the Chengyu region based on the DPSIR model, taking into account the accessibility and timeliness of the data as well as the comprehensiveness and reasonableness of the indices (

Table 1. TESS indicator system.

Standardized Layer	Indicator Layer	Description of Indicator	Weight
Driver	Fixed-asset investment total	Reflecting the investment level and development trend of the economy	0.021
	GDP per capita	Reflecting the regional average economic performance or per capita standard of living	0.037
	Population density	Reflecting the concentration of the population in a region	0.011
	Natural population growth rate	Reflecting demographic changes in potential areas	0.004
	Tourism revenue	Reflecting the level of tourism development and the economic contribution of the region	0.042
	Passenger transport volume	Reflecting the number of people moving through the region	0.032
	Tourism resource density	Reflecting the potential for tourism development or activity in that area	0.012
Pressure	Tourism spatial density	The ratio of the number of tourists to the urban space area in a year	0.013
	Tourism population density	The ratio of the number of tourists to the permanent urban population in a year	0.020
	Tourism transport level	Refers to the number of road miles per 100 square kilometers	0.048
	Annual electricity consumption	Social electricity consumption for the year	0.030
	Annual gas supply total	Social natural gas consumption for the year	0.018
	SO2 concentration	Reflecting the air quality for tourism	0.017
	Urban domestic sewage discharge	Reflecting the consumption level of the urban population (including tourists) and environmental pressures from domestic sewage	0.017
	Urban domestic waste removal capacity	Reflecting the consumption level of the urban population (including tourists) and environmental pressures from domestic waste	0.018
State	Industrial structure	Reflecting the share of tourism in the economy	0.081
	Green development of tourism	Reacting to the natural landscape of the tourist attraction and its quality as a percentage	0.017
	Ecosystem services	Evaluating the total value of various land types using the market value approach	0.032
	Proportion of days with good air	Reflecting the regional air quality	0.030
	Drinking water compliance rate	Reacting to the compliance of water sources for the centralized drinking water supply	0.034
	Proportion of protected areas	Reflecting the levels of nature conservation and scientific civilization in the region	0.087
	Per capita urban green space	Reflecting the urban ecological environment, life quality of the residents, and management level	0.034
Influence	Urbanization rate	Reflecting the transformation of the local social structure	0.040
	Employees in the tertiary sector	Reflecting the structure of the economy and the importance of service activities	0.014
	Retail sales of consumer goods	Reflecting consumer spending patterns and the overall economic activity	0.014
	Product sales revenue	Revenue from purchasing units on sales of finished goods and self-made semi-finished products by industrial enterprises at the sale price	0.025
	Per capita disposable income	Reflecting the standard of living and economic wellbeing of the population	0.067
Response	Environmental investment	Reflecting the strength of government investment in environmental protection	0.065
	Green coverage in built-up areas	Reflecting the level of urban greening	0.037
	Disposal rate of domestic waste	Reflecting the level of technology and efforts of people in dealing with household waste	0.054
	Sewage treatment rate	Reflecting the level of technology and efforts of people in treating domestic wastewater	0.028

). The DPSIR model, which is the full name of the “driver–pressure–state–impact–response” model, provided a systematic approach for analyzing and assessing the impacts of human activities on the ecological environment.

The DPSIR model, which combines the advantages of the PSR and DSR models, was formulated by the Organization for Economic Cooperation and Development (OECD) in 1993 and further developed by the European Environment Agency (EEA) in 1999 [46] and has been widely used in ecological management and sustainable development research [47]. Through the DPSIR model, we can systematically analyze and understand the impacts of tourism activities on the ecological environment, assess the health of the ecosystem, and formulate appropriate management and protection measures to achieve the sustainable development goal of the TES.

3.2. Research Methods

3.2.1. Indicator Weight of the TESS

According to the TESS indicator system, the subjective–objective comprehensive weighting method was used to judge the value of each subsystem, and the coupled coordination degree model was used to assess the TESS-SDL in the Chengyu region. The subjective weight (AHP [48]) is easily affected by personal prejudice, emotion, and other non-objective factors, which may lead to unfairness or errors in decision-making results, while the objective weight (entropy weight method [49]) depends on objective data information. If the data are lacking or inaccurate, this may lead to weight calculation errors or distorted decision-making results. In order to avoid the shortcomings of subjective and objective weights, this study adopts the idea of game theory to integrate subjective and objective weights, dilute the measurement deviation brought by subjective and objective weights, and better improve the rationality and reliability of the index system evaluation [50,51].

3.2.2. Evaluation of the Regional TESS-SDL

TESSs conceptualized using the DPSIR model are systems with complex internal relationships, and the value judgments of their sustainable development levels should not be simple linear sums. In order to reveal this complexity, this study considers each module as a subsystem and judges the TESS-SDL by evaluating the degree of coupling coordination between the subsystems [52], and the coupling coordination degree expressions are as follows:

$$C=5{\left[{\displaystyle \frac{U_d\times U_p\times U_s\times U_i\times U_r}{{\left(U_d+U_p+U_s+U_i+U_r\right)}^5}}\right]}^{1/5}$$

(1)

$$T=\alpha U_d\times \beta U_p\times \gamma U_s\times \epsilon U_i\times \phi U_r$$

(2)

$$D={(C\times T)}^{1/2}$$

(3)

where C is the coupling of the TESS; T is the coordination of the TESS; U_i represents the security degree of subsystem i; $\alpha ,\beta ,\gamma ,\epsilon ,\phi$ are coefficients to be determined; and D represents the coupling coordination degree of the TESS, were D ∈ [0, 1] and the greater the value, the higher the coupling coordination degree of the TESS. In this study, the coupling coordination degree of the TESS was used to measure the TESS-SDL. Based on the actual situation [53], the coupling coordination degree was divided into five levels (

Table 2. Classification criteria of the coupling coordination degree.

Composite Index	0~0.2	0.2~0.4	0.4~0.6	0.6~0.8	0.8~1
Level	I	II	III	IV	V
Coupling Coordination Type	Incoordination	Less Coordination	Critical Coordination	General Coordination	Coordination

).

3.2.3. Evolution Analysis of the TESS-SDL

A Markov process is a kind of stochastic process, which important feature is memorylessness; that is, the future state of the object does not depend on the past state but only on the current state. According to the stochasticity, stability, and posteriority of the Markov chain, Markov processes are used to describe the dynamic evolution of the TESS [54]. The main reasons are as follows: (1) The state transition process of the TESS-SDL is stochastic; it changes the state according to a certain transition probability. The probability matrix describes the likelihood of the system transitioning from one state to another. (2) The state transfer of the regional TESS-SDL depends only on the current state and not on the past state. Finally, the study of the regional TESS-SDL and its path dependence using Markov chains provides a reference basis for the development of relevant policy measures in the future.

$$P\left[X(t+∆t)=e_j\mid X\left(t\right)=e_i\right]=P\left[X(∆t)=e_j\mid X\left(0\right)=e_i\right]=p_{ij}(∆t)$$

(4)

$$\begin{gathered} M=\left[\begin{array}{cccc}{p}_{11}(∆t)& p_{12}(∆t)& \cdots & p_{1n}(∆t)\\ p_{21}(∆t)& p_{22}(∆t)& \cdots & p_{2n}(∆t)\\ ⋮& ⋮& & ⋮\\ p_{n1}(∆t)& p_{n2}(∆t)& \cdots & p_{nn}(∆t)\end{array}\right] \\ \left\{\begin{array}{c}{p}_{ij}(∆t)\ge 0,i,j=1,2\cdots ,n\\ {\displaystyle \sum _{j=1}^n{p}_{ij}}(∆t)=1,i=1,2\cdots ,n\end{array}\right. \end{gathered}$$

(5)

where $p_{ij}(∆t)$ is the probability function of state e_i transitioning to e_j within the time interval Δt. When the time interval is the same, Formula (4) can be written as (5).

A spatial Markov chain introduces a spatial lag operator based on a traditional Markov chain [54,55]. The spatial lag operator is used to express the spatial interaction relationship in the state transition process of the TESS-SDL and defines the relationship between the state transition probability of the regional TESS-SDL and adjacent research units, making up for the spatial deficiencies of the traditional Markov chain.

3.2.4. Spatiotemporal Geographically Weighted Regression of the TESS-SDL

The GTWR model was developed on the basis of the geographically weighted regression model [56], and its advantages lie in its ability to simultaneously consider temporally and spatially localized heterogeneities, overcome the shortcomings brought by artificial regional grouping in the traditional model, and solve temporal and spatial non-stationarities of geographic factors for explanatory variables [57] to improve the explanatory ability and predictive accuracy of the model and provide more detailed localized analyses. Its expression is as follows:

$${\mathrm{y}}_i={\beta }_0\left(u_i,v_i,t_i\right)+{\displaystyle \sum _k{\beta }_k}\left(u_i,v_i,t_i\right)X_{ik}+{\epsilon }_i$$

(6)

$$\widehat{\beta }\left(u_i,v_i,t_i\right)={\left[X^{T}W\left(u_i,v_i,t_i\right)X\right]}^{-1}{X}^{T}W\left(u_i,v_i,t_i\right)Y$$

(7)

where $\left(u_i,v_i\right)$ are the geographic coordinates of the administrative center of a regional city, t_i is the year of the observation, y_i is the explanatory variable for city i, and X_ik denotes the kth explanatory variable for city i. ${\beta }_0\left(u_i,v_i,t_i\right)$ is a constant, and ${\beta }_k\left(u_i,v_i,t_i\right)$ is the regression coefficient of the kth explanatory variable.

According to the principles of typicality and representativeness, the driving factor system of the TESS in the Chengyu region was constructed, and the driving factors mainly included the internal and external factors of the system. The internal factors (green policy level, industrial structure level, economic development level, infrastructure quality, and tourism development level) were mainly selected with reference to the content of the research and the weight of the indicators; the external factors (information technology level, technological innovation level, and openness level) were mainly selected based on combining the relevant references [38,39,40] and the main concerns of today’s social development issues. In the contexts of the transformation of old and new kinetic energies and the green development of the Chinese economy, in order to promote the high-quality and intensive development of tourism, the opening up of the country to the outside world, technological innovation, and the application of information technology can enhance the modernization and intelligence of the system, thus improving the efficiency of tourism management and services. In addition, the current policy environment and changing international situation make it possible to study external drivers to help us to better understand how TESS adapts to these changes and, thus, improve its coping capacity and development. The quantitative and descriptive statistical analyses of the driving factors are summarized in

Table 3. Descriptive statistics of the driving factors of TESS-SDL.

Variable Type	Name	Indicator (Unit)	Mean	S.D.	Min.	Max.
Explained variable	SDL-TESS (Sdl)	Coupling coordination degree of TESS	0.549	0.081	0.285	0.730
Explanatory variable	Information technology level (Tech)	Number of mobile phone users (per 100 persons)	20.637	5.738	10.986	41.606
	Technological innovation level (Inn)	Number of patents granted	60.5402	209.785	0	1490
	Openness level (Ope)	Total exports and imports as a share of GDP (%)	0.1313	0.1385	0.0095	0.5458
	Green policy level (Gre)	Environmental investment as a share of GDP (%)	0.028	0.062	0.008	0.035
	Industrial structure level (Ind)	Value added in tertiary industry as a share of GDP (%)	38.432	9.380	20.16	70.28
	Economic development level (Eco)	GDP per capita (Yuan)	33,226.2	25,119.79	678.977	122,972
	Infrastructure quality (Inf)	Tourism transport level (km/100 km2)	137.961	55.959	25	326.564
	Tourism development level (Tou)	Tourism revenue (108 Yuan)	341.829	567.226	34	4650

.

3.3. Data Sources

This study takes the Chengyu region as the study area. The scope includes 15 cities in Sichuan Province (Chengdu, Zigong, Luzhou, Deyang, Suining, Neijiang, Leshan, Nanchong, Meishan, Yibin, Guang’an, Ziyang, and parts of Mianyang, Dazhou, and Ya’an), 27 districts (counties), and parts of two districts (counties) in Chongqing Municipality, with a total area of 185,000 km², accounting for 1.9% of the national land area. The data for the study are mainly from the Statistical Yearbook of each city (2011–2021), the Statistical Bulletin of National Economic and Social Development of each city (2011–2021), and the Annual Report of Ecological Environmental Statistics of each city (2011–2021). In addition, the data were obtained from the official websites of these cities, such as the Bureau of Ecology and Environment and the Bureau of Culture and Tourism, and some of the missing data were filled in by interpolation. The map data mainly came from the National Information Resources Directory Service System (https://www.webmap.cn, accessed on 30 May 2024).

4. Results

4.1. Development Trend of the TESS

According to the DPSIR model, this study decomposed the regional TESS into five subsystems: D, P, S, I, and R. Figure 2 shows the subsystems of the TESS and their coupling coordination degrees.

As shown, except for the state (S) subsystem, which showed a fluctuating downward trend in the TESS of the Chengyu region from 2011 to 2021, the other four subsystems—driver (D), pressure (P), impact (I), and response (R)—showed upward trends. Among them, the driver (D) subsystem fluctuated from 0.260 to 0.357, with a growth rate of 3.2%, and the pressure (P) subsystem rose from 0.242 to 0.336, with a growth rate of 3.3%. As can be seen in the figure, the graphs of D and P were relatively close to each other, and both showed similar development trends, indicating that the regional tourism industry and socioeconomic development were more consistent. However, it can be seen that the D of the system from 2020 to 2021 underwent an obvious improvement. In January 2020, the sixth meeting of the Central Financial and Economic Commission put forward the proposal for “promoting the construction of the Chengdu–Chongqing Twin Cities Economic Circle”, raising the development of the Chengyu region to a part of the national strategy. This decision provided strong policy support and direction for the development of the Chengyu region and promoted its rapid economic and social growth. The state (S) subsystem decreased from 0.659 to 0.552, changing from more coordinated to critically coordinated, with a change of 0.117, which is within an acceptable range. The impact (I) subsystem fluctuated from 0.372 to 0.432, from generally coordinated to critically coordinated, and the response (R) subsystem fluctuated from 0.322 to 0.443, which promoted the recovery and development of the system because the government and the relevant departments strengthened the supervision of the environmental protection and TES; they also increased their investment in tourism eco-security and problem management.

Figure 2 shows that the coupling coordination degree of the TESS showed a significant and stable growth trend between 2010 and 2021. Specifically, starting from 0.48 in 2010, the coupling coordination degree increased year by year and reached the highest point of 0.73 in 2021. The enhancement of the coupling coordination of the TESS was a complex process that required joint efforts and the support of the government, enterprises, and all sectors of society, resulting in the effective enhancement of the coupling coordination of the TESS and the sustainable development of the tourism.

4.2. Analysis of the Dynamics of and Differences in the Regional TESS-SDL

The coupling coordination degree model was used in this study to calculate the TESS-SDL (

Table A1. TESS-SDL in 16 cities of the Chengyu region (2011–2021).

City	2011	2012	2013	2014	2015	2016	2017	2018	2019	2020	2021	S.D.
Chengdu	0.562	0.577	0.567	0.592	0.601	0.642	0.49	0.616	0.642	0.579	0.646	0.046
Dazhou	0.325	0.309	0.333	0.38	0.369	0.371	0.375	0.374	0.371	0.349	0.48	0.044
Deyang	0.472	0.512	0.533	0.547	0.561	0.533	0.555	0.547	0.562	0.568	0.638	0.041
Guang’an	0.45	0.499	0.54	0.571	0.559	0.597	0.58	0.606	0.537	0.496	0.583	0.049
Leshan	0.522	0.561	0.578	0.604	0.597	0.627	0.595	0.64	0.673	0.686	0.71	0.056
Luzhou	0.44	0.503	0.467	0.406	0.486	0.486	0.558	0.529	0.592	0.468	0.504	0.052
Meishan	0.499	0.568	0.58	0.535	0.521	0.587	0.548	0.562	0.576	0.562	0.583	0.028
Mianyang	0.424	0.464	0.523	0.53	0.514	0.516	0.499	0.479	0.53	0.536	0.572	0.040
Nanchong	0.37	0.434	0.415	0.424	0.441	0.454	0.444	0.482	0.471	0.536	0.59	0.060
Neijiang	0.385	0.459	0.477	0.478	0.45	0.503	0.484	0.473	0.483	0.505	0.578	0.046
Suining	0.589	0.537	0.569	0.589	0.61	0.621	0.587	0.596	0.613	0.659	0.663	0.037
Ya’an	0.51	0.647	0.614	0.625	0.637	0.64	0.598	0.606	0.686	0.654	0.713	0.052
Yibin	0.475	0.445	0.468	0.486	0.512	0.485	0.5	0.548	0.534	0.52	0.58	0.039
Chongqing	0.565	0.472	0.497	0.502	0.608	0.64	0.653	0.675	0.759	0.712	0.841	0.116
Ziyang	0.418	0.503	0.484	0.487	0.502	0.463	0.367	0.455	0.513	0.541	0.609	0.063
Zigong	0.614	0.618	0.62	0.628	0.64	0.633	0.567	0.58	0.561	0.638	0.712	0.042
Mean	0.476	0.507	0.517	0.524	0.538	0.55	0.525	0.548	0.569	0.563	0.625	0.039
S.D.	0.082	0.081	0.076	0.077	0.077	0.084	0.08	0.079	0.094	0.092	0.089	0.006
CV	0.173	0.16	0.147	0.147	0.144	0.153	0.152	0.144	0.165	0.164	0.143	0.010

Note: S.D. is standard deviation; CV is coefficient of variation.

), and a box plot (Figure 3a) was drawn. As shown in the figure, the average value of the TESS-SDL in the Chengyu region from 2011 to 2021 showed a fluctuating increase. On this basis, the standard deviation and coefficient of variation of the regional TESS-SDL in each year were calculated and analyzed (Figure 3b). The standard deviation increased year by year, indicating that the absolute gap between the cities expanded. This meant that some cities were developing a significantly greater TESS-SDL than others. Meanwhile, the coefficient of variation decreased year by year, indicating that although the absolute gap increased, the relative proportion of this gap shrank relative to the overall level of the improvement. This meant that the level of sustainable development in all the cities generally improved and that the overall rate of the progress exceeded the rate at which the gap was widened, further illustrating the unbalanced development of the cities in the region in terms of the TESS-SDL, as well as the overall development progress, emphasizing the need to focus on the rationally tilted distribution of resources and policies for future development.

4.3. Spatial Autocorrelation Analysis

This study analyzed the global autocorrelation level of the TESS-SDL in the Chengyu region using a spatial adjacency matrix, and Global Moran’s I was 0.565 ***, which was significantly autocorrelated. Subsequently, local Moran’s I was employed to assess the similarity or difference of the research attributes between each city and its adjacent cities in the Chengyu region, and scatter plots were generated (Figure 4) to reveal the agglomeration characteristics of each regional spatial unit.

Local Moran’s I was used to test the spatial autocorrelation of the regional TESS-SDL. After the homogenization analysis of the time-series results, it was found that the spatial distributions of Meishan, Suining, and Ya’an had an H-H agglomeration, indicating that their TESS status was better and that they could be maintained and upgraded as a whole. Policymakers need to pay attention to these areas and strengthen the promotion of cooperation among them in order to maintain and expand the scope of high-level areas. The spatial distributions of Dazhou, Mianyang, Nanchong, and Yibin had an L-L agglomeration because low-quality areas were surrounded by other low-quality areas; these cities need more attention and resources to improve the overall sustainable development levels of the regional TESS. Chongqing, Chengdu, Leshan, and Zigong showed an H-L agglomeration, which was characterized by high-value regional areas surrounded by low-value areas. Chongqing and Chengdu, as the core cities in the region, were mainly characterized by their own extremely high levels of TESS, and they had siphoning effects on the surrounding areas, leading to low levels in the surrounding areas. Deyang, Guang’an, Luzhou, Neijiang, and Ziyang exhibited an L-H agglomeration, manifesting itself as low-value areas surrounded by high-value areas, and these cities were particularly characterized by their proximities to Chengdu and Chongqing.

Through this analysis, it was found that Meishan, Suining, and Ya’an were in the high-value TESS agglomeration area, and policymakers can treat them as sub-core cities to assist the core cities (Chongqing and Chengdu) in playing driving roles in the synergistic development of the tourism industry and eco-security. By analyzing the H-L and L-H cities, we found that the radiation-driven effect of the core cities of Chongqing and Chengdu on the neighboring cities was obviously insufficient. As the core cities of the Chengyu region, Chongqing and Chengdu have rich tourism resources and a mature tourism industry, attracting a large number of tourists and resulting in a rapidly developing tourism industry. Compared with Chongqing and Chengdu, the neighboring cities are lagging behind in terms of the economic development, tourism industry development, and protection of the ecological environment. Because of limitations in transportation, infrastructure, and marketing, it is difficult for neighboring cities to fully benefit from the development of the tourism industry in the core cities. This leads to a lack of synergy between eco-security and tourism industry development, affecting the sustainable development of this region. The regional cooperation mechanisms in the Chengyu region, in terms of tourism industry development and eco-security, may not yet have been perfected. This leads to cities working separately in developing their tourism industries and eco-security, demonstrating a lack of overall planning and synergy.

4.4. Spatial Differentiation Characteristics of the Regional TESS-SDL

According to the principles of traditional Markov chains, this study divided the sustainable development levels of the TESS in the Chengyu region from 2011 to 2021 into five types: “disordered”, “less coordinated”, “critically coordinated”, “more coordinated”, and “coordinated”. The principles of traditional Markov chains were used to construct a fifth-order traditional Markov transfer probability matrix (

Table 4. Probability matrix of state transfers (k = 5).

t/t + 1	n	I	II	III	IV	V
I	69	0.594	0.333	0.072	0.000	0.000
II	66	0.227	0.470	0.288	0.015	0.000
III	64	0.016	0.141	0.484	0.344	0.016
IV	68	0.000	0.044	0.191	0.559	0.206
V	69	0.014	0.029	0.014	0.145	0.797

Note: t represents the initial year; t + 1 represents the year after the initial year; n represents the number of cities corresponding to the type of the coupled harmonization; and I, II, III, IV, and V represent dissonance, less coordination, critical coordination, more coordination, and coordination, respectively.

), in which the diagonal line represents maintaining the original level of the coupled coordination; the lower left of the diagonal represents the case of degradation or a jump in degradation, and the upper right of the diagonal is the case of enhancement or a jump in enhancement.

4.4.1. Path Analysis of State Transfers Based on Traditional Markov Chains

(1)

The TESS-SDL, in the Chengyu region, needed to maintain the proportion of the original state is larger, indicating that the system is more stable, regardless of the kind of state transfer the TESS undergoes; the probability for falling on the diagonal is significantly greater than the probability for falling elsewhere;

(2)

In Table 4, the self-locking probabilities of state types I, II, III, IV, and V in the process of the state transfer of the sustainable development level of the regional TESS were 0.594, 0.470, 0.484, 0.559, and 0.797, respectively, indicating that the regional TESS was the most stable in state V, and once it reached this state, the system had the greatest possibility of staying in it. The stability of state I was also higher. In contrast, the self-locking probabilities of the intermediate states (II, III, and IV) were lower, implying that these states were less persistent, and the system was more prone to transfer in these states. Therefore, it can be hypothesized that the system’s dynamics changed more frequently in intermediate states and exhibited greater stability and irreversibility in extreme states;

(3)

Outside the diagonal of the transfer matrix, the transfer power of the TESS-SDL of the cities in the Chengyu region was insufficient, and the mean values of the state elevation (a low-level state transferring to a high-level state) and state regression (falling back from a high-level state to a low-level state) were 8.2% and 12.7%, respectively; the state elevation included continuous elevation and jumping elevation, and the state regression included continuous regression and jumping regression. The level of the state transfer of the sustainable development levels of the TESS in the Chengyu region was greater than the level of the jumping state changes, and continuous regression was more common than continuous elevation. The greater downward compatibility of the regional TESS than the upward compatibility suggested that the regional TESS was more likely to deteriorate and be difficult to recover in the face of unfavorable external conditions, a finding that underscores the importance of preventive and restorative measures.

4.4.2. Path Analysis of State Transfer Based on Spatial Markov Chains

Through spatial correlation analysis, the sustainable development level of the TESS in the Chengyu region was affected not only by its own endogenous factors but also by the influence of neighboring regions. In this study, spatial factors were added to the traditional Markov transfer probability matrix, and the spatial Markov state transfer conditions under the influence of neighboring spatial units were considered to further explore the law of the evolution of the sustainable development of the TESS.

Table 5. Matrix of the probability of spatial Markov state transfers.

Neighborhood Type	t\t + 1	n	I	II	III	IV	V
I	I	37	0.730	0.189	0.081	0.000	0.000
	II	11	0.364	0.455	0.182	0.000	0.000
	III	6	0.167	0.500	0.167	0.167	0.000
	IV	1	0.000	0.000	0.000	0.000	1.000
	V	1	0.000	0.000	0.000	1.000	0.000
II	I	25	0.480	0.520	0.000	0.000	0.000
	II	33	0.242	0.545	0.212	0.000	0.000
	III	12	0.000	0.167	0.583	0.167	0.083
	IV	2	0.000	0.000	0.000	0.500	0.500
	V	6	0.000	0.000	0.000	0.000	1.000
III	I	6	0.167	0.500	0.333	0.000	0.000
	II	14	0.214	0.357	0.357	0.071	0.000
	III	11	0.000	0.091	0.545	0.364	0.000
	IV	6	0.000	0.000	0.000	0.500	0.500
	V	6	0.000	0.000	0.000	0.167	0.833
IV	I	1	1.000	0.000	0.000	0.000	0.000
	II	5	0.000	0.200	0.800	0.000	0.000
	III	25	0.000	0.040	0.480	0.480	0.000
	IV	34	0.000	0.000	0.176	0.618	0.206
	V	36	0.028	0.000	0.000	0.056	0.917
V	I	0	0.000	0.000	0.000	0.000	0.000
	II	3	0.000	0.667	0.333	0.000	0.000
	III	10	0.000	0.200	0.500	0.300	0.000
	IV	25	0.000	0.120	0.280	0.520	0.080
	V	20	0.000	0.100	0.050	0.300	0.550

shows that the spatial Markov process of the regional TESS-SDL had the following characteristics in addition to those found with the traditional Markov chain: (1) The TESS-SDL in the Chengyu region was not spatially isolated, and it was affected by the surrounding spatial units. Spatial environments with different types of TESS-SDLs had different spillover effects on neighboring cities. When neighboring cities had lower TESS-SDLs, the probability of the cities shifting to a higher systemic sustainability stage was lower: P_IIIII/I = 0.182, P_IIIII/II = 0.212, P_IIIII/III = 0.357, P_IIIII/IV = 0.800, and P_IIIII/V = 0.333; P_IIIII/I < P_IIIII/II < P_IIIII/III < P_IIIII/IV; P_IIIIV/I = 0.167, P_IIIIV/II = 0.167, P_IIIIV/III = 0.364, P_IIIIV/IV = 0.48, and P_IIIIV/V = 0; P_IIIIV/I < P_IIIIV/II < P_IIIIV/III < P_IIIV/IV. When the neighboring cities had a higher level of systemic sustainability, the probability that a city would enter a higher stage was higher. (2) Based on studying the spatial transfer probability using the spatial Markov chain in the context of spatial units with the same sustainability, the impact of the upward or downward transfer of the sustainability status was asymmetric. In the Chengyu region, when the neighboring cities had a certain TESS-SDL, the probability of a downward shift (8.5%) was greater than the probability of an upward shift (1.4%). This indicated that the effect for demonstrating the TESS-SDL in regional high-level cities on neighboring regions was significant but obviously insufficient, and the relevant work at this stage failed to positively promote the regional TESS-SDL.

4.5. Analysis of the Driving Factors of the TESS-SDL

In the context of China’s pursuit of sustainable economic growth and environmental management, in order to comprehensively improve the level of ecological security for tourism in the Chengyu region and promote the high-quality and sustainable development of the tourism industry, it is necessary to further explore the driving factors affecting the ecological security for tourism in the region. This study introduced a spatiotemporal geographically weighted regression (GTWR) model to investigate the driving factors. GTWR not only reveals spatial heterogeneity but also captures changes in the temporal dimension, thus providing a dynamic perspective for analysis. By considering both temporal and spatial influences, GTWR is able to provide more detailed and dynamic explanations, contributing to a deeper understanding of the spatiotemporal characteristics of drivers.

4.5.1. Data Verification

During the analysis of the drivers of the TESS-SDL using the GTWR model, it was necessary to standardize all the variables in order to avoid pseudo-regression. To avoid multicollinearity, multiple covariance tests were performed, and variables with a variance inflation factor (VIF) of greater than 10 were excluded (

Table 6. Covariance analysis of the variables.

Variable	Sdl	Tech	Inn	Open	Gre	Ind	Eco	Inf	Tou	VIF	Tolerance
Sdl	1									1.877	0.533
Tech	0.582 **	1								2.762	0.362
Inn	0.289 **	0.425 **	1							5.67	0.176
Open	0.342 **	0.493 **	0.256 **	1						3.927	0.255
Gre	0.195 **	0.305 **	0.470 **	0.434 **	1					1.463	0.684
Ind	0.001	0.163 **	0.835 **	0.195 **	0.169 **	1				12.561	0.079
Eco	0.626 **	0.715 **	0.573 **	0.514 **	0.368 **	0.012	1			3.364	0.297
Inf	0.069	0.355 **	0.249 **	0.247 **	0.091	0.154 **	0.201 **	1		12.291	0.081
Tou	0.424 **	0.513 **	0.659 **	0.781 **	0.539 **	0.082	0.837 **	0.252 **	1	4.981	0.201

Note: ** p < 0.01.

); finally, six variables (the information technology level, technological innovation level, openness level, green development policy level, economic development level, and tourism development level) were considered as the explanatory variables of the GTWR model.

Table A2. Related parameters of GTWR.

Parameter	Bandwidth	Sigma	Residual Squares	AICc	R2	Adjusted R2	Spatiotemporal Distance Ratio
Value	0.1106	0.0333	0.3901	−1182.15	0.8324	0.8290	2.1135

shows the relevant parameters of the GTWR model. In terms of the goodness of fit, both R² and adjusted R² were higher than 0.8, indicating that this GTWR model was able to measure the effects of the explanatory variables on the dependent variables well.

4.5.2. Analysis of the GTWR Results

The regression analysis using the GTWR model and the statistics of the regression coefficients of the drivers of the TESS-SDL in each city from 2011 to 2021 are shown in

Table A3. Regression coefficient statistics of regional GTWR (2011–2021).

	Information Technology					Technological Innovation
	SUNE	Mean	Std. Dev.	Min.	Max.	SUNE	Mean	Std. Dev.	Min.	Max.
2011	3	0.061	0.050	−0.036	0.122	11	0.000	0.102	−0.089	0.300
2012	4	0.021	0.048	−0.085	0.069	10	−0.008	0.051	−0.053	0.131
2013	7	−0.013	0.050	−0.114	0.047	1	0.049	0.057	−0.009	0.170
2014	9	−0.013	0.036	−0.066	0.049	0	0.064	0.058	0.003	0.202
2015	8	−0.006	0.021	−0.046	0.020	4	0.049	0.058	−0.011	0.194
2016	13	−0.007	0.036	−0.049	0.066	8	0.014	0.068	−0.069	0.169
2017	12	−0.021	0.039	−0.072	0.061	12	−0.043	0.087	−0.182	0.126
2018	15	−0.052	0.032	−0.111	0.004	14	−0.118	0.109	−0.315	0.077
2019	16	−0.099	0.027	−0.145	−0.051	16	−0.189	0.114	−0.398	−0.008
2020	16	−0.154	0.037	−0.207	−0.061	16	−0.234	0.101	−0.428	−0.072
2021	16	−0.204	0.054	−0.271	−0.063	16	−0.249	0.103	−0.430	−0.088
Mean		−0.044				Mean	−0.060
	Open to the Outside World					Green Development Environment
	SUNE	Mean	Std. Dev.	Min.	Max.	SUNE	Mean	Std. Dev.	Min.	Max.
2011	16	−0.113	0.052	−0.204	−0.024	7	0.065	0.302	−0.289	0.597
2012	14	−0.042	0.039	−0.118	0.033	9	0.076	0.277	−0.455	0.390
2013	9	0.007	0.037	−0.041	0.084	4	0.206	0.361	−0.945	0.355
2014	2	0.034	0.042	−0.022	0.123	5	0.216	0.292	−0.805	0.303
2015	2	0.057	0.052	−0.011	0.165	6	0.152	0.194	−0.530	0.177
2016	0	0.091	0.068	−0.002	0.224	7	0.102	0.122	−0.311	0.115
2017	0	0.146	0.091	0.023	0.325	8	0.072	0.086	−0.205	0.101
2018	0	0.226	0.116	0.075	0.454	9	0.041	0.077	−0.186	0.105
2019	0	0.311	0.120	0.148	0.544	10	0.011	0.077	−0.151	0.112
2020	0	0.379	0.102	0.244	0.589	11	−0.007	0.069	−0.100	0.126
2021	0	0.425	0.090	0.248	0.603	12	−0.010	0.060	−0.088	0.122
Mean		0.138				Mean	0.084
	Economic Development					Tourism Development
	SUNE	Mean	Std. Dev.	Min.	Max.	SUNE	Mean	Std. Dev.	Min.	Max.
2011	0	0.314	0.152	0.116	0.626	2	0.164	0.121	−0.164	0.251
2012	0	0.167	0.092	0.024	0.330	1	0.128	0.074	−0.092	0.190
2013	3	0.061	0.059	−0.049	0.148	1	0.060	0.039	−0.051	0.124
2014	5	0.020	0.062	−0.093	0.124	4	0.011	0.030	−0.067	0.043
2015	7	0.004	0.075	−0.117	0.149	6	−0.015	0.038	−0.092	0.023
2016	10	−0.006	0.103	−0.122	0.299	9	−0.066	0.046	−0.111	0.016
2017	11	−0.013	0.118	−0.115	0.359	8	−0.075	0.046	−0.111	0.020
2018	12	−0.023	0.120	−0.107	0.359	9	−0.081	0.040	−0.103	0.021
2019	13	−0.041	0.115	−0.123	0.325	11	−0.091	0.033	−0.096	0.013
2020	13	−0.073	0.107	−0.158	0.264	14	−0.127	0.031	−0.098	0.009
2021	14	−0.114	0.101	−0.210	0.187	14	−0.139	0.037	−0.112	0.021
Mean		0.027				Mean	−0.021

Note: SUNE is spatial units of negative effects.

, and Figure 5 was plotted based on Table A3.

Figure 5a shows the multi-year variation curve of the mean value of the GTWR coefficients of each driver of the TESS-SDL. The driving factors for which the mean of the regression coefficients for the TESS-SDL has a positive effect include open-door and green development policies, reflecting the benefits of international integration and environmental protection efforts. Negative trends in the changes in the regression coefficients of the driving factors included influences from the level of information technology, the level of economic development, and the level of tourism development, which showed increasingly negative impacts, indicating that these factors may jeopardize the sustainable development of the regional tourism eco-safety system if there is no proper regulation in line with sustainable practices. The curve of the mean change in the regression coefficient of the technological innovation level showed a fluctuating trend; the level of technological innovation initially had a positive impact, but as time passed, its impact became negative. The impact of technological innovation on the TESS-SDL was multifaceted; the initial period demonstrated positive effects, but with the expansion of the application of the technology and the lagging of environmental management measures, the negative effects gradually appeared and were enhanced.

Figure 5b demonstrates the multi-year movement of the standard deviation of the GTWR coefficients for the drivers of the TESS-SDL, representing the change in the variability of the GTWR coefficients of the drivers in each year. The factors for which the mean standard deviation of the regression coefficients decreased year by year included the information technology level, the green development policy level, and the tourism development level. This indicated that the difference in their impacts’ effects on the TESS-SDL gradually weakened, reflecting the achievements of the region in the levels of information technology, policy support for green development, coordinated development of tourism, consistency in the implementation of policies, and the overall balanced improvement. The factors for which the standard deviation of the regression coefficients increased year by year included the level of technological innovation, the level of opening up to the outside world, and the level of economic development. The gradual increase in the variability of the regression coefficients of the drivers for the TESS-SDL reflected inconsistencies in the policy implementation, resource utilization, attractiveness, and development strategies of the regions in terms of the level of technological innovation, the level of opening up to the outside world, and the level of economic development, leading to imbalanced inter-regional development.

Figure 5c,d show the multi-year curves for the maximum and minimum values of the GTWR coefficients for the drivers of the TESS-SDL, respectively. A comparison showed that the multi-year trends in the maximum and minimum values of the regression coefficients for each driver were categorized into consistent and inconsistent changes. Consistent changes were found for the information technology level, the technological innovation level, the level of opening up to the outside world, and the tourism development level, while inconsistent changes were found for green development policy support and the economic development level, which was mainly characterized by inconsistency in the change rate. It is worth emphasizing that if the change trends of the maximum and minimum values were consistent, this meant that the overall development trend of the driving factor, whether upward or downward, was clear, and the direction of the change in this indicator was consistent across the cities. Inconsistency suggested that there was a significant imbalance in the development of the drivers among the cities and that some cities may have grown rapidly while others were declining or stagnating.

4.5.3. Recommendations Based on the GTWR Results

(1)

The rational application of information technology in the tourism industry should be strengthened to avoid overreliance on the short-term tourism boom brought about by network exposure; in addition, it is necessary to formulate long-term planning, focus on sustainable development, and avoid wasting resources. At the same time, intelligent tourism management systems should be promoted to optimize the allocation of tourism resources and improve management efficiency through the use of big data analysis and artificial intelligence technology;

(2)

While promoting technological innovation, it is necessary to consider its long-term impact, avoid negative effects caused by the expansion of technological applications and the lagging behind of environmental governance measures, and strengthen technological innovation and environmental impact assessment regulations. The application of green technology in tourism should be encouraged and supported to improve ecological protection and reduce the negative impact of technological innovation on the environment;

(3)

Green development policies should be further improved to reduce the negative impacts of high-cost inputs and restrictions on economic activities and promote the adaptation of enterprises and communities through policy incentives. At the same time, environmental protection education should be strengthened to enhance public awareness of environmental protection, encourage public participation in green tourism, and promote the effective implementation of green policies;

(4)

While raising the level of openness to the outside world, attention should be paid to environmental protection and resource management, tourism should be promoted by attracting international tourists and foreign investment, and tourism facilities and services should be upgraded. While promoting economic development, attention should be paid to the sustainable use of resources and environmental protection, and scientific and reasonable development plans should be formulated to avoid the overutilization of resources and environmental pollution problems;

(5)

Strict norms for the development of tourism should be formulated and enforced to avoid the negative impacts of irregular development, and tourism enterprises should be encouraged to adopt a sustainable development model to improve the quality and efficiency of the tourism industry. At the same time, long-term interests should be considered in terms of infrastructure construction to avoid unnecessary construction because of short-term tourism booms, promote the sustainable development of infrastructure, and improve resource utilization efficiency.

5. Discussion

5.1. Main Findings

Taking the Chengyu region as the research object, this study constructed a TESS indicator framework based on the DPSIR model to evaluate the TESS-SDL, explored the dynamic evolution characteristics and driving factors of the TESS-SDL, and proposed policy recommendations. The following findings were obtained:

First, from 2011 to 2021, the TESS-SDL of the Chengyu region was generally at a medium level and showed a trend of steady growth. Although the gap between the cities widened year by year, the improvement rate of the TESS-SDL exceeded the rate at which the gap widened, leading to more consistent regional development directions. The region boosted its overall TESS-SDL through active tourism and economic development policies and investments, especially in core cities, such as Chongqing and Chengdu, by improving infrastructure and environmental governance. The core cities received more support in terms of policies, resources, and technology, while the neighboring cities were relatively slower to develop because of their weak foundations and development levels, leading to increased inter-city disparities; however, the overall pace of development made up for this widening gap. Previous studies [15,42] have discovered similar development trends, indicating that the level of the regional tourism ecological security (TES) is steadily improving; however, interregional development imbalance still exists. Moreover, the change in the TES among the regional cities under the time trend is not the same [42,57] and is mainly influenced by factors such as the level of economic development, strong policy implementation, natural resource conditions, and urban development strategies. Rapidly developing economies, strong policy implementations, abundant natural resources, and scientific management strategies help to enhance TES levels and narrow inter-city gaps; on the contrary, lagging economies, poor policy implementations, lack of resources, and mismanagement may lead to slow improvements or declines in TES levels and widen inter-city gaps. Although this study focuses on evaluating the TESS-SDL, while previous studies mainly concentrated on the TES itself, these results are still comparable because they share a common focus concerning the key factors affecting tourism sustainability and the dynamic changes in the ecological security; therefore, previous studies on the TES can be analogized with this study.

Second, the TESS-SDL had a significant autocorrelation in the Chengyu region, and the core cities had strong siphoning effects on the development of neighboring regions; however, the effect of their leading example on the neighboring cities has not yet been realized, and regional cooperation mechanisms need to be improved. Chongqing and Chengdu have obvious advantages in terms for attracting capital, talent, and technological innovation, which have not effectively diffused to the neighboring cities, resulting in insufficient regional cooperation and resource sharing, affecting green policies and the quality of the infrastructure in the neighboring cities. A comparison of related studies [55] indicates that under the resource agglomeration effect of large cities, the development of neighboring cities faces greater challenges, and more effective regional cooperation mechanisms need to be established to promote balanced development.

Third, according to the Markov process analysis, the state transfer of the TESS-SDL has a self-locking effect; the risk of downward development is greater than the possibility of upward development, and it is significantly influenced by the neighboring cities. Some neighboring cities, because of the poor quality of the infrastructure and the limited level of economic development, easily fall into development dilemmas, and it is difficult to break through the current state of development because they become locked in their current state or undergo downward state transfers; at the same time, the lagging development of these cities also has a certain negative impact on the core cities. This conclusion is verified via comparison with related studies [19,42] in which a lack of vitality in the dynamic evolutionary characteristics of the TES state transfer can be seen, and when spatial factors are taken into account, the state of the TES in neighboring cities affects the probability of the state transfer.

Finally, the analysis of the driving factors further revealed that the average effect of the impacts of open-door and green development policies on the TESS-SDL was positive, while the average effect of the impacts of the level of information technology, economic development, and tourism development had a gradual negative development trend. The average impact of the level of technological innovation went from positive to negative, reflecting the negative effects for expanding technological applications and lagging environmental governance.

The implementation of open-door and green development policies in the Chengdu–Chongqing region promotes international cooperation and environmental protection, enhancing the TESS-SDL level in the region. The level of green development mainly responds to the degree of government support for regional environmental governance and significantly positively promotes the TES level [5,13,15]. Good environmental governance, such as improving the rate of sewage treatment and the comprehensive utilization of solid waste, can effectively enhance the level of ecological security. The positive facilitating effect for opening to the outside world is also supported by relevant studies [15].

Despite the positive impacts of information technology and economic development on the TESS-SDL in the early stages, the burden on the ecosystem was exacerbated over time by the inappropriate application of technology and environmental pressures brought about by economic development, leading to the gradual manifestation of negative effects. The irregular development of tourism has also negatively impacted the environment and increased the ecological pressure. Technological innovation initially promoted the development of the tourism ecological security system, but with the expansion of technology, environmental management measures failed to follow in time, resulting in technological innovations that brought more environmental problems and had negative impacts on the TESS-SDL. Compared with previous studies, this study is significantly different. Most studies assume the positive effects of factors such as tourism development and socioeconomic and technological innovations on the TES [5,13,31]; however, the results of this study show (as shown in Table A3) that over time, these factors show positive effects in the initial stage, which gradually weaken [15], and then have negative effects; the average of the multi-year effects is negative. To analyze the reasons for this key difference, we can consider the different research methodologies used; most previous studies used econometric models or geographic probes, and this study used GTWR for the analysis. Compared with the global linear approach, the GTWR model is able to explore the impacts of drivers on the TES in detail with time and location dimensions, revealing the dynamic changes in drivers across different time periods and different regions and more accurately reflecting the impacts of drivers. In addition, the research location is an important reason for the differences in the results. There are significant differences in the level of economic development, environmental protection policies, and technological innovation capacity across the different regions. Finally, different time periods also affect the results concerning the driving effects.

Information technology as a driver has been relatively less studied because of changes in the characteristics of the times and contents. In recent years, with the rise of social media platforms, the impact of information technology on tourism has become increasingly important. Short videos and online exposure can quickly bring about a short-term tourism boom; however, this phenomenon is often unsustainable and may lead to the overconsumption of resources and increased environmental pressure.

5.2. Application of Innovative Methods

By targeting the Chengyu region, an important strategic region, this study revealed the regional differences and uneven development of the sustainable development levels of its tourism ecological security system, focused on analyzing the spatial development characteristics and their driving factors, and put forward specific policy recommendations. These not only fill the gaps left by existing studies but also provide valuable references for research and practice in other inland urban agglomerations. Through the comprehensive use of multi-dimensional analysis methods, this study provides more precise and systematic insights and offers important theoretical and practical support for promoting the balanced and sustainable development of the Chengyu region and other city clusters.

The application of spatial Markov chains significantly enhanced our understanding of the sustainable development level of the tourism ecological security system (TESS-SDL) in the Chengyu region by integrating geospatial factors. Compared with the traditional Markov chain model, the spatial Markov chain not only captures the spatial dependence among the cities but also reveals the inter-regional evolutionary features, and it has a stronger predictive ability and the ability to identify spatial evolutionary features. This approach provides more precise and effective guidance for policymaking and helps to achieve the goals of balanced development and sustainable development within the region.

Spatiotemporal geographically weighted regression (GTWR) significantly improved our understanding of the driving mechanisms behind the sustainable development levels (TESS-SDLs) of the tourism eco-safety system in the Chengyu region by revealing the spatial heterogeneity of the driving factors and the dynamic trends. Compared with traditional models, GTWR can provide a more precise and systematic analysis, provide strong support for policy formulation, and help to realize the goals of balanced development and sustainable development in the region.

5.3. Limitations and Prospects

5.3.1. Limitations

This study provides a reference for the sustainable development of the tourism ecological security system in the Chengyu region, which consists of tourism activities, the ecological environment, and socioeconomics; however, there are still deficiencies, and subsequent research should continue to deepen and expand upon these findings. Because of the limitations of the data acquisition, indicators such as the tourism reception capacity, tourism development area, and biodiversity index could not be included in the construction of the tourism ecological security system. Despite our efforts to establish a more comprehensive indicator system, the existing indicator system may remain insufficient in fully reflecting the complexity and diversity of the tourism ecological security system. In particular, the existing indicator system may be inadequate in terms for capturing the interactions among tourism activities, the ecological environment, and socioeconomics.

5.3.2. Study Prospects

Considering the insufficiency of the indicator system for measuring the tourism eco-safety system, future research can include some indicators of the correlations between tourism and ecology, tourism and society, and tourism and the economy. To compensate for the difficulty of the data acquisition, the use of geographic data (such as lighting data and satellite remote-sensing data) and survey data can be considered to enhance the accuracy of evaluations with this indicator system. In order to improve the practicality and relevance of the policy recommendations, future research will focus on constructing a model of system dynamics for tourism ecological security systems in order to conduct multi-scenario simulations combining different policies and the basis of urban development to explore the path of sustainable development for regional TESSs.