Mapping the resilience of tourist city: spatial correlation and elasticity of tourism resource value

Table of Contents

Spatial assessment of TRV

TRV can be converted into economic, social, cultural, and ecological benefits for a city. This conversion process exhibits elasticity, with the intrinsic value of tourism resources serving as the foundation for determining its effectiveness. This paper evaluates TRV based on three dimensions: quantity, quality, and diversity. We then analyzes the spatial characteristics of these values and perform cluster analysis using the results from these three dimensions, ultimately identifying five distinct TRV clusters.

Dimensional value assessment

The Getis-Ord G* analysis was employed to investigate the spatial distribution of TRV hotspots across three dimensions. The results revealed that TRV exhibited distinct patterns of spatial agglomeration in each dimension (Fig. 3).

Fig. 3: Distribution of hotspots of TRV.

The figures (a1) and (b1) respectively show the quantity hotspots of Penglai and Qimen; the figures (a2) and (b2) respectively show the quality hotspots of Penglai and Qimen; the figures (a3) and (b3) respectively show the diversity hotspots of Penglai and Qimen.

Quantity dimension. Penglai exhibited “point-axis agglomeration”, with 41% of its tourism resources concentrated within a 5-kilometer northern coastal corridor. Relying on the key attractions such as the Penglai Pavilion, the Eight Immortals Crossing the Sea architectural and cultural landscapes, and the Three Immortal Mountains architectural and cultural landscapes, it has formed hotspots and sub-hotspots agglomeration areas, covering 1.36% of the total area. In contrast, Qimen has developed a “one main and multiple secondary” pattern, with hotspots and sub-hotspots agglomeration areas (1.06% of the total area). Qishan (the county seat) is the primary agglomeration area, relying on the Qihong Cultural Expo Park to form an industrial resource agglomeration.

Quality dimension. In Penglai, part of the quality hotspots and sub-hotspots overlap with quantity-based aggregation areas in the coastal cultural tourism areas (e.g., Penglaige), while another part is located in the ecological industrial zones supported by Aishan National Forest Park and the wine industry chain (e.g., Cunliji, Daliuhang). These hotspots and sub-hotspots cover 11% of the total area. In Qimen, a “quality-quantity” spatial coupling pattern is evident. The hotspots and sub-hotspots aggregation areas exhibit significant spatial overlap with the quantity-based aggregation areas, accounting for ~6% of the total area. This suggests that high-quality resources tend to generate scale economic effects.

Diversity dimension. In Penglai, the area of diversity hotspots and sub-hotspots (accounting for 4.22% of the total) decreased by 62% compared to the quality dimension, reflecting a homogenization trend in resource types. In contrast, Qimen maintained an aggregation scale of 6.43%, demonstrating a more complementary resource combination. These findings indicate that the coastal tourist city’ TRV agglomeration exhibits less significant diversity, predominantly dominated by single-type resources, whereas the mountainous tourist city shows relatively more affluent diversity.

The distribution of tourism resources typically exhibits agglomeration characteristics (Zhang et al., 2023; Han et al., 2024). Across different dimensions, the hotspots and sub-hotspots across all value dimensions have formed distinct aggregation areas, constituting critical zones for converting TRV. With the implementation of supportive policies and market investments, these areas possess significant potential to become the most valuable regions for tourism development. Moreover, the quality agglomeration areas act as the primary carriers for value conversion, with their spatial scale being 8–15 times larger than that of the quantity dimension. This highlights the significant spillover effect of high-quality resources on TRV.

Clusters of TRV

A K-means multivariate cluster analysis was conducted based on the Getis-Ord G* index of TRV’s quantity, quality, and diversity characteristics. In selecting the clustering scheme, this paper combines the Calinski-Harabasz pseudo-F index to validate the clustering validity. The initial global optimal solution is k = 7, but its spatial division shows a fragmented trend, resulting in poor geographical continuity for each category. Therefore, an attempt is made to reduce the number of clusters to achieve a balance between statistical robustness and spatial rationality. Finally, the number of clusters is determined to be 5 (Clusters I–V).The resulting clusters exhibit a gradual increase in TRV from Cluster I to Cluster V, where clusters IV and V represent the advantage clusters, while clusters I-III constitute the general clusters (Fig. 4).

TRV Advantage Clusters. Cluster V demonstrates the highest TRV advantage, with significant quantity, quality, and diversity advantages. The TRV advantage cluster of Penglai is concentrated in the Penglaige, where each grid unit contains an average of ~24 tourism resources across 3 types (the average grade is 2.35); the TRV advantage cluster of Qimen is distributed along the axis of Qishan-Liko, where each grid unit contains an average of ~8 tourism resources across 2 types (the average grade is 1.94). Cluster IV demonstrates the second-highest TRV advantage, primarily distributed around the areas with the highest TRV. In terms of quantity (7.09 per grid in Penglai and 2.22 per grid in Qimen), quality (average grade of 1.72 in Penglai and 1.75 in Qimen), and diversity (1.65 types per grid in Penglai and 1.25 types per grid in Qimen), they exhibit potential value for complementing and supporting the development of Cluster V.

TRV General Clusters. Clusters III-I exhibit less pronounced tourism resource advantages. These clusters’ tourism resources have a smaller scale (<1.18 per unit), lower quality grades (<1.12), and less diverse type combinations (<0.64 per unit). The tourism resources in these areas lack significant clustering effects and are unlikely to generate substantial economic benefits in future market value conversions. However, the inherent ecological and cultural values of these tourism resources can still contribute to maintaining and enhancing the natural and humanistic environment of the tourist city, mitigating the negative impacts of tourism development, and sustaining TCR.

It is worth noting that Penglai and Qimen show different forms of agglomeration. Qimen exhibits a multi-centered pattern, whereas that in Penglai demonstrates a single-centered characteristic. This is because, as a coastal tourist city, Penglai’s core resources are constrained by the rigid limitations of the coastline, thereby forming an irreplaceable resource advantage. Consequently, this leads to the single-pole concentration of resource elements toward Penglaige. In contrast, as an inland tourist city, Qimen’s infrastructure elements, such as major transportation routes, can disrupt spatial resource agglomeration and induce the formation of multiple TRV centers along the infrastructure corridor.

Tourist attractions result from the development and utilization of tourism resources and the conversion of their value into tangible products. All A-level scenic spots are distributed within clusters IV and V, which verifies the value chain transmission mechanism of “resource endowment, product development” (Erislan, 2017). This indicates that, supported by resources, technology, and policies, the TRV advantage cluster can enhance the tourism economic contribution rate (The total tourism revenue of Penglai in 2024 was nearly five times that of Qimen), strengthen cultural revitalization (Penglai has 113 intangible cultural heritage items at or above the county level, compared to Qimen’s 32), and contribute to the construction of TCR.

Spatial characteristics of TCR

Based on the index evaluation system outlined in Table 3, the TCR level was calculated and subsequently visualized spatially using ArcGIS Pro. The natural breaks method was employed to classify the overall resilience index, economic resilience index, social resilience index, cultural resilience index, and ecological resilience index into five categories: high, medium-high, medium, medium-low, and low levels, as illustrated in Fig. 5.

**Fig. 5: Spatial characteristics of dimensional resilience.**

Spatial continuity of overall resilience

By comparing the overall resilience of Penglai and Qimen, it can be observed that both regions exhibit similarities in spatial location and spatial characteristics (Fig. 5a1, b1).

The high-level and medium-high-level areas of overall resilience exhibit a relatively high degree of spatial consistency with the TRV advantage areas. In Penglai, the high-level areas are concentrated in three central townships along the northern coast: Penglaige, Zijingshan, and Dengzhou, covering a total area of 3 km². The medium-high-level areas primarily surround the high-level areas, forming an aggregation area centered on Penglaige along the northern coast. In Qimen, the high-level and medium-high-level areas of overall resilience are relatively small and concentrated within Qishan, where the medium-high-level areas surround the high-level areas to form an aggregation area.

The overall resilience level exhibits a typical spatial continuity, with high-level aggregation areas as the central focus. In Penglai, the resilience level gradually decreases from the northern coastal region toward the south, resulting in a spatial distribution characterized by higher values in the north and lower values in the south. In Qimen, the resilience level diminishes radially outward from Qishan, leading to a spatial distribution marked by higher values in the center and lower values in the surrounding areas. The distribution patterns of TCR in both Penglai and Qimen, along with the “high in the middle, low around” spatial pattern observed for red tourist cities in southeastern China by Zha et al. (2024), collectively reflect the spatial continuity feature of TCR.

However, significant differences exist in the spatial patterns of resilience levels between Penglai and Qimen. Specifically, the resilience level of Qimen exhibits a network-like spatial pattern, whereas that of Penglai demonstrates a surface-like distribution. These spatial pattern differences are shaped by the spatial characteristics of TRV advantage clusters. In Penglai, the high degree of spatial monopoly in tourism resources drives the concentration of resilience-related elements, such as tourism service facilities and socio-economic conditions, toward the core area, forming a surface-like aggregation pattern. Conversely, in Qimen, the multi-centered clustering of TRV advantages promotes the aggregation of resilience elements around each center and enhances the resilience levels of the regions connecting these centers, leading to a network-like distribution structure.

Spatial heterogeneity of dimensional resilience

Economic, social, cultural, and ecological resilience levels exhibit significant spatial heterogeneity.

Economic resilience. While sharing some similarities with overall resilience, the high-level economic resilience areas are more extensive. In Penglai, high-level areas are distributed in a surface-like pattern across four central townships along the northern coastline—Penglaige, Zijingshan, Dengzhou, and Xingang—with a total area of 15.75 km². In Qimen, high-level areas are concentrated in Qishan, covering an area of 2.74 km². The medium-to-high-level areas in both cities predominantly surround the high-level areas, spanning a total area of 50.75 km² in Penglai and 15.5 km² in Qimen (Fig. 5a2, b2).

Social resilience. The high-level and medium-high-level areas exhibit significant clustering, forming small-scale aggregation areas. In Penglai, these areas are concentrated in the three northern central townships with the highest levels of human activity: Penglaige, Zijingshan, and Dengzhou. In Qimen, the high-level and medium-high-level areas are primarily located in Qishan, showing spatial overlap with the high-level areas of overall resilience and economic resilience (Fig. 5a3, b3).

Cultural resilience. The number of high-level and medium-high-level cultural resilience areas is limited, lacking prominent aggregation characteristics. In Penglai, High-level areas are scattered across Penglaige, Liujiagou, and Beigou, covering a total area of 1.5 km². In Qimen, the high-level areas are dispersed in Qishan, Liko, and Xin’an, with a total area of 1 km². Medium-high-level areas are generally adjacent to these high-level areas, covering a total area of 1.5 km² in Penglai and 2.25 km² in Qimen (Fig. 5a4, b4).

Ecological resilience. The overall level of ecological resilience is relatively high, and its spatial distribution pattern differs significantly from those of other resilience dimensions. In Penglai, High-level and medium-high-level areas are predominantly distributed in a band along the northern coastline and the southern Aigu Mountain range, covering 29.25 km² (High-level) and 106.25 km² (medium-high-level). In Qimen, high-level areas are primarily located along the southwestern and southeastern boundaries, covering a total area of 20.75 km². Medium-high-level areas are predominantly clustered around these high-level areas, extending an additional 82.25 km² (Fig. 5a5, b5).

Spatial correlation between TRV and TCR

Overall correlation

As previously discussed, there is a notable correlation between the advantages of TRV and the effectiveness of its conversion. The TRV clusters I-V exhibit significant value gradient disparities. By statistically analyzing the number of resilience grids at each level within each cluster and calculating the average resilience level for each cluster, the relationship between TRV clusters and TCR can be elucidated (Fig. 6a1, b1). It is observed that an enhancement in TRV corresponds to an elastic response in the overall TCR level.

Regarding individual dimensions, economic resilience exhibits the most substantial response sensitivity. This finding validates the central role of the tourism multiplier effect generated by TRV in promoting urban employment growth and optimizing the industrial structure. Social resilience demonstrates a moderate level of responsiveness, with the conversion of TRV contributing to enhancing social resilience in tourist cities. In contrast, cultural and ecological resilience levels across clusters are generally low, and the two forms of resilience show no significant correlation with TRV clusters. This reflects challenges such as inadequate activation and utilization of intangible cultural heritage and increased environmental governance costs associated with tourism resource development.

Dimensional correlation

By analyzing the proportion of grids with varying resilience levels across different grades of hotspots, the coupling and correlation characteristics between the three elements of TRV and the four dimensions of TCR were deconstructed. We found that the responses of TCR in different dimensions to TRV showed significant differences.

**Fig. 6: Spatial correlation between TRV and TCR.**

Economic resilience (Fig. 6a2, b2) exhibits a monotonically increasing trend with the enhancement of hotspots in tourism resource quantity, quality, and diversity. This indicates regions with higher TRV exhibit substantial development potential and can enhance economic resilience by developing upstream and downstream tourism-related industries. This finding validates the value-added principle of the tourism economic system: high-value tourism resources create an element network through industrial chain extension, thereby establishing a positive feedback mechanism of “resource endowment, industrial agglomeration, resilience strengthening”.

Social resilience (Fig. 6a3, b3) continuously correlates with the quality and diversity of tourism resources. There are notable differences in the quantity dimension between the two case studies: In Qimen, a significant correlation is observed, whereas in Penglai, an inverted U-shaped relationship is evident. This suggests that while social resilience is strongly influenced by the quality and diversity of tourism resources, its relationship with quantity is more intricate. Specifically, when tourism resource quality surpasses a certain threshold, the spatial monopoly of high-quality tourism resources may result in the polarized distribution of infrastructure. This phenomenon is particularly pronounced in coastal tourist destinations, thereby corroborating theory of tourism spatial deprivation (Smith, 1990).

Cultural resilience (Fig. 6a4, b4) exhibits a continuous positive correlation with the quality and diversity of tourism resources. However, it shows significant polarization with the increase in the grade of hotspots for tourism resource quantity, where both high and low levels exhibit a relatively high proportion of quantity hotspots. This suggests a significant relationship between cultural resilience and the quality and diversity of tourism resources, but not with their quantity. This phenomenon can be attributed to the interaction between cultural and tourism resources. On the one hand, tourism resources derived from cultural heritage tend to have higher quality. On the other hand, cultural venues are often situated in areas with tourism resource advantages, particularly those characterized by high quality and diversity, leading to the continuous enhancement of cultural resilience.

Ecological resilience (Fig. 6a5, b5) exhibits a continuous positive correlation with the diversity of tourism resources, but does not show a significant upward trend with increases in quantity. Differences in the quality dimension are observed between the two case areas: in Penglai, there is no apparent correlation between ecological resilience and the quality of tourism resources, whereas in Qimen, a notable correlation existed. This suggests that the diversity of tourism resources positively influences the enhancement of ecological resilience in cities.

It is evident that the correlation patterns between economic resilience, cultural resilience, and TRV in Penglai and Qimen are consistent. However, compared to Penglai, Qimen exhibits more pronounced correlations between social resilience and the quantity characteristics of TRV, as well as between ecological resilience and the quality characteristics of TRV. Regarding social resilience, the monopolistic advantage of coastal tourism resources in Penglai causes infrastructure agglomeration to depend on resource quality and combination rather than merely on resource quantity. In contrast, the distributed tourism resource nodes in Qimen activate community networks, establishing a strong correlation mechanism between resource quantity and the enhancement of social resilience. Concerning ecological resilience, the high-intensity development of core scenic areas in Penglai, constrained by the linear distribution of high-quality tourism resources along the coast, compresses ecological buffer zones, leading to a weak correlation between improvements in ecological resilience and tourism resource quality. By contrast, in Qimen, tourism resources are distributed along natural corridors formed by water systems and mountain ranges, where the ecological background and the realization of TRV form a spatially nested structure.

Overall, the difference in resilience response between Penglai and Qimen originates from their distinct distribution patterns of tourism resources. The single-core radiation structure in Penglai reinforces a quality-dominant mechanism, whereas the multi-core network layout in Qimen stimulates a quantity-driven value-added effect. This spatial heterogeneity suggests that enhancing TCR requires constructing models tailored to local natural and humanistic environments, rather than merely transplanting universal development paradigms.