Introduction

The urbanization rate in China has increased from 19.4% in 1980 to 65.2% by 2022, making it the country with the fastest urbanization process in the world since the inception of the reform and opening-up policy in 19781. The impact of this transition on Chinese economic and social structure has been significant2. Tourism urbanization (TU), driven by economic growth and population expansion, has emerged as a significant force in shaping regional development3,4. While it promotes economic diversification and urban expansion, it simultaneously exerts both positive and negative influences on ecosystem services, impacting biodiversity, carbon sequestration, and water resources5. Tourism urbanization facilitates the transformation of land use through the rapid expansion of infrastructure, leading to increased pressure on natural ecosystems6,7. However, the resulting ecological stress often disrupts the balance between urban expansion and ecosystem sustainability, triggering issues such as habitat fragmentation and environmental degradation8. Critically, existing studies predominantly address TU’s economic or ecological impacts in isolation9, leaving a gap in systemic frameworks to balance these dual objectives—particularly in ecologically vulnerable basins like the YRB. This study fills this void by developing a Coupling Coordination Degree (CCD) model that quantifies the dynamic equilibrium between TU and ecosystem services (ESs). This approach not only advances the ‘sustainable tourism’ paradigm10 but also operationalizes the UN SDGs by aligning urbanization pathways with ES resilience in the YRB, a cornerstone of China’s ecological security.

Ecosystem services (ESs) encompass the diverse products that natural ecosystems offer to humans as well as the regulatory and protective functions of ecosystems on the environment11,12. As a bridge and link connecting natural ecosystems and the socio-economic system, ESs have become essential indicators for assessing the sustainable high-quality development and health of ecosystems since the Millennium Ecosystem Assessment13. Nevertheless, preceding studies primarily concentrated on evaluating the value and provisioning capacity of ESs14,15. The balance and synergy of ESs, the heterogeneity of internal relationships in ESs, and factors influencing the alignment of ESs supply and demand have attracted widespread attention from scholars domestically as well as abroad at present16,17. Furthermore, the efficiency and intensity of resource utilization in different locations are mostly influenced by tourism urbanization, especially over-tourism in coastal and ecologically fragile areas, which exacerbates the spatiotemporal dislocation of ESs18.

The coupling coordination relationship between tourism urbanization and ecosystem services has progressively emerged as an academic focal point in recent years19. Coupling Coordination Degree (CCD) refers to the extent to which two or more systems achieve synergistic interactions through coordinated mechanisms during their mutual evolution20,21. The CCD model has been widely adopted to evaluate their interactive dynamics, providing critical theoretical foundations and practical guidance for regional sustainable development. In the context of tourism urbanization and ecosystem services, CCD quantifies the synergistic effects between tourism-driven economic development and ecological conservation. Studies demonstrate that a high CCD signifies efficient mutual promotion between tourism urbanization and ecosystem services, whereas a low CCD reflects contradictions or imbalances between these systems22.

Existing research on the coupling relationship between tourism urbanization and ecosystem services often focuses on individual regions23,24,25 or specific urban agglomerations26,27 rather than urban agglomerations in river basins. Urban agglomerations, as complex systems involving multiple interacting cities, require specific analysis to understand the integrated effects of tourism urbanization and ecosystem services. This study fills this gap by evaluating the coupling coordination degree (CCD) between TU and ES across urban agglomerations in the Yellow River Basin (YRB).

The impact mechanisms of tourism urbanization on ecosystem services have been explored in various regions, but findings are often context-dependent28. For instance, tourism urbanization accelerates the transformation of natural landscapes into urbanized areas, affecting water regulation, carbon storage, and habitat quality29,30,31. On one hand, the rapid conversion of natural ecosystems weakens ecosystem functions, disrupting biodiversity and increasing environmental risks such as soil erosion and desertification32,33. On the other hand, intensified tourism activities elevate the demand for ecosystem services, leading to spatial mismatches between supply and demand, particularly in ecologically fragile areas34,35.

In the Yellow River Basin, the unique ecological and geographical characteristics necessitate a more nuanced understanding of these impacts, which remains underexplored36,37. Due to regional variations in climate and socio-economic factors, the efficiency and intensity of resource utilization differ significantly across the basin. For example, the arid upstream Loess Plateau (annual precipitation < 400 mm)38 faces stringent water constraints that limit conventional tourism infrastructure39, while the flood-prone downstream Shandong Peninsula struggles with seasonal habitat degradation40. Socio-economically, high-income urban clusters like Zhengzhou harness advanced logistics networks41 to develop eco-tourism, whereas pastoral regions (e.g., Inner Mongolia) prioritize low-impact cultural tourism akin42 to the Yangtze Delta’s ‘green villages’43. These disparities underscore the need for adaptive strategies that reconcile TU growth with localized ecological thresholds—a gap this CCD model directly addresses.

The ecosystem of the YRB is fragile and highly susceptible to anthropogenic pressures, including tourism urbanization44. Despite its ecological significance, the interactions between tourism urbanization and ESs remain insufficiently understood, particularly regarding their spatial and temporal variations45. Most existing studies focus on single-factor analyses, overlooking the complex feedback loops between urbanization and ecosystem health. Moreover, systematic assessments of tourism-related land-use transformations and their long-term effects on ESs are lacking46. A more integrated approach is needed to capture the multidimensional impacts of tourism urbanization on ecosystem sustainability15.

This study aims to systematically assess the impact of tourism urbanization on ecosystem services in the YRB through spatiotemporal analysis, revealing its dynamic characteristics and impact mechanisms. The research addresses three core questions: (1) How has the coupling coordination degree between tourism urbanization and ecosystem services evolved spatially and temporally? (2) What are the key drivers influencing their interaction? (3) What policy implications can be derived for sustainable development?

The InVEST model and the Coupling Coordination Degrees (CCD) across the YRB provide a spatially explicit assessment of ecosystem service trade-offs and synergies. Additionally, ecosystem service valuation (ESV) is employed to evaluate both biophysical and socio-economic drivers of ES change47.

China’s national strategy, the “Ecological Protection and High-quality Development of the Yellow River Basin,” calls for a balanced approach to urbanization and ecological conservation. By addressing the spatiotemporal dynamics and impact mechanisms, this study offers practical insights for regional ecological management and sustainable tourism development. The findings provide a scientific basis for spatially adaptive governance and sustainable urban planning in the Yellow River Basin. This study directly informs policymakers by providing empirical evidence on how tourism urbanization can be harmonized with ecosystem services (ES) to achieve regional sustainability goals.

Methods and data

Research area

The Yellow River Basin (YRB) (Fig. 1), spanning approximately 795,000 km2 across four geomorphic units (Qinghai-Tibet Plateau, Inner Mongolia Plateau, Loess Plateau, and North China Plain), is a critical ecological and economic corridor in China (Fig. 1a)48. Characterized by three distinct topographic terraces (Fig. 1b), this region includes the natural ecological corridor and several important functional ecological zones including the Ruoergai, Qilian Mountains and Three Rivers Source46.

Fig. 1
figure 1

Geographical location, elevation, city clusters, and land use of the YRB.

Full size image

Note(a) geographical location of YRB in China; (b) elevation; (c) Yellow River Basin city cluster; (d) Land use and cover in the study area in 2020, illustrating the different land use and cover including grassland, forest, cropland, water bodies, construction land, and barren land. All graphs presented in this study were created by the author using ArcGIS 10.8, developed by the Environmental Systems Research Institute (ESRI) (https://www.esri.com/).

As China’s second-largest economic belt, the YRB hosts seven national-level city clusters (Fig. 1c), including the Guanzhong Plain and Lan-Xi clusters, which collectively account for 63.3% urbanization rate and 412.9 billion CNY tourism revenue (2021)46.

The YRB exhibits significant climatic and land-use heterogeneity (Fig. 1d). (1) Climate: Annual precipitation ranges from 300 mm (northwest) to 600 mm (southeast), with 70% concentrated in July–August29. (2) Land use: Dominated by grassland (42%), cropland (28%), and forest (18%), while construction land (7%) is clustered in downstream urban clusters (e.g., Shandong Peninsula)47. Rapid urbanization and tourism expansion (projected 79% urbanization by 2035) intensify pressures on ecosystem services, particularly in ecologically fragile upper reaches with extensive barren land (Fig. 1d)48.

Data sources

Two main forms of data are primarily utilized in this study, the first includes the integration of multi-year, municipal-scale statistical yearbook data from the YRB, used to evaluate the level of tourism urbanization advancement; Furthermore, this study evaluates the four types of ESs functions based on the multi-source and multi-scale spatial datasets, including land use, DEM, meteorological, soil attribute, and road data. For a more detailed overview of the data sources and data processing methodologies, refer to Table 1.

Table 1 Data for the analysis in study.
Full size table

Methods

Assessment of tourism urbanization

Composite index system

To evaluate tourism urbanization (TU) in the Yellow River Basin (YRB), this study developed a composite index system integrating three subsystems:

Urbanization subsystem Measures socio-economic progress through population urbanization (e.g., urban population density), infrastructure (e.g., per capita GDP), and resource management49.

Tourism industry subsystem Quantifies tourism’s economic impact via income (domestic/international), tourist arrivals, and infrastructure (e.g., hotel density)50.

Eco-environment subsystem Evaluates environmental quality using the Pressure-State-Response (PSR) model, including pollution emissions (pressure), green space coverage (state), and waste utilization (response)51.

Entropy method for objective weighting

To eliminate subjective bias, this study applied the entropy method to assign weights to indicators. The steps are:

Normalization Raw data (e.g., tourist income) were normalized to a 0–1 scale.

Entropy calculation For each indicator j, entropy Ej was computed as:

$$\:{E}_{j}=-\frac{1}{{ln}n}\sum\:_{i=1}^{n}{p}_{ij}{ln}{p}_{ij}$$
(1)

.

where\(\:\:\:{P}_{ij}=\frac{{r}_{ij}}{\sum\:_{i=1}^{n}{r}_{ij}}\), xij is the normalized value of indicator j in region i, and n is the number of regions.

Weight assignment Weight \(\:{\omega\:}_{j}\) for indicator j is:

$$\:{\omega\:}_{j}=\left(1-{E}_{j}\right)/{\varSigma\:}_{j=1}^{m}\left(1-{E}_{j}\right)$$
(2)

.

Higher wj indicates greater impact of j on TU.

State space method for integration

The TU index was derived using a state space model:

Axis definition Three axes represent urbanization, tourism, and eco-environment subsystem scores (calculated via entropy-weighted indicators).

Vector calculation Subsystem scores for each region were plotted as coordinates in the state space.

Index calculation The TU index is the Euclidean distance from the origin to the coordinate point:

$$\:TU=\sqrt{{S}_{urban}^{2}+{S}_{tourism}^{2}+{S}_{eco}^{2}}$$
(3)

.

where Surban, Stourism, and Seco are subsystem scores.

As a result of involving the calculation of tourism urbanization indicators for multiple regions and multiple years and considering the correlations between the various indicators in this study, the entropy method was utilized to eliminate subjective factors in assessing weight, thereby ensuring a more objective determination of the weights of each indicator. The detailed evaluation indicators and their corresponding weights are shown in Table 2.

Table 2 Tourism urbanization evaluation index system.
Full size table

Evaluation of ecosystem services

This study evaluated water yield, soil conservation, habitat quality, and carbon sequestration based on the Integrated Valuation of Ecosystem Services and Tradeoffs (InVEST) model in the YRB over the past 20 years. The detailed calculation process and formulas are shown in the supplementary material.

Measurement of coupling relationship between urbanization and ESs

For a meticulous analysis of the interplay between tourism urbanization and ESs, various functions of ESs and facets of tourism urbanization were considered as distinct subsystems within the broader frameworks of tourism urbanization and ESs (Fig. 3).

Spearman correlation analysis is suitable for assessing the relationship between two non-normally distributed continuous variables, while the coupled coordination degree (CCD) model is more suitable for describing interactions between two or more subsystems. Explore the relationship between tourism urbanization with its subsystem levels and four categories of ESs within Spearman correlation analysis in this study first. Next, the coupling coordination between tourism urbanization and ecosystem services (ESs) was analyzed utilizing the CCD model. The computational procedure involves: determining the coupling degree (CD) between ESs and tourism urbanization; calculating their comprehensive evaluation index, and examining the CCD based on the CD and comprehensive evaluation index, specific calculations are referred to references36,45.

This study utilized the CCD model to analyze the coupling coordination between the four categories of ecosystem services and the level of urbanization. The calculation formula is as follows:

$$Ci=2 \times \sqrt {\frac{{wi \times u}}{{{{\left( {wi+u} \right)}^2}}}}$$
(5)
$$T=\alpha wi+\beta u$$
(6)
$$D=\sqrt {C \times T}$$
(7)

.

Ci represents the coupling degree between ecosystem services and the level of urbanization, where C∈[0, 1].

When C =1, the coupling degree is the highest, indicating a highly coordinated development of urbanization and ecosystem services, signifying their trend toward orderly development; conversely, they tend toward disorderly development.

Wi (i = 1, 2, 3, 4) and u are the comprehensive evaluation values of four categories of ecosystem services and the level of urbanization, respectively;

T is the comprehensive evaluation index of the ecosystem services subsystem and the urbanization subsystem;

α and β are weight values, with α = 0.5 and β = 0.5 based on previous research;

D represents the coupling coordination between ecosystem services and the level of urbanization.

Furthermore, the CCD is divided into the verge of disorder (0.0 ~ 0.2), passive coordination (0.2 ~ 0.4), low-level coordination (0.4 ~ 0.6), middle-level coordination (0.6–0.8), high-level coordination (0.8 ~ 1.0)52.

Results

Land use/land cover structure transformation

Between 2000 and 2020, grassland and barren land were the dominant land cover types in the YRB, accounting for 46.02% (94.23 × 104 km2) and 23.8% (48.73 × 104 km2) of the total land area, respectively (Fig. 2). Cropland and forests followed, occupying 15.83% (32.42 × 104 km2) and 22.62% (22.62 × 104 km2). Notably, forest area increased by 2.31 × 104 km2, whereas barren land and cropland decreased by 3.57 × 104 km2 and 0.95 × 104 km2, respectively.

Fig. 2
figure 2

Changes of land use/land cover in the YRB from 2000 to 2020. Note Grassland is the predominant land use type, while areas allocated for Construction Land and Water Bodies are relatively minimal. Notably, there is a general trend of decrease in Grassland, Cropland, and Forest areas, with an increase in Construction over time.

Full size image

Land use transition Sankey diagrams (Fig. 3) indicate that from 2000 to 2020, different land covers were converted into grassland, cropland, barren, forests, construction, and water bodies with areas of 85,296 km2 (39.18%), 41111.2 km2 (18.88%), 31935.7 km2 (14.67%), 29476.2 km2 (13.54%), 20195.7 km2 (9.27%), and 9643.5 km2 (4.43%), respectively. Among these, cropland was converted into grassland and forests with areas of 47,283 km2 and 10,891 km2, while grassland was converted into cropland, barren, and forests with areas of 35740.7 km2 and 18545.5 km2. Additionally, throughout this period, cropland, barren land, grassland, and water bodies transitioned into construction areas, encompassing 16619.2 km2, 2256.2 km2, 870 km2, and 364.2 km2, respectively. Furthermore, the predominant land use transitions among different land use types were from barren to grassland and from cropland to construction. In summary, between 2000 and 2020, the areas covered by forests and water bodies have increased, whereas the expansion of construction has contributed to the reduction of cropland and grassland areas.

Fig. 3
figure 3

Sankey diagram depicting the transfer flows of various land use/land cover within the YRB between 2000 and 2020. Note The widths of the streams in the diagram represent the scale of area switching between different types of land use/land cover, illustrating the changes in land allocation for cropland, barren land, forests, construction sites, grassland, and water bodies in the preceding 20 years. Data derived from 2000–2020 land-use transition matrices.

Full size image

Spatiotemporal changes in tourism urbanization levels

The levels of tourism urbanization increased by 25% from 2000 to 2020, yet displayed significant spatial disparities (Fig. 4). The urbanization subsystem experienced a substantial surge of 37.59%, while the eco-environmental subsystem recorded a 33.19% increase, indicating improved ecological governance (Fig. 4a). An analysis of the coefficient of variation (Fig. 4b) reveals a reduction in development disparities across the YRB.

Fig. 4
figure 4

Temporal trends and variability of tourism urbanization subsystems in the YRB (2000–2020). Note(a) Development levels of tourism (TS), urbanization (US), and eco-environmental (ES) subsystems, showing a 37.59% increase in US and 33.19% in ES. (b) Coefficient of variation (CV) for TS, US, and comprehensive tourism urbanization (TU), with TU exhibiting the highest variability (CV decline: 13.4%).

Full size image

Spatially, tourism urbanization exhibited an ‘east-high, west-low’ distribution pattern (Fig. 5), with major growth observed in Shandong Peninsula, Central Plains, and Guanzhong Plain city clusters, while upstream regions such as LanXi City Cluster exhibited lower levels of development (Fig. 5a). The tourism subsystem (TS) levels showed a similar trend, expanding significantly in the eastern and central YRB, while remaining stagnant in northwestern regions (Fig. 5b). The eco-environmental subsystem (ES) levels followed an opposite gradient, decreasing from the periphery towards the center (Fig. 5c). Comprehensive tourism urbanization (TU) levels also exhibited spatial heterogeneity, with high values concentrated in urban centers and declining towards rural peripheries (Fig. 5d).

Fig. 5
figure 5

Spatiotemporal evolution of tourism urbanization subsystems in the YRB (2000–2020). Note(a) Urbanization subsystem (US) levels. (b) Tourism subsystem (TS) levels. (c) Eco-environmental subsystem (ES) levels. (d) Comprehensive tourism urbanization (TU) levels. Color thresholds: Low (< 0.2), Lower (0.2–0.4), Medium (0.4–0.6), High (0.6–0.8), Higher (> 0.8). Provincial capitals and city boundaries are marked for spatial reference. All graphs presented in this study were created by the author using ArcGIS 10.8, developed by the Environmental Systems Research Institute (ESRI) (https://www.esri.com/).

Full size image

Spatiotemporal variation characteristics of ESs

Ecosystem services (ESs) displayed heterogeneous trends (Table 3; Figs. 6 and 7). Between 2000 and 2020, water yield (WY) and soil conservation (SC) increased by 64.95% and 23.26%, respectively (Table 3). In contrast, habitat quality (HQ) remained stable, while carbon sequestration (CS) exhibited minor fluctuations.

Table 3 Unit area of water yield (WY), soil conservation (SC), habitat quality (HQ), and carbon sequestration (CS) in 2000, 2010, and 2020 in the YRB.
Full size table

High-value areas for water yield (WY) and soil conservation (SC) were predominantly located in the Sanjiangyuan National Park, Qilian Mountains, and Wei River Basin, whereas low-value areas were identified in the North China Plain and the Loess Plateau (Fig. 6). The spatial distribution of WY (Fig. 6a) reveals an increasing gradient from east to west, while SC demonstrates a increasing trend from north to south (Fig. 6b). Patterns of habitat quality (HQ) and carbon sequestration (CS) were largely consistent, with high-value regions concentrated in forested and mountainous areas, including the Qilian Mountains and Sanjiangyuan National Park. Conversely, low-value zones were predominantly found in industrialized and agricultural regions (Fig. 6c).

Fig. 6
figure 6

Spatiotemporal Trends in Ecosystem Service (WY, SC, HQ, CS) Annualized Change Rates (2000–2020). Note(a) The spatial distribution of Water Yield (WY) shows in 2000, 2010, and 2020. (b) The spatial distribution of Soil Conservation (SC). (c) The spatial distribution of Habitat Quality (HQ). (d) The spatial distribution of Carbon Sequestration (CS). Each map provides a visual representation of the respective ecosystem service distribution within the YRB, with the gradient scales indicating the intensity or quantity of ESs provided. City boundaries are marked to delineate the scope of service distribution. All graphs presented in this study were created by the author using ArcGIS 10.8, developed by the Environmental Systems Research Institute (ESRI) (https://www.esri.com/). *Trends here represent spatially explicit annualized change rates calculated as ΔES/year = (ES2020 – ES2000) / 20, where positive/negative values indicate increasing/decreasing trajectories, respectively. Significance testing (p < 0.05) was applied to validate trend robustness.

Full size image

Long-term trends (Fig. 7a) reveal that WY and SC significantly increased in reforested areas, whereas urbanized regions experienced declining CS levels, highlighting the trade-offs between conservation and development. The temporal evolution of SC (Fig. 7b) further demonstrates significant improvement in erosion control zones, particularly in the Loess Plateau and mountainous landscapes. However, HQ experienced a notable decline in urbanized regions, suggesting a deterioration in ecological integrity (Fig. 7c). Forest-dominated regions in the southwestern YRB have seen substantial carbon sequestration (CS) gains, consistent with reforestation policies and reduced deforestation rates. However, the North China Plain and industrial corridors have exhibited declining CS trends, likely due to cropland expansion and fossil fuel emissions. These spatial heterogeneities highlight the interplay between ecological restoration efforts and human-induced pressures in shaping ecosystem service trajectories across the YRB (Fig. 7d).

Fig. 7
figure 7

Trends of ecosystem services (WY, SC, HQ, CS) in the YRB between 2000 and 2020. Note(a) Shows areas with significant trends in Water Yield (WY) (Moran’s I > 0.35). (b) Depicts significant trends in Soil Conservation (SC) (Moran’s I > 0.16). (c) Illustrates areas with notable trends in Habitat Quality (HQ) (Moran’s I > 0.42). (d) Highlights regions with significant trends in Carbon Sequestration (CS) (Moran’s I > 0.28). The legend indicates areas of increase, no significant change, and decrease, with city boundaries provided for spatial context. All graphs presented in this study were created by the author using ArcGIS 10.8, developed by the Environmental Systems Research Institute (ESRI) (https://www.esri.com/).

Full size image

Spatial regression analysis

Spatial auto-correlation is commonly used to detect potential spatial dependence between observed attribute values within a region. Moran’s I is a spatial autocorrelation statistic that quantifies the degree to which similar or dissimilar values cluster geographically. It answers: (1) Are high/low values of a variable (e.g., habitat quality) spatially clustered? (2) Is the observed spatial pattern random, dispersed, or aggregated? In this study, spatial auto-correlation of tourism urbanization and ESs in the YRB is measured using GeoDa 1.20, the specific calculation process and equation refer to references53. All graphs presented in this study were created by the author using ArcGIS 10.8, developed by the Environmental Systems Research Institute (ESRI) (https://www.esri.com/).

When I > 0: Positive spatial autocorrelation (similar values cluster; e.g., high habitat quality regions adjacent to other high-value areas). I < 0: Negative spatial autocorrelation (dissimilar values cluster; e.g., high-value areas surrounded by low-value zones). I ≈ 0: No spatial autocorrelation (random distribution).

The Global Moran’s I was calculated as:

$$I=\frac{n}{{{S_0}}} \cdot \frac{{\sum\limits_{{i=1}}^{n} {\sum\limits_{{j=1}}^{n} {{w_{ij}}\left( {{x_i} - \bar {x}} \right)\left( {{x_j} - \bar {x}} \right)} } }}{{\sum\limits_{{i=1}}^{n} {{{\left( {{x_i} - \bar {x}} \right)}^2}} }}$$
(4)

.

Where: n: Number of spatial units. wij : Spatial weight between units i and j(queen contiguity weights used). S0=\(\sum\limits_{{i=1}}^{n} {\sum\limits_{{j=1}}^{n} {{w_{ij}}} }\): Sum of all spatial weights. xi: Observed value at unit i. \(\bar {x}\): Mean of all observed values.

Interactive relationship between ESs and tourism urbanization

Correlation between tourism urbanization and ESs

Spearman’s correlation analysis (Fig. 8) showed a negative correlation between tourism urbanization and HQ, SC, and CS, whereas WY exhibited a positive association. However, these negative effects weakened over time, suggesting an evolving balance between tourism growth and ecosystem sustainability.

Fig. 8
figure 8

Correlation matrix showing the relationship between the tourism urbanization level of the Yellow River Basin and ecosystem services from 2000 to 2020. Note The ecosystem services include Water Yield (WY), Soil Conservation (SC), Habitat Quality (HQ), and Carbon Sequestration (CS). The subsystems analyzed are Urbanization Services (US), Tourism Services (TS), Environment Services (ES), and Tourism Urbanization (TU). The correlation coefficient values are shown by a color gradient, where red denotes a positive correlation and blue denotes a negative correlation.

Full size image

Analysis of CCD between ESs and tourism urbanization

Between 2000 and 2020, the coupling coordination degree (CCD) between tourism urbanization and ecosystem services (ESs) in the Yellow River Basin (YRB) showed a declining trend (Fig. 9).

Notably, the CCD between tourism urbanization and water yield (WY) exhibited the least decline, indicating that water resources have demonstrated relative resilience to the impacts of urbanization (Fig. 9a). In contrast, soil conservation (SC) displayed the most significant imbalance, with approximately 86.4% of the YRB experiencing a decrease in CCD values, particularly in the Loess Plateau and North China Plain, where soil erosion and land degradation have intensified (Fig. 9b). Spatially, regions with high coordination were identified in the Shandong Peninsula and Central Plains, while low-coordination zones were predominantly located in the Loess Plateau and North China Plain (Fig. 9c). Carbon sequestration (CS) experienced moderate declines in CCD, primarily in industrialized areas where land conversion and emissions have diminished ecological resilience (Fig. 9d).

Fig. 9
figure 9

City-scale spatial pattern of the coupling coordination degree involving tourism urbanization and ESs in 2000, 2010, and 2020. Note(a) Shows the coordination with Water Yield services (WY), (b) illustrates the coordination with Soil Conservation services (SC), (c) depicts the coordination with Habitat Quality (HQ), and (d) represents the coordination with Carbon Sequestration services (CS). Coordination levels: red = severely imbalanced (CCD < 0.3), orange = marginally coordinated (0.3–0.5), light green = moderately coordinated (0.5–0.7), dark green = highly coordinated (> 0.7). The maps include city boundaries for geographical context. All graphs presented in this study were created by the author using ArcGIS 10.8, developed by the Environmental Systems Research Institute (ESRI) (https://www.esri.com/).

Full size image

Discussion

Spatial imbalances and developmental trajectories in tourism urbanization

The Yellow River Basin (YRB) exhibits pronounced spatial imbalances in tourism urbanization, characterized by an “east-high, west-low” gradient, aligning with regional economic disparities and infrastructure investments (Fig. 5d)14,54. Urban clusters such as the Shandong Peninsula demonstrated rapid expansion, reaching urbanization indices above 0.34 by 2020, whereas upstream regions like LanXi City Cluster stagnated with indices below 0.15. This pattern, partially driven by policy-directed resource allocation, reshaped urban-rural development gaps and land-use transitions, with 16,619.2 km2 of cropland converted into construction areas (Fig. 3).

Tourism urbanization also revealed subsystem asynchronicity, a phenomenon previously noted in the Pearl River Delta’s industrial-ecological systems55. while urbanization and eco-environmental subsystems reached over 68% and 72% of their development potential, the tourism subsystem lagged at 43% efficiency, primarily due to its limited integration with ecosystem services (ESs) (Fig. 4a)14,54. This discrepancy underscores the YRB’s unique challenge in reconciling tourism growth with ES conservation. It resulted in deteriorating coupling coordination degrees (CCD), particularly for soil conservation (SC) and habitat quality (HQ), which declined by 9.3% and 11.6%, respectively, in fragile landscapes such as the Loess Plateau (CCD < 0.41) (Fig. 9)54. Such findings highlight how rapid urbanization can drive ES degradation, necessitating a re-evaluation of land-use policies to mitigate long-term ecological trade-offs.

Recent trends (2018–2022) suggest that certain regions may transition to more balanced urbanization. For example, the Guanzhong Plain cluster (CCD = 0.58 ± 0.07) is poised to reach high-intensity urbanization by 2025, with tourism contributing to 37.6% of system momentum54. Similarly, Lanzhou’s annual urban growth rate exceeding 4.2% indicates a potential diffusion of development benefits beyond core tourism hubs. However, reliance on municipal-scale datasets limits the ability to capture fine-scale ecological responses, suggesting a need for higher-resolution spatial analysis in future studies.

Ecosystem services (ESs) dynamics and multiscale drivers

Spatial heterogeneity in ecosystem services (ESs) across the YRB is influenced by both natural and anthropogenic factors. High-value ES clusters, such as Sanjiangyuan National Park, displayed robust water yield (WY: 64.43 mm by 2020) and soil conservation (SC: 4.15 × 10⁷ t/ha), primarily driven by dense forest (22.62%) and grassland (46.02%) coverage (Fig. 2; Table 3)55. In contrast, low-value ES regions, such as the North China Plain, experienced habitat degradation, with HQ declining to 0.70 in 2020 due to rapid urban expansion (construction land increased by 1.95 × 104 km2, Fig. 3)55.

Land-use transitions played a pivotal role in ES variation. Forest expansion (+ 2.31 × 104 km2, Fig. 3) enhanced carbon sequestration (CS: 80.87 t/ha, Table 3) and improved hydrological regulation. Meanwhile, cropland-to-construction conversions (16,619.2 km2, Fig. 3) in downstream hubs resulted in reduced ES stability, particularly in soil conservation and habitat quality. While erosion control measures in the Loess Plateau improved SC, they failed to compensate for HQ losses in urban peripheries (Fig. 7c)56.

These findings reinforce the spatial dependencies of ESs, consistent with studies in other regions that emphasize vegetation coverage as a key driver of ES provision57,58. However, the YRB’s HQ degradation in urban peripheries diverges from the stable HQ patterns observed in the EU’s Natura 2000 protected areas59, likely due to weaker regulatory enforcement in balancing tourism growth and habitat conservation. Additionally, the study’s exclusion of cultural services (e.g., aesthetic value) contrasts with integrated ES assessments60, suggesting future studies should adopt multi-dimensional frameworks to capture tourism’s full socio-ecological impact.

Coupling coordination between tourism urbanization and ecosystem services

The interaction between tourism urbanization and ESs exhibits complex trade-offs and evolving synergies. Initially, rapid construction expansion (1.95 × 104 km2, Fig. 3) in downstream regions (e.g., Shandong Peninsula) contributed to ES degradation, with CCD values for carbon sequestration (CS) declining by 12.96% (Fig. 9d)19. However, over time, these trade-offs became less pronounced, reflecting adaptive governance measures that improved urban ecological resilience.

The CCD analysis (Fig. 9d) reveals that while water yield (WY) remained relatively stable, soil conservation (SC) exhibited the greatest imbalance, particularly in the Loess Plateau and North China Plain, where urbanization pressures intensified soil degradation54. The Guanzhong Plain’s CCD (0.58 ± 0.07, Fig. 9) indicates a threshold beyond which urbanization may transition towards synergy with ES conservation, contingent on green infrastructure investments and stricter land-use regulations2,19.

Despite these insights, this study’s focus on municipal-scale data may obscure localized ecosystem vulnerabilities. Moran’s I spatial dependencies (HQ > 0.42, Fig. 7c) suggest that fine-scale variations within ES provision remain underexplored. Future research should integrate high-resolution geospatial modeling to capture neighborhood-level ES dynamics, particularly in wetland and riparian ecosystems61.

Policy implications and sustainable governance strategies

Given the observed trade-offs between tourism urbanization and ecosystem services (ESs), targeted policy interventions are essential for achieving sustainable development. This study identifies several key strategies: (1) Green Infrastructure Development. In high-urbanization zones, such as the Shandong Peninsula and Guanzhong Plain, expanding urban green spaces and implementing wetland restoration projects could mitigate declines in ESs, particularly in cultural services (CS) and habitat quality (HQ)61. (2) Differentiated Land-Use Regulations. In areas with medium coordination, such as the Loess Plateau, adaptive zoning policies that restrict construction in high-erosion risk regions could enhance ES resilience while supporting tourism growth54. (3) Ecological Compensation Mechanisms. Implementing regional ecological compensation programs could balance resource demands, ensuring that the degradation of ESs driven by urbanization is offset by conservation investments10. By aligning these strategies with global sustainability initiatives, such as the UN Decade on Ecosystem Restoration, the Yellow River Basin (YRB) provides a valuable case study for integrating tourism-driven growth with ecological stewardship, offering insights for other fragile basins worldwide62.

Conclusion

The spatiotemporal interplay between tourism urbanization and ecosystem services (ESs) in the Yellow River Basin (YRB) is characterized by pronounced spatial polarization and declining coupling coordination degrees (CCD), as revealed through integrated entropy-InVEST modeling (Figs. 2, 3, 4, 5, 6, 7, 8 and 9). Key findings include:

(1) Asymmetric land-use transitions Construction land expanded by 1.95 × 104 km2 (2000–2020) at the expense of cropland and grassland (Fig. 3), driving a 12.96% CCD decline between tourism urbanization and carbon sequestration (Fig. 9d).

(2) Subsystem asynchronicity Urbanization and eco-environmental subsystems achieved 68% and 72% of their potential, respectively, while the tourism subsystem lagged at 43% (Fig. 4a), reflecting policy-driven imbalances in resource allocation.

(3) Transitional zone resilience Despite intensive urbanization, the Guanzhong Plain maintained stable CCD (0.58 ± 0.07; Fig. 9d), suggesting thresholds for sustainable transitions in ecologically vulnerable regions.

These findings extend coupling coordination theory by identifying spatial heterogeneity as a critical mediator of tourism-ES interactions, contrasting with Mediterranean models where policy inertia perpetuates imbalances2,19. The proposed three-tier governance framework—prioritizing green infrastructure in high-coordination zones (CCD ≥ 0.55), ecological compensation in transitional areas (0.35 ≤ CCD < 0.55), and slope-based restrictions in fragile regions (CCD < 0.35) —provides a replicable model for balancing growth and ecology in global river basins. For instance, the Shandong Peninsula’s wetland restoration strategy mirrors the Mississippi River Basin’s floodplain management63, while the Loess Plateau’s slope restrictions align with Nepal’s landslide mitigation policies64. This framework’s emphasis on tourism-specific pressures complements the EU’s Water Framework Directive65, which focuses on industrial and agricultural impacts, thereby addressing a critical gap in basin-scale ES management.

Several limitations warrant attention. First, municipal-scale data, while critical for regional policy alignment, may spatially average fine-grained ES variations (e.g., wetland fragmentation in the Yellow River Delta)66, leading to underestimated local trade-offs—a issue similarly noted in the Mekong Basin’s ES assessments67. Second, the exclusion of cultural services (e.g., recreational aesthetics) limits holistic CCD evaluation, contrasting with integrated frameworks like the UK National Ecosystem Assessment68. Third, our model’s assumption of linear urbanization-ES interactions diverges from threshold-based approaches in socio-ecological resilience studies69, potentially oversimplifying feedbacks in fragile landscapes. Future work should integrate high-resolution remote sensing and dynamic system models to resolve these gaps.