Assessing ecological risk from tourism in human-dominated landscapes: Integrating scenario modelling and nature-based strategies in eastern Qinghai Province, China

• The LER in eastern Qinghai shows a “high-northeast, low-south-central, high-southwest” gradient • Optimization the LER model by integration of ecosystem services theory and multiple methods (buffer gradient/scenario simulation). • TAI and LER have scale-dependent correlations: significantly positive at scenic scale, shifting from weak negative to positive at regional scale. • This study provides a spatialized decision tool for coordinating tourism and ecological protection in plateau areas and advancing SDGs. Amid economic development and urbanization, tourism has emerged as a core anthropogenic disturbance factor in human-dominated landscapes, exerting a prominent threat to the stability of ecosystems especially in ecologically fragile regions. However, existing studies remain limited in three key aspects: they have not clearly defined the spatial scope of how tourism infrastructure and tourism-related human activities impact landscape ecological risk (LER); they lack quantitative analyses of the differential ecological risk mechanisms associated with tourism development units of varying types and grades; and no dynamic balance pathway between tourism development intensity and ecological resilience has been established to date. To address these research gaps, the present study focused on eastern Qinghai province as the study area. Guided by the “Nature-based Solutions (NbS)” framework, it integrated the theory of ecosystem services with the core principles of NbS, and further developed an optimized LER assessment model. This model incorporates multi-buffer gradient analysis, scenic area scenario simulation, a Tourism Activity Intensity (TAI) index, spatial autocorrelation analysis, and a response index, with the specific objectives of quantifying the characteristics of LER induced by tourism activities and providing scientific support for regional sustainable land management. The results show that: (1) From 2000 to 2020, LER in the study area exhibited a distinct gradient distribution pattern, with high values in the northeast, low values in the central-south, and high values in the southwest. The LER evolutionary trajectory was divided into two phases: a period of rapid change (2000–2010) and a phase of gradual adjustment (2010–2020). (2) The buffer zone analysis showed that risk growth slowed beyond the 9 km threshold, with a dynamic balance zone forming within the 11–15 km buffer. Risks remained consistently higher outside scenic areas compared to within them, and 5A-level scenic areas had a 35.6 % lower risk than 4A-level scenic areas. Cultural scenic areas exhibited a 163.2 % higher risk value than natural scenic areas, and road-influenced zones maintained an overall elevated risk level. (3) Scale-dependent variations characterized the relationship between TAI and LER. At the regional scale, the relationship shifted from a weak negative correlation to a positive one, whereas at the scenic area scale, a significant positive correlation persisted. Additionally, between 2010 and 2020, the spatial synergy displayed the dual characteristics of strengthening within core areas and regional differentiation across the broader landscape. By advancing the mainstreaming of NbS through a spatially explicit quantitative framework, this study provides decision-making support for balancing tourism development and ecological protection in human-dominated landscapes, facilitating the achievement of the Sustainable Development Goals (SDGs), and enhancing ecological resilience.