Optimizing blue-green spatial structure to enhance climate resilience under water scarcity

In arid and semi-arid oases, blue-green spaces regulate local thermal conditions, but their expansion is restricted by severe water scarcity. Optimizing the spatial configuration of existing blue-green spaces may therefore offer a feasible strategy for enhancing climate resilience. This study examined four representative oasis groups in the Hexi Corridor, Northwest China, using multi-source datasets of land surface temperature, land-cover, vegetation, albedo, and topography. A Relative Oasis Cooling Intensity index was constructed from an LST-distance gradient to quantify cooling effects under heterogeneous background-temperature conditions. Morphological Spatial Pattern Analysis and CatBoost-SHAP were combined to identify key structural factors and nonlinear cooling responses. The results show that the maximum cooling distance increased significantly with oasis area, with R² = 0.86 and P < 0.01. The landscape-thermal relationship showed clear scale dependence, and 1500 m was identified as the optimal analytical scale. NDVI was the dominant predictor of cooling intensity across the four oasis groups, whereas MSPA Core was the only morphological class with a consistently positive contribution to Relative Oasis Cooling Intensity. In contrast, fragmented or isolated components, particularly Islet, Branch, and Loop, generally showed weak or negative contributions. In three oasis groups, the contribution of Core areas increased nonlinearly after Core proportions exceeded system-specific thresholds of approximately 40%-90%, while the Jiuquan group showed an approximately linear positive response. These results suggest that vegetation intensity provides the basic cooling potential of oases, while spatial integrity enhances the efficiency of this cooling function. The findings provide quantitative guidance for optimizing blue-green spatial structure and improving climate resilience in water-limited dryland regions.