Integrating road network topology, remote sensing, and explainable machine learning to assess urban ecological resilience

Rapid urbanization, driven by expanding road infrastructure and increasing impervious surfaces, has intensified ecological degradation in arid megacities such as Riyadh. This study presents a technology-enabled integrative framework that combines explainable machine learning and remote sensing to evaluate the spatial relationship between road network morphology and urban ecological resilience. Using OpenStreetMap data, eight functional classes of roads were identified, and seven graph-theoretic indices such as Alpha, Beta, Gamma, Eta, Cyclomatic Number, Detour Index, and Nearest Neighbour Index were computed. Ecological quality was assessed using the Remote Sensing-based Ecological Index (RSEI), derived from Landsat 8 imagery by integrating NDVI, LST, NDBSI, and surface wetness through Principal Component Analysis. The Random Forest model achieved an R² of 0.48 and an RMSE of 0.027, indicating that road network features account for nearly half of the spatial variance in RSEI. SHAP (SHapley Additive exPlanations) analysis revealed that network density positively contributed up to +0.08 SHAP units in motorway zones. Compact block geometries and clustered intersections (NNI) consistently enhanced ecological performance across road types. The study results provide actionable directions to urban transport planners through the identification of specific configurations of the road networks which promote ecological resilience. Specifically, the identification of more clustered intersections and denser road networks combined with more direct routes can provide better surface ecological conditions for arid cities, the proposed Random Forest-SHAP framework enables planners to prioritise road redesign and green corridors, supporting evidence-based transport planning aligned with SDGs 11 and 13. • Road network topology explains nearly half of Riyadh’s ecological variance. • Machine learning with SHAP reveals critical transport-ecology interaction pathways. • Compact street geometries and dense intersections enhance urban ecological quality. • Detour indices strongly degrade suburban ecological resilience in arid cities. • Innovative Framework supports SDG-driven climate-resilient and sustainable urban planning strategies.