Abstract
<title>Abstract</title> <p>Rapid urbanization can expose high-value ecosystems to disturbance while leaving other landscapes degraded and vulnerable. This study develops a risk-service mismatch framework that integrates the landscape ecological risk index (LERI), ecosystem service value (ESV), performance-screened machine-learning diagnosis, and bivariate ecological zoning. The framework was applied to the Beijing-Tianjin-Hebei region (BTH), Yangtze River Delta (YRD), and Guangdong-Hong Kong-Macao Greater Bay Area (GBA) for five years between 2000 and 2020. Mean LERI declined and ESV increased in all three regions, but their trajectories and spatial configurations differed. The YRD showed the largest LERI reduction yet retained relatively high risk, whereas the GBA combined the lowest mean risk with concentrated high-risk patches. Regulating services dominated ESV growth, and water bodies and forestland were persistent contributors. Driver importance was interpreted only for models with adequate test-set performance. Ecological advantage zones expanded in all regions, but vulnerable zones persisted in the BTH plains and sensitivity zones increased in the YRD and GBA. By distinguishing high-value ecosystems requiring risk control from low-value landscapes requiring functional restoration, the framework converts joint assessment into region-specific conservation and restoration priorities.</p>