Ecological–environmental transformation efficiency in China: regional disparities, modeling challenges, and prospects for long-term sustainability governance

Amid growing pressure to balance economic growth and environmental sustainability, this study develops a comprehensive framework, ecological sustainability trajectory in dynamic evolution (ECO-STRIDE), to evaluate and forecast regional ecological–environmental transformation efficiency (EETE) in China. Drawing on panel data from 30 provinces (2010–2023), the framework integrates dynamic efficiency modeling, spatial–temporal analysis, deep learning prediction, and model interpretability. The findings reveal: (1) EETE remains low overall, with pronounced interregional disparities. Resource utilization efficiency generally exceeds environmental governance efficiency, with provinces like Guangdong showing marked imbalance between the two stages. (2) While both stages demonstrate an upward trend, regional gaps continue to widen. Resource utilization efficiency steadily improves, but with intensifying structural divergence, whereas environmental governance efficiency lags behind and displays clear polarization. The eastern region maintains a leading position across both stages due to its superior capacity in resource allocation and coordinated governance. (3) A hybrid CNN–LSTM–Attention model delivers high predictive accuracy, projecting continuous improvement in EETE through 2035, although spatial imbalances persist. (4) SHAP-based interpretation identifies urbanization, industrial upgrading, and digital infrastructure as key drivers, while inefficiencies in green finance allocation and underutilized ecological endowments remain critical constraints. By providing a full-period, integrated assessment from historical patterns to future trajectories, this study advances the methodological frontier of EETE research and informs region-specific strategies for long-term resource coordination and adaptive environmental governance under uncertainty.