Sea–Land Heterogeneity of Summer Land Surface Temperature Across Local Climate Zones in Chinese Coastal Cities: Implications for Sustainable Urban Planning

Coastal thermal planning requires understanding how land surface temperature (LST) varies across coastal distances and local climate zones (LCZs). We developed a City–LCZ–Gradient framework for 59 Chinese coastal cities using summer-2023 MOD11A1 observations, rule-based LCZ maps, gradient statistics, XGBoost–SHAP and GeoDetector. Between the 0–10 and 90–100 km bands, pooled mean LST changed from 30.86 to 31.12 °C by day and from 22.40 to 21.60 °C at night. Among 57 cities with valid gradient series, 13 (22.8%) daytime and 19 (33.3%) night-time mean-LST trends met the 0.05 threshold, predominantly indicating inland cooling. Descriptive daytime slopes contrasted between open high-rise LCZ 4 (+0.315 °C per 10 km) and dense-tree LCZ A (−0.129 °C per 10 km). Night-time lights led daytime prediction (mean absolute SHAP contribution: 0.714 °C); at night, their contribution (0.516 °C) was comparable to elevation (0.498 °C). These multilevel results reveal thermal contrasts obscured by pooled averages and provide quantitative support for sustainable coastal planning, including zone-specific greening assessments and night-time thermal monitoring. The findings characterize spatial associations in the sampled summer, rather than isolated causal effects or universal coastal thresholds.