Answers are grounded in this paper and cite their sources.
Intensifying urban heat island and associated heat stress under climate change require a multiscale investigations of land surface drivers of thermal extremes to adapt to heat severity. Yet, how landscape characteristics modulate thermal extremes remains poorly understood. This study examines the scale-dependent influence of landscape structure on extreme surface temperatures in Lusatia, Germany. Using 11 years (2013–2023) of MODIS-derived land surface temperature data, we derived two indicators: annual maximum temperature (AMT) to represent the intensity of heat extremes, and its interannual variability (AMT-var) to capture temporal fluctuations. A multiscale framework (1–400 km 2 ) combining landscape composition, configuration, and topography was analyzed with eXtreme Gradient Boosting (XGBoost) and interpreted using SHapley Additive exPlanations (SHAP). Results showed that while land cover and topography dominated heat dynamics at fine scales, landscape configuration, particularly cohesion, aggregation, core area and shape metrics emerged as the strongest predictor of AMT and AMT-var at regional scales (>100 km 2 ). Nonlinear threshold effects reveal that highly cohesive and simply shaped patches on landscapes can buffer extreme heat and stabilize its variability, whereas highly fragmented, homogeneous, and overly intermixed landscapes amplify thermal extremes. These findings underscore the importance of considering spatial configuration alongside land cover in climate adaptation planning. By linking interpretable machine learning with multiscale landscape indicators, this study provides actionable insights for climate-resilient land-use design and sustainable landscape planning.
No takes yet. Share an insight, caveat, or question.
Aryal et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: