A Reliability-Aware Landsat Framework for Sustainable Geothermal Resource Potential Assessment: A Case Study from Quzhou, Eastern China

Thermal infrared land-surface temperature (LST) anomalies can support early-stage geothermal reconnaissance, but in humid, vegetated and urbanizing low-temperature regions they are readily affected by land cover, terrain and human activity. This study develops a reliability-aware Landsat 8 framework for Quzhou, eastern China. Landsat 8 Collection 2 winter scenes from 2016 to 2025 were cloud-screened, converted to LST, and composited using a per-pixel median strategy. NDVI, NDBI and NDWI masks reduced vegetation, impervious-surface and water interference. Standardized LST anomalies were integrated with a 3 km local-relief constraint, candidate-patch vectorization, and patch-level scoring based on thermal persistence, terrain context, land-cover reliability, structural proximity and lithology/contact favorability. General and strong winter LST anomalies cover approximately 64.4 and 22.3 km 2 , respectively, with impervious surfaces accounting for 55.97% and 67.65% before reliability screening. The terrain-constrained workflow reduces the candidate area to approximately 22.2 km 2 and ranks 267 patches, including 27 very-high relative reconnaissance-priority patches. The highest-ranked patches are characterized mainly by combined thermal persistence, lower land-cover interference, favorable fault proximity and favorable lithology/contact settings, indicating relative priority rather than confirmed resources. Integrated reliability screening is therefore essential before using Landsat LST anomalies for geothermal target prioritization.