Revealing heat patterns of lava flows: a spatial data analysis approach using UAV thermography

The 2021 Tajogaite eruption on La Palma (Canary Islands, Spain) generated extensive lava flows that still exhibit measurable residual surface heat several years after the eruption. Understanding the spatial distribution and persistence of this heat is essential for characterizing post-eruptive cooling processes and for supporting reconstruction activities in affected areas. 

An integrated geospatial workflow was implemented to combine high-resolution UAV-based thermal imagery with lava-thickness models across two sectors affected by the eruption: LPAgricultura (surveyed in February 2024) and LPUrban (surveyed in June 2025). Drone-based radiometric infrared imagery was processed to produce georeferenced thermal mosaics, with emissivity correction (ε = 0.95), and resampled to match the spatial resolution of the corresponding lava-thickness datasets. All data were aligned within a common spatial reference system (REGCAN95 / UTM zone 28N) to ensure pixel-level correspondence. 

Thermal anomalies were defined as surface temperatures equal to or exceeding 30 °C. Lava-thickness values were extracted separately for thermally anomalous and non-anomalous areas, enabling a consistent spatial comparison between the two conditions. Statistical analyses were conducted independently for each sector to evaluate the relationship between residual heat and flow thickness. 

Results reveal a clear,statistically significant association between elevated surface temperatures and thicker lava deposits across the Tajogaite lava field. In the LPUrban sector, characterized by thicker lava accumulations (mean thickness = 21.5 m; maximum = 57.1 m), thermally anomalous areas have a mean thickness of 31.3 m, compared with 21.3 m in non-anomalous zones (p < 0.001). In contrast, the LPAgricultura sector, dominated by thinner flows (mean thickness = 9.2 m; maximum = 51.5 m), shows mean thickness values of 20.6 m in anomalous areas versus 10.0 m elsewhere (p < 0.001). These patterns indicate that residual heat is preferentially concentrated within the thickest portions of the lava flows, where cooling is constrained by reduced surface-to-volume ratios and enhanced thermal insulation. The adoption of relative thickness thresholds (≥ 20 m in urban areas and ≥ 10 m in agricultural areas) captures approximately 95% of the total surface area of detected thermal anomalies, ensuring consistent sensitivity across both sectors.  

The combined use of UAV thermography and lava-thickness models enables a robust characterization of post-eruptive thermal persistence, with direct implications for the assesing lava-flow cooling behavior in complex volcanic terrains.