Mapping Hydrothermal Heat Output with Radiometric UAV-TIR: A New Workflow for Volcanic Geothermal Targeting (Pantelleria, Italy)

Thermal anomalies in volcanic hydrothermal systems provide an early and spatially explicit proxy for changes in permeability, fluid pathways, and magmatic–hydrothermal coupling. However, routine volcano monitoring still faces a critical scale gap: ground-based thermometry and gas surveys provide high-quality point data but limited spatial coverage, whereas satellite thermal products often lack the spatial resolution needed to resolve structurally controlled steaming ground and diffuse degassing structures, where many precursory signals localize. This limitation becomes particularly acute at quiescent calderas and rift-related volcanoes, where small-to-moderate thermal changes can occur over metre-scale fracture networks without producing detectable satellite-scale signals. Pantelleria Island (Sicily Channel Rift, Italy), an active volcanic system with persistent fumaroles, steaming ground, and diffuse degassing, represents an ideal natural laboratory to test high-resolution, repeatable thermal monitoring strategies. Here we present a reproducible workflow for radiometric UAV thermal-infrared (TIR) monitoring that converts centimetre-scale orthomosaics into georeferenced products suitable for operational surveillance: (i) surface temperature maps, (ii) heat-flux density rasters, and (iii) volcanic radiative power (VRP) distributions. We acquired calibrated UAV-TIR imagery under calm conditions (low wind; relative humidity ~80%) at altitudes of 60–100 m and processed the data using structure-from-motion photogrammetry to generate co-registered TIR orthomosaics. We then quantified pixel-wise surface heat loss using the ground-surface energy-balance framework widely applied to geothermal terrains (Sekioka–Yuhara approach), combining net longwave radiative emission (Stefan–Boltzmann law) and sensible convective transfer (Newton cooling). Results reveal a strong, multi-megawatt hydrothermal output concentrated within the main fumarolic sector (Q_pos ≈ 8.44 MW over 0.176 km²; core steaming-ground fluxes ~10¹–10² W m⁻²), whereas the second sector exhibits weak, spatially limited anomalies and an order-of-magnitude lower output (Q_pos ≈ 0.206 MW over 0.031 km²). This quantitative contrast supports a permeability-controlled discharge model in which heat and mass transfer focus along discrete upflow pathways and alteration domains, consistent with independent degassing evidence reported for Pantelleria’s hydrothermal areas. By generating operationally usable heat-flux and VRP baselines at the scale of individual vent fields, this approach strengthens volcano monitoring by enabling (i) objective ranking of thermal anomalies, (ii) structural interpretation of upflow pathways, and (iii) time-lapse detection of subtle hydrothermal changes that may precede or accompany unrest. The workflow is readily transferable to other volcanic islands and caldera systems where hydrothermal signals are spatially focused and temporally variable.