DOTTORATO NAZIONALE IN OSSERVAZIONE DELLA TERRA

Titolo della tesi: Renewing the Role of Thermal InfraRed (TIR) Channels for Remote Sensing of Volcanic Activity and Integration with Multi-Spectral Data for Near-Real-Time Monitoring

For over half a century, satellite-based infrared remote sensing has transformed our understanding of volcanic activity. Yet, research and monitoring have predominantly focused on high-temperature processes observed through Mid-InfraRed (MIR; 3–5 µm) channels. This emphasis has relegated the Thermal InfraRed (TIR; 8–14 µm) spectrum to a marginal role, despite its fundamental significance for accurately assessing and quantifying moderate- to low-temperature volcanic processes. This work revisits and reinvigorates the use of TIR satellite observations in volcanology, demonstrating that TIR data capture unique and otherwise inaccessible signals essential for understanding the full spectrum of volcanic thermal activity. Two major methodological advancements are presented. The TIR-based Volcanic Radiative Power (VRPTIR) method establishes a physically grounded framework for accurately deriving radiative power from a single TIR band. Validation against ground truth data confirms that VRPTIR produces consistent, uncertainty-constrained estimates (±35%) of system radiative output, effectively correcting the underestimations (up to 90%) that affect conventional approaches. By capturing the full range of moderate- to low-temperature emissions, VRPTIR enables the construction of long-term time series of radiative output, allows a better interpretation of eruption dynamics, post-emplacement cooling processes, and supports quantitative assessments of volcanic energy budgets. In this regard, VRPTIR has the potential to redefine global estimates of volcanic heat output. For instance, the TIR-derived annual Volcanic Radiative Energy (VRE) for Vulcano (~1×1014 J) is comparable to MIR-based estimates reported for systems such as Bezymianny and Láscar, highlighting the significant contribution of non-eruptive, low-temperature sources to the global volcanic heat budget. Building on this foundation, the TIRVolcH algorithm was developed as the first dedicated global hotspot detection system operating exclusively on TIR data. Designed for use with VIIRS imagery, TIRVolcH achieves sensitivity to pixel integrated temperature increases as low as 0.5 K above background, with a false positive rate of ~1.8% and provides temporal coverage exceeding that of higher-resolution systems by more than an order of magnitude. This work further extends to complementary applications that enhance both the scientific and operational dimensions of volcanic monitoring. These include the integration of polar and geostationary platforms into established systems such as MIROVA, as well as the fusion of multispectral and multitemporal datasets to investigate multi-decadal variability across diverse volcanic settings. Collectively, this work reframes the role of TIR observations in the study of active volcanoes. It demonstrates that non-eruptive, low-temperature signals are not background noise but fundamental expressions of a volcano’s thermal state, providing early indicators of unrest and refining our quantification of global volcanic energy release. The findings establish a new foundation for the integration of TIR data into multispectral and multiparametric frameworks and open the path for the systematic application of VRPTIR to past and next-generation high-resolution TIR missions.