Name Surname
Decoding pre-eruptive dynamics at Campi Flegrei Caldera: New insights from multimodal petrologic sources combining traditional ML methods and physics informed AI models
Main Supervisor: Prof. Maurizio Petrelli (University of Perugia)
Co-Supervisors: Réka Lukács (Eötvös Loránd Tudományegyetem), Blaž Zupan (University of Ljubljana)
Objectives
The project proposes the development of integrated modeling approaches that combine the knowledge deriving from a large literature record, analytical and optical determination on different phases (minerals, melt, volatiles) of natural samples from explosive volcanic events at Campi Flegrei Caldera with the experimental record and thermodynamic simulations by combining traditional ML methods and physics informed generative AI. The primary objectives are twofold: 1) to decipher pre-eruptive dynamics and 2) to reconstruct the architecture of the volcanic plumbing system and its evolution over time. Achieving these aims will not only mark a significant scientific advancement but also critically improve hazard definition, enabling stakeholders to enhance risk mitigation strategies.
Epected Results
- The development of generative AI models that combine diverse sources of petrologic information (i.e., textual, analytical data on natural samples, thermodynamic simulations, and petrologic experiments) and offer new insights into the pre-eruptive phenomena at Campi Flegrei Caldera;
- The definition of new tools designed to reconstruct the architecture of the volcanic plumbing system at Campi Flegrei Caldera and its evolution over time;
- Advance the knowledge of pre-eruptive dynamics at the Campi Flegrei Caldera and improve the definition of the architecture of the magma feeding system.