Aerial Robotic Studies of Volcanic CO2 Emissions

Volcanic systems are inherently complex, involving dynamic interactions among magma flow, gas emissions, and atmospheric dispersion. This dissertation focuses on developing and analyzing autonomous UAS algorithms for efficiently surveying volcanic CO2 plumes, introducing several novel methods: the LoCUS algorithm, a swarm coordination and self-healing algorithm that supports gradient-based plume tracking, a transect-based technique that employs a 2D Gaussian fit to calculate CO2 plume flux, and the Sketch algorithm for rapid plume boundary tracing. By treating multiple UAS as a single scientific instrument, these methods leverage swarm algorithms to use in-situ data in ways impossible with individual drones. Validated through simulations and field experiments at sites such as the Valles Caldera supervolcano in New Mexico and the Tajogaite eruption in La Palma, these techniques effectively find plume sources, calculate maximum CO2 plume flux, and map plume areas, all the while mitigating operational risks. Conducted under the VolCAN project, this research provides powerful tools for volcano monitoring and hazard prediction, with broader implications for studying environmental phenomena.