UC Santa Barbara

I use infrasound waves to investigate the evolution of the atmosphere and to understand the phenomena that generate and modify the acoustic wavefield at volcanoes. In the first part, I apply machine learning methods and array processing techniques to detect, quantify, and characterize the infrasonic wavefield produced by volcanoes, providing key parameters that help to assess the activity of effusive and explosive systems. Then in the second part of this thesis, I use acoustic ambient noise interferometry to study the structure of terrestrial and Martian atmospheres.

In the second chapter, I employ two infrasound arrays and unsupervised machine learning methods to detect multiyear activity of Tungurahua, El Reventador, and Sangay volcanoes, Ecuador. Using hierarchal clustering applied with the Ward’s criterion, I detect impulsive short-lived signals from these volcanoes in Ecuador. Detections agree with satellite observations of hot spots and with catalogs provided by the local volcano monitoring agency. In the third chapter, I analyze data from a regional (~85 km) infrasound array and a local (~0.3 km) seismo-acoustic station to determine the chronology of the 2018 Sierra Negra eruption (Galápagos). I integrate these seismic and acoustic data streams with satellite data to constrain the eruption onset, track the eruption as it migrates down the north flank, and localize new acoustic vents that have not been mapped on the field yet. I show the capability of seismo-acoustic analysis for enhanced geophysical monitoring. In the fourth chapter, I develop the framework for retrieving atmospheric characteristics using a single infrasound sensor following the autocorrelation interferometry method. I show that relative velocity changes inferred from infrasound autocorrelations can track air temperature and velocity changes. For the propagation geometry at El Reventador (Ecuador) I show that effects from wind velocity can be negligible and provide a theoretical model to derive temperatures from relative velocity changes. In chapter five, I demonstrate that the autocorrelation method can be used to study the Martian atmosphere; specifically, I show that relative velocity changes derived from the pressure sensor on board the InSight lander can track variations of the effective speed of sound. These results also suggest the presence of continuous background infrasound on Mars. These examples highlight the versatility and capability of infrasound to quantify diverse infrasonic phenomena and provide complementary information to other geophysical observations about volcanic activity and atmospheric evolution.