Detection, classification, and localization of seismic signals at the Soufrière geothermal system, Guadeloupe

La Soufriere is an active stratovolcano located in the South of Basse-Terre island, Guadeloupe, and is part of the Lesser Antilles volcanic arc. Eruptions are usually characterised as peléan with the most recent magmatic eruption occurring approximately in 1530 A.D. Since then, activity at La Soufrière has consisted of phreatic eruptions, the last being in 1976. Present-day activity appears to be related to surface fumaroles and shallow-subsurface seismic activity. In 1992, the alert level was updated from green to yellow and the recorded activity began to slowly increase. Since 2018, the hydrothermal system has been subject to overpressure and overheating processes, resulting in a change to the usual fumarole activity and an increase in volcano seismicity.

We analyze the continuous passive seismic record from one month of data recorded on 47 3-component 5Hz nodal stations, deployed around the volcanic summit. We locate emergent seismic signals, including tremors, to have a better understanding of the subsurface structure of the plumbing system, and reveal more information pertaining to the hydrothermal system.

We use the Python package Covseisnet, which uses a network Covariance Matrix Analysis to detect and locate seismic signals that are typically induced by the geothermal systems and volcanic unrest. This method is based on the decomposition of the matrix into eigenvectors and eigenvalues.  Our results are also compared against two complementary methods derived from the same seismic dataset to obtain a more comprehensive interpretation of the subsurface architecture and reduce the uncertainty of our analyses. The combination of our Covariance Matrix analysis, Ambient Noise Tomography, and Local Earthquake Tomography provides an updated image of the shallow plumbing system of the Soufrière Volcano plumbing’s system.