Resonances of a Volcanic Conduit Triggered by Repetitive Injections of an Ash-laden gas

Tungurahua Volcano, located at the center of the Ecuadorian Andes, entered in eruption characterized by ash emissions, vulcanian and strombolian activities in October 1999. There have been 7 to 10 short-period seismic stations and 1 broadband station to monitor the Tungurahua volcanic activity. Between December 5 and 11, 2001, a swarm of long-period (LP) events occurred. The LP waveform is characterized by screw-shaped harmonic oscillations. In this study, we used the LP waveforms to calculate temporal variations in their complex frequencies (frequency, f and quality factor, Q) to understand the source process associated with this activity. We applied the Sompi spectral method to 57 LP waveforms from the station RETU, which provided the best continuous data in the Tungurahua seismic network. We determined the temporal variations in the complex frequencies of the LP events, which indicated that the frequency gradually decreased from 3.5 to 2 Hz, whereas Q gradually increased from 100 to 400 with wide scatter along this trend. Based on the acoustic properties of a crack containing various types of magmatic and hydrothermal fluids, the observed high Q values can only be explained by a large velocity contrast between the surrounding rock and fluid. Such a high velocity contrast suggests ash-gas or water droplet-gas mixtures in a crack developed in the solid rock. Since the LP events are most probably located around 2 km below the summit, a water droplet (liquid water) may not exist at this source depth. We may therefore reasonably assume that the fluid at the source of the LP events is an ash-gas mixture. We propose that an ash-laden gas was injected into a closed conduit (pre-existing crack) just above the magmatic system, causing the resonance of the conduit or an LP event. This process repeatedly occurred and accumulated ash particles in the conduit, generating the temporal variations in the frequency and Q of the LP events. Finally, the conduit was almost filled with ash particles, and as a consequence the resonance of the conduit was no longer excited. The first explosion after this swarm occurred on December 30, and ash emissions from the summit began on January 2, 2002. Our results emphasize the link between LP seismicity and preparatory process for eruptions.