We use infrasound waves generated during eruptions of Volcán Tungurahua (Ecuador) to study both, changing atmospheric conditions and volcanic source characteristics. Analyzed infrasound data were recorded for a 32- month period by a five-station network located within 6.5 km from the vent. We use cross-network correlation to quantify the recurrent eruptive behavior of Tungurahua and results are corroborated by reports from the Ec[1]uadorian monitoring agency. Cross-network lag times vary over short time periods (minutes to days) when vent location is stable and attribute these variations to changes in atmospheric structure. Assuming a fixed source lo[1]cation, we invert for average air temperatures and winds in Tungurahua's vicinity (b6.5 km) and find evidence for diurnal and semidiurnal tropospheric tides. We also use cross-network correlation lag times to compute infrasound source positions with resolutions of ~11.6 m, taking into account coarse NOAA atmospheric models for local winds and temperatures. Variable infrasound-derived source locations suggest source migration during the 32 months of analyzed data. Such source position variability is expected following energetic eruptions that destructively altered the crater/vent morphology as confirmed by imagery obtained during regular overflights.

As part of the Acoustic Surveillance for Hazardous Eruptions project, two infrasound arrays were installed in northern and central Ecuador. The RIOE and LITE arrays were operational between 2006 and 2013, recording thousands of infrasound signals originating from eruptions of Tungurahua, El Reventador, and Sangay. We use Progressive Multi-Channel Correlation array processing together with hierarchical clustering to identify and associate impulsive infrasound signals with each volcano. Infrasound detections correspond to quasi-continuous activity of Sangay between mid 2006 and mid 2012, at least thirteen periods of activity of Tungurahua between 2006 and mid 2012, and strong signals from El Reventador in early 2008. We validate our detections using satellite observations from the MODIS (Moderate Resolution Imaging Spectroradiometer) volcano detection algorithm (MODVOLC) and independent eruption catalogs. For Sangay, we find a good agreement between our infrasound detections and MODVOLC triggers, whereas for El Reventador the agreement is fair. We attribute the fair correspondence at El Reventador to the low-temporal resolution of MODIS data and frequent cloud cover in the region that reduce satellite observational capability. Finally, our study highlights the benefits of infrasound arrays for monitoring volcanoes in Ecuador at regional distances, while further investigating the processes that limit volcanic infrasound detection..

Seismic and infrasound multistation ambient-noise interferometry has been widely used to infer ground and atmospheric properties, and single-station and autocorrelation seismic interferometry has also shown potential for characterizing Earth structure at multiple scales. We extend autocorrelation seismic interferometry to ambient atmospheric infrasound recordings that contain persistent local noise from waterfalls and rivers. Across a range of geographic settings, we retrieve relative sound-speed changes that exhibit clear diurnal oscillations consistent with temperature and wind variations. We estimate ambient air temperatures from variations in relative sound speeds. The frequency band from 1 to 2 Hz appears most suitable to retrieve weather parameters as nearby waterfalls and rivers may act as continuous and vigorous sources of infrasound that help achieve convergence of coherent phases in the autocorrelation codas. This frequency band is also appropriate for local sound speed variations because it has infrasound with wavelengths of ∼170–340 m, corresponding to a typical atmospheric boundary layer height. After applying array analysis to autocorrelation functions derived from a three-element infrasound array, we find that autocorrelation codas are composed of waves reflected off nearby topographic features, such as caldera walls. Lastly, we demonstrate that autocorrelation infrasound interferometry offers the potential to study the atmosphere over at least several months and with a fine time resolution.

Cotopaxi Volcano showed an increased activity since April 2015 and evolved into its eventual mild eruption in August 2015. In this work we use records from a broadband seismic station located at less than 4 km from the vent that encompass data from April to December of 2015, to detect and study low-frequency seismic events. We applied unsupervised learning schemes to group and identify possible premonitory low-frequency seismic families. To find these families we applied a two-stage process in which the events were first separated by their frequency content by applying the k-means algorithm to the spectral density vector of the signals and then were further separated by their waveform by applying Correntropy and Dynamic Time Warping. As a result, we found a particular family related to the volcano’s state of activity by exploring its time distribution and estimating its events’ locations.

We characterize and interpret a new type of infrasound signal originating from the summit of Volcán Cotopaxi (Ecuador) that was primarily observed between September 2015 and March 2016, following the 2015 eruptive period. This infrasound waveform is a slowly decaying sinusoid with exceptional low-frequency (fp = 0.2 Hz) and high quality factor (Q = ~10) and resembles the shape of tornillo seismic waveforms. The repeating events, occurring about once per day in early 2016, are stable in frequency content, and we attribute them to excitation of a vertical-walled crater, with radius of about 125 m and length of 300 m. Spectral properties of the tornillo permit constraints on crater sound speed (335 m/s ± 6%) and temperature (4–32°C). The initial polarity of the tornillos is predominantly a rarefaction and could reflect repeating crater bottom collapse events (implosions) or explosion sources whose infrasound is heavily modulated by the crater’s pipe-like geometry.

El proyecto “Caracterización y estadística de señales volcánicas en los Andes, casos de estudio volcanes Reventador y Cotopaxi en Ecuador” (GEOF 02 2019), formó parte del Programa de Asistencia Técnica del año 2019 del Instituto Panamericano de Geografía e Historia. Contempló trabajos de campo en el volcán Reventador en tres ocasiones, para descarga de datos y mantenimiento de las estaciones acústicas CON, LAV y AZU. Como parte de los trabajos de campo y de las discusiones científicas, se contó con la asistencia del doctor Jeffrey Johnson. También, se llevó a cabo el taller “Principios de Geofísica y Comunicación”, con la asistencia de 19 estudiantes de seis universidades, y siete instructores de diversas áreas de Ciencias de la Tierra. Adicionalmente, se hizo una consultoría para el desarrollo de un sistema de adquisición de datos acústicos, del cual ya se tienen los esquemas y reportes. Finalmente, se publicaron los resultados del proyecto en mayo de 2019, en la revista International Journal of Geophysics, igualmente se enviaron los resúmenes científicos a las conferencias: Fall Meeting 2019-American Geophysical Union, 8th International Symposium on Andean Geodynamics y al 1er Congreso de la Asociación Latinoamericana de Volcanología.