Search for: All records

Abstract Volcano infrasound data contain a wealth of information about eruptive patterns, for which machine learning (ML) is an emerging analysis tool. Although global catalogs of labeled infrasound events exist, the application of supervised ML to local (<15 km) volcano infrasound signals has been limited by a lack of robust labeled datasets. Here, we automatically generate a labeled dataset of >7500 explosions recorded by a five-station infrasound network at the highly active Yasur Volcano, Vanuatu. Explosions are located via backprojection and associated with one of Yasur’s two summit subcraters. We then apply a supervised ML approach to classify the subcrater of origin. When trained and tested on data from the same station, our chosen algorithm is >95% accurate; when training and testing on different stations, accuracy drops to about 75%. The choice of waveform features provided to the algorithm strongly influences classification performance. 

Abstract Over the past two decades (2000–2020), volcano infrasound (acoustic waves with frequencies less than 20 Hz propagating in the atmosphere) has evolved from an area of academic research to a useful monitoring tool. As a result, infrasound is routinely used by volcano observatories around the world to detect, locate, and characterize volcanic activity. It is particularly useful in confirming subaerial activity and monitoring remote eruptions, and it has shown promise in forecasting paroxysmal activity at open-vent systems. Fundamental research on volcano infrasound is providing substantial new insights on eruption dynamics and volcanic processes and will continue to do so over the next decade. The increased availability of infrasound sensors will expand observations of varied eruption styles, and the associated increase in data volume will make machine learning workflows more feasible. More sophisticated modeling will be applied to examine infrasound source and propagation effects from local to global distances, leading to improved infrasound-derived estimates of eruption properties. Future work will use infrasound to detect, locate, and characterize moving flows, such as pyroclastic density currents, lahars, rockfalls, lava flows, and avalanches. Infrasound observations will be further integrated with other data streams, such as seismic, ground- and satellite-based thermal and visual imagery, geodetic, lightning, and gas data. The volcano infrasound community should continue efforts to make data and codes accessible and to improve diversity, equity, and inclusion in the field. In summary, the next decade of volcano infrasound research will continue to advance our understanding of complex volcano processes through increased data availability, sensor technologies, enhanced modeling capabilities, and novel data analysis methods that will improve hazard detection and mitigation. 

Abstract. Surficial mass wasting events are a hazard worldwide. Seismic and acoustic signals from these often remote processes, combined with other geophysical observations, can provide key information for monitoring and rapid response efforts and enhance our understanding of event dynamics. Here, we present seismoacoustic data and analyses for two very large ice–rock avalanches occurring on Iliamna Volcano, Alaska (USA), on 22 May 2016 and 21 June 2019. Iliamna is a glacier-mantled stratovolcano located in the Cook Inlet, ∼200 km from Anchorage, Alaska. The volcano experiences massive, quasi-annual slope failures due to glacial instabilities and hydrothermal alteration of volcanic rocks near its summit. The May 2016 and June 2019 avalanches were particularly large and generated energetic seismic and infrasound signals which were recorded at numerous stations at ranges from ∼9 to over 600 km. Both avalanches initiated in the same location near the head of Iliamna's east-facing Red Glacier, and their ∼8 km long runout shapes are nearly identical. This repeatability – which is rare for large and rapid mass movements – provides an excellent opportunity for comparison and validation of seismoacoustic source characteristics. For both events, we invert long-period (15–80 s) seismic signals to obtain a force-time representation of the source. We model the avalanche as a sliding block which exerts a spatially static point force on the Earth. We use this force-time function to derive constraints on avalanche acceleration, velocity, and directionality, which are compatible with satellite imagery and observed terrain features. Our inversion results suggest that the avalanches reached speeds exceeding 70 m s−1, consistent with numerical modeling from previous Iliamna studies. We lack sufficient local infrasound data to test an acoustic source model for these processes. However, the acoustic data suggest that infrasound from these avalanches is produced after the mass movement regime transitions from cohesive block-type failure to granular and turbulent flow – little to no infrasound is generated by the initial failure. At Iliamna, synthesis of advanced numerical flow models and more detailed ground observations combined with increased geophysical station coverage could yield significant gains in our understanding of these events. 

Infrasound data are routinely used to detect and locate volcanic and other explosions, using both arrays and single sensor networks. However, at local distances (<15 km) topography often complicates acoustic propagation, resulting in inaccurate acoustic travel times leading to biased source locations when assuming straight-line propagation. Here we present a new method, termed Reverse Time Migration-Finite-Difference Time Domain (RTM-FDTD), that integrates numerical modeling into the standard RTM back-projection process. Travel time information is computed across the entire potential source grid via FDTD modeling to incorporate the effects of topography. The waveforms are then back-projected and stacked at each grid point, with the stack maximum corresponding to the likely source. We apply our method to three volcanoes with different network configurations, source-receiver distances, and topography. At Yasur Volcano, Vanuatu, RTM-FDTD locates explosions within ∼20 m of the source and differentiates between multiple vents. RTM-FDTD produces a more accurate location for the two Yasur subcraters than standard RTM and doubles the number of detected events. At Sakurajima Volcano, Japan, RTM-FDTD locates the source within 50 m of the active vent despite notable topographic blocking. The RTM-FDTD location is similar to that from the Time Reversal Mirror method, but is more computationally efficient. Lastly, at Shishaldin Volcano, Alaska, RTM and RTM-FDTD both produce realistic source locations (<50 m) for ground-coupled airwaves recorded on a four-station seismic network. We show that RTM is an effective method to detect and locate infrasonic sources across a variety of scenarios, and by integrating numerical modeling, RTM-FDTD produces more accurate source locations and increases the detection capability.