Published on April 18, 2025 Updated on April 29, 2025
Matoza et al. (2017) introduced an infrasound methodology with an associated tool (IMS_vASC) to achieve automated detection and cataloging of explosive volcanism. Using multiyear (2005- 2010) detections lists from all the available IMS infrasound network stations, a brute-force, gridsearch, cross-bearings approach allowed global discrimination of volcanic eruptions from other natural and human-made sources (e.g., microbaroms and dams). However, the location of the source obtained with IMS_vASC could have a high deviation from true due to the simplified homogeneous atmospheric model it assumes by default (e.g., Matoza et al., 2017, 2018). This is an expected effect of strong seasonal horizontal winds at stratospheric heights causing crosswinds that modify the ray path of the signals, which a can change the apparent azimuth detected at the IMS stations up to ~12° in the most extreme cases (e.g., Le Pichon et al., 2005; De Negri et al., 2022).


 

During my PhD studies, I focused on developing and testing a methodology to reduce the intrinsic mislocation of the sources due to atmospheric cross-winds previously observed. The Automatic Rapid Climatological Azimuth Deviation Estimation algorithm (ARCADE; https://github.com/rodrum/arcade) can be described as an eigenray search algorithm that combines the Horizontal Wind Model (HWM14; Drob et al., 2015) with the NRLMSIS2.0 model (Emmert et al., 2021) to forward model infrasound propagation with a 3D ray tracing algorithm (infraGA; Blom & Waxler, 2012). The method primarily allows finding predictions of the azimuth deviation at any location on the Earth from any source location and time of the year, producing look-up tables accompanied by lists of ground intercepts that also contain other useful arrival parameters (e.g., RMS amplitude, travel time, celerity). It can also be extended for in-deep studies that require a more accurate atmospheric modelling, like range-dependent with hybrid descriptions based on ERA 5 ECMWF (e.g, De Negri et al., 2022). With ARCADE, we achieved a rapid, first-order, and flexible forward modelling approach that could be used towards several goals in infrasound studies (De Negri et al., 2022; De Negri and Matoza, 2023 (in review); and De Negri et al., 2023 (in writing)). The look-up tables of predicted azimuth deviations help to correct the source location estimation of IMS_vASC (Figure 1), showing we could reduce the source mislocation from hundreds to tens of kilometers (De Negri and Matoza, 2023 (in review)). Further, year-long models show promising results that compare well with ground-truth infrasound observations (De Negri et al., 2023 (in writing)), allowing us to discriminate and understand the temporal evolution of the multi-year infrasound signals characteristics for study cases of remote active volcanoes with little direct evidence of activity in the past (De Negri et al., 2022).


 

Given the extensive research on volcanic subaerial eruptive processes, I think it is better to treat the research questions that could be addressed with this study by separating them in three main areas. First, regarding historical analysis of eruptions: how to use the frequency and acoustic magnitude derived from infrasound records to elaborate remote characterization methods. Second, concerning remote parameterization of volcanic eruptions: how the presented methodologies can improve the location, chronology, and ash plume height determination. And third: evaluate the added-value and performance of the combined infrasound+satellite methodology.
 

Figure 1: Schematic of a Plinian eruption, generating seismic waves through the solid Earth, and acoustic waves (infrasound) through the atmosphere.
Figure 1: Schematic of a Plinian eruption, generating seismic waves through the solid Earth, and acoustic waves (infrasound) through the atmosphere.

Figure 1: Schematic of a Plinian eruption, generating seismic waves through the solid Earth, and acoustic waves (infrasound) through the atmosphere.

 

Since 2003, the Comprehensive Nuclear-Test-Ban Treaty Organization (CTBTO ; www.ctbto.org) established the Internationel Monitoring System (IMS) infrasound network, which currently (2025) has 53 (of 60 planned) infrasound stations continuously recording to detect any nuclear explosion on Earth (see Figure 2).



Figure 2: IMS infrasound network (black/grey squares), holocene-active volcanoes (red triangles), and plate tectonic boundaries (dashed/dotted lines).
Figure 2: IMS infrasound network (black/grey squares), holocene-active volcanoes (red triangles), and plate tectonic boundaries (dashed/dotted lines).

Figure 2: IMS infrasound network (black/grey squares), holocene-active volcanoes (red triangles), and plate tectonic boundaries (dashed/dotted lines).

 

A software prototype for long-range volcanic eruption notification called "Volcanic Information System" (VIS) was developed within the Atmospheric dynamics Research InfraStrucure in Europe (ARISE) project (FP7, H2020), in collaboration with the Toulouse Volcanic Ash Advisory Center (VAAC).

The VIS main goal is to detect volcanic eruptions at regional to global distances (15-250 km; >250 km) with sustained ash-columns and provide early warnings to mitigate the risk that eruptions pose to civil aviation. Additionally, it can reconstruct the chronology of eruptions, and provide volcanic source constraints (acoustic intensity, gas flow, etc.).


The VIS is designed to leverage the IMS global infrasound detections and any available local infrasound station to monitor volcanoes. The detections are calculated with the Progressive Multi-Channel Correlation (PMCC) method (Cansi, 1995 ; Cansi and Le Pichon, 2008), which separates coherent infrasound waves (detections) from incoherent signals (noise). The VIS uses the Infrasound Parameter (IP) criterion (Ripepe et al., 2018 ; Marchetti et al., 2019 ; Gheri et al., 2023 ; Gheri et al., 2025) to establish when an eruption is in course, accounting for atmospheric propagation effects (Le Pichon et al., 2012 ; De Negri et al., 2023 ; De Negri et al., 2025), detection persistency, and amplitude (see Figure 3).


Figure 3: The VIS main steps to find a volcanic eruption.
Figure 3: The VIS main steps to find a volcanic eruption.

Figure 3: The VIS main steps to find a volcanic eruption.


Recently, we expanded the capabilities of the VIS to directly use streamlined and standardized IMS-derived infrasound array signal processing data products (Hupe et al., 2023). These are open-access (OA).
We are currently performing regional multi-year test of volcanic notifications with the

VIS for the Toulouse VAAC area (see Figure 4), covering major historical eruptions (e.g., Etna, Piton de la Fournaise, Eyjafjallajoküll), and improving the methodology behind the VIS (e.g., azimuth filter by using back-azimuth deviations ; De Negri and Matoza, 2023, 2025).


Figure 4: Volcanic Ash Advisory areas of the world and the number of advisories for each volcano. Original figure from Engwel et al., 2021.
Figure 4: Volcanic Ash Advisory areas of the world and the number of advisories for each volcano. Original figure from Engwel et al., 2021.

Figure 4: Volcanic Ash Advisory areas of the world and the number of advisories for each volcano. Original figure from Engwel et al., 2021.


With the IMS detections and OA products (2003-2022), added to the available HOTVOLC (https://hotvolc.opgc.fr; Gouhier et al., 2020) webGIS satellite notifications (2010-2022), we built a preliminary catalog showing to what extent infrasound-only, and infrasound+satellite monitoring can achieve reliable eruption notifications in the area (e.g., Figure 5).
 

Figure 5: Left: area of interest (similar to Toulouse VAAC) containing stations (squares/romboid) and volcanoes colored by neaerest station (see colorbar on the left). Right: catalog of preliminary VIS notifications, HOTVOLC notifications, Tolouse VAAC re
Figure 5: Left: area of interest (similar to Toulouse VAAC) containing stations (squares/romboid) and volcanoes colored by neaerest station (see colorbar on the left). Right: catalog of preliminary VIS notifications, HOTVOLC notifications, Tolouse VAAC re

Figure 5: Left: area of interest (similar to Toulouse VAAC) containing stations (squares/romboid), and volcanoes colored by neaerest station (see colorbar on the left). Right: catalog of preliminary VIS notifications, HOTVOLC notifications, Tolouse VAAC red-coded notifications, and eruptive periods of the Global Volcanism Program (see legend on the bottom).
 

The data products of the VIS demonstrator will be available through an application programming interface (API) hosted at the Observatoire de Physique du Globe de Clermont-Ferrand (OPGC, CNRS-INSU and University Clermont Auvergne), where also an archived catalogue of European volcano eruptions and the real-time data products for AMT (Firenze, Italy) will be hosted.
 

As part of the European Geo-INQUIRE project (HORIZON-INFRA-2021-SERV-01), the VIS will be integrated into the Thematic Core Service Volcano Observation (TCS-VO) of the European Plate
 

Observing System (EPOS). Future developments will include integration into web services such as the HOTVOLC web-GIS interface (OPGC, CNRS-INSU) or the EPOS Data Portal.



Rodrigo de NEGRI – LMV