Making Hydrophones Speak: Semi-Automated Hydroacoustic Monitoring of a Seafloor Spreading Event

In April 2024, a major submarine eruption occurred at the I1 segment of the Southeast Indian Ridge (SEIR), near Amsterdam Island in the southern Indian Ocean (see Royer et al., EGU26-GD5.1 and Olive et al., EGU26-GD5.1). Luckily, the OHA-GEODAMS Autonomous Hydrophone (AuH) network had been deployed a few weeks before the event and efficiently monitored the eruptive event. This work presents results from this hydroacoustic monitoring together with an innovative automatic cataloging pipeline that allowed to obtain a full spatio-temporal coverage of the event.

Thanks to the low attenuation of low-frequency hydroacoustic waves, hydroacoustics is known to be an efficient way to monitor earthquake swarms. In this case, the GEODAMS network efficiently monitored the eruptive swarm and detected both direct (P- and S-phases) and indirect (T-phases) seismic waves, as well as H-waves generated by interactions between lava and seawater.

During the first weeks of hydroacoustic activity, more than 500 T-waves and 200 H-waves were detected and their sources located. This enabled a precise relocation of the early swarm of strong, teleseismically-recorded earthquakes (Mw~5) that had been registered in the ISC and GCMT catalogs. It also enabled the detection of smaller events that had been missed by land stations. However, this initial analysis relied on manual annotation, a process that is particularly time-consuming and hinders the possibility to build reliable and near-exhaustive catalogs over extended time periods. To overcome this limitation, a fully automated cataloging pipeline was developed (Raumer et al., 2024 & 2025), enabling systematic detection, association, and location of hydroacoustic signals associated with the swarm. This methodology enabled us to estimate key parameters of the eruptive dynamics, such as the precise timing of sequential dyking events, and subsequent lava outpouring, which was later revealed by diachronous seafloor mapping. Moreover, the seismological catalog showed to be complementary with vertical deformation measurements (detailed in Ballu et al., EGU26-SM3.2).

Beyond its own geodynamical significance, the 2024 I1 eruption constitutes a relevant case study to demonstrate the strong potential of an automated approach for hydroacoustic earthquake monitoring. Future applications of this generic methodology should enable the extraction of geodynamical insights from past and potentially large-scale AuH observatories.

Raumer et al. (2024). Seismica. doi: 10.26443/seismica.v3i2.1344

Raumer et al. (2025). Geochem., Geophys., Geosys. doi: 10.1029/2025GC012572