Skip to content

Using Low-frequency DAS Signals for Early Warning During the Sundhnúksgígur, Iceland, Eruptions in 2024 Authors/Affiliations

1Vala Hjörleifsdóttir, 1Hrefna Marín Gunnarsdóttir, 2Jiaxuan Li, 3Ettore Biondi, 4Kristín Jónsdóttir, 4Sara Barsotti, 4Einar Bessi Gestsson, 5Valey Kamalov,  3Zhongwen Zhan

  • Affiliations: 1 University of Reykjavík, Iceland; 2Houston University, Texas, USA; 3Caltech, Pasadena, USA; 4Icelandic Meteorological Office, Reykjavik, Iceland; 5Valey Kamalov LLC, Gainesville, USA

  • Presentation type: Talk

  • Presentation time: Monday 14:45 - 15:00, Room S160

  • Programme No: 2.4.13

  • Theme 2 > Session 4

Abstract

The Reykjanes peninsula, Iceland, has been experiencing unrest since 2019.  Since Nov 2023 the activity has been focused along the fissure swarm of Sundhnúksgígur, with 10 intrusive events of which 7 were eruptive. The eruptive fissure is located just north of the town of Grindavík, and about 2 km away from the Svartsengi geothermal power plant and the Blue Lagoon spa, a major tourist attraction.  Evacuations, based on increased seismicity at the center of the activity, together with pressure signals on a borehole sensor within the geothermal field, have been issued as little as 30 minutes before the eruption.   From Nov 2023-Nov 2024 a DAS interrogator was running in the peninsula, converting a 100-km-long telecommunication fiber cable crossing Grindavík into a dense array with 10,000 strain-rate sensors. Clear low-frequency signals are visible on the fiber more than 30 minutes before each eruption. We developed a warning system using the amplitude of low-frequency strain rate (LFDAS).  The alert has been running at the Icelandic Meteorological Office (IMO) since April 28th, 2024.  The strain rate observed on the fiber, together with seismicity and pressure changes on a borehole pressure sensor, have caused the IMO to issue and to refrain from issuing warnings, demonstrating the immediate utility of the LFDAS observations.  Furthermore, the similarity of strain signal in the events was used to predict the timing of the August 2024 eruption, to within 10 minutes. In this presentation, we describe the methodology and results from the LFDAS monitoring of the ongoing Sundhnúksgígur sequence.