Locating surface deformation induced by earthquakes using GPS, GLONASS and Galileo ionospheric sounding from a single station

The 15 January 2022 Hunga Tonga Eruption History as Inferred From Ionospheric Observations

“…Also, the spherical wave method is based on the manual determination of the arrival time, which can introduce additional inaccuracy. Knowing that IonoSeis tends to systematically delay the arrival of disturbance with respect to observations (Lee et al., 2018; Mikesell et al., 2019; Zedek et al., 2021), we provide the final ionospheric solution for the onset times by averaging the solutions by two ionospheric methods (Table 1). We obtain 04:05:54 ± 169 s UT for the onset HTHH eruption trigger event and 04:18:10 ± 110 UT for the main big explosion.…”

The 15 January 2022 Hunga Tonga eruption history as inferred from ionospheric observations

“…Also, the spherical wave method is based on the manual determination of the arrival time, which can introduce additional inaccuracy. Knowing that IonoSeis tends to systematically delay the arrival of disturbance with respect to observations (Lee et al, 2018;Mikesell et al, 2019;Zedek et al, 2021), we provide the final ionospheric solution for the onset times by averaging the solutions by two ionospheric methods (Table 1). We obtain 04:05:54 ± 169 s UT for the onset HTHH eruption trigger event and 04:18:10 ± 110 UT for the main big explosion.…”

Near-real-time detection of co-seismic ionospheric disturbances using machine learning

“…In contrast to seismic media, atmospheric velocities, i.e., winds, are time-dependent which introduces further complexity when computing theoretical source-receiver arrival times. Fast simulations of acoustic wave propagation up to the ionosphere with realistic atmospheric specifications would greatly improve the classification between true and false arrivals and enable the localization of the largest surface displacements (Bagiya et al 2019;Inchin et al 2021;Zedek et al 2021). Finally, to confirm the detection of an earthquake across a given network and trigger an alert for human analysts, an additional heuristic could be implemented based, for example, on the number of detections per association class.…”

“…For instance, the amplitude of the CID scales almost linearly with the magnitude of an earthquake (Astafyeva et al 2013b(Astafyeva et al , 2014Cahyadi & Heki 2015;Occhipinti et al 2018;Heki 2021), or -for submarine earthquakes -with the tsunami wave height or volume of water that was displaced due to an earthquake (Kamogawa et al 2016;Rakoto et al 2018;Manta et al 2020). Additionally, CID arrival times and detection coordinates provide strong constraints on the position of the seismic source, or the origin of tsunami (Afraimovich et al 2006;Heki et al 2006;Tsai et al 2011;Lee et al 2018;Bagiya et al 2020;Inchin et al 2021;Zedek et al 2021). Moreover, Astafyeva et al (2011Astafyeva et al ( , 2013a; Astafyeva (2019) showed that the distribution of the first-detected CIDs match the position of the maximum displacement on the ground, and (Kakinami et al 2021) showed that the initial point of CID matches the maximum vertical displacement of the tsunami source.…”