Tsunamis are natural hazards that have already claimed the lives of more than 250000 civilians globally (Mizutori & Guha-Sapir, 2018). Tsunamis are commonly monitored on shores by coastal tide gauges or in deep oceans by tsunami buoys. These instruments provide direct measurements of the tsunami but can be insufficient for early warnings because (1) tide gauges are located on the coasts, giving little to no time for a warning, and (2) tsunami buoys are expensive to deploy and maintain, resulting in a limited sampling of the oceans, not sufficient for near-field warning. An alternative but indirect method centers around the computation of the ionospheric total electron content (TEC) to track tsunami propagation. TEC is a parameter commonly used to study and investigate the state of the ionosphere (Ratcliffe, 1951b), which is the layer containing the ionized part of Earth’s upper atmosphere and stretches from approximately 50km to more than 1000km. The established definition of the total electron content is the total number of electrons integrated between two points along a column of a meter-squared cross-section according to the following expression
$$TEC= \int {n}_{e}\left(s\right) ds$$
1
where \(ds\) is the integration path and \({n}_{e}\left(s\right)\) is the location-dependent electron density (Evans, 1957). There are different methods developed to obtain ionospheric TEC measurements from observations such as the Faraday Rotation effect on a linear polarized propagating plane wave (Titheridge, 1972). However, today TEC measurements are made mostly using GNSS (Global Navigation Satellite Systems) data. By utilizing the delay imposed by the ionosphere on the signal sent by a satellite, TEC values can be computed. For example, in the case of satellites equipped with dual-frequency systems, the ionospheric delay in meters is found according to
$$I= \frac{40.3 ({f}_{1}^{2}- {f}_{2}^{2})}{{f}_{1}^{2}{f}_{2}^{2}} {10}^{16} TEC$$
1
where \(I\) can be computed by taking the difference of the two measurements of pseudo-range or that of carrier phase obtained by a GNSS receiver station and \({f}_{1}\)& \({f}_{2}\) are the two frequencies used by the satellites to transmit signals back to the ground stations (Liu et al., 1996). The first tsunami-induced ionospheric (TEC) signature was presented by Artru et al. (2005) following the tsunami generated by the Jun. 23 2001 8.4 Mw Peruvian earthquake., and since, this technique has been used to identify and characterize the TEC signatures of a variety of tsunamis, all initiated by submarine earthquakes (Liu et al., 2006; Rolland et al.,2010; Galvan et al., 2011; Grawe & Makela, 2015, 2017). Underwater volcanic eruptions and landslides can also trigger tsunamis, except that there haven't been many large instances in the last decades to study them in the light of modern instrumentation. The 2022 explosion of the Hunga Tonga-Hunga Ha’apai (HTHH) submarine volcano provides a unique opportunity to fill this gap and characterize the generated ionospheric perturbations.
According to the US Geological Survey (USGS), the HTHH volcano (20.546°S 175.39°W; Fig. 1a) violently erupted on Jan. 15, 2022, at 4:14:45 UTC (17:14:45 LT). The eruption released a massive ash plume that reached an altitude of ∼55 km (Smart 2022). It also generated a highly-energetic atmospheric Lamb wave observed globally (for a few days after the eruption) in different types of measurements (e.g., barometers, infrasound sensors, satellites images, ionospheric measurements) (Matoza et al.,2022; Wright et al., 2022). According to Themens et al. (2022), large and medium-scale traveling ionospheric disturbances (TIDs) appeared in global TEC measurements following the eruption, with travel speeds ranging from 200 to 1000 m/s. They attributed the two TIDs types to the initial acoustic response of the explosive eruption and the energetic Lamb wave, respectively. The same findings were reported by Lin et al. (2022), where they also reported the presence of conjugate TIDs. In addition, Astafyeva et al. (2022) used the nearfield TEC measurements to identify the presence of several volcanic explosions during the event timeline. Moreover, the Lamb-wave overpressure coupled with the ocean triggering fast traveling air-sea (pressure-forced tsunami-like) waves observed worldwide (Kubota et al., 2022; Lynett et al., 2022; Omira et al., 2022). According to Matoza et al. (2022), the Lamb wave signature appears to be consistent (arrival time, waveform) in both the ionospheric and sea-level observations.
The eruption also produced a classical tsunami, i.e., from direct water mass displacement, detected across the Pacific Ocean (Carvajal et al., 2022), causing four casualties in Tonga (Latu, 2022) and two in Peru (Parra, 2022). The exact mechanism triggering the tsunami is not well-understood yet, but preliminary analysis suggests a combination of submarine explosion and caldera collapse (Hu et al., 2022 and reference therein). An ionospheric signature of this tsunami was reported by Matoza et al. (2022) at near-field. Here, we strengthen the study with a spatial pattern analysis and expand the investigated dataset more globally (Pacific-wide). We seek to isolate the ionospheric signature of the tsunami from the acoustic and Lamb signals. Because of these multiple, partially overlapping signals, we do not expect the discrimination to be straightforward, yet, it is a necessary step to assess the potential of TEC data for tsunami early-warning even in the case of a volcanic eruption.
To support our TEC signal analysis, we first analyze the case of the tsunami produced by the Mw 8.1 Kermadec earthquake, which occurred a year before, on March 4th, 2021 about 1000 km South of Tonga (29.723°S 177.279°W, based on the USGS report) (Fig. 1a). Both events occurred in the Eastern region of Polynesia islands sparsely equipped with GNSS stations installed onland. The size of the tsunami triggered by the Kermadec earthquake was smaller than the one triggered by the HTHH event by less than one order of magnitude (respectively 3 and 20 cm in the near-field after Romano et al., 2021 and Lynett et al., 2022). We thus use the Kermadec event as a test case to help decipher the HTHH tsunami-induced ionospheric signature with a sparse multi-GNSS network.
In addition to presenting the ionospheric signatures of the two tsunamis, we investigate how the tsunami generation mechanism (earthquake vs. volcano) affects their detection. We compare the tsunami sea-level variations to the identified ionosphere disturbance to confirm the tsunami origin of the detected ionospheric imprints. Finally, we examine the ionospheric response of the Lamb wave the HTHH eruption produced and compare it to that of the tsunami. Our goal is to discriminate the tsunami-induced ionospheric signature from the Lamb wave signature. The correct identification of the former is indeed critical for constraining the tsunami wave height in the ocean (Rakoto et al., 2018) and avoiding false alarms.