Magnetic fields generated by tsunamis: Case studies on the 2009 Samoa and 2010 Chile earthquake tsunamis

概要

Tsunami magnetic fields are generated by conductive sea water movement by tsunamis in the presence of the geomagnetic main field. Recent studies have revealed that the arrival of tsunami magnetic fields is earlier than tsunami sea level changes themselves. Moreover, the wave height of tsunamis as well as their propagation directions can also be estimated by the tsunami magnetic fields. These characteristics make the tsunami magnetic fields have two potential applications: 1) tsunami early warning and 2) a new data source of tsunami wave heights. However, the early arrival of the tsunami magnetic fields and their promising potentials to estimate not only propagation directions but also wave heights have not been confirmed by a direct comparison of the simultaneously observed tsunami magnetic fields and sea level changes. In other words, the characteristics of the tsunami magnetic field still need examination before applying them to practical problems such as disaster mitigation and/or tsunami early warning. We, therefore, used the simultaneously observed vector magnetic and pressure data on a deep seafloor for the 2009 Samoa and 2010 Chile tsunamis. Those data were from the “Tomographic Investigation by seafloor ARray Experiment for the Society hotspot (TIARES)” project. This study focuses on examining the possible applications of the tsunami magnetic fields to tsunami early warning, which includes the sensitivity of the tsunami magnetic fields to the tsunami sea level change and the accuracy of both propagation direction and wave height estimated by the tsunami magnetic field.

The ocean bottom electromagnetometer array and a differential pressure gauge at the 4000-5000m depth seafloor maintained by the TIARES project captured the tsunami magnetic and sea level change signals during the 2009 Samoa and 2010 Chile tsunamis. The tsunami magnetic signals were extracted from the observed vector tsunami magnetic data by adopting band-pass filters as well as external field correction. The tsunami sea level changes were obtained from the observed ocean bottom pressure data by a combination of deconvolution and band-pass filters. Comparison of these two different and independent tsunami quantities has revealed not only the indisputable presence of the tsunami magnetic fields but also the early arrival of the vertical tsunami component, bz. Estimation of tsunami propagation directions were also examined by two different methods. One using the tsunami first waves observed by the entire array gave a more accurate direction estimate several minutes after the tsunami arrival at the array, while the other based on the two tsunami horizontal components, bx and by, provided less accurate directions at each observation site on arrival. However, accuracy of the direction estimates depends naturally on signal-to-noise ratios of the tsunami magnetic fields and the activity of the background geomagnetic field of external origin. Tsunami sea level changes converted from the observed tsunami magnetic fields were proved to have very high accuracy by comparing them with the direct observation of seafloor pressure as well. The comparison indicated that both long-wave and linear-dispersion approximations are applicable to the accurate conversion of the tsunami magnetic data to the sea level change on the deep seafloor, provided that the tsunamis in concern are two-dimensional enough with weak dispersion in pelagic environments. The new evidence revealed by this study suggests that the tsunami magnetic field is useful for improvement of the existing global tsunami early warning system.

To further explain the observed tsunami magnetic field well, 3-Demensional time domain tsunami magnetic simulations were conducted. The calculations of tsunami velocity and magnetic fields in the TIARES area were achieved by open-source numerical simulation codes of JAGURS and TMTGEM, respectively. The simulations clarified, for the first time, that the discrepancy between the observed and simulated tsunami magnetic fields stems not from JAGURS or TMTGEM but from inaccurate tsunami source models. The clarification was also made possible by virtue of the simultaneous observation of both tsunami magnetic and sea level changes. An attempt to explain the converted tsunami wave heights from the magnetic signals was conducted to clarify whether they were applied to existing tsunami source model inversion schemes. The result showed us that we were able to incorporate the converted wave heights so as to yield satisfactory fits to the TIARES data. However, it also turned out that the new initial sea surface displacements were incompatible with the tsunami data outside the TIARES region, which implies that more comprehensive inversions are required to construct optimized tsunami source models for both far-field tsunamis using magnetic data.