Signals in the ionosphere generated by tsunami earthquakes: observations and modeling suppor

Forecasting systems failed to predict the magnitude of the 2011 great tsunami in Japan due to the difficulty and cost of instrumenting the ocean with high-quality and dense networks. Melgar et al. (2013) show that using all of the conventional data (inland seismic, geodetic, and tsunami gauges) with the best inversion method still fails to predict the correct height of the tsunami before it breaks onto a coast near the epicenter (< 500 km). On the other hand, in the last decade, scientists have gathered convincing evidence of transient signals in the ionosphere Total Electron Content (TEC) observations that are associated to open ocean tsunami waves. Even though typical tsunami waves are only a few centimeters high, they are powerful enough to create atmospheric vibrations extending all the way to the ionosphere, 300 kilometers up in the atmosphere. Therefore, we are proposing to incorporate the ionospheric signals into tsunami early-warning systems. We anticipate that the method could be decisive for mitigating "tsunami earthquakes" which trigger tsunamis larger than expected from their short-period magnitude. These events are challenging to characterize as they rupture the near-trench subduction interface, in a distant region less constrained by onshore data. As a couple of devastating tsunami earthquakes happens per decade, they represent a real threat for onshore populations and a challenge for tsunami early-warning systems. We will present the TEC observations of the recent Java 2006 and Mentawaii 2010 tsunami earthquakes and base our analysis on acoustic ray tracing, normal modes summation and the simulation code SPECFEM, which solves the wave equation in coupled acoustic (ocean, atmosphere) and elastic (solid earth) domains. Rupture histories are entered as finite source models, which will allow us to evaluate the effect of a relatively slow rupture on the surrounding ocean and atmosphere.