We develop a new method to analyze the total electron content (TEC) depression in the ionosphere after a tsunami occurrence. We employ Gaussian process regression to accurately estimate the TEC disturbance every 30 s using satellite observations from the GNSS network, even over regions without measurements. We face multiple challenges. First, the impact of the acoustic wave generated by a tsunami onto TEC levels is non-linear and anisotropic. Second, observation points are moving. Nevertheless, our method always computes these volumes, along with estimated uncertainties, when applied to the 2011 Tohoku-Oki Earthquake, even with random selections of only 5% of the 1,000 GPS Earth Observation Network System receivers considered here over Japan. The method can warn of a tsunami event within 15 minutes of the earthquake, at high levels of confidence, even with a sparse receiver network. Hence, it is potentially applicable worldwide using the existing GNSS network.

In this paper, we model the full range of possible local impacts of future tsunamis in the Makran subduction zone (MSZ) at Karachi port, Pakistan. For the first time, the 3-D subduction geometry Slab2 is employed in the MSZ, in conjunction with the most refined rupture segmentation to date for this region, to improve the earthquake source definition. Motivated by the massive sediment layer over the MSZ, we also introduce to tsunami modeling the application of the sediment amplification formula, resulting in enhancements of seabed deformation up to 60% locally. Furthermore, we design a new unstructured mesh algorithm for our GPU-accelerated tsunami code in order to efficiently represent flow velocities, including vortices, down to a resolution of 10m in the vicinity of the port. To afford to compute very large number of high resolution tsunami scenarios, for the granularity and extent of the range of magnitudes (occurrence ratios of 1:100,000 implied by the Gutenberg-Richter relation) and locations of source, we create a statistical surrogate (i.e. emulator) of the tsunami model. Our main contribution is hence the largest set of emulated predictions using any realistic tsunami code to date: 1 million per location. We go on to obtain probabilistic representations of maximum tsunami velocities and heights at around 200 locations in the port area of Karachi. Amongst other findings, we discover substantial local variations of currents and heights. Hence we argue that an end-to-end synthesis of advanced physical, numerical and statistical modeling is instrumental in coastal engineering to comprehensively model local impacts of tsunamis.