Tsunami hazard assessment along the U. S. East Coast

In 2005, the National Tsunami Hazard Mitigation Program (NTHMP) was tasked by Congress to develop tsunami inundation maps for the entire US coastline. This work provides an overview of the modeling work related to the development inundation maps along the US east coast. In this region the paucity of historical tsunami records and lack of paleotsunami observations yields a large uncertainty on the source and magnitude of potential extreme tsunami events, and their related coastal hazard. In the Atlantic Ocean basin significant tsunami hazard may result from far-field earthquakes, such as a repeat of the M8.9 Lisbon 1755 event in the Azores convergence zone, or a hypothetical extreme M9 earthquake in the Puerto Rico Trench (PRT). Additionally, it is believed that a repeat of one of the large historical collapses, identified at the toe of the Cumbre Vieja volcano on La Palma (Canary Islands; i.e., with a maximum volume of 450 km³), could pose a major tsunami hazard to the entire US east coast. Finally, in the near-field, large submarine mass failure (SMF) scars have been mapped by USGS, particularly North of the Carolinas (e.g., Currituck), which are believed to have caused past tsunamis. Large SMFs can be triggered by moderate seismicity (M7 or so), such as can occur on the east coast. In fact, one of the few historical tsunamis that significantly affected this region was caused by the 1929 Grand Bank underwater slide, which was triggered by a M7.2 earthquake. In this work we identify and parameterize all potential tsunami sources affecting the US east coast, and perform simulations of tsunami generation, propagation, and coastal impact in a series of increasingly resolved nested grids. Following this methodology, tsunami inundation maps are currently being developed for a few of the most affected areas. In simulations, we use a robust and well-validated Fully Nonlinear Boussinesq long-wave model (FUNWAVE-TVD), on Cartesian or spherical grids. Coseismic tsunami sources are modeled using the standard Okada method. For landslide tsunamis, we first generate tsunami sources using a three-dimensional Navier-Stokes model (THETIS or NHWAVE). These models feature all relevant physical processes, such as frequency dispersion (very important for landslide sources), nonlinear wave effects during shoaling, and dissipation by bottom friction and wave breaking (via a shock-capturing TVD algorithm). In modeling coastal hazard from various sources, we find that tsunamigenic SMFs, which are the nearest tsunami sources and can potentially cause highly focused coastal runup, may control tsunami hazard for many east coast communities north of the Carolinas. In many cases, however, we find that a wide shallow continental shelf may cause significant dissipation of the shorter waves caused by SMFs and hence offer some protection. The accurate modeling of the delicate balance between nonlinear and dissipative processes governing such situations is currently being researched and will be the object of a separate presentation. Additionally, considerable efforts are being devoted to properly parameterizing extreme SMFs, which are also the object of collaborative work with geologists and marine geotechnical experts (reported independently).