Flow speed estimated by inverse modeling of sandy tsunami deposits: Results from the 11 March 2011 tsunami on the coastal plain near the Sendai Airport, Honshu, Japan

Tsunami deposits contain information about the flow that created them which can be interpreted to estimate tsunami magnitude. Sandy deposits near Sendai Airport are modeled using an inverse sediment transport model to explore the spatial and temporal variation of tsunami flow speed in the 11 March 2011 Tohoku-oki tsunami. The 2011 deposits contain grain-size signatures of both sediment transport convergences and sediment settling out of suspension. Sediment transport convergences form massive or inversely graded intervals. Sediment falling out of suspension forms a specific type of normal grading, termed suspension grading. The Jaffe and Gelfenbaum (2007) inverse sediment transport model, which relates the thickness and bulk grain size of suspension-graded intervals to tsunami flow speed, was applied to 24 suspension-graded intervals, ranging in thickness from 2 to 4. cm, at 7 trenches located from about 250 to 1350. m inland from the shoreline. Modeled tsunami flow speeds range from 2.2 to 9.0. m/s, and were strongly dependent on the choice of Manning's n roughness parameterization. Flow speeds were highest from about 75-300. m landward of the forested sand dunes where the tsunami encountered lower roughness in a low-lying area as it traveled downslope, and decreased by 16% as the tsunami moved inland over the subsequent 596. m. As many as 5 suspension-graded intervals, interpreted as forming during the onshore flow of up to 5 waves, were identified at each location, with a tendency for an inland decrease in the number of suspension-graded intervals suggesting fewer waves inland. Tsunami flows tended to be fastest in the lower two intervals (earlier waves) and slowest in the uppermost interval (last wave) of the deposits. Modeled reconstructions of the suspension-graded intervals at a vertical resolution of 1-cm reproduced observed upward fining of the distributions reasonably well, supporting the use of the inverse model. Estimates of the time required to form 99% of the thickness of the suspension-graded intervals ranged 2 to 5. min, allowing that sediment falling from suspension could have formed these portions of the deposits in the time between waves. Application of the inverse model to determine paleotsunami flow speeds from deposits, although now possible, is complicated by uncertainty in roughness. As inverse approaches to estimating roughness and tsunami flow speed improve, the magnitude of paleotsunamis will be better understood and the ability to assess tsunami hazard from paleotsunami deposits will improve.