NTHMP Mapping and Modeling Benchmarking Workshop: Tsunami Currents

For a description of the workshop and all data provided, see the workshop webpage.

For GeoClaw code, see the GitHub repository

Results for other problems

GeoClaw results for Benchmark problem 2

Results presented at workshop:

Problem description
2/3 arcsecond (20 m) resolution: Plot Index
1/3 arcsecond (10 m) resolution: Plot Index
1/6 arcsecond (5 m) resolution: Plot Index ... Vorticity animation
Comparison figures

Notes:

Adaptive mesh refinement was used with the resolution near the top boundary fixed at 2 arcseconds (180 m) to simplify specifying uniform boundary forceing, but the bulk of the domain was refined to the resolution specified above.
The computational domain was longitude \(204.905 \leq x \leq 204.965\), and latitude \(19.71 \leq y \leq 19.758\). The control point where the incident wave information was specified is (204.93, 19.7576). The boundary conditions were set at \(y = 19.758\) based on this data. Each time step, the depth \(h\) and momentum \(hu\) in the grid cell containing this point is used to determine an incoming plane wave moving in the \(y\)-direction necessary to match the data at this point. At other points along the boundary, the ghost cell data is chosen so that the incoming wave matches this.
The control point data provided starts at \(t=7\) hours and shows the peak of the first tsunami arriving at about \(t=8.5\) hours. This is inconsistent with the measured data at TG 7760 and the ADCP gauges, as noted in the problem description. The simulation results were shifted by 0.4 hours to better match the measured data.
The control point data provided starts at a surface elevation of \(-0.13\) m rather than at 0. The script load_offshore_wave.m provided shifts all the data by the first data point before plotting the figure. We also shifted and used the resulting wave (starting at elevation \(\eta = 0\) ) as the input on a simulation where the sealevel was set to 0 relative to the bathymetry provided (corresponding to MHW). In fact the tsunami arrived at a lower tide stage (but not \(-0.13\) relative to MHW as far as we can tell).

Full problem from Tohoku source

We also ran this problem using two different versions of a Tohoku source.

Fujii, Satake, et al. source: Plot Index
Chen Ji, et al. (UCSB) source: Plot Index
Comparison at gauges
Note that these solutions do not agree with the provided data at the control point (labelled Gauge 3333 in our simulations).
The higher frequency signal seen in the Tide gauge 7760 is not captured by these simulations. It is not clear to us where this comes from, since the tsunami arriving at Hilo seems to have a longer wave length, as observed in this animation from the UCSB source run.