Data and Models for "Probabilistic Imaging of Tsunamigenic Seafloor Deformation During the 2011 Tohoku-oki Earthquake"

Directory "waveform_data" includes 13 tsunami time series data from different instruments: TM1, TM2, KPG1, KPG2, GB801, GB802, GB803, GB804, GB806, GB807, D21401, D21413, and D21418. Each data file (*.dat) has two columns for (1) the time since earthquake initiation (min) and (2) ocean surface or seafloor displacement amplitude (m).

Directory “kin_models” includes the following:

1. Seafloor Mesh Geometry

• The entire seafloor mesh consists of two separate parts (422 and 136 nodes each; 558 in total) due to the need to resolve potential discontinuity at the trench. The files “*Pt{1,2}-SM2.PointCoord.txt” consists of six columns for the ID, longitude (deg), latitude (deg), East (km), North (km) and depth (km) of the nodes in each triangular mesh. The E/N coordinates are calculated in UTM projection system, relative to an arbitrary reference point.
• The files “*.ClipPath.txt” includes the ID/lon/lat of mesh boundary nodes, which can be used for plotting.
• Visualization of the mesh parts are provided in PDF files.
• The file “*Total-SM2.PointCoord.txt” excludes boundary nodes and contains seafloor locations (504 nodes) that are directly used in tsunami arrival time calculations.

2. Posterior Mean Models

• Ensemble-averaged models of seafloor displacements and uncertainty estimates, with no spatial averaging (“0R” in the file name) or with one-ring spatial averaging (“1R”). These models are shown in Figures 5 and 6 of Jiang and Simons (2016). The data files “posterior_mean_{0,1}R.txt” have three columns for (1) vertical seafloor displacement (m), (2) one-sigma standard deviation of displacement (m), and (3) corresponding resolution length (km). The model values (558 rows) correspond to nodes in files “*Pt{1,2}-SM2.PointCoord.txt” concatenated in sequential order.
• Tsunami arrival times (in sec) are calculated from the posterior mean values of propagation speeds, with zero sec at the earthquake epicenter. The coordinates (508 nodes) are included in geometry file “*Total-SM2.PointCoord.txt.”

Directory “kin_ensemble” includes the entire posterior model ensemble (98304 samples) in HDF5 format. Using a Linux command h5dump will show the following information about the contained datasets, with their names and dimensions. The main datasets are: (1) Covariance (1008×1008); (2) Data Log-likelihood (98304×1); (3) Posterior Log-likelihood (98304×1); and (4) Sample Set (98304×1008). Each model has 1008 parameters (504 for displacement and 504 for propagation speeds). The source coordinates (504 nodes) are included in geometry file "*Total-SM2-Parameter.PointCoord.txt."

Note: three different geometries files above are used for (1) posterior mean displacements (558 nodes), (2) arrival time calculation (508 nodes), and (3) source inversion models (504 nodes).