Earthquake Source Model Database
The paper in Seismological Research Letters discuss the standards for documenting and disseminating finite‐fault earthquake rupture models, and related data and metadata.

Martin Mai, P., P. Shearer, J.P. Ampuero, and T. Lay (2016). Standards for documenting finite‐fault earthquake rupture models. Seismological Research Letters, 87(3), 712-718

For each source model in the database, the different data files include:

  • evTAG.slp : SLP (Slip) format file, which is an ascii-file with basic header information and simple, single-rupture-plane source-model representation.
  • evTAG.fsp : FSP (Finite Source Parameter) format file, which is an ascii-file with comprehensive list of modeling/inversion parameters, source geometry, and space-time-dependent rupture model.
  • evTAG.mat : MAT (MATLAB struture) format file, which is a binary MATLAB file containing the rupture model as MATLAB structure.
  • Miscellenous : Additional file formats in some cases.
  • SRF format (proposed by Rob Graves, and used in the SCEC rupture-modeling workflows): The .srf format provides a comprehensive space-time description of the kinematic rupture evolution by means of a distribution of point sources for which the slip-rate function (in cm/s) is explicitly given as a time series; two such slip-rate functions allow to include temporal rake variations; if desired, fault opening can be handled by specifying three slip-rate histories for each point on the fault. Thanks to Rob Graves for allowing us to use this format in the SRCMOD and SIV context.

evTAG identifies a specific model and is a character-string composed of the year of the earthquake, an abbreviation for the event location and an abbreviation for the author(s) of the source model. Earlier, the field comprises of 15 characters.

New scheme for evTAG

Presently, we have adopted 17 characters, which include two characters for the count (either number of models or earthquakes in same year from the same author).

The exact algorithm for evTAG is as follows:


  • YYYY: Year of the event
  • NNNNN: first six characters of the event name (which corresponds to the region or location) e.g. 'PARKFI' for Parkfield (Calif.). If name of event has less than six characters, we supplement it by 'x'.
  • CC: A count to define the number of the model. This is a new addition to the old scheme.
  • AAAA: First four letters of the author (citation) e.g. 'DREG' for Dreger et al (2004)
  • If the first author/citation name has less than four characters, we supplement it by 'x'.

©2012 Paul Martin Mai ♦ Disclaimer/Acknowlegments: see About page.