Multidisciplinary Center for Earthquake Engineering Research logo google logo
navigation bar

navigation bar

Instructions to Participants


The objective of this two-day workshop is to evaluate the current methods for developing rock time histories for engineering applications in the Eastern United States (EUS). There are two issues to consider in this regard: the amplitude of the ground motion (e.g. response spectra) and the character of the time history. The following questions will be addressed during the workshop: Which methods can be used to reliably predict the amplitude of the ground motion (median and variability) for defining design spectra? Which methods can be used to define the non-stationary characteristics of the time history? Should time histories be scaled? If so, what are the scaling rules? How can the results of the simulated time histories be evaluated to ensure that they are reasonable?


To allow an evaluation of various models and methodologies, both a validation exercise and a simulation exercise will be conducted. These exercises will be performed by each of the ground motion modelers using their proposed methods and models. The results of the exercises will be submitted by each modeler to Dr. Norman Abrahamson prior to the workshop so that they can be evaluated and presented in a standardized format for use during the workshop.

1. Validation Exercise

The validation is intended to evaluate how well the models can predict the ground motion from previously recorded earthquakes. The comparisons will be made for response spectral values, peak velocity, peak displacement, and duration. Ideally, this validation exercise should include a large number of earthquakes; however, that is beyond the scope of this workshop. For this workshop, only a single event validation is being requested.

There are two moderate-to-large events available for the EUS: the 1988 Saguenay earthquake, and the 1985 Nahanni earthquake. The Saguenay event is the best recorded EUS event with M>5 (distances 40-150 km), but it may have an anomalous source. The Nahanni event has closer distances (8-16 km) with a larger magnitude, but it did not occur in the EUS (although the spectral content looks like EUS ground motion).

For this exercise, the Saguenay event has been selected. A drawback to this selection is that many seismologists consider it to be anomalous. This begs the question that if a model produces a good match to Saguenay, will that model then produce a poor match for other more typical earthquakes? To help address this, if a model has been validated against other earthquakes, that information should be provide by the modeler as well.

The following stations from Saguenay should be modeled:

Code Station Name

Dist (km )

Code Station Name

Dist (km)

SM01 St. Ferreol, Quebec


SM10 Riviere-Quelle, Quebec



Quebec, Quebec


SM16 Chicoutimi-Nord, Quebec


SM05 Tadoussac, Quebec


SM17 St-Andre-du-Lac-St-Jean


SM08 La Malbaie, Quebec


SM20 Les Eboulements, Quebec


SM09 St-Pascal, Quebec



From the validation exercise, the bias of each model will be evaluated (i.e., is the model adequate on average?) along with the variability of the residuals of the model predictions. This is needed to define the total standard deviation of the prediction. (If validations from a larger number of events have already been computed, modelers should include these results as well).

2. Simulation

The simulation exercise is not for a specific location. The event parameters are given below:

  • Site condition: hard rock, kappa = 0.006 sec
  • Moment magnitude: 7.0 (moment = 3.55 E26 dyne-cm)
  • Dip: 45 degrees (E)
  • Strike: 0
  • Rake: 90
  • Dimension: 50 km x 20 km
  • Top of fault: 2 km
  • Velocity structure: EPRI Midcontinent model
  • Distances: 0 - 500 km
  • Source realizations: (as appropriate for each model)
    Each of the source parameters that were optimized in the validation exercise should be varied in this simulation. This may include the following: slip model, hypocenter location, sub-event parameters. A minimum of 10 source realizations should be run to define the parametric variability term (30 is better, but 10 is sufficient for this workshop).
  • Fault Coordinates: (x, y, z) in km
  • Top of fault: (0,25,2) (0,-25,2)
    Bottom of fault: (14.14,25,16.14) (14.14,-25,16.14)
  • Site coordinates:
  • 1 (-10,5) 9 (150,5) 17 (50,15) 25 (5,25)
    2 (5,5) 10 (200,5) 18 (80,15) 26 (5,40)
    3 (10,5) 11 (300,5) 19 (100,15) 27 (5,80)
    4 (25,5) 12 (500,5) 20 (120,15) 28 (5,100)
    5 (50,5) 13 (-10,15) 21 (150,15) 29 (5,120)
    6 (80,5) 14 (5,15) 22 (200,15) 30 (5,150)
    7 (100,5) 15 (10,15) 23 (300,15)    
    8 (120,5) 16 (25,15) 24 (500,15)    
  • Velocity Structure:
  • Vp(km/s)


    Density (gm/cc)

    Thickness (km)

    Depth to top (km)





















  • File Format:

A separate file should be created for each component at each site and for each run. The file should be in ASCII, free format, with a 5 line header followed by acceleration values in g. Use either a ZIP, STUFFIT, or compressed tar file format.

The header should consist of the following lines:

  1. Modeler name
  2. Simulation number (source realization)
  3. Site coordinate
  4. Component (N, E, or Z, corresponding to positive North, East, and Down, respectively)
  5. Number of points, sample interval in seconds (i.e., delta t)


The validation and simulation data files must be provided to Norm Abrahamson by October 9.


top.gif (1113 bytes)

  Contact Us  |  Acknowledgements   |  Disclaimer  |  Copyright© 2007 by the Research Foundation of the State of New York. All rights reserved.