Broadband Data Products: Difference between revisions

From SCECpedia
Jump to navigationJump to search
Fsilva (talk | contribs)
No edit summary
Fsilva (talk | contribs)
No edit summary
Line 127: Line 127:


[[File:Gof-NR-10000000-anderson-summary.png]]
[[File:Gof-NR-10000000-anderson-summary.png]]
== References ==
# Anderson, J. G (2004) Quantitative Measure of the Goodness-of-Fit of Synthetic Seismograms, 13th World Conference on Earthquake Engineering, Vancouver, Canada, August 1-6, 2004, Paper No. 243.

Revision as of 18:42, 24 May 2016

This page contains a basic description of some of the data products produced by the Broadband Platform while running a simulation. All the data products described below will be in the outdata directory after a simulation completes.

Scenario Simulations

Scenario simulations are hypothetical simulations, what-if scenarios. Users provide a source description file describing the rupture and a station list file containing stations for which ground motions will be calculated.

Station Map

A map showing the fault trace along with the stations for which ground motions are calculated. The yellow star shows the hypocenter location.

Rupture Model

Velocity and Acceleration Seismograms

The Broadband Platform will produce both velocity and acceleration seismograms in BBP format for each station. An example is shown below.

# Sim NGAH, stat=2001-SCE
#    time(sec)      N-S(cm/s)      E-W(cm/s)      U-D(cm/s)
0.000000e+00    -4.532020e-05   2.096500e-05    1.781130e-04
2.500000e-02    -8.685250e-05   4.299330e-05    3.587300e-04
5.000000e-02    -1.234730e-04   6.599000e-05    5.400660e-04
7.500000e-02    -1.540990e-04   8.986040e-05    7.203370e-04
1.000000e-01    -1.777220e-04   1.145130e-04    8.978090e-04
...
...
1.022750e+02	9.947510e-04	2.843360e-06	-1.606000e-03
1.023000e+02	9.947510e-04	2.843220e-06	-1.606000e-03
1.023250e+02	9.947520e-04	2.843170e-06	-1.606000e-03
1.023500e+02	9.947520e-04	2.843150e-06	-1.606000e-03
1.023750e+02	9.947520e-04	2.843150e-06	-1.606000e-03

RotD50 for Each Station

A file containing the RotD50 values for 63 periods from 0.01s to 10s.

#  Psa5_N Psa5_E RotD50
#     63    0.0500
   0.0100 .45593E+00 .42564E+00 .40887E+00
   0.0110 .45659E+00 .42605E+00 .40943E+00
   0.0120 .45732E+00 .42650E+00 .41005E+00
   0.0130 .45812E+00 .42700E+00 .41074E+00
   0.0150 .45994E+00 .42813E+00 .41192E+00
...
...
   6.0000 .21332E-01 .63760E-02 .15788E-01
   6.5000 .19264E-01 .61713E-02 .14437E-01
   7.5000 .15063E-01 .50374E-02 .11496E-01
   8.5000 .11519E-01 .37930E-02 .88838E-02
  10.0000 .77129E-02 .28189E-02 .59970E-02

Validation Simulations

Validation simulations can be used to calculate seismograms from historical events and compare these again recorded ground motion data, allowing scientists to develop and validate their methods. For this reason, only locations where recorded ground motions are available can be used. In addition to all data products described above for scenario simulations, the following data products will also be calculated:

Seismogram Comparison Plots

Plots showing calculated data (in red) along with observation data (in black). The two bottom plots show a comparison of the arias duration between simulations and recorded data.

PSA Comparison Plots

Goodness-of-Fit Plot

Map Goodness-of-Fit Plot

Distance Goodness-of-Fit Plot

Anderson GOF

The Anderson GOF produces plots showing results from the following 10 metrics:

Anderson GOF Metrics
Index Metric
C1 Arias Duration
C2 Energy Duration
C3 Arias Intensity
C4 Energy Integral
C5 Peak Acceleration
C6 Peak Velocity
C7 Peak Displacement
C8 Response Spectra
C9 Fourier Spectra
C10 Cross Correlation

Per Station Plots

Summary Plots

References

  1. Anderson, J. G (2004) Quantitative Measure of the Goodness-of-Fit of Synthetic Seismograms, 13th World Conference on Earthquake Engineering, Vancouver, Canada, August 1-6, 2004, Paper No. 243.