Progress Report on Fling
Processing
• Baseline correction
Boore’s method (with judgement)
• Fling model and removal
Sine-Pulse model with 3 parameters
• Minimal effect on SA (T < 10 s)
07/19/04
NGA Workshop: Fling
1
To model fling we need to
accurately
recover the residual
TCU129,
E-W
displacement. Baseline
correction of recorded motion
is important step.
50
t2=30 sec
File: C:\metu_03\rec_proc\129evfit_color.draw;Date: 2003-09-10;Time: 14:25:09
Velocity (cm/sec)
t2=50 sec
t1
0
t2=70 sec
-50
t1: acc > 50 cm/sec2
t2(1): Boore v0
t2(2): acc < 50 cm/sec2
t2(3): 2 x t2(2) – t1
20
07/19/04
40
60
Time (sec)
80
NGA Workshop: Fling
2
Boore’s t2 tends to give reliable results
• Unrealistic results omitted (by judgement)
• Still non-unique (but little significance on fling removal)
07/19/04
NGA Workshop: Fling
3
Fix Tf (pulse period)
using numerical
simulations, then
solve for t1 (start
time)
07/19/04
NGA Workshop: Fling
4
Simulated Near-fault Displacements for Chi-Chi EQ
Measure time between 10% and 90% of final
displacement, adjust to sine pulse to determine Tf
07/19/04
NGA Workshop: Fling
5
Sine-pulse fit and corrected displacement for
TCU052 (090)
07/19/04
NGA Workshop: Fling
6
Sine-pulse fit and corrected displacement for
TCU068 (000)
07/19/04
NGA Workshop: Fling
7
Sine-pulse fit and corrected displacement for
TCU068 (090)
07/19/04
NGA Workshop: Fling
8
Conclusions
• Baseline correction must be done with care.
• Sine-pulse model for fling provides time domain
parameterization and can be related to
underlying physical processes
- residual displacement (D): fault offset
- pulse period (Tf): slip rise time
• Need to explore possible physical constraints on
pulse start time (t1)
- S wave arrival?
07/19/04
NGA Workshop: Fling
9
07/19/04
NGA Workshop: Fling
10