Quick faultplane identification by a geometrical method: The Mw6.2 Leonidio earthquake, 6 January 20

1
Quick fault-plane identification by a geometrical
method The Mw6.2 Leonidio earthquake, 6 January
2008, Greece and some other recent applications
 J. Zahradnik, F. Gallovic Charles University
in Prague E. Sokos, A. Serpetsidaki, G-A.
TselentisUniversity of Patras
2
Why we need to know the fault plane ?

• Shake maps
• Aftershock prediction
• Stress field

3
Why we need to know the fault plane quickly?

• Shake maps
• Aftershock prediction

4
Which nodal plane is the fault plane ?

• Aftershock distribution
• Finite-extent source models waveform modeling
• Geometrical configuration of hypocenter (H) and
  centroid (C)

5
Which nodal plane is the fault plane ?

• Aftershock distribution too slow
• Finite-extent source models waveform modeling
• Geometrical configuration of hypocenter (H) and
  centroid (C)

6
Which nodal plane is the fault plane ?

• Aftershock distribution too slow
• Finite-extent source models waveform modeling
  too slow
• Geometrical configuration of hypocenter (H) and
  centroid (C)

7
Which nodal plane is the fault plane ?

• Aftershock distribution too slow
• Finite-extent source models waveform modeling
  too slow
• Geometrical configuration of hypocenter (H) and
  centroid (C) quick enough !

8
H-C method
H and C are in the same plane (I or II) of the
conjugated fault-plane solutions. H-C distance
must be larger enough Mgt6.
Multiple H and C solutions (uncertainty) help to
prefer one of the two planes.
9
H-C method continuation
Success depends on the particular focal
mechanism

• Easy case of strike slip (Epicenter is
  sufficient)
• Good case of one horiz. plane (H depth not very
  critical)
• Bad case inclined planes
• (H depth is critical)

10
H-C method continuation
Problematic applications

• A segmented fault
• A symmetric case
• H on intersection of I and II

11
H-C method applied to five Mgt6 events in 2008
12
H-C method applied to five Mgt6 events in
2008This presentation 2 examples
13
Example 1
14
M 6.2 Leonidio, Jan 6, 2008 depth 60-80 km
15
Waveform modeling for CMT
10 near-regional BB stations f lt 0.07 Hz
16
CENTROID
HYPOCENTER
17
Collective solutions including uncertainties
of H and CMT
H
H
red, green nodal planes of three CMT solutions
18
The weakly dipping nodal plane identified as the
fault plane
Strike 213 Dip 34 Rake 5
The green nodal plane is the fault
plane because it encompasses the (uncertain)
hypocenter.
19
animation
20
Practical output Report to EMSC within 1 week
after the earthquake report_jan06.pdf (in
Earthquake News Highlights)
21
Consistence with the regional stress
field(Kiratzi Papazachos, 1995)
T of this earthquake T of regional field
22
Slip vector and regional stress field allow us
to resolve thetraction and evaluate the Coulomb
Failure Function.
Validation without aftershocks ?
sub-horizontal slip vector
23
The Coulomb Failure Function supports the
sub-horizontal slip.
TVN negative
TVS tangential traction parallel to slip TVN
normal traction CFFTVSmTVN
TVN positive
24
CFF larger for plane II because TVN is positive
25
Example 2
26
Mw 6.3 Andravida June 8, 2008 depth 20 km
Strike 210 Dip 85 Rake 179
27
Mw 6.3 Andravida June 8, 2008 depth 20 km
H UPSL C Mednet
H UPSL and THE C Harvard
Strike 210, a right-lateral strike slip fault
28
Report to EMSC 7 hours after the
earthquake report_june08.pdf
29
Abundant aftershocks (24-hours, NOA) validate the
quick fault-plane guess (7 hours)
30
Strong-motion accelerograms (NOA) reveal a
different duration
Amaliada dist. 25 km
Patras dist, 35 km
31
Does a simple finite-extent source model based on
the H-C result explain the data?
Patras
Amaliada
32
Typical finite-source synthetics reproduce the
durationand support H-C results
Amaliada backward Patras forward
33
Further support azimuthal variation of the
differential travel time t-t
t hypocentre t asperity first brake
(After Takenaka et al., 2005)
34
Conclusion
 

• H-C method is a simple tool for quick
  identification of the fault plane
• Applicable with manual locations and CMT agency
  solutions (within a few hours)
• Collective solutions account for uncertainty
  through scatter in the H and C solutions
• So far the best validated June 8, 2008
  Andravida (gt rupture propagation to NE)

   
35
 
Full paper and e-supplement Seism. Res. Letters,
79, 653-662, 2008

• Try also a 3D animation tool (hcplot.m).

   
36
H-C geometrical method applied to five Mgt6
events in 2008

Event Fault plane less
likely Report
strike dip rake to
EMSC L
eonidio Jan 6 213 34 5 119 87
124 1 week Methoni Feb 14 311 14 95
126 76 89 1 day Methoni Feb 20
153 78 153 249 64 13 1
day Andravida Jun 8 210 85 179 300 89
5 7 hours Rhodos Jul 15 262 90
-38 352 52 -180 14 days

37
(No Transcript)