Upload
piper
View
37
Download
0
Tags:
Embed Size (px)
DESCRIPTION
S eismic wave P ropagation and I maging in C omplex media: a E uropean network. IVO OPRSAL Experienced Researcher Host Institution: CUP Prague Place of Origin: Czech Republic Appointment Time: July 2005 - PowerPoint PPT Presentation
Citation preview
www.spice-rtn.orgMid-Term Review Meeting, February 13-14, Tutzing
Seismic wave Propagation and Imaging in Complex media: a European network
IVO OPRSALExperienced Researcher
Host Institution: CUP PraguePlace of Origin: Czech RepublicAppointment Time: July 2005
Project: Forward and inverse modelling of finite-extent earthquake source by PEXT and FD method.
Task Groups: Local scale
Cooperation: INGV Rome, GFZ Potsdam, ETH Zurich, AIST Tsukuba
www.spice-rtn.orgMid-Term Review Meeting, February 13-14, Tutzing
Project ScopeKinematic finite-extent source modelling
The Turkey-Flat strong motion "blind" prediction - internationalexperiment
The Turkey Flat strong-motion experiment is the forth of the blind prediction experiments, where the Prague group takes a part:
- Turkey Flat (Parkfield) weak motion test (1987)- Ashigara valley (Odawara, 1992)- Hyogo-ke n Nanbu earthquake (aka Kobe EQ) (Yokohama, 1998)- 2004 Parkfield M6 earthquake (USA, 2006),
The tests are conducted under ESG(Effects of surface geology working group of IASPEI/IAEE), the 2006 ESG meeting will take place in Grenoble.
More about the present Turkey-Flat strong-motion test can be found at the California Geological Survey page www.quake.ca.gov/turkeyflat.htm
www.spice-rtn.orgMid-Term Review Meeting, February 13-14, Tutzing
“Blind” Test Approach• Conduct high quality field and laboratory tests to characterize
the geotechnical properties of the site• Collect high-quality measurements of ground response in
sediment basin and bordering rock• Distribute only rock records and request predictions at basin
recording sites• Release observed basin recordings of and compare with
predictions
Charles R. Real and Anthony F. ShakalCalifornia Geological Survey
Turkey Flat, USA Site Effects Test Area: “Blind” Test of Predicted Ground Response of a Shallow Stiff-Soil Site to the September 28, 2004 M6.0 Parkfield Earthquake
Turkey Flat Working Group
Stay Tuned……..www.quake.ca.gov/Parkfield_2004
www.spice-rtn.orgMid-Term Review Meeting, February 13-14, Tutzing
Turkey Flat Site Effects Test Area
R1, D1
V1, D2, D3
V2R2
B
B’
A
A’
C C’
Geologic Structure
Phase I : R1 predictions (required synthetics: R2,V1,V2,D1,D2,D3)Phase II: D3 predictions (required synthetics: R2,V1,V2,D1,D2)
D3
D2
D1
R1 V1 V2 R2
www.spice-rtn.orgMid-Term Review Meeting, February 13-14, Tutzing
Turkey Flat strong-motion prediction by the finite-extent source modelling
Usual approach is to focus on local site effects while the excitation is represented by plane incident waves.
Composite-source method using a non-uniform fault slip distribution (Ji, 2005) is applied to all stations.
We emphasise the source effect. The reason is that: Firstly, the studied stations (R2, V1, V2, D1, D2, D3) are inside or around a very shallow stiff-soil sedimentary valley (local effects can be treated using individual 1D models below each receiver).Secondly, the dimensions of the ruptured area (W x L = 15 x 40 km, M6) are too large to be approximated by a planar wave.
www.spice-rtn.orgMid-Term Review Meeting, February 13-14, Tutzing
Turkey Flat situation
San Andreas Fault - projection of ruptured area
Turkey Flat
hypocenter
www.spice-rtn.orgMid-Term Review Meeting, February 13-14, Tutzing
Turkey Flat strong-motion prediction by the finite-extent source modelling
Free parameters of the source model are tuned by fitting the available acceleration time history and spectrum at R1 (reference rock site)
Turkey Flat Site Effects Test Area
R1, D1
V1, D2, D3
V2R2
B
B’
A
A’
C C’
Acc
eler
atio
n (m
/s^2
)
www.spice-rtn.orgMid-Term Review Meeting, February 13-14, Tutzing
Turkey Flat strong-motion prediction by the finite-extent source modelling
ADVANTAGE:Synthetics (0-40Hz) at the studied stations R2, V1,
V2, D1, D2, D3 (with individual 1D models) are obtained with the same source model as that used
for R1
i.e. realistic description of the incoming wave filed
www.spice-rtn.orgMid-Term Review Meeting, February 13-14, Tutzing
Turkey Flat strong-motion prediction results
www.spice-rtn.orgMid-Term Review Meeting, February 13-14, Tutzing
Availability of D3 record did not require any additional change.
This is an a posteriori justification of the
approach
Turkey Flat Site Effects Test Area
R1, D1
V1, D2, D3
V2R2
B
B’
A
A’
C C’
Turkey Flat strong motion Prediction, PHASE II
NO MORE WORK NEEDED !
www.spice-rtn.orgMid-Term Review Meeting, February 13-14, Tutzing
Outlook I
Modelling the data for predicted stations after they are released and benchmarking/comparison to solutions of other modellers.
www.spice-rtn.orgMid-Term Review Meeting, February 13-14, Tutzing
Outlook II
Weakness of discussed approach:ground motion modelling was tested at 2 points only (R1, D3)
Hence, general applicability of the source modelling will be tested using accelerograms from more CISN stations distributed around the fault.
0
45
90
135
180
225
270
315
0 4 8 12 16 20
36411
3645036230
36412
3644836176
3645536177
36420
36439
1
23
4
567
8
9
10
Parkfield EQ 2004 - stations 1-10