WP  2.5 Using shear-wave splitting above small earthquakes to monitor stress in the SISZ 

Start date or starting event:

M0

Lead contractor:

UEDIN

Participants:

UEDIN, IMOR

1

Objectives:

1)     Continue monitoring shear-wave splitting (SWS) above small earthquakes in SISZ. 

2)     Evaluation of stress induced SWS changes in SIL-data since 1991, to be correlated with other methods for improved stress-imaging. 

3)       Identify the build-up of stress (from 1) as a base for stress-forecasts.

4)     Develop automatic analysis of shear-wave splitting by artificial neural network (ANN) techniques. 

5)     Develop training sets for ANN that preserve interpreter’s experience for individual seismic stations. 

2

Inputs:

Wave-form data and parameter data from the SIL seismic stations in Iceland.

Methodology and experience gained by P3 in stress-forecasting especially in Iceland.

 

Methodology / work description:

The previous EC Projects PRENLAB-1, PRENLAB-2 and SMSITES monitored shear-wave splitting (SWS) above small earthquakes throughout Iceland and recognized the build-up of stress before earthquakes and volcanic eruptions.  This work will be continued.  This led to a correct stress forecast of the time and magnitude of an mb=5 earthquake in SW Iceland.

 

However, the magnitude Ms=6.6 earthquakes in 2000 in the SISZ were not forecast, because of a gap in source seismicity at the near station BJA.  Detailed analysis of local stations will test whether forecast could have been made without data from BJA.  SWS estimates of stress changes will be correlated with other stress estimates from 1991 before the two Ms=6.6 earthquakes and after in order to build up a more complete image of the behaviour of stress in the SISZ and elsewhere.

 

Shear-wave splitting is subject to so many variables that satisfactory automatic measurement of shear-wave on seismograms by analytical techniques is highly unlikely.  Artificial neural network (ANN) techniques will be developed to preserve interpreter’s experience in SWS.  The skill in applying ANN techniques is developed by selection of suitable data-training sets (at each station) so appropriate experience is inserted into ANN.  Such ANN training sets will be selected for as many individual SISZ stations as possible.  Thus ANN is expected to at least partially automate measuring SWS.  

3

Deliverables including cost of deliverable as percentage of total cost of the proposed project:

D28

Plots of stress variations before earthquakes and volcanic eruptions.

M12/24  Re  PU  0,7%

D29

Stress-forecasts of impending large earthquakes issued to IMOR.

Re  CO  0,6%

D30

Report on stress changes estimates by SWS since 1996.

M12  Re  PU  0,4%

D31

Reports in collaboration with other partners of imaging stress variations.

M12/24  Re  PU  0,8%

D32

Reports on progress of ANN measurements of shear-wave splitting.

M12  Re  PU  0,4%

D33

Reports on experience of selecting training sets for ANN.

M12  Re  PU  0,4%

D34

Program for measuring SWS with ANN.

M24  Re  PU  1,5%

D35

Publication of papers in international research journals.

M24  Re  PU  1,0%

                          

4

Milestones: Delivery of the above items at the date indicated.