spatial organization of foreshocks as a tool for forecasting large earthquakes

16
AGU fall meeting, December 5-9, 2011, San Francisco INGV Spatial organization of foreshocks as a tool for forecasting large earthquakes E. Lippiello 1 , W. Marzocchi 2 , L. De Arcangelis 3 , C. Godano 1 1- Istituto Nazionale di Geofisica e Vulcanologia, Rome, Italy 2- Istituto Nazionale di Geofisica e Vulcanologia, Bologna, Italy 3- Dipartimento Scienze Geologiche, Università "Roma TRE", Rome, Italy The research was developed partially within the Strategies and tools for Real-Time Earthquake Risk Reduction (REAKT; http://www.reaktproject.eu). REAKT is funded by the European Community via the Seventh Framework Program for Research (FP7),with contract no.282862 Lippiello et al, 2012; Scientific Reports, 2, 846

Upload: linh

Post on 23-Feb-2016

29 views

Category:

Documents


0 download

DESCRIPTION

Lippiello et al, 2012; Scientific Reports , 2 , 846 . Spatial organization of foreshocks as a tool for forecasting large earthquakes. E. Lippiello 1 , W. Marzocchi 2 , L. De Arcangelis 3 , C. Godano 1 1- Istituto Nazionale di Geofisica e Vulcanologia , Rome, Italy - PowerPoint PPT Presentation

TRANSCRIPT

Page 1: Spatial organization of foreshocks as a tool for forecasting large earthquakes

AGU fall meeting, December 5-9, 2011, San FranciscoINGV

Spatial organization of foreshocks as a tool for

forecasting large earthquakes

E. Lippiello1, W. Marzocchi2, L. De Arcangelis3, C. Godano1

1- Istituto Nazionale di Geofisica e Vulcanologia, Rome, Italy2- Istituto Nazionale di Geofisica e Vulcanologia, Bologna, Italy3- Dipartimento Scienze Geologiche, Università "Roma TRE", Rome, Italy

The research was developed partially within the Strategies and tools for Real-Time Earthquake Risk Reduction (REAKT; http://www.reaktproject.eu). REAKT is funded by the European Community via the Seventh Framework Program for Research (FP7),with contract no.282862

Lippiello et al, 2012; Scientific Reports, 2, 846

Page 2: Spatial organization of foreshocks as a tool for forecasting large earthquakes

AGU fall meeting, December 5-9, 2011, San FranciscoINGV

Aim of the talk

The aim is to investigate if foreshocks of large earthquakes have distinctive features with respect to regular seismic sequences.

If so, could we profitably use these features to improve our forecasting capabilities?

The two parts of the talk

1. Analysis of the foreshocks and aftershocks sequences of “small” mainshocks in Southern California (SOCAL).

2. Application of the results of point 1 to forecast M>6 earthquakes in SOCAL

Page 3: Spatial organization of foreshocks as a tool for forecasting large earthquakes

AGU fall meeting, December 5-9, 2011, San FranciscoINGV

1. Foreshocks & aftershocks

The catalog: we use the Shearer et al (2005) seismic catalog for Southern California (1984-2002) . Very accurate epicenter location (<0.1Km) and low completeness magnitude (mc=2)

The mainshocks: we use the Felzer & Brodsky (2006) selection criterion. We analyze three mainshock classesM2 maisnhocks with 2<M<3 (black color)M3 maisnhocks with 2<M<3 (red color)M4 maisnhocks with 2<M<3 (green color)

Page 4: Spatial organization of foreshocks as a tool for forecasting large earthquakes

AGU fall meeting, December 5-9, 2011, San FranciscoINGV

1. Foreshocks & aftershocks

Is the method to identify mainshocks reliable?

A perfect model would require that some earthquakes are “really” mainshocks. If not, all mainshock selections will be necessarily model-dependent.

Two questions:

1. Are the mainshocks identified compatible with what we generally consider typical features of a mainshock? (e.g., the time-distribution of foreshocks and aftershocks)

2. Are the results found “real”? Or do they depend on the mainshock selection model?

Page 5: Spatial organization of foreshocks as a tool for forecasting large earthquakes

AGU fall meeting, December 5-9, 2011, San FranciscoINGV

1. Foreshocks & aftershocks

Typical temporal features of aftershocks and foreshocks

Page 6: Spatial organization of foreshocks as a tool for forecasting large earthquakes

AGU fall meeting, December 5-9, 2011, San FranciscoINGV

1. Foreshocks & aftershocks

LINEAR DENSITY DISTRIBUTION in the real Catalog

Page 7: Spatial organization of foreshocks as a tool for forecasting large earthquakes

AGU fall meeting, December 5-9, 2011, San FranciscoINGV

1. Foreshocks & aftershocks

LINEAR DENSITY DISTRIBUTION in the real Catalog

Page 8: Spatial organization of foreshocks as a tool for forecasting large earthquakes

AGU fall meeting, December 5-9, 2011, San FranciscoINGV

1. Foreshocks & aftershocks

LINEAR DENSITY DISTRIBUTION in the real Catalog

Page 9: Spatial organization of foreshocks as a tool for forecasting large earthquakes

AGU fall meeting, December 5-9, 2011, San FranciscoINGV

1. Foreshocks & aftershocks

LINEAR DENSITY DISTRIBUTION in ETAS simulated catalogs

Page 10: Spatial organization of foreshocks as a tool for forecasting large earthquakes

AGU fall meeting, December 5-9, 2011, San FranciscoINGV

1. Foreshocks & aftershocks

The mainshock magnititude is encoded in the foreshocks' spatial organization

Page 11: Spatial organization of foreshocks as a tool for forecasting large earthquakes

AGU fall meeting, December 5-9, 2011, San FranciscoINGV

2. Forecasting M>6 earthquakes

Daily probability for M6+ earthquakes is given by the combination of ETAS probabilities (Zhuang et al., 2004, 2005, 2008) and a factor who takes into account the spatial organization of foreshocks

We use the results of point 1 to forecast the six M>6 earthquakes in the SOCAL seismic catalog

Page 12: Spatial organization of foreshocks as a tool for forecasting large earthquakes

AGU fall meeting, December 5-9, 2011, San FranciscoINGV

2. Forecasting M>6 earthquakes

Comparison of the forecasting performances versus RI model (Rundle et al., 2002)

Model Equivalent

Confidence

level=99%

Average gain = 50.7

Page 13: Spatial organization of foreshocks as a tool for forecasting large earthquakes

AGU fall meeting, December 5-9, 2011, San FranciscoINGV

2. Forecasting M>6 earthquakes

Comparison of the forecasting performances versus ETAS (Zhuang et al., 2004, 2005, 2008)

Average gain = 4.5

Page 14: Spatial organization of foreshocks as a tool for forecasting large earthquakes

AGU fall meeting, December 5-9, 2011, San FranciscoINGV

2. Forecasting M>6 earthquakes

DAILY OCCURRENCE PROBABILITY of M>6 earthquakes within a cell 0.04ox0.04o

Page 15: Spatial organization of foreshocks as a tool for forecasting large earthquakes

AGU fall meeting, December 5-9, 2011, San FranciscoINGV

2. Forecasting M>6 earthquakes

DAILY OCCURRENCE PROBABILITY of M>6 earthquakes within a cell 0.04ox0.04o

Page 16: Spatial organization of foreshocks as a tool for forecasting large earthquakes

AGU fall meeting, December 5-9, 2011, San FranciscoINGV

Final remarks

The organization in space of seismicity before a mainshock contains information about the magnitude of the mainshock itself.

The capability to forecast M6+ earthquakes is significantly improved with respect to a pure ETAS model when the spatial organization of foreshocks is included.

The future...

Verifying the forecasting performances in Japan and Italy (keeping the same rules)

Implementing the code into a formal CSEP testing laboratory