Real-time, probabilistic and evolutionary earthquake location for seismic early warning by Claudio Satriano (I), Anthony Lomax (III) and Aldo Zollo (II) (I) RISSC-Lab, AMRA scarl, Italy (II) RISSC-Lab, Università di Napoli, Italy (III) Anthony Lomax Scientific Software, France
Introduction •
A regional early warning system is based on the ability of a seismic network to provide estimates of the location and size of a potentially destructive earthquake within a few seconds after the event is first detected.
•
This information is then used to characterize the earthquake, and to estimate its severity for the selected target in order to take mitigating actions.
C. Satriano - 2007/07/11
ElarmS • Earthquakes are located using the arrival times of
P-waves. • When the first station triggers an event is located at that station with a depth typical of events in the region. • The earthquake is then located between the first two, and then the first three stations to trigger. • Once four stations have triggered a grid search method is used to locate the event, minimizing the misfit between predicted and observed arrival times.
from Allen, 2005 (www.elarms.org)
C. Satriano - 2007/07/11
Real-time location at NIED, Japan
Tnow − Ti (x, t) < 0
An Automatic Processing System for Broadcasting Earthquake Alarms
from Horiuchi et al., 2005
Figure 2.
Schematic map showing the method of the grid search in the hypocenter location. A threeC. Satriano dimensional grid with 3 ! 3 ! 3 " 27 points (a
-
window order to having s The neously the case or when bances. about 75 entering pocente P-wave produce of Hase timing o roneous is ineffe 2007/07/11 It is ess
Real-Time Earthquake Location (RTLoc) Our methodology is related to that of Horiuchi et al. [2005], which we extend and generalize by:
C. Satriano - 2007/07/11
Real-Time Earthquake Location (RTLoc) Our methodology is related to that of Horiuchi et al. [2005], which we extend and generalize by: a. starting the location procedure after only one station has triggered
C. Satriano - 2007/07/11
Real-Time Earthquake Location (RTLoc) Our methodology is related to that of Horiuchi et al. [2005], which we extend and generalize by: a. starting the location procedure after only one station has triggered b. using the Equal Differential Time (EDT) approach throughout to incorporate the triggered arrivals and the not-yet-triggered stations
C. Satriano - 2007/07/11
Real-Time Earthquake Location (RTLoc) Our methodology is related to that of Horiuchi et al. [2005], which we extend and generalize by: a. starting the location procedure after only one station has triggered b. using the Equal Differential Time (EDT) approach throughout to incorporate the triggered arrivals and the not-yet-triggered stations c. estimating the hypocenter probabilistically as a pdf instead of as a point
C. Satriano - 2007/07/11
Real-Time Earthquake Location (RTLoc) Our methodology is related to that of Horiuchi et al. [2005], which we extend and generalize by: a. starting the location procedure after only one station has triggered b. using the Equal Differential Time (EDT) approach throughout to incorporate the triggered arrivals and the not-yet-triggered stations c. estimating the hypocenter probabilistically as a pdf instead of as a point d. applying a full, non-linearized, global-search for each update of the location estimate.
C. Satriano - 2007/07/11
Equal Differential Time (EDT) approach If the hypocenter is exactly determined, the difference between calculated travel times tt0 and tt1 from the hypocenter to two stations S0 and S1 is equal to the difference between the observed arrival times t0 and t1 at the two stations, since the observed arrival times share the common earthquake origin time:
tt1 − tt0 = t1 − t0 from Font et al., 2004
C. Satriano - 2007/07/11
Evolutionary Earthquake Location - 1/7 stations
Voronoi cell boundaries
C. Satriano - 2007/07/11
Evolutionary Earthquake Location - 2/7 ttB − ttA = 0
B
A
C. Satriano - 2007/07/11
Evolutionary Earthquake Location - 3/7 ttB − ttA ≥ 0
“conditional” EDT surface volume defined by stations without arrivals
hypocenter B
A wavefront
First station detects arrival constraint is Voronoi cells
C. Satriano - 2007/07/11
Evolutionary Earthquake Location - 4/7 ttB − ttA ≥ tnow − tA “conditional” EDT surface
B
A
Wavefront expands EDT surfaces deform, constraint improves
C. Satriano - 2007/07/11
Evolutionary Earthquake Location - 5/7
ttB − ttA = tB − tA “true” EDT surface
B
A
Second station detects arrival constraint includes EDT surface
C. Satriano - 2007/07/11
Evolutionary Earthquake Location - 6/7
Third station detects arrival constraint is mainly EDT surfaces
C. Satriano - 2007/07/11
Evolutionary Earthquake Location - 7/7
Fourth station detects arrival location is well constrained
C. Satriano - 2007/07/11
RTLoc: Algorithm (1/2) We consider N operational stations (S0, …, SN), a gridded search volume V containing the network and target earthquake source regions, and pre-computed travel times from each station to each grid point (i,j,k) in V computed for a given velocity model. When the first station Sn triggers, we compute for each grid point the quantity:
(pn )i,j,k =
!
1 0
if if
(ttSk − ttSn )i,j,k ≥ δtn,k ; k #= n (ttSk − ttSn )i,j,k < δtn,k
δtn,k = tnow − tSn
ttB − ttA ≥ tnow − tA
B A
C. Satriano - 2007/07/11
RTLoc: Algorithm (1/2) We consider N operational stations (S0, …, SN), a gridded search volume V containing the network and target earthquake source regions, and pre-computed travel times from each station to each grid point (i,j,k) in V computed for a given velocity model. When the first station Sn triggers, we compute for each grid point the quantity:
(pn )i,j,k =
!
1 0
if if
(ttSk − ttSn )i,j,k ≥ δtn,k ; k #= n (ttSk − ttSn )i,j,k < δtn,k
δtn,k = tnow − tSn
ttB − ttA ≥ tnow − tA
B A
When a new station triggers, we re-evaluate (pm)i,j,k for all the pairs Sn (triggered) - Sk (not-yet-triggered). C. Satriano - 2007/07/11
RTLoc: Algorithm (2/2) For each pair Sn, Sm of triggered stations, we compute at each grid point (i, j, k), the quantity:
(qm )i,j,k
!
[(ttSm − ttSn ) − (tSm − tSn )]2i,j,k = exp − 2σ 2
"
; m "= n
ttB − ttA = tB − tA B A
C. Satriano - 2007/07/11
RTLoc: Algorithm (2/2) For each pair Sn, Sm of triggered stations, we compute at each grid point (i, j, k), the quantity:
(qm )i,j,k
!
[(ttSm − ttSn ) − (tSm − tSn )]2i,j,k = exp − 2σ 2
"
; m "= n
ttB − ttA = tB − tA
Eventually, for each grid point we define:
Pi,j,k
1 = M
!
"
(pn )i,j,k +
n
Qi,j,k = (Pi,j,k )N
" m
(qm )i,j,k
B
#
A
where M is the number of equations. The quantity Qi,j,k forms a relative probability density function (with values between 0 and 1) for the hypocenter location within the grid cell (i, j, k). C. Satriano - 2007/07/11
RTLoc test at the ISNet
The Irpinia Seismic Network (ISNet) C. Satriano - 2007/07/11
RTLoc test at the ISNet
The Irpinia Seismic Network (ISNet) C. Satriano - 2007/07/11
RTLoc test at the ISNet X(km) −40
−20
Z(km)
0
20
40
0
10
20
30
40
40
40 RSA3
SSB3
BSC3
LPI3
CLT3
Y(km)
LIO3 MNT3 NSC3
20
AND3 RDM3
TEO3
CSG3 SFL3
0 VDS3
SNR3
0
AVG3
BEL3 SCL3
COL3 CMP3
Y(km)
RSF3
20
PCR3 VDP3
−20
−20 PGN3
PST3
CGG3 STN3 SRN3
−40
−40
0
10
20
30
40
Z(km)
0
Z(km)
10 20
Δt = 0.00 s
30
tnow = 3.03 s
40 −40
−20
0
20
40
X(km) 0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
Location Probability
0.8
0.9
1.0
C. Satriano - 2007/07/11
RTLoc test at the ISNet X(km) −40
−20
Z(km)
0
20
40
0
10
20
30
40
40
40 RSA3
SSB3
BSC3
LPI3
CLT3
Y(km)
LIO3 MNT3 NSC3
20
AND3 RDM3
TEO3
CSG3 SFL3
0 VDS3
SNR3
0
AVG3
BEL3 SCL3
COL3 CMP3
Y(km)
RSF3
20
PCR3 VDP3
−20
−20 PGN3
PST3
CGG3 STN3 SRN3
−40
−40
0
10
20
30
40
Z(km)
0
Z(km)
10 20
Δt = 0.20 s
30
tnow = 3.23 s
40 −40
−20
0
20
40
X(km) 0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
Location Probability
0.8
0.9
1.0
C. Satriano - 2007/07/11
RTLoc test at the ISNet X(km) −40
−20
Z(km)
0
20
40
0
10
20
30
40
40
40 RSA3
SSB3
BSC3
LPI3
CLT3
Y(km)
LIO3 MNT3 NSC3
20
AND3 RDM3
TEO3
CSG3 SFL3
0 VDS3
SNR3
0
AVG3
BEL3 SCL3
COL3 CMP3
Y(km)
RSF3
20
PCR3 VDP3
−20
−20 PGN3
PST3
CGG3 STN3 SRN3
−40
−40
0
10
20
30
40
Z(km)
0
Z(km)
10 20
Δt = 0.40 s
30
tnow = 3.43 s
40 −40
−20
0
20
40
X(km) 0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
Location Probability
0.8
0.9
1.0
C. Satriano - 2007/07/11
RTLoc test at the ISNet X(km) −40
−20
Z(km)
0
20
40
0
10
20
30
40
40
40 RSA3
SSB3
BSC3
LPI3
CLT3
Y(km)
LIO3 MNT3 NSC3
20
AND3 RDM3
TEO3
CSG3 SFL3
0 VDS3
SNR3
0
AVG3
BEL3 SCL3
COL3 CMP3
Y(km)
RSF3
20
PCR3 VDP3
−20
−20 PGN3
PST3
CGG3 STN3 SRN3
−40
−40
0
10
20
30
40
Z(km)
0
Z(km)
10 20
Δt = 0.60 s
30
tnow = 3.63 s
40 −40
−20
0
20
40
X(km) 0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
Location Probability
0.8
0.9
1.0
C. Satriano - 2007/07/11
RTLoc test at the ISNet X(km) −40
−20
Z(km)
0
20
40
0
10
20
30
40
40
40 RSA3
SSB3
BSC3
LPI3
CLT3
Y(km)
LIO3 MNT3 NSC3
20
AND3 RDM3
TEO3
CSG3 SFL3
0 VDS3
SNR3
0
AVG3
BEL3 SCL3
COL3 CMP3
Y(km)
RSF3
20
PCR3 VDP3
−20
−20 PGN3
PST3
CGG3 STN3 SRN3
−40
−40
0
10
20
30
40
Z(km)
0
Z(km)
10 20
Δt = 0.80 s
30
tnow = 3.83 s
40 −40
−20
0
20
40
X(km) 0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
Location Probability
0.8
0.9
1.0
C. Satriano - 2007/07/11
RTLoc test at the ISNet X(km) −40
−20
Z(km)
0
20
40
0
10
20
30
40
40
40 RSA3
SSB3
BSC3
LPI3
CLT3
Y(km)
LIO3 MNT3 NSC3
20
AND3 RDM3
TEO3
CSG3 SFL3
0 VDS3
SNR3
0
AVG3
BEL3 SCL3
COL3 CMP3
Y(km)
RSF3
20
PCR3 VDP3
−20
−20 PGN3
PST3
CGG3 STN3 SRN3
−40
−40
0
10
20
30
40
Z(km)
0
Z(km)
10 20
Δt = 1.00 s
30
tnow = 4.03 s
40 −40
−20
0
20
40
X(km) 0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
Location Probability
0.8
0.9
1.0
C. Satriano - 2007/07/11
RTLoc test at the ISNet X(km) −40
−20
Z(km)
0
20
40
0
10
20
30
40
40
40 RSA3
SSB3
BSC3
LPI3
CLT3
Y(km)
LIO3 MNT3 NSC3
20
AND3 RDM3
TEO3
CSG3 SFL3
0 VDS3
SNR3
0
AVG3
BEL3 SCL3
COL3 CMP3
Y(km)
RSF3
20
PCR3 VDP3
−20
−20 PGN3
PST3
CGG3 STN3 SRN3
−40
−40
0
10
20
30
40
Z(km)
0
Z(km)
10 20
Δt = 0.00 s
30
tnow = 5.40 s
40 −40
−20
0
20
40
X(km) 0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
Location Probability
0.8
0.9
1.0
C. Satriano - 2007/07/11
RTLoc test at the ISNet X(km) −40
−20
Z(km)
0
20
40
0
10
20
30
40
40
40 RSA3
SSB3
BSC3
LPI3
CLT3
Y(km)
LIO3 MNT3 NSC3
20
AND3 RDM3
TEO3
CSG3 SFL3
0 VDS3
SNR3
0
AVG3
BEL3 SCL3
COL3 CMP3
Y(km)
RSF3
20
PCR3 VDP3
−20
−20 PGN3
PST3
CGG3 STN3 SRN3
−40
−40
0
10
20
30
40
Z(km)
0
Z(km)
10 20
Δt = 0.20 s
30
tnow = 5.60 s
40 −40
−20
0
20
40
X(km) 0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
Location Probability
0.8
0.9
1.0
C. Satriano - 2007/07/11
RTLoc test at the ISNet X(km) −40
−20
Z(km)
0
20
40
0
10
20
30
40
40
40 RSA3
SSB3
BSC3
LPI3
CLT3
Y(km)
LIO3 MNT3 NSC3
20
AND3 RDM3
TEO3
CSG3 SFL3
0 VDS3
SNR3
0
AVG3
BEL3 SCL3
COL3 CMP3
Y(km)
RSF3
20
PCR3 VDP3
−20
−20 PGN3
PST3
CGG3 STN3 SRN3
−40
−40
0
10
20
30
40
Z(km)
0
Z(km)
10 20
Δt = 0.40 s
30
tnow = 5.80 s
40 −40
−20
0
20
40
X(km) 0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
Location Probability
0.8
0.9
1.0
C. Satriano - 2007/07/11
RTLoc test at the ISNet X(km) −40
−20
Z(km)
0
20
40
0
10
20
30
40
40
40 RSA3
SSB3
BSC3
LPI3
CLT3
Y(km)
LIO3 MNT3 NSC3
20
AND3 RDM3
TEO3
CSG3 SFL3
0 VDS3
SNR3
0
AVG3
BEL3 SCL3
COL3 CMP3
Y(km)
RSF3
20
PCR3 VDP3
−20
−20 PGN3
PST3
CGG3 STN3 SRN3
−40
−40
0
10
20
30
40
Z(km)
0
Z(km)
10 20
Δt = 0.80 s
30
tnow = 6.20 s
40 −40
−20
0
20
40
X(km) 0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
Location Probability
0.8
0.9
1.0
C. Satriano - 2007/07/11
RTLoc test at the ISNet X(km) −40
−20
Z(km)
0
20
40
0
10
20
30
40
40
40 RSA3
SSB3
BSC3
LPI3
CLT3
Y(km)
LIO3 MNT3 NSC3
20
AND3 RDM3
TEO3
CSG3 SFL3
0 VDS3
SNR3
0
AVG3
BEL3 SCL3
COL3 CMP3
Y(km)
RSF3
20
PCR3 VDP3
−20
−20 PGN3
PST3
CGG3 STN3 SRN3
−40
−40
0
10
20
30
40
Z(km)
0
Z(km)
10 20
Δt = 1.00 s
30
tnow = 6.40 s
40 −40
−20
0
20
40
X(km) 0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
Location Probability
0.8
0.9
1.0
C. Satriano - 2007/07/11
RTLoc test at the ISNet X(km) −40
−20
Z(km)
0
20
40
0
10
20
30
40
40
40 RSA3
SSB3
BSC3
LPI3
CLT3
Y(km)
LIO3 MNT3 NSC3
20
AND3 RDM3
TEO3
CSG3 SFL3
0 VDS3
SNR3
0
AVG3
BEL3 SCL3
COL3 CMP3
Y(km)
RSF3
20
PCR3 VDP3
−20
−20 PGN3
PST3
CGG3 STN3 SRN3
−40
−40
0
10
20
30
40
Z(km)
0
Z(km)
10 20
Δt = 1.20 s
30
tnow = 6.60 s
40 −40
−20
0
20
40
X(km) 0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
Location Probability
0.8
0.9
1.0
C. Satriano - 2007/07/11
RTLoc test at the ISNet X(km) −40
−20
Z(km)
0
20
40
0
10
20
30
40
40
40 RSA3
SSB3
BSC3
LPI3
CLT3
Y(km)
LIO3 MNT3 NSC3
20
AND3 RDM3
TEO3
CSG3 SFL3
0 VDS3
SNR3
0
AVG3
BEL3 SCL3
COL3 CMP3
Y(km)
RSF3
20
PCR3 VDP3
−20
−20 PGN3
PST3
CGG3 STN3 SRN3
−40
−40
0
10
20
30
40
Z(km)
0
Z(km)
10 20
Δt = 1.40 s
30
tnow = 6.80 s
40 −40
−20
0
20
40
X(km) 0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
Location Probability
0.8
0.9
1.0
C. Satriano - 2007/07/11
RTLoc test at the ISNet X(km) −40
−20
Z(km)
0
20
40
0
10
20
30
40
40
40 RSA3
SSB3
BSC3
LPI3
CLT3
Y(km)
LIO3 MNT3 NSC3
20
AND3 RDM3
TEO3
CSG3 SFL3
0 VDS3
SNR3
0
AVG3
BEL3 SCL3
COL3 CMP3
Y(km)
RSF3
20
PCR3 VDP3
−20
−20 PGN3
PST3
CGG3 STN3 SRN3
−40
−40
0
10
20
30
40
Z(km)
0
Z(km)
10 20
Δt = 1.60 s
30
tnow = 7.00 s
40 −40
−20
0
20
40
X(km) 0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
Location Probability
0.8
0.9
1.0
C. Satriano - 2007/07/11
RTLoc test at the ISNet X(km) −40
−20
Z(km)
0
20
40
0
10
20
30
40
40
40 RSA3
SSB3
BSC3
LPI3
CLT3
Y(km)
LIO3 MNT3 NSC3
20
AND3 RDM3
TEO3
CSG3 SFL3
0 VDS3
SNR3
0
AVG3
BEL3 SCL3
COL3 CMP3
Y(km)
RSF3
20
PCR3 VDP3
−20
−20 PGN3
PST3
CGG3 STN3 SRN3
−40
−40
0
10
20
30
40
Z(km)
0
Z(km)
10 20
Δt = 1.80 s
30
tnow = 7.20 s
40 −40
−20
0
20
40
X(km) 0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
Location Probability
0.8
0.9
1.0
C. Satriano - 2007/07/11
RTLoc test at the ISNet X(km) −40
−20
Z(km)
0
20
40
0
10
20
30
40
40
40 RSA3
SSB3
BSC3
LPI3
CLT3
Y(km)
LIO3 MNT3 NSC3
20
AND3 RDM3
TEO3
CSG3 SFL3
0 VDS3
SNR3
0
AVG3
BEL3 SCL3
COL3 CMP3
Y(km)
RSF3
20
PCR3 VDP3
−20
−20 PGN3
PST3
CGG3 STN3 SRN3
−40
−40
0
10
20
30
40
Z(km)
0
Z(km)
10 20
Δt = 2.00 s
30
tnow = 7.40 s
40 −40
−20
0
20
40
X(km) 0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
Location Probability
0.8
0.9
1.0
C. Satriano - 2007/07/11
RTLoc test at the ISNet X(km) −40
−20
Z(km)
0
20
40
0
10
20
30
40
40
40 RSA3
SSB3
BSC3
LPI3
CLT3
Y(km)
LIO3 MNT3 NSC3
20
AND3 RDM3
TEO3
CSG3 SFL3
0 VDS3
SNR3
0
AVG3
BEL3 SCL3
COL3 CMP3
Y(km)
RSF3
20
PCR3 VDP3
−20
−20 PGN3
PST3
CGG3 STN3 SRN3
−40
−40
0
10
20
30
40
Z(km)
0
Z(km)
10 20
Δt = 2.20 s
30
tnow = 7.60 s
40 −40
−20
0
20
40
X(km) 0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
Location Probability
0.8
0.9
1.0
C. Satriano - 2007/07/11
RTLoc test at the ISNet X(km) −40
−20
Z(km)
0
20
40
0
10
20
30
40
40
40 RSA3
SSB3
BSC3
LPI3
CLT3
Y(km)
LIO3 MNT3 NSC3
20
AND3 RDM3
TEO3
CSG3 SFL3
0 VDS3
SNR3
0
AVG3
BEL3 SCL3
COL3 CMP3
Y(km)
RSF3
20
PCR3 VDP3
−20
−20 PGN3
PST3
CGG3 STN3 SRN3
−40
−40
0
10
20
30
40
Z(km)
0
Z(km)
10 20
Δt = 2.40 s
30
tnow = 7.80 s
40 −40
−20
0
20
40
X(km) 0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
Location Probability
0.8
0.9
1.0
C. Satriano - 2007/07/11
RTLoc test at the ISNet X(km) −40
−20
Z(km)
0
20
40
0
10
20
30
40
40
40 RSA3
SSB3
BSC3
LPI3
CLT3
Y(km)
LIO3 MNT3 NSC3
20
AND3 RDM3
TEO3
CSG3 SFL3
0 VDS3
SNR3
0
AVG3
BEL3 SCL3
COL3 CMP3
Y(km)
RSF3
20
PCR3 VDP3
−20
−20 PGN3
PST3
CGG3 STN3 SRN3
−40
−40
0
10
20
30
40
Z(km)
0
Z(km)
10 20
Δt = 2.60 s
30
tnow = 8.00 s
40 −40
−20
0
20
40
X(km) 0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
Location Probability
0.8
0.9
1.0
C. Satriano - 2007/07/11
RTLoc test at the ISNet X(km) −40
−20
Z(km)
0
20
40
0
10
20
30
40
40
40 RSA3
SSB3
BSC3
LPI3
CLT3
Y(km)
LIO3 MNT3 NSC3
20
AND3 RDM3
TEO3
CSG3 SFL3
0 VDS3
SNR3
0
AVG3
BEL3 SCL3
COL3 CMP3
Y(km)
RSF3
20
PCR3 VDP3
−20
−20 PGN3
PST3
CGG3 STN3 SRN3
−40
−40
0
10
20
30
40
Z(km)
0
Z(km)
10 20
Δt = 3.00 s
30
tnow = 8.40 s
40 −40
−20
0
20
40
X(km) 0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
Location Probability
0.8
0.9
1.0
C. Satriano - 2007/07/11
Magnitude 6.8 - SOUTHERN GREECE, 2006 January 8 11:34:55 UTC
Z(km)
X(km) −300
−200
−100
0
100
200
0
20
40
60
80 100
100
100
Athens
0
ITM
−100
APE
−100
Y(km)
Y(km)
0
VLI SANT −200
−200
VAM
−300
IDI
−300
LAST 0
20
40
60
80 100
Z(km) 0
Z(km)
20 40 60
Δt = 0.00 s
80 100 −300
−200
−100
0
100
200
tnow = 14.90 s
X(km)
tSAthens = −40.00 s 0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
Location Probability
RTLoc C. Satriano - 2007/07/11
Magnitude 6.8 - SOUTHERN GREECE, 2006 January 8 11:34:55 UTC
Z(km)
X(km) −300
−200
−100
0
100
200
0
20
40
60
80 100
100
100
Athens
0
ITM
−100
APE
−100
Y(km)
Y(km)
0
VLI SANT −200
−200
VAM
−300
IDI
−300
LAST 0
20
40
60
80 100
Z(km) 0
Z(km)
20 40 60
Δt = 4.00 s
80 100 −300
−200
−100
0
100
200
tnow = 18.90 s
X(km)
tSAthens = −36.00 s 0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
Location Probability
RTLoc C. Satriano - 2007/07/11
Magnitude 6.8 - SOUTHERN GREECE, 2006 January 8 11:34:55 UTC
Z(km) 0
100
200
0
20
40
60
100
Athens
Y(km)
0
ITM
−100
Z(km)
X(km) 80 100
APE
−300
100
100
0
0
−100
−200
−100
0
100
200
VAM
−300
0
ITM
−100
APE
−100
IDI
−200
−200
−300
−300
−200
VAM
LAST 20
40
60
IDI
−300
LAST
80 100
0
0
0
20
20
40 60
Δt = 6.00 s
80
−100
0
100
200
0.4
0.5
0.6
0.7
0.8
Location Probability
NLLoc
0.9
1.0
60
80 100
60
Δt = 6.00 s
80 100 −300
−200
−100
0
100
200
tnow = 20.90 s
X(km)
tSAthens = −34.00 s 0.3
40
40
tnow = 20.90 s
X(km)
20
Z(km)
Z(km)
Z(km)
Z(km)
0.2
80 100
SANT
0
0.1
60
Athens
SANT
0.0
40
VLI
−200
−200
20
100
VLI
100 −300
0
Y(km)
−100
Y(km)
−200
Y(km)
X(km) −300
tSAthens = −34.00 s 0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
Location Probability
RTLoc
Lomax, et al., 2000 C. Satriano - 2007/07/11
Magnitude 6.8 - SOUTHERN GREECE, 2006 January 8 11:34:55 UTC
Z(km) 0
100
200
0
20
40
60
100
Athens
Y(km)
0
ITM
−100
Z(km)
X(km) 80 100
APE
−300
100
100
0
0
−100
−200
−100
0
100
200
VAM
−300
0
ITM
−100
APE
−100
IDI
−200
−200
−300
−300
−200
VAM
LAST 20
40
60
IDI
−300
LAST
80 100
0
0
0
20
20
40 60
Δt = 12.00 s
80
−100
0
100
200
0.4
0.5
0.6
0.7
Location Probability
NLLoc
0.8
0.9
1.0
60
80 100
60
Δt = 12.00 s
80 100 −300
−200
−100
0
100
200
tnow = 26.90 s
X(km)
tSAthens = −28.00 s 0.3
40
40
tnow = 26.90 s
X(km)
20
Z(km)
Z(km)
Z(km)
Z(km)
0.2
80 100
SANT
0
0.1
60
Athens
SANT
0.0
40
VLI
−200
−200
20
100
VLI
100 −300
0
Y(km)
−100
Y(km)
−200
Y(km)
X(km) −300
tSAthens = −28.00 s 0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
Location Probability
RTLoc C. Satriano - 2007/07/11
Magnitude 6.8 - SOUTHERN GREECE, 2006 January 8 11:34:55 UTC
Z(km) 0
100
200
0
20
40
60
100
Athens
Y(km)
0
ITM
−100
Z(km)
X(km) 80 100
APE
−300
100
100
0
0
−100
−200
−100
0
100
200
VAM
−300
0
ITM
−100
APE
−100
IDI
−200
−200
−300
−300
−200
VAM
LAST 20
40
60
IDI
−300
LAST
80 100
0
0
0
20
20
40 60
Δt = 14.00 s
80
−100
0
100
200
0.4
0.5
0.6
0.7
Location Probability
NLLoc
0.8
0.9
1.0
60
80 100
60
Δt = 14.00 s
80 100 −300
−200
−100
0
100
200
tnow = 28.90 s
X(km)
tSAthens = −26.00 s 0.3
40
40
tnow = 28.90 s
X(km)
20
Z(km)
Z(km)
Z(km)
Z(km)
0.2
80 100
SANT
0
0.1
60
Athens
SANT
0.0
40
VLI
−200
−200
20
100
VLI
100 −300
0
Y(km)
−100
Y(km)
−200
Y(km)
X(km) −300
tSAthens = −26.00 s 0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
Location Probability
RTLoc C. Satriano - 2007/07/11
Magnitude 6.8 - SOUTHERN GREECE, 2006 January 8 11:34:55 UTC
Z(km) 0
100
200
0
20
40
60
100
Athens
Y(km)
0
ITM
−100
Z(km)
X(km) 80 100
APE
−300
100
100
0
0
−100
−200
−100
0
100
200
VAM
−300
0
ITM
−100
APE
−100
IDI
−200
−200
−300
−300
−200
VAM
LAST 20
40
60
IDI
−300
LAST
80 100
0
0
0
20
20
40 60
Δt = 16.00 s
80
−100
0
100
200
0.4
0.5
0.6
0.7
Location Probability
NLLoc
0.8
0.9
1.0
60
80 100
60
Δt = 16.00 s
80 100 −300
−200
−100
0
100
200
tnow = 30.90 s
X(km)
tSAthens = −24.00 s 0.3
40
40
tnow = 30.90 s
X(km)
20
Z(km)
Z(km)
Z(km)
Z(km)
0.2
80 100
SANT
0
0.1
60
Athens
SANT
0.0
40
VLI
−200
−200
20
100
VLI
100 −300
0
Y(km)
−100
Y(km)
−200
Y(km)
X(km) −300
tSAthens = −24.00 s 0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
Location Probability
RTLoc C. Satriano - 2007/07/11
New Pick
Event association (binding) NO
RM SSi
Is there an event?
M ! 1 2 = [(ttSi − ttSj ) − (tSi − tSj )] M j=1
YES Temporarily associate the pick to the event i
Re-locate event i
NO RMS < RMSmax ?
Declare a new event j
YES Permanently associate the pick to the event i
Compute the location probability for j
Save to disk the new location map for event i
Save to disk the location map for event j
C. Satriano - 2007/07/11
Real-Time Hazard Evaluation/Mitigation
fˆP GA (pga) =
!
M
from Iervolino et al., 2006
!
fP GA|M,R (pga|m, r)fˆM (m)fˆR (r)dmdr
R
C. Satriano - 2007/07/11
Real-Time Hazard Evaluation/Mitigation
fˆP GA (pga) =
!
M
from Iervolino et al., 2006
!
fP GA|M,R (pga|m, r)fˆM (m)fˆR (r)dmdr
R
C. Satriano - 2007/07/11
Real-Time Hazard Evaluation/Mitigation
fˆP GA (pga) =
!
M
!
R
fP GA|M,R (pga|m, r)fˆM (m)fˆR (r)dmdr EEW (function of time)
from Iervolino et al., 2006
C. Satriano - 2007/07/11
Real-Time Hazard Evaluation/Mitigation
fˆP GA (pga) =
!
M
from Iervolino et al., 2006
!
R
fP GA|M,R (pga|m, r)fˆM (m)fˆR (r)dmdr Attenuation Law
EEW (function of time)
C. Satriano - 2007/07/11
Real-Time Hazard Evaluation/Mitigation
fˆP GA (pga) =
!
M
from Iervolino et al., 2006
!
R
fP GA|M,R (pga|m, r)fˆM (m)fˆR (r)dmdr Attenuation Law
EEW (function of time)
C. Satriano - 2007/07/11
Real-Time Hazard Evaluation/Mitigation
fˆP GA (pga) =
!
M
!
R
fP GA|M,R (pga|m, r)fˆM (m)fˆR (r)dmdr Attenuation Law
EEW (function of time)
Estimation of Engineering Demand Parameters (EDP):
fˆEDP (edp) = from Iervolino et al., 2006
!
IM
fEDP |P GA (edp|pga)fˆP GA (pga)dpga C. Satriano - 2007/07/11
Discussion •
We have developed a probabilistic, real-time evolutionary location technique (RTLoc) based on the equal differential-time (EDT) formulation.
•
At each time step, this algorithm makes use of both the information from triggered arrivals and not-yet-triggered stations.
•
Constraint on the hypocenter location is obtained as soon as the first station has triggered and is updated at fixed time intervals or when a new station triggers.
C. Satriano - 2007/07/11
Discussion •
•
We have developed a probabilistic, real-time evolutionary location technique (RTLoc) based on the equal differential-time (EDT) formulation.
•
At each time step, this algorithm makes use of both the information from triggered arrivals and not-yet-triggered stations.
•
Constraint on the hypocenter location is obtained as soon as the first station has triggered and is updated at fixed time intervals or when a new station triggers.
The hypocenter location is estimated as a probability density function defined within a pre-defined search volume.
•
This probabilistic description of the location results is easy to incorporate into a system for real-time hazard evaluation and mitigation.
C. Satriano - 2007/07/11
Discussion •
•
We have developed a probabilistic, real-time evolutionary location technique (RTLoc) based on the equal differential-time (EDT) formulation.
•
At each time step, this algorithm makes use of both the information from triggered arrivals and not-yet-triggered stations.
•
Constraint on the hypocenter location is obtained as soon as the first station has triggered and is updated at fixed time intervals or when a new station triggers.
The hypocenter location is estimated as a probability density function defined within a pre-defined search volume.
• •
This probabilistic description of the location results is easy to incorporate into a system for real-time hazard evaluation and mitigation.
Both synthetic and real data show that useful locations can be obtained within 1-2 seconds for a local earthquake and 6-10 seconds at a regional scale C. Satriano - 2007/07/11
