1. ISTANBUL TEST SITE
Kandilli Observatory and Earthquake Research Institute
Department of Earthquake Engineering Boğaziçi University, Istanbul
SAFER Final Meeting, GFZ, Potsdam
June

2. SAFER Project, Test Site Istanbul
Istanbul Earthquake Rapid Response System
Istanbul Earthquake Early Warning System
Istanbul, Atakoy District SOSEWIN (Self-Organizing Seismic Early Warning Information Network) System
Istanbul Structural Monitoring Networks
Shake and Los Mapping

3. ISTANBUL EARTHQUAKE EARLY WARNING SYSTEM
The Early Warning part of the I-NET 10+2 strong motion stations were located as close as possible to the Great Marmara Fault zone in “on-line” mode.
Data Transmission is provided with Spread Spectrum Radio Modem and Satellite.
The continuous on-line data from these stations is used to provide real time warning for emerging potentially disastrous earthquakes. 3

4. DISTRIBUTION OF EARLY WARNING STATIONS and RADIO-MODEM TRANSMISSION
Satellite Connection

5. REAL-TIME STRUCTURAL DAMAGE RELATED PARAMETERS
Many researchers have investigated the relationships between the earthquake damage and the ground motion parameters such as peak ground motion amplitudes, spectral amplitudes at selected periods.
Arias Intensity, Cumulative Absolute Velocity and Housner’s Spectrum Intensity.
Nakamura (2004) defined the so-called “Destructive Intensity, DI”, represented by the logarithm of the absolute value of the inner product of the acceleration and a velocity vectors:
DI = log (|Σ(a.v)|), MMI=(11/7)*DI 5

6. BCAV – Window Length=1s
BCAV – W (Window Length=8s)

7. Correlation between CAV and computed seismic intensity
(Computed Seismic Intensities are obtained from the FAS of simulated accelerograms using Sokolov (2002)) 7

10. Early Warning times from 280 simulated earthquakes

11. FAST TRAIN AND TUBE TUNNEL
EEW
FAST TRAIN AND TUBE TUNNEL 11

12. IGDAŞ Service Network Earthquake Monıtorıng
BOĞAZİÇİ DISTRICT DIRECTORY
ANADOLU DISTRICT DIRECTORY
İSTANBUL DISTRICT DIRECTORY 12

13. Potential Uses of EW in Istanbul
ELECTRIC POWER DISTRIBUTION SYSTEM
HEAVY INDUSTRY 13

15. ISTANBUL EARTHQUAKE RAPID RESPONSE SYSTEM STATIONS15

16. Expansion by the end of 2009 KOERI IEWRRS 100 Station +
İGDAŞ (Istanbul Gas Company) 100 Station 16

17. After triggered by an earthquake, each station processes the streaming three-channel strong motion data to yield the Spectral accelerations at specific periods, 12Hz filtered PGA and PGV and sends these parameters in the form of SMS messages at every 20s directly to the main data center through the AVEA - GSM communication system by using several base stations, microwave system and landlines.
Spectral displacements obtained from the SMS messages sent from stations are interpolated to determine the spectral displacement values at the center of each geo-cell (0.01 x 0.01).
The seismic demand at the center of each geo-cell is computed using these spectral displacements.
Using the capacities of the buildings (24 types) in each geo-cell the building damage is computed by using the spectral-displacement based fragility curves (Capacity Spectrum Procedure).

18. Capacity Spectrum-type Building Damage Assessment

19. Location of earthquakes recorded by the system
Information on recorded Earthquakes 19

20. Sept. 29, 2004 Marmara Sea Earthquake (M4) ShakeMap (Cambell and Bozorognia, 2008) with no consideration of instrumental data and site response 20

21. Empirical Data Incorporated (Bias adjustment at surface)
Empirical Data and Site Response Incorporated

22. March Earthquake
Shake and Loss Map 22

23. Municipality Governorate 1st Army
Communication of Rapid Response Message (Damage Maps)
(Mobile phones and PDA’s)
Governorate
1st Army
NUMBER OF COLLAPSED BUILDINGS PER CELL (Simulated from random data and communicated to end users every day at 10am)

24. Istanbul, Atakoy District SOSEWIN (Self-Organizing Seismic Early Warning Information Network) System
The development of a new seismic network for earthquake early warning (EEW), made up of low-cost sensors that will eventually be purchasable by a range of end users, giving very dense urban networks
The Seismological SOSEWIN will complement existing EEW networks.
Classical
Seismological Station
City
WLAN
DSL
Central Side:
A normal Node
Visualization
Backup, Monitoring
Gateways (Internet)
Public Node
Low Cost Node
Ref: Picozzi et. al.

25. Testing SOSEWIN: Ataköy district, Istanbul
Example of Earthquake recordings.
18 Sensing Nodes
2 Gateways + Sensing Nodes
Ref: Picozzi et. al.

26. Testing SOSEWIN: Fatih-Mehmet suspension bridge, Istanbul
Experimental array of 24 nodes: 10 along each side of bridge, 2 on either side of each tower.
Node installed in one tower
Node installed along the side
Ref: Picozzi et. al.

27. 24 SOSEWIN nodes recorded bridge vibrations for about 2 hours
about 1000 m
S678C
S6770
S6128
S61BC
S675C
S6754
S6174
S6190
S6124
S61A0
S6778
S6198
S677C
S6168
S61A4
S61D4
S676C
S61C0
S6750
S6794
S6798
S61B4
S61A8
S67A0
Spectra
Time series
In contrast to the high cost standard instrumentation, the low cost SOSEWIN sensors makes it possible the set up of a dense network for the monitoring of structures
Ref: Picozzi et. al.

28. Structural Health Monitoring
BOĞAZİÇİ BRIDGE
FATIH BRIDGE

29. FATIH BRIDGE

30. ENRON-TRAKYA ELEKTRIK
İŞ-KULE
ENRON-TRAKYA ELEKTRIK

31. Structural Health Monitoring
Hagia Sophia
Suleymaniye
Fatih
Mihrimah
Tube Tunnel
Hagia Sophia 31

32. Implementation of online Shake Mapping
The installation and customization of the USGS Shakemap code was completed in 2006
Test runs in scenario mode were made for some of the past earthquakes in Turkey,
Alternatively, an earthquake shaking and loss estimation routine (ELER) has been developed in connection with the EU FP6 NERIES project.

33. NERIES JRA3 ELER SOFTWARE
Now I will give you a brief information on ELER software
The software mainly has two modules which are the Earthquake Hazard Assessment module and the Earthquake Loss Assessment module.
Earthquake Hazard Assessment module provides us estimation of ground motion parameters and intensity distributions using ground motion attenuation relations, correlation between intensity and ground motion parameters and soil condition information.
Earthquake Loss Assessment module uses ground motion parameters and intensity information from Earthquake Hazard Assessment module, demography and building inventory. This module includes three levels (Level0, Level1 and Level2) of analysis for estimation of building damages and casualties. Level0 analysis estimates casualties based on magnitude and intensity information. Level1 analysis estimates casualties and building damages based on intensity information, Level2 analysis estimates casualties and building damages based on ground motion and spectral parameters.

34. Methodologies used in ELER Software
In Hazard Assessment Module
Modified USGS Shake-map algorithm
In Loss Assessment Module
For Building Damage Estimation:
Intensity Based Vulnerability (Level 1)
Macro-seismic method implicitly defined by EMS-98 scale (Giovinazzi S., 2005)
Spectral Displacement- Based Vulnerability (Level 2)
Capacity Spectrum Method (CSM)
Modified Acceleration – Displacement Response Spectrum Method (MADRS)
Reduction Factor Method
Coefficient Method
For Casualty Estimation:
Samardjieva & Badal (2002) – Level 0
Intensity based fatality rates – Level 0
KOERI (with ATC-13) – Level 1
HAZUS99 – Level 2
HAZUS-MH – Level 2
I will give more detailed information on hazard module and loss estimation module

35. Earthquake Hazard Module Calculation Methodology
The user has two options to enter the event data. The first option is to use an XML file containing the event parameters and station information, if available. The second option is to enter event data manually from the Hazard GUI.
The EHA module has three options to define the earthquake source which are: user defined point source, user defined extended source and auto assigned extended source.
The Site Correction panel determines how the effect of the local site conditions will be incorporated into the calculations of ground motion parameters
No Site Correction: In this mode all ground motion estimations are calculated at the engineering bedrock
Borcherds: all ground motion parameters are calculated at the engineering bedrock. The obtained grid based ground motion is then corrected with the site amplification factors (Fa and Fv) given in Borcherdt (1994) according to the selected Vs30 map.
In Eurocode 8 mode only the peak ground acceleration values are modified according to the site condition. Thus in this mode ELER produces only the site corrected PGA distribution
NGA at Surface: In this newly developed approach rather than calculating bedrock values and then amplifying these with respect to site conditions, ELER software uses attenuation relationships taking Vs30 as an input parameter to calculate the ground motion values directly at the surface
Site condition is represented by a parameter: the upper 30-m average shear wave-velocity (Vs30). The user has the following two options in choosing the site condition.
The default site condition map has been compiled from the USGS Global Vs30 Map Server.
Custom site condition maps should be in form of Vs30 grids
The final stage of the input specification is the selection of attenuation relationships. Since different attenuations are derived from different event catalogues the user should select a suitable attenuation taking into account the regional characteristics, magnitude and ground motion parameter of interest.
The selected attenuation is used to estimate measurable ground motion parameters such as PGA and spectral accelerations. Each attenuation function has its unique set of input parameters resulting from the regression analysis.
In addition to the ground motion parameters PGA, PGV, Sa and Sd, ground shaking intensity is also estimated through the regression relationships between MMI and PGA-PGV (Wald and others, 1999), and MMI-MSK and Fourier amplitude Spectrum (FAS).

36. Utilized Attenuation Relations
Ground Motion Estimation
Akkar & Bommer 2007
Boore & Atkinson 2007
Boore et al. 1997
Campbell & Bozorgnia 2007
Instrumental Intensity Estimation
Wald et al. 1999
Sokolov 2002

37. Intensity distribution (ShakeMap) - 1999 Kocaeli earthquake
FAS (Sokolov) Method
Observed Intensity Distribution
PGA/PGV-Intensity Correlations
Wald et al. (1999)
Regional Intensity Attenuation
Relationship

38. Google Earth KML Output of Earthquake Hazard Module
Intensity as Lines
Intensity as Polygons
KML outputs

39. Level 1 - Earthquake Loss Estimation Methodology
Level 1 (Intensity based building damage and casualty assessment)
Grid-based analysis
Estimations in regional and/or country scale
Basic input data available for Europe, for crude estimations of building damage
User defined building inventories are accommodated
Intensity based structural vulnerabilities from Lagomarsiono and Giovinazzi (2006) (EMS-98 Vulnerability Classes, Vulnerability Indices (V, Q), Vulnerability Curves, Damage Probability Matrices)
In Level-1 analysis we estimate casualties through building damage. Here, in addition to the demography we need to have building inventory.

40. Level 1 - Methodology based on Vulnerability Indices
RISK-UE Building Typology Matrix
European building typology
After obtaining the number of buildings we need to also estimate building structural types, height and age. These information were provided in percentages from Pager database, and adapted to the European building typology. The Risk-UE Building Typology Matrix has been used to assign the vulnerability indices for each building type.
The PAGER project provides the percentages of different construction types in all countries of the world for both urban and rural settlements and residential and non-residential occupancy types, making use of a HAZUS type classification. Corresponding European Building Taxonomy classes have been identified for the structural types of the PAGER classification system. Then these percentages have been used to convert the approximated grid based number of buildings to an inventory of different structural types in each county.
Ref: Giovinazzi S.,(2005)

41. Level 1 - Methodology based on EMS98 Vulnerability Classes
D3: Substantial to heavy damage

42. Level 1 – Casualty Estimation
Koeri,2002
Coburn&Spence

43. Case Study for 1999 M6 Athens Earthquake Level 0 and Level 1
Estimated Intensity Distribution
The 7 September 1999 M6 Athens earthquake
In total 143 people were killed, 1600 people were injured and at least 53,000 buildings were reported to be damaged by this earthquake.

44. Case Study for 1999 M6 Athens Earthquake Level 1 Building Damage and Casualty Estimation
Estimated Number of Damaged Buildings (3080 buildings in Damage States D3 + D4 + D5)

45. Estimated Number of Deaths (Total: 236 people)
Case Study for 1999 M6 Athens Earthquake Level 1 Building Damage and Casualty Estimation
Estimated Number of Deaths (Total: 236 people)

46. Stations Data From: http://earthquake. rm. ingv
Fault Data From: )

47. Bias Adjustment at Surface Instead of B/C Boundary

49. Level 2 - Earthquake Loss Estimation Methodology
The spectral capacity-based vulnerability assessment methodology is utilized for the building damage estimation.
The casualty estimation is based on the number of buildings in each different damage state.
Grid- (geo-cell) based urban building inventory and population data.

50. Building Typology Matrix Model Building Types of HAZUS-1999
RISK-UE
Building Typology Matrix
Model Building Types of HAZUS-1999

51. EARTHQUAKE LOSS ESTIMATION FOR ISTANBUL
The "Credible Worst Case" Scenario event:
An Mw=7.5 (similar to 1999 Kocaeli Earthquake in magnitude and in total rupture length) which is assumed to take place on the fault segments 5, 6, 7 and 8.

52. DISTRIBUTION OF DAMAGED BUILDINGS

53. DISTRIBUTION OF CASUALITIES

54. THANK YOU