1. Ground motion simulations and site effect estimation for Istanbul, Turkey Mathilde Bøttger Sørensen 1, Nelson Pulido 2, Sylvette Bonnefoy-Claudet 3, Kuvvet Atakan 1 1 Department of Earth Science, University of Bergen, Norway, 2 Earthquake Disaster Mitigation Research Center EDM, NIED, Kobe, Japan, 3 LGIT, Université Joseph Fourier, Grenoble, France.

2. Outline Introduction Ground motion simulations - technique - results Site effect studies - 4 methods - results Summary and further studies

3. North Anatolian Fault Zone

4. Westward migration of earthquakes on the NAF after Barka et. al. (2002)

5. Coulomb Failure Stress Change after the 1999 Izmit earthquake

6. Ground motion simulations Hybrid method for simulating ground motions due to finite-extent earthquake source in Marmara Sea Pulido and Kubo (2004): Deterministic at low frequencies, semi-stochastic at high frequencies Simulation on regular grid and on irregular grid of Istanbul Earthquake Rapid Response and Early Warning System stations

7. Ground motion evaluation from asperity model

8. Ground motion simulation technique Low frequency: Deterministic wave propagation from an asperity model in a flat layered velocity structure (Discrete Wave Number Method, Bouchon 1981) High Frequency: Semi-Stochastic Simulation based on an asperity model. The model combines the stochastic methodology of Boore (1986) with the empirical Green’s function method of Irikura (1986), and a high frequency radiation pattern model (Pulido et. al 2002).

9. Tectonic Setting Marmara Sea Region Okay et. al 2000

10. Scenario earthquake Active tectonic map of the Marmara Sea (Okay et. al 2000)

11. Source parameters Total seismic momentM 0 = 2.0·10 20 Nm Asperity areaS a /S = 0.22 Average stress drop5 MPa Asperity stress drop10 MPa Rise timeRandom, average 3.0 s Rupture velocityRandom between 2.8 – 3.2 km/s f max 10 Hz Q100 · f 1.5

12. Velocity model

13. PGV results

14. PGA results

15. Istanbul Earthquake Rapid Response and Early Warning System Kandilli Observatory and Earthquake Research Institute Bogazici University, Istanbul

16. Rapid Response system 100 accelerometer stations When triggered, station produces spectral acceleration at a number of frequencies, 12 Hz filtered PGA and PGV Data sent pr SMS every 20 s Main data center produces shake, damage and casualty maps, which are available to the end-users within 5 min

17. Istanbul Earthquake Rapid Response and Early Warning System Kandilli Observatory and Earthquake Research Institute Bogazici University, Istanbul

18. Early Warning system 10 accelerometer stations close to the Marmara fault When several stations triggers an alarm is declared Alarm level is sent to critical facilities, which can then shut down before the earthquake strikes

19. Use of RR/EW simulation results Calibration of Rapid Response system parameters Calibration of Early Warning system parameters Realistic scenario input for producing shake, damage and casualty maps Comparison to recorded earthquakes

20. PGV

21. PGA

22. Waveform example Avcilar district

23. Geology of Istanbul Map from Istanbul Metropolitan Municipality web pages, www.ibb.gov.tr

24. Geology of Istanbul Map from Istanbul Metropolitan Municipality web pages, www.ibb.gov.tr

25. Ataköy area, geology Ataköy From I.T.Ü. Gelistirme vakfi iktisadi isletmesi

26. Ataköy area, accessibility

27. 3 approaches H/V spectral ratio Standard spectral ratio 1D modelling

28. H/V spectral ratio

29. Noise recording Site A019

30. H/V results, Alluvium

31. H/V results, Bakirköy formation

32. Standard spectral ratio

33. Event locations

34. Results, SSR all events

35. Event locations

36. Results, SSR events 2 + 4

37. SSR results comparison

38. 1D modeling site Representative of alluvial site: 5 m Alluvium 8 m Bakirköy formation 80 m Güngören formation Bedrock

39. 1D modeling, sources Point forces with delta- like source time function at 4 and 8 m depth Distributed randomly in space, time, direction and amplitude Green’s functions are calculated using Hisada (1994)

40. 1D modeling, results

41. Summary Agreement between recorded and modelled microtremor H/V results Peaks at 1Hz and 3-6 Hz for alluvial site Peak at 1 Hz for Bakirköy formation SSR indicates high amplification levels, up to a factor of 10

42. Future plans Ground motion simulations for different source parameters Use simulation results as input in RR and EW system Assign site amplification factor to RR and EW sites