Abstract Demircioğlu MSc.
A PROBABILISTIC ASESSMENT OF THE SEISMIC HAZARD ANALYSIS FOR ISTANBUL REGION
Mine Betül Demircioğlu
(Thesis Supervisor: Mustafa Erdik)
ABSTRACT
Timeindependent probabilistic (simple Poissonian) and Timedependent probabilistic (renewal) models assuming characteristic earthquake based recurrence relationships for Istanbul region have been considered. The fault segmentation model was developed assuming that the seismic energy along welldefined faults is released by characteristic earthquakes characterized by magnitude and recurrence time. The used characteristic earthquake model is the Young and Coppersmith (1985) model. The interevent period between large (characteristic) earthquakes in these segments are also consistently gauged by dividing the seismic slip estimated from the earthquake catalog by the GPSderived slip rate of 22 ± 3 mm/yr. In the time independent model, the Poisson process is used in order to model the occurrence of the characteristic earthquakes on each segment. For the renewal model, using the mean recurrence interval of the characteristic earthquake, the elapsed time since the last major earthquake and the exposure period (taken as 50 years), the conditional probability for each segment has been calculated. Cascading of individual segments into longer multisegment ruptures has been considered using the partial cascading procedure as defined in Cramer al. (2000). As for the background, a spatially smoothed seismicity is determined for each cell of a grid composed of cells of size 0.005 x 0.005 degrees. Seismic activity in each cell is determined for earthquakes of magnitude between 5.0 and 7.0 assuming that a GutenbergRichter type recurrence relationship manages the earthquake recurrence in the background for magnitudes ranging 5.0 and 6.0. For background earthquakes with magnitudes 6.0 to 7.0 finite faults are used. The probabilistic methodology used in this study is similar to that used by the U.S. geological Survey to develop the current seismic hazard maps for the United States (Frankel et. al., 1996). Results are presented in terms of peak ground acceleration and spectral accelerations at 1.0 s. and also 0.2 second for the 10% and 2% probabilities of exceedence in 50 years for soft rock (NEHRP B/C boundary(Vs=760m/sn)) conditions. The earthquake hazard results are converted to sitedependent values to reflect the local site effects in Istanbul. For this reason, spectral response acceleration values obtained for NEHRP site class B/C boundary (Vs= 760 m/sec) are adjusted using Fa factors for short period and Fv factors for long period sitecorrection defined in the 1997 NEHRP Provisions (NEHRP 1997). Sitedependent results based on both the timedependent and timeindependent models are provided separately.
