13 Mart 1992 Erzincan Depremi artsarsıntıları kaynak parametreleri
Başlık çevirisi mevcut değil.
- Tez No: 56006
- Danışmanlar: PROF. DR. HALUK EYİDOĞAN
- Tez Türü: Yüksek Lisans
- Konular: Jeofizik Mühendisliği, Geophysics Engineering
- Anahtar Kelimeler: Belirtilmemiş.
- Yıl: 1996
- Dil: Türkçe
- Üniversite: İstanbul Teknik Üniversitesi
- Enstitü: Fen Bilimleri Enstitüsü
- Ana Bilim Dalı: Jeofizik Mühendisliği Ana Bilim Dalı
- Bilim Dalı: Belirtilmemiş.
- Sayfa Sayısı: 143
Özet
Erzincan ili ve civarında 13 Mart 1992 tarihinde oluşan, M=6.9 (NEIC) büyüklüğündeki yıkıcı deprem sonrası, İTÜ ve Fransa Strasbourg Louis Pasteuer Yer Fiziği Enstitüsü (IPGS)'nün birlikte yapmış olduğu çalışmalar sonucu,26 Mart - 4 Mayıs 1992 tarihleri arasında çeşitli türde deprem kayıtçıları ile 2000'e yakın artsarsıntı kaydı alınmıştır. Bu çalışma, 4-10 Nisan 1992 tarihleri arasında, 9 adet telsiz bağlantılı düşey bileşen sayısal hız kayıtçısı tarafından kaydedilmiş 362 artsarsıntıdan 80 tanesine ait kaynak parametrelerinin spektral yöntemlerle saptanmasını içermektedir. Genelde büyüklükleri MD ^3.2 olan bu sarsıntıların, odak derinlikleri 0-14 km arasında, odak-istasyon arası uzaklıkları ise 3-50 km arasında değişmektedir. Kaynak parametrelerinin saptanmasında teknik olarak, Brune'un dairesel kaynak modeli esas alınmış ve sismogramlar üzerine çeşitli spektral analiz yöntemleri uygulanmıştır. Buna göre, hız sismogramlardan P dalga fazlan saptanarak öncelikle alet etkisi giderilmiş, sonra integrali alınarak yerdeğiştirme sismogramlanna geçilmiştir. Daha sonra ise bunların Hızlı Fourier Dönüşüm (FFT) tekniği ile, frekans ortamı görünümleri (spektrumlan) elde edilmiştir. Spektrumlar üzerinden okunan spektral parametreler, düşük frekans seviyesi (Ho) ve köşe frekansı (/ö) yardımıyla incelenen artsarsıntılann dinamik kaynak parametrelerinden sismik moment (Mo), dairesel kaynak yarıçapı (r) ve gerilme düşümü (Ao) hesaplanmıştır. Spektral analiz sonucu, incelenen artsarsıntılann yerdeğiştirme spektrumlanndan; fo'm 4.72 - 20.64 Hz, Oo'uı ise 2.51xlO“7 - 1.03xlO”5 cm.sn arasında değiştiği bulunmuştur. Yüksek frekanslardaki asimtotik azalımın ise af7 ile orantılı ve azalım miktan /nın genelde 3'den büyük, 12 - 20 Hz arasında köşe frekansına sahip sarsıntılar için ise bu değerin 4-5 civarında olduğu gözlenmiştir. Sarsıntılara kalite faktörü (7nun çeşitli değerleri için düzeltmeler uygulanmış, Q değerinin arttikça/ı'ın değişmediği, Oo'm azaldığı, /nın ise Ö'ya bağlı olarak arttığı gözlenmiştir. Spektral parametreler yardımıyla bulunan kaynak parametrelerinden Mo'ın 2.10xl018- 2.45xl020 dyn.cm, Ao'nın 0.1 - 6.9 bar ve r'nin ise 110 - 490 m arasında değiştiği görülmüştür. Ayrıca, M0-MD arasında logA/0=(0.792 ± 0.062)A/D + (17.905 ± 0.111), MO-/O arasında logM0=(-2.774 ± 0.355)log/0 + (22.219 ± 0.383), AÖ--A/D arasında logAo=(0.485 ± Q.066)MD+ (0.852 ± 0.117), ve /0-MD arasında ise logfo=(-0.099 ± Q.Q24)MD+ (1.238 ± 0.042) şeklinde bağıntısal ilişkiler olduğu bulunmuştur.
Özet (Çeviri)
Öne of the majör subject in seismology is to model the earthquake source properties. Numerous theoretical and observational studies have been made nowadays aiming the modeling earthquake source. The purpose of these studies is to determine some kinematics and dynamic parameters of earthquake source and to explain the system of mechanism at the source by using these parameters. For över a decade, theoretical seismologist have developed several theories to model the earthquakes. These theories normally give the near- and far-field displacement spectra as a fünction of stress ör displacement time fünctions at the source (Kasahara, 1957; Archambeau, 1968; Berckhemer and Jacob, 1968; Brune, 1970). As the source models of different theories vary, the predicted displacement spectra vary. Observational seismology has logged behind by not checking empirically which of these theories, if any, correctly describes the far-field displacement spectra of earthquakes. The çare of the problem is to understand the rupture mechanism during an earthquake. Hanks and Wyss (1972) found that Brune's (1970) theory the best estimates the source parameters of vertical strike- slip earthquakes. Although Brune (1970) does not attempt to relate the source dimension to a theoretical P- wave spectrum, it nevertheless seems reasonable that this parameter can be recovered from the P- wave radiation. For several reasons, P- waves are preferable to S- waves for spectral analysis. The advantages of using P- waves are several. They are uncontaminated by earlier arrivals, and they are less sensitive to anelastic attenuation. in addition, the xnspectral shift of/b(P) from^(S) may make the P- wave preferable for the analysis, depending on the recording instrumentation (Hanks and Wyss, 1972). Body- wave spectra approximated in terms of two spectral parameters: Clo, the long-period level and f0, corner frequency and the decay rate of frequencies beyond fa. A general feature of all dislocation models is that fio is proportional to seismic moment M0 (Keilis-Borok, 1959; Maruyama, 1963; Burridge and Knopoff, 1964; Ben-Menahem and Harkrider, 1964; Aki, 1966) and/) is proportional to the reciprocal of the source radius r (Jeffreys, 1931; Kasahara, 1957; Archambeau, 1968; Berckhemer and Jacob, 1968; Brune, 1970). Assuming that the ruptured area was circular, stress drop Aa could be inferred from Mo and r (Wyss and Hanks, 1972) Anelastic attenuation has a major influence on seismic motions recorded at the surface. Anderson and Hough (1984) suggested that above the corner frequency the acceleration spectrum can be approximated by an exponential decay of the form e"*^. The spectral decay parameter y, is most likely controlled by attenuation at the site along the path, although it does not represent the entire contribution from attenuation in the earth (Hough et.al., 1988) and some variability in y may result from variations in seismic source radiation. Anderson (1991b) characterizes y as the sum of a site-dependent term and a distance-dependent variable that increases smoothly from an initial value of zero. At high frequencies, this exponential decay is the dominant, influence on the spectrum, for stronger than the site effect (Humphrey and Anderson, 1992). However both seismic moment and source radius depend critically on the effect of anelastic attenuation along travel paths. An independent estimate of quality factor Q should be performed using standard methods (Aki and Chouet, 1975; Anderson and Hough, 1984) to remove the attenuation filter from spectra or time signals. This study aims to determine source parameters that were calculated by P- wave spectra for 80 aftershocks of 13 March 1992 Erzincan earthquake (M=6.9, NEIC). These spectra were analyzed with respect to the Brune (1970, 1971)'s model as extended by Hanks and Wyss (1972) and yields estimates of seismic moment, source radius and stress drop for events that range in duration magnitude from less than 3.2. These source parameters were then correlated with the magnitude and spectral parameters (Clo and/0) of the aftershock sequences. After 13 March 1992 Erzincan earthquake, İTÜ and Institute of the Physics of the Earth, Strasbourg carried out a field study to record aftershocks. During the xmtime period between 26 March and 4 May 1992, around 2000 aftershocks were recorded by different type instruments. In this study, 80 aftershocks that were selected from 362 aftershocks recorded between 4 and 10 April 1992 by a telemetric network consisting 9 vertical component velocity seismometers, were used for analyzing. The focal depths and hypocentral distances of these aftershocks, range from 0 to 14 km and from 3 to 50 km respectively. All these aftershocks located between 39.5°- 40.0° N latitudes and 39.4°- 40.0° E longitudes. The seismic recording system is a telemetric network. The telemetric network is composed two distinct parts: a set of seismological stations distributed in the field and a central unit including the receiving and recording system. The seismometer output is amplified and converted into an FM signal through a voltage- controlled oscillator (VCO). This signal is transmitted via line-of sight UHF radio (406-470 MHz) to the centralized recording system. The network includes 9 seismological stations, 8 of them equipped with a single component, short-period vertical seismometer. One of the three components stations is installed near the central station. At the receiving center, recording is done after event detection by a triggering system and the signal is sampled rate of 92.308 samples/sec. The signals digitized at the dynamic range of about 70 dB. The digital seismograms were first converted from binary code to ASCII code and then displayed on a PC by using a program called PITSA (Programmable Interactive Toolbox for Seismological Analysis). The portion of the record encompassing the P- wave phase was windowed. At this step, no tapering was applied to the windowed P- wave phase. The next step of the analysis procedure is the deconvolution of the instrument response from the band-passed signal. This correction is carried out in two steps. First the reduced response was removed from the observed velocity signal. Secondly the least significant bit (LSB) correction was done. The LSB of data files corresponds to a velocity of the ground of 0.016 micrometer/second. The flat part of the response curve between 2 and 6 Hz. After instrument correction, the integration was taken according to trapezoid rule. After the integration, to remove the baseline shift or slowly varying components of the seismic traces, a baseline correction was applied. The baseline is calculated from the original trace by sliding an averaging window of a given length over the data series. For each window position, the baseline calculated as the XTVaverage value in data window. The window length controls directly the ruggedness of the baseline. Increasing the window length will increase the smoothness of the corresponding baseline. For all seismic traces, the width of the averaging window in seconds was taken 1.0 sec. The final step of the analysis is the Fast Fourier Transform (FFT) of the displacement seismogram of the P- wave phase. Each time series of the P- wave phase was tapered with cosine bells with a 12.5 %, before being converted to Fourier spectral density. Since the telemetric stations were deployed on rocks sites which may have high Q values and Q? values were not known in the Erzincan region, the spectra were not corrected for the effects of attenuation. However the effect of various Qe values were investigated and Öp was assumed as qo. Seismic source parameters, namely seismic moment, source radius and stress drop, were estimated using the low-frequency amplitude levels and corner frequencies measured directly from P- wave displacement spectra. The low- frequency levels (Qo) and high-frequencies spectral decays were determined by visual fitting of two straight lines to the spectra. The corner frequency (/ö) was obtained as the intersection of these two lines. Following Brune (1970, 1971)'s model that is extended to P- waves by Hanks and Wyss (1972), observed spectral parameters were used to calculate the seismic moment M0, the source radius r and the stress drop Act for each aftershock assuming that p =2.7 gr/cm3, P- wave velocity a - 6.0 km/sec, radiation pattern R^ = 0.4 (for P- waves) and k =2. Following Archuleta et.al (1982), when two or more stations recorded the same event, an average low-frequency spectral level () was estimated and used to obtain the corresponding average seismic moment (). For the corner frequency at each station, a source radius was computed using Brune (1970)' s relation for a circular fault. An average source radius () was computed taking arithmetic average for each aftershock. Finally using these parameters average stress drop () were estimated. At the end of this study, the source parameters and some relationships between these parameters were obtained. Most spectra have a constant long-period level and fall off beyond a corner frequency at a rate proportional to coy. y for these aftershocks is usually greater than 3 and frequently as high as 4 or 5 for XVspectra that have comer frequencies between 12-20 Hz. The seismic moment range from 2.10xl018 to 2.45xl020 dyn.cm. The stress drop range from 0.1 to 6.9 bars. The relationship between seismic moment and source radii is approximately linear. The source radius range from 1 10 to 490 m. The aftershocks with seismic moment greater than about l.SxlO19 dyn.cm tend to have larger stress drops. As a result of this study, a number of equations were found giving the relationships between spectral parameters Qo and/» and source parameters Mo, r, Act and duration magnitude Mb. Following equations are proposed for the aftershocks of 13 March 1992 Erzincan earthquake.. Seismic Moment (Mo) - Magnitude (Md), logM0 = (0. 792 ± 0. 062)MD +(17.905 + 0.111) (MD < 3.2). Seismic Moment (M0) - Corner Frequeny (/o), logM0=(-2.774±0.355)log/0+(22.219±0.383). Stress Drop (Act) - Magnitude (MD), logAcr=(0.485±0.066)MD+(0.852±0.117) (MD
Benzer Tezler
- Erzincan havzası 3-D hız yapısının yerel deprem tomografisi ile belirlenmesi
Determination of 3-D velocity structure of the Erzincan basin by local earthquake tomography
BÜLENT KAYPAK
Doktora
Türkçe
2002
Jeofizik Mühendisliğiİstanbul Teknik ÜniversitesiJeofizik Mühendisliği Ana Bilim Dalı
PROF. DR. HALUK EYİDOĞAN
- 13 Mart 1992 Erzincan depremi artsarsıntılarının kaynak zaman fonksiyonları
Başlık çevirisi yok
E.DİLEK GÜRLER
- Marmara Bölgesi'nde soğurulma yapısının incelenmesi
Investigation of attenuation structure in the Marmara region
AYŞE KAŞLILAR ÖZCAN
- 13 Mart 1992 Erzincan depremi (mw=6.8) için telesismik sonlu kaynak ve nokta kaynak ters çözümü
Teleseismic point and finite source analysis of the March 13, 1992 Erzincan earthquake
ASLI ŞANLI
Yüksek Lisans
Türkçe
2018
Jeofizik MühendisliğiSakarya ÜniversitesiJeofizik Mühendisliği Ana Bilim Dalı
PROF. DR. MURAT UTKUCU
- 13 mart 1992 Erzincan Depremi yapısal hasarları üzerinde yerel zemin koşullarının etkisi
Effects of local site conditions on structural damages of march 13, 1992 Erzincan earthquake
ÖZGE ASLAN
Yüksek Lisans
Türkçe
2015
İnşaat Mühendisliğiİstanbul Teknik Üniversitesiİnşaat Mühendisliği Ana Bilim Dalı
PROF. DR. MUSAFFA AYŞEN LAV