Artalan Sismik Gürültüler (ASG), yerkürenin, gerek kendi içindeki hareketlenmeleri
gerekse de dış etkenlerle etkileşimi sonucunda sismik enerjinin açığa çıkmasıyla
oluşan farklı periyot bantlarına sahip cisim ve yüzey dalgalarıdır. Daha önceleri
veriyi kirlettiği gerekçesiyle istenmeyen ve süzgeçlenerek bastırılmaya çalışılan
gürültü sinyalleri son yıllarda geliştirilen yöntemlerin de katkısıyla, yüzey
dalgalarından Green fonksiyonu eldesi ile kabuk ve manto hız yapısının araştırılması,
petrol aramaları, zemin araştırmaları gibi pek çok çalışmada veri olarak
kullanılmaktadır. Aynı zamanda gürültünün nitelendirilmesi dünyada binlercesi
kurulu olan sismografların performansını daha iyi anlamaya ve veri kalitesini
artırmaya yardımcı olacaktır.
Bu tez çalışmasının amacı, nüfusu yüksek, endüstriyel anlamda gelişmiş ve tektonik
olarak aktif bir coğrafyada bulunan sedimanter Adana Havzası ve çevresinin sismik
gürültü niteliğinin zamana, konuma, jeolojiye ve kültürel yapılarak bağlı olarak
araştırılması ve değerlendirilmesidir. Çalışmada kullanılan 2011 ve 2012 yıllarına ait
sürekli sismik kayıtlar, Boğaziçi Üniversitesi Kandilli Rasathanesi Deprem
Araştırma Enstitüsü (KRDAE) ve T.C Başbakanlık Afet ve Acil Durum Yönetimi
Başkanlığı Deprem Dairesi Başkanlığı (AFAD-DAD)' na ait, Adana Havzası' nda
bulunan 17 adet geniş bantlı Güralp CMG-3T ve Güralp CMG-3ESP marka hız
kayıtçıları ile toplanmıştır.
Havzanın sismik gürültü değerlendirmesine geçmeden önce yöntem ayrıntılarının
verilmesinin yanı sıra programı ve kullanılacak verileri tanımaya yönelik bir takım
çalışmalarda bulunulmuştur. Uygulanan gürültü analizi yönteminde, düşey bileşenli
verilerin güç yoğunluğu spektrumları (GYS), olasılık yoğunluk fonksiyonları (OYF)
ve spektrogramları açık kaynaklı PQLX (Passcal Quick Look eXtented) programı
kullanılarak hesaplatılmış ve çizdirilmiştir. Havzadaki gürültü dağılımının zamana ve
konuma bağlı değişiminin incelenmesi için GMT programı kullanılarak hazırlanan
haritalarda, kültürel gürültülerin yani mikrotremörlerin (kaynak: trafik, insan, makine
vb.) baskın olarak gözlendiği 0.1-1 s bandı ve mikrosalınım olarak da bilinen
mikroseizmların (kaynak: derin ve sığ su etkileşimleri) gözlendiği 4-8 s (İkincil
Yüksek Frekans veya Double Frequency Peak) ve 10-20 s (Birincil Yüksek Frekans
veya Single Frequency Peak) bantlarındaki gürültü verileri, 2011 yılı, yaz-kış
mevsimleri ve gece-güz saatleri için haritalanmıştır. Öte yandan kıyıya yakın bir
istasyon ile karasal ortama kurulmuş ve aynı jeolojik formasyon üzerindeki istasyon
arasındaki farklar spektrogramları hesaplanarak incelenmiştir. Bölgede üç bileşen xviii
kaydı mevcut olan KRDAE verileri kullanılarak gürültülerin mod değerlerinin yatay
ve düşey bileşenlerdeki durumu incelenmiş ve gözlenen farklar tartışılmıştır.
Elde edilen sonuçlara göre gün içindeki gürültü seviyesi değişimlerinin yüksek
frekanslı kültürel gürültülerden kaynaklandığı gözlenmiştir. Mevsimler arasındaki
değişimler ise ancak uzun periyotlarda gözlenebilmiştir. Genel olarak, kültürel
gürültüler yerleşimin ve sediman kalınlığının fazla olduğu havza içinde her zaman
yüksek değerlerde gözlenirken, uzun periyotlu mikrotitreşimlerin daha çok havza
sınırlarını etkilediği söylenebilir.
Çalışmada 2011 yılı düşey bileşen kayıtlar için periyoda bağlı elde edilen en yüksek
ve en düşük mod değerleri grafiklenerek bölgeye ait yüksek ve düşük gürültü modeli
oluşturulmuştur.
Bu çalışma ile bölgedeki istasyonların ve toplanan verinin kalitesi hakkında bilgi
sağlanmıştır. Özellikle bazı istasyonlarda veri kayıtlarının sürekli olmadığı veya alet
kökenli arızalar nedeniyle veri kalitesinin bozulduğu gözlenmiştir. Kullanılan
yöntem, kurulu istasyonların sürekli izlenmesi ve arızaların öngörülüp zamanında
müdahale edilmesine ve dolayısı ile veri kaydı sürekliliğine katkı sağlayabilir.
Çalışmada elde edilen en yüksek ve en düşük gürültü modeli ve gürültü haritaları,
bundan sonra bölgeye yerleştirilecek güncel istasyonlar için referans model olarak
kullanılabilir. Gürültü seviyesi düşük bölgeler daha düşük manyitüdlü depremlerin
kayıtlarına olanak sağlayabilirken, yüksek gürültülü bölgelerde manyitüdü daha
büyük depremler kayıt edilebilecektir.
Bu çalışmada elde edilen bilgiler yine gürültü verilerinin kullanılacağı dizilim
işleme, yatay/düşey spektral oranlama, artalan sismik gürültü tomografisi, zemin
özelliklerini ve sediman kalınlığını belirlemeye yönelik gürültü kökenli birçok
çalışmada referans olarak kullanılabilir.
|
The ambient seismic noise is widely used in imaging the subsurface structure. One of
the popular applications is the retrieval of surface waves Green's functions that are
used to obtain the velocity distributon in the Earth's crust and mantle. Ambient noise
is also used in oil exploration and in near-surface studies for soil characterizations.
In such studies the characterization of the noise can help to better understand the
performance of each seismograph and the noise levels at each station which can
contribute to interpretations of the results obtained by other ambient noise studies. In
this study the characteristics of background seismic noise levels in the Adana Basin
and its surroundings (Southern Turkey) are investigated.
The Adana Basin is an inter-mountain, NE-SW trending basin. The depositional
history, geological formations and thickness of the sediments suggest high potential
for hydrocarbon settings. Besides, it is a developing region both in population and
industry, a tectonically active area and hence prone to seismic hazards. Several
studies are performed or under progress to understand the subsurface structure of the
region. The characterization of the quality of noise data and the ambient seismic
noise levels of the region may contribute to the ongoing and future studies, such as
tomography, array processing and H/V spectral ratio studies, based on ambient noise.
For this purpose the continuous noise data is processed recorded in 2011 by the noise
analysis method given in Chapter 2.
Adana Basin and its surroundings, is an arcuate, open V shape structure extending in
NE-SW direction in the northeastern Mediterranean Sea. The basin is developed on
Tauride basement rocks mainly in Neogene times, affected by the interplay of the
African plate with the Arabian and Anatolian microplates. The basin is primarily
controlled by NE-SW trending faults that are splays of the larger sinistral East
Anatolian Fault. Three main rivers occupy the Adana Basin and form a major deltaic
complex in the region. The previous studies suggest a sediment thickness between 3
km to 6 km. The aim of this study is to characterize the seismic noise and the quality
of the continuous noise data in the region which will contribute to the recent and
future seismic studies. For this reason, the vertical component continuous data
recorded during 2011 by seventeen broadband Guralp CMG-3T, Güralp CMG-3T
and Güralp CMG-3ESP velocity sensors with sampling rates 50 and 100 sps are
used. The national seismic networks in the investigated sites are operated by KOERI
(Kandilli Observatory and Earthquake Research Institute) and ERD (Earthquake
Research Department of Disaster and Emergency Management Presidency of
Turkey). xx
In a continuous seismic record, besides noise, earthquakes, system transients and
instrumental glitches (data gaps, clipping, spikes, mass recenters, calibration pulses)
will be present. For the characterization of the noise, removal of this kind of
distortions is necessary. In the method mentioned in Chapter 2 the discrimination of
these distortions are provided by the calculation of the probability density functions
(PDFs) of the seismic data. As the distortions are represented by low probabilities,
the power values with high probabilities correspond to stationary background noise
at each station. Details of system transients and instrument glitches are given in
Chapter 3 with some demonstrations of seismic traces and PSD curves. In Chapter 4,
investigation of spatial and temporal noise properties in the Adana Basin and its
surroundings are studied. At last, final comments and results are discussed (Chapter
5).
In this study the open source software Passcal-Iris Quick Look eXtented (PQLX) is
use for processing the data. The data is prepared by selecting 1-hr time series
segments with an overlap of 50%. Each 1-hr segment is divided into 13 segments
with 75% overlap to reduce variance of the power spectral density (PSD) estimates.
After removing the mean, trend and instrument response, 10% cosine taper is
applied to the ends of the time series to reduce the spectral leakage in the resulting
Fast Fourier Transform (FFT). After computing the PSD of each time series segment
the results are converted to ground acceleration to compare with the NLNM and
NHNM. After calculating the PSDs and PDFs, first results show that the main
features of noise are coherent in general with the new low-noise model (NLNM) and
new high-noise model (NHNM) of Peterson noise models but different at few
localities. These features are cultural noise (T=0.1-1s, noise source: traffic,
machinery) and microseisms that are classified as Double Frequency Peak (DFP,
T=4-8s, noise source: ocean waves) and Single Frequency Peak (SFP, T=10-20s,
noise source: shallow coastal waters). Before the noise characterization, the usage
and properties of PQLX, its inputs and outputs, arrangement of the suitable response
function format, organization of the data set's header informations etc. are
investigated and applied. In order to investigate the noise characteristics in the basin
and its surroundings, PSD mode values are used for the cultural noise and
microseism periods and noise distribution maps are prepared for annual, seasonal and
diurnal variations by using GMT (Generic Mapping Tools).
To observe the annual noise changes, PSD mode values for the 3 different period
band have mapped for the year of 2011. For the cultural noise band (T=0.1-1s), it is
clearly seen that the noise level is higher than the other bands (DFP, SFP) and high
noise propagates from the coast region towards the basin. For the DFP band (T=4-
8s), relatively higher noise level stands near the border of the basin and also cay
station which has always the highest noise level in the basin. For the SFP band
(T=10-20s), there is a sharp noise decreasing as near as 30 dB through the whole
basin, but still middle of the basin has higher noise values.
To observe the seasonal noise changes, summer months (June, July, August 2011)
and winter months (December 2011, January-February 2012) are processed. In the
cultural noise band, noise distribution in summer is more extensive and propagates
from all coastal side into the basin; in winter, high noise level distribution remains
only on the coastal side. Contrarily, for the DFP band, noise variationss are observed
only on the borders of the basin except cay in summer, while almost whole study xxi
area has affected by the high noise such as -120 dB and -130 dB in winter. Lastly, for
the SFP band, again there is a sharp noise level decreasing for both seasons, but it is
seen that, in winter, especially Adana Basin and around the Taurus Mountain
(Northwest of the study area) has relatively high noise values. In addition to these
noise maps, to observe the seasonal changes with different graphs, spectrograms
were analysed for two different seasons as wet and dry seasons.
To observe the daily noise changes, the day (06:00-20:00) and the night (20:00-
06:00) times data are mapped. For the cultural noise, the day time shows high noise
levels especially near the coast and in the basin region. This high noise level reduces
about 10-20 dB at the night times. For the DFP, there is no significant variation
between day and night times. At SFP both for the day and the night times, the noise
levels in the basin are higher in comparison to mountain sides. The relatively high
noise levels in the basin are spread to wider area in the day time when compared to
the night times. Consequently, in daily basis, the significant variations occur at
cultural noise levels, almost no variation is observed at SFP, and slight variations are
observed at DFP periods.
In addition to three specific period band that defined above, for the 0.5-3s band, we
observe the DFP changes during a year and half yearly. It is seen that, local sea wave
activity affects the noise level, bathimetry changes could also be a good reason for
these changes.
Besides the noise map interpretations, some other investigations are also studied and
some are briefly mentioned as follows. Noise level difference between the coastal
and the crustal based stations which installed on the same geological structure is
studied by the analysis of the spectrograms. Also, as three-component data sets are
available for the KOERI stations, those mode values are graphed for the seven
stations and it is seen that the horizontal components (E, N) have higher noise levels
with respect to vertical (Z) component in high periods. Vertical-component PSD
mode curves for 17 stations are coherent with the Peterson noise models. The Adana
Basin mode high- and low- noise models are estimated which will provide useful
information before the installation of any station and also in earthquake magnitude
interpretations.
In this study, unfortunately it is observed that at some stations the data quality is poor
due to the interruption of data transfer, mass recenter and similar reasons, therefore
significant part of the data could not be used in processing.
For the future studies, the PSD calculations for the year 2012 and for the horizontal
components of the time series are under progress. Afterwards array processing and
the horizontal to vertical spectral ratio (HVSR) studies will be performed for the
region for further characterization of both noise and subsurface structure. |