Eastern Turkey Seismic
Experiment
Eric
Sandvol, Dogan Seber, and Muawia Barazangi, Cornell University, Ithaca,
New York
Niyazi Türkelli,
Cemil Gürbüz, Sadi Kuleli, Hayrullah Karabulut, Ekrem Zor,
Rengin Gök,Tolga
Beklar, and Esen Arpat, Kandilli Observatory and Earthquake Research
Institute,
Bogaziçi
University, Istanbul, Turkey
Salih Bayraktutan,
Atatürk University, Erzurum, Turkey
|
Summary
tectonic map of eastern Turkey (modified after Seber et al., 1997).
The large arrows are the approximate directions of motion of Turkey,
Arabia, and central Iran relative to Eurasia. NAF=North Anatolian
Fault; EAF=East Anatolian Fault; DSF=Dead Sea Fault.
|
Eastern Turkey has
for thousands of years been a region of convergence for European, Asian,
Persian, and Arabic peoples; a region where eastern and western cultures
come together. Similarly the tectonics of Turkey can be described as a
place of convergence or divergence of three continental plates. The Anatolian
sub-plate, the Arabian plate, and Eurasian plate form a diverse suite
of tectonic boundaries. Collision of the Arabian and Eurasian plates is
currently occurring in Eastern Turkey along the East Anatolian Fault Zone
(EAFZ) and the Bitlis suture.
The Bitlis suture/thrust
zone and the EAFZ mark a distributed, irregular, and young continental
collision zone as shown in the figure below. This collision zone is further
characterized by a 1500 meter topographic high (Anatolian plateau) with
corresponding seismicity indicating that this region is still actively
deforming. Tectonic escape is taking place to the west, as evidenced by
the right-lateral strike-slip movement along the North Anatolian Fault
Zone (NAFZ) (e.g., Reilinger et al., 1997) and left lateral movement along
the EAFZ (McKenzie, 1972). The slip along the NAFZ is approximately 2
cm/yr while the slip along the EAFZ is 1 cm/yr. This indicates that the
escaping Anatolian block is rotating counter-clockwise as shown in the
figure above. This rate of westward slip does not account entirely for
the strain induced by the 3.0 cm/yr convergence of the Arabian plate with
respect to Eurasia (Dewey et al., 1986; Reilinger et al., 1997); therefore,
this deformation must be accommodated elsewhere. The North and East Anatolian
faults have been active since the Miocene and are associated with large
pull-apart basins, such as the Karliova Basin located at the junction
of these two fault systems. GPS results from Reilinger et al. (1997) indicate
that the crust which lies to the north and east of the Karliova junction
is deforming very differently from the crust to the west. In the east
there appears to be shortening of the continental crust about 150 km north
of the Bitlis Suture, while in the west escape tectonics seem to dominate.
There is no consensus on precisely when collision between the Eurasian
and Arabian plates began. The age estimates of collision range from 12
Ma, based on stratigraphic
discontinuities in Eastern Anatolia (Sengor and Yilmaz, 1981) and the
beginning of collision related volcanism (Pearce et al., 1990), to 20
Ma based on the convergence rate of the two plates (Dewey et al., 1986).
Tectonic
map of Eastern Anatolia showing the major faults, ophiolites, and
recent volcanics. Also shown are the most recent measurements, relative
to the station near the Black Sea, of crustal deformation from Reilinger
et al. (1997). The majority of the volcanics shown are very young
(< 2 Ma).
|
The East Anatolian plateau
and Bitlis suture offer a unique and excellent opportunity to understand
the early stages of continental collision and its consequences: this understanding
is also essential to develop better models for the later stages of this
process. Until now, the only well studied example for active continental
collision has been that of India colliding with Eurasia and the subsequent
uplift of the Tibetan plateau. In many respects the East Anatolian topographic
uplift can be thought of as a younger version of the Tibetan plateau (e.g.,
Sengor, 1979; Dewey et al., 1986; Barazangi, 1989). In order to characterize
the nature of this collision zone and determine the geodynamics of the
collision we must resolve whether: (a) continental subduction or delamination
is ongoing (Rotstein and Kafka, 1982), (b) the Arabian plate convergence
is accommodated entirely by an escaping Anatolian plate (McKenzie, 1972),
(c) there is an Eurasian lithospheric thickening (Dewey et al., 1986),
or (d) a combination of these processes is taking place.
In order to address
these important questions, we have deployed a temporary 29 station broadband
PASSCAL array in eastern Turkey as shown in the map at right. We designed
our array to greatly improve our understanding of the Bitlis/Zagros thrust
zones as well as the nature of continental escape along the EAFZ and NAFZ,
imaging upper mantle and crustal structure in these two fairly different
tectonic environments. The average station separation is approximately
50 km for the western line of stations and 30 km for the eastern line.
Our western traverse crosses a region where it has been well documented
that the Anatolian block is escaping westward (e.g., Reilinger et al.,
1997). Along our eastern array, recent GPS results indicate that the crust
is being shortened.
Station
locations for the Eastern Turkey Seismic Array.
|
This project is a
collaborative, joint research between Turkish and Cornell researchers.
Scientists from Kandilli Observatory and Earthquake Research Institute
(KOERI), Atatürk University, and Cornell University began site selection
in August 1999 and were able to locate safe and stable sites for almost
all of the stations. Local government officials provided very valuable
logistical support. We chose sites that allow for relatively easy station
access and offer security from vandals and curious children (the latter
evident in the photo below). In many cases we installed seismometers within
government establishments. At four of the stations we arranged for local
residents to make sure that the station would not be disturbed. All station
sites were ready for deployment in mid-September 1999.
Luckily we had excellent
weather for late October and early November on the Anatolian Plateau and
using two teams consisting of three to four people each (including two
PASSCAL engineers) were able to deploy two stations a day. The entire
array was installed in approximately 2 weeks. Most of our stations were
powered by solar panels, however, we were able to arrange continuous AC
power at six of the twenty nine station sites. These six sites are located
on the Anatolian Plateau and experience rough winters. In order to ensure
that all the stations were operating properly, we visited each station
two to three weeks after installation in November 1999. Most were working
well, although we did encounter a few problems with solar panel power
at the easternmost end of the array. Using Linux laptops, we were able
to do all quality control in the field which is essential given that our
field computer and headquarters are located 800 km away in Istanbul.
After the
installation of the final station at Silvan in southeastern Turkey.
|
In order to ensure
maximum data recovery of local, regional, and teleseismic waveforms, we
are recording 24 bit broadband data continuously at 40 sps (samples per
second). Data recorded at 40 sps should provide high enough resolution
for accurately locating local events as well as analyzing regional wave
propagation. Using PASSCAL's 4.1 GByte field disks and the REFTEK 72A's
data compression algorithm, we plan to visit stations every 3 to 4 months.
All of the stations should be accessible throughout the year since we
deployed them within a kilometer of major roads.
The data recorded
by this array will greatly improve our knowledge of seismicity in this
region by revealing seismicity that would otherwise go undetected. The
data we collect will also provide a much improved seismic velocity model
with which we can more accurately locate earthquake hypocenters. A more
complete catalog of seismicity in Eastern Anatolia will provide crucial
information on national earthquake hazards as well as improve our understanding
of the current deformational patterns. The lack of broadband instrumentation
in the northern part of the Middle East makes this data set very important
to understanding seismic hazard and risk along the eastern portion of
the NAFZ. Furthermore, our stations will record the very frequent seismicity
of the Middle East and Eastern Mediterranean. We also plan to use this
extensive regional data set to improve the current regional wave propagation
and Pn velocity models (Hearn and Ni, 1994; Sandvol et al., 2000) in the
northern portion of the Middle East.
Preliminary receiver
functions for our stations located along the Bitlis suture suggest that
there may be evidence for a crustal root beneath the Anatolian Plateau.
Moreover, initial analysis of shear wave splitting shows that there is
no obvious change in mantle azimuthal anisotropy across the EAFZ (see
figure below);
further data and analysis will reveal if this is the case for the eastern
portion of the Bitlis Suture as well. In the long term, data processed
from this project will help to address some fundamental questions about
the extent and mechanism of this continental deformation and escape.
A
record section showing an SKS, pSKS, and SKKS phases recorded by
the western-most stations in the eastern Turkey network. All stations
show very clear evidence of shear wave splitting other than station
ILIC.
|
Given the lack of geophysical measurements in this region, determining
the optimal geodynamic model cannot be treated as a simple binary hypothesis
test. Thus it will be essential to integrate all of our seismological
results with the large number of geologic studies that have been conducted
in this region (geochemical, geochronological, GPS, etc.), creating a
single database. This data integration will allow us to choose the most
reliable tectonic and geodynamic model of this active continental collision
zone. Cornell's Middle East and North Africa GIS databases (see web site:
http://atlas.geo.cornell.edu) provide an excellent starting point and
mechanism for this data synthesis and integration. We plan to create an
East Anatolian GIS database and make it available via a special web site.
These results will constitute the first major broadband seismological
experiment in the Bitlis-Zagros thrust and mountain belt and should help
enormously in our understanding of the continent- continent collisional
processes.
Acknowledgments
This experiment could
not have happened without the enormous help provided by the federal and
local government officials in Turkey. We thank the President of Bogazici
University, Prof. Ustun Erguder, for his help and support during the initiation
of this project. We also thank Mustafa Aktar at KOERI for help and encouragement,
and PASSCAL engineers Bruce Beaudoin and Tom Jackson for their hard work
and helpful advice. Special thanks are due to the local residents of eastern
Turkey who helped in site selection and installation of this array. This
research is supported by the National Science Foundation, Geophysics Program
(Grant No. EAR-9804780). Additional support is also provided for the Turkish
researchers by a Bogazici University Research Fund (Grant No. 99T206).
|