Close
About
FAQ
Home
Collections
Login
USC Login
Register
0
Selected
Invert selection
Deselect all
Deselect all
Click here to refresh results
Click here to refresh results
USC
/
Digital Library
/
University of Southern California Dissertations and Theses
/
Constancy of strain release rates along the North Anatolian fault
(USC Thesis Other)
Constancy of strain release rates along the North Anatolian fault
PDF
Download
Share
Open document
Flip pages
Contact Us
Contact Us
Copy asset link
Request this asset
Transcript (if available)
Content
i
CONSTANCY OF STRAIN RELEASE RATES
ALONG THE NORTH ANATOLIAN FAULT
by
Özgür Kozacı
A Dissertation Presented to the
FACULTY OF THE GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(GEOLOGICAL SCIENCES)
December 2008
Copyright 2008 Özgür Kozacı
ii
ACKNOWLEDGEMENTS
I would like to thank my primary Ph.D. advisor Professor James Dolan for
giving me the opportunity to work on such an interesting problem and making sure
that I stayed on tracks towards putting together this work. In as much as I know that
he loves to micromanage every single detail, he let me do my research and provided
support when I needed.
I would also like to thank Dr. Robert Finkel for his generosity and patience
about sharing his knowledge about cosmogenic nuclide dating while making it possible
for me to process my samples at Lawrence Livermore National Laboratory Centre for
Accelerated Mass Spectrometry (LLNL-CAMS).
Members of my qualification and dissertation committees (Profs. Greg Davis,
Charlie Sammis from Earth Sciences, Dr. Robert Finkel from LLNL-CAMS and
Professor Darrell Judge from Physics & Astronomy) also need to be thanked sincerely
for their constructive questions and inputs regarding my studies.
Special thanks are due to my parents Semra and Cemal who encouraged and
supported me to pursue the answers to the questions on my mind all my life. My lovely
wife Ayşe has been extremely understanding and supportive during the most stressful
days, weeks, and months. If I made it this far I owe this to them.
The work presented in this dissertation was funded mostly by the National
Science Foundation (grant EAR-0409767, EAR-0633489, and EAR-9980564 with
additional support from U.S. Department of Energy (University of California, Lawrence
Livermore National Laboratory) contract W-7405-Eng-48.
iii
TABLE OF CONTENTS
Acknowledgements ii
List of Tables vi
List of Figures vii
Abstract xi
Chapter 1: Introduction 1
1.1 Introduction 1
1.2 The North Anatolian Fault Zone 2
1.3 Overview of Dissertation 5
Chapter 2: A Late Holocene Slip Rate For The North Anatolian Fault, Turkey, 8
From Cosmogenic
36
Cl Geochronology: Implications For The
Constancy of Fault Loading and Strain Release Rates
Abstract 8
2.1 Introduction 8
2.2 The Eksik Study Site 9
2.3 Age Control 14
2.4 Discussion and Conclusions 18
Chapter 3: A Late Holocene slip rate for the central North Anatolian fault, 21
at Tahtaköprü, Turkey, from cosmogenic
10
Be geochronology:
Implications for fault loading and strain release rates
Abstract 21
3.1 Introduction 22
3.2 Tahtaköprü Study Site 23
3.2.1 Site Description 23
3.2.2 Offset Measurement of Karanlık Dere 26
3.2.3 Detailed Geomorphology of the Fan Surface 28
3.3 Age Control 33
3.4 Discussion 38
3.5 Conclusions 49
Chapter 4: Paleoseismologic evidence for the relatively regular recurrence 50
of infrequent, large-magnitude earthquakes on the eastern North
Anatolian fault at Yaylabeli, Turkey
iv
Abstract 50
4.1 Introduction 51
4.2 Yaylabeli Site Description 55
4.3 Trench Results 57
4.4 Stratigraphy 62
4.5 Age Control 62
4.6 Interpretation of Paleo-Surface Ruptures at Yaylabeli 69
4.6.1 Event 1 69
4.6.2 Event 2 70
4.6.3 Event 3 72
4.6.4 Event 4 73
4.6.5 Event 5 74
4.7 Comparison of Yaylabeli and Çukurçimen Paleoseismologic 75
Data; Reproducibility of Trench Results
4.8 Discussion 77
4.9 Conclusions 83
Chapter 5: Slip-Rate Study Sites in Progress 85
5.1 Additional Slip-Rate Study Sites 85
5.2 Çatal Dere Slip-Rate Site 87
5.3 Yatçam Dere Slip-Rate Site 90
5.4 Destek Slip-Rate Site 93
5.4.1 Destek Slip-Rate Site Description 94
5.4.2 Age Control 97
5.4.3 Discussion 98
5.5 Slip-Rate Site at Koçyatağı (Mihar) 102
Chapter 6: Paleoseismologic Trench Sites in Progress 108
6.1 Additional Paleoseismologic Trench Sites 108
6.2 Paleoseismologic Trench Site at Elmalı 108
6.3 Paleoseismologic Trench Site Near Hamamlı 112
Chapter 7: Conclusions and Suggestions for Future Research 114
7.1 Conclusion 114
7.2 Suggestions for Future Research 115
References 116
v
Appendix A: Background Information on Geomorphology 126
and Age Control of the Eksik Study Site
Appendix B: Supplementary Data Figures for Yaylabeli 152
Paleoseismologic Trench Site
vi
LIST OF TABLES
Table 2.1 Cosmogenic (
36
Cl) and Radiocarbon (
14
C) Ages for Eksik 16
Slip-Rate Site
Table 3.1 Input Values Used For Calculating
10
Be Surface Exposure Ages 31
at Tahtaköprü Slip-Rate Site
Table 3.2 Terrestrial Cosmogenic Nuclide (
10
Be) Dating Results for 32
Tahtaköprü Slip-Rate Site
Table 3.3 Age Ranges and Resulting Slip Rates for Tahtaköprü Slip Rate Site 41
Table 4.1 Radiocarbon age dating results from the Yaylabeli site 66
Table 5.1 Cosmogenic
10
Be ages from Destek 100
Table A.1 Location and Size of
36
Cl Surface Samples from Eksik Study Site 145
Table A.2 Chemical Compositions of
36
Cl Samples from Eksik Study Site 146
Table A.3 Radiocarbon Ages from Fault Parallel Trenches at the Study Site 149
Table A.4 Summary of Slip Rate Calculations from Eksik Study Site 150
vii
LIST OF FIGURES
Figure 1.1 Simplified Tectonic Map of the Eastern Mediterranean Region 4
Figure 2.1 Location Map of the Eksik Study Site 10
Figure 2.2 Geomorphologic map of the Eksik Slip-Rate Site 12
Figure 2.3 Plot of
36
Cl Surface Exposure Ages and
14
C Charcoal Ages 17
from Eksik Slip-Rate Site
Figure 3.1 Location of Tahtaköprü Slip-Rate Site 24
Figure 3.2 Panoramic View of the Tahtaköprü Slip-Rate Site 25
Figure 3.3 Detailed Map of the Karanlık Dere Terrace Riser 27
Figure 3.4 A Simple Model To Explain The Effects Of Long-Term 29
Cultivation on an Alluvial Surface
Figure 3.5 Plot of Probability Distribution Functions Of
10
Be Surface Exposure 36
Sample Ages Collected From the Incised Alluvial Surface At
Tahtaköprü
Figure 3.6 Probability Distribution Functions Used to Calculate North 40
Anatolian Fault Slip Rate
Figure 3.7 Compilation of <75-Ka Slip Rates Along the NAF That Are 44
Based On Well-Determined Offsets and Geochronologically
Constrained Ages, Sorted By Time Scale of Measurement
Figure 3.8 Extents of 20
th
Century Earthquake Surface Ruptures and Plot 45
of Geodetic and Geochronologically Constrained Geologic
Slip-Rates along the NAF Sorted by Distance along the Fault
Figure 4.1 Simplified Tectonic Map of the Eastern Mediterranean Region 52
Showing the Yaylabeli Paleoseismologic Site Location
Figure 4.2 Topographic Map of Yaylabeli and Çukurçimen Study Sites 54
Figure 4.3 General View of the Yaylabeli Trench Site 56
viii
Figure 4.4 Map of the Yaylabeli Trench Site, Surveyed With a 58
Laser-Range-Finding Total Station
Figure 4.5 Log of West Wall of Yaylabeli Trench T1 59
Figure 4.6 Log of the Eastern Wall of Yaylabeli Trench T2 60
Figure 4.7 Log of Eastern Wall of Yaylabeli Trench T3 63
Figure 4.8 Logs of Eastern and Western Walls of Yaylabeli Trench T4 64
Figure 4.9 Probability-Distribution Functions for Radiocarbon Ages and 68
Layer-Modeling Results for Events E3, E4, and E5
Figure 4.10 Sediment-Accumulation-Rate Curve for the Yaylabeli Site 76
Figure 4.11 Space-Time Plot of Historical and Paleoseismologically 78
Documented Surface Rupturing Earthquakes along the North
Anatolian Fault
Figure 5.1 Simplified Tectonic Map of the Eastern Mediterranean Region 86
Showing Additional Slip-Rate Study Sites along the North
Anatolian Fault
Figure 5.2 Geomorphologic Interpretation of the Study Site and Detailed 88
Locations of the Exploratory Trenches
Figure 5.3 Destek Slip-Rate Site Location and Topographic map 95
Figure 5.4 A Panoramic View of the Destek Slip-Rate Site 96
Figure 5.5 Photo of a Quartz Gravel Sampled From the Surface and 99
Trench T1 Excavation
Figure 5.6 Probability Distribution Functions of
10
Be Ages (ka) from Destek 101
Figure 5.7 Location Map and Aerial View of Mihar Slip-Rate Site 104
Figure 5.8 Interpretation of Small Drainage Offsets in Mihar 105
Figure 5.9 Panoramic View of Mihar Slip-Rate Site 106
ix
Figure 5.10 Photograph of a Surface Sample and Depth Profile Trench 107
from Mihar Slip-Rate Site
Figure 6.1 Simplified Tectonic Map of the Eastern Mediterranean Region 109
Showing Hamamlı and Elmalı Paleoseismologic Trench Sites
Figure 6.2 A panoramic View of Elmalı Paleoseismologic Trench Site 111
Figure 6.3 Geomorphic Interpretation for the Hamamlı Paleoseismologic 113
Trench Site
Figure A.1 Regional Tectonic Map Showing Major Active Structures 129
Figure A.2 Corona Satellite Image Showing the Region Surrounding the 130
Eksik Study Site
Figure A.3 Three-dimensional Virtual View of the North Anatolian Fault 131
in the Region Near Eksik
Figure A.4 A Panoramic View of the Eksik Slip Rate Site 133
Figure A.5 A Simple Model Suggesting the Sequential Evolution of the 134
Geometry of the Terrace During Recurrent Displacements along
the North Anatolian Fault
Figure A.6 Offset Reconstructions of Geomorphic Features 140
Figure A.7 Logs of Trenches T2, T3, T4 and TP2-06 143
Figure A.8 An Example of the Sampled Limestone Cobbles Collected for 147
Surface Exposure Dating
Figure B.1 View Looking Southward across the Yaylabeli Trench Site 153
Figure B.2 A 3D View of the Faults and Trench Logs at Yaylabeli Trench Site 154
Figure B.3 Photo of the Fault Zone on Western Face of Yaylabeli Trench T1 155
Figure B.4 Photo of Western Wall of Trench T2 at Yaylabeli 156
x
Figure B.5 Photo of Western Wall of Trench T2 at Yaylabeli 157
Figure B.6 Photo Log of the Eastern Wall of Trench T2 at Yaylabeli 158
Figure B.7 Photo of the Fault Zone on the Eastern Face of Trench T3 159
at Yaylabeli
Figure B.8 Photo Log of the Eastern Wall of Trench T3 at Yaylabeli 160
Figure B.9 Photo Log of the Fault Zone in Trench T4 at Yaylabeli 161
Figure B.10 Photo of an Example of Peat Sampling Methodology Applied at 162
the Yaylabeli Trench Site
Figure B.11 A Close Up View of a Peat Sample from T4 in Yaylabeli 163
xi
ABSTRACT
The behavior of major active faults at various temporal and spatial scales is
one of the most fundamental, unresolved problems in modern tectonics. Determining
the degree to which fault loading and strain release rates are constant (or non-constant)
and documenting past earthquake occurrences are key approaches for understanding
this phenomenon. I have employed geomorphic mapping, Quaternary dating methods,
and paleoseismic trenching to generate fault rate and earthquake age data that help
provide a better understanding of the North Anatolian fault’s behavior in various
temporal and spatial scales. Specifically, I mapped offset geomorphic markers along
the North Anatolian fault and used cosmogenic nuclide (
10
Be and
36
Cl) and radiocarbon
(
14
C) dating methods to constrain the ages of these features. Using these and other
published data, I also constructed one of the first compilations of strain-release rates
for the North Anatolian fault. My compilation of these rate data reveals a constant
slip rate of ~15-20 mm yr
-1
over time scales of 10
3
-10
5
years. This result, however,
is slower than the geodetically constrained slip rate of 25 + 2 mm yr
-1
[Reilinger et
al., 2007], possibly indicating a strain transient. In addition to my slip rate studies,
I performed paleoseismologic trenching on the eastern part of the North Anatolian
fault at the village of Lorut. My results from this site demonstrate a relatively regular
occurrence of large earthquakes along this stretch of the North Anatolian fault. I attribute
the relatively constant strain release rates and regular earthquake recurrences to the
mechanical simplicity of the Anatolian-Eurasian plate boundary in northern Turkey,
which is dominated by the slip on the structurally mature North Anatolian fault.
1
CHAPTER 1:
Introduction
Introduction 1.1
The behavior of major active faults at various temporal and spatial scales is
one of the most fundamental, unresolved problems in modern tectonics. Determining
the degree to which fault loading and strain release rates are constant (or non-constant)
and documenting past earthquake occurrences are key approaches for understanding
this phenomenon. At time scales of dozens of earthquakes and hundreds of meters of
fault displacement, data from the San Andreas fault suggest that fault slip rates are
relatively constant over a wide range of time scales along certain sections of the fault
[e.g., Sieh and Jahns, 1984; Weldon and Sieh, 1985]. In contrast, there is increasing
paleoseismological and historical evidence from a number of different regions that strain
release is not constant in time and space at the time scale of individual earthquakes [e.
g., Ambraseys, 1970; 1989; Sieh et al., 1989; Fumal et al., 1993; 2002; Grant, 1996;
Dolan and Wald, 1998; Mann et al., 1998; Rockwell et al., 2000, Hubert-Ferrari et
al., 2002; Weldon et al., 2002]. These observations highlight a number of critically
important questions about the mechanisms by which faults store and release strain
energy: Are fault slip rates constant over all but the shortest time scales (i. e., one to
a few earthquake cycles), as would be expected if faults are loaded at a steady rate by
plate tectonics? Or are slip rates temporally variable at intermediate time scales of a
few to a few dozen earthquakes, as might be expected if loading were dominated by
transient phenomena (e. g., ductile creep events on lower crustal faults, or modulation
of lower crustal loading rates by regional earthquake clusters)? Over what time scales
2
do these different processes operate? And finally, do different faults and regional fault
networks behave differently at different time scales? In particular, do structurally
complex boundaries store and release strain differently than mechanically simpler
margins?
I chose the North Anatolian fault in Turkey as my target study area in order
to answer the above-mentioned fundamental questions for understanding earthquake
behavior. The reason for choosing the North Anatolian fault was its mechanical
simplicity and the availability of ~2000-year-long written records describing historical
earthquakes. I applied detailed geomorphic mapping of displaced landforms, cosmogenic
nuclide dating methods and paleoseismic trenching for quantifying Holocene slip-rates
and earthquake recurrence intervals at various temporal and spatial scales along the
North Anatolian fault. I proposed to test whether strain loading and stress release rates
are constant or non-constant, and whether large earthquake recurrence is relatively
regular or irregular along this major right-lateral strike-slip fault.
I have combined my findings with other data that became available during the
course of this research and I generated a space-time diagram of stress release rates for
the central and eastern sections of the North Anatolian fault.
The North Anatolian Fault Zone 1.2
The North Anatolian fault is an arcuate, right-lateral fault that extends for 1,500
km from the Karlıova triple junction in eastern Turkey westward across northern Turkey
and into the Aegean Sea (Figure 1.1). Together with the left-lateral East Anatolian
fault, the North Anatolian fault accommodates westward motion of the Anatolian block
in response to subduction rollback along the Hellenic trench and collision of Arabia
3
with Eurasia [McKenzie, 1972; Şengör, 1979; McClusky et al., 2000; Reilinger et al.,
2006]. Geologic studies indicate that the slip rate on the North Anatolian fault over
the past 10
3
-10
5
-years is ~15 - 22 mm yr
-1
[Hubert-Ferrari et al., 1997; 2002; Kondo et
al., 2004; Kozacı et al., 2007; in review; Pucci et al., 2007], whereas geodetic studies
reveal somewhat faster rates of strain accumulation of ~24-29 mm yr
-1
[McClusky et al.,
2000; Reilinger et al., 2006]. Both geologic and geodetic data indicate that almost all
of the strain associated with the westward extrusion of the Anatolian block in northern
Turkey is accommodated along the North Anatolian fault system. Specifically, geodetic
measurements indicate that internal deformation rates within central Anatolia are <2
mm yr
-1
[McClusky et al., 2000; Reilinger et al., 2006]. Thus, the North Anatolian fault
is the dominant fault within a relatively simple plate boundary.
Between 1939 and 1999, a >1,000 km length of the North Anatolian fault
ruptured during a generally westward-propagating series of eight M≥7 earthquakes
[Ketin, 1948; Richter, 1958; Allen, 1969; Ambraseys, 1970; Toksöz, et al., 1979;
Barka, 1992, 1996, 1999; Ambraseys and Finkel, 1995; Stein et al., 1997; Barka,
1999]. This sequence included four very large earthquakes: the 1939 M
w
7.9 Erzincan
earthquake, the 1943 M
w
7.7 Tosya earthquake, the 1944 M
w
7.5 Bolu earthquake, and
the 1999 M
w
7.5 İzmit earthquake. Stein et al. [1997] have shown that, in general, the
westward propagation of this sequence is consistent with a model of stress triggering
of each earthquake by earlier events in the sequence.
4
Figure 1.1 Simplified tectonic map of the eastern Mediterranean region [modified
after Barka, [1992] and Armijo et al., [2002]. Arrows indicate generalized relative
motions according to the fixed Eurasian Plate showing rates in millimeters per year
[after McClusky et al., [2000] and Reilinger et al., [2006]. Red boxes indicate slip rate
study site locations. Black boxes indicate paleoseismologic study site locations. NAF-
North Anatolian fault, EAF-East Anatolian fault, NEAF-North East Anatolian fault,
OF-Ovacik fault, EF-Ezinepazari fault, DSF-Dead Sea fault. Satellite image is from
Orbitview-2 satellite (available at http://visibleearth.nasa.gov).
5
Overview of Dissertation 1.3
In this dissertation, I present a number of slip-rate sand paleoseismologic trench
sites that I worked on during the last five years as part of my PhD studies. Three of
these studies are completed and either published or in review in peer-reviewed journals
[Kozacı et al., 2007; Kozacı et al., 2008; Kozacı et al., in review]. Much of the work
presented here requires collaboration as with most of the field-based studies. My
collaborators are listed below.
Chapter 2 presents one of the first geochronologically determined Late-Holocene
slip-rate along the central section of the North Anatolian fault [Kozacı et al., 2007].
This paper was written in collaboration with my co-authors James F. Dolan (University
of Southern California), Robert Finkel (Lawrence Livermore National Laboratory) and
Ross D. Hartleb (University of Southern California) and was published in 2007 in the
journal Geology. I mapped and dated an offset fluvial terrace by using cosmogenic
radionuclide (
36
Cl) and radiocarbon (
14
C) dating. These dating methods yielded two
independent late Holocene slip-rates of 20.5 ± 5.5 mm yr
-1
and 20.5 ± 8.5 mm yr
-1
for
the central part of the North Anatolian fault, respectively [Kozacı et al., 2007].
In Chapter 3, I present another late Holocene slip-rate from a site near the
village of Tahtaköprü, also along the central North Anatolian fault, approximately 200
km to the east of Eksik slip-rate site. This paper was also written in collaboration
with James F. Dolan (University of Southern California) and Robert Finkel (Lawrence
Livermore National Laboratory) and has been accepted for publication in the Journal
of Geophysical Research. At this site, I measured a right-lateral stream offset of 55 +
10 m, with no vertical displacement. Cosmogenic dating (
10
Be) of the alluvial surface
into which the stream incised yield a preferred late Holocene slip rate of 18.6 +3.5/-
6
3.3 mm yr
-1
for the central part of the North Anatolian fault at Tahtaköprü, Turkey; use
of variable cosmogenic production rate (VPR) models yields a slightly slower rate of
~16.4 +6.4/-4.5 mm yr
-1
[Kozacı et al., in review].
In Chapter 4, I present the results from a paleoseismic trench site that I developed
over a course of three years. I excavated paleoseismologic trenches across the eastern
part of the North Anatolian fault at the village of Yaylabeli for documenting historical
earthquake occurrences. These trenches provided evidence for five surface ruptures
during the last 2,000 years. My results are generally similar to those from the nearby
Çukurçimen trench site [Hartleb et al., 2006], located 2 km to the east, demonstrating
reproducibility of the paleoearthquake record. However, the 8
th
to 9
th
century event (E4)
that I document at Yaylabeli was not observed at Çukurçimen. The addition of this event
facilitates the recognition of a previously unnoticed NAF earthquake cluster, during
which the eastern and central parts of the fault appear to have ruptured during a brief
sequence in the 8
th
and 9
th
centuries [Kozacı et al., in review]. This paper was written
in collaboration with co-authors James F. Dolan (University of Southern California),
Önder Yönlü Eskişehir Osmangazi University), and Ross D. Hartleb (University of
Southern California). As of this writing this paper is in peer-review at the Geological
Society of America Bulletin.
In Chapter 5, I present additional slip-rate study sites along the North Anatolian
fault, yet to be concluded.
Two additional paleoseismic trench sites are presented in Chapter 6. At these
sites my trenching efforts did not yield any evidence for documenting historical
earthquakes.
I use these new data, in combination with other published slip-rate and
paleoseismologic results, to generate an updated space-time diagram of earthquake
7
occurrence along the North Anatolian fault over the past 2,000 years. I then discuss
these results in light of their implications for long-term patterns of seismic strain release
on the North Anatolian fault.
8
CHAPTER 2:
A late Holocene slip rate for the North Anatolian fault, Turkey, from
cosmogenic
36
Cl geochronology: Implications for the constancy of
fault loading and strain release rates
Abstract
Geomorphologic mapping and cosmogenic radionuclide (
36
Cl) dating of an
offset fluvial terrace yield a preferred late Holocene slip rate for the central part of
the North Anatolian fault (NAF) of 20.5 ± 5.5 mm yr
-1
, with an independent slip rate
constrained by
14
C ages of 20.5 ± 8.5 mm yr
-1
. These rates are generally similar to,
but possibly slightly slower than, the short-term rate of elastic strain storage of 25
± 2 measured geodetically across this major strike-slip fault [Reilinger et al., 2006],
suggesting that loading and strain release on this part of the NAF have been relatively
constant when averaged over the past ~2–2.5 ka. I attribute this consistency to the
relatively simple structure of the Anatolia-Eurasia plate boundary in north-central
Turkey, where almost all plate boundary strain is accommodated along the NAF. The
absence of other moderate- to high-slip rate faults (and the earthquakes they produce)
leads to a relatively simple stress evolution for the fault dominated by steady tectonic
loading.
2.1 Introduction
The degree to which fault loading and strain release rates are constant (or non-
constant) in time and space is one of the most fundamental, unresolved issues in modern
geodynamics. In particular, it is not clear what processes control strain release at time
scales longer than one to a few earthquakes. Is strain release relatively continuous, as
would be expected if faults are loaded steadily from beneath by plate motions? Or is
9
strain release markedly non-constant, as might occur if fault loading is controlled by
temporally transient mechanisms, such as lower crust creep pulses or feedback related
to earthquake clustering? Over what time scales do all of these processes operate?
And do different plate boundaries store and release strain differently, perhaps as a
consequence of relative structural complexity or simplicity? In this paper, I document
the late Holocene slip rate of the central part of the North Anatolian fault in Turkey. I
discuss this result in terms of its implications for understanding the constancy of strain
storage and release on this major strike-slip fault, and speculate about the causes of the
different behavior of major continental strike-slip faults.
2.2 The Eksik study site
The study site is located near the village of Eksik, Turkey, along the central part
of the North Anatolian fault (NAF) (Figure 2.1; Appendix A1). The NAF, which extends
across the study site as a narrow (2 to 10-m-wide), well-defined N65°E-trending zone,
is oriented nearly perpendicular to several deeply incised, south-flowing drainages
that have been offset by right-lateral slip along the fault [Hubert-Ferrari et al., 2002]
(Figure 2.1b).
Initial incision of these drainages resulted in the development of broad canyons
incised into bedrock. These large canyons, which have been consistently offset ~200
m along the NAF, are thought to have formed at ~10 ka [Hubert-Ferrari et al., 2002].
Inset into the bases of these offset canyons are a set of younger (mid- to late Holocene)
incised canyons that exhibit smaller offsets. Following their incision, these younger,
inset canyons were filled by fluvial gravels composed of white limestone cobbles and
pebbles derived from steep limestone bedrock cliffs situated 1.0-1.5 km north of the
fault. Subsequent incision by active streams into these gravels has isolated several
10
Figure 2.1 (a) Location of the Eksik study site (star) along North Anatolian fault
(NAF) in north-central Turkey. (b) Corona satellite image overlaid with geomorphic
interpretation [modified from Hubert-Ferrari et al., 2002]. Offset fluvial terrace
discussed in text highlighted in yellow. BD = Beygiruçtu Dere, CD = Çatal Dere, AD =
Ağılönü Dere, YD = Yatçam Dere, KD = Kuru Dere. “Dere” means “creek” in Turkish.
11
geomorphically well-defined, nearly flat fluvial terrace remnants nestled within the
much larger, older canyons incised into bedrock.
I mapped in detail the offset of one of these fluvial terraces along Ağılönü
Canyon, one of the major incised drainages located 2 km west of Eksik (Figure 2.2).
Two active streams, which in the vicinity of the fault flow southward near the eastern
and western edges of Ağılönü Canyon, have incised into the fluvial terraces by ~20 m.
Although this recent incision has isolated the terrace into several distinct remnants,
similarities in clast content, surface morphology, elevation, soil development, and age
(discussed below) indicate that these remnants are all parts of a once-continuous fluvial
aggradational fill sequence of Ağılönü Canyon.
In general, the inner edges of the fluvial terrace, where the planar upper surface
of the terrace gravels abuts the canyon walls, are geomorphically well defined over
most of the site. The white limestone-cobble gravels that make up the terrace contrast
markedly with the underlying dark grayish-green schist bedrock and bedrock-derived
colluvium, greatly facilitating mapping of the limits of the terrace deposits. In addition,
the buttress unconformity marking the inner edge of the top of the terrace gravels
is exposed in several natural side drainages along the west side of Ağılönü Canyon.
Subsequent to abandonment of the terrace, the buttress unconformities at the inner
edges of the terrace deposits have been locally covered by bedrock-derived colluvium.
In order to determine the exact location of the terrace inner edges as close to the fault
as possible, I excavated several 2- to 5-m-deep, fault-parallel trenches and test pits
(Figure 2.2; Appendix A1).
There are two questions relevant to determining the slip rate at Eksik: (1) what
is the geometry of the terrace inner edge where it crosses the fault; and (2) what is the
relationship between the offset of the inner edge and the age of the terrace? Offset of
12
Figure 2.2.( a) Geomorphic map of the study area showing the offset fluvial terrace described in text, as well as the locations
of the cosmogenic radionuclide samples (numbered black dots) and my trenches (T1, T2,…etc.). Topographic base map was
surveyed with a laser-range-finding Total Station with electronic compass attachment. Black dashed lines show the inferred
inner edge locations projected from the nearest exposures. Circles with dots indicate surveyed locations of exposures of the
terrace inner edge. Circles with arrows show limiting locations for the terrace inner edge. (b) Restoration of 46 m of right-
lateral slip along the NAF aligns the western terrace inner edge, the active channels that have incised into the terrace deposits,
and the western edge of the preserved mid-canyon terrace remnants. The consistent offsets of these different features suggest
that incision of the terrace deposits occurred soon after stabilization of the terrace surface. Heavy black dashed line highlights
the relatively linear western edges of the two mid-canyon terrace remnants. Inset shows my proposed model for the sequential
development of these features. Bedrock shown as green with white dots denotes area where terrace inner edge has been eroded.
13
the terrace inner edge along the eastern side of Ağılönü canyon is not well defined,
because the upper part of the terrace (including the inner edge) just south of the fault
offset has been eroded by the modern stream. As a result, I concentrated my efforts
on the well-exposed, western inner edge of the terrace. My trenches and the natural
exposures allowed me to map the western terrace inner edge in detail for hundreds of
meters along Ağılönü Canyon, to within a few tens of meters of the fault (burial by
colluvium precludes direct examination of the inner edge closer to the fault). North of
the fault, the inner edge curves gently downstream from a several-hundred-meter-long,
linear reach that trends S20°E, to a S35-40°E trend as it approaches the fault. South
of the fault, the inner edge exhibits a linear, S40°E trend for >200 m downstream.
Restoration of 46±10 m of right-lateral displacement yields a best-fit, gently curved
original channel morphology (see discussion in Appendix A1). This restoration also
restores the incised channels of the two major active streams that have cut down
through the terrace deposits, as well as the linear western edges of the mid-canyon
terrace remnants (Figure 2.2).
Consideration of the similar offsets of the terrace inner edge and the
canyons that have been incised into the terrace gravels suggests an answer to the second
question. As with all fill terraces, the terrace gravels could have been deposited into a
partially pre-existing offset of the canyon, yielding an incorrect over-estimate of the
offset measurement and slip rate. However, in that case the offset of the incised channels
of the active streams, which must have been cut after emplacement and abandonment of
the fluvial terrace, would have to be less than the offset of the buttress unconformities at
the terrace inner edge. The observed concordance between the two offsets indicates that
all, or nearly all, of the inner edge offset has accumulated subsequent to abandonment
of the terrace surface. This implies that the modern streams began incising soon after
14
initial stabilization of the fluvial terrace surface, a likely occurrence given the very
steep gradients of these high-energy mountain streams. I have used the inner edge
offset rather than channel offset to determine the fault slip rate because it is a more
precisely determinable quantity.
2.3 Age Control
I used
36
Cl cosmogenic surface exposure dating to determine the age of the
offset Ağılönü Canyon terrace. I collected a north-south, fault-perpendicular transect
of ten limestone cobbles for cosmogenic dating from the surface of the two largest
and geomorphically best-defined terrace remnants (Figure 2.2). The transect extends
for 50 m north of the fault (4 samples) and 75 m south of the fault (6 samples). The
sampled terrace surfaces are quite planar, suggesting little to no erosion. These surfaces
have <10 cm of local relief, except near the fault offset, where the northern terrace
surface locally exhibits as much as 35 cm of relief. The terrace surface is characterized
by a large-pebble and cobble pavement, through which a thin (2–10 cm), weak A-C
surface soil has developed. I was careful to sample only cobbles that appeared to be
stable within the interlocking pavement surface and that did not exhibit any evidence of
having been moved. Specifically, all of the sampled cobbles had a well-defined upper,
slightly weathered surface that contrasted markedly with a soil-stained, smoother lower
surface.
Ages were calculated using the program CHLOE (CHLOrine-36 Exposure age)
[Phillips and Plummer, 1996], using two different estimates of cosmogenic radionuclide
production rates. Nine of the ten
36
Cl surface samples yielded remarkably clustered
ages for the terrace surface, ranging between 2.1 and 2.9 ka for the production rates of
Stone et al. [1996; for Ca] and Evans et al. [1997; for K], and between 1.6 ka and 2.2
15
ka for the production rates of Phillips et al. [1996] (Figure 2.3; Table 2.1; see Appendix
A1 for discussion). One sample (7) yielded an anomalously young age of 0.9 or 1.2 ka.
I interpret this sample as an outlier, the young age probably reflecting displacement from
its actual exposure site or from physical alteration of the cobble during its exposure history.
There are actually two age peaks within the main cluster of nine
36
Cl ages - a younger
sub-cluster of six samples, and an older set of three samples. I suspect that the older sub-
cluster reflects minor inheritance in these three samples on the order of a few hundred
years, consistent with the short transport distance from the source of the cobbles to the
terrace. If true, this would suggest that the six tightly clustered ages represent more
accurately the age of terrace stabilization. These preferred age ranges are 2.1–2.4 ka
for the production rates of Stone et al. [1996] and Evans et al. [1997], and 1.6–1.8 ka
for the production rates of Phillips et al. [1996].
In addition to the
36
Cl exposure dates from the terrace surface, radiocarbon
dates from four charcoal samples collected from above and below the youngest terrace
gravels exposed at the terrace inner edge in trench T4 along the east edge of Ağılönü
Canyon provide an independent constraint on the age of the terrace surface and on the
general validity of the
36
Cl production rates used in the cosmogenic age calculations. A
charcoal sample collected from colluvium 10 cm above the top of the terrace gravels
yielded a calibrated calendar age of 1.95-2.12 ka (all calibrations were done using
OxCal v.3.10 [Bronk Ramsey, 1995; 2001; using atmospheric data from Reimer et
al, 2004])(Figure 2.3; see Appendix A1). Three charcoal samples recovered from
colluvium beneath the terrace gravels in trench T4 yielded calibrated, calendar ages
of 2.92–3.22 ka, 2.81–2.92 ka, and 2.42–2.92 ka, in correct stratigraphic order. These
four radiocarbon ages thus bracket the age of the easternmost, youngest terrace gravels
to be between 1.95 and 2.92 ka. As shown in Figure 2.3, age range matches closely the
16
Sample #
Philips et al., (1996)
production rates (ka)
Stone et al., (1996) and
Evans et al., (1997)
production rates (ka)
36
Cl-1 2.06+0.05 2.74+0.05
36
Cl-2 1.82+0.06 2.41+0.06
36
Cl-3 2.09+0.07 2.76+0.07
36
Cl-4 1.63+0.05 2.17+0.05
36
Cl-5 2.21+0.07 2.93+0.09
36
Cl-6 1.84+0.08 2.43+0.11
36
Cl-7 0.90+0.03 1.20+0.03
36
Cl-8 1.61+0.05 2.14+0.07
36
Cl-9 1.64+0.05 2.17+0.07
36
Cl-10 1.74+0.05 2.32+0.07
Sample #
14
C ages (yrs BP)
Corrected Calendar
Ages
EX-4-5 2020+30 110BC - 60BC
EX-4-4 2590+80 910BC - 410BC
EX-4-1 2695+35 910BC - 800BC
EX-4-3 2865+45 1210BC - 910BC
Cosmogenic Radionuclide Sample Ages (
36
Cl)
Radiocarbon Sample Ages (
14
C)
Table 2.1 Cosmogenic (
36
Cl) and Radiocarbon (
14
C) ages.
36
Cl ages are reported for
the different
36
Cl production rates of Philips et al. [1996] and of Stone et al. [1996]
and Evans et al. [1997]. Errors on radionuclide ages reflect only 1 sigma analytical
uncertainties in the AMS measurements. Consideration of all sources of uncertainty
would probably result in 10% to 15% standard deviation [Phillips et al., 1966;
Figure A1.4]. Radiocarbon ages reported as 95.4% confidence limits from OxCAL
version 3.10 [Bronk Ramsey, 1995; 2001; using atmospheric data from Reimer et
al., 2004]. See Appendix A1 for locations of
14
C samples collected from trench T4.
17
Figure 2.3 Plot of
36
Cl surface exposure ages and
14
C charcoal ages for samples collected
from beneath (
14
C-1, 3 and 4) and above (
14
C-5) the terrace gravels in the Trench T4. Simple
Gaussian probability distribution functions (pdfs) for
36
Cl data created in Matlab using
Camelplot [Balco, personal communication]. Radiocarbon pdfs data created in OxCAL
[Bronk Ramsey, 1995; 2001; using atmospheric data from Reimer et al., 2004]. For the
36
Cl ages the green and blue lines are Gaussian probability distributions for the individual
36
Cl measurements. Probability values are normalized so that each individual probability
distribution has unit area. Shaded areas are sum of probability distributions for all samples.
For
14
C the plots give calibrated probability distributions from OxCal normalized to
unit probability. The ordinate axis gives relative values independently for
36
Cl and
14
C.
18
2.1-2.4 ka age range of the sub-cluster of six
36
Cl dates determined with the production
rates of Stone et al. [1996] and Evans et al. [1997]. I therefore use the Stone et al.
[1996] and Evans et al. [1997] production rates in my preferred slip rate calculations.
2.4 Discussion and Conclusions
I use the
36
Cl ages from the offset fluvial terrace to calculate two estimates of
the late Holocene slip rate for the North Anatolian fault at Eksik. As discussed, the
36
Cl
ages are not normally distributed, probably because a subset of the cobbles experienced
a few hundred years of exposure prior to being incorporated into the sampled terrace
(Figure 2.3). However, in order to obtain a conservative slip rate estimate, I used all
my data except the single obvious young outlier; and instead of using the mean and
standard deviation to calculate the uncertainty on my measured slip rate, I use the
range. Combining my 46 ± 10 m offset of the terrace inner edge with the 2.1–2.9 ka age
range determined for the nine
36
Cl samples on the basis of the radionuclide production
rates of Stone et al. [1996] and Evans et al. [1997] yields a slip rate of 19 ± 7 mm yr
-1
.
The rate calculated using the 1.6–2.2 ka terrace age based on the production rates of
Phillips et al. [1996] yields a slip rate of 25.5 ± 9.5 mm yr
-1
. If I assume that the very
tight sub-cluster of six
36
Cl ages reflects most accurately the age of terrace stabilization,
the ages approach a normal distribution and I can use the standard deviation to estimate
the uncertainty in the ages, then, propagating the uncertainties, my preferred rates
are 20.5 ± 5.5 and 27 ± 7 mm yr
-1
, respectively. The radiocarbon dates bracket the
youngest terrace gravels and provide an independent, slip rate of 20.5 ± 8.5 mm yr
-1
.
Alternatively, using only the calibrated, calendric
14
C age of the underlying charcoal
sample yields a robust minimum rate of ≥18 ± 4 mm yr
-1
(Table 2.1). All of these rates
are statistically indistinguishable from the short-term, 25 ± 2 mm yr
-1
rate of elastic
19
strain accumulation across this part of the NAF measured by geodetic data [Reilinger
et al., 2006]. The general consistency of these rates suggests that strain storage and
release on this part of the NAF has been relatively constant when averaged over the
past ~2000–2500 yr.
I attribute this approximately continuous loading and release to the relative
structural simplicity of this part of the Anatolia-Eurasia plate boundary. Almost all
(>90%) plate boundary motion in north-central Turkey is accommodated by slip on the
NAF [McClusky et al., 2000; Hubert-Ferrari et al., 2002; Reilinger et al., 2006]. Thus,
there are few other moderate- to high-slip rate faults in this part of the plate boundary.
Specifically, in the region near Eksik, only the poorly known contractional structures
responsible for uplift of the Pontide mountains might have slip rates of more than a few
mm/year (see data in Hubert-Ferrari et al. [2002] and Reilinger et al. [2006]). In such
simple settings, where plate-boundary motion is dominated by slip on a single fault,
the stress evolution on the master fault is likely to be much simpler than in structurally
complicated settings characterized by numerous, moderate- to high-slip rate faults (e.
g., southern California). Not only will each of these faults generate earthquakes of their
own, complicating the stress evolution of the master fault, but if more than one fault
is oriented such that both can accommodate significant plate-boundary motion, then
they may alternate in terms of relative importance such that their loading rates vary
through time [e. g., Dolan et al., 2007]. In contrast, in simple, single-fault-dominated
settings such as the plate boundary in northern Turkey, I suggest that storage of elastic
strain will be dominated by quasi-continuous tectonic loading, and in general, the only
earthquakes that will complicate the stress evolution of the master fault will be those
produced by rupture of other parts of the master fault. It is worth noting, however, that
the
36
Cl terrace age based on the Stone et al. [1996] and Evans et al. [1997] production
20
rates leaves open the possibility that the average late Holocene slip rate for the NAF
is slightly slower than the short-term rate of elastic strain accumulation measured by
geodesy [McClusky et al., 2000; Reilinger et al., 2006], possibly as a result of time-
dependent changes in fault zone loading in response to the recent occurrence of the
1939–1999 earthquake sequence.
Displacement on the Eksik reach of the NAF was 4–4.5 m during the 1943
M
w
7.7 Tosya earthquake [Barka, 1996]. The slip rate data described above, however,
in combination with paleoseismologic data from a nearby trench site [Sugai et al.,
1998], indicate that these relatively small displacements probably are not characteristic
of earthquakes on this reach of the NAF. Specifically, Sugai et al. [1998] observed
evidence for only five surface ruptures during the past 2000 yr at their trench site in
Alıç (~15 km west of Eksik; see Figure DR-1). Although it is possible that events may
be missing from the paleoseismologic record, the Sugai et al. [1998] trench was both
carefully situated and characterized by markedly continuous sediment accumulation in
a marshy, peat-rich, active alluvial fan environment, making this unlikely. Similarly, a
paleoseismologic site located 160 km to the east of Eksik also reveals only five surface
ruptures since 2 ka [Hartleb et al., 2003]. If all of the 46 ± 10 m offset of the terrace
inner edge measured at Eksik accumulated in earthquakes that exhibited “characteristic
slip” [Schwartz and Coppersmith, 1984] of 4–4.5 m/earthquake, then eight to fourteen
surface ruptures would have had to occur since abandonment of the terrace surface at
~2.1-2.4 ka. Thus, either displacements in the 1943 earthquake were much smaller than
is typical for this reach of the NAF, or a cluster of three to eight earthquakes must have
occurred during the 100-400 years preceding the paleoseismologic record of Sugai et
al. [1998] at the nearby Alıç trench site.
21
CHAPTER 3:
A Late Holocene slip rate for the central North Anatolian fault,
at Tahtaköprü, Turkey, from cosmogenic
10
Be geochronology:
Implications for fault loading and strain release rates
Abstract
Measurement of a stream offset and cosmogenic dating (
10
Be) of the alluvial
surface into which the stream incised yield a preferred late Holocene slip rate of 18.6
+3.5/-3.3 mm yr
-1
for the central part of the North Anatolian fault at Tahtaköprü, Turkey;
use of variable cosmogenic production rate (VPR) models yields a slightly slower rate
of ~16.4 +6.4/-4.5 mm yr
-1
. The offset drainage (Karanlık Dere), which flows southward
almost perpendicular to the east-west trace of the NAF at the site, is displaced right-
laterally by 55 + 10 m, with no vertical displacement. A
10
Be age of ~3 ka (~3.5 ka
VPR) from the top of a boulder on the best-preserved part of the incised alluvial surface
provides the most reliable maximum age for the onset of incision; eleven
10
Be ages
from cobbles collected from the cultivated surface to the south yield younger ages,
consistent with exhumation, erosion and mechanical mixing during plowing. My 18.6
+3.5/-3.3 mm yr
-1
rate is similar to other geologic slip rates measured along the NAF,
all of which cluster between 15-20 mm yr
-1
over a wide range of (10
3
-10
5
) time scales.
All of these geological rates, however, are slower than the ~25 + 2 mm yr
-1
short-term
rate of elastic strain accumulation measured geodetically. This disparity suggests the
possibility that the NAF is experiencing a strain transient in which the lower crust
beneath the fault is deforming faster than its long-term rate.
22
3.1 Introduction
The degree to which strain accumulation and release on major continental faults
is constant in time and space is one of the most fundamental, unresolved problems in
modern tectonics. At time scales of dozens of earthquakes and hundreds of meters of
fault displacement, data from some faults (e.g., the central San Andreas fault) suggest
that fault slip rates are relatively constant over a wide range of time scales [e.g., Sieh and
Jahns, 1984; Weldon and Sieh, 1985]. In contrast, there is increasing paleoseismological
and historical evidence from a number of different regions that at the time scale of
individual earthquakes strain release is not necessarily constant in time and space [e.
g., Ambraseys, 1970; 1989; Sieh et al., 1989; Fumal et al., 1993; 2002; Grant, 1996;
Dolan and Wald, 1998; Mann et al., 1998; Rockwell et al., 2000, Hubert-Ferrari et
al., 2002; Weldon et al., 2002; Dolan et al., 2007]. These observations highlight a
number of critically important questions about the mechanisms by which faults store
and release strain energy: (1) Are fault slip rates constant over all but the shortest time
scales (i.e., one to a few earthquake cycles), as would be expected if faults are loaded at
a steady rate by plate tectonics? (2) Or are slip rates temporally variable at intermediate
time scales of a few to a few dozen earthquakes, as might be expected if loading were
dominated by transient phenomena (e.g., ductile creep events on lower crustal faults,
or modulation of lower crustal loading rates by regional earthquake clusters)? (3) Over
what time scales do these different processes operate? (4) And finally, do different
faults and regional fault networks behave differently at different time scales?
In this study I document a late Holocene slip rate for the central part of the
North Anatolian fault (NAF), the 1500-km-long, structurally simple, dextral transform
that accommodates >90% of relative motion between the Anatolian Block and Eurasian
Plate [McClusky et al., 2000; Hubert-Ferrari et al., 2002; Reilinger et al., 2006]. I then
23
discuss the implications of my and other rate data from the NAF for understanding the
behavior of continental strike-slip faults.
3.2 Tahtaköprü Study Site
3.2.1 Site description
My study site is located along the central part of the North Anatolian fault (NAF)
between the villages of Sarıdibek and Tahtaköprü, Turkey, ~225 Km ENE of Ankara
(Figure 3.1). In this area, the fault is characterized by a ~N70 °E trace that extends
parallel to, and just north of, a major east-flowing river (Koca Çay), and along the base
of the south-facing slope of a large mountain (Kunduz Dağı, 1791 m). An extensive,
gently south-sloping, planar alluvial surface north of Koca Çay has been incised by a
number of Koca Çay tributaries that flow southward across, and nearly perpendicular
to the fault, which at the study site is expressed as a single, narrow, geomorphically
well-defined strand. Several of these south-flowing tributaries have been offset by right-
lateral slip on the NAF, as first noted by Allen [1969]. My study focuses on the offset of
Karanlık Dere (Dark Creek), one of these south-flowing streams (Figure 3.1b).
Karanlık Dere is the more westerly of two drainages that flow southward
from headwater cliffs ~1 km north of the fault. The cliffs form a broad, one-km-wide,
concave-to-the-south, amphitheater-shaped drainage basin (Figure 3.2). Both of these
south-flowing drainages have incised into the planar surface of a laterally extensive
alluvial deposit that partially infilled the drainage basin during an earlier aggradational
phase. Currently, Karanlık Dere appears to carry very little bed-load, in contrast to the
unnamed, south-flowing drainage ~150 m to the east, which carries extensive amounts
of cobble to small boulder bed load that is being actively deposited in a small alluvial
24
Figure 3.1 (a) Location of Tahtaköprü study site (star) along the North Anatolian fault
(NAF) in north-central Turkey. (b) Corona satellite photo of the study site overlaid with
10-meter topographic contours. The extent of the incised alluvial deposit that forms
the focus of this study is highlighted in pale yellow. KD is Karanlık Dere (Dere means
“creek” in Turkish); F2 is Vista point of Figure 3.2, F3 is extent of detailed site map in
Figure 3.3.
25
Figure 3.2 Panoramic view of the Tahtaköprü slip-rate site looking northward across the
North Anatolian fault (NAF) (red) and the incised and offset alluvial deposit discussed
in the paper. Karanlık Dere is the western drainage shown in blue (at left). Note active
alluvial fan deposition at and south of fault crossing of an unnamed eastern drainage at
lower right. Black dashed line indicates the combined drainage area of Karanlık Dere
and unnamed eastern drainage (right).
26
fan that is inset by 2 m into the alluvial surface (Figure 3.2).
3.2.2 Offset measurement of Karanlık Dere
At my study site, Karanlık Dere has incised a narrow, ~5-m-deep canyon into
the alluvial deposit. The right-lateral offset of the Karanlık Dere canyon by the NAF is
unusually clear; the incised channel flows southward almost perpendicular to the fault
and is highly linear >200m to the north and for ~150 m to the south of the fault crossing
(Figure 3.2). The canyon walls are quite steep, and the riser that separates the incised,
older alluvial fan surface from the active drainage typically slopes at ~60-70 °, with
locally near-vertical walls. I surveyed the displacement of the edges of the top of the
Karanlık Dere terrace riser using a laser range-finding total station. By incrementally
restoring the resulting map of the offset riser by varying amounts of fault offset, I can
accurately estimate the offset of Karanlık Dere by the NAF, as well as the range of
potential uncertainties in my offset measurement.
As can be seen in figure 3.3, all restorations of offsets between 50 m and 60
m yield equally plausible restored channel geometries. The top of the western riser
of the active drainage is both highly linear and well-preserved south of the fault, in
contrast to the top of the eastern riser, which has been eroded for a distance of 10-15 m
southward from the fault along the fault-parallel reach of the offset channel. My 55 m
reconstruction yields the straightest-possible configuration for the western riser, while
also yielding a plausible geometry for the eastern riser. Restorations of offsets of up to
60 m and as small as 50 m also yield reasonable channel geometries. I therefore take
the average of the range of these most-plausible geometries, and suggest a preferred
offset of 55 m, recognizing that this is just a preferred central value within an acceptable
range, rather than a single, precise offset value.
27
Figure 3.3 (a) Detailed map of the offset Karanlık Dere terrace riser based on our
total station surveys. The test pit for our attempted depth profile is located 5 m east
of the boulder sample location denoted by the yellow dot. Tree locations (green) are
approximate. (b-g) Incremental restorations at 5-meter-intervals of the offset channel
ranging from 45 m to 70 m. Note that the restorations between 45 and 65 m show
plausible channel configurations. Channel offset restorations between 50 m and 60 m
yield the best alignments of the western and the eastern channel risers, and this range
represents our preferred offset value.
28
I can also use these sequential offsets to estimate potential errors. Whereas
restoration of 50 m of offset yields a plausible configuration for the eastern margin
of the drainage (Figure 3.3c). Restoration of 45 m offset requires a sharp downstream
bend to the west of ~5 m at the fault that I consider to be sedimentologically unlikely.I
therefore use 45 m as my minimum offset, as this restoration yields a permissible,
but somewhat unlikely, channel geometry, whereas smaller restorations become
increasingly less plausible. Restorations of offsets >65 m yield increasingly less-
plausible channel configurations. Although larger offsets are possible, restoration of
65 m yields an unlikely pre-offset configuration for the stream, with a very sharp ~8 m
downstream bend to the east of the well-preserved western terrace riser, and a highly
nonlinear configuration to the eastern riser. Restoration of 70 m completely blocks
the drainage (Figure 3.3g). I therefore use 65 m as my maximum-possible offset of
the incised Karanlık Dere drainage. Thus, my preferred estimate of the offset of the
Karanlık Dere canyon is 55 + 10 m, with all offsets between 50 and 60 meters being
considered equally plausible.
3.2.3 Detailed geomorphology of the fan surface
The extensive alluvial surface into which Karanlık Dere has incised is quite
planar, with a relatively constant southward slope of ~10 ° near the fault crossing;
surface slope steepens upstream, within the amphitheater-shaped headwaters drainage
basin. This originally planar surface has, however, long been cultivated, somewhat
complicating the interpretation of the geomorphology and geochronology of the site.
My samples for cosmogenic radionuclide dating (discussed below) were collected from
the alluvial surface adjacent to and just south of the fault crossing along the eastern
side of the Karanlık Dere (Figure 3a). In addition to being plowed, the surface south
29
Figure 3.4 A simple model to explain the effects of long-term cultivation on an alluvial
surface. (a and b) The initial surface (immediately post-abandonment) would have
some quartz cobbles at the surface and at various depths. Samples on the surface
(black) would have higher 10Be nuclide concentrations than samples at depth (gray-
white) as a result of exponential cosmogenic nuclide production rate decrease with
depth. At this pre-cultivation stage, climate-dependent erosion (rain and wind) is the
only physical factor that shapes the surface by removing the finer materials. As a result,
the surface deflates from S0 (initial surface) to S1 (slightly eroded, but uncultivated
surface). (c-f) Erosion is enhanced by the inception of cultivation. Annual plowing
(usually to a depth of ~5-30-cm) weakens the surface soil, making it easier to strip
finer clasts (clays, silts, and sands). In addition to enhancing erosion, plowing will
result in mechanical mixing of clasts from different depths within the plow zone. (g)
Eventually, a mixture of cobbles that have been exhumed from different depths will be
concentrated at the surface. These will yield a set of variable ages (most younger than
the actual stabilization age of the surface) due to their different exposure histories.
30
of the fault has also been slightly terraced by farmers. I estimate that locally, the uphill
sides of some fields south of the fault have been lowered by as much as 50 cm during
terracing and cultivation-associated erosion. A thin strip of apparently original alluvial
surface extends along the top of the very steep riser above the modern drainage along
the east-west, fault-parallel section of Karanlık Dere. The 1-2 m-wide-strip at the top
of the cliff does not appear to ever have been cultivated, presumably because it lies too
close to the cliff edge.
The surface of the incised alluvial fan has been cultivated for centuries, if not
millennia. The modern plow zone within the field to the south extends to ~20-25 cm
depth, resulting in more abundant cobbles at the surface than within the soil profile. In
figure 3.4 I propose a model for the development and evolution of the surface in which
mechanical disturbance during plowing, and associated cultivation-related erosion,
will result in a concentration of large clasts at the surface, as well as mixing of clasts
exhumed from different depths within the original alluvial deposit. Annual plowing
will result in a mechanically weaker surface soil that is more susceptible to erosion,
particularly of wind-blown fines. In contrast, cobbles that are large enough to not be
eroded by wind or surface water flow, but small enough to avoid being cleared from
the field by the farmers, will be concentrated at the surface over time. In addition,
plowing will tend to move cobbles within the plow zone towards the surface, also
resulting in a concentration of large grains at the ground surface. The model shown in
figure 3.4 would thus explain my observations that large clasts are scarce within the
depth profile at my study site, whereas there is an abundance of pebble- to cobble-sized
clasts, consisting mainly of white vein quartz, at the surface. As discussed below, these
effects of long-term cultivation and associated erosion strongly affect my interpretation
of the cosmogenic exposure age data collected from the incised alluvial surface.
31
Sample name Latitude (degrees) Longitude (degrees) Elevation (m) Thickness (cm) Shielding correction Laboratory [
10
Be] (10
3
atoms/g) TAH Ͳ 01 Ͳ 06 41.10059132 35.04921684 973 4 1 LLNL 31.6 + 1.5 TAH Ͳ 02 Ͳ 06 41.10055447 35.04926332 971 4 1 LLNL 5.9 + 0.6 TAH Ͳ 03 Ͳ 06 41.10076025 35.04934117 974 4 1 LLNL 12.7 + 0.7 TAH Ͳ 05 Ͳ 06 41.10091371 35.04932212 973 4 1 LLNL 13.3 + 0.6 TAH Ͳ 12 Ͳ 06 41.10059778 35.04935993 970 4 1 LLNL 25.7 + 1.2 TAH Ͳ 13 Ͳ 07 41.10057956 35.04937127 969 4 1 LLNL 28.0 + 1.3 TAH Ͳ 16 Ͳ 06 41.10072233 35.04944715 970 4 1 LLNL 22.5 + 1.0 TAH Ͳ 17 Ͳ 06 41.10075018 35.04940039 969 4 1 LLNL 17.3 + 0.8 TAH Ͳ 18 Ͳ 06 41.1008025 35.04949728 970 4 1 LLNL 17.2 + 0.8 TAH Ͳ 01 Ͳ 05 41.10076088 35.04934684 969 4 1 LLNL 11.9 + 0.8 TAH Ͳ 06 Ͳ 05 41.10058712 35.04938249 974 4 1 LLNL 24.9 + 1.9 TAH Ͳ 07 Ͳ 05 41.10070869 35.04929481 971 4 1 LLNL 9.8 + 0.6 TAH Ͳ 22 Ͳ 05 41.100 35.049 974 4 1 LLNL 23.0+ 0.9 1 Table 3.1 Input values used for calculating 10Be surface exposure ages in the CRONUS online
10
Be-
26
Al age calculator
[http://hess.ess.washington.edu/math/]. Uncertainties associated with the number of 10Be atoms/g are 1sigma analytical
uncertainties in the accelerator mass spectrometry measurements.
32
Sample ID Age Err Age Err Age Err Age Err Age Err
TAHͲ01Ͳ06 3030 147 3479 468 3533 474 3428 395 3149 324
TAHͲ02Ͳ06 574 66 657 111 636 107 647 100 609 89
TAHͲ03Ͳ06 1219 69 1373 188 1369 186 1350 159 1271 135
TAHͲ05Ͳ06 1281 66 1443 194 1442 193 1419 164 1334 138
TAHͲ12Ͳ06 2472 120 2852 383 2936 393 2808 322 2575 264
TAHͲ13Ͳ06 2698 125 3112 415 3182 423 3065 349 2809 286
TAHͲ16Ͳ06 2170 105 2499 335 2578 344 2459 282 2259 231
TAHͲ17Ͳ06 1665 79 1896 253 1932 257 1863 212 1731 176
TAHͲ18Ͳ06 1655 82 1884 252 1919 256 1851 212 1720 177
TAHͲ01Ͳ05 1139 85 1284 186 1275 184 1261 161 1190 139
TAHͲ06Ͳ05 2383 189 2748 407 2831 419 2705 354 2483 299
TAHͲ07Ͳ05 943 63 1063 150 1044 147 1045 128 991 111
TAHͲ22Ͳ05 2203 90 2538 333 2618 343 2498 279 2294 227
Table 1
CONSTANT VARIABLE
COSMOGENIC NUCLIDEPRODUCTIONRATES
Desilets etal ,(2003, 2006) Dunai, (2001) Lifton etal ,(2005) Lal (1991)/Stone (2000) Lal (1991)/Stone (2000)
Table 3.2 Terrestrial cosmogenic nuclide (
10
Be) dating results. The ages are calculated using the online cosmogenic age
calculator CRONUS [http://hess.ess.washington.edu/math/]. Note that all ages are reported for several different production
rates including both constant-production rate and time-variable production rate models [see Balco et al., 2008]. Errors on
ages reflect only 1sigma analytical uncertainties in the accelerator mass spectrometry measurements. Errors reported for
constant-production-rate model are internal uncertainties, whereas errors reported for variable-production-rate models are
external uncertainties.
33
3.3 Age Control
I used cosmogenic radionuclide (
10
Be) dating of surface samples to determine
the age of the alluvial surface that has been incised by Karanlık Dere. The incised
alluvial deposit consists of sediments eroded from lower and middle Triassic schist
exposed in cliffs ~1 km north of the study area (Figures 3.1 and 3.2). I dated a total of
thirteen samples collected from the alluvial surface. This data set comprises eleven,
5- to 10-cm-diameter (~4 cm thick) quartz cobbles, one in situ quartz vein from the
flat, apparently uneroded top of one of the rare boulders contained within the alluvial
deposit, and one amalgamation sample (TAH-22-05) composed of 14 surface quartz
cobbles combined into one date. The boulder from which I sampled the in situ quartz
vein is located within the well-preserved and apparently never-cultivated strip of
alluvial surface adjacent to the Karanlık Dere canyon edge, whereas the cobbles were
sampled from the cultivated field immediately to the south (Figure 3.3).
All samples were prepared according to the procedures Kohl and Nishiizumi
[1992], and terrestrial cosmogenic nuclide concentrations were measured at the
Lawrence Livermore National Laboratory Center for Accelerator Mass Spectrometry. I
used CRONUS online
10
Be-
26
Al age calculator [http://hess.ess.washington.edu/math/]
to determine sample ages (Table 3.1). Following the recommendations of Balco et al.
[2008], the uncertainties in all ages reported in the following discussion reflect what
are known as “internal errors”. That is, errors from within a group of samples that
are similar to well-dated reference sites in terms of their age, latitude, elevation, and
shielding. External errors reflect comparison to a global data set.
There are two types of production-rate models that can be used to calculate
the ages of my
10
Be samples [Balco et al., 2008]. The first type assumes a constant
34
cosmogenic nuclide production-rate through time [Lal, 1991; Stone, 2000]. These
constant-production-rate (CPR) models yield ages for my 13 samples that range from
0.6-3 ka (Table 3.2). The second type of model attempts to account for temporally
variable production rates caused by variations in magnetic field intensity [ Desilets et
al., 2003 and 2006; Dunai, 2001; Lifton et al., 2005; Lal, 1991 and Stone, 2000]. Of
the four main variable-production-rate (VPR) models, three [Desilets et al., 2003 and
2006; Dunai, 2001; Lifton et al., 2005] yield very similar ages for my samples of (0.6-
3.5 ka), approximately 15% older than constant-production-rate estimates. A fourth
variable-production-rate model [Lal, 1991; Stone, 2000], yields younger ages of (0.6-
3.1 ka) that are similar to the constant-production-rate ages.
Numerous factors such as erosion, inheritance, altitude, shielding, and latitude
must be considered when interpreting the surface exposure age of a sample. My
field observations at the Tahtaköprü site, however, suggest that mechanical mixing
of samples from different depths during plowing, minor removal of surface material
during terracing, and cultivation-related erosion are the main external processes that
affect the ages of my samples (Figure 3.4). Scaling for the altitude and latitude of
the site were made in the CRONUS age calculator [Balco et al., 2008]. The shielding
(skyline) correction factor, on the other hand, is unnecessary at my study site since
the skyline lies at less than 20 ° in all directions. The final factor that could affect
the age of my samples is inheritance. However, given the very short total transport
distance and the relatively steep stream profile in the upstream reach of Karanlık Dere,
I suspect that inheritance in my samples is likely to be small. Specifically, the current
headwaters of Karanlık Dere are only 900 meters north of the fault crossing and the
headwaters of the eastern drainage are only 1.3 km north of the fault. The stream drops
340 meters in the ~1 km distance from headwaters to the fault crossing, resulting in
35
very high stream power during winter rains (which routinely wash away the main road
between Sarıdibek and Tahtaköprü just south of my study site). This situation, with a
short headwaters-to-study-site transport distance via high-energy mountain streams, is
similar to the situation I encountered at my Eksik slip-rate study site, located ~200 km
to the west [Kozacı et al, 2007]. My cosmogenic and radiocarbon age dates from the
Eksik site of an alluvial-fluvial surface that is quite similar to the one that I studied at
the Karanlık Dere site showed only minor inheritance, on the order of no more than a
few hundred years, in similar-aged deposits.
One potential means of checking for inheritance is to sample clasts from the
modern drainage. This could not be done at Karanlık Dere, however, because the
relatively few cobbles observed in the modern stream have all almost certainly been
introduced by farmers throwing and rolling the larger clasts into the drainage while
clearing their fields. Moreover, the paucity of clasts in the active drainage of Karanlık
Dere, relative to the alluvial fan deposits into which the stream is incised, may be
related to capture of the headwaters of the eastern drainage by the currently active
lower reach of the eastern drainage. This inference is supported by the fact that the only
obvious deposition of coarse-grained bed-load (cobbles and small boulders) within
this system lies south of the fault crossing along the eastern drainage in the form of the
small, rapidly aggrading alluvial fan visible in figure 3.2.
Another means of checking for inheritance is by dating a series of samples
within a depth profile excavated into the surface. I excavated a 2-m-deep pit along
the northern edge of the alluvial surface, but I are unable to retrieve sufficient quartz
samples to allow determination of an accurate depth profile.
The ~0.6-3 ka (CPR) age range described above includes all thirteen dated
samples from the incised alluvial surface. However, the relative cumulative probability
36
Figure 3.5 Plot of probability distribution functions of all 13
10
Be surface exposure
sample ages collected from the incised alluvial surface at Tahtaköprü, based on the
constant-production-rate model of Lal [1991] and Stone [2000]; variable-production-
rate models yield slightly older ages, which in turn yield slightly slower slip rates.
Simple Gaussian probability distribution functions (black) were created in Matlab
using Camelplot [J. Balco, pers. comm., 2008]. Probability values are normalized so
that each individual probability distribution function has a unit area. Shaded areas show
the sum of the probability distributions for all samples. The red probability distribution
function shows the age of the lone boulder sample, which was collected from the best-
preserved part of the alluvial surface. This age represents our best estimate of the age of
the incised surface. The dark blue-shaded area of the cumulative probability plot shows
the age ranges of the next-youngest set of samples, consisting of cobbles collected from
just south of the boulder sample from a cultivated and slightly terraced field.
37
plot for these thirteen ages shown in figure 3.5 reveals that there are several distinct
groupings of ages, with an older cluster of six ages, and several clusters of much-
younger ages. The younger clusters of samples are irregularly distributed from 0.6-1.8
ka. Given the nature of my study site, I interpret this complicated pattern of young
ages as reflecting a combination of mechanical mixing of samples from various depths
during plowing, as well as lowering of the surface through cultivation-enhanced erosion
(Figure 3.4), both processes which tend to produce exposure ages which underestimate
the true age of the surface.
I therefore interpret the relatively tight cluster of the six oldest samples as being
generally more reflective of the true age of the incised alluvial surface. In particular,
the oldest sample (TAH-01-06) has several attributes that lead me to suspect that
its ~3 ka (CPR; 3.5 ka [VPR]) age most closely reflects the age of abandonment of
the alluvial surface. Specifically, this sample was collected from a several-cm-thick,
horizontal quartz vein from the flat top of a ~60 cm by ~40 cm boulder embedded in
the alluvial surface (the flat top of the boulder lies only a few centimeters above the
ground surface). Moreover, as described above, this boulder is located within the 2-m-
wide strip of alluvial surface adjacent to the cliff at the top of the stream canyon that I
do not think was ever cultivated, presumably due to its proximity to the cliff edge. The
five cobble samples (TAH-12-06, TAH-13-06, TAH-16-06, TAH-06-05 and TAH-22-
05) that comprise the slightly younger (2.2-2.7 ka) cluster of ages were all collected
from the cultivated field one to ten meters south of the boulder sample. In addition to
the mechanical mixing of samples from different depths during plowing, this field has
also been subject to cultivation-enhanced erosion and minor terracing. Interpolation
in the field of an approximate “envelope” tied to the best-preserved sections of the
pre-agricultural landscape (e.g., the apparently original strip of the alluvial surface
38
containing the boulder sample, and the southern edges of the terraced, 20-30-m-
wide fields) suggested to me that the original surface has been lowered by several
tens of centimeters, with the greatest lowering occurring near the northern edges of
the south-sloping fields. I therefore interpret the cluster of five ages at 2.2-2.7 ka as
reflecting primarily minor erosion (25-50 cm) and lowering of the alluvial surface
during cultivation. The fact that these ages are only a few hundred years younger than
the 3 ka (CPR) boulder age is consistent with erosion of a few tens of cm of material.
Specifically, if I assume that all of these samples were actually deposited at 3 ka, I can
use the CRONUS age calculator to back out the amount of erosion that would have
produced the cluster of 2.2-2.7 ka samples. The resulting amounts of erosion range
from 14 to 42 cm, consistent with my field estimates of the total amount of lowering of
the alluvial surface during cultivation.
The ages reported above are based on constant-production-rate (CPR) models.
Variable-production-rate (VPR) models yield slightly older ages for the same samples
[Balco et al., 2008]. For example, application of the three similar variable-production-
rate models [Desilets et al., 2006; Dunai, 2001; Lifton et al., 2005] yields an age of
~3.5 + 0.5 ka age for the boulder-top quartz vein (TAH-01-06) and a range from 2.5 to
3.2 ka for the cluster of five slightly younger cobble ages (Table 3.2).
3.4 Discussion
As with all such studies, two pieces of information are necessary to determine
the slip rate of the NAF at my Tahtaköprü study site: (1) the displacement of the piercing
line, in this case the top of the Karanlık Dere terrace riser (river bank) that has incised
into an abandoned alluvial surface; and (2) the age of the stabilization of the surface,
39
which provides a maximum-possible age for the incision and subsequent offset of the
stream. My best estimate of the offset of the incised Karanlık Dere canyon is 55 +
10 m, with an equal probability given to all offsets between 50 m and 60 m, because
offsets in this range yield the most sedimentologically plausible reconstructions of the
offset channel. In order to provide a conservative estimate of the total offset of Karanlık
Dere, I follow the methodology established by McGill et al. [2008], and include error
estimates of +5 m on either side of this preferred range of offsets, with the likelihood
of these offsets decreasing to zero at the maximum (65 m) and minimum offsets (45
m). This results in a trapezoidal probability distribution function for the offset, with a
central “boxcar” distribution from 50 to 60 m, flanked on either side by probabilities
that ramp down to zero at 45 and 65 m (Figure 3.6a).
In contrast to the relatively straightforward measurement of the offset, the age
of the incised surface is more complicated. As discussed above, my preferred 3-3.5
ka age for stabilization of the incised alluvial surface comes from the lone boulder
sample that I collected from the least-disturbed part of the alluvial surface; I interpret
all younger ages as reflecting a combination of cultivation-induced erosion and plow-
zone mixing of the cultivated surface from which the samples were collected.
Using the methodology of Bird, [2007], as modified by J. Zechar [unpublished
data], I combine a simple Gaussian probability distribution function for my preferred 3
ka age (CPR) from the boulder sample with the weighted, 55 + 10 m offset to obtain a
slip rate of 18.6 +3.5/-3.3 mm yr
-1
(Figure 3.6); use of the 3.5 ka age based on variable-
production rate models yields a rate of 16.4 +6.4/-4.5 mm yr
-1
. These preferred ages and
resulting slip rates are based on the assumptions that: (1) the boulder age, which is the
oldest of my thirteen
10
Be surface dates, most accurately reflects the stabilization of the
incised alluvial surface; and (2) inheritance is negligible in this short-transport, high-
40
Figure 3.6 Probability distribution functions (pdf) used to calculate our North Anatolian fault slip rate. (a) Simple Gaussian
probability distribution function of the preferred age of the boulder sample from the best-preserved part of the incised alluvial
surface (based on constant production rate; see Table 3.3 and online data repository for variable-production-rate age). (b)
Probability distribution function of our preferred offset, with all offsets between 50 and 60 m being considered equally
plausible, and + 10 m total error estimates (see text for discussion). (c) Probability distribution function (95% confidence
limits) of our preferred 18.6 +3.5/-3.3 mm yr-1 (CPR) slip rate for the NAF based on (a) and (b). Use of the VPR models for
the age of the incised surface yield a preferred slip rate of 16.4 +6.4/-4.5 mm/yr (see text).
41
Max Preferred Min Max Preferred Min Max Preferred Min
Age 2818 2965 3112 2941 3415 3889 2760 3084 3408
Displacement 65 55.0 45 65 55.0 45 65 55.0 45
Slip Ͳrate 22.1 18.6 15.3 22.8 16.4 11.9 23.4 18.0 13.8
Max Preferred Min Max Preferred Min Max Preferred Min
Age 2000 2428 3112 2112 2816 3942 1963 2530 3408
Displacement 65 55.0 45 65 55.0 45 65 55.0 45
Slip Ͳrate 32.5 22.7 14.5 30.8 19.5 11.4 33.1 21.7 13.2
COSMOGENIC NUCLIDEPRODUCTIONRATES
CONSTANT TIMEVARIABLE
Oldestdate (preferred)
Oldest6
Lal/Stone Desilet, Dunai, Lifton Lal/Stone
Table 3.3 Age ranges and resulting slip rates for all samples, including those that
appear to be much younger than the actual age of the incised alluvial surface due to
the effects of cultivation and cultivation-enhanced erosion. The wide age ranges and
consequent wide ranges in slip-rate estimates clearly demonstrate the need for caution
when interpreting cosmogenic radionuclide (CRN) results from cultivated surfaces.
42
energy fluvial-alluvial system. This latter assumption is at least partially supported by
the fact that the cluster of five oldest cobble samples collected from the cultivated and
eroded field just to the south of the boulder sample are: (A) all dated to within a few
hundred years of each other; and (B) only slightly younger than the boulder sample, as
would be expected from samples collected from a slightly eroded surface that has been
subjected to mechanical mixing within the plow zone during cultivation. For the sake
of completeness, in addition to my preferred rates, in table 3.3 I show the slip rates that
would result from the use of various combinations of the cluster of all six oldest ages.
I note, however, that if inheritance is minor in these cobble samples, the observation
that there has been at least a few tens of cm of lowering of the cultivated surface from
which the samples were collected requires that the true age of surface stabilization be
older than the age of the surface cobbles, consistent with my interpretation that my
oldest
10
Be age from the boulder top most closely reflects the age of stabilization and
abandonment of the surface.
My preferred 18.6 +3.5/-3.3 mm yr
-1
constant production rate (CPR) and 16.4
+6.4/-4.5 mm yr
-1
variable production rate (VPR) slip rates are generally similar to
other geological slip rates that have been measured for the NAF over a wide range
of time scales (1 to 75 ka) (Figures 3.7 and 3.8). Specifically, preferred rates from
a number of studies based on a variety of techniques all cluster at ~15-20 mm yr
-1
[Kondo et al., 2004; Kozacı et al., 2007; Hubert-Ferrari et al., 2002; Pucci et al.;
2007]. The data shown in figures 3.7 and 3.8 suggest that the slip rate for the NAF
has been relatively constant during late Pleistocene and Holocene time when averaged
over time scales of more than a few earthquake cycles, reflecting the relative structural
simplicity of both the NAF and the Anatolian-Eurasia plate boundary in northern
Turkey [Kozacı et al., 2007]. It is important to point out, however, that for various
43
reasons many of the geochronologically constrained rates shown in figures 3.7 and 3.8
may be minima. For example, the Pucci et al. [2007] rates, which are based on OSL
and radiocarbon-dating of 100 to 900 m offsets along the section of the NAF that most
recently ruptured in the 1999 M
w
7.1 Düzce earthquake, record the slip rate on only the
northern strand of the NAF, and do not include slip along the southern strand, which
ruptured during the 1957 M
s
= 7.0 Abant earthquake [McKenzie, 1972] and the 1967
M
w
7.0 Mudurnu Valley earthquakes [Stein et al., 1997]; the slip rate of the southern,
Mudurnu Valley strand of the NAF is currently unconstrained. In other instances, in
which the NAF is characterized by a single strand, some slip rates [e.g. Kondo et al.,
2004] may underestimate the total slip rate because they are based on measurements
of slip in 3D trench excavations over a very small fault-perpendicular distance (<5
m), and thus may not account for any possible near-fault distributed strain, such as
that observed after the 1999 M
w
7.5 İzmit rupture [Rockwell et al., 2002]. Some of
the slip rates, however, including the Kozaci et al. [2007] slip rate at Eksik, and this
study, may more fully account for displacement across the entire fault zone because
they are based on wide-aperture geomorphic measurements. Nevertheless, even these
rates may underestimate the total slip rate if there were long elapsed time intervals
between incision and subsequent initiation of fault offset. As noted below, however,
and by Kozaci et al. [2007], the relative temporal regularity of earthquake occurrence
on the NAF suggests that this effect may be relatively small. All geochronologically
constrained fault slip rates measured for the NAF over 10
3
– 10
5
yr time scales, including
my rate at Tahtaköprü, are similar to one another at ~15-20 mm yr
-1
and all of these rates
are slower than the 25 + 2 mm yr
-1
rate of elastic strain accumulation along the NAF
measured by GPS geodesy [McClusky et al., 2000; Reilinger et al., 2006] (Figures
3.7 and 3.8). This apparent discrepancy between short-term geodetic and longer-term
44
Figure 3.7 Compilation of <75-ka slip rates along the NAF that are based on well-determined offsets and geochronologically
constrained ages, sorted by time scale of measurement. The vertical error bar of each datum represents the range of slip-rate
values whereas the horizontal error bar of an individual datum shows the uncertainty in age constraints. Slip-rate data are
color coded as black (
14
C), red (
36
Cl and
10
Be), and purple (OSL). The current rate of elastic strain accumulation on the NAF
measured by GPS geodesy is shown in green [Reilinger et al., 2006]. For completeness, we also show (in blue) the slip rate
estimate of Hubert-Ferrari et al., [2002], which is based on a model of climatically controlled sedimentation and incision
and geomorphic response to climate change. A-Reilinger et al. [2006]; B-Kondo et al. [2004]; C and E-Kozacı et al. [2007];
D-This study; F and H-Hubert-Ferrari et al. [2002]; G, I and J-Pucci et al. [2007]. Arrows pointing up indicate that the slip
rate is a minimum rate (see text for discussion). Geographic location of dated sites is given in Figure 3.8.
45
Figure 3.8 (a) Simplified location map for the North Anatolian fault, showing selected cities near the fault trace. (b) Extents
of 20th century earthquake surface ruptures [Barka, 1996]. (c) Plot of geodetic and geochronologically constrained geologic
slip-rates along the NAF sorted by distance along the fault. Color and letter coding is same as Figure 3.7. The green line
represents rates of elastic strain accumulations from the geodetically constrained block model of Reilinger et al. (2006);
black dots delimit the end points of fault segments used in their block model.
46
geologic rates highlights the possibility that the NAF may currently be experiencing a
strain transient. The cause of this apparent rate discrepancy might be related to: (1) the
transient effects of the 20
th
century earthquake sequence along the North Anatolian fault
(Figure 3.8); (2) the existence of other, unrecognized faults that may accommodate a
component of NAF-parallel slip of ~5 mm yr
-1
; and/or (3) an extended period of time
between either stabilization and initial incision of the alluvial surface, and/or the time
between initial incision and the beginning of fault offset of the incised drainage.
Exploring this third possibility first, I note that paleoseismologic data from the
central NAF on either side of the Tahtaköprü slip-rate site indicate the relatively regular
recurrence of large-magnitude earthquakes every few hundred (~250-623) years [Sugai
et al. 1998; Kondo et al. 2004; Okumura et al. 2003]. In the extreme limiting case in
which the stabilization and initial incision of the alluvial surface at Tahtaköprü occurred
immediately prior to an earthquake, (i.e. at the end of a strain-accumulation cycle), then
I would need to include the period of elastic strain accumulation prior to the earthquake
in my age term. If such an interval was as long as the longest paleoseismologically
defined gap (623 years), this would yield a slower slip rate of ~15.3 mm yr
-1
(55 m in
3588 years).
I note, however, that the 623-year-long inter-event time between the historical
1045 and 1668 earthquakes documented at a trench site 130 km west of Tahtaköprü
[Sugai et al., 1998] includes the interval in which paleo-surface ruptures have been
identified at other sites both to the east and west of the Tahtaköprü site. Specifically,
trenches excavated 250 km west of Tahtaköprü [Kondo et al., 2004; Okumura et al.,
2003], as well as historical and paleoseismologic evidence for a large magnitude
earthquake in 1254 along the eastern NAF [Ambraseys 1970; Ambraseys and Finkel
1995; Hartleb et al., 2006; Kozacı et al., in review] suggest the possibility that a 13-14
th
47
century event may have occurred on the central NAF, as well, but was not recorded at
the Sugai et al. [1998] site. Alternatively, if, as Okumura et al. [2003] have suggested,
earthquake occurrence along the central NAF is almost periodic, then inter-event times
at my Tahtaköprü site are unlikely to exceed ~300 years. If initial stabilization and
incision occurred just prior to an earthquake that followed a 300-year-long period
period of strain accumulation, the fastest slip rate would be ~16.8 mm yr
-1
(CPR; 55 m
in 3265 years) and ~14.7 mm yr
-1
(VPR; 55 m in 3735 years). Given that the rates that
incorporate the possibility of an earthquake occurring immediately after incision are
slower than my preferred rate, This scenario cannot be explained by the apparent rate
discrepancy. As for the possibility that significant time elapsed between abandonment
and initial incision, I consider this to be unlikely at my study site, because Karanlık
Dere is a short-transport, high-energy mountain stream that probably began to incise
soon after abandonment. This reasoning suggests that the 18.6 +3.5/-3.3 mm yr
-1
rate
from Tahtaköprü reported here, and the 20.5 + 5.5 mm yr
-1
rate from the Eksik site of
Kozacı et al. [2007] are probably close to the true rates for the structurally simple,
central part of the NAF.
Another possible explanation for the rate discrepancy is that significant strike-
slip occurs on faults parallel to the central NAF. Arguing against this possibility is the fact
that the stretch of the NAF that includes both the Tahtaköprü and Eksik slip rate sites is
single-stranded, and represents one of the structurally simplest parts of the entire NAF
system. Although other faults have been recognized in the Tahtaköprü region, most of
them for which the kinematics are understood are thought to be predominantly reverse
faults. I think it is unlikely that, even collectively, these reverse faults accommodate >5
mm yr
-1
of NAF-parallel right-lateral slip, although this possibility needs to be tested
rigorously with future field studies.
48
This leaves the possibility that the NAF may be experiencing a strain transient
in which the lower crust beneath the fault is deforming more rapidly than its long-
term average rate. The NAF is notable for the extraordinary sequence of 20
th
century
earthquakes that it has generated. Rupturing mainly from east to west, the NAF
produced large earthquakes in 1939 (M
w
7.9 Erzincan), 1942 (M
w
6.9 Erbaa-Niksar),
1943 (M
w
7.7 Tosya), 1944 (M
w
7.5 Bolu-Gerede), 1957 (M
w
6.8 Abant), 1967 (M
w
7.0
Mudurnu Valley), 1999a (M
w
7.4 İzmit), and 1999b (M
w
7.2 Düzce) [Ketin, 1948; Allen,
1969; Şengör 1979; Stein et al., 1997; Barka, 1996 and 1999]. The western part of
the NAF also ruptured during the 1912 M
w
7.4 Ganos earthquake, and possibly the
1894 earthquake [Ambraseys, 1989]. Thus, at any given point along the NAF, the time
since the most recent large-magnitude event varies from <10 years along the İzmit
and Düzce ruptures to >~70 years along the easternmost and westernmost parts of
the NAF. Interestingly, the current rate of elastic strain accumulation as determined
from a geodetically constrained block model is relatively constant along the entire
length of the NAF at ~25-28 mm yr
-1
, regardless of the elapsed time since the most
recent event [Reilinger et al., 2006]. For example, Reilinger et al. [2006] show that the
rate of elastic strain accumulation along the portion of the central NAF that includes
the Tahtaköprü site, which most recently experienced rupture more than 60 years ago
during the 1943 earthquake, is ~26 + 2 mm yr
-1
. For comparison, their block-model rate
on the northern (main) strand of the NAF beneath the Marmara Sea near İstanbul, which
probably ruptured most recently in the earthquake of 1766, is ~28 + 2 mm yr
-1
. If the
discrepancy between the short-term and long-term data is real and the NAF is currently
experiencing a strain transient, this raises the question of why this is occurring. One
possibility is that coseismic rupture of the seismogenic crust and possible post-seismic
extension of the rupture plane down into the ductile lower crust may release significant
49
fluids [e.g., Hirth and Tullis, 1992; Oskin et al., 2008]. This would have the effect of
reducing strength of the lower crust beneath the NAF, resulting in accelerated lower
crustal deformation that is observed at the surface by GPS as the currently rapid rate of
elastic strain accumulation along the NAF. Alternatively, it is possible that accelerated
deformation of the lower crust beneath major faults may trigger earthquake clusters
[e.g., Dolan et al., 2007]. Potentially both processes could occur, with feedback between
them.
3.5 Conclusions
The well-defined offset of an incised stream that crosses the central part of the
North Anatolian fault, coupled with my cosmogenic dating (
10
Be) of the surface into
which the stream incised, yields a preferred slip rate of 18.6 +3.5/-3.3 mm yr
-1
for the
central North Anatolian fault (preferred rate is 16.4 +6.4/-4.5 mm yr
-1
if I use a variable
cosmogenic nuclide production rate model [VPR]). This ~3 ka (3.5 ka; VPR) slip-rate
is similar to other (10
3
-10
5
yrs) geologic slip rates from the NAF but is slower than
the short-term 25 + 2 mm yr
-1
rate of elastic strain accumulation measured by GPS
geodesy [Reilinger, et al, 2006]. If this apparent rate discrepancy is real, it suggests
the possibility that the NAF is experiencing a strain transient, in which the lower crust
is deforming more rapidly than its long-term average rate, possibly in response to the
occurrence of the well-known sequence of large magnitude events that have ruptured
most of the NAF during the past century.
50
CHAPTER 4:
Paleoseismologic evidence for the relatively regular recurrence
of infrequent, large-magnitude earthquakes on the eastern North
Anatolian fault at Yaylabeli, Turkey
Abstract
Paleoseismologic trenches excavated across the eastern part of the North
Anatolian fault (NAF) at Yaylabeli, Turkey, provide evidence for five surface ruptures
during the last 2,000 years. I interpret these events as: (1) the historical 1939 M
w
7.9
earthquake; (2) the historical 1254 earthquake; (3) the historical 1045 earthquake; (4)
an earthquake that occurred between 660 A.D. and 1020 A.D., and probably between
710 and 850 A.D.; and (5) an earthquake that occurred between 250 A.D. and 800 A.D.,
possibly the historical 499 event. Although one of the inter-event intervals I document
is 685 years long (between the 1254 and 1939 earthquakes), the other three intervals
are between 200 and 350 years long. My results are generally similar to those from
the nearby Çukurçimen trench site [Hartleb et al., 2006], located 2 km to the east,
demonstrating reproducibility of the paleoearthquake record. However, the 8
th
to 9
th
century event (E4) that I document at Yaylabeli was not observed at Çukurçimen. The
addition of this event facilitates the recognition of a previously unnoticed NAF earthquake
cluster, during which the eastern and central parts of the fault appear to have ruptured
during a brief sequence in the 8
th
and 9
th
centuries. Addition of this possible cluster
suggests that the NAF commonly ruptures in brief, system-wide sequences, although
the individual earthquakes in each sequence differ from cluster to cluster in terms of
location, magnitude, and rupture sequence. These paleoearthquake data reinforce the
idea of relatively regular recurrence of infrequent, large-magnitude earthquakes on the
51
eastern section of the NAF. I attribute this relatively simple behavior to the structural
maturity of the North Anatolian fault and its relative isolation from other major seismic
sources within the Anatolia-Eurasia plate boundary.
4.1 Introduction
The North Anatolian fault is an arcuate, right-lateral fault that extends for 1,500
km from the Karlıova triple junction in eastern Turkey westward across northern Turkey
and into the Aegean Sea (Figure 4.1). Together with the left-lateral East Anatolian
fault, the North Anatolian fault accommodates westward motion of the Anatolian block
in response to subduction rollback along the Hellenic trench and collision of Arabia
with Eurasia [McKenzie, 1972; Şengör, 1979; McClusky et al., 2000; Reilinger et al.,
2006]. Geologic studies indicate that the slip rate on the North Anatolian fault over
the past 10
3
-10
5
-years is ~15 - 22 mm yr
-1
[Hubert-Ferrari et al., 1997; 2002; Kondo et
al., 2004; Kozacı et al., 2007; in review; Pucci et al., 2007], whereas geodetic studies
reveal somewhat faster rates of strain accumulation of ~24-29 mm yr
-1
[McClusky et al.,
2000; Reilinger et al., 2006]. Both geologic and geodetic data indicate that almost all
of the strain associated with the westward extrusion of the Anatolian block in northern
Turkey is accommodated along the North Anatolian fault system. Specifically, geodetic
measurements indicate that internal deformation rates within central Anatolia are <2
mm yr
-1
[McClusky et al., 2000; Reilinger et al., 2006]. Thus, the North Anatolian fault
is the dominant fault within a relatively simple plate boundary.
Between 1939 and 1999, a >1,000 km length of the North Anatolian fault
ruptured during a generally westward-propagating series of eight M≥7 earthquakes
[Ketin, 1948; Richter, 1958; Allen, 1969; Ambraseys, 1970; Toksöz, et al., 1979;
Barka, 1992, 1996, 1999; Ambraseys and Finkel, 1995; Stein et al., 1997; Barka,
52
Figure 4.1 Simplified tectonic map of the eastern Mediterranean region [modified after
Barka, [1992] and Armijo et al., [2002]. Arrows indicate generalized relative motions
according to the fixed Eurasian Plate showing rates in millimeters per year [after
McClusky et al., [2000] and Reilinger et al., [2006]. “Yaylabeli” in red box indicates
the location of this study. NAF-North Anatolian fault, EAF-East Anatolian fault,
NEAF-North East Anatolian fault, OF-Ovacik fault, EF-Ezinepazari fault, DSF-Dead
Sea fault. Satellite image is from Orbitview-2 satellite [available at http://visibleearth.
nasa.gov].
53
1999]. This sequence included four very large earthquakes: the 1939 M
w
7.9 Erzincan
earthquake, the 1943 M
w
7.7 Tosya earthquake, the 1944 M
w
7.5 Bolu earthquake, and
the 1999 M
w
7.5 İzmit earthquake. Stein et al. [1997] have shown that, in general, the
westward propagation of this sequence is consistent with a model of stress triggering
of each earthquake by earlier events in the sequence.
These observations raise several fundamental questions. First, does the
seemingly simple structure of the Anatolia-Eurasia boundary along the North Anatolian
fault lead to simple patterns of earthquake occurrence, as witnessed during the 20
th
century sequence? In other words, is the normal mode of earthquake occurrence on the
North Anatolian fault characterized by brief periods (decades to a century in duration)
of intense earthquake activity during which most or all of the fault ruptures? Or, was
the 20
th
century earthquake sequence an anomaly? Does the scarcity of other nearby,
major seismic sources along the central and eastern North Anatolian fault lead to either
simpler patterns of stress evolution? If so, does this simpler loading history result in a
more regular recurrence of large earthquakes?
To address these questions, I present results from a paleoseimologic study of
the North Anatolian fault from the village of Yaylabeli in north-central Turkey along
the 1939 M
w
7.9 surface trace. I use these new data, in combination with previously
published paleoseismologic results, including those of a companion study at a nearby
site [Hartleb et al., 2006] to generate an updated space-time diagram of earthquake
occurrence along the North Anatolian fault over the past 2,000 years. I then discuss
these results in light of their implications for long-term patterns of seismic strain release
on the North Anatolian fault.
54
Figure 4.2 Topographic map of Yaylabeli and Çukurçimen study sites (modified after Hartleb et al., [2006]). Rectangular
box along the fault in Yaylabeli shows location of figure 3. Hartleb et al. [2006] study site is in the large marshy area just
west of the village of Çukurçimen. Contour interval is 10 m.
55
4.2 Yaylabeli Site Description
The Yaylabeli paleoseismic trench site is located ≅65 km west of the city of
Erzincan, along the eastern part of the North Anatolian fault (Figure 4.1). The site is a
~350-meter-long, 40- to 60-meter-wide, semi-enclosed marshy basin that has formed at
a 25-m-wide releasing step in the fault (Figures 4.2 and 4.3). The basin is bounded by
a broad, ~20-m-high shutter ridge to the north, and by 300-m-high peaks to the south.
Silt, sand, and pebble gravel are transported off of the southern ridge and are deposited
as a small alluvial fan that extends northward into the marsh (Figure AP2.1). There is
a gentle down-to-the-east surface gradient at the site, allowing limited water discharge
at the east end of the marsh. In the lowest parts of the marshy depression, local areas
with standing surface water occur throughout the year, whereas the surface of the rest
of the marshy basin is damp during summer, and exhibits standing water only during
the winter rainy season.
To the east and west of the releasing step, the NAF through the site is expressed
as a single, geomorphically well-defined strand. The most recent surface rupture at
Yaylabeli occurred during the 1939 M
w
7.9 Erzincan earthquake. This earthquake was
generated by rupture of an ~400-km-long section of the eastern North Anatolian fault,
from near the city of Erzincan westward to the 10-km-wide extensional Niksar step-
over, where the rupture extended southwestward along the Ezinepazarı fault [Barka,
1996] (Figure 4.1). Surface displacements in the 1939 earthquake ranged up to ~10
m near my study site [Hartleb et al., 2006]. Although no significant scarps or mole
track remnants from the 1939 earthquake are preserved at the site, my interviews with
local eyewitnesses confirmed the presence and location of the surface rupture within
the marshy basin. Moreover, my easternmost trench was located ~10 m west of a dirt
road that was offset 9.5 + 1.5 m during the 1939 earthquake [Hartleb et al., 2006; their
56
Figure 4.3 General view of the Yaylabeli trench site, looking northeast. The North Anatolian fault extends across the marsh
that has formed between the shutter ridge to the north and mountains to the south. Exposures of the fault zone in trenches
T-2 (west) and T-3 (east) reveal a 25-m-wide right (releasing) step-over.
57
figure 5].
4.3 Trench Results
I excavated four fault-perpendicular trenches at Yaylabeli (T1-T4) (Figure 4.4).
The fault zone was exposed in all four trenches, allowing me to delineate the fault
geometry across the site (Figure AP2.2). Trenches T3 and T4, in particular, revealed
well-defined event stratigraphy, as well as numerous peat samples for radiocarbon
dating.
In order to expose the deepest-possible stratigraphic section, trenches T1, T2
and T3 were benched due to shallow ground water and resulting problems with the
stability of the trench walls. After encountering very shallow groundwater in my trench
T1 (excavated during the Summer of 2004), a ~300-m-long, fault-parallel drainage
trench was excavated to divert the ground water flowing into the basin in the beginning
of summer 2005. This drainage trench successfully reduced groundwater input from
the mountains to the south, enabling me to excavate trench T3 to a depth of 4.5 meters
below ground surface.
Trench T1 was excavated as an exploratory trench to locate the fault and
determine the width of the fault zone. I was not able to log this trench in detail due
to unstable trench walls associated with a very shallow water table (<0.5 m depth). I
did, however, locate the fault in the northern end of the trench, where it is expressed
as a <1-m-wide zone of fault gauge and multiple sub-vertical fault strands (Figure 4.5
and Figure AP2.3). I also recovered peat samples from T1 for stratigraphic correlation
purposes.
Trench T2 was excavated during the Summer of 2005, 45 m west of T1, and two
months after I dug the drainage trench. In trench T2, the fault lies within the heavily
58
Figure 4.4 Map of the Yaylabeli trench site, surveyed with a laser-range-finding total
station. Pale orange swath encompasses all identified strands of the North Anatolian
fault. Red lines show individual fault strands observed in trenches. The 0-m-contour
line corresponds to approximately 1873 m above sea level.
59
Figure 4.5 Log of west wall of Yaylabeli trench T1. The figure is mirrored for ease of comparison with other trench exposures.
The log shows the upper and lower faces of this benched exposure at correct relative elevations. The deeper face is ~1 m to
the east of the upper wall. Faults are shown as red lines, dashed where inferred. Stratigraphic units are color coded: peats are
shown in black, organic-rich clays are shown in gray, gravels and pebbles are shown in dark blue, sands are shown in blue,
and clays and silts are shown in white. This color-coding is applied uniformly to all trench logs in this paper. Peat units that
are correlative between T1 and other trenches are labeled in units of 10 (e.g., P00, P10, …). Radiocarbon sample locations
are shown as yellow squares and ages are given as calibrated calendric dates (OxCal 4.0 program [Bronk Ramsey, 1995;
Bronk Ramsey, 2001; Reimer et al., 2004]).
60
Figure 4.6 Log of the eastern wall of Yaylabeli trench T2. The lower face of the benched trench is ~1.2 m to the west of the
upper wall. Dark gray units shown towards the northern end of the log are heavily altered schist bedrock exposures. Other
stratigraphic color-coding is the same as figure 5.
61
altered schist bedrock (now a pale-gray clay) and bedrock-derived soil exposed at the
northern end of the trench (Figures 4.6, AP2.4, AP2.5, and AP2.6). Because the fault
is not located within the marsh deposits, I was not able to identify individual paleo-
earthquake event horizons, with the exception of an extensive extrusive sand sheet,
which provides evidence of a paleo-liquefaction event (Figures AP2.4 and AP2.5). As
with T1, I located the fault in T2 and collected peat samples for stratigraphic correlation.
Trench T3 was excavated 28 m to the east of T1, where the fault extends across the
active marsh. Trench T3 was a 4.5-m-deep, benched exposure with three, 1.5-m-high
faces. I was able to map the eastern wall of T3 in detail (Figures 4.7, AP2.7, and AP2.8).
However, despite my drainage trench, groundwater still entered the trench from near
the base of the west wall, causing it to cave repeatedly. Consequently, I was unable to
log the west wall of T3 in detail.
Trench T4 was excavated in the Summer of 2006, eight meters to the east of
T3. Trench T4 was excavated as a 90-cm-wide slot trench designed to encompass the
entire fault zone and provide reproducibility of the event horizons documented in T3. I
exposed the fault zone in T4 and logged both walls in detail (Figures 4.8 and AP2.9).
All trenches were logged on a 1 m x 1 m grid system, which was set arbitrarily for
each trench. Locations of specific features in the trenches are described in the following
sections according to their trench coordinates, with a horizontal meter measurement
starting at the north end of each trench, followed by a depth below an arbitrary “0”
datum, which is different for each trench. Stratigraphic units are numbered to facilitate
discussion, with numbers increasing with depth. Peat horizons that extend across the
entire trench exposure are labeled as multiples of 10 (e.g., Peat 10, 20, 30), whereas the
peat units that are limited in lateral extent are labeled as P15 and P45. Individual units
that I can confidently correlate between different trenches are labeled with the same
62
numbers in all trenches. All trench locations were surveyed with a laser-range-finding
total station.
4.4 Stratigraphy
All four trenches exposed interbedded, fine-grained alluvial units (gravels,
sands, and silts) and marsh deposits (peats and organic-rich clays) (Figures AP2.10
and AP2.11). These stratigraphic relations are continuous throughout the Yaylabeli
trench site. I was able to correlate major peat horizons between trenches by their depth,
thickness, color, texture and the results of
14
C dating. Layer thicknesses range from
a few millimeters to a few decimeters, with generally sharp stratigraphic contacts
associated with the near-complete absence of bioturbation in the marsh deposits due to
the perennially wet conditions. Stratigraphic contacts, even where they are gradational,
can be determined to within five centimeters. An increased proportion of distal fan
deposits were observed in the southern ends of all trenches. These deposits become
coarser-grained southward towards the fan source. They are composed of coarse-grained
sand, pebbly sand, and pebble gravel. Towards the center of the basin, the materials
become finer grained. In this area, lacustrine deposits and peat horizons thicken into the
trough at the axis of the marshy low. These deposits are composed of clays, silty clays,
organic-rich clays, and peats. Towards the northern end of my trenches, the lacustrine
deposits and peat horizons onlap and pinch out against the altered bedrock that makes
up the shutter ridge to the north.
4.5 Age Control
I radiocarbon (
14
C) dated 34 peat samples from the Yaylabeli trenches. All of these
samples were pretreated and analyzed at the Center for Accelerator Mass Spectrometry
63
Figure 4.7 Log of eastern wall of Yaylabeli trench T3. The uppermost face of the benched trench is 1.5 m east of the middle
face, which is 1.5 m east of the lowest face. Color coding as in figure 5.
64
Figure 4.8 Logs of eastern and western walls of Yaylabeli trench T4. The western wall is shown as a mirror image for ease
of correlation. The trench walls were 90 cm apart. Color coding as in figure 5.
65
(CAMS) at Lawrence Livermore National Laboratory. I examined the peat samples
under a microscope and separated the “flattened”, partially degraded, leafy plant
material, which I think represents the youngest, in-situ component of each peat, from
detrital charcoal and roots. Following this physical separation, I performed standard
acid-alkali-acid (AAA) chemical pretreatments on these selected leafy components
of each sample. In addition to radiocarbon dating of the resulting residual carbon,
I collected the alkali-treated washes in a separate container and dated the resulting
humic fractions of some samples in order to obtain the youngest-possible in-situ ages
for the peat samples, following the methodology established previously by others at
stratigraphically similar peat-rich sites [e.g., Schoute et al., 1981, 1983; Johnson et al.,
1990; Fumal et al., 2002]. Accelerator Mass Spectrometry (AMS) radiocarbon dating
of the paired AAA-treated fractions and associated humic fractions from the same peat
samples yielded similar ages. Specifically, eight out of ten of the humic fraction ages
returned results that are very similar to those of the radiocarbon ages of the AAA-
treated peat samples. Two samples (T4-07 and T4-08), however, returned ages for the
AAA-treated fractions that are ~100-200 years older than those of the associated humic
fractions from the same sample, as might be explained as a result of some detrital
charcoal left in the residuals. Thirty one out of 34 radiocarbon dating results of these
samples revealed ages in correct stratigraphic order (Table 4.1). The remaining three
samples (T3-1-top, T3-1-base, and T3-6) apparently contained significant reworked
detrital carbon.
In my descriptions of event ages in the following section, I use OxCal 4.0 [Bronk
Ramsey, 1995; Bronk Ramsey, 2001; Reimer et al., 2004] to determine calibrated,
calendric ages for each radiocarbon date. In addition to presenting the calendric date
range of each sample, I also use OxCal 4.0 to model the distribution of sample ages
66
Table 4.1 Radiocarbon age dating results from the Yaylabeli site. Pretreatment
method “R” indicates measurement results for carbon residue fractions based on AAA
pretreatment of hand-picked “leafy” components of each peat sample. “H” indicates
that the radiocarbon measurement was performed on the humic fraction (from the alkali
wash in pretreatment) of the same sample.
67
Table 4.1
68
Figure 4.9 Probability-distribution functions for radiocarbon ages and layer-modeling
results for events E3, E4, and E5, based on OxCal v4.0 [Reimer et al., 2004; Bronk-
Ramsey, [2005]). In this program, individual probability distribution functions (pdf)
for each radiocarbon sample age are trimmed according to stratigraphic relationships.
Phases P20, P30, P40, and P50 represent age groups from peat layers that are identifiable
across the entire trench site. The label corresponding to each pdf indicates the trench
number, followed by the sample number (see Table 4.1).
69
that are in correct stratigraphic order (Figure 4.9). This program re-weights the age
distributions of adjacent layers which overlap by using stratigraphic ordering and
incorporating the sample probability density functions to directly determine layer and
event ages [e.g., Biasi and Weldon, 1994; Bronk-Ramsey, 1995; Hartleb et al., 2006].
Finally, I compared these trimmed age ranges with historical earthquake records to
determine if each of my paleoseismologically defined event horizons could plausibly
be correlated with a major historical earthquake.
4.6 Interpretation of Paleo-Surface Ruptures at Yaylabeli
I identified evidence for five surface ruptures in my trenches, referred to as
Events 1 through 5, from youngest to the oldest. Evidence for these paleoearthquake
event horizons includes upward fault terminations, fissure fills, the geometry of growth
strata, and liquefaction features. Most of the event horizons were observed in multiple
exposures. The predominance of sharp contacts and laterally continuous beds allows
me to define the precise stratigraphic level of many of these event horizons. As noted
above, most of my event evidence comes from trenches T3 and T4 (Figures 4.7 and
4.8), with additional evidence from trenches T1 and T2 (Figure 4.5 and 4.6). In the
following section, I present structural and stratigraphic evidence, age control, and
pertinent historical data for each of these events.
4.6.1 Event 1
The most recent surface rupture at the Yaylabeli site was the 1939 M~8 Erzincan
earthquake. This event extends to within a few centimeters of the surface in T1, T2,
and T4, and eyewitnesses to the earthquake corroborate the location of the surface
rupture exposed in my trenches. Specifically, in trench T1, at least one, and possibly
70
two, fault strands terminate ~20 cm below the ground surface between m4 and m5 on
the western wall, within unit P00 (Figure 4.5). In trench T4, Event 1 is expressed as a
single, well-defined upward fault termination at m7, <5 cm below the ground surface
on the eastern wall. On the western wall of trench T4, Event 1 is marked by two fault
strands that bound a 25-cm-wide, 30-cm-deep fissure fill at ~m5.5. These fault strands
terminate upward within the unit peat P00, as in trench T1. Interestingly, in trench T3,
which was located between trenches T2 and T4, no fault strands extend to the surface.
Instead, the upper meter of section in T3, encompassing peats P00 and P10, is folded
into a syncline-anticline pair (Figure 4.7). This young folding appears to reflect warping
during the 1939 surface rupture within the releasing step between T1 and T4 (Figure
4.4). The syncline exposed at m25.4 in T3 was expressed as a topographic swale prior
to opening the trench, attesting to the recency of this deformation.
I do not have direct age constraints for the occurrence of Event 1. However,
based on the proximity to the ground surface of the multiple upward fault terminations,
the topographic expression of the near-surface syncline in T3, and the eyewitness
descriptions, I conclude that Event 1 is the historical 1939 Erzincan earthquake.
4.6.2 Event 2
Event 2 is defined by upward fault terminations in trench T3 and T4, at m27.47
and m6, respectively, and by liquefaction of Peat 20. In trench T3, a clearly defined
fault strand at m27.47 offsets peat P15, but terminates upward at the top of the peat
P15, beneath peat P10; P10 and the underlying clay layers are not offset. At m6 on the
western wall of trench T4, a fault strand offsets peat P20, which exhibits liquefaction
features, and terminates below peat P10. On the eastern wall of T4, however, a clear
fault strand, distinct from the strand that ruptured in Event 1 (1939), could not be
71
observed. In trench T2, I interpret a laterally extensive sheet sand at this stratigraphic
level as an extrusive sand blow deposit. At m6.5, this sand can be traced down into an
apparent sub-vertical feeder dike that cross-cuts stratigraphy. Although this extrusive
sand deposit occurs at approximately the same stratigraphic level as the E2 event
horizon observed in my other trenches, I cannot unequivocally correlate the extrusive
sand with Event E2 because peat P15 is not present in trench T2.
I dated samples T3-2, T4-06, and T4-11 from Peat 20 to provide a maximum
age constraint on the timing of Event 2. Dating of peat sample T3-P2 from trench T3
returned a calibrated, calendric age range of 990 A.D. to 1190 A.D. Dating of T4-06
and T4-11 from peat P20 in trench T4 returned calendric age ranges of 990 A.D. to 1160
A.D., and 980 A.D. to 1160 A.D., respectively. These dates confirm my stratigraphic
correlation of peat P20 between trenches T3 and T4.
Although I do not have a minimum age constraint for event E2, the event horizon
for this event is only ~10 cm above the top of peat unit P20. Given the lack of evidence
for any major hiatuses in this section, this would suggest that E2 probably occurred
soon after the deposition of P20 (i.e., soon after the 980-1190 A.D. age range of P20).
Moreover, E2 occurred after event E3, which I interpret below as the historical 1045
earthquake. The historical 1254 earthquake, which caused substantial damage and loss
of life in Erzincan was likely generated by rupture of an ~150 km of the NAF between
Erzincan and Suşehri [Ergin et al., 1967; Ambraseys and Melville, 1995; Barka, 1996;
Ambraseys and Jackson, 1998]. My results corroborate the findings of Hartleb et al
[2006], and suggest that there were no major surface ruptures on this stretch of the NAF
between the 1254 earthquake and the 1939 Erzincan earthquake.
72
4.6.3 Event 3
Evidence for Event 3 is best observed in trench T3 between m23 and m27.2.
There, nine fault strands terminate upward at ~75 cm depth, just below the base of
P20. Liquefaction of peat P30 in trench T3 probably also occurred during this event.
Moreover, the greater vertical separation of P30 relative to P20 at m7 in trench T4 is
consistent with an event horizon at this depth. Specifically, P30 exhibits 35 cm of down-
to-the-south separation on the fault at m6.25 in T4, whereas P20 exhibits only 14 cm
of down-to-the-south separation. The absence of isolated, upward fault terminations at
the same stratigraphic level in T4 suggests that Event 3 ruptured only the main strand
of the fault at m7, which subsequently re-ruptured in later events.
The ages of peat P20 and peat P30 bracket the occurrence of Event 3. Three
samples (T3-3, T4-05 and T4-10) from P30 provide a calibrated, calendric maximum
age range for Event 3 of 660 A.D. to 1160 A.D. As noted above, samples T3-2, T4-06
and T4-11 from peat P20, which was deposited after Event 3, yield a calendric age
range of 980 A.D. to 1190 A.D. A chronological model of layer and event ages based on
OxCal 4.0 [Bronk Ramsey, 1995; Bronk Ramsey, 2001; Reimer et al., 2004] suggests
that Event 3 likely occurred between 910 A.D. and 1110 A.D. The historical, large-
magnitude earthquake of April 5, 1045, which is thought to have ruptured a long section
of the eastern part of the North Anatolian fault in the Erzincan-Suşehri region [Ergin
et al., 1967; Ambraseys, 1970; Barka, 1996; Ambraseys and Jackson, 1998; Hartleb
et al., 2006], occurs in the middle of the likely age range for event E3. I therefore
interpret Event 3 as the historical 1045 earthquake. At their Çukurçimen paleoseismic
site, Hartleb et al. [2006] also saw evidence for a surface rupture that occurred soon
after 880 A.D., which they also interpreted as the 1045 earthquake.
73
4.6.4 Event 4
Event 4 is recorded by fissure fills and upward fault terminations in trenches
T3 and T4. Specifically, in trench T3, two fault strands terminate upward at m24.81 at
180 cm depth. These strands are ~25 cm apart at ~180 cm and merge at ~230 cm depth,
forming a wedge-shaped fissure. The fissure is filled with fine-grained sand derived
from a thin sand layer between peats P40 and P45. The faults bounding the fissure
terminate upward within this sand bed. Similarly, in trench T4, 20 and 30 cm-wide,
wedge-shaped fissures are present on east and west wall exposures respectively. These
fissures are centered at m8.4 on the east wall and at m8.21 on the west wall. Faults
bounding the fissure merge at m8.4,-1.74 on the east wall whereas the faults bounding
the fissure on the west wall merge at m8.39,-2.07.
The ages of peats P40 and P45 bracket the timing of Event 4. Three calibrated,
calendric dates from P40 (T3-5, T4-04, T4-09) yielded an age range of 680 to 1020
A.D., providing a minimum age for Event 4. Sample T4-07 from Unit 45 yielded a
calendric age range of 660 to 870 A.D., providing a maximum age. Thus, Event 4
occurred between 660 and 1020 A.D. My layer- and event-age modeling of these age
constraints indicate that Event 4 likely occurred between 710 and 850 A.D (Figure 8).
This 8
th
-9
th
century surface rupture at Yaylabeli was not observed by Hartleb et
al. [2006] at their nearby Çukurçimen site, which had no recorded events between 540
A.D. and 930 A.D. My re-evaluation of the Çukurçimen trench data reveals that this
interval in their trenches is stratigraphically and structurally complex. In addition, this
stratigraphic interval was located at a horizontal bench at the fault zone in their key
trenches. These observations lead me to suspect that this event horizon may exist, but
went unrecognized at Çukurçimen.
74
In marked contrast to the other four earthquakes that I identify (E1, E2, E3,
E5), to the best of my knowledge there are no known historical accounts for a large-
magnitude earthquake in this part of Anatolia during the 710 to 850 A.D. period when
Event 4 likely occurred. This was a period of unrest in eastern Anatolia, with ongoing
wars between Byzantines and Arabs [e.g., Haldon, 2002], and it is possible that
earthquake catalogs may be incomplete for this time period.
4.6.5 Event 5
Event 5 occurred after deposition of peat P60, and before the deposition of P50.
In trench T3, the event horizon is marked by two fault strands that terminate upward
at m22.66 and m22.98, just below peat P50 (Figure 4.7). In trench T4, a fault strand
at m6.09 terminates at the base of peat P50, at the same stratigraphic level as in T3.
In trench T1, however, I could not observe a distinct upward fault termination for this
event. Instead, the fault is marked by a ~30-cm-wide gouge zone, suggesting repeated
rupture of the same strand. On the other hand, a fine-grained sand unit that exhibits
liquefaction features is exposed between peats P50 and P60 in trenches T1 (Figure 4.5;
m3,-1), T2 (Figure 4.6; m10,-2), T3 (Figure 4.7; m23,-2), and T4 (Figure 4.8; east wall;
m6.5,-2), providing additional evidence for event E5.
Calibrated, calendric radiocarbon dates from five peat samples (T1-6A, T1-6B,
T3-7, T4-03 and T4-08) from peat P50 in trenches T1, T3, and T4, indicate that Event
5 occurred before deposition of P50 at 650 to 970 A.D. Sample T4-02, from Peat 60
in trench T4, yielded an age of 250 A.D. to 420 A.D., providing a maximum age limit.
Event 5 therefore occurred between 250 and 970 A.D. My stratigraphic layer modeling
narrows the likely age range for Event 5 to between 320 A.D. and 690 A.D.
75
It is possible that Event 5 is the historical 499 A.D. earthquake, which caused
substantial damage in the Niksar-Suşehri region ~75-175 km to the west of the Yaylabeli
site [Ergin et al., 1967; Ambraseys and Jackson, 1998] (Figure 4.1). This inference is
supported by interpolation of my sediment accumulation rate curve, which suggests
that Event 5 occurred at ~500 A.D. (Figure 4.10).
4.7 Comparison of Yaylabeli and Çukurçimen Paleoseismologic data;
Reproducibility of trench results
The Yaylabeli trenches provide a record of five surface-rupturing earthquakes on
the eastern part of the North Anatolian fault during the last 2000 years. Relatively rapid,
continuous sediment accumulation at the site at 1-3 mm yr
-1
, and wide stratigraphic
separation of the different earthquake event horizons, suggest that this record is likely
to be complete. My results are generally similar to those of Hartleb et al. [2006] from
the nearby (2 km to the east) Çukurçimen trench site, which was also located in a
semi-enclosed, peat-rich marsh fed by small alluvial fans. The reproducibility of the
results from these two sedimentologically independent trench sites, coupled with
the generally close correspondence between my paleoseismologic observations and
historical earthquake records, lends confidence to my results.
Specifically, both the Yaylabeli and Çukurçimen trench sites contain evidence
for: (1) the historical 1939 M
w
7.9 Erzincan earthquake (Event 1 at Yaylabeli; “Event
A” at Çukurçimen), (2) an event (Event 2) that occurred soon after the 990-1160 A.D.
age of a well-dated peat, and after the historical 1045 earthquake at Yaylabeli, and
between 980 A.D. and 1420 A.D. at Çukurçimen (Event B). I follow Hartleb etal.
[2006] in interpreting this event as the historical 1254 earthquake; (3) a surface rupture
dated to between 910 to 1110 A.D. at Yaylabeli (Event 3) and 930 A.D. to 1070 A.D.
76
Figure 4.10 Sediment-accumulation-rate curve for the Yaylabeli site. The continuity
of the curve demonstrates that there are no major hiatuses during the past ~2000 years.
Colored horizontal bars show the depth and age ranges of individual samples from all
trenches (see Table 4.1 and Figure 4.8). Red arrows indicate the depths of the event
horizons relative to ground surface in trench T4 for the five paleo-surface ruptures I have
identified. The vertical lines are interpolations of these events based on my sediment-
accumulation-rate curve (dashed line). The age ranges on the vertical interpolation
lines are the trimmed age ranges from my layer modeling in OxCal v4.0 (see text for
discussion).
77
at Çukurçimen (Event C). I concur with Hartleb et al. [2006] that this event is likely
the historical 1045 A.D. earthquake; (4) an event marked by well-defined fissure fills
in multiple exposures at Yaylabeli that occurred between 710 and 850 A.D. Although
the Çukurçimen study did not reveal evidence for an earthquake during the 7
th
to
9
th
centuries, recognition of this event fills in one of the two long temporal gaps in
earthquake occurrence along the eastern NAF suggested by Hartleb et al. [2006],
(specifically, the 390 to 710-year-long interval between their events C and D); and
(5) an event that likely occurred between 320 and 690 A.D. at Yaylabeli (Event 5) and
between 250 and 540 A.D. at Çukurçimen (Event D). This event probably represents
the historical 499 A.D. earthquake. The addition of the new Yaylabeli paleoseismic
data allows me to update the space-time history of the earthquake occurrence along the
NAF of Hartleb et al. [2006] (Figure 4.11).
4.8 Discussion
The Yaylabeli and Çukurçimen paleoseismologic data indicate that only five
surface ruptures have occurred on the eastern part of the NAF over the past ~1800
years. Assuming that my correlations with the historical records are correct, three of the
four intervals between these five earthquakes are 200 to 350 years long (1045 to 1254 =
209 years; ~770? + ~70 A.D. to 1045 ≅ 200 to 350 years; 499? to ~770? + 70 = ~200 to
350 years). This suggests that the timing of earthquake occurrence on the eastern part
of the NAF is commonly relatively regular. The fourth interval, however, is 685 years
long (1254 to 1939), two to three times longer than the others.
As noted by Hartleb et al. [2006], the historical 1579 and 1583 earthquakes,
which caused widespread damage in parts of northern Anatolia, occurred during this
long interval. The 1579 earthquake, in particular, caused damage that may have extended
78
Figure 4.11 Space-time plot of historical and paleoseismologically documented surface
rupturing earthquakes along the North Anatolian fault (modified from Hartleb et al.
[2006] with the addition of new data). The simplified fault map at top shows the location
of the paleoseismic sites used in this compilation. Vertical black bars indicate the
bracketing age constraints from paleoseismologic studies and circles indicate potentially
correlative historical earthquakes. References for the paleoseismologic results shown
above are: (A) Rockwell et al. [2001a], (B) Klinger et al. [2003], (C) Rockwell et al.
[2001b], (D) Pantosti et al. [2008], (E) Kondo et al. [2004], (F) Okumura et al. [2003],
(G) Sugai et al. [1998], (H) Yoshioka et al. [2000], (I) Hartleb et al. [2003], (J) this
study, (K) Hartleb et al. [2006].
79
over a 400-km-long region of northern Turkey, from Çorum to Erzincan [Ambraseys
and Finkel, 1995], reminiscent of the damage pattern that might be expected from a
large-magnitude NAF earthquake. Thus, it is a potentially attractive idea to assume that
either the 1579 or 1583 earthquakes (or both) occurred on the NAF. Yet, like Hartleb et
al. [2006], who found no evidence for a 16
th
century surface rupture in their Çukurçimen
trenches, I find no evidence for a surface rupture during this period in my trenches at
Yaylabeli. Given the absence of any other sources large enough to generate an earthquake
capable of causing damage over such a wide area, the lack of evidence for either of
these earthquakes in the Yaylabeli and Çukurçimen trenches is puzzling. It is, of course,
possible that an event may have gone unrecognized in the paleoseismologic trenches,
as was the case for my Yaylabeli Event 3 at the nearby Çukurçimen site. But in light of
the large number of trench exposures, the rapid sediment accumulation rates, absence
of bioturbation, and the excellent stratigraphy at these sedimentologically independent
sites, this seems rather unlikely. In Yaylabeli, for example, the sediment accumulation
between events E1 (1939) and E2 (1254?) appears to be continuous throughout this
interval in both T3 and T4, without any evidence for erosional features suggestive of
significant hiatuses during this interval. In addition, the E1 and E2 event horizons are
separated by a meter of section characterized by continuous, thin bedding with sharp
depositional contacts. At the Çukurçimen site, Hartleb et al. [2006] carefully assessed
the possibility that 1579 and/or 1583 earthquake(s) were missed. As at Yaylabeli,
their Çukurçimen trenches showed no evidence for any of these earthquakes, despite
continuous exposure of laterally continuous, thin-bedded strata across the fault zone
within this stratigraphic interval. These observations suggest that the 1579 and/or 1583
earthquakes did not cause surface rupture on the NAF west of the Erzincan basin.
80
Alternatively, it is possible that the 1579 and 1583 earthquakes were local,
smaller-magnitude events. Historical data indicate that the 1579 earthquake caused
widespread damage in and around the cities of Çorum and Amasya, in north-central
Anatolia; a separate, near-contemporary account suggests that damage may also have
occurred in Erzincan, ~350 km east of Amasya [Ambraseys and Finkel, 1995] (Figure
4.1). If the report of damage in Erzincan relates to a different earthquake, the 1579
earthquake may not have caused damage throughout northern Anatolia. Rather, the
well-documented damage in the Çorum-Amasya area could have been caused by a
moderate-magnitude earthquake on the Ezinepazarı fault, part of which ruptured as the
westernmost section of the 1939 rupture between Ilgaz and Niksar (Figures 4.1 and
4.11), or one of the reverse faults associated with uplift of the Pontide mountains north
of Ilgaz. Similarly, historical evidence for the 1583 earthquake comes mainly from the
Erzincan area, which was heavily damaged in this earthquake with great loss of life.
The 1583 earthquake could have been a moderate-magnitude earthquake that occurred
near Erzincan, similar to the 1992 M
s
6.8 Erzincan earthquake, which caused 2000
fatalities and was generated by rupture of a ~30-km-long section of the NAF that did
not cause any surface rupture [Barka and Eyidoğan, 1993; Fuenzalida et al., 1997].
Nevertheless, the sources of the 1579 and 1583 earthquakes remain unknown,
and if future studies unearth unequivocal evidence for one or both of these earthquake(s)
on the eastern NAF, this would cut the 685-year-long gap between the 1254 and 1939
earthquakes nearly in half. The resulting intervals would be 325 years (1254 to 1579)
and 360 years (1579 to 1939). Thus, if 1579 did occur on eastern NAF, the recurrence
intervals between all five of the most recent earthquakes would range from 200-350
years, indicating very regular recurrence of large-magnitude earthquakes on this
section of the NAF. Even if the 1579 (or 1583) earthquake(s) did not occur on the North
81
Anatolian fault, as indicated by my paleoseismic results, the occurrence of earthquakes
on the eastern part of NAF appears to be relatively regular, with at least three of the
four intervals between last five events varying by less than factor of two, from ~200 to
350 years.
The Yaylabeli and Çukurçimen paleoseismologic data demonstrate surface
rupturing earthquakes on the eastern part of the NAF have been rare events during at
least the past two millennia, with recurrence intervals ranging from ~200 to almost
700 years. Geologic data from the central part of the NAF indicate a slip-rate over this
time interval of ~18-22 mm yr
-1
[Kozacı et al., 2007; in review]. These observations
indicate that large amounts of elastic strain accumulate during interseismic periods on
this part of the NAF. This accumulated strain energy is released by large, infrequent
earthquakes, such as the 1939 M
w
7.9 Erzincan earthquake, which generated ~10 m of
surface displacement in the Yaylabeli-Çukurçimen area [Hartleb et al., 2006].
The North Anatolian fault is well known for the occurrence of a mainly westward-
propagating sequence of moderately large to great earthquakes (Mw 7.0-8.0) during
the 20
th
century [Ketin, 1948; Richter, 1958; Allen, 1969; Ambraseys, 1970; Toksöz et
al., 1979; Barka, 1992, 1996, 1999; Ambraseys and Finkel, 1995; Stein et al., 1997].
Historical and paleoseismological evidence shows that the 20
th
century sequence is not
the only such cluster generated by the NAF. For example, historical and paleoseismic
studies indicate that the entire central and eastern NAF ruptured during a relatively
brief cluster in the 10
th
-11
th
century, including the historical 967, 1035, 1045, and 1050
earthquakes [Barka, 1996; Ambraseys and Mellvile, 1995; Stein et al., 1997; Sugai et
al., 1998; Yoshioka et al., 2000; Pantosti et al., 2008; Kondo et al., 2004; Okumura et
al., 2003; Hartleb et al, 2006] (Figure 4.11). Similarly, the entire central and western
sections of the NAF ruptured during a sequence of eight moderately large to great
82
earthquakes between 1666 and 1766, including the great M~8 1668 earthquake, which
probably ruptured an ~400-km-long section of the central NAF from near Bolu to the
Niksar step-over [Barka, 1996; Ambraseys and Finkel, 1995; Stein et al., 1997] (Figure
4.1). As discussed above, my paleoseismologic data suggest that this cluster did not
extend along the eastern NAF, which had ruptured earlier in the 1254 earthquake.
My identification of a 710-850 A.D. surface rupture at Yaylabeli, combined
with other paleoseismologic data [Sugai et al., 1998; Yoshioka et al., 2000; Okumura
et al., 2003], suggests the occurrence of yet another cluster during the 8
th
-9
th
centuries.
This previously unrecognized earthquake sequence possibly ruptured more than 600
km of the central and eastern NAF. However, the behavior of the western NAF during
this period is difficult to discern because the historical earthquake record before the
6
th
century becomes difficult to interpret, particularly in terms of magnitude and exact
location of the earthquakes [e.g., Ambraseys 2002; 2006]. This problem is exacerbated
by the absence of any paleoseismologic constraints for earthquakes from the western
NAF prior to the ~9
th
-10
th
centuries [e.g., Rockwell et al., 2001a].
Although these observations emphasize that relatively brief clusters of
earthquakes are common along the NAF, the earthquakes that occur during these
clusters are not the same from sequence to sequence, in terms of location, magnitude,
and sequence of rupture. For example, the largest event in the 20
th
century cluster
was the M~8 1939 Erzincan earthquake, and events in this sequence occurred in a
predominantly westward-propagating sequence. The sequence of individual earthquakes
in the cluster has been explained as resulting from a combination of the stress evolution
of the fault system, as modeled by changes in Coulomb Failure Function [Stein et al.,
1997], and by fault-zone material contrasts, which may play a major role in controlling
rupture propagation and arrest [e.g., Dor et al., 2008]. Likewise, the 17
th
-18
th
century
83
sequence propagated mainly westward, but comprised earthquakes with magnitudes
and locations that were significantly different from those of the 20
th
sequence (e.g.;
the great 1668 earthquake appears to have ruptured approximately the 1943 and 1944
rupture segments). In contrast to these mainly westward propagating sequences, the
10
th
-11
th
sequence appears to have propagated both to east (the 967, 1035, and 1050
earthquakes, which ruptured in sequence eastward from the Bolu area) and west (the
1045 earthquake occurred along the eastern NAF, prior to the 1050 earthquake on the
central part of the fault) (Figure 4.1).
Moreover, although the NAF appears to commonly rupture in approximately
century-long clusters of earthquakes, it is important to note that not all large-magnitude
earthquakes on the NAF occur as part of such sequences. For example, both historical
observations [Ergin et al., 1967; Ambraseys and Melville, 1995; Barka, 1996; Ambraseys
and Jackson, 1998] and paleoseismologic data [Hartleb et al., 2006; this study] indicate
that the 1254 earthquake ruptured a long stretch of the east-central part of the NAF. Yet
this earthquake does not appear to be part of any obvious sequence similar to the 20
th
-
century cluster. Similarly, the 1509 earthquake in the Marmara region near İstanbul
[Ambraseys and Finkel, 1995] does not fit neatly into any well-documented cluster.
The relationship of the 1912 M 7.4 [Pondard et al., 2007] Gulf of Saros earthquake to
the 20
th
-century cluster likewise remains unclear.
4.9 Conclusions
My paleoseismologic trenches at Yaylabeli, on the eastern North Anatolian
fault, provide evidence for five paleo-surface ruptures during the past two Millennia,
including a previously unknown 8
th
-9
th
century surface rupture. The timing of these
earthquakes indicates the relatively regular recurrence of large-magnitude events for
84
this part of the NAF, with three of four intervals between events varying between
200-350 years. The recognition of one long (685 years) interval, however, indicates
that recurrence is not truly quasi-periodic on this stretch of the NAF. These results, in
combination with other paleoseismologic data, indicate that the NAF is characterized
by the occurrence of infrequent, large-magnitude earthquakes. Moreover, my results,
particularly the identification of a previously unrecognized 8
th
-9
th
century cluster on the
central and eastern NAF, strengthen the idea that the North Anatolian fault commonly,
but not always, ruptures in clusters lasting on the order of a century.
I attribute this relatively simple behavior to: (1) the relative structural simplicity
of the plate boundary in northern Turkey, in which >90% of plate boundary motion is
accommodated by slip along the NAF, with few other significant, high-slip rate faults;
and (2) the mechanical efficiency of the structurally mature North Anatolian fault. Thus,
the stress evolution of the NAF is generally not complicated by stress interactions
from earthquakes generated by other faults. These results emphasize the importance
of the degree of tectonic complexity in controlling earthquake occurrence, and suggest
that earthquake occurrence on mature, structurally isolated strike-slip faults will be
dominated by the relatively regular recurrence of large-magnitude events.
85
CHAPTER 5:
Slip-Rate Study Sites in Progress
5.1 Additional slip rate sites
In this chapter, I present my preliminary results from four additional slip-rate
sites (Figure 5.1). The first two sites are located to the east and west, respectively, of
my earlier Eksik slip-rate site (the results from that site are presented in Chapter 2). I
mapped, sampled, and started dating samples from these two slip-rate sites. The third
site that I present in this chapter is the Destek slip rate site, which is located ~220 km to
the east of Eksik. The fourth site, which is the focus of an ongoing project, is near the
village of Koçyatağı (Mihar), ~270 km to the east of Destek and ~20 km west of the
city of Erzincan along the eastern part of the North Anatolian fault. I received much-
appreciated contributions from my collaborators James Dolan (University of Southern
California), Robert Finkel (Lawrence Livermore National Laboratory), and Ziyadin
Çakır (Istanbul Technical University) during these studies.
86
Figure 5.1Simplified tectonic map of the eastern Mediterranean region [modified after
Barka, [1992] and Armijo et al., [2002]. Arrows indicate generalized relative motions
according to the fixed Eurasian Plate showing rates in millimeters per year [after
McClusky et al., [2000] and Reilinger et al., [2006]. “Çatal Dere”, “Yatçam Dere”,
“Destek”, and “Mihar” in red boxes indicate the locations of sites presented in this
chapter. NAF-North Anatolian fault, EAF-East Anatolian fault, NEAF-North East
Anatolian fault, OF-Ovacik fault, EF-Ezinepazari fault, DSF-Dead Sea fault. Satellite
image is from Orbitview-2 satellite (available at http://visibleearth.nasa.gov).
87
5.2 Çatal Dere slip rate site
The west Eksik slip-rate site focuses on a right-lateral offset of Çatal Dere (Çatal
Dere means “Forked Creek” in Turkish; Figure 5.2). This offset is located ~500 m to
the west of my earlier Eksik slip rate site [Kozacı et al., 2007]. An older, larger offset
of Çatal Dere was recognized by Hubert-Ferrari et al. [2002], as part of a series of 200
m offsets they mapped along the North Anatolian fault. At the Çatal Dere offset site
a fluvial terrace made up of limestone cobbles and boulders is nestled into this major
bedrock canyon, partially infilling the older canyon, in a situation that is very similar
to the Eksik slip rate site. The south-flowing Çatal Dere is offset right-laterally along
the North Anatolian fault ~1.5 km south of its headwaters. My preliminary mapping
demonstrated that the displacement of the eastern inner edge of this alluvial terrace is
~50 m. The bedrock canyon into which these terrace gravel were deposited, however, is
offset by ~200 m, as suggested by Hubert-Ferrari et al. [2002]. As at my earlier Eksik
slip rate site, the terrace gravels were deposited into a deeply incised bedrock canyon
that had already accumulated ~150 m of offset before terrace deposition.
At this site, two south-flowing drainages have incised into the fluvial terrace
gravels that were deposited within the much older bedrock canyon. I focused my
mapping on the eastern-most inner edge of the terrace deposits because the western
extent of the terraces is buried by the material transported by Beygiruçtu Dere (which
means Donkey Flew off [the Cliff] Creek in Turkish), the next major canyon to the west
of Çatal Dere. Beygiruçtu Dere is currently capturing the southern extent of Çatal Dere
as a result of progressive right-lateral displacement along the North Anatolian fault.
88
Figure 5.2 (a) Corona satellite image overlaid with geomorphic interpretation [modified
from Hubert-Ferrari et al., 2002]. Offset fluvial terraces discussed in text highlighted
in yellow. Site “1” is the Çatal Dere site discussed in this chapter. Site “2” is the
Eksik slip-rate site [Kozacı et al., 2007; Chapter 2]. Site “3” is the Yatçam Dere offset
discussed in this chapter. BD = Beygiruçtu Dere, CD = Çatal Dere, AD = Ağılönü Dere,
YD = Yatçam Dere, KD = Kuru Dere. “Dere” means “creek” in Turkish. White box
at site 3 shows location of figure 5.2b. (b) 3D survey map with trench locations north
of the fault, west of Yatçam Dere. Trench depths at this location range from ~2 m to
2.5 m. as shown in 5.2b, the NAF offsets a small, fault-perpendicular ridge by 5.83 m,
probably as a result of displacement during the 1943 earthquake.
89
Mapping the surface extent of the terrace deposits to the east of Çatal Dere is
facilitated by the lithologic contrast of the white/pale gray limestone gravels and the
underlying dark green schist. I was able to map this contact at the surface in detail for
>100 m northwards from the fault. The terrace to the north of the fault is exposed over a
wide area between the active drainage and the buttress unconformity at its eastern edge,
where it onlaps bedrock. The surface of the terrace deposits exhibits relatively planar
topography (with local relief of up to 40cm) north of the fault.
To the south of the fault, the eastern limit of the terrace gravels is somewhat
harder to define as it lies near the eastern edge of a stand of pine trees and the gravels
are buried by 10-20 cm of soil and leaf litter. We were able to identify the approximate
location of the inner edge, however, by digging numerous, shallow (30 cm) pits to
expose the presence of the distinctive white limestone cobbles that make up the terrace
deposit. The active drainage converges downstream with the eastern inner edge of the
terrace deposit and the base of the terrace gravels is visible in several natural side-
drainages (50 m south of the fault).
I sampled the terrace remnant between Beygiruçtu Dere and Çatal Dere north
of the fault. I dug a 1.5-m-deep depth profile (EX-TP-13 coordinates: UTM 555787,
4540923) and recovered five
36
Cl samples from depths of 10, 30, 50, 70, and 100 cm. I
also recovered a charcoal sample (EX-TP-13-C100-06) at 100 cm depth. In addition, I
collected five surface samples within a 10-meter radius of the depth profile (EX-TP-13)
location. The
36
Cl samples have not yet been processed. The radiocarbon dating result
for my EX-TP-13-C100-06 sample returned a calibrated age range of 2470-2200 BC
(95% confidence). I performed the radiocarbon age calibration by using OxCal v.3.10
[Bronk Ramsey, 1995; 2001; using atmospheric data from Reimer et al, 2004].
90
The combination of my preliminary mapping of the eastern inner edge offset
of Çatal Dere and the one radiocarbon age constraint (~55 m and 4143-4413 calender
years old) yields a minimum slip rate of 12.9 + 0.4 mm yr
-1
. This rate is significantly
slower than my earlier Eksik slip rate of ~20.5 + 5.5 mm yr
-1
from the next drainage
to the east. The Çatal Dere rate is a minimum, however, for two reasons. First of all,
stabilization and abandonment of the surface of the terrace from which I recovered the
charcoal sample predates the subsequent incision and initiation of displacement. In
addition, as with any charcoal dating, the amount of inheritance for this specific sample
is unknown. Dating of both the
36
Cl surface exposure samples and
36
Cl depth profile
samples should yield a more robust slip rate constraint for the Çatal Dere site.
5.3 Yatçam Dere slip rate site
The second slip rate site that is currently under development is the Yatçam Dere
offset (“Leaning Pine Creek” in Turkish). This site is located ~1 km to the east of my
Eksik slip-rate site [Kozaci et al., 2007], and ~150 m to the east of the village of Eksik
(Figure 5.2). As with both the Kozacı et al. [2007] and Çatal Dere sites to the west, at
the Yatçam Dere site an extensive, flat-topped fluvial terrace that partially infilled a
pre-existing bedrock canyon has been deeply incised by the active drainage of Yatçam
Dere. This terrace is composed of a thick sequence of white to gray limestone gravels
that were deposited into an older deeply incised canyon, which has been offset ~500 m
along the North Anatolian fault (Figure 5.2a). The offset inner edge of this extensive,
younger fluvial terrace deposit forms the focus of this study.
At the Yatçam Dere site, two drainages, Yatçam Dere and an unnamed drainage
to the east, flow southward, perpendicular to the fault. North of the fault, Yatçam Dere
has incised into the terrace deposits by ~20 to 30 m, isolating a sliver of terrace deposits
91
to the west from the main terrace body, which is located in the center of the canyon
between the two active drainages (Figure 5.2). For ~300 m southward from the fault,
Yatçam Dere flows southward and continues to incise into the terrace deposits before
it merges with the unnamed drainage to the east. The unnamed drainage in the east,
however, flows along the contact of the fluvial terrace deposits and bedrock. This
drainage is deflected to the west by the offset bedrock canyon walls by ~500 m. At a
point ~300 m south of the fault, the unnamed eastern drainage merges with the Yatçam
Dere and the combined river flows along the bedrock-terrace deposit contact on the
east side of the canyon.
I concentrated my mapping efforts to map the displacement of the western inner
edge of the Yatçam Dere terrace deposits. In addition to surface mapping, I excavated
a series of trenches both north and south of the fault to locate the terrace inner edge at
these localities. I located a terrace inner edge in my trenches EX-TP-23 and EX-TP-24,
north of the fault (Figure 5.2b). In these trenches the limestone terrace cobbles show a
buttress unconformity against the underlying schist bedrock. I also exposed a fluvial/
bedrock contact in my trench EX-TP-20 north of the fault. However, I believe that this
might be the lower contact between the terrace deposits and bedrock, where the surface
has been eroded away. Although this could be another inner edge for a younger, inset
terrace, topographic relationships suggest that the fluvial material exposed in trenches
EX-TP-20, EX-TP-21, EX-TP-22, EX-TP-23, and EX-TP-24 are all partial exposures
of the same continuous deposit. In addition, I located the fault and the extent of the
terrace deposits in my trenches south of the fault. The displacement between these
inner edge exposures both south and north of the fault is ~200 + 10 m, according to my
preliminary mapping results.
92
I excavated four additional trenches, both north (EX-TP-6 and EX-TP-5) and
south (EX-TP-7 and EX-TP-8) of the fault. These trenches were situated within the
planar parts of terrace deposits. I sampled limestone cobbles from these trenches
for dating
36
Cl depth profiles. In addition, I sampled
36
Cl surface exposure samples
within a ~20 m radius of each trench. As of this writing, these samples have not yet
been processed. However, I recovered a total of eight charcoal samples from trenches
EX-TP-5, EX-TP-8, EX-TP23, and EX-TP-24. I pretreated these charcoal samples at
Lawrence Livermore National Laboratory using a standard acid-base-acid method. The
radiocarbon dating of these results yielded a somewhat scattered age range. Samples
EX-TP-5-1 and EX-TP-5-2 from trench EX-TP-5 yielded calibrated calendric age
ranges of 1400-1130 BC and 1430-1260 BC, respectively (all
14
C ages calibrated in
OxCal v.3.10 [Bronk Ramsey, 1995; 2001; using atmospheric data from Reimer et al,
2004]). Samples EX-TP-8-1 and EX-TP-8-2 from trench EX-TP-8 returned ages of
3940-3660 BC and 3790-3630 BC, respectively. Two samples from trench EX-TP-23
returned age ranges of 5210-4850 BC and 3340-3100 BC. Two samples from trench
EX-TP-24 returned a more homogenous age interval of 8540-8260 BC. I should note
that these results, as with any radiocarbon dating of charcoal, represent maximum ages
for the deposits in which they are found. The inheritance amount on each sample is
unknown.
In order to calculate the slip rate for any site, the amount of displacement and
corresponding time interval must be known. However, my preliminary investigations
and results demonstrated a potential for complexity in determining the offset at this
site. The wide range of dating results could also be a result of unnoticed multiple inset
terrace deposits with various amounts of displacements. If I assume that my ~200 +
10 m offset measurement is correct and that all the surfaces that I dated are parts of
93
the same single terrace surface, than the combination of these results yield a slip rate
ranging from ~19.1 mm yr
-1
to ~62 mm yr
-1
, depending on the radiocarbon dating result
used. If I assume that all of these dates are coming from the same terrace deposit, then
the older samples are probably reworked, with significant inheritance. This assumption
would make the slip rates at the faster end of the 19-62 mm yr
-1
range more plausible.
However, such fast slip rates (~62 mm yr
-1
) are incompatible with extensive geologic
and geodetic rate data from the North Anatolian fault (see discussion in Chapter 3). On
the other hand, if I assume that there are multiple inset terraces, than I should expect
larger displacements for older deposits and smaller displacements for younger deposits.
As of this point, I have not yet recognized an unequivocal offset smaller than the 200 +
10 m at Yatçam Dere. However, other possible terrace remnants might be obscured and/
or modified by the development of Eksik village and long-term cultivation activities. If
in fact there is such a smaller offset, it is likely to be a ~50 m displacement similar to
the two ~50 m displacements that I have documented within 2 km west of this site. This
could be a signal of a minor climate change ~2-3 ka that resulted in an aggradational
response within the deeply incised bedrock canyons into which the terrace gravels
were deposited. These potential complexities could potentially be solved by additional
mapping and dating as long as the key geomorphic markers have not been modified
significantly by the local inhabitants.
5.4 Destek Slip-rate site
Destek slip-rate site is located on the central section of the North Anatolian
fault, just east of the town of Destek (Figure 5.3). Allen [1969] first noted that numerous
incised, south-flowing channels were consistently offset right-laterally by the NAF at
Destek. Hubert-Ferrari et al. [2002] mapped a series of 180- to 200-m right-lateral
94
offsets at a number of sites along the NAF, including a 5-km-long stretch of the fault
near the village of Destek (Figure 5.3b). These authors suggested that offset began
to accrue at all of these sites when incision began ~10,000 years ago in response to
climatic changes. Based on this inference, they suggested that the Holocene slip rate
of the NAF is ~ 18 ± 3.5 mm yr
-1
, somewhat slower than the geodetically measured
rate of elastic strain accumulation across the NAF. Although this rate has been quoted
in a number of recent publications, it is not constrained by geochronologic data. In an
effort to place firm constraints on the average Holocene slip rate of the fault, I dated
the surfaces into which the offset drainage incised by using
10
Be cosmogenic nuclides.
5.4.1 Destek slip-rate site description
At the study site, nine river valleys are right-laterally offset along the North
Anatolian fault. Although Hubert-Ferrari et al. [2002] suggested that the average
displacement that best restores all the valley offsets back into a plausible original
configuration is ~200 m, their measurements of individual canyon offsets actually
range from 125 to 250 m.
The upper reaches of the drainages are very steep, as demonstrated by the 500m
drop in elevation of the headwaters 1.8 to 2.0 km north of the offset, to the 650 m
elevation of the incised alluvial surface. In addition, although I did not perform any
further mapping at this study site, my reconnaissance indicates that there are multiple
inset fluvial terraces along many of these offset drainages that could provide shorter-
term offsets.
The extensive alluvial surface into which the streams have incised is quite planar
and exhibits rounded shapes around the edges, indicating the initiation of erosion. These
originally planar surfaces have, however, long been cultivated (hundreds of years, if
95
Figure 5.3 (a) Map showing location of Destek slip-rate site, (b) topographic map
showing nine south-flowing drainages that are offset ~200 m along the North Anatolian
fault [modified from Hubert-Ferrari et al., 2002]. Red dots (T1 and T2) indicate
sampled surfaces and trench locations for depth profiles.
96
Figure 5.4 A panoramic view of Destek slip-rate site. The view is to the north. A series of drainages are offset along the
NAF by ~200 m as noted by Hubert-Ferrari et al. [2002]. The atomic signs show the sampled surfaces at the study site (T-1
on left [west] and T-2 on right [east]).
97
not for millennia). Long-term cultivation, along with rain and wind-related erosion,
somewhat complicates the interpretation of the geomorphology and geochronology of
the site (see discussion in Chapter 3).
I collected quartz cobbles from the alluvial surface, which can be found
abundantly, for
10
Be dating. In addition, I excavated two short, 2.5-m-deep trenches
across the alluvial surfaces into which the ~200-m-offset streams are incised (Figures
5.3b, 5.4 and 5.5). Trenches T1 and T2 exposed massive silts and clays without any
distinctive bedding, along with some local carbonate concentrations (Figure 5.5b).
Gravel- and cobble-sized quartz clasts, however, were rare within Trench T1, and
nonexistent in Trench T2. Nevertheless, I was able to find a sufficient number of quartz
clasts in T-1 to collect a depth profile to 2.5 m depth.
5.4.2 Age control
I used cosmogenic radionuclide (
10
Be) dating of surface samples to determine
the age of the alluvial surface. I processed and dated ten quartz cobbles and large
pebbles from the surface, and four quartz cobbles and pebbles from the depth profile.
As in my Tahtaköprü slip-rate site, I prepared all samples according to the procedures
Kohl and Nishiizumi [1992], and terrestrial cosmogenic nuclide concentrations were
measured at the Lawrence Livermore National Laboratory Center for Accelerator Mass
Spectrometry. I used the CRONUS online
10
Be-
26
Al age calculator (http://hess.ess.
washington.edu/math/) to determine sample ages (Table 1). Uncertainties of all ages
reported for this study site reflect what are known as “internal errors” as suggested
by Balco et al. [2008]. That is, errors from within a group of samples that are similar
to well-dated reference sites in terms of their age, latitude, elevation, and shielding
(“external errors” reflect comparison to a global data set).
98
The dating results of the ten surface samples yield a somewhat scattered range
of ages varying from 10 ka to ~45 ka at the study site (Table 5.1). Five out of ten
samples, however, returned ages between ~21 ka and ~25 ka. In addition, the 10-ka-
old, youngest sample (DES-11-04), and the two oldest samples (DES-10-04, ~32.6 ka
and DES-08-04, ~44.7 ka) appear to be outliers (Figure 5.6). If these three “outlier”
ages are excluded, then the average age for my seven samples is ~21.2 ka.
A number of factors such as erosion, inheritance, altitude, shielding, and latitude
can affect the age of a sample. Scaling for the altitude and latitude of the site were made
in the CRONUS age calculator [Balco et al., 2008]. The shielding (skyline) correction
factor, on the other hand, is unnecessary at my study site since the skyline lies at less
than 20 ° in all directions. Inheritance and/or erosion, however, remain as possible
causes for the scatter in the dating results.
One potential way of checking inheritance is sampling the modern drainage
clasts. This has not yet been done at the Destek site. Another way of checking inheritance
is dating samples from a depth profile. I was able to recover four depth profile samples
(DES-07A-04, DES-07B-04, DES-07C-04, and DES-07D-04) large enough to date
from the short trench T1. Samples 07C, 07D, 07A, and 07B were recovered from depths
of 43 cm, 60 cm, 70 cm, and 116 cm, respectively. The dating results of the T-1 depth
profile samples returned ages of 65.8 ka, 88.5 ka, 87.5 ka, and 75.4 ka, from shallowest
to deepest. There were not any samples large enough to date within trench T2.
5.4.3 Discussion
Two values are needed to calculate the slip-rate at Destek: (1) the displacement
measurement of the piercing line; and (2) the age of stabilization of the surface, which
provides a maximum-possible age for the incision and subsequent offset of the stream.
99
Figure 5.5 (a) Photo of a quartz gravel sampled from the surface near T1 (b) I excavated
Trench T1 in order to collect depth profile samples.
100
Table 5.1 Cosmogenic
10
Be ages from Destek shown. Ten surface samples of quartz
gravels and four depth profile samples are processed. These ages are calculated by
using constant-production-rate models [Lal, 1991; Stone, 2000]. The uncertainties
reflect only internal 1-sigma analytic uncertainties.
101
Figure 5.6 Probability distribution functions of
10
Be ages (ka) from Destek are plotted
using Camel plot. Individual samples are color coded. The error of each sample is 1
sigma and represents analytical errors only. In the camel plot diagram entire range of
possibilities is shown.
102
I use Hubert-Ferrari et al. [2002]’s average displacement measurements of ~190 m.
In contrast to the relatively straightforward of the offset measurement, the
10
Be
age data are extremely complicated and difficult to interpret. Specifically, the depth
profile ages suggest a deposit on the order of >85 ka. These ages contrast markedly
with the ages of the surface cobbles. For example, the average of all ten surface sample
ages is 23,626 years. The average of the five clustered ages is 23,155 years. Thus,
the surface samples suggest a deposit that is <~25 ka. It is difficult to reconcile these
different age data. One possibility is that a >85 ka alluvial deposit has been resurfaced
during progressive fault offset by a much younger deposit. Alternatively, it is possible
that the depth profile ages exhibit major inheritance problems. This seems unlikely,
however, because the depth profile ages are all relatively similar and this fails to explain
the much younger ages from the surface samples.
In light of these observations, I do not feel that I can generate an unequivocal slip
rate for Destek. The ~200 m offsets of Hubert-Ferrari et al. [2002] appear to be robust,
yet the complicated age data suggest the possibility that there may be alternative larger-
scale restorations. Based on the very thick argillic horizon exposed in my trenches, I
suspect that these “terraces” may in fact be something more akin to pediments, albeit
pediments eroded into older alluvium.
5.5 Slip rate site at Koçyatağı (Mihar)
Hubert-Ferrari et al. [2002] mapped a series of 150 to 220 m offsets along the
North Anatolian fault near the village of Mihar. These offsets are measured on streams
incised into an older terrace of which only small erosional remnants remain (none of
which are sufficiently well preserved for cosmogenic radionuclide dating). They also
identified a well-defined, much younger 19-28 m offset ~ 2 km west of Mihar. This
103
offset, southward-flowing incised drainage is located ~2 km south of its headwaters
(Figure 5.7). As suggested by Hubert-Ferrari et al. [2002], the offset measurement
of the western riser (28 m) is a better constraint for the true displacement because the
eastern riser is being eroded by the active drainage. The Mihar offset is the smallest
offset (Figure 5.8) that I examined during this study, and the resulting (presumably)
shorter-term rate will represent an important point of comparison in my efforts to
determine the relative constancy (or non-constancy) of strain release along the North
Anatolian fault.
I sampled the fluvial terrace (t1) associated with this offset by collecting
limestone pebbles and cobbles from both the surface and in depth profiles in two short
2.2-m-deep trenches (Figures 5.9 and 5.10). I collected a total of 20 surface samples
from both north and south of the fault. In addition to these individual pebble- and
cobble-sized limestone samples, I collected an amalgamation sample of 15 limestone
pebbles and cobbles from the surface south of the fault. I also recovered samples from
each trench exposure at every ~20 cm depth interval.
I have processed nine surface samples, along with the amalgamation sample from
the terrace surface south of the fault. I also processed a depth profile sample recovered
from the short trench south of the fault at ~100 cm depth. As with all cosmogenic
sample preparations, I performed these chemical analyses at the Lawrence Livermore
National Laboratory under the supervision of Dr. Robert Finkel. This project is ongoing
and we are currently in the process of calculating ages for this initial set of samples.
104
Figure 5.7 (a) Map showing the location of Mihar slip-rate site. (b) Oblique aerial
photo showing the Mihar area. The North Anatolian fault displaces southward-flowing
drainages incised into an extensive alluvial terrace to the west of Mihar village. Dashed
white box indicates the location of ~30 m offset mapped by Hubert-Ferrari et al.
[2002].
105
Figure 5.8 (a) Interpretation of small (19–26 m) drainage offsets in Mihar [after
Hubert-Ferrari et al., 2002, modified to show the locations of the two short trenches
that I excavated for depth profile sampling (yellow squares)]. (b) aerial view of the
offset from Hubert-Ferrari et al. [2002] showing their survey points.
106
Figure 5.9 Panoramic view of Mihar slip-rate site. View looking to the south. The short trench locations are shown with an
atomic symbol.
107
Figure 5.10 (a) Photograph of a limestone cobble sampled for
36
Cl dating taken from
the surface. View is to the north with the headwaters visible in distance (b) photo of
the short trench south of the fault. Limestone gravels were collected every ~20 cm for
depth profile modeling.
108
CHAPTER 6:
Paleoseismologic Trench Sites in Progress
6.1 Additional paleoseismologic trench sites
In this chapter I present the results of initial paleoseismic trench investigations
at two sites. The Elmalı Trench site is located on the northern part of the East Anatolian
fault. This site remains an attractive paleoseismic site suitable for future investigation.
V olkan Karabacak (Osmangazi University) provided much appreciated help in the field
during my preliminary studies at this site. The Hamamlı site, on the other hand, is
located on the North Anatolian fault. Metin Gürcan (Istanbul University) assisted with
trenching at Hamamlı site. By reporting my preliminary results at these two sites I hope
to provide useful information for future studies.
6.2 Paleoseismic trench site at Elmalı
The Elmalı trench site is located on the East Anatolian fault, ~30 km northeast
of Bingöl and ~45 km south of Karlıova (Figure 6.1). The East Anatolian fault is the
conjugate fault to the North Anatolian fault, accommodating the shortening between the
Caucasus and the Arabian Plate. It is a major left-lateral fault that allows the westward
escape of the Anatolian Block. Although it is a major fault that has produced large
historical earthquakes, there is only one paleoseismic trench site along this fault [Çetin
et al., 2003].
The most recent surface rupturing earthquake at the Elmalı site was the 1971
M6.7 Bingöl earthquake [Seymen and Aydın, 1972; Arpat and Şaroğlu, 1972]. The
trench site is located in a marshy area. The fault is oriented ~N-S, approximately
109
Figure 6.1 Simplified tectonic map of the eastern Mediterranean region [modified after
Barka, [1992] and Armijo et al., [2002]. Arrows indicate generalized relative motions
according to the fixed Eurasian Plate showing rates in millimeters per year [after
McClusky et al., [2000] and Reilinger et al., [2006]. Black boxes indicate Hamamlı and
Elmalı paleoseismologic trench site locations. NAF-North Anatolian fault, EAF-East
Anatolian fault, NEAF-North East Anatolian fault, OF-Ovacık fault, EF-Ezinepazarı
fault, DSF-Dead Sea fault. Satellite image is from Orbitview-2 satellite (available at
http://visibleearth.nasa.gov).
110
parallel to the river valley ~150 m to the west of my trench site. In an attempt to
document the ages of paleoearthquakes, I excavated a ~20-m-long, ~2-m-deep, slot
trench near Elmalı Village, north of Bingöl (Figure 6.2). The trench exposed silt- and
sand- dominated stratigraphy without distinct bedding. However, if I had had enough
time to let the trench walls dry out I may have been able to pick out fine layers within
this stratigraphy. The same conditions probably prevented the recognition of the fault
within this trench. Unfortunately, my attempt to develop this site coincided with
the second Iraq War, and the security-related instability affecting this remote region
prevented me spending more than one day at the site. However, the Elmalı trench site
remains an attractive location for future studies.
111
Figure 6.2 A panoramic view of Elmalı paleoseismologic trench site. The view is to the west. A series of back-scarps and
ponds are forming along the East Anatolian fault. The trench is located within a marshy area at the center of the image.
112
6.3 Paleoseismic trench site near Hamamlı
The Hamamlı trench site is located on the North Anatolian fault, ~10 km east of
İsmetpaşa (Figure 6.1). The Gerede River flows along the North Anatolian fault for several
kilometers in this region (Figure 6.3a). During my initial field investigation, I located
this site as a potential paleoseismic study site and proceeded with the reconnaissance
trench. At the study site, a marshy area is located south of an east-west shutter ridge
(Figure 6.3b). Wet conditions and two small alluvial fans located in the southern part of
this marsh appeared to be promising for finding well-bedded stratigraphy and datable
material for paleoseismic trenching. I excavated the trench within this marshy area, ~2
m east of the pond. The trench was 35 m long and ~2 m deep. Alluvial fan material
was exposed at the southern end of the trench with an increase in gravel size towards
the mountains in the south. The northern end of the trench, however, exposed mostly
clay with silt and some pebbles/gravels. The stratigraphy showed very poor bedding.
An uncut, well-defined pebbly layer exposed at 1.9 m depth showed no deformation.
At the time of opening the reconnaissance trench at my Hamamlı trench site there
were no topographic maps of any scale or aerial photographs available. Lack of aerial
photography for this site prevented from realizing that the fault is a few meters to the
north of my trench, just south of the shutter ridge. As a result, I was not able to expose
the fault within this trench exposure. Nevertheless, I confirmed that the active trace
of the North Anatolian fault is not within the marsh. The marsh is forming behind the
shutter ridge passively while the shutter ridge is being displaced along the fault.
113
Figure 6.3 (a) Google map image shows the location of study site (white box, Figure
6.2b). (b) Interpretation of the North Anatolian fault in the vicinity of my study site
based on Google map image. The trench location is shown in yellow.
114
CHAPTER 7:
Conclusions and Suggestions for Future Research
7.1 Conclusions
I have employed geomorphic mapping, Quaternary dating methods, and
paleoseismic trenching to generate fault rate and earthquake age data that help provide
a better understanding of the North Anatolian fault’s behavior in various temporal
and spatial scales. Specifically, I mapped offset geomorphic markers along the North
Anatolian fault and used cosmogenic nuclide (
10
Be and
36
Cl) and radiocarbon (
14
C)
dating methods to constrain the ages of these features. The results I present in this
dissertation are among the first geochronologically constrained slip-rate data for the
North Anatolian fault. Using these and other published data, I also constructed one of the
first compilations of strain-release rates for the North Anatolian fault. My compilation
of these rate data reveals a constant slip rate of ~15-20 mm yr
-1
over time scales of
10
3
-10
5
years. This result, however, is slower than the geodetically constrained slip
rate of 25 + 2 mm yr
-1
[Reilinger et al., 2007], possibly indicating a strain transient.
In addition to my slip rate studies, I performed paleoseismologic trenching on the
eastern part of the North Anatolian fault at the village of Lorut. My results from this
site demonstrate a relatively regular occurrence of large earthquakes along this stretch
of the North Anatolian fault. I attribute the relatively constant strain release rates and
regular earthquake recurrences to the mechanical simplicity of the Anatolian-Eurasian
plate boundary in northern Turkey, which is dominated by the slip on the structurally
mature North Anatolian fault.
In addition to providing new slip-rate and paleoseismological data for the North
Anatolian fault, I tested dating and trenching techniques. I employed multiple dating
115
methods within a single site and compared similar sites in order to provide a test for
various cosmogenic nuclide production rate models. In addition, the comparison of
my results from the Yaylabeli study site and Hartleb et al., [2006] Çukurçimen site
(Chapter 4) provide a significant data source for confirming the reproducibility of
paleoseismologic results.
I believe that this dissertation will be of interest to not only tectonic
geomorphologists and paleoseismologists, but also geodynamicists, geochronologists,
earthquake engineers, and natural hazard planners, both in Turkey and worldwide.
7.2 Suggestions for Future Research
The North Anatolian fault remains as an open laboratory with vast research
potential for Earth scientists. Additional paleoseismologic studies that document
earthquake occurrence information and longer-term (10
5
-10
6
years) slip-rate data would
further extend the limits of our understanding of earthquake occurrence and evolution
of associated mechanisms, such as fault loading. This is especially true for the East
Anatolian fault, where paleoseismological and geological slip-rate data are extremely
scarce. Similar data from this fault would lead to a better understanding of the possible
interactions between the East and North Anatolian faults [Hubert-Ferrari et al., 2003;
Hartleb, 2006].
116
REFERENCES
Allen, C.A., (1969), Active faulting in northern Turkey: California Institute of
Technology, Division of Geological Sciences, Contribution No. 1577, 32 p.
Ambraseys, N.N., (1970), Some characteristic features of the North Anatolian fault:
Tectonophysics, v. 9, p. 143– 165, doi: 10.1016/0040–1951(70)90014–4.
Ambraseys, N.N., (1989), Temporary seismic quiescence: SE Turkey, Geophys. Jour.
Int., v. 96, p. 311-331.
Ambraseys, N.N., (2002), The Seismic Activity of the Marmara Sea Region over the
Last 2000 Years, Bulletin of the Seismological Society of America, 92, 1, pp.
1–18.
Ambraseys, N.N., (2006), Comparison of frequency of occurrence of earthquakes with
slip rates from long-term seismicity data: the cases of Gulf of Corinth, Sea of
Marmara and Dead Sea Fault Zone, Geophysics Journal International, 165,
516–526, doi: 10.1111/j.1365-246X.2006.02858.x.
Ambraseys, N.N., and Finkel, C.F., (1995), The seismicity of Turkey and adjacent
areas: A historical review, 1500–1800: Istanbul, Muhittin Salih Eren, 240 p.
Ambraseys, N.N., and Jackson, J.A., (1998), Faulting associated with historical and
recent earthquakes in the eastern Mediterranean region: Geophysical Journal
International, v. 133, p. 390–406, doi: 10.1046/j.1365–246X.1998.00508.x.
Ambraseys, N.N., and Melville, C.P., (1995), Historical evidence of faulting in eastern
Anatolia and northern Syria: Annali di Geofisica, v. 38, no. 3–4, p. 337–343.
Argus, D.F., and Gordon, R.G., (2001), Present tectonic motion across the Coast
Ranges and San Andreas fault system in central California: Geological Society
of America Bulletin, v. 113, p.1580-1592.
Armijo, R., Meyer, B., Navarro, S., King, G., and Barka, A., (2002), Asymmetric slip
partitioning in the Sea of Marmara pull-apart: A clue to propagation processes
of the North Anatolian fault?: Terra Nova, v. 14, no. 2, p. 80–86, doi: 10.1046/
j.1365–3121.2002.00397.x.
117
Arpat, E. and Şaroğlu, F., (1972), The East Anatolian fault system: Thoughts on its
development, Bulletin Mineral Research and Exploration Institute of Turkey,
78, 33-39.
Balco, G., Stone, J.O., Lifton, N.A., Dunai, T.J., (2008), A complete and easily accessible
means of calculating surface exposure ages or erosion rates from
10
Be and
26
Al
measurements, Quaternary Geochronology, in press.
Barka, A.A., (1992), The North Anatolian fault zone: Annales Tectonicae, v. 6, p. 164–
195.
Barka A.A., (1996), Slip distribution along the North Anatolian fault associated with
large earthquakes of the period 1939-1967: Bulletin of the Seismological
Society of America, v. 86, p. 1238-1254.
Barka, A.A., (1999), The 17 August 1999 Izmit Earthquake, Science V ol. 285. (no.
5435), pp. 1858 – 1859 DOI: 10.1126/science.285.5435.1858
Barka, A.A., and Eyidoğan, H., (1993), The Erzincan earthquake of 13 March 1992
in Eastern Turkey Terra Nova 5 (2), 190–194 doi:10.1111/j.1365-3121.1993.
tb00245.x
Biasi, G., and Weldon, R.J., II, (1994), Quantitative refinement of calibrated
C-14 distributions: Quaternary Research, v. 41, p. 1–18, doi: 10.1006/
qres.1994.1001.
Bird, P., (2007), Uncertainties in long-term geologic offset rates of faults: General
principles illustrated with data from California and other western states,
Geosphere, v. 3, no; 6; p. 577-595; doi: 10.1130/GES00127.1.
Bronk Ramsey, C., (1995), Radiocarbon Calibration and Analysis of Stratigraphy: The
OxCal Program: Radiocarbon, v. 37(2), p. 425-430.
Bronk Ramsey, C., (2001), Development of the Radiocarbon Program OxCal:
Radiocarbon, v. 43 (2A), p. 355-363.
Çetin, H., Güneyli, H., Mayer, L., (2003), Paleoseismology of the Palu-Lake Hazar
segment of the East Anatolian fault zone, Turkey, Tectonophysics, 374, p. 163-
197.
118
Desilets D., and Zreda, M., (2003), Spatial and temporal distribution of secondary
cosmic-ray nucleon intensities and applications to in situ cosmogenic dating,
Earth and Planetary Science Letters, 206, pages 21-42.
Desilets D., M. Zreda, P. F. Almasi, D. Elmore, (2006), Determination of cosmogenic
36Cl in rocks by isotope dilution: innovations, validation and error propagation,
Chemical Geology, 233, p. 185-195, doi:10.1016/j.chemgeo.2006.03.001
Dolan, J.F., and Wald, D., (1998), The 1943-1953 north-central Caribbean earthquake
sequence: Active tectonic setting, seismic hazards, and implications for
Caribbean-North America plate motions: GSA Special Paper 326 Active
tectonics of the north-central Caribbean, (eds.) Dolan, J. F., and P. Mann, p.
143-169.
Dolan J. F, Bowman D. D, Sammis C. G (2007), Long-range and long-term fault
interactions in Southern California. Geology: V ol. 35, No. 9 pp. 855–858
Dor, O., Yildirim, C., Rockwell, T.K., Ben-Zion, Y ., Emre, O., Sisk, M., and Duman,
T., (2008), Geological and geomorphologic asymmetry across the rupture
zones of the 1943 and 1944 earthquakes on the North Anatolian Fault: possible
signals for preferred earthquake propagation direction, Geophys. J. Int. doi:
10.1111/j.1365-246X.2008.03709.x
Dunai, T.J., (2001), Influence of secular variation of the geomagnetic field on production
rates of in situ produced cosmogenic nuclides, Earth and Planetary Science
Letters, 193, pages 197-212.
Ergin, K., Güçlü, U., and Uz, Z., (1967), A catalog of earthquakes for Turkey and
the surrounding area: 11 A.D. to 1964 A.D.: Istanbul, Istanbul Technical
University, Faculty of Mining Engineering, 169 p.
Evans, J. M., Stone, J.O.H., Fifield, L.K., and Creswell, R.G.,(1997), Cosmogenic
chlorine-36 production in K-feldspar. Nuclear Instruments and Methods in
Physics Research: v. 123, p. 334-340.
Fuenzalida, H., Dorbath, L., Cisternas, A., Eyidogan, H., Barka, A.A., Rivera, L.,
Haessler, H., Philip, H., and Lyberis, N., (1997), Mechanism of the 1992
Erzincan earthquake and its aftershocks, tectonics of the Erzincan Basin and
decoupling on the North Anatolian Fault Geophysical Journal International
129 (1), 1–28 doi:10.1111/j.1365-246X.1997.tb00935.x
119
Fumal, T.E., Pezzopane, S.K., Weldon, R.J.II, and Schwartz, D.P., (1993), A 100-
year recurrence interval for the San Andreas fault at Wrightwood, California,
Science, v. 259, p. 199-203.
Fumal, T.E., Weldon, R.J.II, Biasi, G.P., Dawson, T.E., Seitz, G.G., Frost, W.T., and
Schwartz, D.P., (2002), Evidence for Large Earthquakes on the San Andreas
Fault at the Wrightwood, California, Paleoseismic Site: A.D. 500 to Present,
Bulletin of the Seismological Society of America, V ol. 92, No. 7, pp. 2726–
2760.
Grant, L. B., (1996), Uncharacteristic earthquakes on the San Andreas fault: Science,
v. 272, p. 826-827.
Gosse, J.C. and Phillips, F.M., (2001), Terrestrial in-situ cosmogenic nuclides: Theory
and application: Quaternary Science Reviews 20 1475-1560.
Haldon, J., (2002), Byzantium at War 600 - 1453, Oxford: Osprey, ISBN 978-
1841763606
Hartleb, R.D., Dolan, J.F., Akyüz, H.S., and Yerli, B., (2003), A 2,000 year-long
paleoseismologic record of earthquakes along the central North Anatolian
fault, from trenches at Alayurt, Turkey: Bulletin of the Seismological Society
of America, v. 93, p. 1935-1954.
Hartleb, R.D., Dolan, J.F., Kozacı, O., Akyüz, H.S., and Seitz, G.G., (2006), A 2500-
yr-long paleoseismologic record of large, infrequent earthquakes on the
North Anatolian fault at Çukurçimen, Turkey, Geological Society of America
Bulletin, v. 118; no. 7/8; p. 823–840; doi: 10.1130/B25838.1
Hirth, G., and J. Tullis, (1992), Dislocation creep regimes in quartz aggregates, Journal
of Structural Geology, v. 14, p. 145-159.
Hubert, A., Armijo, R., King, G., Gasse, F., and Barka, A.A., (1997), Slip rate of the
North Anatolian fault, Turkey: Eos (Transactions, American Geophysical
Union), v. 78, p. F715.
Hubert-Ferrari, A., Armijo, R., King, G., Meyer, B., and Barka, A.A., (2002), Morphology,
displacement, and slip rates along the North Anatolian fault, Turkey: Journal
of Geophysical Research, v. 107, p. 2235, doi: 10.1029/2001JB000393.
Johnson, R.H., Tallis, J.H., and Wilson, P., (1990), The seal edge combes, North
Derbyshire—a study of their erosional and depositional history, J. Quat. Sci.
5, 83–94.
120
Ketin, I., (1948), Ueber die tektonisch-mechanischen Folgerungen aus den grossen
Anatolischen Erdbeben des letzten Dezenniums: Geologische Rundschau, v.
36, p. 77–83, doi: 10.1007/BF01791916.
Klinger, Y ., Sieh, K., Altunel, E., Akoglu, A., Barka, A.A., Dawson, T., Gonzalez, T.,
Meltzner, A., and Rockwell, T., (2003), Paleoseismic evidence of characteristic
slip on the western segment of the North Anatolian fault, Turkey: Bulletin
of the Seismological Society of America, v. 93, no. 6, p. 2317–2332, doi:
10.1785/0120010270.
Kohl, C.P., and Nishiizumi, K., (1992), Chemical isolation of quartz for measurement of
in-situ-produced cosmogenic nuclides, Geochimica et Cosmochimica Acta,56,
3583-3587.
Kondo, H. Ozaksoy, V ., Yildirim, C., Awata, Y ., Emre, O., and Okumura, K., (2004),
3D trenching survey at Demir Tepe site on the 1944 Bolu-Gerede earthquake
ruptures, North Anatolian fault system, Turkey: Seismological Research
Letters, v. 75, no. 2, p. 292.
Kozacı, O., Dolan, J.F., and Finkel, R., (in press), Late Holocene slip rate for the North
Anatolian fault, Turkey, from cosmogenic
10
Be geochronology: Implications for
the constancy of fault loading and strain release rates, Journal of Geophysical
Research.
Kozacı, O., J. F. Dolan, R. C. Finkel, R. D. Hartleb, (2007), Late Holocene slip rate
for the North Anatolian fault, Turkey, from cosmogenic 36Cl geochronology:
Implications for the constancy of fault loading and strain release rates, Geology,
October, v. 35, no. 10; p. 867–870.
Lal, D., (1991), Cosmic ray labeling of erosion surfaces: in situ nuclide production
rates and erosion models, Earth and Planetary Science Letters, V olume 104,
Issues 2-4, pages 424-439.
Lifton N.A., Bieber, J.W., Clem, J.M., Duldig, M.L., Evenson, P., Humble, J.E., and
Pyle, R., (2005), Addressing solar modulation and long-term uncertainties in
scaling secondary cosmic rays for in situ cosmogenic nuclide applications,
Earth and Planetary Science Letters, V olume 239, Issues 1-2, 30, pages 140-
161.
121
Mann, P.,C.S., Prentice, G. Burr, L. Pena, and Taylor, F.W., (1998), Tectonic
geomorphology and paleoseismology of the Septentrional fault system,
Dominican Republic, GSA Special Paper 326 Active tectonics of the north-
central Caribbean, (eds.) Dolan, J. F., and P. Mann, p.63-124.
McClusky, S., Balassanian, S., Barka, A., Demir, C., Ergintav, S., Georgiev, I., Gurkan,
O., Hamburger, M., Hurst, K., Kahle, H., Kastens, K., Kekelidze, G., King,
R., Kotzev, V ., Lenk, O., Mahmoud, S., Mishin, A., Nadariya, M., Ouzounis,
A., Paradissis, D., Peter, Y ., Prilepin, M., Reilinger, R., Sanli, I., Seeger, H.,
Tealeb, A., Toksöz, M.N., and Veis, G., (2000), Global positioning system
constraints on plate kinematics and dynamics in the eastern Mediterranean
and Caucasus: Journal of Geophysical Research, v. 105, no. B3, p. 5695–5719,
doi: 10.1029/1999JB900351.
McGill, S., Wells, S., Fortner, S., Kuzma, H., McGill, J., (in press 2008), Slip rate of
the Western Garlock fault, at Clark Wash, near Lone Tree Canyon, Mojave
Desert, California, Geological Society of America Bulletin, in press.
McKenzie, D., (1972), Active tectonics of the Mediterranean region: Geophysical
Journal of the Royal Astronomical Society, v. 30, p. 109–185.
Meriaux, A.S., Ryerson, F.J., Tapponier, P., Van der Woerd, J., Finkel, R.C., Xu, X., Xu,
Z., and Caffee, M.W., (2004), Rapid slip along the central Altyn Tagh Fault:
Morphochronologic evidence from Cherchen He and Sulamu Tagh: Journal of
Geophysical Research, v. 109, B06401, doi: 10.1029/2003JB002558.
Okumura, K. and Kondo, H., (2006), Variability of long-term fault activity along the
North Anatolian fault, Turkey: Seismological Research Letters, v. 77, p. 250
[abstract].
Okumura, K., Rockwell, T.K., Duman, T., Tokay, F., Kondo, H., Yildirim, C., and
Ozaksoy, V ., (2003), Refined slip history of the North Anatolian fault at Gerede
on the 1944 rupture: Eos (Transactions, American Geophysical Union), v. 84,
no. 46, p. F715.
Oskin, M., and Iriondo, A., (2004), Large magnitude transient strain accumulation on
the Blackwater fault, Eastern California shear zone: Geology, v. 32, p. 313-
316.
Oskin, M., Perg, L., Shelef, E., Strane, M., Gurney, E., Singer, B., and Zhang, X.,
(2008), Elevated shear-zone loading rate during an earthquake cluster in
eastern California, Geology, in press.
122
Pantosti, D., Pucci, S., Palyvos, N., De Martini, P.M., D’Addezio, G., Collins, P.E.F,
and Zabci, C., (2008), Paleoearthquakes of the Düzce fault (North Anatolian
Fault Zone): implications for earthquake recurrence, Journal of Geophysical
Research, 113, B01309, doi:10.1029/2006JB004679
Phillips F.M. and Plummer M.A., (1996), CHLOE: A program for interpreting in-situ
cosmogenic nuclide data for surface exposure dating and erosion studies:
Radiocarbon (Abstr. 7
th
Int. Conf. Accel. Mass Spectrom.), v. 38, p. 98-99.
Phillips F. M., Zreda M. G., Flinsch M. R., Elmore D., and Sharma P., (1996),
A reevaluation of cosmogenic
36
Cl production rates in terrestrial rocks:
Geophysical Research Letters, v. 23, p. 949-952.
Phillips, F.M., Stone, W.D., and Fabryka-Martin, J.T., (2001), An improved approach
to calculating low-energy cosmic-ray neutron fluxes near the land/atmosphere in -
terface: Chemical Geology 175 689-701.
Pondard, N., Armijo, R., King G.C.P., Meyer, B., and Flerit, F., (2007), Fault interactions
in the Sea of Marmara pull-apart (North Anatolian Fault): earthquake clustering
and propagating earthquake sequences, Geophysics Journal International, 171,
1185-1197, doi: 10.1111/j.1365-246X.2007.03580.x
Pucci, S., De Martini, P.M., and Pantosti, D., (2007), Preliminary slip rate estimates for
the Düzce segment of the North Anatolian Fault Zone from offset geomorphic
markers, Geomorphology, doi:10.1016/j.geomorph.2007.09.002
Reilinger, R., McClusky, S., Vernant, P., Lawrence, S., Ergintav, S., Cakmak, R.,
Ozener, H., Kadirov, F., Guliev, I., Stepanyan, R., Nadariya, M., Hahubia, G.,
Mahmoud, S., Sakr, K., ArRajehi, A., Paradissis, D., Al-Aydrus, A., Prilepin,
M., Guseva, T., Evren, E., Dmitrotsa, A., Filikov, SV ., Gomez, F., Al-Ghazzi,
R., Karam, G., (2006), GPS Constraints on Continental Deformation in the
Africa-Arabia-Eurasia Continental Collision Zone and Implications for the
Dynamics of Plate Interactions, Journal of Geophysical Research - Solid Earth
111 (B5): Art. No. B05411.
Reilinger, R., McClusky, S.C., Oral, M.B., King, R.W., Toksöz, M.N., Barka,
A.A., Kinik, I., Lenk, O., and Sanli, I., (1997), Global positioning system
measurements of present-day crustal movements in the Arabia-Africa-Eurasia
plate collision zone: Journal of Geophysical Research, v. 102, p. 9983–9999,
doi: 10.1029/96JB03736.
123
Reimer, PJ, Baillie, M.G.L., Bard, E., Bayliss, A., Beck, J.W., Bertrand, C.J.H.,
Blackwell, P.G., Buck, C.E., Burr, G.S., Cutler, K.B., Damon, P.E., Edwards,
R.L., Fairbanks, R.G., Friedrich, M., Guilderson, T.P., Hogg, A.G., Hughen,
K.A., Kromer, B., McCormac, G., Manning, S., Bronk Ramsey, C., Reimer,
R.W., Remmele, S., Southon, J.R., Stuiver, M., Talamo, S., Taylor, F.W., van
der Plicht, J. and Weyhenmeyer, C.E., (2004), IntCal04 terrestrial radiocarbon
age calibration, 0-26 cal kyr BP, Radiocarbon, 46 (3) 1029-1058
Richter, C.F., (1958), Elementary seismology: San Francisco, W.H. Freeman, 768 p.
Rockwell, T. K., Lindvall, S., Herzberg, M., Murbach, D., Dawson, T., and Berger,
G., (2000), Paleoseismology of the Johnson Valley, Kickapoo, and Homestead
Valley Faults: Clustering of Earthquakes in the Eastern California Shear Zone:
Bulletin of the Seismological Society of America, v. 90, no. 5, p. 1200-1236.
Rockwell, T.K., Akbalık, B., Akoglu, A., Aksoy, E., Barka, A., Dier, D., Ferry, M.,
Klinger, Y ., Langridge, R., Meghraoui, M., Meltzner, A., Ragona, D., Satir, D.,
Seitz, G., and Uçarkus, G., (2001b), Results of paleoseismic studies after the
1999 Izmit earthquake:
Rockwell, T.K., Barka, A.A., Dawson, T., Akyüz, S., and Thorup, K., (2001a),
Paleoseismology of the Gazikoy-Saros segment of the North Anatolia fault,
northwestern Turkey: Comparison of the historical and paleoseismic records,
implications of regional seismic hazard, and models of earthquake recurrence:
Journal of Seismology, v. 5, p. 433–448, doi: 10.1023/A:1011435927983.
Rockwell, T.K., S. Lindvall, T. Dawson, R. Langridge, W. Lettis, Y . Klinger, (2002),
Lateral Offsets on Surveyed Cultural Features Resulting from the 1999 İzmit
and Düzce Earthquakes, Turkey, Bulletin of the Seismological Society of
America; February 2002; v. 92; no. 1; p. 79-94
Schoute, J.F.T., Griede, J.W., Mook, W.G., and Roeleveld, W., (1981), Radiocarbon
dating of vegetation horizons, illustrated by an example from the Holocene
coastal plain in the northern Netherlands, Geologie en Minjinbouw 60, 453–
459.
Schoute, J.F.T., Mook, W.G., and Streurman, H.J., (1983), Radiocarbon dating of
vegetation horizons: methods and preliminary results, in Proceedings of the
1st International Symposium on 14C and Archaeology, PACT, W. G. Mook
and H. T. Water (Editors), 295–311.
124
Schwartz, D. P., and Coppersmith, K., (1984), Fault behavior and characteristic
earthquakes -- Examples from the Wasatch and San Andreas fault zones:
Journal of Geophysical Research, v. 89, p. 5681-5698.
Seymen, I. and Aydın, A., 1972, The Bingol earthquake fault and its relation to the North
Anatolian fault zone, Bulletin Mineral Research and Exploration Institute of
Turkey, 79, 1-8.
Sieh, K.E., and Jahns, R.H., (1984), Holocene activity of the San Andreas fault at
Wallace Creek, California: Geological Society of America Bulletin, v. 95, p.
883-896.
Sieh, K.E., Stuiver, M., and Brillinger, D., (1989), A more precise chronology of
earthquakes produced by the San Andreas fault in southern California: J.
Geophys. Res., v. 94, p. 603-623.
Stein, R.S., Barka, A.A., and Dieterich, J.H., (1997), Progressive failure on the North
Anatolian fault since 1939 by earthquake stress triggering: Geophysical Journal
International, v. 128, p. 594–604.
Stone, J.O., Allan, G.L., Fiefield, L.K., and Cresswell, R.G., (1996), Cosmogenic
chlorine-36 from calcium spallation. Geochimica et Cosmochimica Acta: v.
60(4), p. 679-692.
Stone, J.O., (2000), Air pressure and cosmogenic isotope production, Journal of
Geophysical Research, 105, (B10), 23753-23759.
Stone J.O.H., Evans J.M., Fifield L K., Allan G.L., and Cresswell R.G., (1998),
Cosmogenic chlorine-36 production in calcite by muons: Geochimica et
Cosmochimica Acta 62 433-454.
Sugai, T., Emre, Ö., Duman, T., Yoshioka, T., and Kusçu, I., (1998), Geologic evidence
for five large earthquakes on the North Anatolian fault at Ilgaz, during the last
2000 years; a result of GSJ-MTA international cooperative research: Interim
Report on Active Fault and Paleoearthquake Research in the 1998 Fiscal Year:
GSJ Interim Report no. EQ, v. 99, no. 3, p. 263–273.
Şengör, A.M.C., (1979), The North anatolian transform fault: Its age, offset and tectonic
significance, Journal of the Geological Society of London, v. 136, p. 269-282.
125
Toksöz, M.N., Shakal, A.F., and Michael, A.J., (1979), Space-time migration of
earthquakes along the North Anatolian fault zone and seismic gaps: Pure and
Applied Geophysics, v. 117, p. 1258–1270.
Wang, Q., Zhang, P., Freymueller, J. T., Bilham, R., Larson, K. M., Lai, X., You,
X., Niu, Z., Wu, J., Li, Y ., Liu, J., Yang, Z., and Chen, Q., (2001), Present-
day crustal deformation in China constrained by Global Positioning System
measurements: Science, v. 294, p. 574-577.
Weldon, R. J., and K. E. Sieh, (1985), Holocene rate of slip and tentative recurrence
interval for large earthquakes on the San Andreas fault, Cajon Pass region,
southern California, Geol. Soc. Amer. Bull., v. 96, p. 793-812.
Weldon, R. J., Scharer, K. M., Fumal, T. E., and Biasi, G., (2004), Wrightwood and
the earthquake cycle; what a long recurrence record tells us about how faults
work: Geological Society of America Today, v. 14, p. 4-10.
Weldon, R. J., T. E. Fumal, T. J. Powers, S. K. Pezzopane, K. M. Scharer, and J. C.
Hamilton, (2002), Structure and earthquake offsets on the San Andreas fault
at the Wrightwood, California, paleoseismic site: Bull. Seism. Soc. Amer., v.
92, p. 2704-2725.
Yoshioka, T., Okumura, K., Kusçu, I., and Emre, Ö., (2000), Recent surface faulting of
the North Anatolian fault along the 1943 Ladik earthquake ruptures: Bulletin
of the Geological Survey of Japan, v. 51, no. 1, p. 29–35.
126
APPENDIX A:
Background Information on Geomorphology and Age Control of the
Eksik Study Site
A1. Geomorphologic setting of the Eksik Study Site
The Eksik study site lies along the central part of the North Anatolian fault
(Figure A1). This reach of the fault ruptured most recently during the 1943 M
w
7.7
Tosya earthquake. The surface trace in the Eksik area is generally linear, as can be seen
in figures A2 and A3. Figure A4 shows a general view of the study site looking towards
the south-southwest, with the fault extending across the site as a narrow, well-defined,
west-southwest-trending zone. The offset fluvial terrace that forms the basis for our slip
rate determination (highlighted in green on Figure A4) is composed of packed, white
limestone cobbles and large pebbles derived from limestone cliffs along the southern
face of Büyük Dağ (“Big Mountain” in Turkish) 1.0-1.5 km north of the study site
(Figures A2 and –3). The surface of this fluvial terrace, which is quite planar, has been
incised ~20 m by active streams that flow southward near the western and eastern
edges of Ağılönü Canyon. These steep, high-energy mountain streams currently carry
as bed load cobbles and large pebbles similar to those that make up the offset terrace.
Figure A5 shows our model for the sequential development of the mid-canyon terrace
remnants (also shown as an inset in Figure 2.2 of main text). Note that during right-
lateral slip on the North Anatolian fault the southeastern corner of the terrace remnant
north of the fault and the northwestern corner of the terrace remnant south of the fault
will be exposed to enhanced erosion by the modern, post-terrace streams. This model
predicts that continued offset will yield rhomb-shaped terrace remnants that are very
127
similar to those observed. The final drawing shows capture of the headwaters of the
western stream by the eastern stream. This has occurred very recently at the study site,
as shown by the extremely steep dip of the captured part of the stream system (see
topographic map in Figure 2.2 of main text).
Projection of the surface slope of the northern terrace fragment southward
across the fault indicates 3.5 ± 0.5 m of north-side-up separation. Relative to our offset
measurement of 46 ± 10 m, this yields a strike-slip: reverse slip ratio of ~13:1, indicat-
ing predominantly right-lateral strike-slip motion along the fault at Eksik, with a minor
component of north-side-up contraction.
A2. Discussions on offset measurement
There are two different offset features, of very different ages, present in Ağılönü
Canyon; (1) a major bedrock canyon that has been offset ~200 m [Hubert-Ferrari et
al., 2003]; and (2) a much younger, ~46-m-offset of the inner edge of a late Holocene
fluvial terrace (the focus of this study) that was deposited in a canyon incised into the
base of the larger, older canyon.
The geomorphic evolution of the study site started with incision of the major
bedrock canyon, inferred by Hubert-Ferrari et al. [2003] to have occurred ~10 ka in
response to climatic changes at the end of the last glaciation. Following incision, this
major canyon began to be offset by the NAF. Much later, probably during mid- or late
Holocene time, a second downcutting event resulted in the development of a linear,
incised valley inset into the base of the partially offset, older early Holocene canyon.
This second incision event was followed by an aggradational pulse that backfilled the
younger inset canyon with fluvial gravels, forming a flat-topped, planar surface. It is
128
this surface that we have dated in this study. Subsequent incision of this surface by
the currently active drainages resulted in preservation of the fluvial terrace remnants
described in this study. In light of the extremely steep stream gradient (>10º in the 60
m upstream of the fault, increasing gradually to >20º one km upstream) and resulting
high-energy nature of these drainages, we think it likely that incision began soon after
stabilization of the aggradational terrace surface. As discussed in the main text and
below, the similar offsets of the buttress unconformity at the terrace inner edge and the
incised, currently active drainages argue strongly for near-simultaneous terrace surface
stabilization and subsequent incision.
The focus of this study is the younger, ~46 m offset of the late Holocene fluvial
terrace. Specifically, we used the well-defined location of the inner edge of the terrace
along the western edge of the canyon, in combination with the geometries of the currently
active streams and the remaining terrace remnants, to determine the displacement along
the North Anatolian fault at the Eksik site. The buttress unconformity at the terrace inner
edge provides an easily recognized geomorphic feature that allows precise estimation
of the offset. Moreover, because in most cases colluvial deposits cover the terrace inner
edge at the Eksik site, it has been protected from post-terrace-abandonment erosion. In
general, our mapping reveals that the terrace inner edge along the western margin of
Ağılönü Canyon is relatively linear for several hundred meters both to the north and
south of the fault, although the trend of the inner edge exhibits a gentle curve centered
~40 m north of the fault. Remnants of the terrace surface preserved along the edges of
the canyon and within the canyon are all at approximately the same elevation (relative
to downstream position), indicating that the fluvial terrace was planar prior to incision
by the current active streams.
In order to precisely determine the offset of the terrace inner edge, we
129
Figure A1 Regional tectonic map showing major active structures [modified from
Barka, 1996]. Eksik study site along the central North Anatolian fault denoted by
white circle. Large arrows show GPS rates generalized from McClusky et al. [2000].
Background satellite image provided by NASA Earth Observatory. DSF=Dead Sea
fault; EAF=East Anatolian fault.
130
Figure A2 Corona satellite image showing the region surrounding the Eksik study
site. (a) Uninterpreted Corona image. (b) Corona image showing the trace of the North
Anatolian fault, and part of the Çankırı-Kastamonu highway. Note location of Alıç
paleoseismologic study site of Sugai et al., [1998].
131
Figure A3 Three-dimensional virtual view of the North Anatolian fault in the region
near Eksik looking ENE. The image was created by draping a Landsat image over
90 m DEM (Digital Elevation Model). No vertical exaggeration. Büyük Dağ (“Big
Mountain”) is the source of the limestone gravels that form the fluvial terrace discussed
in this paper.
132
created a detailed (50 cm contour interval) topographic map and surveyed all the
major topographic and geologic features at the study site with a laser-range-finding
theodolite with digital compass attachment. Specifically, we surveyed the limits of the
preserved terrace surface remnants, the precisely determined positions of the buttress
unconformity at the terrace inner edge revealed by both natural and trench exposures,
the locations of incised, active drainages, the positions of all of our trenches and test-
pits, and the locations of all geologic features that provide constraints on the location
of the terrace gravels (e.g., bedrock exposures at the expected elevation of the terrace
deposits, particularly the terrace surface).
On the detailed topographic map of our study site, shown in Figure 2.2, the
exact locations of the terrace inner edge are plotted as circles at locations where the
buttress unconformity between the terrace deposits and the underlying bedrock-derived
colluvium was exposed directly, either in natural, stream-cut exposures or our trenches.
Other geomorphic and geologic features that provide limits on the location of the terrace
inner edge, such as exposures of bedrock or bedrock-derived colluvium at the elevation
of the terrace deposits, are shown with arrows pointing towards the contact.
A2.1 Measurement of the western terrace inner edge
North of the fault (extending north from near the northwest corner of Figure
2.2), the western inner edge of the terrace can be traced along a linear, south-flowing
drainage that has incised along the buttress unconformity between the underlying
bedrock-derived colluvium and the upper surface of the onlapping terrace gravels.
Bedrock and bedrock-derived colluvium are exposed at the elevation of the terrace
surface to the west of the drainage. To the east of the drainage, however, a narrow, linear
133
Figure A4 A panoramic view of the Eksik slip rate site looking south-southwest across the North Anatolian fault (NAF)
(red) and dissected and offset fluvial terrace discussed in the paper. The bright green overlay shows the extent of the fluvial
terrace deposits.
134
Figure A5 A simple model suggesting the sequential evolution of the geometry of
the terrace during recurrent displacements along the North Anatolian fault. Note the
similarity between the final stage of the model and the mapped mid-canyon terrace
remnants.
135
ridge composed of fluvial gravels locally preserves the planar terrace surface along its
crest. The terrace inner edge lies between these two continuous exposures, within the
~10-m-deep incised drainage. This ~S20°E-trending, linear relationship can be traced
northward for >200 m from the northernmost direct exposure of the terrace inner edge
~130 m north of the fault. Between the southern end of this linear drainage and the
fault, the terrace inner edge is exposed in several natural, east-flowing side drainages,
as well as within trench T2 (Figure 2.2). Trench TP2-06 is located 20 m north of the
fault, about halfway between trench T2 and the fault. This 5-m-deep trench extended
below the expected depth of the terrace surface at this location, as projected from the
planar, mid-canyon terrace surface, yet exposed only bedrock-derived colluvium. If we
have correctly projected the expected terrace elevation from the mid-terrace remnant to
the site of trench TP2-06, this would indicate that the terrace inner edge lies to the east
of trench TP2-06. In addition, we excavated trench TP1-06 at a lower elevation ~15 m
to the east of TP2-06. Trench TP1-06 exposed a section of the basal contact between
terrace fill deposits and underlying bedrock-derived colluvium. The base contact of the
terrace gravels is ~2-3 m below the expected elevation of the terrace surface, indicating
that the terrace inner edge lies to the west of trench TP1-06. Based on this constraint,
and the exposure in trench T2, the trend of the inner edge must be more easterly than
S34ºE, indicating that the inner edge exhibits a gentle downstream change in trend
from S20ºE to ~S35-40ºE south of trench T2, as it approaches the fault. Thus, the trend
of the inner edge just north of the fault is similar to the linear, S40ºE trend of the inner
edge south of the fault.
South of the fault, right-lateral offset of the east-facing, western side of Ağılönü
Canyon has resulted in the development of a steep, ~15-m-high, south-facing fault
scarp (as noted in the main text, true vertical uplift of the terrace surface is only ~3.5
136
m). Colluvium derived from this fault scarp has deeply buried the terrace inner edge
immediately south of the fault, such that we could not expose it in our deepest trenches
within ~35 m of the fault. The northernmost direct exposure of the terrace inner edge
south of the fault is in trench T3, which is located ~2 m south of the break in slope at
the base of the fault scarp (Figure 2.2). The buttress unconformity at the terrace inner
edge exposed in trench T3 trends S40°E. Forty meters south of T3, the terrace inner
edge is exposed in two southeast-flowing drainages. Fifty meters south of these natural
exposures, the inner edge is not covered by colluvium and can be mapped directly
at the surface for several hundred meters downstream. Collectively, these exposures
reveal a markedly linear, S40°E trend for the terrace inner edge south of the fault.
Although the area between the fault and trench T3 is covered by thick, fault-
scarp-derived colluvium, a bedrock outcrop ~5 m south of the fault provides a constraint
on the orientation of the terrace inner edge south of the fault. Specifically, the terrace
inner edge just south of the fault cannot lie more than ~9 m west of the projection of the
terrace inner edge based on its S40°E trend and location in and south of T3.
Taking all of these observations into consideration, we examined a range of
potential offset reconstructions. As shown in Figure 2 in the main text and in Figure A6,
restoration of 46 m of right-lateral offset yields a geometrically simple alignment of
the western, terrace inner edge. The bedrock exposure 5 m south of the fault, together
with the projection of the terrace inner edge north of the fault, suggest a maximum-
possible offset of ~55 m. The minimum-possible offset of the inner edge is not tightly
constrained by available data.
In addition to our observations of the terrace inner edge, the incision associated
with the active eastern and western drainages, as well as the linear western edge of the
mid-canyon terrace remnant, provide additional, independent estimates of fault offset.
137
North of the fault, the western modern drainage has incised a 30-m-wide, S30ºE linear
canyon. Similarly, downstream of the ~50 m-wide zone of colluvial deposition off the
south-facing fault scarp, this drainage has cut a linear, S30ºE-trending canyon (Figure
2.2). The eastern modern drainage has incised a similar, S15ºW canyon north of the
fault, which gently curves downstream to a S20ºE trend south of the fault. Moreover,
the western edges of the terrace remnants both north and south of the fault exhibit
similar, linear S30ºE trends.
As shown in Figure 2.2 in the main text and in Figure A6, our preferred 46 m
offset, which is based on restoration of the western terrace inner edge, also restores the
canyons incised by the currently active drainages. Moreover, this preferred reconstruction
also approximately aligns the linear, S30ºE-trending western edges of the mid-canyon
terrace remnants. Figure A6 illustrates restoration of these geomorphic features at five
different offsets ranging from 36 m to 56 m. This range is based on our preferred,
46 m offset, together with symmetric epistemic ±10 m uncertainty. The maximum
potential offset (preferred 46 ± 10 m) is based on a conservative amplification of the 55
m estimate of the maximum possible offset described above. The minimum possible
offset is not as well constrained but, as shown in Figure A6 a suggested minimum
offset of ~36 m (based on application of symmetric ±10 m error estimates) yields a
geomorphically less likely reconstruction that does not properly align any of the major
features discussed above.
A3. Data from Fault-Parallel Trench
We excavated four fault-parallel trenches at the Eksik site to locate more accu-
rately the terrace inner edges where they are not clearly exposed at the surface (Figure
A7). The trenches were located as close as possible to the fault zone within the limita-
138
tions of topography and thickness of overlying colluvium (see discussion below) (Fig-
ure 2.2). Trenches T1, T2 and T3, which were excavated by hand and trenches TP1-06
and TP2-06, which were excavated by backhoe, were located on the western side of
Ağılönü Canyon. Trench T4, which was excavated by backhoe, is located on the east- ğılönü Canyon. Trench T4, which was excavated by backhoe, is located on the east- Canyon. Trench T4, which was excavated by backhoe, is located on the east-
ern terrace inner edge, north of the fault.
Trench T1, which was 6 m long and 50 cm wide, was excavated 22 m north of
the fault on the western inner edge of the terrace. Trench T1 exposed bedrock at ~60
cm depth, indicating that the western inner edge of the terrace deposits is located to
the east of the trench. This trench thus provides a limiting data point for the location of
the western terrace inner edge north of the fault (shown by eastward-pointing arrow in
Figure 2.2).
Trench T2 was hand-excavated 10 m to the north of T1 and ~35 m north of
the fault. The trench was 4.5 m long and 2.5 m deep. The western inner edge of the
terrace gravels was exposed in three dimensions at 1.8 m depth, one meter from the
western end of the trench (Figure A7). The terrace deposits are identical to those
exposed elsewhere along Ağılönü Canyon, consisting of white- to pale-gray limestone
cobbles and large pebbles. The limestone clasts are angular-sub angular, consistent
with relatively limited transport from the limestone cliffs 1.0-1.5 km upstream of the
site. In trench T2, the terrace gravels are overlain along a planar contact by massive,
dark-brown colluvium with bedrock pebbles in a silty clay matrix. Similar colluvium
underlies the terrace gravels. The colluvium above the terrace gravels is overlain by
coarser-grained, massive colluvium composed of abundant schist pebbles and angular
to sub-angular schist boulders set in a silty sand matrix. The shallowest unit exposed in
the trench is the 10- to 30-cm-thick weakly developed surface soil.
139
Trench T3 was hand-excavated parallel to, and 38 m south of the fault on
the western inner edge. The trench was sited at the break in slope at the base of the
prominent south-facing fault scarp. The trench was 10 m long and 2.5- to 2.8-m-deep.
The surface soil is very thin (< 5 cm) and very weakly developed. The weak surface soil
has developed down into dark brown, silty-sand colluvium that contains local bedrock
clasts. As shown in Figure 2.3, this massive colluvium overlies the terrace gravels along
an irregular contact that contrasts markedly with the near-horizontal upper contacts
of the terrace gravels observed in the other three trenches and in the exposures of
the terrace deposits throughout Ağılönü Canyon. Specifically, the upper surface of the
terrace gravels slopes sharply downward to the west, at a pronounced angle relative to
near-horizontal layering within the gravels. We interpret these relationships as being
indicative of erosion of the upper surface of the western-most part of the terrace gravels
at the site of T3, probably during landsliding off the steep, south-facing fault scarp
immediately north of the trench. The western-most exposure of the terrace gravels
is located at meter 3 in the trench, at a depth of 2 m. In contrast, the terrace gravels
extend as shallow as 0.5 m depth at the eastern end of trench T3. In our reconstructions
of the offset of the inner edge of the terrace gravels, we have used the actual western-
most limit of the gravels at meter 3 in our minimum offset calculation, even though
we strongly suspect that the original terrace inner edge was located farther west. By
extrapolating the upper and lower contacts of the terrace gravels exposed in T3, we
can estimate the approximate, pre-erosion location of the inner edge at ~meter 0 at the
location of T3. We use this extrapolated location for our calculation of the preferred
displacement of the western terrace inner edge.
Trench TP1-06 was excavated on the southwest-facing-slope along the western
side of Ağılönü Canyon, parallel to and north of the fault. TP1-06, which was 10 m long
140
Figure A6 Offset reconstructions of four geomorphic features; (1) western inner edge
of the youngest terrace deposits, (2) and (4) the active flood plain, and (3) the edges of
the flat-topped mid-canyon terrace remnant with 5 m increments. Please note that the
46 m preferred offset reconstruction is based on the precisely determined positions of
the buttress unconformity at the terrace inner edge west of the canyon (1).
141
and ~2 m deep, was excavated to provide an additional constraint on the terrace inner
edge location between trench T2 and the fault. A section of the basal contact between
terrace fill deposits (white limestone cobbles and small boulders) and the underlying
bedrock-derived colluvium (dark gray colluvium containing highly altered, dark-green
schist clasts) was exposed within TP1-06. The surface of this contact, however, is ~2 m
below the expected elevation of the upper surface of the fluvial terrace gravels. Thus,
data from TP1-06 indicates that the inner edge of the youngest terrace deposits is to the
west of the trench.
Trench TP2-06 was excavated on the east side of Ağılönü Canyon, parallel
to and 20 m north of the fault. TP2-06, which was 2 m long and ~5 m deep, exposed
only highly altered, dark green bedrock and bedrock derived colluvium. The trench
extended below the expected depth of the terrace surface at this location, as projected
from mid-canyon, planar terrace surface. The absence of terrace gravels in TP2-06 thus
indicate that edge of the youngest terrace deposits is located to the east of the trench.
Trench T4 was excavated on the eastern side of Ağılönü Canyon, parallel to
and 8 m north of the fault. The trench was 12 m long and 2.5 m deep. In contrast to
the exposures on the western side of Ağılönü Canyon described above, the ground
surface on the east side of Ağılönü Canyon has been plowed during crop cultivation.
The uppermost unit in this trench is a massive plow zone composed of silty-sandy dark
brown soil containing bedrock clasts. The terrace deposit consists of a relatively thin
gravel sheet that pinches out eastward at m1.5 and thickens westward from ~20 cm at
meter 2 to 40-55 cm at the west end of the trench (m 12). The top of the terrace deposit is
relatively planar and nearly horizontal. The terrace gravels are similar to those exposed
elsewhere at the site, consisting of angular to sub angular, pale gray-white limestone
cobbles and large pebbles. The youngest terrace gravel unit is underlain by bedrock-
142
derived colluvium with a silty-sandy clay matrix. A thin (~10 cm thick) yellowish-
brown, silty clay lens within this unit that extends from meter 3 to 10. An older gravel
unit is exposed from meter 10.8 to the western end of the trench. In marked contrast to
the terrace gravels, this unit includes both limestone and bedrock-derived clasts. We
interpret this deeper gravel as an older phase of fluvial deposition in which colluvium
derived from bedrock slopes ~70 m to the north of the trench was mixed with river-
borne limestone gravels prior to deposition at the trench site.
A4.
36
Cl Surface Exposure Dating Results
The
36
Cl ages for the ten limestone cobbles collected from the surface of the offset
fluvial terrace were calculated using the program CHLOE (CHLOrine-36 Exposure
age) [Phillips and Plummer, 1996](Figure A8, Tables A1 and A2). This version of
CHLOE employed the thermal and epithermal neutron distribution equations of
[Phillips et al., 2000] and production of
36
Cl by muons according to [Stone et al., 1998].
The
36
Cl production parameters of [Phillips et al., 1996] were used, as corrected by
[Phillips et al., 2001] for the new neutron distribution equations and the incorporation
of production from muons. Production rates of
36
Cl from Ca and K were determined by
principal factor analysis of whole rock data [Phillips et al., 1996] and by measurement
of mineral separates [Stone et al., 1996] and [Evans et al., 1997]. The values of the
three critical production parameters were 66.8 atoms
36
Cl (g Ca)
-1
yr
-1
, 154 atoms
36
Cl
(g K)
-1
yr
-1
, and 626 epithermal neutrons (g air)
-1
yr
-1
, based on Phillips et al. [1996].
We also calculated the ages of the
36
Cl samples using the production rates of Stone et
al. [1996] for Ca , (48.4 atoms 36ck (g Ca)
-1
yr
-1
), and Evans et al. [1997] for K, (170
atoms
36
Cl (g K)
-1
yr
-1
(Table A4). Snow shielding, shielding by surrounding topography,
and effects of non-horizontal surfaces were negligible and no correction was made
143
Figure A7 Logs of trenches T2, T3, T4 and TP2-06. Terrace gravels are shown in
yellow and yellow-orange. Olive green denotes bedrock-derived colluvium. C1 and C2
represent different bedrock-derived colluvium units within T-2 based on change in clast
size. Dashed lines on logs of trench T-3 show the probable original extent of the terrace
inner edge. This interpretation is based on the unusual west-dipping upper surface of
the terrace deposits and trucation of sub-horizontal layering within the terrace gravels,
which suggest that the upper contact of the terrace gravels has been disturbed, probably
by shallow landsliding. Although the extrapolated terrace inner edge location at m 0
is used for our preferred offset measurement, the actual measured minimum exposed
extent of the inner edge (i. e. m 2.5) is used for the calculation of the minimum offset
discussed in the text. Trench T1 and TP2-06 exposed only dark-grayish-green bedrock-
derived colluvium and schist bedrock Therefore, the logs for these trenches are not
shown.
144
for these parameters. All analytical uncertainties are reported as plus-or-minus one
standard deviation and incorporate only the reported AMS analytical uncertainty in the
36
Cl measurement. Consideration of all sources of uncertainty would probably result in
10% to 15% standard deviations [Phillips et al., 1996]. A more complete description of
the dating methodology and shielding calculations can be found in Gosse and Philips
(2001). The relevant exposure conditions and chemical compositions of the samples
are shown in Table A1 and A2, respectively.
A5. Radiocarbon Ages
A total of eight charcoal samples were radiocarbon dated to independently
constrain the age of the offset Eksik terrace deposits (Table A3 and 4). Five of these
samples were collected from trench T4 along the east side of Ağılönü Canyon, two were
collected from trench T2, and one was collected from T-3, both along the west side of
the canyon. All samples were dated at the Center for Accelerator Mass Spectrometry
(CAMS) at Lawrence Livermore National Lab. All samples were calibrated using the
program OxCAL [Bronk Ramsey, 1995; 2001, using atmospheric data from Reimer et
al., 2004]. OxCAL is available online at: http://c14.arch.ox.ac.uk/oxcal.php.
As discussed in the main text, we recovered three samples (EX-4-1, EX-4-3,
and EX-4-4) from beneath the youngest terrace gravels. As can be seen from the trench
log of trench T4 (Figure A7), these samples were collected from beneath the thin
sheet of limestone cobbles and pebbles that represent the youngest phase of fluvial
gravel deposition, and hence the age of initial stabilization of the terrace surface,
along the eastern edge of Ağılönü Canyon. The radiocarbon ages for these samples
were in correct stratigraphic order, with the shallowest samples (EX-4-1 and EX-4-4)
145
Sample Lab Latitude (N)* Longitude (E)* Elev (m)* Rock Type Thickenss (cm)
# ID (d m.mmm) (d m.mmm) (cm)
36
Cl-1 TE-S-10 41 01.158 33 40.199 1453 Limestone 5
36
Cl-2 TE-S-9 41 01.134 33 40.206 1449 Limestone 5
36
Cl-3 TE-S-8 41 01.120 33 40.211 1442 Limestone 4
36
Cl-4 TE-S-7 41 01.112 33 40.209 1440 Limestone 5
36
Cl-5 TE-S-6 41 01.073 33 40.204 1426 Limestone 5
36
Cl-6 TE-S-5 41 01.062 33 40.206 1426 Limestone 5
36
Cl-7 TE-S-1 41 01.062 33 40.208 1427 Limestone 5
36
Cl-8 TE-S-2 41 01.056 33 40.210 1426 Limestone/Marble 6
36
Cl-9 TE-S-3 41 01.052 33 40.214 1424 Limestone 5
36
Cl-10 TE-S-4 41 01.040 33 40.226 1420 Limestone 4
Eksik, Turkey Surface Samples (
36
Cl)
Table A1 Location and size of
36
Cl surface samples from Eksik study site.
146
36
Cl/Cl (at/at in rock) Cl (ppm)
Sample Sample Wt Carrier Wt Carrier Conc
36
Cl/
35
Cl SiO
2
TiO
2
Al
2
O
3
Fe
2
O
3
-T MnO MgO CaO Na
2
O K
2
O P
2
O
5 LOI Total Ba Rb Sr Pb Th U B Gd
g g mg/g (10
-15
) at/at in rock ppm % % % % % % % % % % % % ppm ppm ppm ppm ppm ppm ppm ppm
TE-S-1 150.06 3.023 0.993 94 165 20.3 21.55 0.02 0.81 0.53 0.07 0.83 40.61 0.12 0.08 0.03 35.48 100.13 66 ND 2865 ND ND 1 8.2 0.7
TE-S-2 111.40 2.992 0.993 99 218 18.8 45.41 0.02 0.88 0.55 0.08 0.44 27.67 0.14 0.08 0.02 24.96 100.25 63 ND 2335 ND 2 ND 16.5 1.5
TE-S-3 150.67 3.014 0.993 148 191 38.5 11.87 0.02 0.71 0.16 ND 0.74 47.64 0.06 0.04 0.02 39.27 100.53 68 ND 2511 ND ND 1 8.4 0.7
TE-S-4 150.50 3.002 0.993 201 325 23.8 5.58 0.02 1.33 0.08 0.01 0.59 49.31 0.06 0.05 0.02 41.78 98.83 ND ND 2424 ND ND 1 1.4 ND
TE-S-5 146.50 3.001 0.993 175 230 37.6 3.94 0.02 0.97 0.84 0.12 0.88 50.44 0.14 0.05 0.02 41.91 99.33 ND ND 3760 ND ND 1 3.1 ND
TE-S-6 96.05 3.035 0.993 168 301 31.0 14.32 0.04 1.23 0.26 0.01 0.71 44.87 0.08 0.17 0.05 38.32 100.06 60 6 2542 ND ND 2 10.2 0.8
TE-S-7 150.50 3.029 0.993 179 250 32.3 4.66 0.02 0.14 0.11 ND 0.72 51.83 0.06 0.04 0.02 42.23 99.83 ND ND 2873 ND ND 3 2.1 ND
TE-S-8 150.18 3.018 0.993 203 266 37.5 3.79 0.02 1.09 0.2 ND 0.85 50.5 0.06 0.04 0.02 42.48 99.05 76 ND 2812 ND ND 1 2.1 ND
TE-S-9 150.40 3.023 0.993 170 209 43.6 2.54 0.02 0.69 0.96 0.11 1.11 52.19 0.14 0.04 0.03 42.35 100.18 44 ND 2316 ND ND 1 5.8 1.0
TE-S-10 150.30 3.020 0.993 222 317 30.8 3.44 0.01 0.59 0.09 ND 1.16 51.01 0.06 0.05 0.04 42.65 99.1 ND ND 2511 ND ND ND 7.6 ND
Major Elements Trace Elements
EKSIK, TURKEY
36
Cl Surface Samples - Chemical Compositions
Table A2 Chemical compositions of
36
Cl samples. Note that major elements are in weight percent. Trace elements are in parts
per million. LOI is loss on ignition. Fe
2
O
3
-T is total iron expressed as Fe
2
O
3
. ND is below the lower limit of determination.
147
Figure A8 An example of the sampled limestone cobbles collected for surface exposure
dating. The samples were collected where they were embedded in a pavement composed
of small pebbles and soil. They have smoother and soil-stained lower surfaces indicating
little or no movement during their exposure history. Dashed line shows location of the
sample before it was collected.
148
yielding overlapping, calibrated dates of 2.81-2.92 ka and 2.42-2.92 ka, respectively
(Figure A7 and 8). Sample EX-4-3, collected 95 cm beneath EX-4-4, yielded a slightly
older calibrated age of 2.92-3.22 ka. In addition to these three dates from beneath
the shallowest terrace gravels, we also recovered charcoal samples from within, and
just above, the terrace deposits. The only radiocarbon sample recovered from within
the terrace gravels (EX-4-2) yielded an anomalously old, calibrated calendric age of
3.50-4.05 ka, reflecting reworking. However, a charcoal sample recovered from 20
cm above the top of the shallowest terrace gravels in trench T4 (EX-4-5) yielded a
calibrated age of 2.06-2.12 ka. The consistency of these radiocarbon dates (with the
exception of sample EX-4-2 from within the gravels) indicates minimal reworking,
suggesting that these samples may provide a reliable estimate of the age of the colluvial
deposits that immediately over-and underlie the youngest, shallowest part of the terrace
gravels. These radiocarbon age data thus provide an independent constraint on the age
of the terrace surface and on the general validity of the
36
Cl production rates used in the
cosmogenic age calculations discussed in the main text.
In addition to the five samples dated from trench T4, we also collected two
samples from trench T2. Both were collected from near the western inner edge of the
terrace. Sample EX-2-2 was collected from colluvium 10 cm below the base of the ter-
race gavels, whereas EX-2-1 was collected from colluvium 20 cm above the top of the
gravels, at a depth of 1.4 m (Figure A7). Both of these samples yielded much older ages
than the
14
C samples recovered from T4, suggesting that the T2 samples are reworked
from older deposits. Specifically, sample EX-2 –2 yielded a calibrated age of 5.00-5.48
ka, and sample EX-2-1 yielded a calibrated age of 4.53-4.93 ka. We also recovered one
sample from trench T3. Sample EX-3-1 was collected at 85 cm depth, from colluvium
just above the terrace gravels (Figure A7). However, as discussed above, we suspect
149
CAMS # Sample Sample G 13
C fraction ± D
14
C ±
14
C age ±
Name Location Modern
118239 EX-4-1 Trench T4 -25 0.7149 0.0027 -285.1 2.7 2695 35
120517 EX-4-2 Trench T4 -25 0.6503 0.0054 -349.7 5.4 3460 70
118240 EX-4-3 Trench T4 -24.5 0.6999 0.0038 -300.1 3.8 2865 45
120518 EX-4-4 Trench T4 -25 0.7244 0.0066 -275.6 6.6 2590 80
118241 EX-4-5 Trench T4 -24.8 0.7776 0.0025 -222.4 2.5 2020 30
118242 EX-2-1 Trench T2 -24.2 0.6022 0.0020 -397.8 2.0 4075 30
118243 EX-2-2 Trench T2 -22.2 0.5718 0.0019 -428.2 1.9 4490 30
118244 EX-3-1 Trench T3 -25.0 0.7613 0.0026 -238.7 2.6 2190 30
Eksik, TURKEY,
14
C Samples
Table A3 Radiocarbon ages from fault parallel trenches at the study site.
150
min
age (ka) <2.42-2.92
offset (m) 46+10
slip rate (mm/yr) >12-23 (17.5+5.5)
Stone et al., (1996) and Evans et al., (1997)
36
Cl
production rates
Philips et al., (1996)
36
Cl production rates
2.14-2.43
46
15-26 (20.5+5.5)
age range for 9 samples age range for 6 samples
14
C
Slip Rate Calculation Summary Table
age range for 9 samples age range for 6 samples
1.61-1.84
46+10
1.95-2.92
46+10
preferred
46+10
2.14-2.93
12-26 (19+7) 12-28 (20.5+8.5)
46+10
1.61-2.21
16-35 (25.5+9.5) 20-35 (27.5+7.5)
Table A4 Summary of slip rate calculations including estimates using all nine
36
Cl samples as well as preferred sub-cluster
of six youngest
36
Cl ages for production rate of Stone et al., (1996) and Evans et al., (1997), and Philips et al., (1996). Rates
based on radiocarbon ages are shown both for minimum rate based solely on age range of youngest charcoal fragment
recovered from beneath the shallowest terrace gravels, and for the full slip rate based on over-and underlying samples.
151
that the interval from which EX-3-1 was recovered is part of a shallow landslide that
has disturbed the top of the terrace gravels at this location. Thus, we do not have con-
fidence that the 2.06-2.47 ka age of this sample provides a precise constraint on the
age of the terrace gravels, although it is generally consistent with both the
14
C sample
ages from trench T4 and with the cosmogenic surface exposure ages from the terrace
deposit.
152
APPENDIX B:
Supplementary Data Figures for Yaylabeli Paleoseismologic Trench
Site
153
Figure B1 View looking southward across the Yaylabeli trench site. The alluvial fan that provides the primary sediment
input into the marsh is highlighted in yellow.
154
Figure B2 A 3D view of the faults and trench logs at Yaylabeli trench site.
155
Figure B3 Photo of the fault zone on western face of Yaylabeli trench T1.
156
Figure B4 Photo of western wall of trench T2 at Yaylabeli.
157
Figure B5 Photo of western wall of trench T2 at Yaylabeli.
158
Figure B6 Photo log of the eastern wall of trench T2 at Yaylabeli.
159
Figure B7 Photo of the fault zone on the eastern face of trench T3 at Yaylabeli.
160
Figure B8 Photo log of the eastern wall of trench T3 at Yaylabeli.
161
Figure B9 Photo log of the fault zone in trench T4 at Yaylabeli.
162
Figure B10 Photo of an example of peat sampling methodology applied at the Yaylabeli
trench site. This particular photo is from the eastern wall of trench T3 centered at
m22.5,-0.5.
163
Figure B11 A close up view of a peat sample (T4-11; east wall; m10,+0.21) from
eastern wall of trench T4 in Yaylabeli.
Abstract (if available)
Abstract
The behavior of major active faults at various temporal and spatial scales is one of the most fundamental, unresolved problems in modern tectonics. Determining the degree to which fault loading and strain release rates are constant (or non-constant) and documenting past earthquake occurrences are key approaches for understanding this phenomenon. I have employed geomorphic mapping, Quaternary dating methods, and paleoseismic trenching to generate fault rate and earthquake age data that help provide a better understanding of the North Anatolian fault's behavior in various temporal and spatial scales. Specifically, I mapped offset geomorphic markers along the North Anatolian fault and used cosmogenic nuclide (10Be and 36Cl) and radiocarbon (14C) dating methods to constrain the ages of these features. Using these and other published data, I also constructed one of the first compilations of strain-release rates for the North Anatolian fault. My compilation of these rate data reveals a constant slip rate of ~15-20 mm yr-1 over time scales of 103-105 years. This result, however, is slower than the geodetically constrained slip rate of 25 + 2 mm yr-1 [Reilinger et al., 2007], possibly indicating a strain transient. In addition to my slip rate studies, I performed paleoseismologic trenching on the eastern part of the North Anatolian fault at the village of Lorut. My results from this site demonstrate a relatively regular occurrence of large earthquakes along this stretch of the North Anatolian fault. I attribute the relatively constant strain release rates and regular earthquake recurrences to the mechanical simplicity of the Anatolian-Eurasian plate boundary in northern Turkey, which is dominated by the slip on the structurally mature North Anatolian fault.
Linked assets
University of Southern California Dissertations and Theses
Conceptually similar
PDF
Fault slip rates, constancy of seismic strain release, and landscape evolution in the eastern California shear zone
PDF
Implications of new fault slip rates and paleoseismologic data for constancy of seismic strain release and seismic clustering in the eastern California shear zone
PDF
Paleoseismologic and slip rate studies of three major faults in southern California: understanding the complex behavior of plate boundary fault systems over millenial timescales
PDF
Observations of temporal and spatial patterns of strain accommodation and earthquake occurrence along strike-slip faults of New Zealand and southern California, USA
PDF
Towards an understanding of fault-system mechanics: from single earthquakes on isolated faults to millenial-scale collective plate-boundary fault-system behavior
PDF
Revealing spatiotemporal patterns in fault-system behavior: insights from multi-millennial earthquake clustering and slip rate variability on the Garlock fault
PDF
Characterizing fault behavior and earthquake surface expression on timescales of single events to multiple millennia
PDF
Heterogeneity of earthquake stress drops, focal mechanisms and active fault zones
PDF
Paleoseismology of blind-thrust faults beneath Los Angeles, California: implications for the potential of system-wide earthquakes to occur in an active fold-and-thrust belt
PDF
Quantifying ground deformation of large magnitude earthquakes using optical imgaging systems
PDF
Determining fault zone structure and examining earthquake early warning signals using large datasets of seismograms
PDF
Seismicity distribution near strike-slip faults in California
PDF
The impact of faulting and folding on earthquake cycles in collisional orogens
PDF
Microseismicity, fault structure, & the seismic cycle: insights from laboratory stick-slip experiments
PDF
Analysis of waveform and catalog data of aftershocks for properties of earthquakes and faults
PDF
Stress-strain characterization of complex seismic sources
PDF
Structural and thermobarometric constraints on the exhumation of the northern Snake Range metamorphic core complex, Nevada
PDF
Elements of seismic structures near major faults from the surface to the Moho
PDF
Detection and modeling of slow slip events as creep instabilities beneath major fault zones
PDF
Extreme events as drivers of landscape evolution in active mountain belts
Asset Metadata
Creator
Kozaci, Özgür
(author)
Core Title
Constancy of strain release rates along the North Anatolian fault
School
College of Letters, Arts and Sciences
Degree
Doctor of Philosophy
Degree Program
Geological Sciences
Publication Date
12/16/2008
Defense Date
06/01/2008
Publisher
University of Southern California
(original),
University of Southern California. Libraries
(digital)
Tag
active tectonics,cosmogenic nuclide dating,OAI-PMH Harvest,paleoseismology,tectonic geomorphology,the North Anatolian fault
Place Name
faults: North Anatolian fault
(geographic subject)
Language
English
Contributor
Electronically uploaded by the author
(provenance)
Advisor
Dolan, James F. (
committee chair
), Davis, Gregory A. (
committee member
), Judge, Darrell L. (
committee member
)
Creator Email
kozaci@usc.edu,ozgurkozaci@yahoo.com
Permanent Link (DOI)
https://doi.org/10.25549/usctheses-m1932
Unique identifier
UC1140132
Identifier
etd-KOZACI-2532 (filename),usctheses-m40 (legacy collection record id),usctheses-c127-149733 (legacy record id),usctheses-m1932 (legacy record id)
Legacy Identifier
etd-KOZACI-2532.pdf
Dmrecord
149733
Document Type
Dissertation
Rights
Kozaci, Özgür
Type
texts
Source
University of Southern California
(contributing entity),
University of Southern California Dissertations and Theses
(collection)
Repository Name
Libraries, University of Southern California
Repository Location
Los Angeles, California
Repository Email
cisadmin@lib.usc.edu
Tags
active tectonics
cosmogenic nuclide dating
paleoseismology
tectonic geomorphology
the North Anatolian fault