Volume 32, Issue 5 p. 1107-1120
Regular Article
Free Access

Tectonic implications of fluvial incision and pediment deformation at the northern margin of the Central Anatolian Plateau based on multiple cosmogenic nuclides

Cengiz Yildirim

Corresponding Author

Cengiz Yildirim

İstanbul Technical University, Eurasia Institute of Earth Sciences, Istanbul, Turkey

Deutsches GeoForschungsZentrum (GFZ), Potsdam, Germany

Institut für Erd- und Umweltwissenschaften and DFG Leibniz Center for Surface Process and Climate Studies, Universität Potsdam, Potsdam, Germany

Corresponding author: C. Yildirim, İstanbul Technical University, Eurasia Institute of Earth Sciences, Maslak/Sar𝚤yer 34469, Istanbul, Turkey. ([email protected])Search for more papers by this author
Taylor F. Schildgen

Taylor F. Schildgen

Institut für Erd- und Umweltwissenschaften and DFG Leibniz Center for Surface Process and Climate Studies, Universität Potsdam, Potsdam, Germany

Search for more papers by this author
Helmut Echtler

Helmut Echtler

Deutsches GeoForschungsZentrum (GFZ), Potsdam, Germany

Search for more papers by this author
Daniel Melnick

Daniel Melnick

Institut für Erd- und Umweltwissenschaften and DFG Leibniz Center for Surface Process and Climate Studies, Universität Potsdam, Potsdam, Germany

Search for more papers by this author
Bodo Bookhagen

Bodo Bookhagen

Department of Geography, University of Santa Barbara, Santa Barbara, California, USA

Search for more papers by this author
Attila Çiner

Attila Çiner

Department of Geological Engineering, Hacettepe University Ankara, Ankara, Turkey

Search for more papers by this author
Samuel Niedermann

Samuel Niedermann

Deutsches GeoForschungsZentrum (GFZ), Potsdam, Germany

Search for more papers by this author
Silke Merchel

Silke Merchel

Helmholtz-Zentrum Dresden-Rossendorf (HZDR), Dresden, Germany

Search for more papers by this author
Martin Martschini

Martin Martschini

Faculty of Physics, Isotope Research, VERA Laboratory, University of Vienna, Vienna, Austria

Search for more papers by this author
Peter Steier

Peter Steier

Faculty of Physics, Isotope Research, VERA Laboratory, University of Vienna, Vienna, Austria

Search for more papers by this author
Manfred R. Strecker

Manfred R. Strecker

Institut für Erd- und Umweltwissenschaften and DFG Leibniz Center for Surface Process and Climate Studies, Universität Potsdam, Potsdam, Germany

Search for more papers by this author
First published: 01 August 2013
Citations: 30
This article was corrected on 29 NOV 2016. See end of the full text for details.

Abstract

[1] We document Quaternary fluvial incision driven by fault-controlled surface deformation in the inverted intermontane Gök𝚤rmak Basin in the Central Pontide mountains along the northern margin of the Central Anatolian Plateau. In-situ-produced 10Be, 21Ne, and 36Cl concentrations from gravel-covered fluvial terraces and pediment surfaces along the trunk stream of the basin (the Gök𝚤rmak River) yield model exposure ages ranging from 7 ± 1 ka to 346 ± 45 ka and average fluvial incision rates over the past ~350 ka of 0.28 ± 0.01 mm a-1. Similarities between river incision rates and coastal uplift rates at the Black Sea coast suggest that regional uplift is responsible for the river incision. Model exposure ages of deformed pediment surfaces along tributaries of the trunk stream range from 60 ± 5 ka to 110 ± 10 ka, demonstrating that the thrust faults responsible for pediment deformation were active after those times and were likely active earlier as well as explaining the topographic relief of the region. Together, our data demonstrate cumulative incision that is linked to active internal shortening and uplift of ~0.3 mm a-1 in the Central Pontide orogenic wedge, which may ultimately contribute to the lateral growth of the northern Anatolian Plateau.

Key Points

  • active deformation at the northern margin of the Central Anatolian Plateau
  • fluvial incision at the northern margin of the Central Anatolian Plateau
  • Quaternary Uplift Rate at the northern margin of the Central Anatolian Plateau

1 Introduction

[2] The pattern and rate of fluvial incision along major rivers are often useful proxies to reconstruct the history of surface uplift and deformation within a landscape, as fluvial incision typically sets the pace of landscape response to tectonic or climatic forcing [e.g., Whipple and Meade, 2004]. Large rivers and their tributaries that have incised bedrock channels in areas such as the Apennines [Wegmann and Pazzaglia, 2009], Tibet and the Himalaya [Hetzel et al., 2006; Pan et al., 2003; Wobus et al., 2005; Kirby and Ouimet, 2011], and the Andes [ Tosdal et al., 1984; Garcia and Hérail, 2005; Barke and Lamb, 2006; Hoke et al., 2007; Schildgen et al., 2007] have been used as proxies for orogen-scale uplift and regional deformation. In actively deforming regions, planar geomorphic features such as fluvial strath terraces, and pediments are also useful strain markers to quantify deformation related to more localized shortening and uplift along specific structures [Strecker et al., 1989; Benedetti et al., 2000; Lavé and Avouac, 2000].

[3] Yildirim et al. [2011] demonstrated that the northern margin of the Central Anatolian plateau (Figures 1a and 1b), defined by the Central Pontide mountains (Figure 1c), is an orogenic wedge with a positive flower-structure geometry between the restraining bend of the North Anatolian Fault (NAF) and the Black Sea (Figure 1d). They further inferred active deformation within the wedge based on deformed Quaternary sediments, GPS-derived slip vectors along the trace of the NAF, and morphometric indices including transient river profiles. Despite this evidence for active deformation, the timing of deformation, its pattern, and the uplift rates have not yet been quantified in the region, making it difficult to assess the importance of the restraining bend of the NAF in terms of long-term topographic development and exhumation along the plateau's northern margin. Fortunately, fluvial terraces and pediments associated with northward-flowing rivers sourced from the interior part of the plateau are well suited features to evaluate the rate of regional uplift and patterns of local deformation at the northern plateau margin. Information from the northern margin can then be linked to observations within the central part of the plateau, thus helping to elucidate the characteristics of plateau-margin development on timescales spanning the Quaternary.

Details are in the caption following the image
(a) Simplified neotectonic map of Turkey [Barka and Reilinger, 1997; Dilek, 2006]. CAP: Central Anatolian Plateau, EAF: East Anatolian Fault, KR: K𝚤z𝚤l𝚤rmak River, NAF: North Anatolian Fault, TF: Tuz Gölü Fault. Dashed white line denotes boundary of the Central Pontides. Small box on the K𝚤z𝚤l𝚤rmak River shows terrace locations of Doğan [2011]. (b) Swath profile of topography across the northern margin of the CAP. The top and bottom of the gray shaded area shows maximum and minimum elevations, respectively. Dashed line shows mean elevation. (c) Topography, active faults [Şaroğlu et al., 1992; Dirik, 1993; Andrieux et al., 1995; Barka and Reilinger, 1997; Şenel, 2002], drainage system, and morphotectonic units in the study area. Gray box denotes location of the swath profile in Figure 1b. Pn: Peninsula. White line denotes watershed of the Gök𝚤rmak Basin. Numbers denote surface samples, and DA and DB show the locations of the depth profiles. Arrows labeled with single capital letters indicate locations and view directions of photos in Figure 2. (d) Conceptual model for the Central Pontides in the form of an active orogenic wedge related to a positive flower structure developed over a shallow detachment surface linked to the North Anatolian Fault at depth (modified after Yildirim et al. [2011]). BF: Bal𝚤fak𝚤 Fault, EF: Erikli Fault, EkF: Ekinveren Fault, KF: Kay𝚤 Fault.

[4] In this study, we quantify the rate of late Quaternary fluvial incision in the intermontane Gök𝚤rmak Basin in the central part of the Pontide mountains (Figure 1c) by measuring in situ-produced 10Be, 21Ne and 36Cl concentrations (Tables 1-4) on gravel-covered fluvial strath terraces (Figures 2 and 3). The Gök𝚤rmak Basin also hosts structures related to young tectonic activity, whose local contributions to deformation we investigate through cosmogenic nuclide exposure dating of deformed pediment surfaces (Figure 4). The area thus allows us to constrain the rate and timing of regional uplift that is related to local deformation along key structures at the northern plateau margin.

Table 1. Locations, AMS Results, and Model Ages (Inheritance Is Not Corrected) of 10Be Samples
Sample

Latitude

(oN)

Longitude

(oE)

Elevation

(m asl)

Production Ratea

Quartz

(g)

Be Carrier

(mg)

10Be/9Be

(x10-13)

10Be Concentraion

(104 atoms/g

SiO2)

10BeModel Ageb

(ka)

Spallation Muons
1 41.4075 35.0941 219 5.03 0.193 98.94 0.20 2.66±0.07 3.5±0.1 7±1
2a 41.5597 34.6714 350 5.65 0.202 114.36 0.20 39.00±0.75 44.6±0.9 81±8
5 41.6314 34.3954 508 6.50 0.213 106.77 0.19 12.77±0.22 15.5±0.3 24±2
6 41.4844 34.1255 618 7.07 0.221 106.34 0.19 24.77±0.47 30.2±0.6 43±4
7 41.4960 33.8898 776 8.08 0.233 100.44 0.20 64.65±0.91 83.9±1.2 110±10
8 41.4634 33.6730 842 8.51 0.239 113.50 0.20 43.17±0.60 51.2±0.7 178±18
10 41.4674 33.4949 903 8.88 0.243 103.18 0.20 134.70±1.00 170.5±1.3 219±23
11 41.4902 33.4740 902 9.01 0.244 86.08 0.20 148.60±3.10 226.9±4.7 304±34
12a 41.4955 33.7673 882 8.81 0.242 100.00 0.19 44.78±0.72 58.6±0.9 60±5
14a 41.2510 33.9668 1274 11.94 0.275 104.20 0.20 55.30±0.64 69.8±0.8 60±5
  • a Production rates (atoms/g/a) calculated using Lal [1991]/Stone [2000] constant production scaling for surface samples, assuming negligible shielding.
  • b 10Be ages were calculated using the Cronus online calculator [Balco et al., 2008], assuming no surface denudation. Uncertainties are reported at the 1σ confidence level. 10Be/9Be ratio for standard = 2.85E−12.
Table 2. Locations, Cosmogenic 21Ne Concentrations and Model Exposure Ages (Inheritance Is Not Corrected) of Pediment Surfaces Calculated From 21Ne Analyses
Sample Latitude (°N) Longitude (°E) Elevation (m asl) 10Be Production rate (atoms/g/a) 21Ne Production ratea (atoms/g/a) 21Ne Concentrationb (104 atoms/g) Exposure Ageb (ka)
13 41.2393 33.7869 1130 10.94 48.23 413 ± 23 86 ± 7
14b 41.2510 33.9668 1274 12.22 53.48 372 ± 29 70 ± 7
  • a 21Ne production rates were calculated using a value of (0.232 ± 0.009) for the 10Be/21Ne production ratio [Goethals et al., 2009].
  • b Uncertainties are reported at the 1σ confidence level.
Table 3. Locations, AMS Results, and Model Ages (Inheritance Is Not Corrected) of 36Cl Samples
Sample

Latitude

(oN)

Longitude

(oE)

Elevation

(m asl)

Production Ratea

(atoms/g*a)

36 Cl Concentration

(104 atoms/g)

36Cl Model Ageb

(ka)

2b 41.5597 34.6714 350 69.2 230.6±6.6 133±13
3 41.5776 34.6571 360 709 220.4±7.0 118±12
4 41.6129 34.5552 500 78.3 640±12 438±64
9 41.4734 34.5767 931 105.9 460.4±4.6 178±20
12b 41.4955 33.7673 882 109.9 291±10 108±11
  • a 36 Cl production rates were calculated using Stone et al. [1996] constant production scaling factor with negligible shielding.
  • b 36Cl ages were calculated by using Excel Add-on spreadsheet from Schimmelpfennig et al. [2009] assuming no surface denudation. Uncertainties are reported at the 1σ confidence level.
Table 4. Locations and AMS Results of 10Be Samples From Depth Profiles DA and DB
Sample

Latitude

(oN)

Longitude

(oE)

Elevation

(m asl)

Depth

(cm)

Production Ratea Quartz

Be Carrier

(mg)

10Be/9Beb

(x10-13)

10Be Concentrationb

(104 atoms/g SiO2)

Spallation Muons
DA_0 41.3933 35.1170 240 0 5.08 0.194 93.56 0.1959 n/a n/a
DA_1 41.3933 35.1170 100 1.46 0.168 140.83 0.1907 14.27±0.32 12.91±0.29
DA_2 41.3933 35.1170 200 042 0.149 199.09 0.1939 24.46±0.55 15.91±0.36
DB_0 41.4912 33.9836 652 0 7.26 0.224 112.98 0.1942 69.72±1.60 80.07±1.83
DB_1 41.4912 33.9836 120 2.08 0.196 200.11 0.1915 44.78±0.72 28.62±0.46
DB_2 41.4912 33.9836 220 0.6 0.171 101.11 0.1919 12.31±0.20 15.62±0.25
  • a Lal [1991]/Stone [2000] constant production for surface samples.
  • b Uncertainties are reported at the 1σ confidence level.
  • We assumed a range of bulk densities (ρ) from 1.6 to 2.0 g cm-3 and a neutron-attenuation length (Λ) of 160 g/cm2 [Gosse and Phillips, 2001].
Details are in the caption following the image
(a) Terrace surface and outcrop of the depth profile DB. (b) Outcrop sampled for depth profile DB. (c) View of terrace staircase. T3 is the terrace surface sampled for sample 4. The rest of the terrace surfaces were not sampled because of anthropogenic impacts on their surfaces. FP: Floodplain.
Details are in the caption following the image
(a) Geomorphological map of the Gök𝚤rmak and K𝚤z𝚤l𝚤rmak confluence. Arrows with capital letters indicate point of views in Figures 3b, 3d, and 3e. FP: Floodplain. (b) River terrace T4 is sampled for depth profile DA. (c) Outcrop sampled for depth profile DA. (d) View of steep, convex slopes of the K𝚤z𝚤l𝚤rmak Valley at the confluence with the Gök𝚤rmak River; white lines show scar of a landslide. (e) Steeply dipping alluvial fan-delta foreset deposits in the K𝚤z𝚤l𝚤rmak Valley.
Details are in the caption following the image
(a) View of a pediment surface and cemented nature of top gravels. (b) Close-up view of cemented pediment surface. (c) View of pediment surface staircases (see Figure 1c for locations of photographs) from the upstream part of the basin. (d) Deformed pediment surface in the midstream part of the basin. The surface is flexed and back tilted toward the mountain front (left) of the Ekinveren Fault. (e) View of pediment surface staircases from the downstream part of the basin.

2 Regional Tectonic Setting

[5] The northern margin of the Central Anatolian plateau spans the northward-convex Central Pontide mountains between the Black Sea and the curved sector of the North Anatolian Fault (NAF) (Figure 1a). Following closure of the Neotethys Ocean and collision of the Pontide island arc with the Sakarya continent to the south mainly during the late Cretaceous, the Central Pontides were shortened and uplifted along a north-vergent detachment and thin-skinned foreland fold-and-thrust-belt during Paleocene-Eocene time [e.g., Şengör, 1995; Okay and Tüysüz, 1999; Dilek, 2006]. North of the Pontides, the southern Black Sea margin is currently under compression due to collision between Africa-Arabia and Eurasia [Barka and Reilinger, 1997; Cloetingh et al., 2003]. South of the Pontides, the NAF is the most prominent active member of the dextral North Anatolian Fault Zone [Şengör et al., 2005]. The NAF changes strike along its middle portion to form a broad, asymmetric restraining bend (Figure 1a). The Central Pontides are located at the apex and western continuation of this broad bend, where higher contractional strain is likely to be accumulated (Figure 1a). Yildirim et al. [2011] found that the NAF-normal component of motion based on GPS-derived slip vectors reaches up to 8 mm a-1. The north-south asymmetric topography of the Central Pontides (Figure 1b) and transient river profiles, combined with the spatial distribution and geometry of Quaternary thrust faults (Figure 1c), imply that the Central Pontides comprise an active orogenic wedge that accommodates NAF-normal contraction on a south-dipping low-angle detachment splaying from the NAF and extending to the Black Sea (Figure 1d) [Yildirim et al., 2011].

3 Gök𝚤rmak Intermontane Basin

[6] The ~7000 km2 Gök𝚤rmak Basin is the largest intermontane basin within the Central Pontides, with elevations ranging from 170 to 2590 m above sea level (asl) (Figure 1c). The semiarid basin is located in the lee of the coastal ranges and receives an average of 350–500 mm a-1 precipitation, which is similar to the central part of the Central Anatolian plateau [Bozkurt and Sen, 2011]. The Gök𝚤rmak River, which is the main river that drains the Gök𝚤rmak Basin, and its tributaries are often flanked by fluvial strath terraces. The straths are covered by 3 to 5 m thick gravel deposits (Figures 2a and 2b), which contain mostly well-rounded pebbles. Their heights vary between 18 and 130 m above the channel (Figure 2c and Tables 5 and 6). The K𝚤z𝚤l𝚤rmak River is sourced from highlands rising in the eastern part of Anatolia and flows across the Central Anatolian plateau with a crescent-like channel geometry (Figure 1a). The Gök𝚤rmak River reaches the Black Sea via the gorge and delta of the K𝚤z𝚤l𝚤rmak River, approximately 120 km downstream from the study area (Figure 1c). The K𝚤z𝚤l𝚤rmak gorge is deeply incised, locally very narrow (<100 m) and hosts a series of fluvial terraces and deposits at successively lower elevations (Figures 3a–3d), and steep, landslide-prone slopes and fan delta deposits (Figures 3d and 3e). Along the Black Sea coast, the K𝚤z𝚤l𝚤rmak delta region is characterized by uplifted Quaternary paleo-deltas at 20 to 30 m and 60 to 70 m asl. [Akkan, 1970].

Table 5. Statistics for the Simulation of the Depth Profiles DA and DB
Age (ka) Inheritance (104 atoms/g) Erosion Rateb (cm/ka)
DA
Mean 132.5 10.53 0.11
Median 131.8 10.57 0.11
Mode 130.8 10.36 0.16
Lowest χ2 142.1 10.41 0.2
Maximuma 179.9 14.46 0.26
Minimuma 99.9 5.36 0
DB
Mean 99.3 8.63 0.15
Median 99.2 8.75 0.15
Mode 100.4 9.38 0.26
Lowest χ2 104.4 10.14 0.24
Maximuma 115.0 12.03 0.29
Minimuma 84.4 4.10 0
  • a Maximum and minimum values represent the 95% (2σ) confidence window for each parameter.
  • b Maximum total erosion confined to 30 cm.
Table 6. Inheritance Corrected Ages and Incision Rates of the Strath Terracesa
Sample TCN Height Above River (m) Age (ka) Incision Rate (mm/a)
1 10Be 45 ± 5 7 ± 1 6.43 ± 1.16
2a 10Be 30 ± 3 81 ± 8 0.37 ± 0.05
2b 36Cl 30 ± 3 78 ± 7 0.38 ± 0.05
3 36Cl 32 ± 3 68 ± 6 0.47 ± 0.06
4 36Cl 130 ± 10 346 ± 45 0.38 ± 0.06
6 10Be 18 ± 2 43 ± 4 0.42 ± 0.06
DB 10Be 25 ± 3 99 ± 8 0.25 ± 0.04
8 10Be 54 ± 5 178 ± 18 0.30 ± 0.04
10 10Be 52 ± 5 219 ± 23 0.24 ± 0.03
11 10Be 90 ± 9 304 ± 34 0.30 ± 0.04
  • a Uncertainties are reported at the 1σ confidence level.

[7] The Gök𝚤rmak Basin is bounded to the north by the Sinop range, a push-up structure between the Erikli and Ekinveren thrust faults, and to the south by the Ilgaz range, another push-up structure between the NAF and its transpressional splays (Figures 1c and 1d). The Ekinveren Fault, at the northern margin of the Gök𝚤rmak Basin, is the principal structure bounding the mountain front of the Sinop Range (Figure 1c). This N-dipping, E-W striking thrust fault comprises multiple en échelon-segments along the northern flanks of the Gök𝚤rmak Basin (Figure 1c) [Tüysüz, 1999].

[8] River channels incise through smooth, extensive, low-sloping erosional bedrock surfaces that are mantled by a thin regolith blanket along the mountain fronts of the Gök𝚤rmak Basin (Figures 4a–4e). These surfaces are pediments that were sculpted into limestone, marls, and sandstone by erosive processes associated with the transport of the capping regolith [e.g., Cooke et al., 1993; Pelletier, 2010]. Localized deformation of the pediments has led to incision by the trunk channel or by tributary streams, implying that the surfaces were last active prior to their deformation. The top 1 to 2 m of gravels/regolith capping the higher pediments include well-developed carbonate K-horizons (Figure 4a). The limestone and quartzite clasts are mostly angular and subrounded cobbles or boulders (Figure 4b).

[9] Some of the gently inclined pediment-surface remnants are deformed by active faults and folds along the mountain fronts, leading to different elevations along the southern mountain front compared to those along to northern mountain front. This asymmetry is also reflected by the trunk channel of the basin, which flows closer to the northern mountain front, especially in the central part of the basin (Figure 1c).

4 Methods

4.1 Terrestrial Cosmogenic Nuclide Exposure Dating

[10] The accumulation of in situ-produced terrestrial cosmogenic nuclides [TCNs], such as 10Be, 21 Ne, and 36Cl, is related to the time that a fluvial terrace or pediment surface is exposed to secondary cosmic rays and can be used to derive an exposure age for the surface [cf. Siame et al., 1997; Gosse and Phillips, 2001; Niedermann, 2002]. Cosmogenic nuclides are produced in near-surface rocks by neutron-induced spallation, slow (or stopping) muon capture, fast muon-induced reactions, and capture of epithermal and thermal (low-energy) neutrons [Gosse and Phillips, 2001, and references therein]. The nuclide 10Be is produced through spallation reactions, stopped negative muon, and fast muon interactions; 21Ne is mainly produced through spallation reactions, and 36Cl is produced through low-energy neutron capture mechanisms, spallation reactions, stopped negative muon, and fast muon interactions [Zreda et al., 1991; Ivy-Ochs and Schaller, 2009]. Currently, the best-constrained production rates for these nuclides include the production of 10Be and 21Ne in quartz and 36Cl in calcite.

[11] For landforms capped by alluvial or colluvial covers, like river terraces and pediments, exposure ages from TCN concentrations in surface clasts can be complicated by inherited TCNs generated during exhumation and transportation prior to final deposition [Anderson et al., 1996; Hancock et al., 1999; Perg et al., 2001]. Specifically, the measured concentration N in a profile at depth z and time t, assuming uniform inheritance through the profile, can be described by a simplified equation (which can be expanded to include production mechanisms other than spallation):
urn:x-wiley:02787407:media:tect20066:tect20066-math-0001(1)
where Ninher is the inheritance at the time of deposition, Λspall is the attenuation length for spallation-induced reactions, τ is the radioactive mean life, ρ is the density of the material, and P(0) is the production rate at the surface [Repka et al., 1997]. Because the production of TCNs by spallation and muon-induced reactions decreases approximately exponentially with depth, deeper samples are effectively shielded from postdepositional accumulation, meaning that measurements of nuclide concentrations from different depths (or one sample at great depth) can be used to estimate the nuclide inheritance in a sedimentary deposit. If that inherited component can be quantified, it can be subtracted from the surface concentration to provide a more reliable exposure age for the surface [Hancock et al., 1999].

[12] We collected TCN samples from twelve fluvial strath terraces incised by the Gök𝚤rmak River to estimate the timing and rate of regional incision and seven samples from locally deformed pediments associated with active thrusts or folds along the southern and northern mountain fronts to constrain the timing of deformation. We measured 10Be and 21Ne in quartz-rich samples and 36Cl in carbonate samples. Sample sites were selected in areas that had no apparent anthropomorphic modification, away from the edges of the surfaces to minimize postdepositional erosion, and away from surrounding hillslopes to minimize postdepositional burial. On terrace surfaces, we collected at least 30 pebbles with 1 to 3 cm diameter to obtain a representative mean surface concentration [e.g., Repka et al., 1997]. At two locations (samples 2a,b and 12a,b), we collected samples of both quartz and carbonate to measure both 10Be and 36Cl, and at one location (sample 14a, b), we measured 10Be and 21Ne nuclides from the same quartz sample to compare exposure ages of the surface derived through different nuclides. We also collected samples along two depth profiles at terrace sites DA and DB (Figures 1c, 2a, 3b, and 3c). For the depth profiles, we similarly collected at least 30 pebbles with 1 to 3 cm diameter at each depth interval so that if there is a grain-size dependence of the inherited nuclide concentration [e.g., Brown et al., 1998], our calculated inheritance values from the depth profiles should still be comparable to those of the terrace surface samples. Our samples from the pediment surfaces consist of 5 amalgamated cobbles, to increase the likelihood to get a representative mean surface concentration since the time of pediment abandonment.

4.2 TCN Sample Preparation and Analysis

[13] Chemical preparation of targets for accelerator mass spectrometry (AMS) and quartz samples for noble gas analysis followed variations of established protocols for 36Cl from carbonate [Stone et al., 1996; Conard et al., 1986] and for 10Be and 21Ne from quartz-rich samples [Kohl and Nishiizumi, 1992]. AMS measurements of 10Be were made at Lawrence Livermore National Laboratory. We calculated 10Be surface exposure ages using the CRONUS online calculator version 2.2.1 (available at http://hess.ess.washington.edu/, with documentation in Balco et al. [2008]), which includes variations in altitude, latitude, and topographic shielding in calculating production rates from spallation and muons and assumes reference 10Be production rates that depend on the selected scaling scheme (see explanation in the study by Balco et al. [2008]). We used the Lal [1991]/Stone [2000] constant production scaling scheme. Uncertainties in the reference production rates are estimated at ca. 10% [Balco et al., 2008]. We rely on the original ICN standard [Nishiizumi et al., 2007] and used a 10Be half-life of (1.387 ± 0.012) Ma [Chmeleff et al., 2010; Korschinek et al., 2010]. Tables 1-3 list the measured 10Be, 21Ne and 36Cl concentrations with analytical uncertainties and uncorrected model exposure ages, respectively. To demonstrate the impact of different production rate scaling methods on our results, model exposure ages of 10Be samples based on various scaling schemes are compared in Table S1 in the supporting information.

[14] AMS measurements of 36Cl were performed at the University of Vienna, VERA Laboratory [Martschini et al., 2013] and normalized to SM-Cl-standards [Merchel et al., 2011]. We calculated surface exposure ages using the Excel Add-on provided by Schimmelpfennig et al. [2009], which includes variations in altitude, latitude, and topographic shielding in calculating production rates from spallation and muons, and assumes a SLHL 36Cl spallation rate of (48.8 ± 3.4) at gCa-1 a-1 [Stone et al., 1996] and a neutron attenuation length of 160 g cm-2 [Gosse and Phillips, 2001]. Production rates based on the calibration data set of Stone et al. [1996] and used for age calculation are based on inductively coupled plasma mass spectrometry (ICP-MS) measurements of major and trace elements of the samples, which are given in Tables S2 and S3 (supporting information), and a 36Cl half-life of 0.301 Ma [Endt, 1990].

[15] Two samples (samples 13 and 14b) were analyzed for cosmogenic 21Ne at the facilities of the GFZ Potsdam. The purified quartz was first crushed to a grain size of ~100 µm to open part of the fluid inclusions and thereby reduce the contribution of atmosphere-like trapped Ne. Samples were then wrapped in Al foil and loaded into the sample carrousel above the extraction furnace, which was baked at 100°C for about one week. The noble gases were extracted by stepwise heating (at 400, 600, 900, and 1200°C) for 20 min each. In addition, one aliquot of sample 14b was crushed in vacuo to check whether trapped Ne was indeed of atmospheric composition. After gas extraction by either heating or crushing, chemically active gases were removed in two Ti sponge and two SAES (ZrAl) getters, and He, Ne, and Ar-Kr-Xe were separated from each other by trapping in a cryogenic adsorber at 11 K and subsequent sequential release. Noble gas concentrations and isotopic compositions were determined in a VG5400 sector field mass spectrometer and were corrected for isobaric interferences, instrumental mass fractionation, and analytical blanks. Further details about the analytical procedures can be found in Niedermann et al. [1997].

[16] The crushing extraction yielded a Ne isotopic composition consistent with that of air [22Ne/20Ne = 0.1020, 21Ne/20Ne = 0.002959]; therefore, all 21Ne excesses were calculated relative to atmospheric composition. In the three-isotope diagram (Figure 5), most data plot on the “spallation line” [Niedermann et al., 1993], i.e., they are consistent with a two-component mixture of atmospheric and cosmogenic Ne; some data additionally show a slight isotope fractionation. Excess 21Ne in the 1200°C step is negligible, confirming that the samples do not contain any nucleogenic Ne. Therefore, total cosmogenic 21Ne concentrations were calculated by summing up the 21Ne excesses up to 900°C. We calculated 21Ne production rates using a value of (0.232 ± 0.009) for the 10Be/21Ne production ratio [Goethals et al., 2009] to correspond to the 10Be half-life value applied in CRONUS version 2.2.1 [Balco et al., 2008], and based on a total 10Be SLHL total production rate derived from the Lal [1991]/Stone [2000] scaling of the CRONUS online calculator of 4.58 ± 0.43 atoms g-1 a-1.

Details are in the caption following the image
Ne three-isotope diagram for samples 13 and 14b. Light gray symbol with white square inside denotes crusher extraction of sample 14b, white symbols are 400°C, light gray 600°C, dark gray 800°C, and black 1200°C steps. “Spallation line” is the mixing line of atmospheric and cosmogenic Ne according to Niedermann et al. [1993]. mfl: mass fractionation line. 2σ uncentertainties are shown.

[17] Uncertainties in final reported ages, after the inheritance correction described below, include uncertainties in 10Be and 36Cl production rates, the 10Be/21Ne production ratio, uncertainty in the mass of quartz dissolved, AMS or noble gas mass spectrometric measurements, blank corrections, and Be carrier weighing.

5 Results

5.1 Nuclide Inheritance and Constraints on Exposure Ages

[18] We determined the 10Be nuclide concentration in depth profiles at two locations (Figure 1c) at road cuts (Figures 2a, 2b, 3b, and 3c) to estimate the nuclide inheritance of our samples, which allows us to calculate a more reliable exposure age from the surface-sample TCN concentrations. Advanced carbonate cementation of the gravels precluded depth-profile sampling at most locations. Table 4 lists 10Be concentrations of the samples from depth profiles. The depth profiles consist of amalgamated quartz clasts from the surface, 1.0 (or 1.2) m, and 2.0 (or 2.2) m depth below the surface at two terraces (Figures 2a, 2b, 3b, and 3c) located at the center and eastern end of the Gök𝚤rmak Basin (Figure 1c).

[19] We modeled the 10Be concentration in the two depth profiles (Figures 6a and 6b) by varying the surface age, deposit density, and erosion rate with a Monte Carlo simulator developed by Hidy et al. [2010] to calculate the most probable values for inherited nuclide concentration and exposure age. The spallogenic nuclide production is accounted for by using the Lal [1991]/Stone [2000] constant production scaling model, and the muogenic component of the production rate is accounted for in the simulator by applying the theoretical production equations of Heisinger et al. [2002a, 2002b]. The uncertainty in total production rate was estimated to be 10% according to Balco et al. [2008]. We explored a range of bulk densities ( ρ) (assumed to be uniform throughout the profile) from 1.6 to 2.0 g cm-3, assumed a neutron-attenuation length (Λ) of 160 g cm-2 [Gosse and Phillips, 2001], a maximum erosion rate of 1 cm ka-1, and a total erosion of 30 cm. Despite the maximum erosion rate constraint being quite low, the results of the Monte Carlo simulations (summarized in Table 5) demonstrate that the best fit erosion rates are lower than our maximum constraint, which is likely the result of the essentially flat (hence, very slowly eroding) slopes that cap the terraces.

Details are in the caption following the image
(a) Concentration versus depth plots illustrating the (right) 2σ profile solution spaces and (left) best fits curve for the depth profile DA. (b) Concentration versus depth plots illustrating the (right) 2σ profile solution spaces and (left) best fits curve for the depth profile DB. The 1.2-m depth sample was not part of the simulation due to discordance with the theoretical change in nuclide concentration with depth. Statistics for these simulations are shown in Table 5.

[20] The three samples of depth profile DA were taken from gravels capping a strath terrace at the confluence of the Gök𝚤rmak and K𝚤z𝚤l𝚤rmak rivers (Figures 3a–3c). Unfortunately, the surface sample collected from the depth profile DA was lost during laboratory processing. Although losing the surface sample precludes making a well-constrained estimate of the age of the surface, the two samples at 1.0 and 2.0 m depth still allow us to make a reasonable estimate of the 10Be nuclide inheritance (Figure 6a). The depth-concentration simulation yielded the most probable mean values of 133 +47/-33 ka for the exposure age and (10.5 +4.0/-5.1) × 104 atoms g-1 for the inherited concentration of 10Be within the gravel deposit (Table 5). Due to the very large error on the modeled age, we do not use the age in any further analysis.

[21] The three samples of depth profile DB were taken from the midstream part of the basin (Figures 1c, 2a, and 2b). The sample from 1.2 m depth shows discordance with the theoretical change in nuclide concentration with depth (Figure 6b). Because we observed no evidence for compositional changes or multistep deposition in the stratigraphy of the gravel, the higher concentration of the 1.2 m depth sample may indicate variable inheritance in the amalgams of heterogeneous clasts [e.g., Codilean et al., 2008; Le Dortz et al., 2012]. We tested various possibilities to calculate the most probable values for inherited nuclide concentration, exposure age, and erosion rate. The Monte Carlo simulation found no set of parameter values to match the data at a confidence interval lower than 10σ when we included all three samples (from surface and 1.2 m and 2.2 m depths). After discarding the sample at 1.2 m below the surface, the simulation ran well at a 2σ confidence level and yielded the most probable mean values of 99 +16/−15 ka for the age and (8.6 +1.5/−4.5) × 104 atoms g-1 for the inherited concentration (Figure 6b and Table 5). If we exclude the top sample instead, the model finds no set of parameter values to match the data at a confidence interval lower than 9σ, even when allowing for terrace ages up to 1 Ma. Alternatively, if we exclude the 2.2 m sample, the resulting modeled inherited concentration is ~5 times the inherited concentration derived from depth profile DA and is higher than the concentration of several of the surface samples from other terraces. For these reasons, we believe that the middle sample is the most likely outlier and that it should be excluded when estimating the inherited concentration.

[22] Despite sample site DA being located at the downstream portion of the basin, with a probable contribution from the K𝚤z𝚤l𝚤rmak River catchment, the inherited concentration calculated from depth profile DA overlaps within error of that calculated from depth profile DB (Table 5). Nonetheless, because the inherited concentration that we obtained from the depth profile DA might contain a signal from the whole K𝚤z𝚤l𝚤rmak River catchment, including the internal part of the plateau, we preferred to use the inherited concentration that we obtained from the depth profile DB because it reflects inheritance only related to the Gök𝚤rmak River catchment. We assume that same inheritance for all fluvial strath terrace surfaces in the catchment.

[23] Because we only measured 10Be in the depth-profile samples, we have no direct measurements of inheritance values for 36Cl samples taken from fluvial strath terrace surfaces. The production mechanisms for 36Cl are different from those for 10Be, therefore, the depth-dependency of production rates differs for the two nuclides. This difference could lead to slight variations in the ratio of the inherited concentration of the two nuclides, particularly if the samples experience any complex exposure histories [Gosse and Phillips, 2001]. Braucher et al. [2011] measured concentrations of 10Be and 36Cl from a core drilled down to 11 m below the surface and found that 36Cl concentrations between 1 m and 4.3 m depth are consistently around 10 times higher than 10Be concentrations. Although there are numerous 36Cl production pathways and they are strongly dependant on bulk chemical composition, due to the lack of experimental 36Cl data for individual samples at depth from our sampling site, we estimate a rough 36Cl inheritance value by following the observations of Braucher et al. [2011] and multiplying our inherited 10Be concentrations (8.6 +1.5/-4.5 x104 atoms g-1) by a factor of 10 (the approximate production rate ratio of 36Cl/10Be). It must be noted that, although these corrections to the measured 36Cl concentrations are quite large (therefore, lend greater uncertainty to the final model 36Cl exposure age), they considerably reduce the discrepancy between our 10Be and 36Cl ages (Tables 1 and 3). These results suggest that inheritance corrections are an important component for obtaining accurate 36Cl exposure ages of transported gravels.

5.2 Surface Exposure Ages of the Strath Terraces

[24] Because the surface exposure ages calculated for each sample represents the time elapsed since the respective surface was active gravel transport environments, we consider them to represent “abandonment ages” for the formerly active fluvial channels or pediment surfaces. The surface exposure ages of the fluvial terraces are listed in Table 6. The 10 samples collected from strath terrace surfaces along the trunk channel yield inheritance-corrected model exposure ages that range from 7 ± 1 ka to 346 ± 45 ka (Table 6). The terrace exposure ages generally increase in age with greater height, from 43 ± 4 ka at 18 m above the river (sample 6) to 346 ± 45 ka at 130 m (Figure 3c; sample 4) above the river (Figure 7). One exception is sample 1, which yields a very young age (7 ± 1 ka) for a relatively high terrace (45 m), thus plotting as an outlier from the regional trend (Figure 7).

Details are in the caption following the image
Terrace abandonment ages (based on Lal [1991]/Stone [2000] scaling; Table 6) versus heights above the river. Bars on symbols show 1σ uncertainties. Odd-numbered MIS are highlighted with white colored numbers and dark gray shading. Black numbers are numbers of samples. Bold line is the median regression line of 10Be and 36Cl ages versus elevation, while dashed thick dark gray lines represent the 95% confidence intervals based on n = 1000 bootstrap simulations of an ordinary least squares regression.

5.3 Surface Exposure Ages of the Pediment Surfaces

[25] In calculating exposure ages for the pediment surfaces, we did not include inheritance corrections, since our inheritance estimate is derived from fluvial sediments, and the gravels on the pediment surfaces experience a different exhumation and transport history, with potentially long transport times across the pediment surface [e.g., Nichols et al., 2005]. For this reason, the surface exposure ages of the pediment surfaces should yield maximum ages for when these surfaces were last active. The three samples from different fragments of a pediment surface that is deformed by the Kay𝚤 Fault along the southern mountain front (samples 13, 14a, and 14b; Figure 1c) yield exposure ages of 86 ± 7 ka and 70 ± 7 ka from 21Ne analyses, and 60 ± 5 ka from 10Be analyses (Tables 1 and 2). The other three samples analyzed for 10Be (samples 7 and 12a) and 36Cl (sample 12b) from different pediment surfaces deformed by the Ekinveren Fault (Figures 8a and 8b) and a frontal fold along the northern mountain front (Figure 8c) yield ages of 110 ± 10 ka, 60 ± 5 ka and 108 ± 11 ka, respectively (Tables 1, 3, and 7). The maximum ages of the surfaces demonstrate Quaternary (and possibly earlier) activity of faults and folds within the basin.

Details are in the caption following the image
(a) Map view of deformed pediment surface of sample 12ab; bedding planes from Quaternary lacustrine limestone. Contour interval is 5 m. Orange polygons represent the pediment surfaces. (b) Swath profile (light gray outlined box in Figure 8a) showing tilting of pediment surface and knickpoints along channel as a result of fault propagation toward the south. Red, orange, and blue lines indicate maximum, mean, and minimum elevations, respectively. EkF: Ekinveren Fault. (c) Oblique view of deformed pediment surface of the sample 7. Dashed lines with arrows indicate interfluves of channels deflected as a result of fold propagation toward west. Yellow lines outline the pediment boundaries.
Table 7. Surface Exposure Ages of the Pediment Surfacesa
Sample TCN Age (ka)
5 10Be 24 ± 2
7 10Be 110 ± 10
9 36Cl 178 ± 20
12a 10Be 60 ± 5
12b 36Cl 108 ± 11
13 21Ne 86 ± 7
14a 10Be 60 ± 5
14b 21Ne 70 ± 7
  • a Uncertainties are reported at the 1σ confidence level.

5.4 River Incision Rates From Strath Terraces

[26] To calculate river-incision rates based on an individual fluvial strath terrace, we divide the terrace height by its inheritance-corrected exposure (abandonment) age. The individual terraces yield incision rates ranging from 0.24 to 6.43 mm/a (Table 6).

[27] To calculate mean incision rates based on all of the fluvial strath terraces along the trunk stream of the Gök𝚤rmak River, we used a boot-strapped ordinary least squares regression of the relationship between terrace heights above the river and the inheritance-corrected abandonment ages. To quantify potential uncertainties brought about by propagating both field and laboratory measurement errors (including the effect of potential outliers), we replicated n = 1000 data subsets from (a) 10Be and (b) 36Cl by randomly drawing with replacement each data point from a normal distribution according to the means and standard deviations specified by the AMS measurements. We then calculated an ordinary least squares model for each of these 1000 boot-strap data subsets and estimated the probability density functions (pdfs) of the resulting regression slopes. We used both the original and centered data to account for the effect of a fixed intercept t = 0 (see Table S4 in the supporting information for summary). From these pdfs, we derived the medians and the 95% confidence intervals. We skipped T-tests to check for significant differences between the different models, as the inferred incision rates are rather well constrained to 0.27–0.29 mm a-1 (Figure 7).

6 Discussion

6.1 Tectonic and Climatic Controls on Fluvial Terrace Formation

[28] Fluvial incision or aggradation may result from tectonic, climate, or other local factors. These effects may include uplift or subsidence, climatically driven base-level changes, changes in discharge, changes in sediment load, and stream capture [e.g., Schumm, 1977; Hancock and Anderson, 2002; Vandenberghe, 2003; Bookhagen and Strecker, 2012]. Nonetheless, the staircase morphology of fluvial strath terraces only forms in regions where long-term uplift raises older abandoned surfaces to progressively higher elevations [e.g., Bridgland, 2000; Maddy et al., 2001; Westaway et al., 2004; Wegmann and Pazzaglia, 2009; Schildgen et al., 2012]. Because climate-related changes in sediment load and discharge tend to occur in cycles, particularly glacial-interglacial cycles, fluvial incision applied as a proxy for uplift should integrate over at least one glacial-interglacial cycle [Hancock and Anderson, 2002; Wegmann and Pazzaglia, 2002].

[29] Doğan [2011] suggested that terrace formation along the K𝚤z𝚤l𝚤rmak River in the central part of the Central Anatolian plateau was forced by orbital cycles and that the main terrace abandonment phases occurred during glacial periods. Although a number of other studies have also demonstrated correlations between MIS stages and terrace formation times in other regions [e.g., Benedetti et al., 2000; Pazzaglia and Brandon, 2001; Schildgen et al., 2012], the ages from the Gök𝚤rmak Basin (even considering the nontrivial errors) show a variable relationship with glacial and interglacial stages (Figure 7). This difference in the timing of terrace abandonment between the central part of the plateau and its northern margin suggests that variations in sediment flux and/or discharge resulting from glacial-interglacial cycles are not the only factor that controls the change between lateral and vertical river incision in the Gök𝚤rmak Basin. Rather, the Gök𝚤rmak Basin might experience vertical uplift rates that vary on the time scale of individual terrace formation, or other minor climate cycles could influence the formation of terraces in that region. Currently, however, no high-resolution uplift-rate estimates allow us to test this possibility further.

6.2 Local Influences on Incision Record

[30] Stream capture is a geomorphic phenomenon that may trigger and/or accelerate incision due to additional discharge obtained from captured streams or water bodies. This process is particularly important to consider along major strike-slip fault systems, where lateral motions through time may cause downstream river segments to capture larger or smaller drainage areas on the opposite side of the fault [e.g., Hubert-Ferrari et al., 2002] and subsequently lead to aggradation or incision as the new channel adjusts to its new length and drainage area. Hubert-Ferrari et al. [2002] and Şengör et al. [2005] proposed that the K𝚤z𝚤l𝚤rmak River established its present course across the NAF and to its outlet at the Black Sea at ca. 2 Ma. Prior to that time, the modern headwaters of the K𝚤z𝚤l𝚤rmak River appear to have been internally drained [Ardel, 1967]. Because the Gök𝚤rmak Basin drains into the lower part of the K𝚤z𝚤l𝚤rmak River between the NAF and the Black Sea, it must have experienced any changes in the elevation of its outlet to the K𝚤z𝚤l𝚤rmak River at approximately the same period of time. The presence of incised lacustrine deposits within the subbasins of the Gök𝚤rmak Basin that are connected with each other via deep gorges could have resulted from such transient adjustments following the establishment of the modern K𝚤z𝚤l𝚤rmak channel. Nonetheless, the ~2 Ma estimate for the timing of the capture of the K𝚤z𝚤l𝚤rmak headwaters is much earlier than our oldest terrace abandonment age of 346 ka, implying that the incision recorded in the Gök𝚤rmak Basin since 346 ka is not likely associated with capture-related processes.

[31] Although we rule out effects of stream capture as an important influence on the incision record, other local effects may have been significant, particularly along the very narrow valleys of the K𝚤z𝚤l𝚤rmak River (Figure 3d). For example, recent processes such as landsliding and/or rapidly changing sediment loads within the K𝚤z𝚤l𝚤rmak Valley may have led to temporary blocking/damming and reincision of the Gök𝚤rmak Valley. The alluvial-fan delta deposits (Figure 3e) that we observed on the upstream side of landslides at the outlet of the Gök𝚤rmak Basin indicate that a paleo-fan delta was deposited in a water body that was probably dammed at the K𝚤z𝚤l𝚤rmak Valley. The very young exposure age (~7 ka) and high incision rate (6.43 mm a-1) that we obtained from a strath terrace at the outlet of the Gökirmak Basin might be related to either shielding of the terrace surface from cosmic rays as a result of such damming and subsequent reincision of the gorge or from the creation of a new epigenetic gorge if the river incised through the bedrock in a different location compared to the earlier valley [e.g., Ouimet et al., 2008]. For that reason, we excluded that sample when performing the boot-strap regression to determine the mean regional incision rate.

6.3 Deformed Pediment Surfaces and Regional Patterns of Crustal Deformation

[32] Pediment surfaces are generally smooth and gentle surfaces inclined toward the basin. They are geomorphic markers of the base-level stabilities across the northern margin. The presence of the deformed pediment surfaces (Figures 8a and 8b) along the Ekinveren Fault indicate upper crustal shortening in the Central Pontide wedge related to thrust faults. Our TCN data from deformed pediment surfaces (Figures 8a and 8b) reveal that thrust faulting continued at least through Late Quaternary time.

[33] This deformation is not only restricted to the Gök𝚤rmak Basin but extends across the array of ranges between the NAF and the Black Sea (Figure 1d). Seismic reflection profiles [Finetti et al., 1988] combined with morphometric data [Yildirim et al., 2011] show hinterland propagation of the deformation by out-of-sequence faulting, implying that upper-crustal shortening takes place on faults spanning the width of the wedge. The out-of-sequence nature of deformation is underscored by undeformed Pliocene and Quaternary sediments covering the outermost thrust front [Finetti et al., 1988].

[34] According to wedge theory, an orogenic wedge will advance toward its foreland and lengthen itself when the sum of its basal and upper slopes exceeds a critical taper value; conversely, if the sum of the slopes is less than the critical value, internal shortening will thicken the wedge and build topography until it reaches the critical taper [Dahlen, 1990; Mitra, 1997; DeCelles and Mitra, 1995]. Especially after focused erosion reduces topographic slope and drives the wedge into a “subcritical” state, the wedge may attempt to regain taper through out-of-sequence thrusting or duplexing before propagating outward [Dahlen, 1990]. Given our new evidence for ongoing out-of-sequence deformation in the Central Pontides and the inferred subcritical state of the orogenic wedge, we conclude that the Central Pontides are currently in a state of rebuilding topography that may have been lost earlier through basin exhumation and spatially focused erosion.

[35] Finally, with respect to the overall evolution of the Central Anatolia Plateau, our observations of uplifted and deformed geomorphic reference horizons unambiguously show that the lateral growth of the plateau has continued into the Quaternary and is ongoing. Continued vertical growth of the interior part of the wedge may ultimately result in continued northward orogenic growth and possible reactivation of thrust systems at the periphery of the Pontide wedge.

6.4 Differential Uplift Between the Northern Margin and the Central Part of the Plateau

[36] Even in the absence of local influences on the incision record, the mean fluvial incision rate does not always match the mean uplift rate in actively deforming areas, particularly if fluvial response times are slow, as may happen in the upper reaches of a large river catchment. Because no significant knickpoints occur between our terrace locations and the river outlet, the river incision rate should be adjusted to recent uplift rates [e.g., Whipple and Meade, 2004], implying that our mean incision rate derived from fluvial strath terraces reflects the mean uplift rate in the Gök𝚤rmak Basin. Also, uplifted Quaternary paleo-deltas of the K𝚤z𝚤l𝚤rmak River, 120 km downstream from the outlet of the Gök𝚤rmak Basin, yield an inferred uplift rate of 0.2–0.3 mm a-1 during the last 346 ka based on correlations between delta formation and relative sea level highstands [Demir et al., 2004]. The similarity between this coastal uplift rate and our 0.27–0.29 mm a-1 mean incision rate along the trunk stream of the Gök𝚤rmak Basin over the last 346 ka (Figure 7) further supports our inference that the mean incision rate reflects the mean uplift rate along the northern margin of the Central Anatolian Plateau over the past ~350 ka.

[37] In contrast to the correspondence between our incision rates from the northern plateau margin and uplift rates along the coast, river terraces along the K𝚤z𝚤l𝚤rmak River in the central part of the plateau (Figure 1a) record an incision rate that is three times lower, i.e., 0.08 mm a-1 averaged over the last 400 ka (Figure 9) [Doğan, 2011]. This difference between the central part of the plateau and its northern margin indicates that either the northward-flowing K𝚤z𝚤l𝚤rmak River is currently in a transient state, with portions of the river upstream from the Gök𝚤rmak Basin not yet adjusted to the faster recent uplift rates or that faster uplift has occurred at least since the Late Pleistocene time along the northern margin of the plateau compared to the central region. Without independent estimates of the uplift rate in the central plateau region, we cannot currently discriminate between these two possibilities.

Details are in the caption following the image
Cross section illustrating conceptual model for the Central Pontides in the form of an active orogenic wedge (modified after Yildirim et al. [2011]) together with river incision rates from the interior part of the Central Anatolian Plateau (from Doğan [2011]) and from the Gök𝚤rmak Basin. BF: Bal𝚤fak𝚤 Fault, EF: Erikli Fault, EkF: Ekinveren Fault, KF: Kay𝚤 Fault, NAF: North Anatolian Fault, TF: Tuz Gölü Fault. Arrows indicate shortening in the Pontides and extension in the Central Anatolian Plateau.

7 Conclusions

[38] Despite potential complexities involving climate cycles and local river incision processes, the patterns of deformation and rates of river incision in the Gök𝚤rmak Basin support the notion that active deformation and uplift are the primary factors driving fluvial incision of the northern margin of the Central Anatolian plateau. Terrestrial cosmogenic nuclide-derived abandonment ages of fluvial strath terraces along the trunk channel of the Gök𝚤rmak River imply that long-term average incision rates over the past ~350 ka were 0.27–0.29 mm a-1. Surface exposure ages of deformed pediment surfaces ranging from ~60 to 110 ka demonstrate that thrust faults responsible for the pediment deformation have been active at least since that time and were likely active much earlier as well to explain the amount of topographic relief in the region. Any inference about earlier activity, however, implies that the pediments formed during a relatively recent phase of base level stability and/or tectonic quiescence. Because the average incision along the Gök𝚤rmak River matches uplift rates inferred for deltaic deposits at the coast, and because the deformed pediment surfaces correspond to locations above thrust faults, we interpret incision within the Gök𝚤rmak Basin to be driven primarily by differential uplift through thrusting. This active shortening, coupled with a regional out-of-sequence faulting pattern, implies that the Central Pontides orogenic wedge is actively growing by internal shortening.

8 Acknowledgments

[39] This work is part of the Vertical Anatolian Movements Project, funded by the TOPO-EUROPE initiative of the European Science Foundation, including contributions by the German Science Foundation (DFG) to MS (grants STR373/20-1 and 25-1). CY was supported by Potsdam University and Deutsches GeoForschungsZentrum Potsdam; HE was supported by DFG. TS was supported by the Leibniz Center for Surface Processes and Climate Studies at the University of Potsdam (DFG: STR373/20-1) and the Alexander von Humboldt Foundation. BB was supported with grants from NASA (NNX08AG05G), AC was supported by TÜBITAK (107Y333). We thank E. Schnabel for performing noble gas analyses; A. Musiol, K. Hahne, and H. Rothe for major and trace element measurements; A. Landgraf, I. Schimmelpfennig, and A. Hidy for discussions on cosmogenic nuclides and depth profiles; B. Rojay, U. Doğan, P. Ballato, S. Hoth, and C. Kunze for discussions of regional geology and geomorphology; and Y. Garcin and O. Korup for help with the statistical analysis and discussions.

    Erratum

    In the originally published version of this article, the addresses of affiliations were correct but numbering of authors according to their affiliations was not. In Tables 1, 3, and 4, coordinates of three samples and digits of cosmogenic nuclide concentrations were incorrect. Although the results of the paper are correct the wrongly written coordinates and digits of concentrations may mislead the readers about the calculations. The errors have since been corrected and this version may be considered the authoritative version of record.