Earth and Planetary Science Letters 477 (2017) 168–182

Contents lists available at ScienceDirect

Earth and Planetary Science Letters

www.elsevier.com/locate/epsl

Early Neogene foreland of the Zagros, implications for the initial closure of the Neo-Tethys and kinematics of crustal shortening

Mortaza Pirouz a,∗, Jean-Philippe Avouac a, Jamshid Hassanzadeh a, Joseph L. Kirschvink a, Abbas Bahroudi b,c

a Division of Geological and Planetary Sciences, California Institute of Technology, Pasadena, CA 91125, USAb Exploration Department, Mining Engineering Faculty, University College of Engineering, University of Tehran, North Karghar Street, P.O. Box 1439, Tehran 957131, Iranc Department of Geoscience, University of Calgary, 2500 University Drive NW, Calgary, AB T2N 1N4, Canada

a r t i c l e i n f o a b s t r a c t

Article history:Received 27 May 2017Received in revised form 23 July 2017Accepted 25 July 2017Available online 9 September 2017Editor: A. Yin

Keywords:Neo-Tethys closureArabia–Eurasia collisioncrustal shorteningZagros foreland dynamicsmagnetostratigraphystrontium isotope stratigraphy

We study the transition from passive margin to foreland basin sedimentation now exposed in the High Zagros belt to provide chronological constraints on the initial stage of Arabia–Eurasia collision and closure of the Neo-Tethys. We performed magnetostratigraphy and strontium isotope stratigraphy along two sections near the Zagros suture which expose the oldest preserved foreland deposits: the Shalamzar section in the west and the Dehmoord section in the east. The top of the passive margin Asmari formation has an age of 28–29 Ma in the High Zagros and is overlain by foreland deposits with a major basal unconformity representing 7 Myr of hiatus. The base of the foreland deposits has an age of 21.5 Ma at Dehmoord and ca. 26 Ma at Shalamzar. The sedimentation rate increased from 30 m/Myr in the passive margin to 247 m/Myr in the foreland. Combined with available age constraints across the Zagros, our results show that the unconformity is diachronous and records the southwestward migration of the flexural bulge within the Arabian plate at an average rate of 24 ± 2 mm/yr over the last 27 Ma. The time evolution of sediment accumulation in the Zagros foreland follows the prediction from a flexural model, as the foreland is thrust beneath the orogenic wedge and loaded by the wedge and basin fill. We detect the onset of forebulge formation within the Asmari Formation around 25 Ma. We conclude that closure of the Neo-Tethys formed the Zagros collisional wedge at 27 ± 2 Ma. Hence, the Arabia–Eurasia collision was probably not the main driver of global cooling which started near the Eocene–Oligocene boundary (ca. 33.7 Ma). We estimate 650 km of forebulge migration since the onset of the collision which consists of 350 km of shortening across the orogen, and 300 km of widening of the wedge and increasing flexural rigidity of Arabia. We conclude the average rate of shortening across the Zagros to be ca. 13 mm/yr over the last 27 Myr; a value comparable to the modern rate. Palinspastic restoration of structural cross-sections and crustal volume conservation comprise only ca. 200 km of shortening across the Zagros and metamorphic Sanandaj–Sirjan belt implying that at least 150 km of the Arabian crust was underthrust beneath Eurasia without contributing to crustal thickening, possibly due to eclogitization.

© 2017 Elsevier B.V. All rights reserved.

1. Introduction

The timing of the Arabia–Eurasia collision along the Alpine–Himalayan orogen (Fig. 1) is highly controversial, with estimated ages ranging between 5 and 35 Ma (e.g., Allen and Armstrong, 2008; Bird et al., 1975; Dewey et al., 1973; McQuarrie and van Hinsbergen, 2013; Mouthereau et al., 2012; Saura et al., 2015). A more robust constraint of the timing is crucial for establishing

* Corresponding author.E-mail addresses: mpirouz@caltech.edu, mortaza.pirouz@me.com (M. Pirouz).

http://dx.doi.org/10.1016/j.epsl.2017.07.0460012-821X/© 2017 Elsevier B.V. All rights reserved.

correlations with shifts in climate and plate tectonic patterns. For example, form a climate perspective, it has been proposed that the closure of the Neo-Tethys drove Cenozoic cooling (Allen and Armstrong, 2008) because of the impact on volcanism, the draw-down of atmospheric CO2 due to chemical weathering and burial of organic carbon associated with erosion of the collisional orogen, and the shutdown of the oceanic connection between the Indian and Atlantic Oceans. From a geodynamic perspective, several ma-jor regional changes occurred in the time span of the collision: (1) the upwelling of the Afar plume and the opening of the Red Sea around 30 Ma (Bosworth et al., 2005), (2) the slowdown of

M. Pirouz et al. / Earth and Planetary Science Letters 477 (2017) 168–182 169

Fig. 1. Geological setting of the Zagros collision zone with locations of the dataset used in this study. White squares show locations of the studied magnetostratigraphic sections. Locations of previous published magnetostratigraphic studies are shown with yellow circles (Emami, 2008; Homke et al., 2004; Khadivi et al., 2010; Ruh et al., 2014). Abbreviations are: Af, Afrineh; Cg, Chaman–Goli; Cm, Chahar–Makan; Cn, Changuleh; Go, Gol–Gol; Zn, Zarinabad. Green circles show borehole locations and outcrops used for strontium isotope stratigraphy of the Asmari Formation (Ehrenberg et al., 2007; Van Buchem et al., 2010). Labels 1, 2, 3, 4, 9 and 12 refer to the boreholes and Aw, As, Ch, Iz, Kt and Sh stand for Ahwaz, Asmari, Chidan, Izeh, Katoola and Shalamzar sections. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

northward motion of Arabia relative to Eurasia between 28 Ma and 20 Ma (McQuarrie and van Hinsbergen, 2013), and (3) a possible more recent 15–20% slowdown of Arabia suggested by the compar-ison of modern kinematics from GPS geodesy with late Quaternary plate kinematics (e.g. Nocquet et al., 2006). At the more local scale of the Zagros and Iranian Plateau, there is geological evidence for a Paleogene volcanic flare-up that ceased around 40 Ma (Verdel et al., 2011) possibly due to slab steepening and final break-off in the Miocene (Omrani et al., 2008). These various events could all be related to a reorganization of mantle convection linking the Afar plume and slab roll back in the Aegean (e.g., Faccenna et al., 2013), and to the effect of the collision and slab break-

off (Austermann and Iaffaldano, 2013). Establishing the causative links between these various geodynamic events and their relation to climate relies critically on their relative chronology. Improved chronology of the collision is also important for calculating the amount of crustal shortening and underthrusting of Arabian litho-sphere since the closure of the Neo-Tethys.

In this study, we focus on improving the constraints on the initial stage of mountain building in the Zagros by investigating its earliest foreland basin deposits. We assume that the foreland evolves in relation to the orogenic wedge following a framework established from studies in other tectonic contexts (e.g., DeCelles and Horton, 2003). This model postulates that the foreland basin

170 M. Pirouz et al. / Earth and Planetary Science Letters 477 (2017) 168–182

Fig. 2. Schematic representation of forebulge migration due to the thrusting of Eurasia over Arabia following the Zagros collision. Initial (a) and advanced (b) stages of a foreland basin development are shown in a Arabia fixed framework. The section (b) stems from seismic data. “L” shows migration of the bulge from its initial location “Bi” to its present location “Bf”. The forebulge results in an unconfromity (a hiatus represented by the dashed line) between the foreland basin and the passive margin rock units. “F” stands for foreland width and “�W” stands for wedge shortening. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

forms in response to flexure of the lithosphere as it is loaded by the topographic wedge and sediment accumulation (Fig. 2). At the onset of the collision, the foreland must have developed as a result of loading by the Zagros wedge topography and sed-iment filling. A forebulge, separating a back-bulge basin from the foredeep, must have formed and lead to the development of an erosional unconformity. The forebulge must have migrated south-westward within the Arabian plate as a result of underthrusting of the Arabian lithosphere and growth of the Zagros wedge. Sed-iment accumulation at any given location of the foreland con-tinued before incorporating into the wedge. The ancient foreland has thus been partly incorporated into the Zagros wedge and has archived this history which is schematized in Fig. 3. Various studies have documented elements of this history (Emami, 2008;Fakhari et al., 2008; Homke et al., 2004; Khadivi et al., 2010;Pirouz et al., 2016; Ruh et al., 2014; Saura et al., 2015).

In this study, we focus on the erosional unconformity at the base of the foreland deposits which can be investigated from field observations and subsurface data from exploration wells or seis-mic profiles. The stratigraphic location of this unconformity al-lows tracking of the contact between the foreland deposits as it gradually transgressed over the underlying passive margin de-posits, and aids deciphering of the formation of the back-bulge and forebulge syn-kinematic deposits (Fig. 3). To constrain this kinematics, we have spotted the exposures of the rarely pre-served oldest foreland deposits and the underlying shelf sediments along two sections in the High Zagros belt. We collected samples from these sections to constrain their ages by magnetostratigra-phy and strontium isotope stratigraphy. We report these results hereafter and combine them with data compiled from the bib-liography across the Zagros wedge (e.g., Ehrenberg et al., 2007;Fakhari et al., 2008; Khadivi et al., 2010; Pirouz et al., 2016;Van Buchem et al., 2010) to derive a kinematic model of the fore-

land evolution, post-collisional shortening and timing of the Neo-Tethys Ocean closure.

2. Tectonic and stratigraphic setting

The Zagros collisional zone includes deformed margins of Ara-bia and Eurasia, and is divided into three main tectonic units parallel to the suture (Fig. 1). From north to south, they are the Urumieh–Dokhtar Cenozoic magmatic arc and the Sanandaj–Sirjan Mesozoic magmatic arc on the Eurasian plate, and the sedimentary Zagros fold–thrust belt of the Arabian plate. The Zagros basin itself can be divided into four main structural units separated by ma-jor faults, which are the High Zagros imbricated belt, the Zagros simply folded belt, the Zagros foredeep (Dezful and Kirkuk em-bayments) and the Mesopotamian–Persian Gulf basin toward the southwest (Fig. 1). The High Zagros belt forms the edge of Ara-bia and includes peaks reaching elevations >4 km. It is bounded by the Zagros suture in the north and the High Zagros fault in the south (Fig. 1). The Paleozoic rocks are largely thrust over the younger deposits and the structures are mostly tight or over-turned representing accommodation of significant shortening (see schematic section in Fig. 2), while it is not seismically active at present.

Neoproterozoic–Phanerozoic stratigraphy of the northern mar-gin of the Arabian plate is divided into four main cycles (e.g., Alavi, 2004). The late Neoproterozoic to Devonian cycle con-sists of a thick unit of evaporites, volcanic, and sedimentary strata at the base, overlain by siliciclastic and carbonate rocks, which were deposited on an epicontinental platform and within pull-apart basins. The Permian to Triassic cycle consists of epi-Pangean platform deposits with conglomerates and sandstones at the base, followed by coastal to shallow marine and suprati-dal and sabkha deposits. The Jurassic to Upper Cretaceous cy-

M. Pirouz et al. / Earth and Planetary Science Letters 477 (2017) 168–182 171

Fig. 3. Conceptual model for sedimentary facies and basal unconformity ages as a function of distance from the forebulge. a) Sediments coarsen from forebulge toward the foredeep and wedge-top. b) Non-linear increasing sedimentation rate in the basin as a function of distance from forebulge. In a steady-state system, the sedimentation rate is determined by the shape of flexure basement and the rate of forebulge migration. Color gradation shows upward younging. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

cle consists of the Neo-Tethyan continental shelf deposits rep-resented by mostly carbonate and marl deposits underlain by the Early Jurassic lagoonal, near-shore and sabkha deposits. The Cenozoic cycle consists of two distinct Paleogene and Neogene foreland basins separated by the massive Oligo-Miocene As-mari platform carbonates. This Formation is a thick, long-lived limestone unit which accumulated between 34 and 16.5 Ma (Ehrenberg et al., 2007; Homke et al., 2009; Pirouz et al., 2016;Van Buchem et al., 2010). The Asmari Formation is separated from the older units by a major hiatus between 60 and 35 Ma, in Fars and Lorestan regions (Mouthereau et al., 2012 and refer-ences therein). This unconformity probably formed as a result of the Neyriz and Kermanshah ophiolite obduction. The base of the Asmari is more transitional in the Dezful embayment and Izeh zone (Ehrenberg et al., 2007; Van Buchem et al., 2010) (see loca-tions in Fig. 1).

The Paleogene foreland deposits include fluvial to shallow marine sediments uniquely dominated by detritus from the ob-ducted oceanic lithosphere (i.e., Neyriz and Kermanshah ophi-olites). Other than the contrasting provenance, the two sys-tems have quite different extensions; the Paleogene foreland

deposits are located around the ophiolites (i.e. Homke et al., 2010), whereas the Neogene system is orogen-wide (Pirouz et al., 2011). The Neogene distal foreland deposits exhibits a con-tinuous and largely gradational passage from basal supratidal and sabkha sediments (Gachsaran Formation) to carbonates and marine marls (Mishan Formation with its basal Guri carbonate member) followed by coastal plain and meandering river de-posits (Agha Jari Formation) and finally to braided river gravel sheets (Bakhtiari Formation) in the simply folded belt and to-ward the south. The Neogene proximal foreland deposits in the High Zagros belt exhibit almost the same stratigraphy, except that the basal sediments, i.e. fluvial red marls with conglomer-ate interbeds (Razak Formation), were deposited in a continental environment (Pirouz et al., 2011). The Neogene vertical succes-sion of facies is interpreted to represent the southward migra-tion of foreland basin depozones (from back-bulge to foredeep and wedge-top, respectively) as the Zagros fold–thrust belt mi-grated progressively southwestward (Pirouz et al., 2011). As a result of this process the contacts between these stratigraphic units, which are classified based on facies, are diachronous. A number of studies have already provided chronological con-

172 M. Pirouz et al. / Earth and Planetary Science Letters 477 (2017) 168–182

Fig. 4. Geological maps of Dehmoord (a) and Shalamzar (b) with sampling sites for magnetostratigraphy (black dots). Locations are shown in Fig. 1.

straints on this stratigraphy (Emami, 2008; Fakhari et al., 2008;Homke et al., 2004; Khadivi et al., 2010; Pirouz et al., 2015;Ruh et al., 2014).

3. Age constraints of the foreland deposits in the High Zagros

3.1. Description of investigated sections (Dehmoord and Shalamzar)

Previous magnetostratigraphic sections are located in the sim-ply folded belt in Fars and Lorestan at distances of 50–120 km from the suture (Emami, 2008; Homke et al., 2004; Khadivi et al., 2010; Ruh et al., 2014) (Fig. 1). We focused our own investiga-tions on sections closer to the suture zone, in the High Zagros, where the foreland deposits are presumably older. We identified two areas very close to the Zagros suture with relatively well-exposed outcrops of the transition from pre-collisional to post-collisional deposits (Fig. 4). One section is located to the northwest of the Neyriz ophiolites and adjacent to the village of Dehmoord in the Fars Province (Figs. 1 and 5a). The stratigraphic section is ca. 1200 m thick and exposes a continuous depositional record. These sedimentary units are mapped as Neogene flysch deposits in the geological map of Neyriz area (Sabzehei et al., 1993). They are strikingly similar to the Zagros foreland deposits. The sequence consists of basal fine grained red fluvial deposits with conglomer-ate intercalations similar to the Razak Formation, followed by shal-low marine deposits similar to the Mishan Formation with basal

Guri limestone, delta plain to meandering river and flood plain de-posits similar to the Agha Jari formation and braided river deposits comparable to the Bakhtiari Formation. The Dehmoord section thus represents the typical progressive evolution from distal to proximal facies observed in the entire foreland basin of the Zagros.

The Shalamzar section, in the west, is located about 30 km south of Shahr-e-Kord (Figs. 1 and 5b). It exposes a ca. 800 m thick sequence showing clear upward coarsening. We identified the Razak Formation at the base of the section, which mainly consists of fine grained multi-colored mudstone with evaporitic interbeds. This formation is deeply eroded and largely covered by quaternary deposits forming farmlands. The Agha Jari Formation is also eroded and covered by younger deposits and not suitable for core sam-pling, while the Bakhtiari Formation has an extraordinary exposure with marine intercalations including reefal limestone. This section has been investigated by Fakhari et al. (2008), who provide fos-sil age constraints for the Bakhtiari marine interbeds. The shallow marine foredeep deposits equivalent to the Mishan Formation do not exist in the western High Zagros.

3.2. Magnetostratigraphic constraints

For the magnetostratigraphic investigation of the Dehmoord section in the eastern Zagros, we collected a total of 114 oriented core samples using portable drills and standard magnetic and solar compass techniques. Of these, 68 cored samples (2.5 cm diameter

M. Pirouz et al. / Earth and Planetary Science Letters 477 (2017) 168–182 173

Fig. 5a. Stratigraphy, declination and chronology of the Dehmoord section in the eastern Zagros determined from the magnetostratigraphic results and the strontium (Sr) isotopes. The horizontal time line shows the reference geomagnetic polarity time scale (GPTS, Gradstein et al., 2012). Periods of normal (inverse) polarities are shown in black (white). Constrained time zones with one data point are shown with gray. Details of the magnetic measurements are given in supplements (Figs. S1 to S5). Strontium isotopic measurements are converted to ages using a global reference curve (McArthur et al., 2001) and are shown with green dots and their relevant error bars. Details are given in the supplements (Fig. S6 and Table 1). The misfit between the ages derived from the Sr isotopes and the magnetostratigraphy is ca. 0.5 Ma. The average sedimentation rate is estimated at 185 m/Myr to 247 m/Myr. This rate is not corrected for compaction which could be of the order of 25–30% due to large volume marl units. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

cylinders) were taken from indurated strata and 46 tube samples were obtained from soft and fragile units. Tube sampling was used because a large fraction of the Neogene strata consists of soft marl deposits in the High Zagros. In the western sector at Shalamzar, 62 oriented core samples were collected from the Bakhtiari (ca. 200 m) and upper part of the Asmari (ca. 145 m) Formations. The average sampling intervals were 7 m for the Bakhtiari, 4 m for the Asmari Formations in the western sector, and 11 m in the eastern sector. Geomagnetic polarity chrons have had average durations longer than 100 ka since Early Miocene, therefore the sampling strategy would detect most polarity intervals. However, partially covered outcrops and the lack of fine-grained matrix and sand-stone interbeds in the Bakhtiari conglomerate prevented uniform sampling along the section, and caused some sampling gaps. In both sections fine-grained units were sampled, and in conglomer-atic intervals we collected fine-grained matrix or sandstone lenses when present.

Paleomagnetic measurements were conducted at Caltech using 2G™ Enterprises model 755 superconducting rock magnetometers,

with 3-axis DC SQUIDS with a sensitivity of ca. 2 × 10−13 Am2, coupled with automatic sample changers and quartz-glass vacuum sample holders housed in a double-layer, mu-metal and soft-steel shielded rooms (Kirschvink et al., 2008). We followed a hybrid demagnetization strategy designed to isolate secondary and vis-cous magnetic components from possible characteristic directions (ChRMs). First, we measured the natural remanent magnetization (NRM) of samples after a few days storage in the magnetically-shielded rooms. Next, we conducted one or more low-temperature thermal cycles in liquid nitrogen to reduce the contribution of multi-domain magnetite by cycling it through the Verwey transi-tion, followed by low-intensity progressive 3-axis alternating field (AF) demagnetization at 2.3, 4.6 and 6.9 mT steps to help re-move viscous components of multi-domain magnetite grains and soft magnetic components that might have been acquired by ex-posure to moderate magnetic fields during sample preparation. After these steps, samples were subjected to 30 min of progres-sive thermal demagnetization under a gentle flow of N2 gas in a magnetically-shielded ASC oven (<25 nT residual field). Stepwise

174 M. Pirouz et al. / Earth and Planetary Science Letters 477 (2017) 168–182

Fig. 5b. Stratigraphy, declination and chronology of the Shalamzar section in the eastern Zagros determined from the magnetostratigraphic results and the strontium isotopes. Average sedimentation rate is about 30 m/Myr in the Asmari Formation, and 180 m/Myr in the Bakhtiari Formation. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

demagnetization started at 50 ◦C and increased in steps of 10◦to 25 ◦C until the magnetic field of the samples were too weak, or they displayed unstable behavior (ca. 400 ◦C to 650 ◦C). Re-manence data after each demagnetization procedure were viewed using standard orthogonal projections (Figs. S1 and S2) and an-alyzed with principal component analysis (Kirschvink, 1980). At this point, direction of the preserved magnetic field either main-tains or gradually moves toward more characteristic remanence (ChRM) directions, N–S direction in both sections (Fig. S1 and Ta-ble S1). Most specimens present a large unidirectional magnetic component up to 250–300 ◦C when ca. 80% of the magnetic field is altered (Figs. S2 and S3). Characteristic remanent magnetiza-tion (ChRM) directions were selected using the standard criteria

in which the maximum angular deviation (MAD) is less than 10◦(Fig. S4). We also conducted rock magnetic investigations to re-veal the nature of the present magnetic minerals (Fig. S5), and to help interpret the demagnetization patterns. Demagnetization of the ChRM displays both linear and arc trends toward the ori-gin with maximum unblocking temperatures of typically ca. 350 ◦C for carbonates and 580 ◦C for sandstone and red marls (Fig. S1). Apparently, their remanence was carried by magnetite and/or ti-tanomagnetite with Curie temperatures near 580 ◦C, in agreement with the coercivity spectra from the rock-magnetic data (Fig. S5). The second magnetic component, high temperature, is short and is limited somewhere between 300◦ and 500 ◦C, with a few spec-imens remaining stable up to 650 ◦C (Fig. S2). Combined line and

M. Pirouz et al. / Earth and Planetary Science Letters 477 (2017) 168–182 175

plane analysis of McFadden and McElhinny (1988) was applied to obtain the best sable end point on the arcs for specimens with re-verse polarity. Results of the Principal Component Analysis (PCA) on the demagnetization data from the Shalamzar and Dehmoord sections are shown in Fig. S6. A reversal test, conducted for both geographic and tilt corrected ChRMs, reveals an angle γc−g = 13.6◦for geographic and γc−t = 13.3◦ for tilt corrected data of the west-ern (Shalamzar) section. According to 10◦ < γc < 20◦ , the reversal test of the western section is passed and classified as “C class”. In the eastern (Dehmoord) section, γc−g = 8.9◦ and γc−t = 9.3◦ and according to 5◦ < γc < 10◦ , is classified as “B class” (McFadden and McElhinny, 1990). More details are provided in the supplementary materials.

3.3. Chronostratigraphic constraints based on strontium isotopes

The 87Sr/86Sr ratio of sea water has changed significantly through time in response to the input of varying proportions of strontium derived from continental crust and upper mantle sources and carbonate recycling (e.g., Elderfield, 1986). The vari-ation of the 87Sr/86Sr ratio of sea water versus geological time can be used for stratigraphic dating of marine carbonates based on the correlation with a global standard curve (e.g., McArthur et al., 2001). The technique is most effective for the Cenozoic due to the rapid monotonous increase of the 87Sr/86Sr ratio caused by a grad-ual increase of terrestrial input (Fig. S7a). Hence, it provides age constraints within the range of ±1 Ma (e.g., McArthur et al., 2001). The method applies to marine minerals based on the assumption that the world ocean is homogeneous with respect to 87Sr/86Sr. Such uniformity of 87Sr/86Sr is expected, because the present residence time of Sr is ca. 1 Ma (Gradstein et al., 2012). This technique has already been applied to constrain the age of car-bonates in the Zagros (Ehrenberg et al., 2007; Pirouz et al., 2015;Van Buchem et al., 2010).

We collected fifteen samples of limestone and shell fragments from the top of the Asmari–Jahrom Formation and the Guri Mem-ber and marine sub-lithofacies association of the Bakhtiari for measuring strontium ratios. Details of sample preparation and laboratory procedure are described by Pirouz et al. (2015). The numerical ages of the samples were determined from the mea-sured 87Sr/86Sr values on the 87Sr/86Sr LOWESS (LOcally WEighted Scatterplot Smoothing) curve (Table S2, Fig. S7b) calibrated by McArthur et al. (2001). These ages represent the maximum ages of the fossils and they correspond to depositional ages of sediments if no reworking has taken place. The LOWESS best-fit curve gives robust statistical uncertainties on any measured age (McArthur et al., 2001). The uncertainty of ages includes uncertainty of the Sr global seawater curve, lab measurements, measuring tools and any isotopic heterogeneity.

3.4. Age constraints for the foreland basin initiation in the High Zagros

Correlation of our observed magnetic polarity pattern with the standard geomagnetic polarity time scale (GPTS), is shown in Figs. 5a, 5b and S6. Constraints from Sr isotope stratigraphy have been used as anchor points. Our field observations and strontium chronological constraints along the Dehmoord section (eastern Za-gros) show that the top of the Asmari–Jahrom platform carbon-ate is marked by an erosion surface that formed after 27.6 Ma. This hiatus has a regional character across the eastern Zagros and displays diachronous ages between 25.5 and 18 Ma in the simply-folded Fars and between 17 and 9 Ma in the distal Fars (Pirouz et al., 2016). Above the unconformity, the magnetostrati-graphic constraints imply an age range of 21.5 Ma to 20 Ma for the basal foreland strata, i.e., the Razak Formation. The upper con-tact of the Razak fluvial deposits shows transition to shallow open

marine deposits (i.e., the Mishan Formation) for which we have obtained another set of Sr ages (Table S2, Fig. S7b). This unit is diachronous similar to the Agha Jari and Bakhtiari foreland de-posits. Transition from the Razak formation to the younger Mis-han formation is easy to identify due to the change from red to green marls with a basal fossil-rich limestone. The marine strata have an age of 20 Ma to 18.5 Ma in the Dehmoord region of the High Zagros, however, to the south it varies from 16.5 Ma to 1 Ma (Pirouz et al., 2015). The Agha Jari Formation mainly consists of repeated red marl and sandstone beds representing a fluvial environment, but also includes grayish–greenish marl and thick sandstone units of delta environments, deposited between 18.5 Ma and 16.7 Ma. The uppermost foreland sediments, the Bakhtiari conglomerate, yield an age range from 16.7 Ma to 16 Ma (Fig. 5a). The diachroneity of the foreland deposits documented in this investigation is consistent with the diachronous ages re-ported in previous studies which suggest younger ages toward the southwest in Fars (Khadivi et al., 2010; Pirouz et al., 2015;Ruh et al., 2014).

Correlations between the measured sections at Shalamzar in the western Zagros and the GTPS were also straightforward, thanks to the anchor points provided by the Sr isotope data. The top of the Asmari formation has an age of 29 Ma (Fig. 5b), and is separated from the foreland deposits by the above mentioned unconformity. The basal foreland strata, i.e. the Razak and the Agha Jari Forma-tions with total thickness of around 500 m, are mostly covered by the Quaternary deposits in the western High Zagros. The base of the Bakhtiari conglomerate has an age of 23.5 Ma, while its marine middle part has an age of 22 Ma (Fig. 5b). Our obtained ages agree with the fossil ages reported by Fakhari et al. (2008). The covered 500 m of basal foreland strata represents a 2.7 Myr record by as-suming an average sedimentation rate of 180 m/Myr. Therefore, the unconformity between the carbonate unit and the foreland de-posits is constrained between 29 Ma and ca. 26 Ma.

The age constraints additionally show a rapid increase in sed-imentation rate from about 30 m/Myr within the Asmari to aver-age of 180 m/Myr within the Bakhtiari Formation in the western High Zagros, and up to average of 185 m/Myr in the Razak, and 247 m/Myr in the Aga Jari and Bakhtiari Formations at the Dehmo-ord section in the eastern High Zagros (Figs. 5a and 5b). This increasing sedimentation rate with time is consistent with our in-terpretation that the erosion surface above the Asmari formation developed in the forebulge (Fig. 3), and that the foreland deposits gradually accumulated on top of the unconformity as the Zagros wedge was thrust southwestward over the Arabian plate.

4. Discussion

4.1. Nature and age of the forebulge unconformity

The unconformity associated with the forebulge which is doc-umented in this study is a hiatus of deposition, a disconfor-mity, with diachronous age younging cratonward. The unconfor-mity reaches variable stratigraphic depths within the Asmari For-mation (Asmari–Jahrom in Fars, Asmari–Shahbazan in Lorestan). In the Dehmoord section, eastern Zagros, the unconformity is mani-fested as karstification in the Asmari–Jahrom limestone and laterite formation seen at the base of the foreland deposits. The fore-bulge unconformity separates the Asmari–Jahrom limestone from the Razak fluvial deposits at 27.6 Ma in the Dehmoord section, and at 25.5 Ma in the Kaftar section. Near the coast of the Per-sian Gulf, in the Asaluyeh section, the unconformity lies between the Champeh member of the Gachsaran Formation and the Mis-han Formation and is younger, ca. 17 Ma (Fig. S8). Our estimate of the duration of the hiatus is about 7 Myr in the eastern Zagros. In the western Zagros, in the Dezful embayment and Lorestan arc,

176 M. Pirouz et al. / Earth and Planetary Science Letters 477 (2017) 168–182

Fig. 6. Balanced structural cross-sections across the Zagros wedge and the Persian Gulf foredeep. Sections are based on seismic data. Section a (between suture and Shadeghan borehole) and section b are modified after Sherkati et al. (2006). Section c is modified after Jahani et al. (2009). Locations are shown in Fig. 1.

the forebulge unconformity is situated at the boundary between the Asmari and Gachsaran Formations. The hiatus duration is not well constrained in the western sector due to ductile behavior of the Gachsaran evaporites; strontium isotope stratigraphy however suggests the hiatus could be much shorter (ca. 1 Myr) or nonexis-tent (Ehrenberg et al., 2010; Van Buchem et al., 2010).

The forebulge unconformity is not angular as expected, given that the amplitude of the bulge does not exceed above 300 m, and the width is ca. 150 km (Pirouz et al., 2017; Saura et al., 2015). Our interpretation is that the disconformity formed in the forebulge area due to a protracted period of time (say 7 Myr) with neither significant uplift nor erosion. It was a surface of non-deposition (a transport surface) equivalent to the surface now found south of Mesopotamia and the Persian Gulf (Fig. 1). The erosion could have been modulated by climate and sea-level variations, and re-moved some of the underlying back-bulge and possibly early fore-land basin fill.

4.2. Closure of the Neo-Tethys and implications for climate and geodynamics

The timing of the closure of the Neo-Tethys has been debated since the advent of plate tectonics. Dewey et al. (1973) proposed

a collisional age of about 5 Ma. Such a young age was also advo-cated by Bird et al. (1975). However, in the 1980s other authors claimed older ages between 20 Ma and 45 Ma (e.g., Dercourt et al., 1986), and in the last decade between 20 Ma and 35 Ma (e.g., Agard et al., 2011; Ballato et al., 2011; Fakhari et al., 2008;Mouthereau et al., 2012; Saura et al., 2015). Our study provides new and tighter time constraints. In the one hand, direct measure-ments of the oldest foreland deposits preserved above the fore-bulge unconformity near the suture in the western Zagros imply a collision older than 23.5 Ma. In addition, the chronostratigraphy of the Asmari formation in the Dezful embayment show the onset of the forebulge formation and subsequently orogenic wedge loading prior to 25 Ma. On the other hand, the Asmari Formation is trun-cated by the forebulge unconformity at 29 Ma toward the north in the Shalamzar section. Therefore, these observations suggest a time range of 27 ± 2 Ma for the collision. Our results indicate possible variations of collisional age along the strike of the orogen that we have not analyzed in detail here.

This chronology weakens the argument of Allen and Arm-strong (2008) that the Arabia–Eurasia collision around 35 Ma could have driven Mid-Cenozoic global cooling which was near the Eocene–Oligocene boundary, ca. 33.7 Ma (Liu et al., 2009). The older unconformity which lies below the Asmari was used

M. Pirouz et al. / Earth and Planetary Science Letters 477 (2017) 168–182 177

Fig. 7. Spatiotemporal distribution of the diachronous unconformity, foreland (orange) and pre-foreland (green) deposits across the Zagros basin. Observed data of the sections are shown with bars, and modern location of the forebulge axis with dots. The propagation rate of unconformity is 24 ± 1.5 mm/yr in the eastern sector and 22.6 ± 1.5 mm/yr in the western sector. Transition between pre-foreland and foreland deposits is almost continuous in the Asmari anticline and is shown with the gray box. Each site is located with respect to distance from the suture (shown with gray numbers) after restoration for post depositional deformation. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

to support an older age of the collision more consistent with the scenario of Allen and Armstrong (2008). As we mentioned ear-lier, this unconformity corresponds to a hiatus from 60 Ma to 35 Ma, and the overlying Asmari Formation is younging northward(i.e. Ehrenberg et al., 2007; Van Buchem et al., 2010). In view of this timing, and the low sedimentation rate during the Asmari car-bonate deposition, we favor the hypothesis that the unconformity at the base of the Asmari formation developed as a result of the ophiolite obduction in the Late Cretaceous.

If the continent–continent collision occurred at 27 Ma, then at 35 Ma the Neo-Tethyan oceanic basin must have been about 160 km wide given the 20 mm/yr convergence rate between 35 and 27 Ma (McQuarrie and van Hinsbergen, 2013). The seaway connecting the Indian Ocean to the Atlantic Ocean was probably much wider as it must have included the epi-continental sea over the thinned Arabian plate (which could have been typically 150 km wide). So it is likely that the discussed seaway was not disrupted significantly by 35 Ma. It may even have sustained for some time after the continent–continent collision started.

On the geodynamic side, our proposed timing supports the hy-pothesis that the Zagros collision contributed to the slow-down in Arabia–Eurasia plate convergence between 28 Ma and 20 Ma (McQuarrie and van Hinsbergen, 2013). This scenario conflicts with the notion that the Zagros wedge developed over the last 5 Ma and slowed down the convergence between Eurasia and Arabia (Austermann and Iaffaldano, 2013). The fact that the opening of the Gulf of Aden between Arabia and NE Africa does not show any significant slowdown over that period (Fournier et al., 2011)suggests that the inferred recent slowdown could be an artifact. Finally our study supports the view that the Paleogene volcanic flare-up, which ended around 40 Ma (Verdel et al., 2011), shut down much earlier than the onset of the collision. This find-ing is in line with the interpretation of Verdel et al. (2011) that this volcanic activity was associated with an episode of adjust-ment of the slab dynamics unrelated to the collision. Horton and Fuentes (2016) documented a correlation between coupling along the South America subduction zone and magmatism in the An-des, where no continent–continent is involved. It thus possible that the Paleogene magmatic period in Iran, which is associated with widespread crustal extension (Verdel et al., 2011), would corre-spond to a period of low coupling, possibly associated with slab roll back and back arc extension along the Eurasian active mar-gin. In any case, the wane of the volcanism is probably not a good proxy for the age of the collision in the case of the Zagros orogen.

4.3. Kinematics of the foreland basin evolution and its initial formation

Here we combine our results with the published constraints on the geometry and chronology of the Asmari limestone and foreland deposits to quantify the evolution of the foreland (e.g., Ehrenberg et al., 2007; Pirouz et al., 2016; Van Buchem et al., 2010). Given the current distance of about 600 km between the present forebulge and the paleo-forebulge at the Dehmoord sec-tion, we can determine by how much the forebulge has mi-grated southwestward within Arabia. The present distance, which is the sum of the widths of the wedge (W = 280 km) and fore-land (F = 320 km), needs to be corrected for the amount of crustal shortening across the Zagros, i.e., �W in Fig. 2. To es-timate this amount of shortening, we used the published bal-anced cross-sections based on subsurface data (Jahani et al., 2009;Sherkati et al., 2006); this implies ca. 49 km of shortening in the wedge along the eastern cross-section (Figs. 1 and 6), and about 1 km additional shortening in the Persian Gulf foredeep (Fig. 6). To take into account lateral variations of F, W and the uncertainty on �W, we estimate the migration to 650 ± 50 km. We therefore conclude that the forebulge has migrated southwestward at a rate of 24 ± 2 mm/yr on average since 27 Ma.

The estimated ages of the erosion surface between the Asmari Formation and foreland deposits at different locations are plotted in Fig. 7, as follows: at Shalamzar between 29 Ma and ca. 26 Ma, at Dehmoord between 27.6 Ma and 21.5 Ma, at Kaftar between 25.5 Ma and 18 Ma, and at Asaluyeh between 17 Ma and 9 Ma (Pirouz et al., 2016). The hiatus is shorter in the western Zagros, especially in the Dezful embayment and the adjacent wedge to the North, where it is mostly marked by the Gachsaran evapor-ites. In the Haftkel (well 12) and Asmari anticlines, the top of the Asmari Formation has an age of 19 Ma (Ehrenberg et al., 2007;Van Buchem et al., 2010), and 60 km westward in the Masjed–Solayman anticline the top of the Gachsaran Formation/base of the Mishan Formation is about 16.75 Ma old (Pirouz et al., 2016)(Figs. 1 and 7). These data show that the available chronological constrains on intermediate positions of the forebulge between the suture and its modern location are consistent with a migration rate of the forebulge of 24 ± 1.5 mm/yr in the eastern Zagros. We ob-tain similar results in the western sector indicating a migration rate of 22.6 ± 1.5 mm/yr since 27 ± 2 Ma (Fig. 7).

In principle, the incorporation of a section of foreland deposits into the wedge should be accompanied with a change of sedi-mentation rate, of sedimentary facies and growth strata. Such a

178 M. Pirouz et al. / Earth and Planetary Science Letters 477 (2017) 168–182

Fig. 8. Depth (a) and rate (b) of sediment accumulation as a function of time for different sections in the Zagros (see Fig. 1 for locations). Observed sedimentation rates are compared with predictions of simple models of foreland migration using the present geometry of the Zagros foreland with southwestward migration at rates of 21 and 26 mm/yr. Time “0” is set at the onset of sediment accumulation at each section.

rationale was used for example by Ruh et al. (2014) to describe the kinematics of the Zagros wedge in the Fars. Sediment accumu-lated along the Dehmoord section at a fast rate (185–247 m/Myr), consistent with foreland subsidence, until about 16 Ma. This sug-gests that the site reached the tip of the fold-thrust belt and was incorporated into the Zagros wedge at a later date. The tip of the Zagros continental wedge now lies at a structurally restored dis-tance of 330 km south of this site, at the Asaluyeh anticline (Figs. 1and 7). We infer that the tip of the wedge has migrated southwest-ward at a rate of at least 24.4 mm/yr, similar to the migration rates of the forebulge and foreland basin (Figs. 7 and 8). The similarity of these three rates is remarkable and supports the interpretation that the flexure of the foreland migrated southwestward in rela-tion to the migration of the load applied by the Zagros wedge. The Z shape of the Zagros deformation front suggests significant vari-ations (in space and/or time) either in initiation of the shortening or widening rate of the wedge. We ignore these variations for sim-plicity.

We now assess how the history of sediment accumulation at various sections in the Zagros compares with the history expected from a simple model of foreland migration. The model assumes that subsidence pattern has been migrating southward together with the forebulge, as Arabia was underthrusting the Zagros wedge (Fig. 2), and that the geometry of the basement of the foreland has been steady-state. For this calculation, we use the modern ge-

ometry that was found to best fit subsurface constraints on the depth to the basement and gravity anomalies (Pirouz et al., 2017). The model corresponds to the lithosphere with a uniform width of the wedge and equivalent elastic thickness of 50 km. Sediments fill up the accommodation space formed by migrating flexure of the Arabian lithosphere. This assumption seems reasonable given that the sedimentary facies indicate subaerial deposition on an alluvial plain similar to Mesopotamia with rare shallow marine episodes. To compare with the observations from different sec-tions, we calculate the time evolution of the sedimentation rate since the passage of each section over the forebulge. In the steady-state assumption, all sections should have experienced the same accumulation history with a time offset determined by the mi-gration of the forebulge. We used the sections described in this study as well as sections described in previous studies at Cha-har Makan (Khadivi et al., 2010), Kaftar, and Asaluyeh (Pirouz et al., 2016) (Fig. 8). We tested different migration rates and find a qualitative match between the model prediction and the observed accumulation sedimentation rates for migration rate of 26 mm/yr in agreement with the migration rate derived from the diachronic-ity of the erosion surface itself, as described above. For example, the agreement is quite remarkable given that the assumption of a steady-state foreland and forebulge geometry is questionable. The Dehmoord section which spans depositional ages of 21.5 Ma to 16 Ma, and the Asaluyeh section spanning 9.1 Ma to 2.6 Ma, both

M. Pirouz et al. / Earth and Planetary Science Letters 477 (2017) 168–182 179

show a good agreement to the model with 26 mm/yr. The section at Kaftar, which spans depositional ages between 17.5 Ma and 8.5 Ma, has somewhat lower sedimentation rates, which fit best a mi-gration rate of 21 mm/yr. However, the migration rate should have been lower between 16 Ma and 9 Ma. This interpretation agrees with the decreasing propagation rate of the sedimentary facies and existence of a transgressive system tract sometime between 9 Ma and 14 Ma in the Fars region based on strontium isotope stratigra-phy (Pirouz et al., 2015).

The model shown in Fig. 8 implies that the deposition of the first foreland sediment would lag behind the formation of the erosion surface at the forebulge by about 7 Myr. The first order agreement between the observations and 2-D model predictions is quite remarkable, as some mismatch between sedimentation rates should arise from various factors. It is improbable that the system evolves in steady-state, especially in the early days of the forma-tion of the foreland. Sediments might not fill up to the crest of the forebulge. The initially very low sedimentation rate predicted by the model might not be resolvable from the available chrono-logical constraints (Figs. 3 and 8). Eustatic sea-level variations must also have influenced sedimentation rates. Given the 21.5 Ma for the lowermost foreland sediments at Dehmoord, the time lag would imply a passage over the forebulge no later than 28.5 Ma. In the western sector, this time lag of the passage is much shorter (ca. 1 Myr) to nonexistent due to continuous record between back-bulge and foredeep, where the forebulge is occupied by the Gach-saran evaporites for a period of 3 Myr. That interpretation agrees with the geometry of isochrones and ages of the broad synform in the upper 150 m of the Asmari Formation north of Well 12 (Fig. 9). We interpret this pattern to reflect sediment accumulation in the nascent back-bulge basin. Elastic flexure would then have started after 25 Ma. This interpretation is in favor of the scenario that the erosion surface associated with the forebulge first developed after 25 Ma in the western sector.

4.4. Kinematics of the Zagros orogen

The conceptual stratigraphic model in Figs. 2 and 3 shows how the facies and isochrones of foreland basin deposits relate to the kinematics of underthrusting and loading by the growing oro-genic wedge in the case of the Zagros Neogene foreland. Given the large orogenic advance and the removal of the earliest prox-imal basin, the preserved remnants of the ancient foreland basin represent distal records and could postdate the onset of forma-tion of the foreland basin as has been argued for a similar case in the Andes (DeCelles and Horton, 2003). In the previous section, we argued that the onset of flexural bending of the Arabian plate was recorded in the Asmari formation at about 25 Ma, suggest-ing that flexing of the Arabian plate by the Zagros orogenic wedge started at about 27 Ma. Since then the forebulge has migrated by ca. 650 km southwestward to its present location southwest of the Mesopotamian–Persian Gulf. Assuming that the width of the Zagros and the foreland has remained constant during the migra-tion, the same length of the Arabian lithosphere would have been thrust beneath Eurasia. If the wedge widened, as is likely, the mi-gration would have been augmented by the migration of the center of mass of the wedge associated with its widening. It is probable that the foreland width was initially much less than the present value due to the smaller elastic thickness of the stretched Arabian passive margin. The foreland basin width probably increased and contributed to the southward migration of the foreland deposits over Arabia. In the following we take these two factors into ac-count to relate the migration of the foreland to the convergence across the Zagros.

The elastic thickness was probably significantly less than the value of 50 km suggested by the current geometry of the foreland

and the gravity anomalies (Pirouz et al., 2017). For comparison the elastic thickness of the South China passive margin over which the foreland basin of the Taiwan orogen has formed is estimated at Te = 13 km based on gravity data and seismic constraints on the geometry of the 120 km wide foreland basin (Lin and Watts, 2002). Assuming that the passive margin of Arabia had a similar elastic thickness, the width of the initial foreland might have been ca. 120 km, compared to ca. 320 km at present. Out of the 650 km south-westward migration of the forebulge, possibly as much as 200 km could then reflect the widening of the flexural basin. For taking the widening of the Zagros wedge into account, we hypothesize that there was no significant orogenic wedge by 27 Ma. Assuming a triangular wedge, the center of mass of the wedge would have mi-grated by about 100 km relative to the suture, as the wedge grew to its present width of 330 km. Given a 100 km migration due to the widening of the wedge, we get that the convergence between Arabia and the Zagros suture, i.e. the Eurasian margin, is estimated to be 350 km. This implies that an equivalent length of Arabian lithosphere was thrust north of the suture. This is roughly consis-tent with the value of 270 km of underthrust Arabian lithosphere estimated from seismological observations (Paul et al., 2010). It also implies an average shortening rate of ca. 13.5 mm/yr south of the Zagros suture. This rate compares well with the modern shortening rate of 11 ± 2 mm/yr across the Zagros which is about half of the 25 ± 2 mm/yr convergence between Arabia and Eurasia (Masson et al., 2005).

Now it is instructive to compare our estimated 350 km of convergence between Arabia and the suture with crustal short-ening calculated from the volume of the thickened crust of the Zagros and palinspastic restorations since the onset of the colli-sion. Verges et al. (2011) calculated shortening of Arabia along a 2-D crustal section of the Zagros wedge in Lorestan assuming con-servation of crustal volume. For their calculation, they used the top of the Sarvak Formation (Mid-Cretaceous carbonate) as the up-per boundary, the seismic Moho depth as the lower boundary, and the suture as the northern boundary. The initial crustal geometry of the Arabian margin prior to shortening was reconstructed using paleo-bathymetric constraints assuming local isostasy and the cur-rent Moho depth along the southern part of the Lorestan section (see their Fig. 8). They balanced the crustal area after removing sediments postdating the Sarvak Formation and found ca. 180 km of shortening for the sedimentary cover and ca. 149 km for the basement since the Late Cretaceous.

Pirouz et al. (2017) presented a similar analysis at 3-D orogen-scale including the Zagros foreland system and the Sanandaj–Sirjan metamorphic belt. They calculated the volume confined between modern topography and Moho obtained from a flexural model consistent with observed gravity data. Assuming an initial 38 km crustal thickness for both Arabia and Eurasia, the crustal volume stored in the thickened Zagros together with the crust recycled into the basin account for averaged along-strike shortening of 126 ± 18 km across the orogen; i.e., ca. 60 km in the Zagros and ca. 65 km in the Sanandaj–Sirjan. Another similar 2-D analysis repre-sents ca. 120 km of crustal shortening at orogen scale (Allen et al., 2013).

The Arabian rifted passive margin involved in the initial Zagros wedge was probably <38 km thick. Additional shortening across the Zagros must therefore have been accommodated by tectonic inversion in the rifted passive margin (Fig. 10). To estimate that quantity, we assume that the passive margin of the Neo-Tethys was similar to the passive margin of the modern Atlantic Ocean, which is thinned by ca. 50% and stretched ca. 150 km on each side (Peron-Pinvidic et al., 2013). Such a thinning is consistent with the paleo-depth of the passive margin sediment estimated by Verges et al. (2011) based on sedimentary facies. In this case, the in-version of the stretched Arabian and Eurasian margins could have

180 M. Pirouz et al. / Earth and Planetary Science Letters 477 (2017) 168–182

Fig. 9. Cross-section of the Asmari basin deposits (including Asmari limestone, Ahwaz sandstone, and Pabdeh–Gurpi marls) with time lines across the Dezful embayment. For location, see Fig. 1. The color shading shows depositional ages based on age constraints from strontium isotope ratios (Ehrenberg et al., 2007; Van Buchem et al., 2010and this study). Isochrones are plotted every 1 Myr and labeled every 5 Myr. Distribution of the Ahwaz sandstone Member is shown with lithologic pattern between wells 1 and 9. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

Fig. 10. a) Schematic cross-section showing the present crustal geometry. We propose that out of ca. 350 km of shortening that has been accommodated since the onset of the collision ca. 135 km contributed to crustal thickening of the Arabian margin, and ca. 65 km was taken by thickening of the Eurasian margin. The remaining 150 km of shortening, that did not contribute to crustal thickening, is proposed to have been underthrust and eclogitized. b) Schematic representation of our calculation, assuming that the crust was 38 km thick before opening of the Neo-Tethys Ocean. Restoring the present crustal volume to a crust of uniform thickness of 38 km implies ca. 126 ± 18 km (rounded to 125 km for simplicity) of shortening (Pirouz et al., 2017). We assume that pre-collisional stretching of the Arabian rifted margin due to opening of the Neotethys Ocean had resulted in about 75 km of extension. The total shortening deduced from crustal volume conservation is thus 200 km.

M. Pirouz et al. / Earth and Planetary Science Letters 477 (2017) 168–182 181

absorbed ca. 150 km of shortening (75 km on each side). The in-version in Eurasia is likely to have occurred during subduction of the Neo-Tethyan lithosphere beneath Eurasia, prior to the conti-nental collision (Fig. 10). Therefore, the shortening of the Arabian crust includes two components: (1) 60 km of shortening in nor-mal 38 km thick crust, and (2) ca. 75 km of inversion. That much shortening in the Zagros plus 65 km of shortening in the Sanandaj–Sirjan adds to 200 km, which is 150 km short from the 350 km shortening implied by the forebulge migration over the last 27 Myr(Fig. 10).

We might speculate on possible interpretations of this discrep-ancy. Our reasoning assumes plane strain deformation with no strain perpendicular to the N–S convergence across the Zagros. Al-ternative explanation is to invoke significant out-of-plane flow as has been proposed for the Andes (e.g., Gerbault et al., 2005). This seems improbable given that the estimated 150 km crustal volume budget averages out the Zagros over >1000 km between Makran and Anatolia. The outflow scenario is ruled out, because balancing the large crustal budget would require squeezing out a very large volume beneath or onto the Anatolian plateau for which there is no topographic or Moho depth evidence. Therefore, we favor the alternative possibility that the crustal mass was not conserved due to eclogitization of the 150 km underthrust plate. The same mech-anism has been advocated in the Himalayas (e.g., Hetenyi et al., 2007).

Note that 350 km of shortening over the last 27 Ma represents only about half of the Arabia–Eurasia convergence since the col-lision and might be an underestimate. The shortening absorbed in other areas (i.e., Alborz, Caspian Sea, Kopet Dag) indeed add to no more than 100 km (Allen et al., 2004), and could not have played a big role in accommodating the post-collisional conver-gence. Therefore a larger amount of eclogitized crust than is ex-pected would have been subducted into the mantle. Finally, we note that a continent–continent collision older than 27 Ma would imply a larger inconsistency between the amount of the shortening inferred from cumulative convergence between Arabia and Eurasia and crustal volume conservation since the collision.

5. Conclusions

The closure of the Neo-Tethys Ocean due to the Arabia–Eurasia collision in the Zagros resulted in the development of the Zagros orogenic wedge and associated foreland basin starting ca. 27 Myr ago. The chronology of sediment accumulation on top of the fore-bulge unconformity, which caps the carbonate of the Asmari For-mation, indicates that the forebulge of the foreland basin migrated southwestward since then and at a rate of 24 ± 1.5 mm/yr in the eastern Zagros and 22.6 ± 1.5 mm/yr in the western sector within the Arabia. Shortening across the Zagros is estimated to be 13.5 mm/yr and accounts for about half of the migration rate; the re-maining fraction is being due to the widening of the wedge and foreland basin. The closure of the Neo-Tethys seaway postdates the Mid-Cenozoic shift toward a cooler climate, although it may well have contributed to a global cooling trend due to the var-ious mechanisms described by Allen and Armstrong (2008). The Arabia–Eurasia collision could have contributed to a slowdown of the Arabia–Eurasia convergence at the onset of the collision. Short-ening in the Zagros and Sanandaj–Sirjan amounts to 350 km since the continent–continent collision started. This value exceeds the cumulative shortening estimated from palinspastic restorations and crustal volume implying that a significant fraction of the Arabian crust (i.e. more than 40%, equivalent to 150 km of shortening) was thrust beneath Eurasia, eclogitized and did not contribute to the crustal thickening.

Acknowledgements

This Research has been partially funded by Swiss National Sci-ence Foundation Grant No. P2GEP2-148801, and Grant PRF No.53814-ND8 of the American Chemical Society. We are grateful to the people of Dehmoord and Shalamzar for assistance during the field work. Seyed Abdolnabi Abtahi and Yaser Gharakhani are thanked for hospitality. Tahmoures Yousefi, Ebrahim Khademi, and Ramin Elyaszadeh from Geological Survey of Iran–Shiraz, and Reza Monsef, Rasoul Esmaeili helped with logistics and sampling. Isaac Hilburn, Sarah Slotznick and Ross Mitchell are thanked for tech-nical assistance with paleomagnetic measurements. We benefited from constructive comments from Brian Horton, Jonas B. Ruh and an anonymous reviewer. Associate Editor An Yin is thanked for handling the manuscript. Mark Allen and Frédéric Mouthereau pro-vided valuable comments and suggestions.

Appendix A. Supplementary material

Supplementary material related to this article can be found on-line at http://dx.doi.org/10.1016/j.epsl.2017.07.046.

References

Agard, P., Omrani, J., Jolivet, L., Whitechurch, H., Vrielynck, B., Spakman, W., Monié, P., Meyer, B., Wortel, R., 2011. Zagros orogeny: a subduction-dominated process. Geol. Mag. 148, 692–725.

Alavi, M., 2004. Regional stratigraphy of the Zagros fold–thrust belt of Iran and its proforeland evolution. Am. J. Sci. 304, 1–20.

Allen, M., Jackson, J., Walker, R., 2004. Late Cenozoic reorganization of the Arabia-Eurasia collision and the comparison of short-term and long-term deformation rates. Tectonics 23.

Allen, M., Saville, C., Blanc, E., Talebian, M., Nissen, E., 2013. Orogenic plateau growth: expansion of the Turkish–Iranian Plateau across the Zagros fold-and-thrust belt. Tectonics 32, 171–190.

Allen, M.B., Armstrong, H.A., 2008. Arabia–Eurasia collision and the forcing of mid-Cenozoic global cooling. Palaeogeogr. Palaeoclimatol. Palaeoecol. 265, 52–58.

Austermann, J., Iaffaldano, G., 2013. The role of the Zagros orogeny in slowing down Arabia–Eurasia convergence since ∼5 Ma. Tectonics 32, 351–363.

Ballato, P., Uba, C.E., Landgraf, A., Strecker, M.R., Sudo, M., Stockli, D.F., Friedrich, A., Tabatabaei, S.H., 2011. Arabia–Eurasia continental collision: insights from late Tertiary foreland-basin evolution in the Alborz Mountains, northern Iran. Geol. Soc. Am. Bull. 123, 106–131.

Bird, P., Toksöz, M.N., Sleep, N.H., 1975. Thermal and mechanical models of continent–continent convergence zones. J. Geophys. Res. 80, 4405–4416.

Bosworth, W., Huchon, P., McClay, K., 2005. The red sea and gulf of Aden basins. J. Afr. Earth Sci. 43, 334–378.

DeCelles, P.G., Horton, B.K., 2003. Early to middle Tertiary foreland basin develop-ment and the history of Andean crustal shortening in Bolivia. Geol. Soc. Am. Bull. 115, 58–77.

Dercourt, J., Zonenshain, L., Ricou, L.-E., Kazmin, V., Le Pichon, X., Knipper, A., Grandjacquet, C., Sbortshikov, I., Geyssant, J., Lepvrier, C., et al., 1986. Geolog-ical evolution of the Tethys belt from the Atlantic to the Pamirs since the Lias. Tectonophysics 123, 241–315.

Dewey, J.F., Pitman, W.C., Ryan, W.B., Bonnin, J., 1973. Plate tectonics and the evolu-tion of the Alpine system. Geol. Soc. Am. Bull. 84, 3137–3180.

Ehrenberg, S.N., McArthur, J.M., Thirlwall, M.F., 2010. Strontium isotope dating of spiculitic Permian strata from Spitsbergen outcrops and Barents Sea well-cores. J. Pet. Geol. 33, 247–254.

Ehrenberg, S.N., Pickard, N.A.H., Laursen, G.V., Monibi, S., Mossadegh, Z.K., Svånå, T.A., Aqrawi, A.A.M., McArthur, J.M., Thirlwall, M.F., 2007. Strontium isotope stratigraphy of the Asmari Formation (Oligocene–Lower Miocene), Sw Iran. J. Pet. Geol. 30, 107–128.

Elderfield, H., 1986. Strontium isotope stratigraphy. Palaeogeogr. Palaeoclimatol. Palaeoecol. 57, 71–90.

Emami, H., 2008. Foreland Propagation Folding and Structure of the Mountain Front Flexture in the Posht-e-Kuh Are (NW Zagros, Iran). Universitat de Barcelona, Barcelona, p. 181.

Faccenna, C., Becker, T.W., Jolivet, L., Keskin, M., 2013. Mantle convection in the Middle East: reconciling Afar upwelling, Arabia indentation and Aegean trench rollback. Earth Planet. Sci. Lett. 375, 254–269.

Fakhari, M.D., Axen, G.J., Horton, B.K., Hassanzadeh, J., Amini, A., 2008. Revised age of proximal deposits in the Zagros foreland basin and implications for Cenozoic evolution of the High Zagros. Tectonophysics 451, 170–185.

182 M. Pirouz et al. / Earth and Planetary Science Letters 477 (2017) 168–182

Fournier, M., Chamot-Rooke, N., Rodriguez, M., Huchon, P., Petit, C., Beslier, M.O., Zaragosi, S., 2011. Owen fracture zone: the Arabia–India plate boundary un-veiled. Earth Planet. Sci. Lett. 302, 247–252.

Gerbault, M., Martinod, J., Hérail, G., 2005. Possible orogeny-parallel lower crustal flow and thickening in the Central Andes. Tectonophysics 399, 59–72.

Gradstein, F.M., Ogg, J.G., Schmitz, M., Ogg, G., 2012. The Geologic Time Scale 2012. Elsevier. 2-volume set.

Hetenyi, G., Cattin, R., Brunet, F., Bollinger, L., Vergne, J., Nabelek, J., Diament, M., 2007. Density distribution of the India plate beneath the Tibetan plateau: geo-physical and petrological constraints on the kinetics of lower-crustal eclogitiza-tion. Earth Planet. Sci. Lett. 264, 226–244.

Homke, S., Vergés, J., Garcés, M., Emami, H., Karpuz, R., 2004. Magnetostratigraphy of Miocene–Pliocene Zagros foreland deposits in the front of the Push-e Kush Arc (Lurestan Province, Iran). Earth Planet. Sci. Lett. 225, 397–410.

Homke, S., Vergés, J., Serra-Kiel, J., Bernaola, G., Sharp, I., Garcés, M., Montero-Verdú, I., Karpuz, R., Goodarzi, M.H., 2009. Late Cretaceous–Paleocene formation of the proto-Zagros foreland basin, Lurestan Province, SW Iran. Geol. Soc. Am. Bull. 121, 963–978.

Homke, S., Vergés, J., Van Der Beek, P., Fernàndez, M., Saura, E., Barbero, L., Badics, B., Labrin, E., 2010. Insights in the exhumation history of the NW Zagros from bedrock and detrital apatite fission-track analysis: evidence for a long-lived orogeny. Basin Res. 22, 659–680.

Horton, B.K., Fuentes, F., 2016. Sedimentary record of plate coupling and decoupling during growth of the Andes. Geology 44, 647–650.

Jahani, S., Callot, J.-P., Letouzey, J., Frizon de Lamotte, D., 2009. The eastern ter-mination of the Zagros Fold-and-Thrust Belt, Iran: structures, evolution, and relationships between salt plugs, folding, and faulting. Tectonics 28, 1–22.

Khadivi, S., Mouthereau, F., Larrasoaña, J.C., Vergés, J., Lacombe, O., Khademi, E., Beamud, E., Melinte-Dobrinescu, M., Suc, J.P., 2010. Magnetochronology of syn-orogenic Miocene foreland sediments in the Fars arc of the Zagros Folded Belt (SW Iran). Basin Res. 22, 918–932.

Kirschvink, J., 1980. The least-squares line and plane and the analysis of palaeomag-netic data. Geophys. J. Int. 62, 699–718.

Kirschvink, J.L., Kopp, R.E., Raub, T.D., Baumgartner, C.T., Holt, J.W., 2008. Rapid, pre-cise, and high-sensitivity acquisition of paleomagnetic and rock-magnetic data: development of a low-noise automatic sample changing system for supercon-ducting rock magnetometers. Geochem. Geophys. Geosyst. 9.

Lin, A.T., Watts, A.B., 2002. Origin of the West Taiwan basin by orogenic loading and flexure of a rifted continental margin. J. Geophys. Res. 107.

Liu, Z., Pagani, M., Zinniker, D., DeConto, R., Huber, M., Brinkhuis, H., Shah, S.R., Leckie, R.M., Pearson, A., 2009. Global cooling during the Eocene–Oligocene cli-mate transition. Science 323, 1187–1190.

Masson, F., Chéry, J., Hatzfeld, D., Martinod, J., Vernant, P., Tavakoli, F., Ghafory-Ashtiani, M., 2005. Seismic versus aseismic deformation in Iran inferred from earthquakes and geodetic data. Geophys. J. Int. 160, 217–226.

McArthur, J.M., Howarth, R.J., Bailey, T.R., 2001. Strontium isotope stratigraphy: LOWESS Version 3: best fit to the marine Sr-isotope curve for 0–509 Ma and accompanying look-up table for deriving numerical age. J. Geol. 109, 155–170.

McFadden, P., McElhinny, M., 1988. The combined analysis of remagnetization circles and direct observations in palaeomagnetism. Earth Planet. Sci. Lett. 87, 161–172.

McFadden, P., McElhinny, M., 1990. Classification of the reversal test in palaeomag-netism. Geophys. J. Int. 103, 725–729.

McQuarrie, N., van Hinsbergen, D.J.J., 2013. Retrodeforming the Arabia–Eurasia col-lision zone: age of collision versus magnitude of continental subduction. Geol-ogy 41, 315–318.

Mouthereau, F., Lacombe, O., Verges, J., 2012. Building the Zagros collisional orogen: timing, strain distribution and the dynamics of Arabia/Eurasia plate conver-gence. Tectonophysics 532, 27–60.

Nocquet, J.-M., Willis, P., Garcia, S., 2006. Plate kinematics of Nubia–Somalia using a combined DORIS and GPS solution. J. Geod. 80, 591–607.

Omrani, J., Agard, P., Whitechurch, H., Benoit, M., Prouteau, G., Jolivet, L., 2008. Arc-magmatism and subduction history beneath the Zagros Mountains, Iran: a new report of adakites and geodynamic consequences. Lithos 106, 380–398.

Paul, A., Hatzfeld, D., Kaviani, A., Tatar, M., Péquegnat, C., 2010. Seismic imaging of the lithospheric structure of the Zagros mountain belt (Iran). Geol. Soc. (Lond.) Spec. Publ. 330, 5–18.

Peron-Pinvidic, G., Manatschal, G., Osmundsen, P.T., 2013. Structural comparison of archetypal Atlantic rifted margins: a review of observations and concepts. Mar. Pet. Geol. 43, 21–47.

Pirouz, M., Avouac, J.-P., Gualandi, A., Hassanzadeh, J., Sternai, P., 2017. Flexu-ral bending of the Zagros foreland basin. Geophys. J. Int 210 (3), 1659–1680. http://dx.doi.org/10.1093/gji/ggx252.

Pirouz, M., Simpson, G., Bahroudi, A., Azhdari, A., 2011. Neogene sediments and modern depositional environments of the Zagros foreland basin system. Geol. Mag. 148, 838–853.

Pirouz, M., Simpson, G., Castelttort, S., Gorin, G., Bahroudi, A., 2016. Controls on the sequence stratigraphic architecture of the Neogene Zagros foreland basin. Spec. Pap., Geol. Soc. Am. 525, SPE525-512.

Pirouz, M., Simpson, G., Chiaradia, M., 2015. Constraint on foreland basin migra-tion in the Zagros mountain belt using Sr isotope stratigraphy. Basin Res. 27, 714–728.

Ruh, J.B., Hirt, A.M., Burg, J.P., Mohammadi, A., 2014. Forward propagation of the Za-gros Simply Folded Belt constrained from magnetostratigraphy of growth strata. Tectonics 33, 1534–1551.

Sabzehei, M., Ravan, J.R., Amini, B., Eshraghi, S.A., Mahabadi, S.A., Seraj, M., 1993. Geological map of the Neyriz quadrangle H-11, scale 1:250,000. Geological Sur-vey of Iran.

Saura, E., Garcia-Castellanos, D., Casciello, E., Parravano, V., Urruela, A., Verges, J., 2015. Modeling the flexural evolution of the Amiran and Mesopotamian fore-land basins of NW Zagros (Iran–Iraq). Tectonics 34, 377–395.

Sherkati, S., Letouzey, J., Lamotte, D.F.D., 2006. Central Zagros fold–thrust belt (Iran): new insights from seismic data, field observation, and sandbox modeling. Tec-tonics 25.

Van Buchem, F.S.P., Allan, T.L., Laursen, G.V., Lotfpour, M., Moallemi, A., Monibi, S., Motiei, H., Pickard, N.A.H., Tahmasbi, A.R., Vedrenne, V., Vincent, B., 2010. Regional stratigraphic architecture and reservoir types of the Oligo-Miocene de-posits in the Dezful Embayment (Asmari and Pabdeh Formations) SW Iran. Geol. Soc. (Lond.) Spec. Publ. 329, 219–263.

Verdel, C., Wernicke, B.P., Hassanzadeh, J., Guest, B., 2011. A Paleogene extensional arc flare-up in Iran. Tectonics 30.