Ophiolite emplacement in NE Oman: constraints from magnetotelluric sounding

Transcription

1 Chapter 4 Ophiolite emplacement in NE Oman: constraints from magnetotelluric sounding 4.1 Foreword This following chapter differs slightly in format compared to the other chapters. The contents represent the current status of a revised manuscript which has originally been submitted to Geophysical Journal International. Co-authors on this paper are Graham Heinson, David Gray (Melbourne University) and Robert Gregory (Southern Methodist University, Texas). The coauthors have initiated and organised the trip, and was conducted by all four authors at the same time. While most work on the manuscript has been done by myself, parts of section 4.2, 4.3 and 4.8 which relate to the geology of the area, had input from David Gray and Robert Gregory. Graham Heinson s input has been that of a PhD supervisor role and is not different to the input of other chapters presented in this thesis. 31

2 4.2. Introduction 32 Ophiolite emplacement in NE Oman: constraints from magnetotelluric sounding S. Thiel 1, G. Heinson 1, D.R. Gray 2 and R.T. Gregory 3 1 CERG, School of Earth & Environmental Sciences, University of Adelaide, SA 55, Australia, 2 School of Earth Sciences, The University of Melbourne, Melbourne, Victoria 31, Australia 3 Geological Sciences, Southern Methodist University, Dallas Texas 75275, USA Abstract Magnetotelluric (MT) data have been acquired across the Samail Ophiolite, Oman mountains, along a 115 km transect in January 25. Twenty-five MT stations were deployed approximately every 4km along major valleys crossing the Saih Hatat window, a Tertiary domal culmination that folds a major NE-facing recumbent fold nappe, and the boundary between the Dasir and Ibra ophiolite blocks. The survey aimed to investigate the tectonic evolution of the emplacement of the Samail Ophiolite by delineating major faults and geological boundaries on a crustal scale. Induction arrows indicate a different response between the ophiolite to the south and the Saih Hatat window in the north-east of the profile. Arrows are larger and point towards the ocean across the Saih Hatat, whereas arrows on top of the ophiolite are generally small and rotate by 18 across the period bandwidth. Due to distinct topography the impedance data appear very distorted locally and are statically shifted. However, robust phase tensor analysis shows that the regional field is two-dimensional (2-D) for short periods of up to 3s. Rotated MT impedances were inverted using a two-dimensional (2-D) code. Major resistivity interfaces coincide with the upper plate-lower plate (UP-LP) shear zones near the surface, and dip towards the Arabian margin suggesting a south-west orientated underthrusting prior to the ophiolite emplacement. The para-autochthonous and less deformed upper plate is a resistor, whereas the strongly deformed lower plate is more conductive. 4.2 Introduction While the common type of subduction of oceanic plates has been scrutinised with the MT method (Unsworth et al., 25; Spratt et al., 25; Soyer and Unsworth, 26; Ritter et al., 23; Pous et al., 24; Hamilton et al., 26; Weckmann et al., 27), obduction environments, where oceanic lithosphere has been emplaced onto continental lithosphere, have not been studied in detail before. This study presents the results and data analysis of a MT survey across the Samail Ophiolite of the Oman Mountains, Arabian Peninsula, one of the best exposed ophiolites on Earth. It is a pilot study, aimed at defining the crustal structure of ophiolite environments to help distinguish between different emplacement scenarios for these large slabs of oceanic lithosphere.

3 4.2. Introduction E 58 E 6 E 26 N IRAN 26 N 1 UNITED ARAB EMIRATES N 1 1 Hawasina window 1 Muscat 24 N 22 N MT station Maastrichtian & Tertiary sediments Samail ophiolite Hawasina and Haybi allochthonous units Permian to Cretaceous autochthonous shelf carbonates Lower plate units Pre- Permian basement 56 E 1 Jebel Akhdar Window OMAN 58 E Saih Hatat Window a As Sifah Figure 4.1: MT stations on top of regional geology map. The Oman Ophiolite Mountains and major shear zones run approximately parallel to the coast line. Adapted from Gray et al. (25). 1 6 E 22 N

4 4.2. Introduction a N Samail Detachment 22 MT station Dasir km Ibra Jabal Muscat Saih Hatat 1 1b 4 5 4b 6 7 Abyad b 23 3'N 23 24'N Huw l/meeh subwindow Ocean Iraq UAE Saudi Arabia Oman Jemen Somalia 58 24'E 58 3'E 58 36'E 9 5 Hatat Shist sz Iran Ocean 1 1b 4 4b 5 6 Huw l/meeh UP-LP sz UP-LP sz Tertiary Samail Ophiolite Ophiolite sequence Hawasina melange Permian-Cretaceous carbonates Ordovician Amdeh Formation Hatat Schist and Hijam Formation (upper plate) Lower plate Figure 4.2: a) MT stations on top of an interpretation of the geology. Twenty-five stations, out of which 2 were utilised, were deployed in January 25 along a two-dimensional survey line stretching approximately 115 km. The survey line followed main valleys crossing the Saih Hatat window and the Jabal Abyad mountains onto the boundary between the Dasir and Ibra ophiolite blocks. Geology interpretation is redrawn from Gray et al. (2). b) Station locations on a digital elevation image with focus on the north-western part of the profile. Semitransparent gray-shaded areas show location of lower and upper plate material (Huw l/meeh and Saih Hatat, respectively). The locations of the major shear zones (sz) are also shown. The Oman Mountains have been the focus of a number of recent studies on obduction mechanisms (Chemenda et al., 1996; Gray and Gregory, 23; Breton et al., 24) and are cited as the classic example of Tethyan ophiolite obduction (Moores et al., 2; Wakabayashi and Dilek, 23; Robertson, 24). Despite almost 1% exposure and a lack of surface vegetation, interpretation of the Oman surface geology allows more than one permissible tectonic scenario of ophiolite emplacement (see discussion in Searle et al. (23); Gray and Gregory (23); Gray et al. (24)). The Saih Hatat window, south-east of Muscat, contains a major refolded fold-nappe that underlies the ophiolite sequences to the south (Figure 4.1). It holds vital information about the obduction of the Samail Ophiolite sequences (Gray et al., 25). Models describe either subduction towards (Gregory et al., 1998; Gray et al., 2, 24) or away from the margin (Chemenda et al., 1996; Searle et al., 1994, 23). This MT survey, involving twenty-five stations from Muscat to interior Oman across the Saih Hatat window and Dasir-Ibra ophiolite block (Figure 4.2), seeks to resolve this uncertainty by providing constraints on the resistivity distribution of the subsurface, and therefore crustal structure beneath the Oman Mountains. The topography of the area, time constraints

5 4.3. Survey details and geology 35 and therefore logistical circumstances led to the deployment of the stations along a line running northeast-southwest. Two-dimensional (2-D) modelling is able to reproduce first-order regional features but cannot account for three-dimensional (3D) complexity in the dataset. Careful analysis of the dataset using robust processing (Chave and Thomson, 1989) and phasesensitive dimensionality and strike analysis (Caldwell et al., 24) were followed by an inversion (Rodi and Mackie, 21) of a subset of the data that was considered 2-D. This paper presents resistivity imaging that suggests an inland-dipping zone of high conductivity correlating with outcrop of the lower plate in the Huw l/meeh subwindow. 4.3 Survey details and geology Geology The Oman ophiolite is believed to have formed at a Tethyan spreading center between 94 and 97 Ma (Tilton et al., 1981). Subsequent detachment was caused by intraoceanic thrusting along the metamorphic sole (94-9 Ma) with final emplacement of the Samail Ophiolite nappe onto the Arabian margin dated at around Ma (Warburton et al., 199; Hacker et al., 1996). The final positioning of the Samail Ophiolite is characterised by gravity-driven emplacement along pseudo-extensional faults due to a rising upper plate between 76 and 7 Ma (Gregory et al., 1998). The Oman Mountains stretch over 7 km along the coast of Oman and yield elevations up to 3 m (Figure 4.1). The Oman ophiolite (Searle and Cox, 1999; Nicolas et al., 2) was formed when former Tethys oceanic lithosphere was emplaced on top of Arabian continental lithosphere. The Cretaceous Samail Ophiolite slab is a major allochthonous unit and delineates the structurally highest sheet represented by the Dasir-Ibra blocks along the MT profile (Figure 4.2). Gravity measurements indicate a presently thin (< 5 km thick) ophiolite sheet bounded by a flat-lying basal fault (Shelton, 199; Gray et al., 2). Near Muscat, the Muttrah peridotite is assumed to be part of the ophiolite sheet outboard of the Dasir-Ibra blocks (Figure 4.2). The Oman Mountains show a thrust stack of former Tethys basinal sediments (Hawasina melange) overlying para-autochthonous continental-shelf carbonates and pre-permian basement units (Hatat schist). These features crop out in the Saih Hatat window, a Tertiary domal culmination that folds a major NE-facing recumbent fold nappe of the upper plate that is truncated by the UP-LP shear zone (Miller et al., 22; Gray et al., 25) (Figure 4.2b). The core of the major fold-nappe is made up of intensely foliated Hatat Schist. The structurally lowest parts of the mountains are represented by the outcrops of high-pressure blueshists and eclogites of the lower plate in the Huw l/meeh and As Sifah subwindows. To the north and east the pre-permian to Cretaceous units are onlapped by Maastrichtian and Tertiary shallow-water carbonate sediments, which are clearly post-ophiolitic.

6 4.4. Transfer function estimation 36 The tectonic evolution of the Samail Ophiolite is under vigorous debate. Sm-Nd and Zircon U-Pb ages of eclogites and the position of regional isoclinal folds in the Saih Hatat window and the Huw l/meeh subwindow hold clues to the understanding of the ophiolite emplacement (Miller et al., 22) and are therefore covered by the northern part of the MT line. To the south of the Saih Hatat, the MT line extends across the Dasir Ibra blocks consisting of the Samail Ophiolite (Figure 4.2). Due to difficult accessibilty in the rugged and mountaneous landscape, the profile follows major valley floors, which are perpendicular to the coastline and major geological boundaries, such as the UP-LP shear zone and the Hatat Shist shear zone (Figure 4.2b) Measurement details Five MT instruments, which have been built by Flinders and Adelaide Universities, were used in the experiment. The instruments recorded time-series of electric field in the two horizontal directions and of magnetic field in the horizontal and vertical directions. The instruments were left out for durations of 2 h to 45 h, sampling at 2 Hz. The electric and magnetic field were recorded with Cu CuSO 4 electrodes and fluxgate magnetometers, respectively. In the field, the instruments were oriented to geomagnetic coordinates. Subsequent rotation to geographic coordinates is negligible, since the magnetic declination is only.7. Site spacing is 2 4 km in the northern part of the profile between stations OMN1 to OMN6 in order to ensure a good spatial resolution across the UP-LP shear zone and the Huw l/meeh subwindow. Spacing was increased to about 5 km across the Saih Hatat (stations OMN7 to OMN11) and reaches 1 km from OMN18 to OMN22. The increased spacing allows the coverage of the entire ophiolite block to its southern margin with the MT profile. At least two stations recorded simultaneously in order to use remote-reference schemes that reduce noise across the entire period bandwidth (Gamble et al., 1979). 4.4 Transfer function estimation Data quality In order to assess the data quality, it is useful to initially investigate the time-series data. The quality or the signal-to-noise ratio is influenced by signal strength and length of deployment. Time constraints only allowed instrument deployments up to two days at a time, which reduces the resolution to periods greater than 1 s. During the experiment, the measured magnetic fields have occasionally shown spikes in the time-series, which is most likely due to cultural noise. During pre-processing, a median filter with adaptive length and threshold value significantly reduced the amplitude of the noise. Even though the median filter can only produce

7 4.4. Transfer function estimation 37 site 1 site 1b site 4 site 4b App. resistivity [Ohm.m] Phase [ ] site 5 site 6 site 7 site App. resistivity [Ohm.m] Phase [ ] site 9 site 1 site 11 site App. resistivity [Ohm.m] Phase [ ] Period [s] Figure 4.3: Apparent resistivity ρ a and phase φ plots for all stations utilised in the inverse modelling. Shaded triangles and squares denote ρ a and φ of the Z x y and Z y x component of the impedance tensor, respectively. Open circles and inverse triangles represent the Z x x and Z y y components of the impedance tensor. The coordinate frame {x, y} of the impedance tensor is rotated 34 counter-clockwise from geographic north. For a hypothetical 2-D environment Z x y (shaded triangles) denotes the TE-mode and Z y x (shaded squares) denotes the TM-mode.

8 4.4. Transfer function estimation 38 site 17 site 18 site 19 site 21 site App. resistivity [Ohm.m] Phase [ ] Period [s] Figure 4.3: (cont d) Apparent resistivities and phase plots for the remaining stations. For a detailed description see previous page. a replacement value that is only near the actual value of the varying magnetic field alone, the MT impedance responses across the entire period bandwidth have been improved and their respective errors reduced. Stations that produced responses with a high level of noise have been excluded from further analysis MT responses For most sites, robust processing (Chave et al., 1987; Chave and Thomson, 1989) generated good estimates of the MT impedance Z and magnetic transfer function (T zx and T zy ) responses. E x E y H z = Z xx Z yx T zx Z xy Z yy T zy ( Hx H y ), (4.1) with E the electric field, H the magnetic induction and {x, y, z} denote geographic north and east and the vertical direction, respectively. Z is complex and can be separated into a real (X) and imaginary (Y ) part: Z = X + iy, (4.2) where i = 1 is the imaginary number. Remote referencing (Gamble et al., 1979) has been utilised and resulting responses were compared with the results of single site robust processing. In some cases, longer time-series data and single site processing generated better-quality responses than shorter time-series and remote referencing due to voids in the data of the remote site. Stations that produced very noisy responses of the impedance tensor have been excluded from further analysis.

9 Transfer function estimation 'E 58 'E 58 3'E 59 'E 24 'N 57 3'E 58 'E 58 3'E 59 'E 24 'N 'N 23 3'N 'N 'N 22 3'N 24 'N.5 85 s s 22 3'N 24 'N 'N 23 3'N 'N 'N s 22 3'N 57 3'E 58 'E 58 3'E 59 'E s 57 3'E 58 'E 58 3'E 59 'E Figure 4.4: Real (red) and imaginary (blue) Parkinson induction Arrows for 84, 341, 682, 1365 s on top of the geology of the region and bathymetry of the ocean. For a description of the morphological units see Figure 'N

10 4.4. Transfer function estimation 4 Figure 4.3 shows the apparent resistivities ρ a (T) and phases φ(t) of all stations utilised in the inversion calculated from the four elements of the rotated impedance tensor. In anticipation of section 4.5, the rotation angle is 34 counter-clockwise from north, which results in the rotated coordinate system {x y }. Hence, we assign the x y -component of the impedance tensor to the E-polarisation (TE mode for transverse electric) and the y x -component to the B-polarisation (TM mode for transverse magnetic). The x -coordinate is therefore approximately parallel to the lateral conductivity contrast. The data show that the TM-mode responses have small error bars in comparison to the TE-mode responses. The TM-mode phases stay within the -9 quadrant. As expected, the TM-mode has higher apparent resistivities than the ones calculated from the diagonal components Z x x and Z y y of the impedance tensor. However, this is not true for the TE mode with responses being noisier and having a lower apparent resistivity connected with very high phases. At some stations the TE-mode phases exceed 9 for periods above 1 s (Figure 4.3). A similar behaviour, however at much shorter periods of 1 1 s, has been reported and attributed to current gathering and deflection in the shallow crust due to strong 3D effects (Weckmann et al., 23). Phases larger than 9 can also occur in environments, where an anisotropic block overlies another anisotropic layer of different strike (Heise and Pous, 23). In our example, phases reach values larger than 9 at much longer periods, indicating that 3D effects occur at much larger depths. Both the TE- and TM-mode indicate an area of high conductivity beneath stations 4b to 6. At stations 1 to 4, the TM mode indicates an increase in apparent resistivity for longer periods. The subsurface beneath stations is generally more conductive than for the northern part of the profile. Induction arrows are a graphical representation of the Geomagnetic Depth Sounding (GDS) transfer functions T zx and T zy and the real and imaginary parts have been plotted in Figure 4.4 for four representative periods. The real induction arrows in the Parkinson convention will point towards zones of higher conductivity and away from resistive blocks. For short (85 s) and intermediate periods (341 s), the real arrows are very large in the northern part of the profile (stations 1-11) and point mostly north. Their respective imaginary counterpart is nearly antiparallel, which usually occurs in a 2-D subsurface. In the southern part of the profile (stations 15-22), the real arrows are comparatively smaller and point west to south-west for short periods. As expected, for intermediate periods, the arrows begin to rotate toward the conductive ocean. However, the imaginary arrows are perpendicular to the real induction arrows indicating a complex three-dimensional (3D) subsurface. For longer periods (682 s and 1365 s), eventually all induction arrows rotate north-west-north toward the abyssal plains of the Arabian Sea, with increasing magnitude closer to the coast. The imaginary arrows are also perpendicular to their real counterparts, a sign for 3D structures

11 4.5. Dimensionality and strike considerations 41 at greater skin-depths. 4.5 Dimensionality and strike considerations Before attempting a 2-D inversion, data need to be analysed in terms of its dimensionality. In the case of a 2-D setting a consistent strike direction has to be found to which the geographic coordinate frame of measurement can be rotated. In order to avoid galvanic distortion of shallow 3D structures affecting the dimensionality analysis, we used the phase tensor approach of Caldwell et al. (24). This method of analysing the phase information of the impedance tensor does not require presumptions about the underlying dimensionality as do decomposition methods (Bahr, 1988, 1991; Groom and Bailey, 1989). This method is applicable where both the regional conductivity distribution and the conductivity heterogeneity close to the surface are three-dimensional. The MT phase is defined as the ratio of the real and quadrature parts of the impedance tensor Z. This expression allows a generalisation to the entire tensor, specifically its real and imaginary parts X and Y, respectively. Φ = X 1 Y. (4.3) Here, Φ defines the real phase tensor and X 1 is the inverse of the real part of the impedance tensor Z. In a 2-D environment, a coordinate system {x, y } can be found, where the offdiagonal components Φ xy and Φ yx in Equation (4.3) vanish. With increasing complexity (e.g. 3D) of the subsurface those components will increase independent of the chosen coordinate frame. The skew angle β is a measure of the tensor s asymmetry and hence of threedimensionality: β = 1 ( ) Φxy Φ yx 2 tan 1. (4.4) Φ xx + Φ yy We believe that a threshold of β < 5 is a good approximation to the assumption of twodimensionality of the subsurface. Figure 4.5 is a contour plot of the skew angle β as a function of period across the profile. For stations 8-22 β is small for a period band between 2 s-3 s suggesting a 2-D resistivity distribution beneath those sites. Between stations 4-7 a localised 3D inductive structure causes the skew angle to decrease below 5. Stations 4-7 cross or are adjacent to the upper plate material of the Huw l/meeh subwindow (see Figure 4.2b). In general, the skew increases with period and reaches values higher than 5 for periods longer than 5 s. Based on the skew analysis, station 1 has been excluded from modelling altogether because of its extremely large skew angles of 15 and more throughout the entire period bandwidth (Figure 4.5). The other two invariants of the phase tensor are the maximum and minmum phase, Φ max and Φ min, respectively. The common graphical representation of the phase tensor is in the

12 4.5. Dimensionality and strike considerations b 4 4b b e t a Figure 4.5: Contour plot of the phase tensor skew β. Data points utilised in the triangulation have been marked with black dots. Skew values β < 5 indicate periods, where we assume two-dimensionality of the data set. Generally, station responses for periods longer than 3 s are 3D. form of an ellipse (Figure 4.6), with the major and minor axis depicting the Φmax and Φmin (Caldwell et al., 24). The direction of the major axis of the ellipse indicates the direction of major horizontal current flow, e.g. in a 2-D scenario, phase tensors will be parallel (on the conductive side) or perpendicular (on the resistive side) to electric strike. Figure 4.6 shows the phase tensor ellipses for four different periods of stations 1-1. For short periods (42 and 85 s), the ellipses are aligned with the strike of the valleys. Furthermore, stations in the valleys demonstrate a large discrepancy between the maximum and minimum phase suggesting that current flow is significantly larger along strike of the valleys. Sites situated in the open Saih Hatat plain (e.g. site 8) show different behaviour at short periods with almost circular phase tensor ellipses (see also Figure 4.3). At intermediate to long periods (341 and 682 s), the phase tensors rotate with the major axis aligned east-west. The rotation together with large skew angles for periods longer than 5 s (Figure 4.5) suggest 3D resistivity distribution at greater depths. We have determined a period-averaged strike direction for all stations between periods 2 to 3 s (Figure 4.7), following the insight from the skew angle analysis and the rotation of the ellipses that the subsurface is 3D for periods longer than 3 s. We used the direction of the phase tensor major axis to determine the strike. In this period range, the orientation of the major axes are mostly period-independent and around N56 E (Figure 4.7). For longer periods the orientation of the phase tensor ellipses rotate to around N15 E. However these orientations are by far not as consistent as the orientations for the period range 2 to 3 s. The induction