Applicability of magnetic fabrics in rocks associated with the emplacement of salt structures (the Bicorb-Quesa and Navarrés salt walls, Prebetics, SE Spain)

This work assesses the applicability of the analys is of the anisotropy of magnetic susceptibility (AMS) to rocks associated with salt structures. Magnetic fabrics from 16 sites sampled on Upper Triassic clays (Keuper facie s) interbedded within the salt layers of the Bicorb-Quesa and Navarrés salt walls (Prebet ic Zone, SE Spain) are mostly characterized by oblate magnetic ellipsoids with th eir magnetic foliation parallel to the bedding and magnetic lineation contained in the bed ding plane but without a constant relationship with dip direction. To evaluate the qu ality and significance of the low-field AMS, several experiments were carried out to charac te ize the magnetic mineralogy of representative samples and enable separation of the contributions from paramagnetic and ferromagnetic (s.l.) minerals. The comparison b etween the orientation of the observed magnetic lineation and that of the structu al elements characterizing the internal geometry of the studied salt structures (a ntiforms, minor folds, faults and shear zones) and flanks indicates that the magnetic fabri c is controlled by deformation processes affecting the salt walls during their gro wth and/or evolution. Samples were AC C EP TE D M AN U SC R IP T ACCEPTED MANUSCRIPT 2 also analyzed from sixteen sites from Miocene rocks outcropping in several syn-diapiric Miocene half-grabens related to the salt walls stud ied. Their magnetic fabric is interpreted to reflect the dominant stretching dire ct on at the syn-diapiric half grabens during and/or shortly after deposition. Our results indicate that caution is required in interpreting magnetic lineations related to salt fl ow mechanics from rocks associated with salt structures.


Introduction
The analysis of the geometry of salt structures through direct and/or indirect methods has attracted substantial attention during the last three decades due to their importance as geological reservoirs. Less attention has been paid, however, to the study of the kinematics and/or internal fabric of this kind of structure mainly because of their high internal complexity. This complexity results from the presence of diapiric (salt) rocks, which undergo viscoelastic deformation under geological conditions (Weijermars et al., 1993) (Martínez-González et al., 1997;Lago et al., 1999).
Towards the east, this structure continues with the NNW-SSE Navarrés salt wall (1 x 15 km), also bounded by two syn-diapiric Miocene half-grabens, the Escalona and Playamonte half-grabens, to the east and west, respectively (Fig. 2). Around these salt structures, most Jurassic and Cretaceous overburden rocks lie subhorizontally.
The Bicorb-Quesa and Navarrés salt walls result from a complex Alpine polyphase deformation history including three stages of successive extensional and contractional events (Roca et al., 1996(Roca et al., , 2006Rubinat et al., 2013); an initial Mesozoic The low-field AMS was measured at room temperature (RT-AMS) using a static KLY-2 kappabridge susceptibility meter (AGICO) at the Paleomagnetic Laboratory of Barcelona CCiTUB-ICTJA CSIC. Magnetic susceptibility is a physical property of materials, representing the capacity of the material to be magnetized in a given magnetic field, and is described by a second-rank tensor (K) that relates the applied magnetic field (H) to an induced magnetization (M): M = K x H. The AMS at room temperature in rocks depends mostly on crystallographic preferred orientation, shape of grains, composition and sometimes the distribution-interaction of magnetic minerals (Tarling and Hrouda, 1993). Three axes define the susceptibility ellipsoid: maximum (k max ), intermediate (k int ) and minimum (k min ). The orientations of these axes correspond to the eigenvectors of the susceptibility tensors and characterize the magnetic fabric.
The statistical procedure to obtain the directional and tensor data was based on Jelinek's method (Jelinek, 1977) using Anisoft 4.2 (Chadima and Jelinek, 2009). The magnetic fabrics have been described using parameters defined by Jelinek (1981): (1) the corrected anisotropy degree, P′, that provides a first indication of rock deformation and preferred mineral orientation, and (2) the shape parameter, T, ranging from -1 (prolate ellipsoid) to +1 (oblate ellipsoid).
In order to assess the quality and significance of the low-field AMS, we analyse the contributions from paramagnetic and ferromagnetic (s.l.) minerals to the total AMS.
Two sites representative of Upper Triassic (QB18 with the highest Km value) and Miocene (QB08) rocks were analyzed by means of low-temperature AMS (LT-AMS) and anisotropy of the anhysteretic remanent magnetization (AARM) to compare their paramagnetic and ferromagnetic (s.l.) subfabrics, respectively. In addition, three more sites were analyzed by LT-AMS to better characterize the total AMS (the Upper Triassic QB03 site and the Miocene QB04 and QB19 sites). LT-AMS of spinning A C C E P T E D M A N U S C R I P T

ACCEPTED MANUSCRIPT
8 specimens was measured in a KLY-3 kappabridge susceptibility meter (AGICO) in the Magnetic Fabric Laboratory at the University of Zaragoza. The samples were measured at 77 K (i.e. they were immersed in liquid nitrogen for one hour and for 10 minutes again between changing positions for the same sample) using the same method as in Parés and van der Pluijm (2002). Reducing the temperature of this type of sample generates an increase in the magnetic susceptibility, values at 77 K being approximately 3.8 times higher than at room temperature (298/77), assuming purely paramagnetic phases with a paramagnetic Curie temperature around 0 K (Lüneburg et al., 1999). This increase in the AMS at low temperature is not, however, symmetric along all axes, the k max values increasing by a larger factor than k min (Parés and van der Pluijm, 2002). A LT-AMS ellipsoid with the orientation and magnitude of its k min ≤ k int ≤ k max axes is also obtained from this kind of measurement and can be compared with results obtained at room temperature in order to estimate the paramagnetic contribution to the magnetic anisotropy (e.g. Richter and van der Pluijm, 1994; Parés and van der Pluijm, 2002).
The AARM was measured in four specimens per site. The anhysteretic remanent magnetization (ARM) was induced by a degausser with a DC field and the magnetization was measured in a SRM-755 cryogenic magnetometer (2G Enterprises) in the Paleomagnetic Laboratory at Burgos University. This procedure was performed along nine different axes for every sample. For each position, samples were previously demagnetized by alternating fields (AF) in a peak field of 100 mT and the remaining magnetization was measured. Subsequently, an ARM was induced in the sample by the application of a DC field of 0.05 mT together with a decreasing alternating field with a peak field of 90 mT. For each position, the remaining magnetization measured before the application of the ARM was subtracted. The AARM ellipsoid was calculated using A C C E P T E D M A N U S C R I P T  (1990) using an IM10-30 pulse magnetizer (ASC Scientific) and a TSD-1 thermal demagnetizer (Schonstedt), and (4) variation of magnetic susceptibility with temperature (K-T curves) of lutitic samples using a KLY3 susceptibility meter combined with a CS-L/CS-3 cryostat and furnace apparatus (all from AGICO). Two kinds of K-T curves were obtained: heating/cooling curves and low temperature measurements. The heating/cooling curves were acquired from room temperature to 700ºC and cooling back to room temperature in an argon atmosphere.
The low temperature measurements were taken by cooling the samples to -190ºC and then heating them to room temperature in air. These K-T curves were also used to estimate the percentage of the paramagnetic/ferromagnetic (s.l.) content of the sample

Magnetic properties and ferromagnetic (s.l.) mineralogy
On the basis of magnetic properties at room temperature, samples can be divided into two groups: (1) Group 1 is composed of the red clays Upper Triassic in age, and (2) Group 2 comprises the Miocene rocks that include brownish to grey clays and fine sandstones. The magnetic susceptibilities of the Group 1 samples range from 102 to 358 x 10 -6 SI, being between 150 and 250 x 10 -6 SI in the majority of specimens (Fig. 3), whereas Group 2 shows in general lower magnetic susceptibilities, most lying between 15 and 100 x 10 -6 SI (except for site QB11 with values of 882-1409 x 10 -6 SI) (see Fig.   3 and Table 1). In both groups, the magnetic ellipsoids were predominantly oblate ( Fig.   4a and Table 1). Site-mean T values range between 0.3 and 0.9 in Group 1 samples, most of them indicating fabric ellipsoids that are clearly oblate (T>0.7) (Table 1), whereas in Group 2 samples range between -0.281 and 0.706 (except QB11), showing magnetic ellipsoids from clearly oblate to triaxial (T~0.0) to weakly prolate in shape.
The corrected degree of anisotropy P' for both groups is low (Table 1) have a wasp-waisted shape, confirming this bimodal distribution of grains with contrasting coercivity (Fig. 5hi). The final decay of the magnetic susceptibility of these samples occurs at ∼560-580ºC (Fig. 5kl). The cooling curves of these samples also show a dramatic increase in the susceptibility (Fig. 5kl).
In summary, the ferromagnetic (

Low-temperature AMS and AARM
Samples analyzed by LT-AMS and AARM are all clays. They differ in color and age: red in the Upper Triassic samples and brownish (QB04 and QB19) and pale grey (QB08) in the Miocene samples. Regardless of these differences, the increase in bulk susceptibility at low temperature with respect to its value measured at room temperature is similar in all samples (Fig. 6a), the ratio between the bulk susceptibility at low and room temperature (LT/RT ratio) being around 2.2 in all samples (Fig. 6a). These LT/RT ratios indicate the predominance of the paramagnetic minerals controlling the AMS.
Only samples from site QB03 (red bed) show a higher ratio (LT/RT=3.8) close to the perfect paramagnetic behavior. The shape of the LT ellipsoid is oblate in all cases

ACCEPTED MANUSCRIPT
13 (T>0.5) ( Fig. 6c and Table 2). These results are similar to that obtained at room temperature, with the exception of site QB08 that shows a triaxial magnetic ellipsoid (T=0.102, Table 1) with k min and k int axes girdling in a plane perpendicular to a fairly well clustered k max at room temperature. As in the measurements at room temperature, both the bulk susceptibility and the corrected anisotropy degree (P') at low temperature were higher in the Upper Triassic red bed samples than the Miocene clays ( Fig. 6c,b and Table 2).
The AARM ellipsoids display a different behavior depending on the site analyzed. Specifically, the axes of the AARM ellipsoid coincide with those of the RT-AMS for the Upper Triassic site (QB18), but they do not match for samples from the Miocene site (QB08) (Fig. 7). In fact, this disagreement could explain the better grouping of the k min and k int axes at low temperature than at room temperature for site QB08 (Fig. 7). In the case of the Triassic rocks, the three axes of the RT-AMS, LT-AMS and AARM ellipsoids coincide, suggesting that hematites (the main carrier of the ferromagnetic s.l. subfabric) have a similar subfabric to that shown by the paramagnetic minerals. For Miocene rocks, with a more variable ferromagnetic (s.l.) mineralogy, the axes of the RT-AMS and LT-AMS ellipsoids coincide in site QB19 (Fig. 7), but results from sites QB04 and QB08 point out the influence of a different ferromagnetic (s.l.) subfabric to the RT-AMS in those samples.

Magnetic fabric
Regardless of the rock type, color and age, all sites can be grouped into four types of magnetic ellipsoids according to their directional properties ( Fig. 8 and Table 1). The first type (Fig. 8a), observed in 32% of the analyzed sites, is characterized by an oblate

ACCEPTED MANUSCRIPT
14 magnetic ellipsoid with the minimum susceptibility axes (k min ) perpendicular to the bedding plane and the maximum (k max ) and intermediate (k int ) susceptibility axes scattered within the bedding plane. In this type of magnetic fabric, it is not statistically justifiable to identify a magnetic lineation based on the site-mean orientation of k max axes as the semi-angle (ε 12 ) of the confidence ellipse around the k max axes in the k maxk int plane is large (>42º). It is representative of a sedimentary fabric related to compaction. Type 1 ellipsoids have not been taken into consideration for further structural analysis. The second type (Fig. 8b) consists of clearly oblate to oblate-totriaxial susceptibility ellipsoids which also have k min axes perpendicular to the bedding.
In this case, k max and k int are in the bedding plane and they are slightly more closely grouped than in the first type as indicated by smaller ε 12 values (see Table 1). This indicates that, although weak, a magnetic lineation can be distinguished in these sites.
Most sites sampled in the Upper Triassic rocks and three Miocene sites show this type of magnetic fabric (see Table 1). In the third type (Fig. 8c), again with a defined magnetic lineation in the bedding plane, the k min and k int axes are scattered forming a girdle perpendicular to k max . Thus in this magnetic fabric type, the semi-angle (ε 13 ) of the confidence ellipse around the k min axes in the k max -k min plane is larger (>42º) than in the two previous types. The magnetic ellipsoid is triaxial to weakly prolate in shape.
Only Miocene sites display this type of magnetic fabric. The fourth group (9% of sampled sites) contains all the types of magnetic ellipsoids not covered by the last three types. It includes ellipsoids with k min axes not perpendicular to the bedding plane, meaning that their magnetic foliation does not coincide with bedding, and also magnetic ellipsoids with a very high dispersion in the orientation of their magnetic axes. Like type 1, sites classified into this last group have not been taken into consideration for further structural interpretations. In addition, site QB11 has not been considered due to its high magnetic susceptibility signal (Km=1140 x 10 -6 SI), probably dominated by ferromagnetic (s.l.) minerals (as ferromagnetic s.l. minerals can have 2 to 3 orders of magnitude higher magnetic susceptibilities than paramagnetic minerals).

Orientation of the magnetic ellipsoids
Diapiric structure For the magnetic ellipsoids of the Upper Triassic sites located in areas where the structure is relatively simple and consists of beds folded parallel to the salt-wall axis (sites QB07, QB10, QB15, QB56, QB60, QB62, QB66, and QB69), tectonic correction was performed by simple bedding plane restoration. In areas with superimposition of two differently-oriented trains of folds and local tilting, the tectonic correction was performed following the inverse sequence of tectonic events that were deduced, from field data, to have occurred at each site (see Table 1). In particular, six of these sites were affected by local tilting after the formation and rising of the salt structure and two sites were first folded and then tilted. Thus, the magnetic foliation of all sites is parallel to bedding ( Fig. 9 and Table 1). All sites, except 4 at which the k max and k int axes are scattered within the bedding plane (i.e., magnetic fabric type 1; sites QB12, QB56, QB66 and QB69), have a type 2 magnetic fabric (see the previous section). As In order to investigate this potential relationship, we also analyze in detail the internal structure of the area of two of the three sites where magnetic lineation is not parallel to the flanks of the salt structure (sites QB01 and QB10) (see Fig. 11). Both areas display numerous structural elements indicating they underwent a complex deformation. Site QB01 is located close to a fault system oriented NE-SW and site QB10 close to a fold-and-thrust system oriented WNW-ESE (Fig. 11). In both cases, the magnetic lineation is oriented perpendicular to the closest thrust and/or fold indicating the local tectonic transport direction and that these structures probably control the orientation of the magnetic ellipsoids locally.
In summary, the magnetic foliation observed inside the studied salt walls is controlled by the bedding, being parallel to this plane in most cases. The orientation of the magnetic lineation is controlled by the existing structural elements; it is seen to be parallel to the axes of the antiforms arranged along the salt-wall axes or perpendicular to the trend of fold-and-thrust systems deforming areas locally (sites QB01 and QB10).

Syn-diapiric Miocene half-grabens
The magnetic fabric type of the Miocene sites is very variable with type 1, 2 and 3 magnetic ellipsoids (Table 1). After tectonic correction by simple bedding restoration, the magnetic lineation of most Miocene specimens is found to be subhorizontal and its orientation is spread across a wide range with a maximum oriented NW-SE ( Fig. 9 and Table 1). Most sites with an identifiable magnetic lineation had a down-dip lineation (i.e. parallel to the dip direction) (Fig. 10) perpendicular or highly oblique with respect to the direction of the principal normal faults delimiting both the Bicorb and Quesa halfgrabens (Fig. 9).

Magnetic fabric origin inside the salt structures
All sites sampled on Upper Triassic rocks and, therefore, inside the salt structures of the study area, have type 1 (sedimentary fabric related to compaction) or 2 (oblate to triaxial) magnetic ellipsoids, with exception of sites QB02 and QB60 (both being type 4) ( Fig. 8 and Table 1). Despite these magnetic ellipsoids being very similar to those described for other weakly deformed rocks (e.g. Borradaile and Tarling, 1981;Mattei et al., 1997;Parés et al., 1999;Cifelli et al., 2004;Larrasoaña et al., 2004), the orientation of the magnetic lineation with respect to bedding is highly variable (down-, oblique-

Magnetic fabric origin in related syn-diapiric half-grabens
With respect to the Miocene rocks from the associated syn-diapiric half-grabens, these samples show examples of all four types of magnetic ellipsoids found in the study area ( Fig. 8 and Table 1). Most Miocene sites have a magnetic lineation perpendicular or highly oblique to the bedding strike (i.e. parallel to the dip direction) (Fig. 10). This

Magnetic fabric in rocks associated with salt structures
Most studies related to salt structures have dealt with the characterization of their external geometry and mechanics of formation (e.g. Jackson and Talbot, 1986;Koyi, 1998) and the geometry and kinematics of related salt-structures generated in the sedimentary overburden (e.g. Alsop, 1996;Buchanan et al., 1996). However, much less attention has been paid to their internal fabric (e.g. Kupfer, 1968; Talbot and Jackson,

A C C E P T E D M A N U S C R I P T
The difficulty in interpreting salt fabrics by traditional methods lies in the rapid recrystallization of salt rocks, which removes the original fabric and can make it difficult to correctly determine the strain ellipsoids (Talbot and Jackson, 1987;Miralles et al., 2001). Therefore, the application of AMS analysis to the internal geometry of salt structures by studying salt rocks (e.g. halite) could be very successful and constitutes a powerful tool if they contain disseminated magnetic particles with a known origin.

Conclusion
In this work, we set out to assess the applicability of the AMS analysis in rocks associated with salt tectonics, not extensively investigated previously. With this objective, AMS analysis has been used to study: (1) the interbedded layers of Upper