Paleotemperature investigation of the Variscan southern external domain: the case of the Montagne Noire (France)

The Montagne Noire located in the southern part of the French Massif Central represents the 43 northern part of the South-Variscan Foreland. It is subdivided into three parts. The granite- 44 migmatite Axial Zone dome is surrounded by non- or weakly metamorphosed Paleozoic 45 sedimentary series. Both northern and southern flanks of the Montagne Noire dome are 46 deformed by km-scale, south


66
The knowledge of quantitative constraints, such as pressure, temperature, duration of 67 heating, strain, strain rate, exhumation rate, and uplift rate, is essential to understand the 68 formation and evolution of a mountain belt. In recent years, many studies (e.g Beyssac et al., are often not achieved (Frey, 1987;Merriman and Frey, 1999;Merriman and Peacor, 1999). 77 7 syn-tectonic gravity flow deposits such as greywacke turbidites and olistostromes, which 160 recycle pre-Visean sedimentary rocks of the advancing nappe pile (Engel et al., 1978(Engel et al., , 1980(Engel et al., -161 1981Feist and Galtier, 1985;Poty et al., 2002;Vachard and Aretz, 2004;Cózar et al., 2017;162 Vachard et al., 2017). During the Carboniferous, the progressive deepening of the 163 sedimentary basin argues for a foreland trough, filled by syntectonic deposits, and coeval 164 with the formation of recumbent folds. The entire Paleozoic series has been deformed into 165 several km-scale, southeast-verging, thrust sheets and recumbent folds (Arthaud 1970;166 Echtler and Malavieille, 1990). 167 synforms deform the inverted series of the entire nappe stack. The poorly exposed contact 184 between the lower recumbent fold and the foreland basin led to controversial interpretations. 185

Tectonic subdivisions
According to Arthaud (1970), and Alabouvette et al. (1982) the inverted limb is in 8 contact with Visean-Serpukhovian turbidites. However, according to (Engel et al., 1978(Engel et al., , 187 1980(Engel et al., -1981, the flysch sequence is part of the lower recumbent (Mt. Peyroux) fold. The 188 normal and inverted limbs of that fold nappe are connected by a D1 fold hinge. Whatever the 189 structural interpretation, this does not change the RSCM results presented below, and their 190 interpretations. The presence of Paleozoic olistoliths in the Visean-Serpukhovian turbidites 191 basin in which the recumbent folds are emplaced documents the syn-sedimentary character 192 of this event. It provides an important time constraint for the tectonic evolution of the 193 southern flank: the nappe thrusting is contemporary or older than 318Ma (Engel et al., 1978(Engel et al., , 194 1980(Engel et al., -1981Alabouvette et al., 1982Vachard et al., 2017). Carboniferous sedimentary rocks exposed in the normal stratigraphic order. This domain is 200 observed between the lower recumbent fold and the Axial Zone metamorphic rocks. Close to 201 the Axial Zone, for instance near Saint-Pons (Fig. 2), the Para-autochthonous unit consists of 202 Devonian marbles and micaschists (Engel et al., 1980(Engel et al., -1981Feist and Galtier, 1985;Poty et 203 al., 2002;Vachard and Aretz, 2004;Cózar et al., 2017;Vachard et al., 2017). 204 The autochthonous turbiditic basin represents the foreland basin into which the Mt-205 Peyroux and Mts-de-Faugères recumbent folds were emplaced. The basin substratum is 206 unknown but might probably be similar to that observed in the northern para-autochthonous 207 unit. To the west, the basin underlies the stack of recumbent folds, and to the south, it is 208 hidden below the Cenozoic formations. 209

The Axial Zone 210
The Axial Zone is composed of orthogneiss, paragneiss, amphibolite, micaschist, and rare 211 marble that experienced a partial melting giving rise to migmatites and granites. The Axial 212 Zone presents a dome architecture of about 90 km long, and 20 km wide, with a N70E long     The Raman measurements were carried out on a Renishaw inVia reflex system belonging to 291 the ISTO-BRGM analytic platform. The Wire 3.4 software was used for the data acquisition.

12
The argon-ion laser source excitation of 514.5nm was set at a power of about 5% of its 293 capacity (2.5mW). The monochromatic ray was coupled to a reflection microscope with an 294 x100 objective. Before each series of measurement, the spectrometer was calibrated using 295 an internal silica standard for the wavenumber (520,4 cm -1 ) and the signal intensity (at least 296 30 000 counts per second). Fifteen carbonaceous matter spectra were systematically 297 acquired for each sample within the thin section in order to obtain a representative and 298 reliable temperature. Only organic particles located below transparent quartz, a few µm 299 under the thin section surface, were analysed, to avoid any mechanical damaging of their 300 crystalline structure due to thin section preparation and polishing. The acquisition duration 301 was set to at least 60s and adapted depending on the quality of the spectrum. in the wavenumber range between 1100-1800 cm -1 and 2500-3100 cm -1 . The first order 307 region of the carbonaceous matter spectrum for 1100-1800 cm -1 provided the data of this 308

study. 309
Depending on its crystallinity (hence on the temperature experienced), carbonaceous 310 matter presents a variable number of Raman peaks, which can be used to discriminate  The RSCM method reliability depends on the richness of the carbonaceous matter 370 enclosed in the sample. In the field, the dark lithologies, i.e. the Early Ordovician and Visean-371 Serpukhovian turbidites, turned out to be the most appropriate lithologies (Fig. 6a, b, d, f). On 372 the contrary, the Early Cambrian Marcory green sandstone, and the Early Cambrian and 373 Devonian carbonates did not yield sufficient organic matter (Fig. 6c, e), and thus, were 374 avoided for sampling. 375

376
The main goal of the present study was to establish the paleotemperature field associated 377 with the regional tectonic-metamorphic events. Several hypotheses may be formulated., 378 The temperature development was coeval with the recumbent folding and thrusting 379 event for this purpose, samples were collected in each structural unit except for the South 380 siltstone formation of the Upper (Pardailhan) recumbent fold, only the S1 slaty cleavage is 385 observed (Fig. 4a).

401
Despite the low-grade metamorphism of the area, both high and low temperature methods 402 were used during this study. The results are well distributed in the 230 to 400°C range 403 covered by both methods (Fig. 5 and 7). Seventy-three temperature values were calculated 404 in the Montagne Noire southern flank. Forty-eight RSCM spectra were processed using 405  (Fig. 8, 9 cross sections 1, 2, 3, 4, 5). Globally, the temperature decreases from the 419 southern edge of the Axial Zone dome towards the southeast in the Paleozoic sedimentary 420 rocks (Fig. 10). The maximum temperature is located in the micaschists series that form the 421 dome envelope with a value close to 400°C. The lowest temperature, about 220°C, was 422 obtained in the Visean turbidites. In the western part of the southern flank, the gradient is 423 less apparent that in the eastern part. It decreases from 400°C to 315°C towards the 424

Cenozoic sedimentary rocks. 425
The RSCM method was also used within the dome micaschist envelope (Fréville et  426 al., 2016). The measured temperatures, ranging between 450°C and 600°C are higher than 427 those obtained in the sedimentary rocks of the southern flank (Fig. 11). The temperature 428 gradient is steeper in the dome northern edge than in the micaschist series of the eastern 429 area. Indeed, the deformation is more intense in the dome northern edge, as also shown by 2016). This is probably a consequence of a brittle shearing that poste-dates the dome 432 exhumation (Thompson and Bard, 1982). The data set obtained via the RSCM method 433 documents a continuous evolution from the dome core southward with a temperature 434 decrease from 580°C to less than 300°C (Fig. 11)  The RSCM temperature map of the Montagne Noire southern flank (Fig. 10) is clearly 448 incompatible with these assumptions. Since the isotherms cut across the different fold units 449 irrespective of their contacts, the bulk thermal structure of the Montagne Noire southern flank 450 cannot be linked to the pre-doming thickening stage. Moreover, the RSCM temperature is not 451 correlated with the structural position of the analyzed samples since the highest nappes of 452 the tectonic pile record temperatures similar to those measured in the para-autochthonous 453 units (Fig. 10). Thus, the link of the temperature field with nappe emplacement, and tectonic 454 burial is unlikely. The measured temperature is also not in agreement with shear heating 455 along major contacts, as that observed in Japan (Mori et al., 2015). 456 In contrast, there is a clear temperature gradient decreasing with the distance to the 457 dome. Whatever the thrusting direction of the nappe, southeastward (Arthaud, 1970)   Two temperature anomalies are visible within the N-S temperature gradient (Fig. 8). To the 471 east of the study area, within the Visean-Serpukhovian turbiditic basin, several samples 472 present relatively high temperatures around 360°C, similar to those found near the dome 473 micaschists (Figs. 6, 9 cross section 5). However, in this area, the Axial Zone dome is not 474 exposed. The other temperature anomaly is located in the southern part of the Pardailhan 475 upper recumbent fold (Fig. 9 cross section 3). The gradient is decreasing southward from ca 476 380°C to 270°C, but at the very end of the exposed Paleozoic series, the temperature 477 increases to reach ca 320°C. Since this area is located about 15 km away from the Axial 478 Zone dome south margin, it is unlikely that these high temperatures result from the direct 479 dome thermal influence. 480 Three hypotheses might explain these singularities. 481 First, the effect of a late fault might be taken in account as it could reorganize the 482 distribution of paleotemperatures. It is especially noticeable for the anomaly within the upper 483 recumbent fold (Pardailhan unit). The N-S striking sinistral fault at the east of the points H25 484 an H26 Fig. 2) separates high temperature in the west >315°C (H27-H28-H29) from lower 485 temperatures in the east <290°C (H36-H39-H40; Fig. 8; Fig. 9). This effect is also shown by 486 the bending of the 300°C isotherm (Fig. 10). Finally, at the regional scale, the district is riddled by Late Carboniferous granitic 492 intrusions, such as the Sidobre or Folat plutons (Fig. 2) can explain the isotherm pattern in the easternmost part of the southern flank, more than 502 15km away from the dome (Fig. 10). The illite crystallinity has been widely studied in the eastern part of the Montagne Noire 506 southern flank to unravel low-grade metamorphic gradients. In our study area, two previous 507 works (Engel et al., 1980(Engel et al., -1981Doublier et al., 2015) documented a north to south 508 decreasing thermal gradient (Fig. 12). The latter study used the Árkai index (Árkai, 1991;509 Guggenheim et al., 2002)  Weber index (Weber 1972) in Engel et al. (1980Engel et al. ( -1981 to the diagenetic domains defined in 512 Frey and Robinson, (1999). In agreement with our own results, these works show the same 513 N-S variations as those documented by the RSCM method. 514 In the same area, the conodont color alteration has been used to estimate the 515 paleotemperature repartition (Wiederer et al., 2002). The method is based on the color of 516 apatite crystals that form the conodonts (Epstein et al., 1977). During a metamorphic event, 517 the crystallinity of organic matter trapped within the apatite grains increases, allowing a 518 possible correlation with the metamorphic grade experienced by the fossil. As shown in Fig.  519 13, the dome thermal effect is also visible by this method. The intensity of the metamorphism 520 is globally higher to the northwest and decreases away to the southeast (Fig. 13). These 1-to 10-cm sized quartz lenses were inferred to be related to fluid circulation coeval 526 with the emplacement of the upper recumbent fold (Guiraud et al., 1981). Analyses show a 527 temperature around 275±25°C, which is in agreement with our results at ca 300°C, but 528 higher than those provided by illite crystallinity at 200°C (Doublier et al., 2015). This 300°C 529 temperature has been related to the thickening event (Guiraud et al., 1981), however, fluid 530 circulation might have occurred also during the doming. The tectonic contact between the 531 upper and lower recumbent folds has been reworked during doming, as indicated by E-W 532 striking slickenlines on flat lying surfaces, and N70E striking fibers infilling tension gashes 533 (Arthaud, 1970;Sauniac, 1980;Harris et al., 1983). Therefore, the deformation observed 534 along the contact suggests that the ca 300°C temperature might be related to the Axial Zone 535 dome thermal effect or even to a late event as documented by Aerden (1998)  depending on the method considered. 549 As example of these differences, the temperature range estimated using the 550 conodont approach is the widest, with temperatures comprised between 75°C and 475°C 551 against 85°C to 300°C for IC and 230°C to 400°C for RSCM. Furthermore, for a given 552 location, the results are sometimes different between the two illite crystallinity studies by 553 Engel et al. (1980Engel et al. ( -1981 and Doublier et al. (2015). This is remarkable at the contact 554 between the lower recumbent fold and the Visean basin where the data from Engel et al. 555 (1980Engel et al. 555 ( -1981  are also large differences in the temperature gradients derived from the different 571 geothermometers. Between 1 to 6 km away from the dome, the illite crystallinity seems to 572 reach a plateau around 275°C (dashed line in Fig. 14). This plateau could correspond to the 573 temperature above which small muscovite or sericite have completely replaced the original 574 clays and would be the upper limit of the method, while carbonaceous matter crystallinity 575 (hence its Raman spectra) continue its evolution at higher temperature. Hence, for 576 comparing the RSCM temperature with illite crystallinity temperature, we shall exclude the 577 highest temperature domain in the vicinity of the dome and consider the section at a distance 578 from 5 km to 15 km from the dome. In this distance range, the gap between RSCM 579 temperatures and illite crystallinity temperatures increases with the distance to the dome 580 (Fig. 14). This divergence is apparently irrespective of the RSCM calibration used (Fig. 15a,  581 b) or the calibration to convert KI in temperature (T): 582 -(i) With Merriman and Frey (1999)  Therefore, these disparities suggest that organic matter record the dome exhumation thermal 588 impact on a much broader area than illite crystallinity. 589 To interpret these discrepancies, one has to consider the nature of the physical and Hence, the disparities between illite crystallinity and RSCM in Montagne Noire 603 suggest that the dome emplacement, which controls IC and RSCM evolution, was a relatively 604 short event in terms of heat source. Accordingly, it left a more significant imprint on 605 carbonaceous matter than on illite. 606 One can note that the much smaller-scale T anomaly in the basin, recorded by RSCM but 607 not by illite crystallinity, can be similarly explained by hidden magmatic bodies that provided 608 a local heat source for a short period of time. We therefore propose that a regional high heat-609 flow linked to a hot crust is present over a long period 330-295Ma, as recorded by Illite and 610 CAI. However, local events resulting from this regional heat such as plutonic intrusion and on 611 a larger scale the migmatization mostly affected the carbonaceous material. In spite of a common North to South decreasing temperature gradient, the different 634 thermometers present quantitative discrepancies in the estimated temperature. IC 635 temperatures decrease away from the dome to temperatures below 150°C, while RSCM 636 temperatures remain above 250°C. In this case, the fast kinetics of carbonaceous material 637 crystalline evolution, with respect to illite evolution, could then account for the record of the 638 heating event over a larger area by RSCM than by IC. 639 The link between the bulk temperatures and the recumbent folding is not supported 640 by this study. Furthermore, anomalies are observed within the thermal gradient. Since these 641 anomalies are not documented by the IC method, the most likely explanation would be that 642 hidden bodies such as granitic plutons, would have emplaced beneath the Paleozoic 643 sediments after the dome exhumation. In addition, the presence of a hot crust underlying the 644 Paleozoic series cannot be discarded.

Acknowledgments 655
This study has been funded by the Labex VOLTAIRE, and the Observatoire des Sciences de 656 l'Univers en Région Centre (OSUC). Jean-Gabriel Badin, Sylvain Janiec, and Ida di Carlo are 657 thanked for their assistance during the thin section preparation and Raman spectroscopy 658 measurements. We aslo thank Stanislas Sizaret for his help during sampling session.