Geo-thermochronology of the Saint Antonin basin , south-eastern France

This is anOp Abstract – The clastic sedimentary formations of the Saint Antonin basin in the French Maritime Alps contain the record of the Early Oligocene erosional history of the Maures-Esterel massif, Sardinia and Corsica. Detrital apatite fission-track dating and zircon fission-track/U-Pb double dating of samples collected from the Saint Antonin basin confirm sediment provenance and allow obtaining first-order estimates of drainage basin maximum and long-term average exhumation rates. Whereas average exhumation rates were on the order of 0.1–0.2 km/Myr during the Early Oligocene, small parts of the Saint Antonin basin source areas may have experienced maximum exhumation rates on the order of 0.4–0.7 km/Myr. Although zircons and apatites with Early Oligocene fission-track cooling ages make up between 11–15% of the dated grains, a possible volcanic contribution is negligible, as only one single volcanic zircon grain was identified by fission-track/U-Pb double dating. Regional geodynamic processes with convergence in the Western Alps to the east and the end of the Pyreneo-Provençal compression phase by the early Oligocene controlled the differences in basin fill history and sediment provenance between the Saint Antonin basin and the largely contemporaneous Barrême basin in southeastern France.


Introduction
The Haute Provence -Maritime Alps region in southeastern France is characterized by a complex geodynamic evolution.North-south directed deformation during the Eocene-early Oligocene was driven by the Pyreneo-Provençal compression phase, which had started during the Late Cretaceous, followed by late Oligocene-early Miocene opening of the Ligurian Sea (Jolivet et al., 1990(Jolivet et al., , 1991;;Rosenbaum et al., 2002;Lacombe and Jolivet, 2005).Eastwest directed compressional deformation initiating during the early Oligocene was related to Alpine convergence between the Apulian and European plates (e.g., Lacombe and Jolivet, 2005;Dumont et al., 2012 and references therein).The intersection of the two N-S and E-W deformation patterns is well preserved in the Arc de Castellane (Fig. 1a).Eoceneearly Oligocene deposits derived from the Corsica-Sardinia block are preserved in the Saint Antonin basin, which experienced synsedimentary deformation during the final stages of the Pyreneo-Provençal N-S shortening phase (Campredon and Giannerini, 1982).Eocene to late Oligocene/early Miocene sediments of the Alpine foreland basin are preserved in basin remnants of the Barrême and Annot basins, among others.These basins show east-west shortening (Fig. 1b  and c).The Saint Antonin, Barrême and Annot basin remnants have received attention over the past 100 years, which is reflected in a series of publications, culminating in a special publication by Joseph and Lomas (2004 and references therein) on the Grès d'Annot turbidite systems and their equivalence in the Barrême and Saint Antonin basins.
Reconstructions of sediment provenance and potential source areas in these basins are important for understanding source-to-sink relationships and the evolution of the Western Alps foreland basin in general.Provenance studies based on heavy mineral analysis and sediment petrology in the Barrême, Annot and Saint Antonin basins (e.g., Stanley, 1964;Bodelle, 1971;Evans and Mange-Rajetzky, 1991;Evans et al., 2004) provided important information on sediment source areas and routing systems.These studies showed that during the Eoceneearly Oligocene sediments were delivered to these three basins from the Maures-Esterel massif and the Corsica-Sardinia block located to the south.A change of sediment provenance between about 30-29 Ma has been reported for the Barrême basin in terms of sediment petrology, paleocurrent transport directions and zircon fission-track and U-Pb data (e.g., Evans and Mange-Rajetzky, 1991;Evans et al., 2004;Jourdan et al., 2013), but not for the Saint Antonin or the Annot basin, because the Saint Antonin basin never received sediments from the Western Alps, and deposition in the Annot basin had ended by that time.
Detrital apatite fission-track (AFT) and zircon fission-track (ZFT) and U-Pb single grain double dating can provide additional information on sediment provenance, as well as the exhumation history of sediment source areas (e.g., Zeitler et al., 1986;Cerveny et al., 1988;Carter, 1999;Carter and Moss, 1999;Garver et al., 1999;Bernet and Garver, 2005;Bernet et al., 2006;Jourdan et al., 2013).The purpose of this study is combine sediment provenance information with rates of source area exhumation.Therefore, we present geothermochronological data from six samples of detrital apatite and zircon collected from the three main sedimentary units of the latest Eocene to early Oligocene sedimentary sequence preserved in the Saint Antonin basin syncline (Tab.1).Similar data have already been published for the Barrême and Annot basins (Bernet et al., 2009;Jourdan et al., 2013), which allows for a good comparison of the thermochronological signal in the different foreland basin remnants.The times of deposition of the sedimentary sequences in the Saint Antonin basin have previously been constrained by biostratigraphy (Callec, 2001).We use the minimum age approach for comparing the youngest age population of each sample with the biostratigraphically constrained depositional ages of the three formations for estimating maximum source area exhumation rates.Central ages of the AFT and ZFT grain-age distributions of our samples are used for first-order estimates of average long-term exhumation rates of the paleo-drainage areas.Zircon U-Pb ages of grains already dated with the fission-track method were determined in order to identify a possible Paleogene volcanic contribution, which may perturb the exhumation signal, as it was done in a previous Barrême basin study (Jourdan et al., 2013).

The Saint Antonin basin
The Saint Antonin basin is located in the French Maritime Alps, within the Castellane arc, 30 km northwest of Nice (Fig. 1a).The east-west striking basin is about 4.5 km wide and 20 km long and more than 1000 m of detrital sediments accumulated in this basin during the late Eocene to early Oligocene (e.g., Bodelle, 1971;Stanley, 1980 and references therein;Callec, 2001).The Saint Antonin basin is a structurally complex syncline with three minor sub-basins, formed between two anticlines (Fig. 1b and c; Campredon and Giannerini, 1982).
Typical for the Tertiary basin deposits in the Alpine realm, the stratigraphy of the Saint Antonin basin represents the classic trinity of Nummulitic limestone, Blue (or Globigerina) marls, and relatively coarse grained clastic sedimentary deposits (Fig. 2).The borders of the Saint Antonin basin syncline are outlined by outcrops of Nummulitic limestone (Fig. 1b).The coarse clastic deposits have been subdivided into three units called the lower, middle and upper member of the Saint Antonin conglomerate by Stanley (1980), or simply Formations 1, 2 and 3 by Bodelle (1971) and Callec (2001); we use the latter nomenclature here (Fig. 2).A series of different depositional environments have been proposed for the three formations, ranging from alluvial/lacustrine, fluvialto tidedominated shallow marine and deep marine sediments (e.g., Stanley, 1980 and references therein;Callec, 2001).Following Callec (2001), the sampled units were deposited between 33 and 28 Ma.
In general, the three formations are composed of poorly sorted coarse conglomerates, fine to medium grained sandstones and silt, and locally in the upper part, also of andesitic breccias (Fig. 2).Formation 1, which is about 200-400 m thick, begins with medium to coarse grained, well-sorted sandstones that contain sedimentary structures including ripple marks, mud drapes and flaser bedding, typical of a tidal   Campredon and Giannerini (1982).flat depositional environment.Formation 1 shows a coarsening upward trend into conglomeratic layers.In the southeastern part of the basin decametric to metric granite blocks have been described in the literature (Toure, 1981).

Saint Antonin basin
The 220-350 m thick Formation 2 represents another sandyconglomeratic coarsening upward sequence, which is characterized by the occurrence of andesite breccias and conglomeratic lenses (Bodelle, 1971;Stanley, 1980).Andesite pebbles at the base and the top of Formation 2 have been dated with the 40 Ar/ 39 Ar dating method to 31.1 ± 0.4 Ma and 30.8 ± 0.4 Ma, respectively (Montenat et al., 1999).Formation 3 is the third coarsening upward unit composed of sand and conglomerates, but with a lower proportion of andesitic breccias and pebbles than in Formation 2. Formation 3 is between 180-500 m thick and shows continental characteristics including fluvial channel and overbank deposits particularly in its upper part in the south-eastern corner of the Saint Antonin basin (Campredon and Giannerini, 1982;Callec, 2001).The Saint Antonin basin area experienced a first phase of deformation during the Late-Middle Paleogene, which is reflected by shortening of the underlying Nummulitic limestones and the unconformity with the overlying clastic sedimentary rocks.Late Paleogene syn-sedimentary northsouth directed tectonic shortening caused uplift of the southern flank of the Saint Antonin basin syncline and exposure above sea level accompanied by fluvial sedimentation during deposition of Formation 3. A third regional northwest-southeast directed shortening phase affected the Saint Antonin basin area during the Miocene (Beaudoin et al., 1977;Campredon and Giannerini, 1982;Schreiber et al., 2010).

Sediment provenance of the Saint Antonin basin clastic formations
In the field, the first indication of sediment provenance is the pebble lithology present in the conglomeratic layers of the three formations.In Formation 1, granite, gneiss, sandstone, limestone, rhyolite and basalt pebbles can be observed.For most of the pebbles, the origin is unknown, but the limestone and red rhyolite pebbles in the Saint Antonin conglomerates may have been derived from the sedimentary cover of the Maures-Esterel massif (Bodelle, 1971).Furthermore, it has been proposed by Bodelle (1971) that large, metric-scale boulders of fine grained porphyritic monzogranite, particularly abundant in Formation 1, were derived from a local source to the south of the basin, which has been completely eroded, as no equivalent monzogranites are known in the Corsica-Sardinia block or the Maures-Esterel massif (Toure, 1981).
In Formation 2, Bodelle (1971) identified gabbro with characteristics typical of gabbros found in Alpine Corsica, whereas in Formation 3, Bodelle (1971) identified syenite pebbles that correspond to syenite from outcrops in Hercynian Corsica.These observations of southern sources in the Maures-Esterel massif and the Corsica-Sardinia block are consistent with paleocurrent data (Bodelle, 1971;Stanley, 1980;Callec, 2001) and paleogeographic reconstructions as presented in Joseph and Lomas (2004).

The Maures-Esterel massif and the Corsica-Sardinia block
The crystalline rocks of the Maures-Esterel massif consist mostly of Proterozoic to Paleozoic quartzite, mica schist, gneiss and migmatite, but also of Carboniferous granite and Permian rhyolites.The intrusions, volcanic activity, and part of the metamorphism in the Maures-Esterel massif are associated with the Hercynian orogeny, which also affected part of the crystalline basement of Corsica and Sardinia.The bedrock geology of Corsica is divided into two parts.With the exception of a few outcrops of pre-Hercynian basement rocks, the so-called Hercynian Corsica consists of the Carboniferous Corsica batholith, Permian intrusive and volcanic rocks, and Mesozoic-Cenozoic cover units (e.g., Zarki-Jakni et al., 2004;Danisik et al., 2007;Malusà et al., 2016 and references therein).The second and younger part is Alpine Corsica, consisting of an Alpine metamorphic core complex (e.g., Jolivet et al., 1990;Fournier et al., 1991).During the Jurassic, Corsica and Sardina were tectonically juxtaposed, and during the Cretaceous and Eocene the Maures-Esterel massif, Tanneron massif, Corsica, and Sardinia supposedly formed one block in the Pyreneo-Provençal relief (Malavieille et al., 1998;Andreani et al., 2010, Malusà et al., 2015).Between 33 and 22 Ma Corsica and Sardinia separated from the French Maritime Alps and the Maures-Esterel massif during rifting and opening of the Ligurian Sea (Séranne, 1999;Brunet et al., 2000;Lacombe and Jolivet, 2005;Zarki-Jakni et al., 2004).This rifting started in the western Gulf of Lion (Jolivet et al., 2015).

Fission-track analysis
Seven samples were collected in the field, six from sandstone layers and one from an andesite breccia of Formation 2 (Tab. 1, Fig. 2).Apatite and zircon crystals were separated from bulk rock sandstone samples after crushing and sieving (80-160 mm fraction), using standard magnetic and heavy liquid separation techniques.The andesite breccia sample yielded neither apatite nor zircon.Apatite aliquots from the sandstone samples were mounted in an epoxy resin, polished and etched for 20 s at 21 °C in 5.5M HNO 3 .The samples, covered with muscovite sheets as external detectors, were irradiated together with Durango and Fish Canyon Tuff age standards and IRMM540R dosimeter glasses at the FRM II Research Reactor in Garching, Germany.
Zircon aliquots were mounted in Teflon ® sheets, polished and etched in a laboratory oven at 228 °C between 10 and 35 h in a eutectic melt of NaOH and KOH.The zircon samples, also covered with white mica detectors, were irradiated together with Buluk and Fish Canyon Tuff age standards and CN1 dosimeter glasses.After irradiation, all mica detectors were etched at 21 °C for 18 min in 48% hydrofluoric acid.
All fission-track analyses were done in the ISTerre fissiontrack laboratory in Grenoble, France.Fission-tracks were counted dry at 1250Â magnification in grains mounted parallel to their C-axis, using an Olympus BX51 microscope and the FTStage 4.04 system of T. Dumitru.The grain age distributions of all samples are presented in radial plots, and minimum and central ages (Galbraith and Laslett, 1993) were calculated with the RadialPlotter program of Vermeesch (2009).Given the relatively close age of deposition of about 31 ± 2 Ma of the sampled units, we combined all AFT and ZFT single grain age data for binomial peak fitting using BINOMFIT of Brandon (see Stewart and Brandon, 2004;Ehlers et al., 2005) to determine the main grain age components for provenance analysis (e.g., Bernet et al., 2004;Bernet and Garver, 2005).
The first-order drainage basin maximum and average exhumation rates are estimated from the 1-D thermal advection model age2edot by Brandon (see Ehlers et al., 2005 for details), using the lag-time concept (e.g., Garver et al., 1999;Bernet and Garver, 2005).In this study, lag times are calculated: for the maximum exhumation rate estimates by subtracting the depositional age from the minimum age, and; for the average exhumation rate estimates by subtracting the depositional age from the central age of each sample.

Zircon U-Pb dating and single grain fission-track/ U-Pb double dating
A selection of zircons mounted in Teflon ® sheets for fissiontrack analysis were dated with the laser-ablation ICP-MS U-Pb dating method at the Laboratoire Magmas et Volcans, Blaise Pascal University, Clermont-Ferrand, France.An Agilent 7500 ICP mass spectrometer coupled to a fully computer-controlled 193 nm Resonetics M-50E excimer laser were used with a 20 mm diameter spot size (Hurai et al., 2010 and2012).Some of the same grains had been dated first with the fission-track method for double-dating of single grains.The objective was to analyse the core of the grains to determine if they have 30-38 Ma or much older (>> 100 Ma) crystallization ages, in order to identify a possible Paleogene volcanic contribution.Certain zircon crystals experienced lead-loss, but this is of limited relevance in this study, as no high precision ages are needed to distinguish between Eocene-Oligocene and Hercynian, Pan African or older crystallization ages.All U-Pb data and Concordia plots can be consulted in the supplementary data files.

Fission-track results
AFT ages were determined for samples 10SJ06 and 10SJ07 collected from Formation 2, and sample 10SJ11 collected from Formation 3. The AFT data are shown in Table 2 and presented in radial plots in Figure 3.All these samples show similar results with minimum ages between about 37 and 48 Ma and central ages between 73 and 86 Ma.The spontaneous versus induced track-density plot shows a wide range of values, which is common for the analysis of detrital grains that were not reset after deposition (Fig. 4a), and the uranium concentration versus AFT cooling age plot (Fig. 4b) shows that the majority of grains have uranium concentrations between 10 and 40 ppm but with extreme values of < 5 ppm and > 120 ppm.
The ZFT data are shown in Table 3 and in radial plots in Figure 5.The ZFT minimum ages between 33 and 49 Ma and central ages between 74 and 85 Ma are very similar to the AFT results.Similar to the apatite grains, the spontaneous versus induced track-density plot (Fig. 6a) shows a wide range of values, as it would expected from zircons of detrital samples.The uranium concentration versus ZFT cooling age plot (Fig. 6b) shows that most analysed grains have uranium concentration between 100 and 400 ppm.
Lag-times calculated from AFT and ZFT minimum and central ages are given in Table 4. AFT lag times vary between 6.6 and 15.3 Myr, with a mean around 12.2 ± 4.3 Myr for minimum ages, and between 42.8 and 54.1 Myr, with a mean of 50.3 ± 4.3 Myr for central ages.ZFT lag times range from 4.4 to 19.0 Myr, with a mean around 12.3 ± 4.8 Myr for minimum ages, and from 41.0 to 54.4 Myr, with a mean of 46.3 ± 5.6 Myr for central ages.
Binomial peak fitting results of the AFT and ZFT data are shown in Figure 7 (also see the data repository for single grain data).Given the close range of depositional ages of sampled units to obtain a general detrital age signal of the Saint Antonin basin the grain ages can be combined and treated as one sample.The combined 244 single grain AFT ages of the three samples can be decomposed into four major age components at about 37, 59, 85, and 175 Ma.The best-fit solution for the combined 206 single grain ZFT ages of the four samples results in three major age peaks at about 33, 61, and 134 Ma.

Zircon U-Pb and fission-track/U-Pb double dating results
A total of 109 zircons were dated with LA-ICP-MS analysis (Fig. 8a) and 62 of these zircons were previously dated with the fission-track method (Fig. 8b).With the exception of one single zircon from sample 10SJ10 in Formation 3, with a crystallization age of 30 ± 1 Ma, all other zircons have U-Pb ages ranging between 277-2812 Ma (Fig. 8a).Almost 40% of   the grains show Hercynian ages (∼360-280 Ma) and another 20-25% Pan African ages (∼460-660 Ma).
The zircon grain with a 30 Ma U-Pb age also has a 34 Ma ± 25 ZFT cooling age, which means that this grain has to be considered as the only Oligocene volcanic zircon grain observed in our samples.The ZFT data show that a few grains with Hercynian and Pan African crystallization ages have Eocene-Oligocene cooling ages, but otherwise show a large spread of apparent cooling ages between ca.55-255 Ma (Fig. 8b).

Apatite and zircon fission-track ages in the source areas
The grain age distribution of detrital apatite and zircon samples from sandstone reflects a mixture of apparent cooling ages in the drainage area at the time of deposition.The spread of the observed grain age distribution depends, among other factors, on the size of the drainage area, relief, erosion rates, bedrock lithologies, and of course the bedrock cooling history (e.g., Garver et al., 1999;Bernet et al., 2004).If fission-track data are available from the present-day outcrops in the drainage area, they can be compared with the detrital grain age distributions to provide valuable information and constraints on potential paleo-source areas.To the best of our knowledge, no bedrock ZFT data have been published for the Maures-Esterel massif, but AFT ages between 21-37 Ma were determined for the Plan de la Tour granite, Moulin Blanc granite, and Bornes gneiss in the Maures massif; of about 93 Ma for the Saint Tropez granite; and around 140 Ma in the Tanneron massif west of Cannes (Fig. 9; Lucazeau and Mailhé, 1986;Morillon, 1992Morillon, , 1997;;Jakni, 2000).
Paleozoic basement and Triassic-Jurassic cover rocks in eastern and western Sardinia have AFT ages over 50 Ma and increase to 100 to 306 Ma in southwestern Sardinia (Rossi et al., 2005;Zattin et al., 2008, Malusà et al., 2016).AFT cooling ages of 30 Ma and younger can be found along the central graben in Sardinia, but the majority of AFT ages in Sardinia ranges from about 40 to 170 Ma (Rossi et al., 2005;Malusà et al., 2016).ZFT ages throughout Sardinia are between 140-240 Ma (Fig. 9; Malusà et al., 2016).Single grain apatite (U-Th)/He data for Sardinia are as variable as the AFT ages, ranging from 26-204 Ma (Malusà et al., 2016).
Late Oligocene and Miocene AFT ages in the Maures-Esterel massif, Sardinia and Corsica are related in the literature to extension during the phase of rifting in the Provençal-Ligurian basin and counter-clockwise rotation of the Corsica-Sardinia block away from the Maures-Esterel massif (Vigliotti and Langenheim, 1995;Séranne, 1999;Brunet et al., 2000;Rollet et al., 2002;Zarki-Jakni et al., 2004;Fellin et al., 2006;Danisik et al., 2007;Malusà et al., 2016).At the same time the Saint Antonin area experienced a phase of compression (Campredon and Giannerini, 1982), with gradual surface uplift and eventual exposure of the basin syncline.Late Triassic ZFT ages in western Corsica were interpreted to reflect partially annealed zircons during Jurassic rifting, and Early Jurassic to Late Cretaceous ZFT in central Corsica relate to the Liguria-Piedmont ocean opening according to Danisik et al. (2007) and Malusà et al. (2016).Nonetheless, the Late Cretaceous ZFT ages are debatable, as Fellin et al. (2006) considered such ZFT ages as well partial annealed ages.

Provenance information from geo-thermochronological data
Based on structural, paleomagnetic and petrographic data, Westphal et al. (1976) had already shown that the Corsica-Sardinia block was directly connected with the Porvence area of southern France, with Corsica in a location close to where Nice and Cannes are today (Lacombe and Jolivet, 2005).In addition, from the sediment petrological and sedimentological evidence of the Saint Antonin basins mentioned above, the sediment source areas were located to the south and south-west of the basin, including the Maures-Esterel massif, Corsica and Sardina (Stanley, 1980;Campredon and Giannerini, 1982).
The question is which of the three potential source areas is the most important?We first look at the zircon U-Pb data to elaborate this question.Zircons with Pan-African and older U-Pb ages are most likely derived from Permo-Triassic sedimentary cover rocks in the Maures-Esterel massif, Sardinia and Corsica, or from meta-sedimentary rocks of the Alpine Corsica zone.However, large fractions of zircons dated in this study have Hercynian U-Pb ages between 280 and 350 Ma.Hercynian granitic intrusions and migmatites are widespread throughout the Maures-Esterel massif, Sardinia and Corsica (e.g., Morillon et al., 2000;Duchesne et al., 2013).Therefore, the zircon U-Pb data alone do not provide solid criteria for distinguishing between the three source areas.Only one zircon grain with an Oligocene crystallization age was identified, which is, given its corresponding ZFT cooling age, interpreted to be derived from a volcanic source.Because this contribution is very small, and because the andesite breccia we sampled in the Saint Antonin basin did not yield any apatite or zircon, we think that Oligocene volcanic rocks are not a major source and the detrital AFT and ZFT cooling age signals discussed below represent source area exhumation.Here the questions is how can the AFT and ZFT ages be linked to source areas?A basic assumption is that for apatite and zircons with Eocene and early Oligocene cooling ages only bedrock exposed at the surface today which have early Oligocene or younger cooling ages can be considered as potential source areas (Fig. 9).In the Saint Antonin basin, we interpret apatites with early Oligocene cooling ages being mainly derived from source areas in Corsica and possibly from the Maures-Esterel massif, but there outcrops with early Oligocene AFT cooling ages are rather limited, at least from what is known from the published data (Fig. 12).Zircons with early Oligocene cooling ages have most likely a source area limited to Alpine Corsica, in the northeast of the island (Fig. 9).Alpine Corsica is separated from Hercynian Corsica by what is regarded as the southern extension of the Penninic front (Lacombe and Jolivet, 2005).
Apatites and zircons with Late Cretaceous to Early Tertiary apparent cooling ages may have been derived from very slowly cooled Hercynian intrusive rocks or from partially reset (cover?) units, which do not provide a clear age signal.Jurassic to Early Cretaceous apparent cooling ages most likely come from source rocks in the Maures-Esterel and Tanneron massifs, Sardinia and Corsica (Fellin et al., 2006;Danisik et al., 2007;Malusà et al., 2016).
The zircon double-dating results in Figure 8b show that these different sources need to be considered.The lack of a dense fission-track dataset in the Maures-Esterel massif prohibits more precise constraints.However, the published ZFT data from Corsica and Sardinia support potential source areas in the eastern, central and western parts of Hercynian Corsica for zircons with Early Cretaceous and older apparent fission-track cooling ages and Alpine Corsica for zircons with Late Cretaceous and Paleogene fission-track cooling ages (Fellin et al., 2006;Danisik et al., 2007;Malusà et al., 2016).We think that all of these Corsican sources were exposed and eroded during the early Oligocene and some of these sediments were deposited in the Saint Antonin basin, before Corsica and Sardinia drifted away during the late Oligoceneearly Miocene (Lacombe and Jolivet (2005), and were removed as potential source areas, and the Saint Antonin basin was fully inverted and uplifted (Campredon and Giannerini, 1982).Given the overall younging age trend in the AFT and ZFT data that can be observed from southern Sardinia to northern Corsica suggests that northern Corsica was possibly the most prominent sediment source area for the Saint Antonin basin (Fig. 9; Malusà et al., 2016).

Source area exhumation rates
The minimum and central ages of the individual AFT and ZFT samples are clearly older than the given depositional ages between about 33 and 29 Ma.This, together with the relatively shallow burial of the sampled sedimentary rocks, suggests that fission-tracks in apatite and zircon were not affected by strong post-depositional annealing.As outlined above with the zircon double-dating results, a volcanic contribution seems to be negligible.Therefore, the AFT and ZFT data presented in this study can be used for estimating drainage basin average and maximum exhumation rates.The lag times shown in Table 4 can be used with Figure 11 to obtain first-order estimates of exhumation rates.Drainage basin average long-term exhumation rates determined from the mean central-age lag times are on the order of 0.1 to 0.2 km/Myr (Fig. 10).The maximum exhumation rate estimates based on the minimum-age lag-time calculations are on the order of 0.4-0.7 km/Myr (Fig. 10).Only about 10 to 15 percent of the dated apatite and zircon grains have Eocene to Oligocene cooling ages.While ignoring potential lithologic bias, this could mean that probably only small areas were affected by faster erosion.Overall slow erosion in the proximate source area of the Maures-Esterl massif is consistent white mica 40 Ar/ 39 Ar data from the Tanneron massif, which are in the 320-300 Ma range (Corsini et al., 2010), indicating that these rocks had cooled below temperatures of about 400-350 °C during the Carboniferous and where slowly exhumed afterwards, as no white micas with younger 40 Ar/ 39 Ar cooling ages were detected.Because the Pyrenees-Provence belt formed during the collision of the Corsican-Sardinia-Iberia block with the European plate during the Eocene (Arthaud and Séguret, 1981;Lacombe and Jolivet, 2005;Jolivet et al., 2015), the erosional response followed.Erosional exhumation was slow in the sources areas and may have accelerated only during the late Eoceneearly Oligocene, when the erosional products were transported and deposited in the Saint Antonin basin, at the end of the Pyreneo-Provençal compression phase, and before late Oligocene-early Miocene Ligurian basin rifting had reached the region (Lacombe and Jolivet, 2005;Jolivet et al., 2015).

Comparison with the Barrême basin
If we compare detrital zircon U-Pb ages of the Saint Antonin basin with zircon U-Pb ages from sandstone deposited between 30 and 29 Ma in the Barrême basin, we see a similar age spread and a large number of zircons with Hercynian crystallization ages in both basins (Jourdan et al., 2013, and this study).However, using the Kolmogorov-Smirnov (KS) test (e.g., Press et al., 1992), we can determine if the difference between the zircon U-Pb age distributions for the two basins is significant or not.A P(KS) value of < 5% indicates that the difference is systematic and a P(KS) value of >> 5% indicates that the difference is most likely due to random chance alone.As shown in Figure 11a, with a P(KS) = 0% the two distributions are significantly different in as much as many more Oligocene volcanic zircons were identified in the Barrême basin deposits of the Clumanc and Saint Lions conglomerates (Jourdan et al., 2013) than in the Saint Antonin basin sedimentary rocks.
In the ZFT data, the differences are more striking, as the 30 to 29 Ma Barrême basin deposits contain large numbers of zircons with ∼30 Ma cooling ages (Fig. 11b).The P(KS) = 0 value emphasizes this significant difference.Jourdan et al. (2013) had shown with single zircon grain double-dating that only some of these grains are of volcanic origin, but the majority seems to reflect rapid source area exhumation in the Western Alps.Whereas zircons with ∼33 Ma fission-track cooling ages in the Saint Antonin basin make up only about 11% of the dated zircons, in the Clumanc and St. Lions conglomerates ∼30 Ma zircons dominate, comprising 67-68% of the dated grain-age distributions (Jourdan et al., 2013).In contrast, when comparing the ZFT grain-age distribution of the Grès de Ville, deposited between 31-30 Ma in the Barrême basin, with the ZFT grain-age distribution of the Saint Antonin basin, we see that the two distributions are very similar and a P
(KS) = 31.1% value indicates that differences are most likely due to random chance alone (Fig. 11c).This result is not surprising as the Grès de Ville was sourced from the Maures-Esterel massif to the south (Fig. 12; Evans and Mange-Rajetzky, 1991;Joseph and Lomas, 2004), similar to the Saint Antonin deposits and in contrast to the Clumanc and St. Lions conglomerates of the Barrême basin, which had their sources in the Western Alps (Evans and Mange-Rajetzky, 1991;Schwartz et al., 2012;Jourdan et al., 2013).

Regional tectonics and basin evolution
Formations 1, 2, and 3 of the Saint Antonin basin were deposited at roughly the same time as the Grès de Ville, Clumanc and Saint Lions conglomerates in the Barrême basin (Callec, 2001).Both basin remnants were affected by synsedimentary compressional deformation.The Barrême basin experienced east-west shortening, principally on its eastern flank between 30-28 Ma, when the first sediments arrived from the internal Western Alps (e.g., Evans and Mange-Rajetzky, 1991;Schwartz et al., 2012;Jourdan et al., 2013).The syn-sedimentary deformation in the Saint Antonin basin from the late Eocene to the early Oligocene was northsouth directed and mainly affected the southern flank of the basin.Therefore, the two basins show similar deformation styles but with different orientation of tectonic structures while being in the same Western Alps pro-side foreland basin stress field (Figs. 1, 9 and 12;e.g. Ford et al., 1999;Ford and Lickorish, 2004).The main phase of compressional tectonics in the Saint Antonin basin is linked to the final stages of the Pyreneo-Provençal N-S shortening phase, indicating that in this area compression occurred until about 28 Ma, during    et al., 2009;Jourdan et al., 2013).A P(KS) = 0% value shows that the two distributions are significantly different, as is also the case with the zircon U-Pb ages.c) Cumulative probability plots of the combined Saint Antonin ZFT data plotted against the Grès de Ville ZFT data of the Barrême basin (Jourdan et al., 2013).A P(KS) = 31.1% value indicates that the two distributions are very similar and differences are probably only due to random chance variations.
deposition of the 3rd clastic formation in the basin.This makes the Saint Antonin basin being a kind of foreland basin of the Pyreneo-Provençal thrust belt which includes the Maures-Esterel massif, Corsica and Sardinia.Furthermore, the timing of deformation and rate of exhumation is comparable to what has been observed in the Pyrenees (e.g., Fitzgerald et al., 1999;Beamud et al., 2011;Vacherat et al., 2016).From about 28 Ma on rifting, which had already started in the Gulf de Lion region before, affected the region south of the Saint Antonin basin.This interpretation is consistent with the model proposed by Malusà et al. (2016) who based on AFT, ZFT and apatite (U-Th)/He data concluded that the late Oligocene and Miocene apparent cooling ages in Corsica are related to extensional cooling during rifting, but that northern Corsica experienced exhumational cooling caused of erosion during the late Eocene and early Oligocene times.As mentioned above, Alpine Corsica is separated from Hercynian Corsica by the southern extension of the Penninic front.Shortening along the Penninic front stopped here during the Oligocene (Lacombe and Jolivet, 2005), after the deposition of the Saint Antoni basin units, when the regional stress regime changed to extension during the opening of the Ligurian basin.This change coincides with the change in convergence between the Adriatic and European plate in the Western Alps as outlined by Dumont et al. (2012), and rollback of the Adriatic slab in the Mediterranean (e.g., Malusà et al., 2015).Lacombe and Jolivet, 2005;Dumont et al., 2012;Jourdan et al., 2013;and Malusà et al., 2015).White arrows indicated main sediment supply to the two basin remnants.
As mentioned above, only the Barrême basin recorded a significant change in sediment provenance from a southern to a north-eastern direction, whereas the Saint Antonin basin was sourced always only from the south (Fig. 12).The change in provenance in the Barrême basin and the east-west shorting has been explained with a change in convergence directions in the Western Alps between the Apulian and European plates during the mid-Oligocene (Fig. 12; e.g.Dumont et al., 2012).Whereas, the Barrême basin received molasse sedimentation until the early Miocene from the Western Alps (e.g., Evans and Mange-Rajetzky, 1991;Callec, 2001;Schwartz et al., 2012), and was continuously affected by east-west shortening, thrusting and transport on the Digne thrust-sheet (Fig. 12; e.g.Artoni and Meckel, 1998;Lickorish and Ford, 1998;Evans and Elliott, 1999;Evans et al., 2004), sedimentation stopped in the Saint Antonin basin and no late Oligocene to early Miocene molasse sediments were deposited, as Corsica, the most important sediment source area, was removed by the opening of the Ligurian basin.Therefore, regional tectonics were the main driver of basin subsidence, depositional environments, sediment provenance, syn-sedimentary deformation, and eventual basin inversion.

Conclusions
The geo-thermochronologic data of the Saint Antonin basin presented in this study are consistent with sediment provenance from the Maures-Esterel massif, Sardinia and Corsica.The Saint Antonin basin clastic formations were deposited between 33 and 28 Ma, at the end of the Pyrenean-Provencal collision phase and just before the initiation of the opening of the Ligurian Sea and rifting of Corsica and Sardinia away from the Maures-Esterel massif, and the single grain zircon fission-track and U-Pb dating shows that a volcanic contribution is negligible, and apatites and zircons with early Oligocene fission-track cooling ages are in the Saint Antonin basin, even if rare, are regarded as the result of erosional exhumation at maximum rates of up to 0.4-0.7 km/Myr.The majority of the AFT and ZFT cooling ages however indicate overall relatively slow long-term average erosional exhumation rates on the order of 0.1-0.2km/Myr in the Maures-Esterel, Sardinia and Corsica, consistent with the rather limited amount of sediments deposited in the Saint Antonin basin.In comparison, the published geo-thermochronologic data of the Barrême basin, show similar age ranges but with the difference of a large proportion of zircons with early Oligocene fissiontrack cooling ages, partly of volcanic origin and partly derived from rapid erosional exhumation in the Western Alps.The Saint Antonin basin deposits did not record a significant change in sediment provenance at about 30 Ma, as it has been documented in the Barrême basin.This means that the Saint Antonin basin was not affected by the same changes in foreland basin dynamics and sediment routing systems as the Barrême basin, even if the two foreland basin remnants are less than 50 km apart.In the Saint Antonin basin sedimentation ended around 28 Ma, with eventual basin inversion and surface uplift and removal of the main source areas in Corsica-Sardinia block by continued rifting in the Ligurian Sea.

Fig
Fig. 1. a) Regional overview map of southeastern France showing the location of the Saint Antonin basin, Annot basin, and Barrême basin.TM: Tanneron massif; PTG: Plan de la Tour granite; STG: Saint Tropez granite.b) Simplified geologic map.c) Cross-section of the Saint Antonin basin modified afterCampredon and Giannerini (1982).

Fig. 3 .
Fig. 3. Detrital AFT age data shown in radial plots for samples 10SJ06 and 10SJ07 of Formation 2, and sample 10SJ11 of Formation 3. Graphs were plotted and central and minimum ages were calculated with the RadialPlotter program by Vermeesch (2009).
Fig 4. a) Spontaneous versus induced track density plot, and b) uranium concentration versus AFT age plot.
Fig 8. a) Combined data set of detrital zircon U-Pb ages from all three clastic formations of the Saint-Antonin basin.b) Double dating results of single grain detrital ZFT and U-Pb analyses.

Fig. 11 .
Fig. 11.a) Cumulative probability plots of the combined Saint Antonin zircon U-Pb data plotted against the combined zircon U-Pb data of the Clumanc and St. Lions conglomerates of the Barrême basin(Jourdan et al., 2013).A P(KS) = 0% value shows that the two distributions are significantly different.b) Cumulative probability plots of the combined Saint Antonin ZFT data plotted against the combined ZFT data of the Clumanc and St. Lions conglomerates of the Barrême basin(Bernet et al., 2009;Jourdan et al., 2013).A P(KS) = 0% value shows that the two distributions are significantly different, as is also the case with the zircon U-Pb ages.c) Cumulative probability plots of the combined Saint Antonin ZFT data plotted against the Grès de Ville ZFT data of the Barrême basin(Jourdan et al., 2013).A P(KS) = 31.1% value indicates that the two distributions are very similar and differences are probably only due to random chance variations.

Fig. 12 .
Fig. 12. Simplified paleogeographic model of the of the Western Alps south-eastern foreland basin with the Saint Antonin and Barrême basins at A-B) 31 Ma and about 25 Ma C-D) with respect to the rising internal Western Alps, opening of the Ligurian Sea and rotation of the Corsica-Sardinia block during the early to mid-Oligocene (modified afterLacombe and Jolivet, 2005;Dumont et al., 2012;Jourdan et al., 2013;and Malusà et al., 2015).White arrows indicated main sediment supply to the two basin remnants.

Table 2 .
Vermeesch (2009)track data.Note: Samples were counted dry with a BX51 Olympus microscope at 1250Â magnification.Central and minimum ages were calculated with the RadialPlotter program ofVermeesch (2009), using a zeta value of 214.63 ± 18.5 and the IRMM540R dosimeter glass.rs: spontaneous track density; ri: induced track density; rd: track density of the fluence monitor.
Vermeesch (2009)re counted dry with a BX51 Olympus microscope at 1250Â magnification.Central and minimum ages were calculated with the RadialPlotter program ofVermeesch (2009), using a zeta value of 104.39 ± 3.32 and the CN1 dosimeter glass.
Detrital ZFT age data shown in radial plots for sample 09SJ29 of Formation 1, sample 09SJ33 of Formation 2, and samples 10SJ10 and 10SJ11 of Formation 3. Graphs were plotted and central and minimum ages were calculated with the RadialPlotter program by Vermeesch (2009).

Table 4 .
Lag-time calculations.Bold means average values.
First order estimates of source area exhumation rates from detrital apatite and zircon fission-track minimum-age and central-age lag times are derived from the 1-D steady-state thermal advection model age2edot by Brandon (seeEhlers et al., 2005 for details).Model parameters given in the plot are: Ts: surface temperature; Tgrad: thermal gradient; Zlw: crustal thickness; Tlw: temperature at the base of the crust; kappa: thermal diffusivity.The lag time (Myr) is the time difference between the age of deposition and the fission-track age.The black dotted lines show the mean lag-time estimates of the combined AFT and combined ZFT samples.The exhumation rate is estimated from the Y-axis.