Geochemical and Dynamic Model of Repeated Hydrothermal Injections in Two Mesozoic Successions, Provençal Domain, Maritime Alps, SE-France

: A ﬁeld, petrographic and geochemical study of two Triassic–Jurassic carbonate successions from the Maritime Alps, SE France, indicates that dolomitization is related to episodic fracturing and the ﬂow of hydrothermal ﬂuids. The mechanism governing hydrothermal ﬂuids has been documented with the best possible spatio-temporal resolutions specifying the migration and trapping of hydrothermal ﬂuids as a function of depth. This is rarely reported in the literature, as it requires a very wide range of disciplines from facies analysis (petrography) to very diverse and advanced chemical methods (elemental analysis, isotope geochemistry, microthermometry). In most cases, our di ﬀ erent recognized diagenetic phases were mechanically separated on a centimetric scale and analyzed separately. The wide range of the δ 18 O VPDB and 87 Sr / 86 Sr values of diagenetic carbonates reﬂect three main diagenetic realms, including: (1) the formation of replacive dolomites (Type I) in the eogenetic realm, (2) formation of coarse to very coarse crystalline saddle dolomites (Types II and Type III) in the shallow to deep burial mesogenetic realm, respectively, and (3) telogenetic formation of a late calcite cement (C1) in the telogenetic realm due to the uplift incursion of meteoric waters. The Triassic dolomites show a lower 87 Sr / 86 Sr ratio (mean = 0.709125) compared to the Jurassic dolomites (mean = 0.710065). The Jurassic calcite (C1 J ) shows lower Sr isotopic ratios than the Triassic C1 T calcite. These are probably linked to the pulses of the seaﬂoor’s hydrothermal activity and to an increase in the continental riverine input during Late Cretaceous and Early Cenozoic times. This study adds a new insight into the burial diagenetic conditions during multiple hydrothermal ﬂow events.


Introduction
The formation of low-versus high-temperature dolomites in carbonate successions is attributed to separate diagenetic conditions and fluid origins [1,2]. Low temperature dolomitizing fluids are associated with near surface diagenetic fluids [3][4][5]. Hot dolomitizing fluids are either the result of Figure 1. General geological map of the south-west Alps. The red rectangle shows a part of the study area (Provençal Domain) (after [14]).
The European paleo-margin experienced extensional and transtensional tectonics during the opening of the Alpine Tethys during the Jurassic and Cretaceous [16]. A subsequent regional unconformity followed the Triassic succession and encompasses the Late Triassic-Early Jurassic succession [14]. The emersion of the Provençal Domain during Early Jurassic times led to the formation of a carbonate platform followed by a continuous decrease in the sea level and the consequent establishment of a peritidal environment during the Berriasian [17,18]. The deposition of shallow-water marine carbonates continued during the Cretaceous [14]. The Mesozoic succession is separated from the Cenozoic deposits by a regional unconformity that is equivalent to the late Cretaceous-Middle Eocene interval [14]. A rapid lateral facies variation is observed in both Provençal and Dauphinois Domains and the Mesozoic succession is thinner toward the Provençal area [14].   (1) is extensively brecciated and the floated fragments have been cemented by saddle dolomite (see the enlarged part of this breccia on the right part "1"). The breccia is followed by another structure, i.e., zebra dolomites.

Middle Triassic Section
The micrite and replacive microdolomite (D1 T ) composition is more common in breccia and zebra fabrics of the Triassic than in the Jurassic carbonates. Despite strong dolomitization, the depositional features are still preserved including laminated algal/microbial mats and slightly compacted micritized ooids with darker micritic envelopes in grain-supported fabrics ( Figure 4A,B). Dolomite 2 (D2 T ) is composed of medium to coarse grained anhedral-subhedral crystals (70-200 µm) in a nonplanar to planar-s fabric. D2 T replaced the D1 T dolomite and display relict ooids ( Figure 4A,B). Saddle dolomite (SD1 T ), which occurs as irregular compacted crystals (0.3 to 0.6 mm) with nonplanar-s fabric, replaced D2 T . SD1 T commonly occurs in the dolostones and, less commonly, as vug-lining cement in fractures. Coarser saddle dolomite SD2 T crystals (0.5 up to 4 mm) consist of homogeneous euhedral slightly curved, sometimes planar, crystals. In contrast to SD2 T , SD1 T appears as a cement in irregular fenestrae and breccias and as zebra dolomite fabric. The brecciated dolostones D2 T (clast sizes up to a few cm) are cemented by various dolomitic types. SD1 T is distinguished from SD2 T crystals by its fabric retentive, inclusion-free rims and inclusion-rich cores. SD1 T and SD2 T saddle dolomites are characterized by a similar growth zoning with bright red and dark red luminescence. The gray bands of D2 T alternate with the white bands of the saddle dolomites (SD1, SD2), forming the zebra fabric and are thus genetically related to the fracturing of the dolostones. SD1 T and SD2 T are crosscut by low amplitude stylolites ( Figure 3C) that were aligned parallel to fracture-filling saddle dolomite.
Light brown colored, coarse-grained, euhedral, slightly planar, Fe-rich saddle dolomite crystals (SD3 T ) are found in the breccias ( Figure 4D-F). The breccias vary from clast-supported (jigsaw puzzle) to cement-supported textures ( Figure 2B). SD3 T is associated with stylolites with sets of fractures or tension gashes. The tension gashes are parallel to the columns of stylolites and terminated at the wider end of the stylolite seam. SD3 T exhibited the typical saddle characteristics with a wavy extinction and curved crystal faces under a cross-polarized light, and has a uniform bright red luminescence under CL.
Euhedral, Fe-poor SD4 T saddle dolomite ( Figure 5A) displayed alternating inclusion-rich cores and -free rims. The margins of SD4 T crystals were corroded and replaced by calcite (C1 T ; Figure 5B,C). Similar calcite filled the vugs and fractures and replaced D2 T , SD3 T and SD4 T too.
Another coarse-crystalline, euhedral, nonplanar, multizoned, red luminescent fracture and vug-filling saddle dolomite (SD5 J ) ( Figure 6E,F) was observed. SD5 J crystals are characterized by alternating inclusion-rich and inclusion-free zones. High-amplitude stylolites cross cut the fracture-filling saddle dolomites; no tension gashes were observed. The SD6 J dolomite is distinguished from the previous saddle dolomites by its elongated crystal shapes ( Figure 7A-C). This SD6 J dolomite is densely packed and systematically associated with D4 ( Figure 6D). CL revealed an alteration of thick red luminescence (SD5 J ) and thick nonluminescent SD6 J ( Figure 8B,C). In addition, thin bright red luminescence is associated with fractures that postdated SD6 J .  A Fe-poor calcite cement (C1 J ) fills the central parts of the fractures and vugs, hence the latest cement ( Figure 8A,B). The calcite sizes are ranged from fine to coarse crystals (>400 µm), precipitated and postdated saddle dolomite formation, and had crystals that were transparent with whitish and yellowish colors. CL revealed an alteration of thick red luminescence (SD5 J ) and thick nonluminescent SD6 J ( Figure 8B,C). In addition, thin bright red luminescence was associated with small veins and fractures that postdated SD6 J .
Dolomite recrystallization (Dr) occurred after all saddle dolomitic phases, where the ghosts of the pristine oolitic facies still resisted diagenesis, and in places the coarse sized poikilotopic cement fabric (>700 µm) developed. Dr displays the same characteristics as those observed in SD6 J under CL, i.e., nonluminescent thick with thin bright red luminescent bands ( Figure 8E,F).
Because of the small sizes of fractures and cavities, and also the fact that sometimes two dolomite phases occur together in a fracture during mixing the carbonate phases are categorized into four groups to obtain precise data: type 1 with D J , D1 T and D2 T ; type II with saddle dolomites from SD1 T/J and SD2 T/J ; type III with SD3 T/J , SD4 T , D4 J , SD5 T/J , Dr, and SD6 J and type IV with C1 T/J . This phase succession will also make it easier to interpret the geochemical composition, especially when two carbonate phases are mixed.

Early and Late Diagenetic Stylolites
Most stylolites are used as a key of the burial depth, especially those aligned parallel to the bedding planes. The Triassic-Jurassic carbonate succession is characterized by two types of stylolites: (i) low amplitude stylolites (no more than 1 cm in length) crosscut Type I dolomites. The low amplitude stylolites (early stylolites) were the first ones produced during the Triassic-Jurassic diagenesis and are associated with a jump-up in McArthur's graph [19]; (ii) early stylolites are followed by larger stylolites (several centimeters in length with high amplitudes) associated with a set of fractures and tension gashes. Stylolite-related fractures (tension gashes) are rare diagenetic features formed by increasing and progressive mechanical loading during burial, evidenced by the parallel orientation with the bedding plane [20,21].

Chemical Composition (Major and Trace Elements)
Nineteen samples were selected for trace element analyses from both outcrops (Table 1).

NM4
Type Generally, Ca (mean: 22.1 wt. %; n = 15) and Mg (mean: 12 wt. %; n = 15), Ca (mean: 39.2 wt. %; n = 2) and Mg (mean: 0.6 wt. %; n = 2) are the major elements in dolomite and calcite phases, respectively. Other elements like Sr (mean: 112 ppm; n = 15 for dolomite; mean: 330 ppm; n = 2 for calcite), Fe (mean: 580 ppm; n = 15 for dolomite; mean: 116 ppm; n = 2 for calcite), Mn (mean: 95 ppm; n = 15 for dolomite; mean: 180 ppm; n = 2 for calcite), Ba (mean: 3 ppm; n = 15 for dolomite; mean: 4 ppm; n = 2 for calcite), represent crucial "trace elements" in the carbonate diagenetic setting ( Figure 9). The relative abundance of major elements and scarcity of trace elements in dolomite and calcite, are likely the recorded evidence of diagenetic processes (e.g., [22]). to SD3T, which was overlain by the calcite cement, i.e., due to mixing of SD3T and C1T during preparation.  The Ca/Mg ratio of the saddle dolomites (Types II and III) ranges between 1.72 and 2.14 from the Triassic samples, and 1.69 and 2.08 from the Jurassic samples. The Ca/Mg ratio of the late calcite phase in the Triassic is 68 and in the Jurassic is 85 (Figure 9). The Jurassic dolomites show a gradual increase in Fe and Mn, starting from 34 going up to 226 ppm for Mn, and 78 to 730 ppm for Fe. The Triassic dolomites exhibit high Fe and Mn values (mean: 846 and 92 ppm, respectively), with a wide range. Rarely do the Mn contents of the early dolomites in both sections overlap the Mn contents of the later dolomites. The early dolomites display the highest Fe contents. The Fe and Mn contents in the Triassic and Jurassic limestones share the same values with a few late dolomite samples. In contrast, late calcite-Type IV (C1 J and C1 T ) contents exhibit the lowest Fe contents (C1 T : Fe = 204 ppm; Mn = 146 ppm) (C1 J : Fe = 27 ppm; Mn = 13 ppm). Higher Fe and Mn contents in C1 T could be related to SD3 T , which was overlain by the calcite cement, i.e., due to mixing of SD3 T and C1 T during preparation.

δ 18 O VPDB and δ 13 C VPDB Isotopes
The stable isotopic compositions of ten limestone samples (

Paragenetic Sequence Related to Dolomitizing Fluids
The petrographic and geochemical data obtained suggest that dolomitization has been accomplished by various episodes of hydrothermal flow events under various burial diagenetic conditions.
The earliest dolomitization, which affected mainly the micrite mud matrix, probably occurred at near-surface conditions. This is evidenced by the anhedral, non-saddle dolomite fabric [28] and the uncompacted ooids. The nonplanar crystal shape of dolomite Type I D2 T could be linked to a diagenesis at a higher temperature (e.g., ≥ 50 • C, [28]). However, these nonplanar dolomites are probably related to impurities in the growth media, which stabilize the crystal faces [29], and not to high temperature fluids. This is supported by petrographic evidence-for example, D2 T crystals do not show any characteristics of hot fluid involvement "i.e., lack of curved faces and sweeping extinction".
It is well reported in the literature that saddle dolomite forms under high temperatures typically by flow of hydrothermal fluids [30]. Type II is dissolved by low-amplitude, early stylolites, in high temperature conditions (senso, [31]) in the shallow-intermediate depth without necessarily implying deep burial conditions "i.e., eogenesis".
Brecciation and concomitant dolomitization (Type II dolomites) suggest a flow of hydrothermal fluids along fractures ( Figure 10). Brecciation was accompanied by the dissolution of the host carbonates, which resulted in the creation of new pathways for fluid of hydrothermal fluids and cementation by producing the Type III saddle dolomite. Further cementation by the latter dolomite was probably related to hydrofracturing and resetting of the zebra and breccia fabrics. Stylolite-related fractures are parallel to the amplitude of stylolite ( Figure 4F), and according to Nelson [20] the tension gashes are not true tension fractures, but are extensional fractures derived from the same compressive state of stress as the fractures of stylolites [32]. Filling the tension gash fractures with SD3 confirmed that the late stylolites are coeval with Type III saddle dolomite formation (Figures 11 and 12). Stylolite-related fractures (tension gashes), which were produced during deep burial mesogenesis, were cemented by Type III saddle dolomite ( Figure 4F). Hence, Type III saddle dolomite formation was presumably formed by the flux of hot basinal fluids along stylolites during compressional basin tectonics [33].   Subsequent openings of the veins or fractures in Type II facies led the hot secondary brine fluid to percolate along these pathways, which resulted in the precipitation of Type III saddle dolomites. The hydrothermal activity associated with hydrofracturing affecting the studied Mesozoic series has also been reported by [34][35][36].
The homogenization temperature of saddle dolomites from the Dauphinois-Provençal Domains varied between 170 and 260 • C ( [14]; Figure 10), which rules out a low temperature origin for dolomite. The final melting temperature of fluid inclusions have been reported as highly saline fluids, and the salinity was approximately equivalent to 20% to 23% CaCl 2 [14], this confirms the interaction of hot circulating fluids with Argentera basement. The high temperature dolomitizing fluid is further supported by low δ 18 O VPDB values. The depletion of the δ 18 O VPDB values follows the increase in the size of saddle crystal from Type II to Type III (up to −11.6% in Type III dolomites). Therefore, the plausible explanation that can adequately explain the range of homogenization temperatures and considerable range of oxygen isotope values is that the precipitation of dolomite phases occurred at different diagenetic settings during which a repeated injection of hydrothermal fluid occurred (Figures 13 and 14). This interpretation is further supported by the following lines of evidence (Figures 4 and 6): (1) suture intergranular contacts between the ooids are engulfed by the saddle dolomite cement; (2) occurrence of sulfides and dissolution of stylolite-related fractures in type III saddle dolomites; (3) co-occurrence of Type III saddle dolomites and late stylolite-related fractures, later the cements inside the fractures were dissolved and etched; (4) coarse-crystalline poikilotopic cement fabric (>700µm) etched and corroded the Dr boundaries; (5) dissolution, compacted saddle dolomite and association with sulfides; (6) the presence of high-amplitude stylolites parallel to the overburden force.  Therefore, according to the abovementioned points, our interpretation agrees with the one of Wierzbicki et al. [37], who suggested that the partial dissolution of dolomite crystal outlines fits the deep burial mesogenesis. This late evolution could be linked to the subsidence and recrystallization processes in the Provençal Domain during Late Cretaceous-Early Eocene times, where the main extensional and transtensional tectonics affected the European paleomargin of the Alpine Tethys [38].
The calcitization of saddle dolomite and the precipitation of the fracture-filling Fe-poor calcite cement (C1 T/J , Type IV; Figure 14), which cross cut the high-amplitude stylolites, are the last diagenetic event affecting the Triassic series by the incursion of meteoric waters. Furthermore, the precipitation of bladed calcite cement around saddle crystals implies that a new activity led to the uplift of the studied series and could reactivated the Alpine transcurrent shear zone during Cenozoic [39].

Geochemical Evolution of Diagenetic Fluids
A crucial question in any study of carbonate chemical composition is whether a primary marine signal is preserved or not. Carbon and oxygen isotopic compositions of the regional limestones span a considerable range with δ 13 C values between −0.5% and +3.1% and δ 18 O VPDB values between −4.2% and 0.2% . Their isotopic compositions are similar to the values expected for Triassic and Jurassic marine carbonates [23][24][25][26][27].
Type I dolomite and the last event of calcite precipitation (Type IV) have considerably lighter δ 18 O VPDB compared to those expected for the carbonate precipitated in equilibrium with ambient seawater (Figure 13). The δ 18 O and δ 13 C values in both types of dolomite and calcite fit the inverse "J" Lohmann curve [40] pointing to a diagenetic alteration near surface conditions. The initial nucleation of dolomite in the muddy facies was therefore driven by dolomite supersaturation caused by the mixing of fresh and marine groundwater [40][41][42]. Mixed groundwater filled the pore spaces in the oolite facies and led to undersaturation of calcite, and the dissolution and precipitation of dolomite in the case of Type I dolomite. The Triassic and Jurassic rocks developed on a shelf in which the water depth was extremely shallow, probably a few meters deep, as suggested by the oolites and algae. In such a shallow depositional setting these carbonates were probably altered and dolomitized by mixed water, as no evaporitic minerals were observed in this study. In the Triassic and Jurassic sections, the δ 18 O VPDB shifts toward a lower value in comparison with the original depositional facies, are quite abrupt and vary around their respective average (−6.7% and −6.4% ) in the range of the "meteoric" [3]. The strong deviation in δ 13 C VPDB values from this line is related to the rock-water interaction in the Type I Triassic samples. Very weak variations or near constant δ 13 C VPDB values in the Type I Jurassic samples suggest decreased rock-water interactions in a distal setting from the meteoric recharge with a consequent buffering by the rocks.
Types II and III have depleted δ 18 O values (up to −10.4% for Type III in the Triassic and up to −11.6% for Type III in the Jurassic) with more or less constant δ 13 C VPDB values in the Jurassic (Types II and III) and a wider range of δ 13 C VPDB values in Type III in the Triassic. The δ 13 C values are similar to the ones of the early dolomitic facies suggesting that the saddle dolomitizing fluid was mainly buffered by interactions with early dolomite rocks (Type I). The same buffered system has been inferred from our trace element analyses and showed that positive and gradual variation in Fe and Mn contents observed between Types II and III saddle dolomites due to water-rock interactions ( Figure 13).
The significant shift to lower δ 18 O VPDB values from Type II to Type III in both sections is probably due to the interaction of the dolomitized rocks (Type I) with hydrothermal fluids (e.g., [12,43]). Similar carbon and oxygen isotopic values (−9.0% to −7.3% and +1.8% to +2.4% , respectively) were reported by Nader et al. [42] in the Jurassic rocks at Lebanon as the result of hydrothermal hot fluids under deep conditions. The tension-gashed stylolitization phase during the saddle emplacement between Type II and Type III saddle dolomites suggests that the chemical compaction or burial was coeval with the late phase of hot fluids (Type III dolomite). This phase is associated the most widespread depleted δ 18 O VPDB values ( Figure 14). Triassic Type II facies show some similarities and overlapping δ 18 O VPDB values, such as those of Type I, suggesting that saddle precipitation started in the same burial conditions of the latter, but after early stylolitization. These complex scenarios of paragenetic models show that Types II and III correspond to complex polyphased dolomitizing fluids and Type III saddle dolomite was precipitated under deep burial-mesogenetic diagenesis. Their geometries (irregular, strata-discordant, zebras, even faults on the field) may suggest that the faults constituted a hydrothermal fluid conduit operating during two major steps with increasingly 18 O VPDB depleted values in the saddles as evidenced by two different isotopic domains in both sections. Further evidence for deep burial setting and mesogenesis includes the dissolution of saddle dolomite along stylolites (e.g., [44]) and sutured intergranular contacts between ooids.
Type IV Fe-poor calcite records a dedolomitization and cementation by similar calcite due to the incursion of meteoric waters [40,45]. The wide variations of the δ 18 O VPDB and δ 13 C VPDB values reveals an inverse "J" Lohmann curve [40,45], at least in the Jurassic and probably in the Triassic, supporting a telogenetic meteoric origin. The δ 18 O VPDB and δ 13 C VPDB values from calcite samples overlap with the values of the HT dolomite; a similar situation has been reported from NW Spain by Shah et al. [46] and Nader et al. [42] in Lebanon on hydrothermal-hosted Jurassic rocks. These authors concluded that the wide range of depleted isotopic oxygen values with a slight shift in the isotopic carbon values of pristine facies indicate different temperatures of the dolomitizing fluids, while the extra-shifted carbon isotope values and highly depleted isotopic oxygen values in the calcite facies indicate surface-derived fluids. The abrupt Fe and Mn drops, Sr rise and weak Mn variations only reported in the Triassic samples that could be related to the SD3 overgrowth by C1 T suggest a significant lack in carbonate-buffering system. In addition, the saddle dolomitic contribution to calcite oxygen contents is very low compared to the contribution of the meteoric fluids [46] due to high δ 18 O water /δ 18 O rock ratios. These arguments support that the calcite cement marks the end-member of the diagenetic sequence as a result of a late uplift of the carbonate platform, which initiated this late telogenetic phase ( Figure 14).

Impact of Basement-Fluid Interactions, Continental Riverine and vceanic Fluxes on Strontium Isotope Signatures
Multiple parameters controlling the radiogenic strontium isotope in carbonate rocks, including: (i) the isotopic composition of seawater from which they precipitated and terrigenous (i.e., siliciclastics) sedimentary rocks; (ii) the 87 Sr/ 86 Sr ratio is mainly related to the nature of the parent rocks and their continental weathering rates (riverine) and oceanic fluxes of nonradiogenic strontium. However, diagenesis will be the main factor controlling the significant variation of the 87 Sr/ 86 Sr ratio in marine carbonates as it will be affected by hydrothermal or cold fluids that control the strontium ratio. 87 Sr/ 86 Sr in dolomites (Types I, II, III) vary from 0.707735 and 0.712023 in the Triassic and from 0.708680 up to 0.711203 in the Jurassic ( Figure 15). These ratios are associated to the early and late stylolites, prior to the latest deformation during the uplift ( Figure 14). The 87 Sr/ 86 Sr ratios are higher than those of ambient Triassic and Jurassic seawater [26,27]. A few of these ratios in the Triassic dolomites are lower than those of the marine carbonate obtained from host limestone samples (0.708766; Figure 15), while in the Jurassic dolomites the ratios are higher than those in the marine carbonate samples (0.707862; Figure 15). When comparing our radiogenic values with the ones of the Triassic 87 Sr/ 86 Sr signals of the Lombardic Alps, the strontium values of the Lombardic Alps increased from 0.70805 in the Anisian to 0.708148 in the Ladinian as a result of the various rates of weathering and erosion of young volcanic rocks and old granitic rocks [47], are lower than our values in Provençal Domain (0.708766; Figure 15). Thus, it is more likely that an increase in the riverine flux was the dominant factor causing the 87 Sr/ 86 Sr seawater signature rise during the Triassic times in Provençal Domain (cf. [26]), as well as the involved Triaasic carbonates by meteoric waters.  [30]). The high Sr isotopic ratios of saddle dolomite suggest that the hydrothermal fluids interacted with the crystalline basement and/or siliciclastic deposits [12,30].
The radiogenic enrichment of the Jurassic samples could be deviated from open ocean evaporite, however, the dolomitizing fluids are not in agreement with the involvement of the evaporites which are missing here. If a high strontium ratio represents the involvement of evaporites, then it is expected that a high strontium ratio would correlate with the depletion of oxygen values. Mostly, the dolomitizing fluids that have high strontium ratios would have been enriched with heavier oxygen isotope values, confirming the involvement of evaporites (e.g., [48]). However, this scenario is not in agreement with our Jurassic dolomite samples and the regional geology, therefore the interaction of hot dolomitizing fluid mostly interacted with the Argentera crystalline basement before flowing up to the Jurassic carbonate rocks.
In contrast to Jurassic dolomites, some Triassic dolomites have significantly lower 87 Sr/ 86 Sr values with respect to the open marine facies values. The depletion of 87 Sr/ 86 Sr ratios started from the Late Triassic up to the Turonian recording pulses of mid-oceanic hydrothermal fluxes that related to the opening and closure of the North Atlantic Ocean [27]. There were several short-term excursions of radiogenic depletion of 87 Sr/ 86 Sr ratios highlighting these pulses. The authors of [49] reported excursions of very low 87 Sr/ 86 Sr ratios during three short periods (Early Jurassic, Early Cretaceous and Late Cretaceous) that are linked to the rate of hydrothermal activity as a consequence of increased ocean-crust production at the mid-oceanic ridges. Therefore, some of our data in the Triassic could be partly influenced by the short-term excursions of strontium depletion after the deposition of Triassic sediments. In addition, following McArthur's Graph, our data display that Type II saddle dolomites are associated with a jump-down of the 87 Sr/ 86 Sr isotope values. The repetition of the jump-down and jump-up graph of each type of dolomite follows the same trends in the Triassic and Jurassic samples ( Figure 15). The graph suggests that the Triassic and Jurassic sections had the same stratigraphic position (not separated) prior to the precipitation of the calcite phase (Type IV).
The ultimate diagenetic phase (Type IV) records the highest enrichment radiogenic composition (0.712023) from the Triassic and the lowest 87 Sr/ 86 Sr ratio (0.708680) from the Jurassic samples. This phase attributed to the late stage of calcite cementation and was consistent with the very depleted δ 18 O VPDB values and the lowest δ 13 C VPDB values (cf. [50]). The meteoric water has the lowest δ 18 O VPDB values and highest 7 Sr/ 86 Sr ratios (0.709630 to 0.70948, [51]). The increasing recharge from meteoric water could apply to the late stage of Fe-poor calcite precipitation and is probably related to an uplift of the buried Triassic-Jurassic rocks (Figures 11-14). The late calcite cement C1 J in the Jurassic section shifted toward decreasing Sr isotope values, and appears to be more affected by other sources besides meteoric water (in contrast to Triassic calcite value). On the basis of the tectonic graph, C1 T and C1 J are attributed to two different fluids, which were stratigraphically separated during the uplifting of these units (outcrops). Consequently, the lowering of Sr isotope compositions (C1 J ) in this study, especially in the Jurassic section, could be partly linked to the pulses of the seafloor hydrothermal activity that lowered the 87 Sr/ 86 Sr ratios. During the late calcite event, the former dolomitized rocks were strongly brecciated with an in situ accumulation of crushed fragments. This event reveals that an intense tectonic activity was ongoing during the late fluid migration [30,52,53], which could be associated to Late Cretaceous-Eocene times, where general strike-slip fault systems occurred in the European paleomargin [52,53].
The Triassic-Jurassic carbonate seawater in Provençal Domain could partly have been influenced by the input of various rates of weathering and erosion of young volcanic rocks and/or old granitic rocks. These inputs probably persisted after the deposition of Mesozoic succession in Provencal Domain. Finally, fluctuations in the rising and falling of 87 Sr/ 86 Sr seawater signatures have already been reported [26,27], therefore, the values of the 87 Sr/ 86 Sr fluctuations are the important factors and may have influence the diagenesis products beside the hydrothermal fluids and meteoric waters in Provencal Domain ( Figure 15).

Conclusions
The main conclusions derived from this integrated field, petrographic and geochemical study of dolomitization in the Mesozoic carbonate succession, Provençal Domain, Maritime Alps, SE France show that:

1.
The succession is extensively dolomitized with abundant breccia and zebra textures, and complex fracturing paths due to the influence of repeatedly injections of hydrothermal fluids.

2.
Three diagenetic settings are recognized: (i) An eogenetic realm with weak diagenetic imprints. This setting is confirmed by δ 18 O VPDB and δ 13 C VPDB values that fit the inverse "J" Lohmann curve and meteoric line models. (ii) A second mesogenetic setting in two phases, the first one started with the generation of wispy stylolites ("early dolomite"), and various networks of fractures as well as zebra and breccia fabrics. The fabrics are associated with a first generation of medium-to coarse-grained dominantly euhedral saddle dolomites (Type II) formed through the migration of hot dolomitizing fluids. This occured during early Cretaceous times and was related to the extensional and transtensional tectonics of the European paleomargin of the Alpine Tethys. In the second phase (our Type III   Funding: This research received no external funding.