Evaluating new fault-controlled hydrothermal dolomitization models: Insights from the Cambrian Dolomite, Western Canadian Sedimentary Basin

Fault-controlled hydrothermal dolomitization in tectonically complex basins can occur at any depth and from different ﬂuid compositions, including ‘deep-seated’, ‘crustal’ or ‘basinal’ brines. Nevertheless, many studies have failed to identify the actual source of these ﬂuids, resulting in a gap in our knowledge on the likely source of magnesium of hydrothermal dolomitization. With development of new concepts in hydrothermal dolomitization, the study aims in particular to test the hypothesis that dolomitizing ﬂuids were sourced from either seawater, ultramaﬁc carbonation or a mixture between the two by utilizing the Cambrian Mount Whyte Formation as an example. Here, the large-scale dolostone bodies are fabric-destructive with a range of crystal fabrics, including euhedral replacement (RD1) and anhedral replacement (RD2). Since dolomite is cross-cut by low amplitude stylolites, dolomitization is interpreted to have occurred shortly after deposition, at a very shallow depth ( < 1 km). At this time, there would have been sufﬁcient porosity in the mudstones for extensive dolomitization to occur, and the necessary high heat ﬂows and faulting associated with Cambrian rifting to transfer hot brines into the near surface. While the d 18 O water and 87 Sr/ 86 Sr ratios values of RD1 are comparable with Cambrian seawater, RD2 shows higher values in both parameters. Therefore, although aspects of the ﬂuid geochemistry are consistent with dolomitization from seawater, very high ﬂuid temperature and salinity could be suggestive of mixing with another, hydrothermal ﬂuid. The very hot temperature, positive Eu anomaly, enriched metal concentrations, and cogenetic relation with quartz could indicate that hot brines were at least partially sourced from ultramaﬁc rocks, potentially as a result of interaction between the underlying Proterozoic serpentinites and CO 2 -rich ﬂuids. This study highlights that large-scale hydrothermal dolostone bodies can form at shallow burial depths via mixing during ﬂuid pulses, providing a potential explanation for the mass balance problem often associated with their genesis.

Recent studies have shown that the convection of seawater along fault planes and basal clastic aquifers can explain the formation of fault-related dolomite from hydrothermal fluids (Martin-Martin et al., 2015;Hollis et al., 2017;Al-Ramadan et al., 2019). These models require that faults are open to the sea floor, as often occurs within extensional basins Hirani et al., 2018). Other more recent studies have also invoked fluid interaction with mafic and ultramafic rocks (for example, serpentinized peridotites) as the source of magnesium and demonstrated the thermodynamic viability of ultramafic carbonation to increasing the potency of dolomitization (Lavoie et al., 2014;Falk & Kelemen, 2015;Robertson et al., 2019). These two new ideas have provided a different view of potential sources for water and magnesium in the formation of large-scale HTD bodies that might be appropriate for reconstruction of dolomitization processes in the WCSB.
The dolomitized sequences of Cambrian strata in the Canadian Rocky Mountains provide excellent exposure of fault-controlled dolostone bodies which are easily accessed, allowing for detailed, methodical sampling across the dolostone bodies. Prior studies of the Cambrian (Series 2) Cathedral Formation (Yao & Demicco, 1997;Jeary, 2002;Vandeginste et al., 2005;Powell et al., 2006) have shown that dolomitization involved hydrothermal fluids in the vicinity of the Kicking Horse Rim (Fig. 1A) and surrounding area. These studies invoked a range of models for explaining dolomitization, including post-Silurian topographically-driven fluid (Yao & Demicco, 1997), focused fluid flow of hot and saline brines along the fold thrust belt associated with Devono-Carboniferous Antler Orogeny (Vandeginste et al., 2005), Cambrian-aged Mgrich brine seeps sourced from Middle Proterozoic magnesite (Powell et al., 2006) and multiple dolomitization events during the Cambrian period (Series 2 to Furongian) when early dolomitization was overprinted by hydrothermal dolomite, producing several hydrobrecciated zones (Jeary, 2002). Of these models, the suggestion by Powell et al. (2006) that Middle Proterozoic magnesite sourced dolomitizing brines have received relatively little attention; this is evaluated here with reference to new models of ultramafic carbonation to drive dolomitization.
This study focuses upon the Mount Whyte Formation, which immediately underlies the Cathedral Formation, and which has been remarkably understudied. It was described principally from the area around Whirlpool Point (Fig. 1A) where it is separated from the Cathedral Formation by a shale unit. The Mount Whyte Formation differs in character from the Cathedral Formation because it does not exhibit well-developed zones of brecciation or zebra dolomite textures. It therefore offers an ideal opportunity to evaluate dolomitization processes within the WCSB, since it appears to have a simpler history of dolomitization than the overlying Cathedral Formation. In particular, the study aims to identify: (i) the potential source of magnesium for dolomitization; (ii) the mechanisms that drive the circulation of dolomitizing fluids and timing of dolomitization; and in particular (iii) to assess the potential of ultramaficderived carbonation as a source of magnesium for extensive HTD bodies in the Cambrian, and potentially other Palaeozoic, dolomitized successions in the WCSB.

GEOLOGICAL BACKGROUND
The WCSB is a large sedimentary basin, ranging from south-western Manitoba to south-west of Northwest Territories, with a complex tectonic history (Creaney & Allan, 1990;Gabrielse et al., 1991). It is comprised of a sedimentary wedge up to 6 km thick. The sedimentary succession is predominantly composed of Palaeozoic to Mesozoic carbonate rocks, after which it evolved into a clastic-dominated basin, from the Cretaceous until Recent (Creaney & Allan, 1990;Wright et al., 1994). Changes in sedimentation patterns through time were closely associated with the tectonic evolution of the basin, from a rift basin to a passive margin into a foreland basin (Wright et al., 1994). Across the WCSB, the sedimentary sequence overlies a range of Precambrian metamorphic terranes, including a Paleoproterozoic subduction complex, as revealed from lithoprobe transects (Eaton & Cassidy, 1996;Fig. 1B and C). The Terreneuvian succession was deposited in an extensional basin, formed during tectonic reactivation of the Neoproterozoic passive margin (Bond & Kominz, 1984), with preservation of the syn-rift conglomeratic sandstone of the Gog Group ( Fig. 2) (Kubli & Simony, 1994). Subsequently, regional subsidence led to several marine transgression-regression cycles (Grand Cycles) during the late-syn rift to post-rift stage, in the Series 2 through to the Furongian period (Slind et al., 1994). This resulted in the growth of several carbonate platforms, including the Mount Whyte, Cathedral and Eldon formations (Aitken, 1997;Powell et al., 2006;Norford, 2012). The Mount Whyte Formation underlies the Cathedral Formation and is considered to be at the base of the second 'Grand' cycle in this region (e.g. Norford, 2012). The Mount Whyte is characterized by interbedded limestone-dolostone and shale in the Front Ranges, where the thickness can reach up to 176 m (Aitken, 1989(Aitken, , 1997Collom et al., 2009;Fig. 2).
The Mount Whyte Formation is well-exposed at Whirlpool Point (Fig. 1A), where it is bounded by two major north-west/south-east thrust faults; Pipestone Pass to the east and Bourgeau in the west (Fig. 1A). Previous studies have included the limestones and dolostones of the Mount Whyte in this study location into the Cambrian Peyto Formation (Terreneuvian,  Davies & Smith, 2006). (B) and (C). Lithoprobe transect ('3' to '5') near the Snowbird Tectonic Zone where it shows the appearance of relict of Proterozoic subduction zone and complex underneath WCSB based on the presence of LVZ (low velocity zones) and dip of the reflection (Eaton & Cassidy, 1996;Van der Velden & Cook, 2005). Stage 2; Aitken, 1997;Jeary, 2002). However, Aitken (1997) reported the thickness of the Peyto Formation to be only 7.6 m which is significantly thinner than measured at Whirlpool Point (up to 80 m thick; Fig. 3A and B). Because Aitken (1997) does not report explicit measurements of the remaining stratigraphic section in this location, it is possible that the rest of the section (ca 70 m) belongs to the Mount Whyte Formation while the lower part belongs to the Peyto Formation.
The reference section of the Mount Whyte Formation is located in the south-west flank of Mount Weed, Mistaya Canyon which is ca 50 km south-west from the study location (Aitken, 1997). Here, the Mount Whyte Formation (126 m thick) consists of two members: (i) greenish siltstones of the Weed Member; and (ii) limestone-dominated of the Chephren Member (Aitken, 1997). At Whirlpool Point, the Terreneuvian Gog Group sandstone is partlyexposed (ca 10 m thick), comprising mediumgrained cross-bedded sandstone to conglomeratic sandstone overlain by greenish siltstones, and is possibly therefore part of the Peyto Formation (see description in Aitken, 1997). Several studies have indicated that the stratigraphic relationship between the Mount Whyte and Peyto formation is unconformable based on the gap in trilobite zones found in both formations (Tremblay, 1996;Aitken, 1997). The Mount Whyte Formation consists of shale and carbonate sediments, comprising stromatolites and thick bedded partially dolomitized mudstones with oncoidal-ooidal grainstone in its upper part ( Fig. 3A to D) consistent with deposition in the platform interior (Fig. 3A). It is conformably overlain by the Cathedral Formation. Distinct trilobite faunas, including the Plagiura-Albertalla zone in the Main and Front Ranges indicated that the Mount Whyte and Cathedral formations were deposited during the Series 2 (516.0 to 513.0 Ma; Deiss, 1941;Collom et al., 2009;Fig. 2), which is taken to be the assumed age of the Mount Whyte Formation in this study.

ANALYTICAL METHODOLOGY
A total of 104 samples were collected from the stratigraphic section and various dolostone bodies in the Cambrian Mount Whyte Formation at the Whirlpool Point locality. Thin section analysis was conducted on 74 blue-epoxy resin dyed samples, stained with Alizarin Red and potassium ferrocyanide, using standard petrographic  Collom et al., 2009). This study included the Gog group sandstone as part of the syn-rift succession while the Mount Whyte Formation belongs to the post-rift sequence. Note the presence of platform-scale dolostone bodies in the three carbonate formations, Mount Whyte, Cathedral and Eldon formations (Aitken, 1978;Yao & Demicco, 1997;Collom et al., 2009). This study focuses on the formation of pervasive dolostone bodies in the Mount Whyte Formation.
techniques (Dickson, 1966) to qualitatively discriminate between calcite and dolomite. Cathodoluminescence analysis was performed using a [Cambridge Image Technology Limited (CITL), Cambridge, UK] Mk5 Cold cathode stage at the University of Manchester to determine the cement generations and origins. A cold cathode luminescope was used with a beam voltage of ca 17 kV and a current of 600 lA.
Scanning electron microscopy (SEM) was performed on selected samples (n = 16) by using FEI Quanta FEG 450 (Thermo Fisher Scientific, Waltham, MA, USA) in the Williamson Research Centre, University of Manchester to investigate different mineral microfabrics. The bulk mineralogy of all samples was determined by X-ray Diffraction (XRD) analysis using Bruker D8Advance Diffractometer (Bruker Corporation, Billerica, MA, USA). High-resolution quantitative major, minor and trace elemental analysis was conducted by using Cameca SX 100 Electron Microprobe (EPMA; Cameca, Gennevilliers, France) at the University of Manchester. The analysis was conducted using wavelength dispersive spectrometry (WDS) for Fe, Mn, Ca and Mg, and energy dispersive spectrometry (EDS) for other elements (for example, Ti and Cr) with an accelerating voltage of 15 kV and a beam current of 10 to 20 nA. Replicate measurements of different dolomite crystals show that the analytical precision (1r) was estimated at ≤5%.
Nine double-polished, 50 to 100 lm thick sections of selected dolomite samples were selected for fluid inclusion studies. Standard, transmitted-light petrographic analysis was conducted to map the presence of fluid inclusions within the dolomite before detailed examination using the fluid inclusion stage. A heating-freezing stage (LINKAM MDSG 600 system; Linkam Scientific Instruments, Tadworth, UK) attached to a Nikon microscope (Nikon, Tokyo, Japan) at the University of Manchester and a LINKAM THMS-600 stage (À196 to 600°C) coupled to an Olympus BX53 microscope (Olympus, Tokyo, Japan) at the University of Alberta were utilized to perform fluid inclusion analysis, using a standard method proposed by Goldstein & Reynolds (1994).
For strontium isotope ( 87 Sr/ 86 Sr) analysis, 100 to 150 mg of powder of ten whole-rock samples were dissolved a mixture of 1 ml HNO 3 and 1 ml HF in a crucible at 190°C for 48 hours. Strontium produced by reaction between carbonate powder and acid was extracted by using conventional ion exchange procedures using resin cation Bio-Rad AG50 X12 (Baadsgaard, 1987). The resulting Sr isotope ratios were measured using a Triton Plus thermal ionization mass spectrometer (Thermo Fisher Scientific) at the PetroChina Hangzhou Research Institute of Geology, Hangzhou, China. Analytical precision of individual runs is at 0.00005 (2r). All 87 Sr/ 86 Sr values were normalized and reported to a NBS-987 standard (0.710253).
For elemental, d 13 C and d 18 O, and clumped isotope analyses, thin and thick section counterparts were micro-drilled under a binocular microscope to extract different dolomite fabrics and limestone matrix and cement, in order to avoid mixing of different diagenetic phases and verify the homogeneity of the samples. Trace element (Fe, Mn, Ba and Sr) and rare earth element (REE+Y) concentrations in the samples were analysed using Perkin-Elmer Optima 5300 dual view Inductively Coupled Plasma Atomic Emission Spectroscopy (ICP-AES; Perkin-Elmer, Waltham, MA, USA) and Agilent 7700x Inductively Coupled Plasma Mass Spectrometry (ICP-MS), respectively. A subset of samples used for stable isotope analyses were utilized in both analyses [trace element (n = 15); REEs (n = 10)]. The samples were reacted with 6 M hydrochloric acid (HCl) and 10 ml de-ionized water were added to ensure that total dissolved solids will be less than 0.1%. The liquid samples were then acidified by typically 2% HNO 3 and filtered to remove particles > 0.45 lm. Detection limits for both analyses are as low as 0.01 ppb in solution under the usual operating conditions. The REE concentrations were normalized to both Post-Archean Australian Shales (PAAS) (McLennan, 1989) and chondrite (Anders & Grevesse, 1989), and their anomalies (Ce, Pr, La and Eu) were determined following the methods proposed by Bau & Dulski (1996) and Webb & Kamber (2000).
A total of 42 samples were analysed for their d 13 C and d 18 O. The powder samples (5 to 10 mg) (n = 20) were reacted with anhydrous phosphoric acid at 25°C and ca 16 h for calcite and 50°C and ca 48 h for dolomite (McCrea, 1950) in a carbonate preparation line connected to a VG SIRA 12 gas-source mass spectrometer (Elementar GmbH, Langenselbold, Germany) at the University of Liverpool, UK. An additional 22 samples were analysed at the Scottish Environmental University Research Centre (SUERC), Glasgow using A VG OPTIMA mass spectrometer (Isoprime Limited, Cheadle, UK). A carbonate-phosphoric acid correction factor was applied on the oxygen isotope ratios for both calcite and dolomite by using fractionation factors (i) of 1.01025 (Kim & O'Neil, 1997) and 1.01066 (Rosenbaum & Sheppard, 1986), respectively. All values are reported as delta values with respect to the Vienna PeeDee Belemnite (VPDB) and standardized to NBS-19 and Carrara marble. Average analytical precision and reproducibility for both calcite and dolomite was checked by replicate analysis and it is better than AE0.1&.
The measurement of clumped isotope on both limestone (n = 5) and dolostone samples (n = 17) was conducted at the Stable Isotope Laboratory, University of Miami by using dual inlet Thermo Scientific 253 ultra-high-resolution isotope ratio mass spectrometers. The CO 2 extraction procedures followed the description of Murray et al. (2016) and Swart et al. (2016) for the D 47 measurements. For each analysis, 10 to 15 mg of sample was reacted in a common acid bath (103% phosphoric acid at 90°C) for 30 and 45 min on calcite and dolomite, respectively. The yield CO 2 was then frozen in a sealed vessel and transferred into a mass spectrometer where it was measured against a working cryogenically cleaned reference gas. Replicate analyses (two to four times) were performed to correct the error/drift over time and ensure that the results are consistent. The results were then converted to the absolute reference frame (ARF) by following a Carbon Dioxide Equilibrated Scale (CDES) calibration scheme described by Dennis et al. (2011). The Δ 47 value was adjusted and converted to a temperature value without applying acid fractionation factor by using the equation of Staudigel et al. (2018) modified for dissolution at 90°C. In addition, during the measurement of clumped isotopes, d 13 C and d 18 O values were measured by calculating the masses 45/44 and 46/44. The correction method proposed by Craig (1957) was followed to correct for the isobaric interferences. As the reaction was conducted at 90°C, the d 18 O value of the dolomite was constantly corrected by -0.8& to account for the differential fractionation of CO 2 yielded from dolomite relative to calcite (Sharma & Clayton, 1965;Land, 1980;Swart et al., 2016). The 18 O/ 16 O fractionations of calcite-water and dolomite-water published by Kim & O'Neil (1997) and Horita (2014), respectively, were utilized to calculate the oxygen isotopic compositions of parent (i.e. diagenetic fluids (d 18 Ow).

Lithofacies
The Gog Group at Whirlpool Point is characterized by the presence of medium to fine grained cross-bedded sandstone. The sandstones are typically cemented by dolomite with the occasional presence of dolomite cemented fractures (Stacey et al., 2017). The Mount Whyte Formation directly overlies the Gog Group and is principally composed of five different lithologies that make up a total succession of ca 78 m thickness with both lateral and vertical diagenetic contacts between dolostone (brown) and limestone (grey) ( Fig. 3A and B).
1 Grey-greenish shale to siltstone facies occurs only in the lower part of this section and is either well-bedded (centimetre-scale) or structureless (Fig. 3C).
2 Microbial laminites that overlie the siltstone facies and are characterized by ca 5 m thick light grey coloured centimetre-scale crinkly laminated mudstone (Fig. 3D).
3 Thinly bedded, burrowed bioclastic wackestone which is mud-supported. The occurrence of dolostone-limestone transitions and preferentially dolomitized burrows within this facies suggests that it was the parent rock of the dolomitized unit (Fig. 4A).
4 Oncoidal packstone to grainstone facies containing oncoid grains (0.2 mm to 2.0 cm in diameter) which remain undolomitized. This unit dominates the upper part of this Mount Whyte section and forms the upper boundary for the dolomitized unit (Fig. 4B).
5 Well-bedded medium to coarse crystalline dolostone that constitutes more than half of the total thickness of the Mount Whyte Formation in the study area. The dolostone is light to dark orange-brown colour with centimetre to metrescale bedding preserved (Fig. 4C). Thinly bedded shale is often observed intercalated with this dolostone unit. Zebra dolomite fabrics are occasionally associated with this facies but are generally rare (Fig. 4D). North-west/south-east to west-east trending fracture sets cemented by either calcite or dolomite spar are present throughout the outcrop and more pronounced in the dolostone unit (Fig. 4E). In addition, bedding parallel vuggy porosity and low amplitude stylolites can also be recognized in the outcrop ( Fig. 4F and G).

Petrography
Three different dolomite fabrics were identified based on their petrographic and cathodoluminescence (CL) characteristics. Visual estimation of their relative volume with respect to other fabrics is also described here.
Replacive dolomite (RD1) RD1 is a fabric destructive dolomite and characterized by a fine-sized dolomite crystal with relatively unimodal crystal size distribution (50 to 250 lm) ( Fig. 5A and B). Under plane polarized light, phase RD1 exhibits planar-e to planar-s textures (sensu Sibley & Gregg, 1987) with cloudy, inclusion-rich cores and clear rimmed crystals (Fig. 5A). This phase has distinct CL characteristics, in comparison with other dolomite phases, exhibiting either a dark purple luminescence with no to thin orange bright zones or dull red core or thick, bright orange luminescent zones in the outer rim (Fig. 5C). The presence of RD1 is mostly observed close to dolostone-limestone transitions and occasionally observed as isolated rhombs in the limestone or accumulating along pressure solution seams. Intercrystalline porosity is often found associated with the RD1. This fabric constitutes approximately 25% of the total dolomite volume in the Mount Whyte Formation.

Replacive dolomite (RD2)
This replacive dolomite phase is also strongly fabric destructive and shows a polymodal crystal size distribution (100 to 1000 lm) ( Fig. 5D and E). The dolomite crystals exhibit a non-planar texture and closely packed mosaic of subhedral to anhedral crystals with occasional saddle dolomite textures (sensu Sibley & Gregg, 1987;Fig. 5E). While this fabric often exhibits coarse crystal sizes (up to 1000 lm), RD2 only rarely shows curved crystal faces and sweeping extinction. RD2 has a characteristic dull red, unzoned CL with some mottled brighter orange-yellow spots in the central part (Fig. 5F). Only very slight variations in CL colour or intensity were

Saddle dolomite
The saddle dolomite (SD) fabric exhibits undulose extinction, curved crystal cleavage and faces, and megacrystal size (can be >2 mm) which is a typical of saddle-dolomite (sensu Radke & Mathis, 1980;Fig. 5G and H). This dolomite fabric is recognized filling interparticle and fracture pore space and is typically related to zebra textures in outcrop. Under CL, SD is compositionally zoned and has bright to dull red luminescent zones and thin to thick bright orange luminescence (Fig. 5I). SD is only present in subordinate amounts, up to 10% of the total dolomite volume.

Host limestone
In the Mount Whyte Formation, the host limestone can be divided into two groupsmud- dominated and grain-dominated facies. Petrographically, the mud-dominated facies are predominantly composed of micrite and a minor amount of microspar (Fig. 6A); in outcrop it is pervasively bioturbated, accentuated by dolomitization of the burrows. Micrite shows bright orange luminescence under CL (Fig. 6B). In contrast, the grain-dominated facies are mainly composed of oncoids and peloids with pervasive marine cements occluding most of the pore space ( Fig. 6C and D). The CL characteristic of grain-dominated facies is dull to non-luminescence (Fig. 6E). In the Mount Whyte Formation, the bioturbated mud-dominated facies is preferentially dolomitized whilst the graindominated, oncoidal grainstone facies remains as limestone.
Calcite cement Calcite cement occurs as two different phases: (i) a pore filling phase as observed in the grainstone facies (Fig. 6C); and (ii) a fracture filling phase ( Fig. 6A to D). These two phases typically display a similar blocky texture, with crystal sizes ranging from 20 to 500 lm. However, their CL characteristics are different; the pore filling phase is dull to non-luminescence while the fracture-filling phase is either bright orange or dull to non-luminescence with no zonation (Fig. 6B to E).

Authigenic quartz
This phase is difficult to identify under transmitted light; however, SEM and XRD analyses  reveal the presence of a small percentage of microquartz (<3%) in almost all limestone matrix and dolostone samples (Fig. 6F). Authigenic, hexagonal quartz crystals are observed in the samples using SEM and often occludes macro-porosity and microporosity (Fig. 6F).

Stylolite
Low-amplitude stylolites are the dominant pressure solution feature in the Mount Whyte Formation (Fig. 6G). In some cases, these low amplitude stylolites have amalgamated to create an anastomosing pressure seam morphology (Fig. 6H). They are often characterized by the accumulation of less soluble material (for example, dolomite rhombs, organic material) along the surfaces. The dolomite crystals are mainly characterized by microspar-sized crystals (10 to 30 lm) with planar-e to planar-s texture (similar to RD1), embedded within a black matrix (Fig. 6H). Higher amplitude microstylolites are recognized in a few thin sections, characterized by the presence of only minor residual material (Fig. 6I). These different pressure solution morphologies post-date all dolomite fabrics ( Fig. 6G and I).

Trace elements
Minor and trace elemental concentrations (Fe, Mn, Sr, Ba, Cr and Ti) were analysed for the two replacive dolomite fabrics (RD1 and RD2) and host limestone (Table 1).

Iron and manganese
Both RD1 and RD2 display significant enrichment of iron (Fe) and manganese (Mn) concentrations when compared to the limestone matrix (Fig. 7A). The mean of Fe and Mn concentrations in the limestone matrix are 2997 ppm and 136 ppm, respectively. In contrast, the mean of Fe and Mn concentrations of RD1 dolomite are ca 8225 ppm and ca 796 ppm, respectively (Table 1). Similarly, the RD2 fabric also display high mean values of both Fe and Mn concentrations 9053 ppm and 843 ppm, respectively, which were comparable with the RD1.

Strontium and barium
The mean of strontium (Sr) values measured from two different dolomite fabrics (RD1 and RD2) in the Mount Whyte Formation are considerably lower than the host limestone samples, between 28 ppm and 31 ppm, respectively, while the limestone mean value is 808 ppm (Table 1; Fig. 7B). In contrast, the mean of barium (Ba) concentration in the host limestone (246 ppm) is comparable to RD1-RD2, 194 ppm and 226 ppm, respectively (Table 1).

Titanium and chromium
Comparable mean values of Titanium (Ti) concentration in both RD1 and RD2 dolomites are observed in the Mount Whyte Formation, 540 ppm and 482 ppm, respectively (Table 1). This is considerably higher than the limestone, where Ti concentrations in micrites and cements are below the detection limit (<1 ppm). Similarly, no values can be obtained for the chromium (Cr) concentration in the host limestone samples whereas the RD1 and RD2 dolomites show mean values between 50 and 62 ppm (Table 1).

Rare earth elements
Distribution of rare earth elements (REEs)+Y in the two dolomite fabrics and host limestone are presented in Table 2 and Fig. 8 and described below.

Replacive dolomite (RD1)
The mean of total content of REEs (ΣREE) for this dolomite phase is 6.8 ppm. The mean of ΣLREE (6.6 ppm) displays an enrichment in comparison with the ΣHREE (0.25 ppm). The mean ratio between Y/Ho is 21.8 (Table 2). Positive Eu anomalies (Eu/Eu* = [Eu/(0.67Sm + 0.33 Tb] SN ) range from 1.8 to 3.2 and are clearly illustrated in both PAAS and chondrite normalized REE profiles ( Fig. 8A and B). There is no correlation between Ba and Eu suggesting the absence of Ba interference during the REE analysis (Fig. 8C). In addition, the calculation of both Ce anomalies (Ce/Ce* = [Ce/(0.5 La + 0.5 Pr] SN ) and Pr anomalies (Pr/Pr* = [Pr/ (0.5 Ce + 0.5 Nd] SN ) display values close to 1 (Fig. 8D).

Replacive dolomite (RD2)
The RD2 phase REE profile displays a mean of ΣREE (14.7 ppm) that is almost three times higher than the RD1. There is a similar enrichment pattern of ΣLREE (mean: 14.1 ppm) with respect to the ΣHREE (mean: 0.64 ppm), and a comparable trend between RD1 and RD2 with positive anomalies of Eu (Eu/Eu* > 1) ( Fig. 8A and B) and non-anomalous Ce and Pr (Ce/Ce* and Pr/Pr* ca 1) (Fig. 8D). The mean of Y/Ho ratio of this dolomite fabric is 26.9 (Table 2).

Host limestone
The host limestone shows a slightly lower ΣREE mean value (5.9 ppm) than the RD1 and RD2, which have mean values of ΣLREE and ΣHREE of 5.6 ppm and 0.33 ppm, respectively. The mean value of Y/Ho ratio of the host limestone is 27.8 (Table 2). Normal (i.e. nonanomalous) Eu (Eu/Eu* ca 1), Ce, Pr anomalies (Ce/Ce* and Pr/Pr* ca 1) are also observed in the PAAS-normalized profile of the host limestone samples, they are lower than the two different fabrics (RD1 and RD2) (Fig. 8A to D).

Fluid inclusion microthermometry
Fluid inclusion microthermometry was conducted in dolomite fabrics RD1, RD2 and SD (Table 3). The primary fluid inclusions (i.e. within growth zones or parallel to the crystal facet) in these samples are usually very small in size (4 to 8 lm) (Fig. 10A to C), but a small number of unusually large fluid inclusions (up to 20 lm) can also be observed (Fig. 10B). The majority of the fluid inclusions observed in this study are two-phase with 10 to 20% vapour-liquid ratio. Statistically, the homogenisation temperature (Th) in the dolomite samples can be grouped into (Fig. 10): (i) RD1 (mean: 124°C); (ii) RD2 in the middle part (mean: 147°C); (iii) RD2 in the core of the body (mean: 181°C) and (iv) saddle dolomite cement associated with either fracture or zebra dolomite fabrics (mean: 181°C). The Th of the smaller fluid inclusions cannot be obtained due to the equipment limitation. Analyses of the melting temperature (Tm) of different samples were also attempted, however none of the large fluid inclusions (10 to 20 lm) were frozen during the analysis (up to À180°C).

Strontium isotope analysis
Strontium isotope analysis was conducted on eight dolomite samples and two whole rock limestone samples. Overall, the values show two different populations with the dolomite samples   obtained from clumped isotopes are higher than the fluid inclusion (Table 3 and

Diagenetic environments and paragenesis
Outcrop observations and petrographic studies are used to define the relative timing of diagenetic features (Figs 5 and 6). The earliest diagenetic phases include micritization of skeletal and non-skeletal grains and precipitation of blocky calcite cementation. This blocky calcite cement preserves an open, uncompacted texture, filling the primary interparticle pore spaces between oncoidal grains ( Fig. 6C and D). It is therefore interpreted as marine, with the blocky texture reflecting precipitation from low-magnesium calcite seawater during the Cambrian (Horita et al., 2002). High concentrations of Sr and low concentrations of Mn and Fe alongside seawater-like d 13 O calcite values (mean: À0.7& VPDB; Figs 7 and 9) support this interpretation. Near-surface, marine diagenesis may also be inferred from the presence of burrow-selective dolomitization in the host limestone (Fig. 4A), suggestive of microbially-mediated dolomitization during bioturbation (e.g. Gingras et al., 2004;Corlett & Jones, 2012). However, the original geochemical signatures of the burrow-selective dolomitization appear to have been altered during subsequent high-temperature dolomitization events, evidenced by more negative d 18 O dolomite , elevated Fe-Mn concentrations, d 18 O water and temperature than would be expected from near-surface, microbially mediated dolomitization (Table 3). Subsequent burial diagenetic events can be divided into two stages; (i) shallow burial (<1 km) and (ii) deep burial (1 to 6 km). Mechanical compaction can only be observed in some parts of the host limestone facies and is notably absent in the grain-supported facies due to the presence of marine cements which appear to have inhibited compaction. The cross-cutting relationship between low amplitude stylolites and dolomite suggests that dolomitization predated stylolitization (Fig. 6G to I).  and Beaudoin et al. (2016) suggested that formation of stylolites in mud-dominated carbonate facies can begin at between 500 m to 1 km, implying that dolomitization occurred at depths of less than 1 km, although the possibility of slightly greater depths for the onset of stylolitization cannot be ruled out. Dolomitization produced fabric-destructive dolomites (RD1, RD2 and SD), which are typically associated with formation above the critical roughening temperature of 55°C (Sibley & Gregg, 1987 Swart, 2015;Veizer & Prokoph, 2015;. In addition, both fluid inclusion (Tm) and Δ 47 temperatures are very high (up to 235°C). Water/rock ratios (W/R), calculated following the method proposed by Banner & Hanson (1990) for the different dolomite fabrics (using the original composition of d 18 O rock , d 18 O fluid and crystallization temperature) were moderately high, particularly for RD1 (0.3 to >10 W/R, depending on the Cambrian seawater values) ( Fig. 11B and C).

Burial history of Mount Whyte Formation
Burial history reconstruction indicates continuous burial of Cambrian strata, depending on their location, that can reach up to approximately 4 km, with increasing burial depths towards the  (Veizer & Prokoph, 2015;Henkes et al., 2018; and the overlying Cambrian Cathedral dolomites at Whirlpool Point and Kicking Horse locations (Yao & Demicco, 1997;Jeary, 2002;Vandeginste et al., 2005). Note that the dolomite d 18 O values of the Mount Whyte Formation are more negative than the host limestone counterparts and Cathedral dolomites at Whirlpool Point.
Mountain Ranges during the onset of the Laramide Orogeny in the Jurassic-Cretaceous (Wright et al., 1994;Jeary, 2002). Fracturing and cementation by bright orange luminescence calcite and creation of high amplitude micro-stylolites, which all cross-cut the dolostone fabrics, could have occurred at any point from the Devonian onward, when burial depths exceed 1 km.
To date, the maximum burial depth of the Mount Whyte Formation in this part of the basin has still not been confidently determined. Although d 13 C calcite values and trace elemental concentrations of micrite and marine calcite cements are within the expected range of Cambrian marine values, some of the geochemical signals of (for example, light d 18 O calcite , high Δ 47 -derived temperature and the absence of a negative Ce anomaly) are inconsistent with seawater precipitation. Water/rock ratios are very low, calculated to range from 0.05 to 0.3, depending on the original values of Cambrian seawater ( Fig. 11B and C). Banner & Hanson (1990) suggested that modification of REE would require high water/rock ratios, but Wallace et al. (2017) suggested that Cambrian seawater did not have a Ce anomaly since seawater was not as oxygenated as it is today. It is possible therefore that d 18 O calcite and Δ 47 signatures were altered during burial recrystallization, which usually occurs within a semi-closed (rock-buffered) system and could occur during deep burial (Brand & Veizer, 1980;Czerniakowski et al., 1984;Banner & Hanson, 1990). In other words, the recorded Δ 47 ratios may reflect solid-state reordering during geothermal heating or rockbuffered crystallization during burial which can be used to approximate the burial depth of the limestone (Ghosh et al., 2006;Passey & Henkes, 2012;Winkelstern & Lohmann, 2016;Lawson et al., 2018).
Cambrian limestone formed under ambient seawater conditions should display high Δ 47 values (0.58 to 0.57; 30-35°C using Staudigel et al. (2018) temperature conversion), but instead they show much lower values (mean: 0.365; mean: 164°C using similar temperature conversion). Winkelstern & Lohmann (2016) suggested that a threshold of Δ 47 calcite reordering under rockbuffered diagenesis system in micrite-dominated rocks could occur at 70°C [<2 km if using normal geothermal gradient; 25°C/km (Weides & Majorowicz, 2014) and 35°C surface temperature (Henkes et al., 2018)]. The Δ 47 temperatures of the Mount Whyte micritic limestones range between 146 to 185°C which may suggest that either: (i) the rock has been buried to at least 5 to 6 km, i.e. deeper than previously thought, assuming a normal geothermal gradient; or (ii) a maximum burial depth of 4 km, as suggested by Wright et al. (1994). but with a higher geothermal gradient. The latter is feasible because the Mount Whyte Formation was deposited during rifting, when geothermal gradients would have been higher (ca 35 to 40°C/km; Bond et al., 1985). Although there is likely to have been a decrease of geothermal gradient during thermal sag subsidence (potentially to 20°C/km; Bertotti et al., 1999), the presence of thick shales above and surrounding the Mount Whyte Formation might have acted as an insulator that retained the high heat flow within the rock sequence (Cercone et al., 1996;Nunn & Lin, 2002).

Temperature of dolomitization
While the fluid inclusion (Th) and Δ 47 temperatures are broadly similar, they are not totally compatible (Fig. 12) since Δ 47 temperatures are typically much higher (+10 to 40°C) than the Th temperatures (Fig. 12). Such differences could reflect thermal equilibration of Δ 47 temperatures through solid-state reordering either by cryptic dissolution or recrystallization (Mangenot et al., 2017). The temperature at which solid-state reordering occurs in dolomite is still not fully understood, but most studies agree that it happens at much slower rates and higher temperatures in dolomite than calcite (Ryb et al., 2017;Lloyd et al., 2018; and references therein) probably >250 to 300°C (Ferry et al., 2011;Ryb et al., 2017). Since Δ 47 values in calcite the Mount Whyte Formation suggest a temperature of 165°C at maximum burial, the Δ 47 values for dolomite are most likely not thermally re-equilibrated or altered by solid-state reordering. Hence the values should reflect the crystallization temperature during the dolomitization processes. Mangenot et al. (2017) showed that the temperatures measured by clumped isotope analysis and fluid inclusion analysis generally show a good agreement for temperatures ranging from 70 to 100°C, whether the Th temperatures were pressure corrected or not (Fig. 12). In contrast, different studies of high-temperature dolomite (105 to 175°C) show a similar temperature relationship to that observed in this study (e.g. Came et al., 2017;Honlet et al., 2018), where the majority of the Δ 47 temperatures are higher than the fluid inclusion temperatures, even following pressure correction (Fig. 12). In addition, there is also an increase in the deviation between the temperature values at higher temperature (>200°C) (Fig. 12).
While pressure correction (Pc) analysis of the Th could also yield the 'true' entrapment temperature (Tt) (Roedder & Bodnar, 1980), the difficulty in obtaining salinity values in the Mount Whyte Formation (i.e. fluid inclusions did not freeze at the minimum freezing temperature of À180°C) and uncertainty as to the actual pressure regime of the basin, means that pressure correction was not possible. This is potentially further compounded by the interpreted shallow burial depth of dolomitization which creates a relationship between pressure and temperature that is not straightforward. However, if dolomitization in the Mount Whyte and the overlying Cathedral Formation occurred from similar fluids and pressure regime, the fluid salinity (25 wt. % CaCl 2 ) and basin pressure conditions determined by Vandeginste et al. (2005) can be used. In this case, the pressure-corrected temperatures calculated using FLUIDS (Bakker, 2009) are +9 to +17°C higher (based on Zhang & Frantz, 1987) than the measured homogenization temperatures, which is still below the Δ 47 temperatures. In this case, such temperature differences between Th and Δ 47 may reflect the pressure differences during fluid inclusion entrapment, such that the Th temperatures reflect only the minimum entrapment temperatures whereas the Δ 47 temperatures show the true entrapment temperatures (Honlet et al., 2018).

Source of magnesium
The origin of dolomitizing fluids in the burial realm has been widely discussed but a suitable source of Mg cannot always be adequately identified (Wilson et al., 1990;Davies & Smith, 2006;Hollis et al., 2017). Utilizing information on the tectonic evolution of the basin, field geometry and different geochemical signals, four possible sources of magnesium-enriched fluids are considered here: (i) dissolved evaporites (for example, chloride and sulphate groups); (ii) fluid the Stephen/Burgess Shale in the distal part of the platform; (iii) seawater; and (iv) crustal fluids.
While dissolved evaporites are often invoked to provide the necessary magnesium for HTD dolomite and to explain the high fluid salinity (Hendry et al., 2015;Martin-Martin et al., 2015;Fig. 12. Comparison between crystallization temperature from fluid inclusion analysis (Th) and clumped isotopes (D 47 ) of different dolomite phases in the Mount Whyte Formation. Majority of the samples show higher D 47 temperatures than the Th. Increased in deviation at high temperature is also observed from this cross-plot. The data obtained in this study are also compared with previously published data. Koeshidayatullah et al., 2016), the absence of Cambrian, or older, evaporites in the WCSB nullifies the possibility of dissolved evaporites as the main source of magnesium.
If dolomitization took place during shallow burial (<1 km), Mg-rich fluids might have been derived from compactional dewatering of either intraformational Mount Whyte or the Burgess/ Stephen shales. Most of the 87 Sr/ 86 Sr values of the dolomites are beneath the Maximum Sr Isotope Ratio of Basinal Shale (MASIRBAS) value (0.7120; Machel & Cavell, 1999) (Fig. 13B) which may suggest a possible interaction with the shale. However, derivation of large volumes of Mg-enriched brines from normal compaction results in slow flow rates creating only localized dolostone bodies (Drivet & Mountjoy, 1997;Machel & Cavell, 1999). Even during overpressure release, magnesium can be fixed within the shale as chlorite (Frazer et al., 2014). High salinity brines can be derived from shales (Hanor & McIntosh, 2007), but rarely as high as determined in this study. In addition, the maximum burial temperature of the Stephen/Burgess shales when the dolomitization occurred (approximately ca 1 to 2 km; Aitken, 1997;Jeary, 2002) would be only around 70-100°C even with an elevated geothermal gradient (>30°C/km), which is much lower than the recorded crystallization temperature.
Seawater is the volumetrically most viable source of water and magnesium, especially when considering the volume of dolostone bodies produced in the Mount Whyte Formation. Furthermore, the presence of high Fe and Mn concentrations, elevated Sr isotope ratios and evolved d 18 O water can be explained by the interaction of seawater with the underlying Cambrian Gog Group (Terreneuvian, Stage 2) or Precambrian Miette Group. This is particularly plausible since dolomitization occurred during shallow burial, when basal sandstone aquifers have been shown to provide important fluid flow pathways for convection of seawater (Martin-Martin et al., 2015;Hollis et al., 2017;Stacey et al., 2018). In addition, the lower range of d 18 O water values of RD1 (Fig. 13A) intersects with the expected upper range of Cambrian seawater. However, there is still much discussion as to the d 18 O water values for Cambrian seawater, in particular whether it was similar to Modern seawater (mean: 0& SMOW; Henkes et al., 2018; or more negative (mean: À4 to À6& SMOW; Veizer & Prokoph, 2015;V erard & Veizer, 2019). In either case, calculation of d 18 O water for RD2 fabric shows that the mean of isotopic compositions of the dolomitizing fluids are more positive (+2.96& SMOW) than Cambrian seawater ( Fig. 13A; Table 3). Nevertheless, the lack of a seawaterlike REE profile, very high crystallization temperatures and very high fluid salinity (fluid inclusions remained unfrozen at À180°C; e.g. Baldassaro & Bodnar, 2000) of dolomites suggest that there was another significantly hotter and deeper source of fluids. Previous studies of HTD across Canada have proposed fluid interaction with ultramafic rocks (for example, peridotite and serpentinite) and magnesite deposits which have potentially increased the Mg/Ca of dolomitizing fluids (e.g. Powell et al., 2006;Lavoie et al., 2014). More recently, Robertson et al. (2019) have demonstrated the significance of carbonation processes of ultramafic rocks [i.e. listwaniterock consisting of typically magnesite/Mg-bearing carbonate and quartz intergrown with minor fuchsite (Halls & Zhao, 1995;Hansen et al., 2005)] to drive the presence of HTD in Western Canada.
The study area is underlain by complex Palaeometamorphic accreted terranes (Ross et al., 1994), and lithoprobe data taken from around 200 km to the north-east of the study location ( Fig. 1B and C) shows the presence of relict of Palaeoproterozoic (1.75 to 1.85 Ga) subduction zone below the WCSB associated with the Snowbird Tectonic Zone (Eaton et al., 1995;Eaton & Cassidy, 1996;Van der Velden & Cook, 2005). The presence of a low velocity zone (LVZ) within this dataset (>12 km depth at present day) has been interpreted to represent ultramafic rocks with a high degree of serpentinization (i.e. subducted serpentinites) (Eaton & Cassidy, 1996) (Fig. 1B). Carbonation of serpentinite (listwanitization) can lead to the formation of Mg-rich minerals, such as magnesite, together with quartz and water. This process has been suggested to sequester large amounts of CO 2 and could also release large volumes of magnesium to drive the formation dolomite (Halls & Zhao, 1995;Scambelluri et al., 2001;Hansen et al., 2005;Schandl & Gorton, 2012;Falk & Kelemen, 2015;Chen et al., 2016;Robertson et al., 2019). Recent work by Robertson et al. (2019) indicates that quartz and magnesite precipitation is favoured at <300°C and that CO 2 and H 2 S are required to liberate magnesium from ultramafic rocks. These authors also suggested that fluids in equilibrium with magnesite and quartz should have high Mg concentrations. Furthermore, their PHREEQC modelling demonstrates that these fluids not only have a high Mg concentration but also a significantly improved potential to dolomitize limestones due to the modification of Ca 2+ . Therefore, of the different basement rocks beneath the WCSB, serpentinites and their associated by-products could be considered as a potential source of magnesium and high-temperature fluids for dolomitization in the Mount Whyte Formation.
Trace element analysis of R1 and R2 dolomite in this study has shown a positive Eu anomaly and enrichment of LREE compared to HREE, comparable to the REE profiles of hydrothermal fluids hosted in ultramafic rocks (for example, serpentinites) ( Fig. 8A and B; Douville et al., 2002). Dolomite within the Mount Whyte Formation also has high levels of metal enrichment (Table 1), including Fe, Ti and Cr compared to the host limestone, which has also been reported from dolomites associated with ultramafic rocks (Hansen et al., 2005). The positive d 18 O water and radiogenic 87 Sr/ 86 Sr ratios are also within the range of values derived from carbonated serpentinites and their by-products (for example, magnesite) (Falk & Kelemen, 2015;Hinsken et al., 2017) (Fig. 13A and B). Further supporting evidence comes from the cogenetic relationship between dolomite and quartz in all of the Mount Whyte Formation samples (see XRD data, Koeshidayatullah, 2019) since Schandl & Gorton (2012) and Robertson et al. (2019) demonstrated that Mg and Si would be released during the reaction between CO 2 -rich fluids and serpentinites. A recent study by Debure et al. (2019) has also shown that serpentinization derived brines should have a very high salinity, as observed in this study. Finally, the presence of subducted serpentinites at 5 to  (Veizer & Prokoph, 2015;Henkes et al., 2018) and crustal fluids (Schulze et al., 2003, and references therein et al., 1996). The 87 Sr/ 86 Sr ratios of the host limestones are fall within and slightly enriched than the expected values. MASIRBAS and listwanitemagnesite values were obtained from Machel & Cavell (1999) and Hinsken et al. (2017), respectively. 6 km depth beneath the Mount Whyte Formation in the WCSB is in good agreement with the fluid temperatures obtained from Th and D 47 analyses, which were interpreted to have been derived from 5 to 6 km depth, assuming a geothermal gradient of up to 40°C/km. Nevertheless, carbonation of serpentinites requires interaction with CO 2 -rich fluids (Hansen et al., 2005;Falk & Kelemen, 2015;Debure et al., 2019;Robertson et al., 2019) and there is no obvious source of CO 2 for dolomitization during the interpreted period. It is possible that CO 2 was sourced by: (i) degassing during volcanism related to the Terreneuvian to Series 2 Cambrian rifting; or (ii) metamorphism of deeply buried carbonate and clastic strata due to high geothermal gradient. Several studies have reported the presence of mafic and tholeiitic volcanism during the same period as rifting in Western Canada (Stewart, 1972;Cecile et al., 1997;Beranek, 2017;Campbell et al., 2019) and it is possible that CO 2 generated during volcanism could have reacted with the deep-seated serpentinite to liberate large amount of magnesium. Sandiford et al. (1998) and Falk & Kelemen (2015) have interpreted CO 2 generation during metasomatism of clastic sediments resulting in enriched 87 Sr/ 86 Sr values (up to 0.7135), such as observed in the Mount Whyte Formation (Table 3). Since the clastic sediments would have overlain the serpentinites and not been in direct contact with them, however, it is unclear how the CO 2 released by this process, could have caused ultramafic carbonation.
Ultimately, although several lines of geochemical evidence support the interpretation that Mgrich fluids could have been partially derived from ultramafic rocks, there is no direct evidence of serpentinite beneath the Mount Whyte Formation [i.e. lithoprobe data showing the presence of subducted serpentinites is ca 200 km to the north-east of study location (Fig. 1B)]. However, the complex tectonic history of the basin means that the actual underlying stratigraphyprior to thrustingis unknown; hence the possibility remains open. The widespread presence of ultramafic-mafic rocks, carbonated serpentinite (listwanite) and their by-products (for example, magnesite) in the WCSB, British Columbia and Eastern Canada (Eaton & Cassidy, 1996;Hansen et al., 2005;Lavoie & Morin, 2004;Lavoie et al., 2014) suggests that ultramafic carbonation could provide an alternative source of magnesium to dolomitizing fluids within the WCSB and beyond. Several studies have reported a close association between dolomite and magnesite mineralization, a common by-product during the carbonation process, in the Cambrian of the WCSB (Nesbitt & Muehlenbachs, 1994;Powell et al., 2006). A recent study by Robertson et al. (2019) also discussed the mechanism in which the carbonation process of listwanite sourcing the magnesium for fault-controlled dolomitization in Atlin, British Columbia. Lavoie et al. (2014) discussed the presence of hydrothermal dolomitization in Palaeozoic sedimentary strata, Eastern Canada and briefly discussed the possible contribution of Mg-rich ultramafic rocks as a magnesium source of hydrothermal dolomitization. Furthermore, Falk & Kelemen (2015) discussed the presence of dolomite as part of listwanite mineral assemblages in the Samail ophiolite, Jabel Akhdar, Oman, which may indicate the importance of this carbonation process of ultramafic rocks. The widespread presence of HTD in Permian and Jurassic carbonate platforms in the Jabel Akhdar area (Vandeginste et al., 2013;Beckert et al., 2015) may need to be further investigated in order to support the wider application of carbonation of ultramafic rocks in driving the dolomitization process.

Fluid flow model
It has been proposed that RD1 has an isotopic signature that is suggestive of dolomitization from seawater, whilst RD2 is more positive. Strontium isotopic signatures are suggestive of fluid interaction with radiogenic rocks (for example, clastic or basement). Although seawater would have been abundant and able to convect at the proposed burial depths of <1 km, potentially utilizing the basal Gog Group or Miette sandstones as pathways, dolomitization has been shown to occur from very hot, saline fluids that are inconsistent with seawater. Without a confident structural reconstruction of the study area it is not possible to show that it was underlain by serpentinized basement, but it is possible that mixing of deep-seated hydrothermal fluids derived from ultramafic carbonate occurred within the Mount Whyte Formation during convection of seawater in the shallow subsurface, particularly during formation of RD2. This is corroborated by the lower 87 Sr/ 86 Sr ratios and lower d 18 O water (closer to seawater values) of RD1 than RD2 ( Fig. 13B; Table 3). It is possible that these deep-seated, ultramaficderived brines further increased their Mg/Ca ratio during fluid-rock interaction within the Miette or Gog Groups or by interaction with Mesoproterozoic magnesite, as suggested for the overlying Cathedral Formation (Jeary, 2002;Davies & Smith, 2006;Powell et al., 2006;Stacey et al., 2017). Overall, however, fluid mixing could explain the fact that 87 Sr/ 86 Sr ratios of RD1 and RD2 are lower than MASIRBAS since previous studies have indicated that interaction between Mg-rich fluids with clastic rocks should have yielded 87 Sr/ 86 Sr ratios that exceed the MASIRBAS value (Drivet & Mountjoy, 1997;Duggan et al., 2001;Packard et al., 2001). Furthermore, considering the very high salinity and temperature of fluids recorded from both RD1 and RD2 fabrics may suggest that high volumes of fluids were released from deep-seated brines during the coseismic dilatancy pumping associated with Cambrian rifting and mixed with the seawater.

Mechanism and timing of dolomitization
In the Mount Whyte Formation, it has been shown that the dolomitizing fluids were varied in isotopic composition (for example, d 18 O water and 87 Sr/ 86 Sr), very hot and very saline, potentially as a result of mixing between seawater and Mg-enriched fluids supplied from carbonation of deeply seated (ca 5 to 6 km) serpentinites at shallow burial depths (<1 km). However, the driving mechanism for fluid flow is not straightforward to unravel due to complex tectonic overprinting, particularly during the Mesozoic Laramide Orogeny that has obscured most of the precursor structural grain. The formation of HTD in the Mount Whyte Formation has been interpreted to involve a large gap between the depth of dolomitization (<1 km) and fluid source (5 to 6 km), and therefore the most likely drive mechanism for fluid migration is tectonic activity, since to transport such deepseated fluids into the near surface requires high angle, permeable conduits. Since dolomitization is interpreted to have occurred shortly after deposition, during the Cambrian, it could have coincided with Cambrian rifting and thermal subsidence (Powell et al., 2006). Consequently, normal faults would have been dilatant and heat flow high due to ongoing extension (Fig. 14A). Therefore, hot fluids could have been quickly supplied by seismic pumping along faults into the shallow sedimentary succession. This process could have occurred from multiple pulses of large volumes of Mg-rich fluids supplied during coseismic dilatancy.
The interpretation of dolomitization at a shallow depth is corroborated by the alteration of original Cambrian seawater REE and trace element signatures in the dolostone suggest high water/rock ratios during dolomitization. This is in a good agreement with water/rock ratio calculations ( Fig. 8A and B) (Banner et al., 1988;Banner & Hanson, 1990). The localization of dolomitization to bioturbated mudstones is perhaps noteworthy, since although mudstones provide a high reactive surface area, their permeability is typically low (Gabellone & Whitaker, 2016). However, at burial depths of <1 km, the total porosity of the mudstone would still be high (potentially up to 30%; Schmoker & Halley, 1982;Goldhammer, 1997). This condition coupled with the presence of permeability-enhancing burrows may therefore provide sufficient flow pathways for circulation of dolomitizing fluids. Furthermore, organically mediated dolomitization of burrows might have provided 'seeds' that accelerated dolomitization. In contrast, the pervasively cemented grainstone that overlies the dolostone body, and the underlying shale could have focussed fluid flow laterally ( Fig. 14B) to form the observed stratabound terminations.
Ultimately, this study provides an unambiguous example of 'true' hydrothermal dolomitization where the fluid temperature is significantly hotter (>150°C) than the ambient rock temperature. In this case, ambient temperature of the host rock should be around 60 to 75°C (based on 35 to 40°C/km geothermal gradient and 35°C surface temperature) while the dolomitization temperature ranges from 120 to 235°C. These recorded crystallization temperatures may have been significantly lower than the actual fluid source temperature due to rapid cooling of fluids during thermal convection and also mixing processes. Therefore, these temperature contrasts (>50°C) have significantly exceeded the temperature difference threshold for hydrothermal dolomitization (>5-10°C; Machel & Lonnee, 2002). The model infers that dolomitization occurred much earlier (Series 2, Stage 3 to stage 5 Cambrian) than previously proposed (for example, Furongian to Cretaceous; Yao & Demicco, 1997;Jeary, 2002;Vandeginste et al., 2005;Davies & Smith, 2006). This earlier timing of dolomitization is supported by Powell et al. (2006), who also reported that mobilization of Mg-rich brines along the Kicking Horse area was associated with Cambrian rifting and the observation of large eroded blocks of dolostone that occur within the Stephen/Burgess Shale (Collom et al., 2009).

CONCLUSION
Dolomitization of the Mount Whyte Formation is interpreted to have occurred shortly after deposition, in a shallow burial setting (<1 km), during the late syn-rift. Petrographically, dolomitization in the Mount Whyte Formation is characterized by fabric-destructive dolomite with a wide range of crystal fabrics (RD1, RD2 and SD). All dolomite fabrics are post-dated by bedding-parallel, low amplitude stylolite. These dolomite fabrics also exhibit distinct geochemical signatures compared to the host limestone, including: (i) more negative d 18 O dolomite and more positive d 18 O water ; (ii) low concentrations of Sr and high Fe and Mn concentrations; (iii) positive Eu anomaly with enriched LREE and depleted HREE. These signatures are suggestive of hydrothermal dolomitizing fluids. Formation temperatures of the dolomite estimated from both Tm and Δ 47 analyses also indicate a very hot fluid (up to 235°C). Such temperatures, if considering a high geothermal gradient during Cambrian rifting (35 to 40°C/km), should indicate a deep fluid source, potentially from approximately 5 to 6 km depth.
Although aspects of the geochemical fingerprint of RD1 and RD2 dolomite are supportive with dolomitization from heated and modified seawater, high fluid temperatures and salinities are interpreted to potentially reflect a high degree of mixing with Mg-enriched fluids supplied by carbonation of subducted serpentinites below the Western Canada Sedimentary Basin (WCSB). These fluids could have been supplied by seismic pumping during rifting. This interpretation is corroborated by typical ultramaficderived fluid rare-earth element (REE) patterns, the enrichment of metal concentrations and cogenetic relationship between dolomite and authigenic quartz in the dolomite phases compared to the host limestones. Nevertheless, the complex tectonic history and absence of clear evidence for a source of CO 2 by which carbonation could be initiated means that aspects of this model require further investigation. Ultimately, this study shows that the formation of largescale hydrothermal dolomitization (HTD) dolostone bodies preferentially took place in shallow burial depth and required multiple pulses and sources of fluids in order to satisfy the mass balance problem.
supported by a NERC-CDT grant to CH and JS. HR was funded by the PD3 consortium at University of Manchester, supported by Tullow Oil, Woodside Energy and Wintershall DEA. We thank IAS and AAPG postgraduate grants for AK in providing additional support for this study. Help from colleagues at the Williamson Research Centre, University of Manchester, Alberta Geological Survey and Stable Isotope Laboratory, University of Miami, are gratefully acknowledged. We are grateful for the kind support from Matthew Steele-MacInnis and Pillar Lecumberri-Sanchez for the fluid inclusion analysis. Stable isotope analysis was funded through a NERC grant IP-1759-1117 at the NERC Isotope Geoscience Facility in East Kilbride to AB, CH and AK. Anping Hu is acknowledged for the help with strontium isotope analysis. Constructive comments from Editor Giovanna Della Porta, Associate Editor Hairuo Qing, Jeff Lonnee and two anonymous reviewers are greatly acknowledged and have improved the clarity of the paper.

DATA AVAILABILITY STATEMENT
The data that support the findings of this study are available from the corresponding author upon reasonable request.