deletions | additions       diff --git a/untitled.tex b/untitled.tex  index ab7160a..7eb2245 100644  --- a/untitled.tex  +++ b/untitled.tex 
...
There is abundant evidence that pore fluids and fracture processes in the upper crust are physically and chemically coupled (\citep{Hubbert:1959ea, Frank:1965wi, Nur:1973ju, Sibson_1981}. Increases in pore fluid pressure, for example due to fluid injection, can lead to rupture and an increase in seismic activity. This is readily explained by the Navier-Coulomb criteria for brittle failure and the decrease in effective stress as a result of the elevated pore fluid pressure \citep{Price:1966uh, Sibson_1981}. Conversely, the changes in groundwater levels and the surface effusions of warm water that sometimes occur along fault traces following earthquakes show that failure can also have a profound effect on fluid flow, heat and mass transport \citep{Nur:1974ht, Sibson:1975cn, Sibson_1981}. The flow is interpreted to result from either seismic pumping as a result of dilatancy diffusion type processes or a seismic valve mechanism in which fault rupture leads to leakage of an overpressured aquifer or reservoir of fluid \citep{Nur_1972, Sibson_1981}.   In the geological record evidence for episodic and rapid flow of fluids associated with fracture and faulting is more equivocal. The presence of banded hydrothermal mineralisation of varying degrees of complexity in exhumed faut systems is often taken as evidence of pulsed fluid flow driven by seismic activity. Failure may result in brecciation of earlier generations of veins and the opening of large dilation voids that are then cemented by mineral precipitation from upwelling hydrothermal fluids \citep{Wright:2009ej}. The extent, however, to which mineral precipitation in the veins is contemporaneous with, and directly coupled to failure is open to question. Observations of epithermal mineralisation associated with, for example, dilation jogs between en-echelon fault segments suggests that in this structural setting fluid flow is rapid. It may also be accompanied by rapid pressure fluctuations which triggers high level boiling, or effervesence of hydrothermal fluids. This results in precipitation of common gangue quartz and calcite as well as economic metalliferous deposits \citep{Sibson:1975cn} Sibson:1981kt, Sibson:1987dq, Henley:2000wk}). Henley:2000wk).  A defining feature of systems with episodic pulsing of warm to hot fluid is the development of a thermal anomaly along the high permeability paths in which flow is focused. A good example is the perturbation of the temperature field that has been observed in Mississippi Valley Type (MVT) mineralisation districts that lie around the upper Mississippi Valley and major Palaeozoic sediment basins in the continental USA (add references). Cathles and co-workers have identified a difference in mineral precipitation temperatures reported for Mississippi Valley Type (MVT) deposits in the basal Cambrian quartzite aquifer of the Arkoma Basin and that typical of the host rock for the depth of burial at the time of mineralisation. Similar temperature differences have been reported between lead-zinc mineralisation temperatures and the host rock for deposits located in other sedimentary basins, for example to the south east of the Massif Central in France (Charef and Sheppard, 1986). In both cases a cogent argument can be made that the difference is due to heat advection associated with episodic release of hot fluids from deeper regions of the basins. In these example the fluids are thought to originate from overpressured brines in compacting sedimentary basins. Importantly, the fluid velocities required to produce the thermal anomaly are more than 1000 times greater than could be produced by the steady subsidence, compaction, and dewatering of the basins (Cathles and Smith, 1983). This suggests to us that the pulses result from a coupling between the pore fluid pressure and rock failure. We envisage that when pore fluid pressures approach lithostatic pressure either hydraulic fracturing or shear failure and development of dilation jogs along fault surfaces allows rapid dewatering of the sediment pile with channel flow of fluids into the basal aquifers.  A corollary of the model proposed by Cathles and others is that the fluids involved in the mineralisation are trapped formation and not meteoric waters. Additionally the mineralisation involves restricted water volumes at consequent low water:rock ratios. Supporting evidence for formation waters is found in the fact that in the modern upper Mississippi Valley basins the deep waters are saline (Cathles, 1993). Were the waters to have been meteoric one expects the salinity to have been flushed out as recharge in areas of elevated terrain drives cross-basin flow along path lengths of several hundred kilometres. Thus the evidence of a thermal anomaly, coupled with remnant salinity in the sedimentary basin runs counter to the current paradigm for MVT formation which involves cross-basin flow of large volumes of gravity driven meteoric water (Garven plus other references). Rather a more dynamic regime involving fluid overpressure, fracture and fluid flow is implicated.  
...
(iii) The calcites exhibit zoned development characterised by cyclic and pulsed evolution of precipitation temperatures and fluid compositions as a result of varying mixing ratios of the two fluid end-members.  (iv) The fluids have evolved under low fluid:rock ratios.    \section{Geological setting of the Variscan faults and vein fill}  We studied samples of hydrothermal vein calcite from three areas of Carboniferous limestone in the British Isles: (i) The Peak District of the southern Pennines in Derbyshire and Staffordshire; (ii) The Gower Peninsula, South Wales, and (iii) The Burren, County Clare, Ireland, Figure 1. All the locations lie to the north of the Variscan front in the British Isles with faulting and vein fill in each area associated with foreland deformation during the late Carboniferous and early Permian periods.  There are distinct similarities in the general geologic setting of the three areas. During mid-Devonian to the end of the lower Carboniferous Britain was subject to a north-south back arc-extension as a result of subduction of the Rheic ocean to the south. Extension was largely accommodated by the development of a series of graben and half graben deep water basins bounded by normal growth faults and with foot wall topographic highs. Amongst the basins are: the Edale Gulf, Goyt Trough and Widmerpool Gulf in the Peak District; the Culm basin to the south of the Gower Peninsula, and; the Clare Basin along the axis of the Shannon estuary in the west of Ireland. Dinantian shallow water platform and ramp carbonates accumulated on the topographic highs whilst deep water facies limestones and shales were deposited in the basins (Figure 1). Both extension and carbonate sedimentation ended at the end of the Dinantian and was succeeded by Namurian shales and later a range of Westphalian river and deltaic sandstone facies. The total thickness of post Dinantian sediment accumulation on the platform limestones ranged from a 1 to 2km, with correspondingly thicker sequences of 3-4km accumulating in the basins.   Variscan shortening resulted in basin inversion and in the foreland reactivation of extensional faults with both significant strike slip and reverse components of movement. There is evidence for high pore fluid pressures during the period of faulting with the development of mode I fractures and in places a pervasive mesoscale fracture network. Fracture dimensions range from sub-mm to fault widths of several metres. Fracture and fault vein fill at all the locations is dominated by white, sparry calcite often showing a syntaxial growth pattern with varying degrees of complexity as a result of repeated episodes of movement, brecciation and renewed hydrothermal mineral growth.  Details of the sample localities are given below.  (i) Peak District. The Derbyshire platform is largely composed of shallow water Dinantian shelf limestones that show little evidence of deformation. The structural evolution of the area is summarised by Quirk (1991). During the lower Carboniferous NE-SW extension, accompanied by normal faulting produced a series of WNW-ESE trending deep water basins that border the platform. The basins are infilled with thick sequences of Dinantian basin limestones, shales and sands. Rotation of the stress field during the upper Carboniferous as a result of the Varsican orogeny to the south resulted in fault re-activation with a significant strike-slip component of movement. It is these re-activated faults that host the largest mineral veins exposed on the North Derbyshire platform.  The thickness of overburden on the platform is not well constrained. Estimates based on reconstruction of post Dinantian upper Carboniferous stratigraphy range up to 2km, with maximum sediment thickness in the basins of 3km (Colman, 1989). Similarly the temperature conditions during vein formation are not well constrained. Fluid inclusion temperatures obtained from hydrothermal calcite suggest temperatures ranging from 50° to as high as 240°C (Rogers, 1977; Masheder and Rankin, 1988; Rodger, 1996; Kendrick et al., 2002).  We sampled hdrothermal calcite from five sites in the Peak District.  (i) Dirtlow Rake (GR) is a major WSW-ENE trending strike slip fault lying just to the south of Castleton. The width of the exposed fault is greater than 10m and it has been extensively worked over a greater than 10km length for Pb (galena) and Zn (sphalerite). Hydrothermal calcite occurs as large syntaxial and elongate, sparry crystals. The growth form often exhibits dog-tooth terminations indicative of growth into a void.  (ii) Pindale  (iii) Hucklow Edge (GR)  (iv) Blakelow Lane quarry (GR)   (v) Ecton Mine.  (ii) Gower Peninsula. The Gower Peninsula lies to the south of the Wales-Brabant Massif and just north of the Variscan Front which is marked by the Bristol Channel Fault Zone, a major south dipping thrust (refs). The peninsula is composed of moderately deformed Devonian and Carboniferous rocks, Figure 2. The deformation is consistent with a north-south compression associated with the Variscan orogeny. Structural elements include folded open, ESE-WNW trending folds that gently plunge to the east. Associated with the folding are a series of contemporaneous SSW dipping thrust faults (e.g. Bishopston thrust) that verge towards the north and the Port Eynon thrust, a NNE dipping backthrust (Wright et al., 2009). A transition in the stress regime from vertical σ3 to vertical σ2 is marked by the development of SSW-NNE trending, vertical wrench faults with development of strike slip and oblique slip slickensides. Fault displacements range from several tens to hundreds of metres. Where these faults cut the lower Carboniferous limestones they are characterised by mineral filled fissures several metres wide with a wide range of infill textures (Wright et al., 2009). These include dendritic haematite, elongated syntaxial calcite growth, spar ball and cockade breccia formation, sediment infill and void collapse breccias (Wright et al., 2009, Frenzel and Woodcock, 2014).   The overburden thickness and P-T conditions during deformation and mineralisation are poorly constrained. There are no available estimates of temperature based on either isotope geothermometry or fluid inclusion homogenisation temperatures. Reconstruction of sediment stratigraphy by comparison with adjacent areas where upper Carboniferous sediments are preserved suggests maximum burial depths of 3 to 4km (Sibson, 1981, Wright et al., 2009) at the onset of Variscan inversion. These values are somewhat greater than estimates of Cenozoic denudation which range between 0.5 and 1.5km for South Wales and the Bristol Channel (Jones et al., 2002, Hills et al., 2008). Though not directly comparable because of the lack of data for Mesozoic cover and erosion the estimates place lower and upper bounds on the depth at the time of deformation of 0.5 to 4km.  We sampled hydrothermal calcites from three vertical wrench faults exposed on the foreshore at Limeslade (GR), Oxwich East (GR) and Oxwich West (GR).   (iii) Burren, County Clare. The Burren is composed of a sequence of relatively undeformed, shallowly dipping (2° to the south) Asbian and Brigantian limestones that are uncomformably overlain by the condensed Clare Shales of Namurian age. To the south lies the Clare Basin filled with a thick sequence of Namurian and Westphalian shales and sandstones. Exposures of the Brigantian Slievenaglasha limestones at Fisherstreet (GR) comprise 3-20m thick cyclic units of crinoidal grainstones with abundant corals, cherty layers and a well developed joint system. In addition the limestones are cut by vertical veins ranging in width from a few mm to tens of cm. The veins are persistent and, unlike the jointing, cross several limestone beds and have a near north-south strike (Gillespie et al., 2001). Most of the veins are mode I hydraulic fractures with little or no apparent displacement. The veins are typically filled with white sparry calcite though locally both fluorite and galena can be found.   Estimates of the depth of burial at the time of vein formation are poorly constrained. Gillespie et al. (2001) have attempted to estimate the burial depth by: (i) reconstructing the missing Silesian overburden prior to the culmination of Variscan activity, and; (ii) using the mechanical characteristics of veins crossing discontinuities between beds of similar elastic properties. They determine a likely minimum and maximum depth of burial of 1.25km and 2km respectively. Previous fluid inclusion homogenisation temperatures obtained from fluorite in the veins range from 85 to 200°C (O’Connor et al., 1993) .  We sampled hydrothermal calcite from fractures in the Slievenaglasha Limestones that are exposed on the coastal section to both the north and south of Fisherstreet (GR).  \section{Methods}  \subsection{Sample preparation and mass spectrometry}  Samples of vein calcite were prepared for clumped isotope analysis by either gently crushing small crystal fragments in an agate pestle and mortar, or by drilling using a dental bit in a Dremel drill and avoiding undue use of pressure so as to minimise frictional heating of the sample. Analyte CO2 is produced by reacting 6-8mg of each sample with 102% ortho-phosphoric acid in vacuo at 25° for a period of 12 hours. The evolved CO2 is then dried, collected by cryo-distillation into a calibrated volume manometer to check reaction yields and then stripped of potential hydrocarbon contaminants before collection in a valved glass gas tube. The drying stage involves freezing the CO2 into a glass spiral trap at liquid nitrogen temperatures before sublimation at -120°C, passing the gas through a second trap at -120°C, and freezing with liquid nitrogen into the manometer. We strip any potential hydrocarbon contaminants from the CO2 by cryo-distillation into a gas tube via a 20cm glass tube packed with porapak Q ion exchange resin held at a temperature of -20°C.  The sample gases were analysed for their clumped isotope values, δ45 - δ49 on the UEA MIRA dual-inlet isotope ratio mass spectrometer (Dennis, 2015). Samples were analysed at a major beam (m/z=44) intensity of 7.5×10-8A and data collected for each cardinal mass of the CO2 molecule (m/z = 44 - 49). Each sample measurement consists of 4 acquisitions each of 20 reference-sample gas pairs. Before analysis and between each acquisition the sample and reference gas volumes and signal strengths are balanced to within 1%. Each sample or reference cycle consists of a 10s ‘dead time’ after switching of the changeover valve followed by a 20s integration period. The total measurement time, including sample and reference gas balancing is approximately 90 minutes. The total integration time is 1600s each for the sample and reference gas. Internal precisions (±1σ) for δ45 and δ46 are better than 0.001‰, for δ47 better than 0.008‰, for δ48 better than 0.03‰ and for δ49 better than 10‰.  The reference gas used in MIRA is CO2 produced by reaction of BDH marble chips with 85% ortho-phosphoric acid and subsequently equilibrated with water at 20°C for a period of 1 month. This is to ensure that the Δ47 value of the reference gas is in equilibrium at the laboratory temperature. The nominal composition of the reference gas is: δ13C = 2.007‰VPDB; δ18O = 34.899‰VSMOW, and Δ47 = 0.94URF. To ensure a robust calibration of scale compression and transfer function between the local reference frame for Δ47 and the universal reference frame (URF) both 1000°C heated and 20°C water equilibrated reference gas samples are measured on a daily basis (Dennis et al., 2011). Data quality and long term stability of measured values is monitored by daily measurement of two laboratory standards that bracket the range of Δ47 values for samples in this study: UEACMST (Δ47 = 0.39±0.01, n= ) and UEAHTC (Δ47 = 0.56±0.01, n=).  MIRA is inherently linear with no variation in the calculated Δ47 and Δ48 values of samples as a function of their bulk isotopic composition as represented by the δ47 and δ48 values of samples (Huntington et al., 2010). Not-with-standing this we regularly check for linearity by measurement of 1000°C heated cylinder CO2 (BOC) that is depleted in δ47 with respect to the reference gas by approximately 65‰. We have not recorded any change in the linear behaviour of MIRA over a four year period since commissioning.  (ii) Data handling and calculation of Δ values.  The clumped isotope Δi value of a sample is defined as:   (1)  where Ri is the measured ratio of isotopologue i to the non-isotopically substituted isotopologue and Ri* is the expected ratio assuming a stochastic distribution of all isotopes over all possible sites in the lattice (reference). For CO2 we are largely concerned with the isotopologue 13C18O16O (i = 47) but also determine Δ values for 13C18O17O (i = 48) and 13C18O18O (i = 49).  The ratios Ri and Ri* are determined from the measured δi, δ13C and δ18O values of the sample CO2. For R47:   (2)  where R47wrg is the 47/44 ratio of the working reference gas (wrg) and is determined as:   (3)  Note that implicit in this treatment is an assumption that the mass spectrometer working reference gas has a stochastic distribution of isotopes. It is self evident that this is incorrect since the working reference gas has been equilibrated with water at the laboratory temperature. However, since the Δ values are <1‰ we can make this assumption and carry out a later linear transformation of the data to take account of the actual reference gas R47 value without introducing any significant errors.  The ratios R13, R17 and R18 are determined from the δ13CVPDB, δ17OVSMOW and δ18OVSMOW values of the working reference gas:   (4)  and similar equations for R17 and R18. The reference gas ratios are R13VPDB = 0.0112372 (reference), R17VSMOW = 0.0004023261 (reference) and R18VSMOW = 0.0020052 (reference).  Similarly the R47* ratio for a sample corresponding to a stochastic distribution of the isotopes is given by:   (5)  with R13, R17 and R18 determined using equation 4 and the measured δ13C, δ17O and δ18O values for the sample.  Substitution of R47 (equation 2) and R47* (equation 5) into equation 1 allows determination of Δ47. Evaluation of Δ48 and Δ49 follows the same steps as above. The complete data reduction algorithm and it’s implementation in a Mathematica program for data obtained on the MIRA mass spectrometer is included in the supplementary information.  Finally, using the method outlined by Dennis et al. (2011) and the heated and water equilibrated gas standards we determine a transfer function between measurements made on the local scale with reference to the mass spectrometer working reference gas the absolute reference frame (ARF). All the results are reported with respect to the ARF. The full data set including results reported on both the local and ARF scales are included with the supplementary information.  (iii) Temperature estimation using Δ47  Using the clumped isotope composition of carbonate minerals as a geothermometer is a relatively immature technique. Of particular importance is the lack of agreement on the calibration between the deviation from a stochastic distribution of the isotopes and temperature. There exists a large number of calibrations with varying degrees of sensitivity. Moreover, most of these calibrations have been made over a restricted range of temperatures based on the collection of biogenic carbonates from environments with known temperature (e.g. references). More recently calibrations have become available for inorganically precipitated calcites. However these have not served to resolve the origin of the differences.  What is clear from these studies is that differences between calibrations are laboratory dependent. i.e. measurements of unknown samples when converted to a temperature only make geological sense when the local Δ47 - T calibration is used. If another laboratories calibration is used then unrealistic temperatures are often estimated. This points to significant methodological issues associated with sample preparation and mass spectrometry that have yet to be resolved. We don’t discuss these here. For this study we have used the UEA determined calibration to translate Δ47 values for carbonates into temperature.    (6)  This calibration is based on (i) biogenic carbonates (bivalves and foraminifera) collected from environments with known environmental temperature; (ii) natural inorganic calcites from travertines precipitated at spring heads with known temperature, and (iii) experimental inorganic carbonates precipitated or recrystallised at known temperatures. A single calibration is consistent across the temperature range covered by the calibration from 0 - 600°C. We note that within measurement error this calibration is indistinguishable with the theoretical line of Guo et al. (2009).