Structural and stable isotope studies of calcite veins and host rocks in a 1000 m thick, Devonian limestone sequence in the Taemas area of the Lachlan Orogen in SE Australia indicate that externally derived fluids migrated through the sequence over a protracted interval during folding and associated reverse faulting at temperatures of ∼200°C. The evolution of fluid pathways was governed by growth of fold-related and fault-related fracture networks at transiently supralithostatic fluid pressures. The internal structures of veins indicate growth during repeated cycles of permeability enhancement and sealing. Systematic increase in δ18O of vein calcite upward through the limestone sequence, and the presence of an O isotope alteration front in the vein system, is related to buffering of fluid compositions by fluid-rock reaction during upward migration of overpressured, low δ18O fluids. Repeated switching of fractures between high-permeability and low-permeability states promoted episodic flow and changes in flow paths during progressive deformation. A lack of significant migration of the isotopic front during the flow history is consistent with fluid-rock interaction being associated with numerous, separate pulses of fluid ascending through the percolation network. Irregular variations in fluid δ18O with time during vein growth are interpreted to be driven by repeated changes in reactive path lengths, reaction rates, and flow rates in a discontinuous flow regime. Systematic upward changes in vein δ18O over an area of 20 km2 requires that on a time-averaged basis, most of the vein system was hydraulically well connected to the external fluid reservoir and that growth of the fracture network occurred by fluid-driven, invasion percolation processes. Repeated failure and sealing events in the fracture system are interpreted to have been driven by episodic migration of fluid pressure waves up through the fracture network, possibly in response to deeper level fault rupture events repeatedly breaching an overpressured fluid reservoir.
If you can't find a tool you're looking for, please click the link at the top of the page to "Go to old article view". Alternatively, view our Knowledge Base articles for additional help. Your feedback is important to us, so please let us know if you have comments or ideas for improvement.
 Growth of fracture networks during crustal deformation plays a critical role in generating fluid pathways in intrinsically low-permeability rocks, and facilitates fluid redistribution between various crustal reservoirs. The formation of fracture networks at depth can be driven by changes in both stress states and fluid pressures [Cox et al., 2001; Sibson, 2001]. Particularly where fracture networks breach overpressured fluid reservoirs at depth in the crust, “fluid-driven failure” [Sibson, 1996] has potential to be the major factor controlling growth and repeated reactivation of fracture systems that govern reservoir drainage and the architecture of crustal-scale flow systems [Cox, 2005].
 In actively deforming regions, fluid pathways in fracture-controlled flow networks are governed by the evolution of fracture connectivity between the upstream (reservoir) and downstream parts of the system. Percolation concepts provide a valuable framework in which to understand the distribution of fluid flux in fracture systems [Guéguen et al., 1991; Sahimi, 1994; Berkowitz, 1995]. In terms of hydraulic connectivity to a fluid reservoir, fracture networks can be considered in terms of “backbone,” “dangling,” and “isolated” elements. Most flow occurs along backbone elements of the network which connect the upstream and downstream parts of the system (Figure 1). Dangling fractures connect to the flow backbone, but if they terminate in a low-permeability matrix, they do not host high fluid fluxes. Isolated fractures are not connected to the flow backbone; accordingly they host no flow if they occur within intrinsically low-permeability rock.
 Two end-member scenarios for growth of fracture systems can be envisaged [Cox, 2005].
1.1. Ordinary Percolation
 If fracture growth is driven predominantly by stress states, nucleation and growth of fractures occur throughout a deforming, homogeneous rock mass. Early formed fractures are isolated. However, as more fractures nucleate and grow during ongoing deformation, connectivity between fractures progressively increases (Figure 1a). The system reaches a critical point, known as the “percolation threshold,” when connectivity between the upstream and downstream parts of the system is first reached. With further deformation, the proportion of fractures forming the flow backbone increases, and the proportion of isolated fractures decreases.
 As fractures tend to be isolated during the early stages of growth of fracture systems by “ordinary percolation,” mass transfer associated with vein sealing at this stage is dominated by diffusional mass transfer or short-range advective fluid redistribution. At this stage veins have isotopic compositions which reflect buffering by proximal host rocks [e.g., Cartwright et al., 1994]. When fracture networks reach the percolation threshold, veins formed along the flow backbone and dangling parts of the network will have isotopic compositions influenced by the externally derived fluids. As the proportion of the network belonging to the flow backbone increases, vein fillings whose isotopic compositions reflect buffering by externally derived fluids will increase in abundance. Transitions from rock-buffered conditions to more fluid-buffered conditions recorded in isotopic studies of some veins [e.g., Dietrich et al., 1983; Rye and Bradbury, 1988; Lefticariu et al., 2005] could reflect fracture systems crossing a percolation threshold.
1.2. Invasion Percolation
 If fracture-controlled flow systems grow by fluid-driven failure, formation of fractures and veins is dominated by preferential growth of those parts of the network which connect directly to high pore fluid factor reservoirs (pore fluid factor λv is the ratio of fluid pressure to vertical stress). During invasion percolation, fracture networks grow progressively away from the overpressured reservoir, much like the spread of a virus or bushfire (Figure 1b). From the earliest stages of growth of a percolation network, most fractures are connected to the external fluid reservoir. Accordingly, isotopic compositions of most veins will reflect buffering by externally derived fluids. The evolution of isotopic compositions of veins in fracture-controlled flow systems therefore has potential to determine whether they have grown by ordinary percolation, or by invasion percolation processes.
 Although fracture networks will tend to grow during progressive deformation, competition between permeability enhancement (fracturing) and permeability destruction (fracture sealing) can result in components of the percolation network episodically switching between a high-permeability state and a low-permeability state [Miller and Nur, 2000]. This should lead to dynamic changes in fluid pathways as failure events migrate around fracture systems. As a consequence, flow at any one site is likely to be episodic rather than continuous. In the upper crustal seismogenic regime, episodic changes in permeability distribution in fracture systems are likely to be coupled with changes in fluid pressure and stress states associated with seismic cycles [Cox, 2005].
 The spatial and temporal variations in isotopic compositions of fluids, recorded by accretionary growth of veins, have potential to provide insights about the evolution of fracture-controlled flow systems. As externally sourced fluids react progressively with rocks along the flow path, they become more rock-buffered toward the downstream parts of the system. Reactive transport modeling of isotopic profiles in veins and wall rocks along flow paths can be used to constrain flow directions and fluid fluxes [e.g., Lassey and Blattner, 1988; Bowman et al., 1994; Barnett and Bowman, 1995; Gerdes et al., 1995; McCaig et al., 1995; Abart and Sperb, 1997; Abart and Pozzorini, 2000; Knoop et al., 2002]. The isotopic compositions of vein minerals formed during incremental growth of veins may also preserve a record of changes in fluid chemistry during the evolution of a flow system [Dietrich et al., 1983; Rye and Bradbury, 1988]. Potentially available information includes a record of migration of isotopic fronts through percolation networks, as well as insights about evolution of flow rates, reaction rates and pathways in fracture systems. Accordingly, the integration of stable isotope data from incrementally grown veins, with understanding of deformational controls on the development of fluid pathways and their permeability history, offers potential to explore the dynamics of coupling between deformation processes, fluid migration and fluid-rock reaction during crustal deformation.
 This paper combines structural and C/O stable isotope studies of veins and their host rocks to examine the development of a fracture-controlled, advective flow system that formed during progressive folding of a limestone sequence at upper crustal levels. Spatial and temporal variations in C/O isotope compositions of calcite veins, along with reactive transport modeling, are used to explore how progressive fluid infiltration and changes in reactive path lengths, reaction rates, flow rates and fluid fluxes can all influence the dynamics of fluid-rock reaction and stable isotope chemistry of veins. The study highlights the role of fluid-driven failure in driving the growth of percolation networks, and emphasizes the importance of episodic fluid flow and reaction processes in upper crustal, fracture-controlled flow systems as they drain overpressured fluid reservoirs.
2. Geological Setting
 The Taemas area is located within the Yass zone of the eastern Lachlan Orogen [Glen, 1992] in New South Wales, southeastern Australia. The area is approximately 50 km north west of Canberra (Figure 2). Here, the folded and faulted Murrumbidgee Group comprises an approximately 1000 m thick sequence of Lower Devonian limestone with some interbedded sandstone, mudstone and marl (Table 1). The basal Cavan Bluff Formation is up to 125 m thick and dominated by massive to thinly bedded limestone with minor interbedded mudstone and sandstone. It is overlain by the Majurgong Formation, a red bed sequence containing 30 to 120 m of interbedded sandstone, mudstone and minor limestone. The overlying Taemas Formation comprises up to 750 m of massive limestone and sequences of thinly interbedded limestone and mudstone. The Murrumbidgee Group overlies up to several thousand meters of Lower Devonian felsic to intermediate volcanics, volcaniclastics and shale known as the Black Range Group [Cramsie et al., 1978]. This sequence, in turn, overlies a regionally extensive Silurian volcano-sedimentary sequence and a thick, Ordovician quartz-wacke turbidite succession.
Table 1. Stratigraphy of the Murrumbidgee Group
Approximate Thickness, m
thinly bedded limestone and minor shale
Massive, coarsely bedded limestone
thinly interbedded limestone and shale
Spirifer Yassensis Limestone
thinly interbedded limestone and shale
interbedded sandstone and shale red bed sequence
Cavan Bluff Formation
thinly bedded limestone and minor shale
 In the Taemas area, the Murrumbidgee Group lies on the eastern limb of the regional-scale Narrangullen Anticline [Cramsie et al., 1978] and forms a 5 km wide synclinorial inlier (Figure 2). This inlier occurs in the footwall of a major, steeply dipping fault system (Warroo and Deakin–Devil's Pass Faults) along which the Silurian volcano-sedimentary sequence apparently has overthrust the Murrumbidgee Group to the west. The structure of the Murrumbidgee Group is controlled largely by gently plunging, upright, NNW trending folds (Figure 3) [see also Hood and Durney, 2002]. Folds in the thinly bedded lower to middle parts of the Murrumbidgee Group predominantly have interlimb angles in the range 60°–110°, with wavelengths between 200 and 1000 m. However, thick, massive limestone sequences in the upper Taemas Formation are associated with more open, rounded, longer wavelength folds (Figure 3).
 Folding in interbeddeed limestone and mudstone units involved bedding-parallel flexural slip, together with flexural flow and bedding-parallel stretching in fold limbs. Bedding-parallel flexural slip is indicated by the presence of laminated, bedding-parallel calcite veins. These are common locally in the Cavan Bluff Formation and thinly interbedded limestone-mudstone sequences within the Taemas Formation (e.g., Spirifer Yassensis, Bloomfield, Warroo Limestones). Minor flexural slip is also evident in some of the massive limestone members of the Taemas Formation (e.g., Currajong Limestone). Most limestone beds are internally unstrained to weakly strained, but locally have coarsely spaced, stylolitic solution cleavage. Bedding-parallel flexural flow occurred predominantly on fold limbs in mudstone, immature sandstone and marl layers, and was associated with weak to moderate cleavage development. Bedding-parallel stretching during folding of competent limestone layers is localized mainly on steeply dipping fold limbs and is indicated by the presence of abundant extension veins at a high angle to bedding.
 Fold growth in the Murrumbidgee Group was associated with reverse faulting. Low-displacement faults (<20 m slip) with strikes subparallel to fold trend are abundant. They are particularly common near fold hinges and indicate that fault growth was related to strain accommodation during folding of interbedded competent and incompetent units [see Ramsay, 1974]. Some faults probably also facilitated further crustal shortening after folds began to frictionally lock-up. Rare minor faults have late slickenfibers indicating local sinistral reactivation. Other than the high-displacement Warroo and Deakin–Devil's Pass Faults, there are few faults with displacements higher than about 20 m in the Murrumbidgee Group in the Taemas area.
 The age of crustal shortening at Taemas is poorly constrained. Possible conformable relationships between the Murrumbidgee Group and the Middle Devonian Hatchery Creek Formation, 30 km west of Taemas, indicate that regional deformation could have occurred during the Carboniferous Kanimblan Orogeny [Hood and Durney, 2002].
 Temperatures during deformation and associated vein formation in the Taemas area are constrained only approximately by calcite-quartz oxygen isotope geothermometry. The δ18O compositions quartz-calcite pairs from veins at various levels in the Murrumbidgee Group were analyzed (Figure 4). On the basis of the calcite-quartz geothermometer of Clayton and Kieffer , the data scatter closely around isothermal lines in the temperature range 130°–240°C. There are no consistent changes in relative isotopic temperatures vertically through the folded limestone sequence. The temperature constraints provided by the Sharp and Kirschner  calcite-quartz geothermometer are in the range 240°C–350°C, but most of this range is inconsistent with the presence of illite in pelitic units in the Murrumbidgee Group. The depth of the Murrumbidgee Group during regional deformation is poorly constrained. For temperatures in the vicinity of 150°C–250°C, and for reasonable geothermal gradients, depths may have been as much as 5 to 8 km.
3. Vein System at Taemas
 Vein development in the Taemas area is related to low-displacement faulting and extension fracturing during fold growth and tightening. Veins in limestones are composed predominantly of calcite, although quartz and dolomite are present locally. Minor fluorite and barite have a limited occurrence. Fluid inclusions, which are locally abundant in vein quartz, typically are water-rich, two phase (H2O liquid + vapor) inclusions.
 Veins are common in most limestone units in much of the northern half of the Taemas synclinorial zone (Figure 2). An exception in this area is the thick and massive Receptaculites Limestone in which veins are much less abundant. Veins are also much less common in the southern part of the synclinorium. They also are rare in the siliciclastic Majurgong Formation throughout the region. Quartz veins are locally developed near the top of the Sugarloaf Creek Formation in the Tate's Straight area (63200E, 30150N).
 Several structural controls on vein development and distribution are recognized (Figure 5). Bedding-parallel laminated veins are up to 30 cm thick and consist of coarse-grained, calcite-rich laminae 1–20 mm thick; these alternate with thinner septa of layer silicate-rich material (Figure 6a). Laminated veins are most abundant in thinly interbedded limestone and mudstone sequences in the northern half of the Taemas synclinorium. Well-developed slickenfiber lineations in laminated veins plunge at a high angle to the local fold axis, as expected for flexural slip during folding. Crack-seal inclusion bands in calcite laminae are spaced 0.1 to1 mm apart and have geometries indicating vein growth during reverse slip (Figure 6b). The crack-seal textures indicate that formation of individual laminated veins involved thousands of microslip and sealing events during fold growth.
 Some laminated veins are overprinted by weak axial surface cleavage and by decimeter-scale, open, parasitic folds, indicating that they initiated prior to, or early during fold growth. In other cases, laminated veins thicken toward fold hinges, forming small saddle reef veins which must have been active during fold tightening. Less common, lenticular, bedding-parallel veins have fibrous to elongate internal textures at a high angle to vein walls. Their lenticular nature and lack of slickenfibers indicates formation by interlayer dilation, rather than by bedding-parallel slip.
 Fault-related veins occur in nearly all low-displacement, bedding-discordant reverse fault zones (Figure 6c). Fault-fill veins typically have coarse-grained, massive internal textures. However, coarsely laminated or fibrous textures are developed locally. Matrix-supported breccias occur in some fault veins. Although fault veins are usually less than about 15 cm thick, they can be up to 1.5 m thick within dilatant jogs. One low-displacement fault zone near Duffy's Point (65300E, 32600N) contains texturally distinctive banded, crustiform textures and calcite-cemented breccia (Figure 6d). This fault vein postdates adjacent, fold-related veins. Fault-related extension veins are usually abundant for up to a meter or so around bedding-discordant faults (Figure 6c). They are typically gently to moderately dipping, indicating formation during reverse faulting. The veins generally overprint the regional axial surface cleavage, although some veins are gently folded. This indicates that fault slip and associated extension veining occurred mainly late during folding. Extension veins are seldom more than 10 cm thick and are continuous along strike for distances up to a meter or so. They contain various internal textures, including fibrous to elongate grain shapes formed by antitaxial, syntaxial, and stretched crystal growth processes. Massive, coarse-grained, equigranular textures tend to be dominant in veins thicker than about 2 cm, even though vein margins can have fibrous to elongate textures. Numerous wall rock inclusion bands are locally present in hybrid antitaxial/stretched crystal veins. The inclusion bands commonly are up to 20 μm thick, separated by up to 100 μm of vein fill, and are subparallel to the vein walls. The inclusion bands faithfully mimic the irregularities in the vein/wall rock interface and are interpreted to have formed by repeated crack-seal processes. Fiber-stepping textures are commonly present in fibrous, stretched crystal veins, and are also indicative of crack-seal processes [Cox et al., 1987].
 Small, irregularly oriented extension veins occur in several places in the immediate footwall of the high-displacement Warroo Fault, but appear to be rare along most of this structure. Similarly, veins are uncommon along the several moderate displacement (20–300 m) reverse faults in the region.
 Fold-related extension veins are particularly abundant within limestone beds which are interbedded with mudstone on steeply dipping fold limbs (Figure 6e). These veins typically are moderately to gently dipping, up to 10 cm thick, and up to a meter or so long. They occur in ladder-like arrays, as well as conjugate, en echelon arrays. Formation of these veins is related to layer-parallel stretching and incipient boudinage of competent units sandwiched between more ductile mudstone layers during fold tightening. These veins can both predate and overprint stylolitic axial surface cleavage.
 Arrays of en echelon veins, in which the arrays are parallel to bedding and individual veins are inclined at 30°–50° to bedding, are present locally on fold limbs in impure limestone beds (Figure 5). This geometry indicates vein formation was associated with bedding-parallel shear during fold growth by flexural flow in moderately incompetent beds.
4. Stable Isotopes
 Carbon and oxygen isotope compositions were analyzed in 328 calcite samples. Two hundred and eighty-one samples were from calcite veins, 39 samples were from limestone wall rock less than 10 mm from vein margins (“proximal” samples), and 8 samples were from limestone remote from veins (“distal” samples, >50 m from significant vein development). The oxygen isotope compositions of 7 quartz samples from calcite-quartz veins in the Murrumbidgee Group were also analyzed. Additional oxygen isotope analyses were conducted on 3 quartz veins from the Sugarloaf Creek Formation, and on one quartz vein from the Majurgong Formation.
 Calcite samples selected for analysis were typically less than 1–2 mm3 in volume. These were crushed and ∼200 μg of powder was separated for analysis. The measurements were made in a Finnigan MAT 251 mass spectrometer with a Kiel microcarbonate device using 105% phosphoric acid at 90°C for 12 min. 18O compositions are given relative to VSMOW; 13C analyses are relative to VPDB. The external precision (1σ standard deviation) for isotopic analyses of calcite is <0.06‰ for δ18O and <0.04‰ for δ13C; this is estimated from results for the NBS-19 and NBS-18 standards run concurrently. Standards were analyzed after every 8 samples. δ18O analyses on quartz (10–20 mg) were done by laser fluorination and have a 1σ analytical uncertainty <0.1‰.
4.1. Distal Host Rocks
 Limestones which are remote from significant vein development have 18O and 13C compositions which are tightly clustered (Figure 7a). The δ18O ranges between 23‰ and 25‰, and δ13C ranges between −0.5‰ and 3‰.
4.2. Proximal Host Rocks
 Limestone immediately adjacent to vein margins has a broader distribution of δ18O compositions than the distal limestones (Figure 7b). The δ18O compositions of the former range from values typical of distal host rocks, down to 14‰. 18O depleted zones around veins are typically only a few centimeters wide, except near the base of the Murrumbidgee Group where 18O depleted zones locally are up to several meters wide. The δ13C compositions of proximal host rocks are similar to those of distal host rocks.
4.3. Vein Calcite
 Calcite in veins displays a much broader spread of 18O and 13C compositions than both the proximal and distal host rocks (Figure 7c). Several trends for variation of 18O and 13C are recognizable. The predominant trend involves 18O depletion relative to unaltered host rocks. Maximum values of δ18O in veins are the same as those of distal limestones. Minimum values are as low as −2‰. 13C compositions remain similar to those of distal host rock.
 Departures from the main 18O depletion trend are restricted to a few analyzed veins. Meter-scale dilatant step overs in two laminated, bedding-parallel veins (C-2, TS-04-102) in the upper part of the Cavan Bluff Formation in the Tate's Straight area (63325E, 30650N) have marked 13C depletion relative to other nearby veins (Figure 7c). Here, δ13C has a minimum of nearly −11‰. However, the moderately depleted δ18O compositions of these veins are similar to those of the more typical veins in this area. Extension vein CL-12, from the Spirifer Yassensis Limestone near Shark's Mouth (65600E, 29300N), also contains a zone with strong 13C depletion relative to the texturally identical adjacent parts of the vein (Figure 7c).
4.4. Spatial Variations in Stable Isotope Compositions
 At any one stratigraphic level in the Murrumbidgee Group, there are modest variations in the oxygen isotope compositions of veins and isotopically altered proximal host rocks. However, there are substantial variations related to structural position in the Taemas synclinorium. These variations are illustrated by examples from (1) the Tate's Straight area in the Cavan Bluff Formation near the base of the Murrumbidgee Group on the western side of the synclinorium, (2) the Shark's Mouth area within the Spirifer Yassensis and Currajong Limestones in the middle of the Taemas synclinorium, and (3) the area immediately west of the Warroo Fault around Warroo Creek (67000E, 30500N) in the upper part of the Murrumbidgee Group on the eastern side of the Taemas synclinorium.
 In the Tate's Straight area, proximal altered limestone is markedly 18O depleted relative to unaltered limestone (Figure 8a). The δ18O ranges between 21‰ and 15‰. Although veins typically have δ18O in the range 19‰ to 14‰, several veins have δ18O as low as 1‰. Calcite in veins in an inlier of the Cavan Bluff Formation in the central part of the Taemas Peninsula has similarly low δ18O compositions. In contrast, at the same stratigraphic level in the Cavan area, in the southeastern part of the synclinorium, vein calcite usually is less 18O depleted than in the Cavan Bluff Formation at Tate's Straight. Apart from the rare veins with strongly depleted 13C compositions (C-2 and TS-04-102), δ13C values in vein calcite in the Tate's Straight area are typically in the range −3‰ to 3‰. This range is up to 2‰ lighter than δ13C values in the distal limestones.
 In the Spirifer Yassensis and Currajong limestones in the Shark's Mouth area, veins and proximal host rocks are much less 18O depleted than is the case lower in the stratigraphy at Tate's Straight (Figure 8b). The δ18O in proximal host rocks ranges from undepleted values of 24‰, down to 21‰. Calcite in veins mostly has δ18O in the range 19.5‰ to 23.5‰. The lowest δ18O value, 15.7‰, occurs in sinistral strike-slip slickenfibers recording the youngest part of the slip history in a low-displacement fault with an earlier reverse slip history (SM-74). The δ13C values of proximal host rocks and most veins in the Shark's Mouth area are clustered tightly between 0‰ and 1.5‰ (Figure 8b). Only vein CL-12 exhibits a substantial depletion in 13C (minimum value of −8‰).
 There is considerable variability in oxygen isotope compositions in veins and proximal host rocks in the Murrumbidgee Group adjacent to the Warroo–Devil's Pass–Deakin Fault system (Figure 8c). In several sites less than about 5 m west of the Warroo Fault, veins have δ18O compositions in the range 4–10‰. Although the associated alteration haloes have δ18O compositions as low as 18‰, several highly 18O-depleted veins have proximal host rocks with δ18O compositions similar to those of unaltered limestone. More than 10 m west of the fault, most veins and proximal alteration zones have 18O compositions within 1–2‰ of those of unaltered limestone. The δ13C compositions of all analyzed veins in this area are similar to those of distal limestones and range from −1‰ to 2‰.
 Immediately adjacent to the southern termination of the Devil's Pass Fault, and nearby parts of the Deakin Fault in the Cavan area, 18O compositions of calcite veins are also strongly depleted. Very low δ18O compositions also occur in calcite veins in two minor faults up to 200 m west of the Warroo Fault. The texturally distinctive, crustiform calcite fault fill in a moderately SW dipping fault near Duffy's Point has δ18O in the range −2‰ to 4‰. The other minor fault with low δ18O calcite (4.2‰) occurs 100 m west of the Warroo Fault and has massive white calcite fill.
4.5. Vertical Variations in 18O in the Murrumbidgee Group
 Isotopically unaltered limestones throughout the Murrumbidgee Group have nearly uniform O isotope compositions which are clustered at 24.0 ± 1.2‰ (Figure 9a). However, excluding the 18O depletion zone adjacent to parts of the Warroo Fault, there is a systematic increase in δ18O of vein calcite and proximal limestone upward though the Murrumbidgee Group (Figures 9b and 9c). Vein δ18O compositions increase sharply upward within the Cavan Bluff Formation, then exhibit a more gradual increase with increasing stratigraphic height. Above an isotopic front, defined by the zone of steep increase in δ18O, there is typically a scatter of approximately 4‰ in δ18O compositions of vein calcite at any one stratigraphic level. Oxygen isotope compositions of vein quartz in the Murrumbidgee Group also have an upward enrichment in 18O, consistent with the vertical changes in oxygen isotope compositions of vein calcite (Figure 9d).
4.6. The 18O and 13C Relationships Between Veins and Proximal Wall Rocks
 C/O isotopic compositions were measured for both vein calcite and the immediate wall rock for 40 samples. For the majority of vein-rock pairs, both δ18Ovein and δ18Orock are significantly 18O depleted relative to the distal wall rocks (Figure 10). In 35 cases the δ18O of vein calcite is lower than that of the proximal host rock. Although depletion ranges up to 19‰, in 29 samples the depletion is less than 4‰. In 4 cases only, the vein calcite δ18O is higher than the proximal wall rock by up to 4‰. In all of these examples, the proximal host rock is significantly 18O depleted relative to unaltered limestone, whereas the vein calcite has a 18O composition similar to that of unaltered limestone. There are no systematic relationships between δ13C of calcite veins and their immediate host rocks. Nearly all veins have δ13C compositions within 1‰ of that of the immediate host rock.
4.7. Within-Site Variations in δ18O and δ13C
 Overprinting relationships are well developed between veins formed at various stages within the folding history. The 18O compositions of these veins can be used to explore the evolution of fluid oxygen isotope chemistry during progressive deformation.
 Although changes in 18O with time among veins are usually small at any one stratigraphic level (<4‰), changes are most marked near the base of the Murrumbidgee Group. Significantly, both increases and decreases in vein δ18O with decreasing age are recorded during vein growth. For example, in the Cavan Bluff Formation along Tate's Straight, early, bedding-parallel veins tend to have δ18O compositions which are lighter than those of younger extension veins which cut across weak cleavage (Figure 11a). Veins associated with synfolding boudinage in this area also exhibit systematic changes in C and O isotopic compositions during progressive deformation. For example, a swarm of thin (<1 cm), calcite extension veins occurs on both sides of a 15 cm wide boudin separation zone which is filled by coarse-grained calcite. The extension veins record distributed layer-parallel stretching prior to onset of pervasive boudinage. The early syntaxial veins show a progression to heavier δ18O and δ13C values from the fibrous vein margin to the younger, massive vein core (Figure 11b). The later calcite fill in the boudin separation zone has even heavier δ18O, but the δ13C compositions are nearly unchanged relative to the youngest parts of the earlier veins.
 An example of decreasing δ18O with decreasing age is provided by vein TS-34 (white calcite) which has a lighter δ18O signature than TS-33 (grey calcite) which it crosscuts (Figure 11c). Such differences in stable isotope compositions among crosscutting veins indicate that early veins have not been isotopically reset during subsequent vein growth. Another example of decreasing δ18O with decreasing age occurs higher in the stratigraphy near Shark's Mouth (65925E, 29030N). Here, early reverse and oblique slip calcite slickenfibers in a fault zone are overprinted by sinistral strike-slip slickenfibers which are 18O depleted by 4‰ relative to earlier calcite (Figure 11d).
4.8. Within-Vein Variations in δ18O and δ13C
 While many individual veins have minor internal variation in C/O isotopic compositions (<1‰), others exhibit substantial variations in δ18O and/or δ13C during vein growth. Again, these variations are most apparent near the base of the Murrumbidgee Group, where veins have markedly depleted 18O compositions relative to unaltered host rocks. For example, laminated vein TS-31 exhibits 3.3‰ variation in δ13C, but only 0.4‰ variation in δ18O (Figure 12a). Nearby laminated vein TS-15 has a δ18O range from 9.3‰ to 14.2‰, but negligible variation in δ13C. Although another laminated vein (TS-04-10) has minor, nonsystematic variations in δ18O between 17.0‰ and 17.3‰, δ13C has a systematic change between 0.6‰ and 1.1‰ from west to east across six laminae. In the same area, δ18O in syntaxial extension vein TS-34 decreases by over 2‰ between vein margin and younger vein interior (Figure 11c). A nearby extension vein (TS-28), with a coarse fibrous texture (fibers 1–3 mm in diameter, 10 cm long), has nearly uniform δ13C across the vein (<0.2‰ variation), but has fluctuations in δ18O larger than 3‰ (Figure 12b).
 Toward the middle of the Murrumbidgee Group veins usually exhibit smaller internal variations in isotopic compositions than occur below the isotopic front. For example, in CF-11, which is the thickest (30 cm wide) bedding-parallel vein found in the region, the total range in δ18O is less than 0.3‰ and δ13C has a range of only 0.2‰. Similarly, fibrous fault vein SM-500 has a δ13C variation less than 0.04‰ over 10 cm. The variation in δ18O is less than 0.2‰ (Figure 12c). Fibrous extension vein CL-12, from the Spirifer Yassensis Limestone, is unusual. In this case, vein calcite has a relatively uniform C/O isotopic composition over much of the vein width. However, within a 2 cm wide zone, there is a strong coupled variation in C/O isotope compositions. Here 13C is depleted by up to 7‰ relative to the rest of the vein, and 18O is enriched by as much as 3.5‰ relative to the rest of the vein (Figure 12d).
 In the low-displacement fault zone 200 m west of the Warroo Fault near Duffy's Point, crustiform grey and white calcite fault fill exhibits repeated fluctuations in C/O isotope compositions with decreasing age (Figure 13). There is over 5‰ variation in δ18O, and up to 1‰ variation in δ13C. Although variations in δ13C are decoupled from δ18O fluctuations and close to rock-buffered, 18O is depleted by up to 25‰ relative to the host rocks. The grey calcite bands have δ18O less than 1.9‰, whereas the white calcite bands have δ18O greater than 1.9‰.
5.1. General Aspects
 Fracture growth and vein formation at Taemas during deformation of the Murrumbidgee Group was related to strain accommodation processes during folding. These processes include bedding-parallel flexural slip, bedding-parallel stretching and flexural flow on fold limbs, low-displacement reverse faulting and mechanically related extension fracturing. Flexural slip commenced early during fold growth, and remained active during fold tightening. Most fault-related extension veins overprint cleavage and are interpreted to have formed late during fold growth. Veins associated with layer-parallel stretching and flexural flow formed during fold growth.
 Although individual veins probably grew over only a fraction of the crustal shortening history, opening of fracture arrays enhanced permeability and facilitated fluid migration during most of the deformation history. Stable isotope compositions of vein calcite therefore record evolution of a fluid flow system over an extended period of crustal deformation. Shortening rates associated with folding in orogens are poorly constrained, but if strain rates were in the range 10−14 s−1 to 10−15 s−1 [e.g., Pfiffner and Ramsay, 1982; Mueller et al., 2000], then vein growth at Taemas potentially records 1.5 × 106 to 1.5 × 107 years of fluid flow history.
 The abundance of gently dipping extension veins, as well as the occurrence of severely misoriented reverse faults and bedding-parallel slip surfaces is significant. The presence of these structures requires that vein growth occurred in a fluid pressure regime with λv > 1 (i.e., supralithostatic fluid pressures) during substantial parts of the deformation history [see Sibson, 1985]. Such overpressured regimes provide a strong upward driving force for fluid migration [Cox, 2005].
 The structures which localized fluid flow are interpreted to have formed a mesh-like network of transiently permeable fluid pathways comprising bedding-discordant faults, zones of flexural slip, and fold-related extension fracture arrays. However, the abundance and connectivity of low-displacement faults and veins in most areas is insufficient to provide a fully hydraulically interconnected fracture network. Accordingly, fracture arrays must have been linked hydraulically by localized fluid flow through the intervening rock mass.
 The high-displacement Warroo fault is also implicated as a fluid pathway. However, the low abundance of veins associated with this structure, as well as with the several moderate displacement (20–200 m slip) faults in the area, indicates that these structures probably contained pore fluids at lower average pore fluid factors than the low-displacement faults and fold-related vein arrays. Similarly, the abundance of all types of veins in the central and northern part of the Taemas synclinorium, relative to their lower abundance further south, requires that average pore fluid factors in the southern part of the synclinorium were lower than to the north.
 The occurrence of laminated textures in bedding-parallel veins and some bedding-discordant faults, together with the widespread presence of crack-seal textures in bedding-parallel veins and extension veins, indicates that vein growth involved repeated episodes of fracture-controlled permeability enhancement and permeability destruction. For example, the spacing of crack-seal inclusion bands (0.1–1 mm) in laminated veins indicates that net slips of several meters during bedding-parallel flexural slip typically involved thousands of slip, permeability enhancement and sealing events. The slip magnitudes are consistent with slip events being Mw < 4 ruptures with rupture areas up to approximately 100 m × 100 m [Wells and Coppersmith, 1994]. The spacing of crack-seal inclusion bands in extension veins indicates that growth of these structures typically involved hundreds of crack-seal cycles. Repeated switching between high-permeability states and low-permeability states, hundreds to thousands of times, along fracture-controlled fluid pathways, requires that flow through individual veins was episodic.
5.2. Oxygen Isotope Systematics
 Throughout the Murrumbidgee Group, limestones remote from veins have nearly uniform 18O and 13C compositions which are typical of Lower Devonian marine limestones [Veizer et al., 1999]. The 18O compositions of vein calcite tend to be depleted relative to the compositions of unaltered limestones and indicate that veins are not in isotopic equilibrium with the limestone host rocks. The most extreme 18O depletions occur near the base of the Murrumbidgee Group, as well as immediately west of some segments of the Warroo Fault. The amount of 18O depletion systematically diminishes upward from the base of the Murrumbidgee Group, so that veins near the base of the sequence have 18O compositions which are more fluid-buffered than those at structurally higher levels. Limestones within a few millimeters of vein margins typically have 18O compositions intermediate between those of unaltered limestone and the adjacent vein. These narrow, 18O-depleted alteration haloes record discharge of 18O-depleted fluid from veins.
 The systematic changes in oxygen isotope compositions of veins and immediate wall rocks upward through the Murrumbidgee Group, as well as the presence of a marked oxygen isotope alteration front approximately 25 m stratigraphically above the base of the limestone sequence, indicate that externally sourced fluids have migrated upward via fracture-controlled pathways during crustal shortening. The narrowness of the 18O depleted alteration envelope around veins, compared with the vertical extent of 18O depletion in vein networks, indicates that upward advection was much more rapid than lateral dispersion of fluids from fractures into wall rock.
 The vertical 18O depletion profile in veins is recorded over a structural thickness of approximately 2000 m in the Murrumbidgee Group. However, fluid fluxes are unlikely to have been uniform through the fracture system across the entire area. Oxygen isotope compositions of veins near the base of the Murrumbidgee Group in the Cavan area are notably less 18O depleted than in the northern half of the Taemas synclinorium. This indicates significantly lower fluid flux in the southern part of the synclinorium than at comparable structural levels to the north. This is consistent with the paucity of vein development and inferred lower, time-averaged pore fluid factors in the southern part of the synclinorium. Accordingly, the fracture network in the southern area was probably less well connected to the external fluid reservoir than the network further north in the Taemas synclinorium. The narrow 18O depletion zone, and paucity of vein development adjacent to parts of the Warroo Fault, also indicates substantially less fluid migration out of this structure than was associated with upflow from the base of the Murrumbidgee Group.
 Assuming calcite-fluid isotopic equilibrium during vein growth, temperatures of vein formation of approximately 200°C, and that temperatures remained approximately constant during operation of the flow system, the most 18O-depleted veins indicate the O isotope composition of fluid entering the base of the limestone sequence was approximately −7‰ [O'Neil et al., 1969; Friedman and O'Neil, 1977]. Fluids entering the Murrumbidgee Group from the Warroo Fault had similar 18O compositions.
5.3. Reactive Transport Modeling
 If calcite deposition in veins was a near-equilibrium process, the δ18O composition of vein calcite monitors the change in δ18O of the pore fluid in response to isotopic exchange between fluid and wall rock along fluid pathways during the evolution of the flow system. The δ18O compositions of the proximal wall rock monitor the extent of reaction of the fluid with the wall rock, the kinetic controls on exchange, and the extent of fluid dispersion from fractures into the wall rock.
 As externally derived fluids progressively infiltrate a percolation network, an isotopic alteration front migrates through the system. Quantitative interpretations of isotopic alteration fronts are based on transport theory [e.g., Bickle and McKenzie, 1987; Lassey and Blattner, 1988; Bowman et al., 1994; Baker and Spiegelman, 1995; Abart and Sperb, 1997; Abart and Pozzorini, 2000]. There are two main approaches to analysis of reactive transport in an advective flow system [Abart and Pozzorini, 2000]. First, if rock-fluid exchange rates are fast relative to advective transport rates, rock-fluid equilibrium is maintained and the position of the stable isotope front records the extent of fluid infiltration. Distension of the front during coupled transport and equilibrium isotopic exchange is related to molecular diffusion and hydrodynamic dispersion processes [Ogata, 1970; Ingebritsen et al., 2006]. Alternatively, if isotopic exchange between fluid and rock is slow relative to advective transport rates, the system can deviate from local equilibrium and rock-fluid stable isotope exchange is kinetically controlled. In this case the position of the isotopic front is retarded relative to the infiltration front and distension of the isotopic front is related to kinetically controlled isotope exchange and hydrodynamic dispersion [Abart and Sperb, 1997].
 The length scales and temperatures of isotopic transport in the Murrumbidgee Group require that transport was dominated by advective flow [Cole and Ohmoto, 1986]. The preservation of millimeter-scale isotopic variations in veins confirms that molecular diffusion was unimportant at the scales of interest. Furthermore, the observation that calcite in the immediate wall rock adjacent to veins is not in isotopic equilibrium with the vein calcite indicates that mineral-fluid isotopic exchange was kinetically controlled, rather than the result of equilibrium exchange and diffusive processes [Knoop et al., 2002].
Lassey and Blattner  present numerical solutions to the transport equations for coupled, purely advective transport and kinetically controlled fluid-mineral oxygen isotope exchange in a monomineralic isothermal system. Exchange is governed by two partial differential equations. Mass balance is described by
and kinetic exchange is given by
where τ is dimensionless time, Φ′ “oxygen porosity,” is the ratio of oxygen in pore-bound fluid to that in the rock. Z is dimensionless position in the aquifer, ND is the Damköhler number, δ18Owater is the isotopic composition of the fluid, and δ18Orock is the isotopic composition of the rock. The variable α is the isotopic fractionation between the rock and the fluid, and Δ is the per mil departure of α from 1, i.e., 103(α-1). In the present study, the numerical solutions to these equations were obtained by evaluation of the K functions in equations (5a) and (5b) of Lassey and Blattner  using Mathematica™.
 For flow through a total infiltrated aquifer length L, over time t, with flow rate q, dimensionless time, τ is equal to qt/L. L is defined as the position of the infiltration front at dimensionless time, τ = 1. Dimensionless position Z in the aquifer is equal to z/L, where z is the distance from the fluid inlet site and 0 ≤ Z ≤ τ. The Damköhler number can be expressed as
where κ is the rate constant for 18O exchange between water and rock, i.e., for the reaction
ND is a measure of the relative rates of advective transport and mineral-fluid isotopic exchange. For small ND, mineral-fluid isotopic exchange is slow relative to rates of advective fluid transport (for nonreactive flow, ND = 0). Large ND indicates that the rate of isotopic exchange is fast relative to advective transport (for instantaneous reaction, ND = ∞).
Figure 14 illustrates the main features of the evolution of isotopic profiles along transport paths for veins formed during purely kinetically controlled isotopic exchange and continuous, isothermal, advective transport. At initial time t0, the isotopic profile is a step function located at the fluid inlet site to the reactive rock mass. The initially steep isotopic front migrates in the fluid flow direction but its position becomes retarded relative to the position of the infiltration front. The amount of distension of the isotopic front increases with slower isotope exchange kinetics (i.e., lower Damköhler numbers). A notable consequence of continuous migration of an isotopic front through a rock mass is that the isotopic composition of minerals accreting in veins should change monotonically from a rock-buffered composition to a more fluid-buffered composition with decreasing age (Figure 14c).
 The isotopic front (F) is located approximately at
Comparison of the extent of displacement and distension of an isotopic alteration front with reactive transport models allows estimation of the Damköhler number, as well as the time-integrated fluid flux (TIFF) during infiltration and reaction [Abart and Pozzorini, 2000; Knoop et al., 2002].
 Reactive transport modeling requires an estimate of the value of ϕ, the “effective porosity” of the flow path. The time-integrated fluid flux, Q, is given by
The distance to the infiltration front L is given by
where zF is the distance between the fluid inlet site and the isotopic front at τ = 1. From (5), (6) and (7) it is clear that L is dependent on Φ′ and ϕ and it can be shown that
for α ≈ 1. For calcite, Φ′ = 1.85ϕ if ϕ ≪ 1. Accordingly, the TIFF can be expressed as
Note that although the position of the isotopic front is critical to the determination of the TIFF, the assumed value of ϕ does not have a significant effect on the estimate of the TIFF, provided ϕ ≪ 1. On this basis, it is assumed that ϕ ≈ 0.05. This value is a reasonable approximation for the macroscopic value of effective porosity for flow paths involving a combination of low-displacement faults, extension fracture arrays and intervening rock mass connecting fractured zones.
 Assuming the vertical variation in 18O in veins and host rocks at Taemas approaches kinetically controlled rock-fluid stable isotope exchange in an advective, relatively nondispersive, isothermal flow regime, δ18O compositions of veins and their immediate host rock were calculated for τ = 1, for 0 ≤ Z ≤ 1, and for various values of ND, 0.1 < ND < 100 (Figure 15). Assuming the most 18O-depleted calcite vein compositions at the base of the Murrumbidgee Group record the O isotope composition of the inlet fluid, we use an initial fluid δ18O composition of −7‰. The δ18O composition of the unaltered limestone is taken as 23‰, and the fluid flow is assumed to occur at 200°C.
 To compare the modeled isotopic profiles with the measured data, the reactive path length from the inferred fluid inlet at the base of the Murrumbidgee Group to the sites of vein formation for each sample must be estimated. This cannot be well constrained because the structural thickness of the Murrumbidgee Group progressively increased during crustal shortening. Accordingly, we assume a “model” path length double that of the stratigraphic height above the base of the Murrumbidgee Group. This approximates the vertical distance between the base of the limestone sequence and vein sites at the end of folding, but does not account for potential path tortuosity, or changes in path length due to gradual thickening of the sequence during folding. In determining the model path length, the thickness of the Majurgong Formation has been reduced to zero because this siliciclastic unit is largely devoid of limestone, and apparently was not a reactive part of the fluid pathway (Figure 15).
 The scatter in vein δ18O values at any one model path length likely reflects a number of causes. Clearly, some of the data scatter, between different sites at the one stratigraphic level, is an artifact of using stratigraphic level as an estimator of reactive path length. Despite this, variations in O isotope compositions comparable to this scatter occur among groups of veins in individual outcrops. The causes of such “within-site” variations are discussed in the next section.
 From Figure 15 the isotopic front is located at ZF = 0.09 at τ = 1. Although the isotopic front is approximately 50 m structurally above the base of the Murrumbidgee Group, comparison between the shapes of the measured and modeled isotopic profiles (Figure 15) suggest that the front must be approximately 100 to 200 m along the fluid pathway from the notional fluid inlet site. This may indicate that the reactive flow path involved a major component of bedding-parallel flow near the base of the limestone sequence.
 Assuming the reactive path length between the fluid inlet and the isotopic front, zF, is 150 m (±50 m), and using relationship (8), the total infiltrated path length is estimated as 1700 m (±550 m). This is consistent with the infiltration front penetrating to the upper parts of the Taemas Formation. Note, however, that as ZF is a function of Φ′ (equations (5) and (6)), changes in the assumed value of ϕ lead to proportional changes in the estimate of L.
 Comparison of vein δ18O results with the reactive transport models (Figure 15a) indicates Damköhler numbers approximately in the range 1–10 for ϕ ≈ 0.05. Decreasing the assumed value of ϕ leads to corresponding increases in the estimated value of ND. The δ18O data for the proximal altered wall rocks are also consistent with Damköhler numbers in the same range (Figure 15b).
Equation (10) indicates that for 100 m < zF < 200 m, the model TIFF along fluid pathways is in the range 60 to 120 m3 m−2 through Z = 0. This is equivalent to 280–560 mol H2O cm−2 at a temperature of 200°C and pore fluid pressure of 100 MPa.
5.4. Significance of Within-Vein and Within-Site Variations in O Isotope Compositions With Time
 During continuous flow and reaction, the isotopic front should migrate upward within the vein system producing a monotonic decrease in vein δ18O with time as total fluid flux increases and as compositions of proximal wall rocks along fluid pathways become progressively more fluid buffered (Figure 14). However, two observations are inconsistent with a simple continuous flow and reaction model. First, the observed variations in δ18O within veins, and among veins at any one outcrop, are small (<4‰) relative to the total vertical variation in vein δ18O (24‰) across the entire vein field. This is regardless of whether veins formed early or late during fold growth. This requires that the isotopic front has not migrated a significant vertical distance during a protracted history of folding, faulting and fluid flow. Secondly, there are not uniform trends to changes in δ18O with time, either in individual veins, or among overprinting veins at the one site. Even though δ18O is nearly constant during the growth of many veins, in other cases 18O exhibits depletion with time, or enrichment with time. Some veins (e.g., Figure 12b) display episodic fluctuations in δ18O compositions of calcite during progressive vein growth.
 During continuous flow and reaction, two factors can cause departures from the predicted rates of decrease in δ18O with time. First, if early generations of calcite on vein walls effectively “armour” veins, fluid-rock reaction can be inhibited [Cathles, 1991]. In this case, the rate of upward migration of the isotopic front within veins would be enhanced, relative to that predicted by steady state transport models. This leads to monotonic decrease in vein calcite δ18O with time at rates faster than predicted by simple continuous flow models. Second, during thickening of the stratigraphic sequence by folding the reactive path length should increase. This promotes more rock-buffered fluid compositions at any one point in the flow system, and will lead to vein δ18O decreasing at rates less than predicted by constant path length models. Neither of these effects is consistent with the observed variable changes in vein δ18O with time.
 The changes in δ18O in veins with time, particularly below the isotopic front, support the earlier conclusion, based on the internal structures of veins, that flow was episodic. In seismogenic, upper crustal regimes, faults and related extension fractures tend to fail episodically and promote transitory permeability enhancement [Cox et al., 2001]. In this scenario, fluid pressure cycling and episodic migration of fluid pressure pulses and associated fluid batches up through fracture-controlled percolation networks is expected [Cox, 2005]. At Taemas, fluid migration episodes could be related to fault rupture events intermittently breaching an overpressured fluid reservoir at depth. In this case each rupture event would drain a batch of fluid from the reservoir and generate a fluid pressure wave which then migrates up through the fracture network, driving a transitory cascade of fracture reactivation and growth at the fluid pressure front.
 In such a fracture-controlled, discontinuous flow regime, the percolation network comprises a mesh of transiently hydraulically interconnected, low-displacement faults and extension fractures. Because individual veins along fluid pathways repeatedly cycle between high-permeability states and low-permeability states, the instantaneous hydraulic connectivity between components of the fault/vein network will evolve dynamically as successive failure events migrate around the fracture system. Such complex behavior will cause repeated changes to fluid pathways and reactive path lengths during successive rupture cycles (Figure 16). This dynamic evolution of pathways will influence the bulk permeability of the network, driving changes in hydraulic gradients across the network, and causing variations in the timescales on which pathways become sealed during the migration of successive fluid batches up through the percolation network. Consequent changes in flow rates and fluid fluxes, as well as reactive path lengths, between successive pulses of fluid migration can all influence the evolution of vein δ18O with time (Figure 17). As flow repeatedly shifts between different components of the percolation network, changes in effective porosity, fracture density, or fracture surface roughness along fluid conduits, are likely to influence the efficiency of isotopic exchange at constant flow rate. Additionally, if pathways episodically switch between different parts of the fracture network, changes in the mean isotopic compositions of reactive wall rocks along transitory pathways can also influence the isotopic evolution of fluids.
 At constant flow rate, effects on vein δ18O due to changes in reactive path length will be most significant near the isotopic front where the vertical δ18O gradient is highest. For example, from the model data in Figure 15a, for Z = 0.1 and ND = 4, a 20% decrease in path length can lead to a 5‰ decrease in vein δ18O. However, above the isotopic front, for example at Z = 0.6, a 20% decrease in reactive path length decreases vein δ18O by less than 1‰.
 Changes in flow rates at constant path length and reaction rate influence the effectiveness of isotopic exchange and may particularly influence isotopic evolution during vein growth in discontinuous flow regimes (Figure 17). Faster flow rates allow the isotopic front to penetrate to higher structural levels and produce more fluid-buffered vein compositions. As with changes in reactive path length, the effects of changes to flow rates are greatest near the isotopic front. For example, at Z = 0.1 and ND = 4, a doubling of flow rate decreases δ18O of veins by approximately 6‰. On this basis, the several per mil fluctuations in vein δ18O found in, and among, veins near the isotopic front (e.g., Figures 11b, 11c, 12a, and 12b) could equally well be associated with modest changes in reactive path length or flow rates. The 4‰ decrease in δ18O with time found in a fault above the isotopic front at Z ≈ 0.3 (Figure 11d), could, for example, be associated with a threefold increase in flow rate, or a 60% decrease in reactive path length. However, well above the isotopic front, very large changes in flow rates and/or reactive path length are necessary to drive fluctuations in vein δ18O of several per mil. Increasing duration of flow events at constant fluid flow rate allows fluid infiltration to higher structural levels and has effects similar to increased flow rates. As the original δ18O compositions of limestones in the Murrumbidgee Group range from approximately 23‰ to 25‰, variations up to 2‰ in δ18O of vein calcite also could result from variations in O isotope compositions of limestones along changing flow paths.
 At constant flow rate, an increase in the effective reactive surface area in a fluid pathway (e.g., via higher fracture density, or larger component of flow path involving porous flow through limestone) increases the effective rate of isotopic exchange. For example, with an initial ND = 10 near the isotopic front at Z = 0.1, a tenfold decrease in the reaction rate (i.e., “effective ND” decreases from 10 to 1) reduces δ18O in veins by 6‰, whereas at Z = 0.4, the decrease is substantially less (Figure 15a). Notably, near the fluid inlet, and for some values of ND, a decrease in reaction rate locally has the effect of increasing vein δ18O relative to that associated with high reaction rates (Figure 15a).
 The presence of quartz- and illite/muscovite-bearing shales along flow paths (e.g., especially within the Spirifer Yassensis Limestone and Bloomfield Limestone) is unlikely to have substantially influenced the O isotope chemistry of fluids migrating along pathways. The apparent lack of influence of the Majurgong Formation on vein δ18O indicates that silicates had only a minor influence on stable isotope exchange between fluids and rocks at the temperatures in the Murrumbidgee Group during flow. Accordingly, the 4‰ variation in δ18O found in veins well above the isotopic front could result from the combined effects of δ18O variations in limestone wall rock along fluid pathways, and fluctuations in path length, flow rates, reaction rates, and duration of flow in successive flow events.
5.5. Implications of Episodic Flow for Estimates of Time-Integrated Fluid Flux
 The recognition that veins accreted episodically during numerous successive pulses of fluid flow has implications for interpreting the time-integrated fluid flux in the Taemas vein field. The oxygen isotope composition of individual crack-seal growth increments in calcite veins is interpreted to reflect equilibrium between the fluid and precipitating calcite at their position in the flow network during the passage of a batch of fluid through that site. Successive crack-seal increments accordingly reflect episodic calcite growth during migration of successive fluid batches. In this case, the vertical oxygen isotope profile recorded for small volumes of vein calcite (Figure 15a) provides a snapshot of vertical variations in vein δ18O at the time those volumes accreted. A time-integrated fluid flux of approximately 102 m3 m−2, calculated on the basis of this data, accordingly estimates the mean fluid flux for individual fluid batches. Given that individual veins may record growth during hundreds to thousands of crack-seal cycles, and that veins formed at all stages of a protracted crustal shortening history, the time-integrated fluid flux over the active lifetime of the vein field is likely to be several orders of magnitude greater than the estimate for one cycle. Accordingly, the time-integrated fluid flux along fluid pathways is likely to have been at least 105 m3 m−2. Converting these estimates into net fluxes during fracture-controlled flow is problematic. However, using a conservative estimate that flow paths are 10 cm wide and have an across-strike spacing of 100 m over an area of 20 km2, net fluxes could be as high as 2 km3 of fluid.
5.6. C-H-O Fluid Compositions
 If calcite precipitating in veins was predominantly in exchange equilibrium with the pore fluid, the relative positions of the C isotope front and O isotope front in the Murrumbidgee Group reflects carbon and oxygen partitioning between solid and fluid phases in the veins, as well as the relative abundances of carbon and oxygen in the fluid [Abart and Pozzorini, 2000]. The observation that the C isotopic front is essentially at the base of the Murrumbidgee Group indicates X(CO2) of the fluid was very low and effectively buffered by the host rock sequence as soon as the fluid entered and reacted with the limestones. For an O isotope front located approximately 150 m from the fluid inlet sites in the limestone sequence, and a C isotope front occurring less than 30 m from the fluid inlet, the C isotope front is retarded relative to the O isotope front by a factor >5. As the C/O ratio in calcite is 1/3, the C/O ratio in the fluid typically must have been <1/15. This indicates X(CO2) < 0.08, in accord with the occurrence only of two phase, water-rich fluid inclusions in veins.
5.7. Fluid Reservoirs
 Most vein formation at Taemas was driven by upflow of near-lithostatic pressured fluids with an initial δ18O composition of approximately −7‰. Fluids with this composition are most likely either meteoric fluids or formation waters [Taylor, 1997]. Potential sources of overpressured formation waters during the folding of the Murrumbidgee Group include the underlying Black Range Group and the Silurian volcano-sedimentary sequences. A fluid source involving topographically driven recharge of meteoric waters, followed by overpressuring and upward leakage via fault/fracture networks, is possible, but requires large topographic relief to drive such a flow system. It is also possible that the Warroo-Deakin fault system was involved in driving the flow system. Large, episodically rupturing faults have potential to repeatedly drawn-down hydrostatic-pressured, surface-derived fluids after rupture events. During interseismic compaction of damage products in rupture zones, contained fluids can be pressurized, then expelled into the wall rocks [Sleep and Blanpied, 1992]. Over the lifetime of a major fault system, involving hundreds to thousands of rupture events this process can generate time-integrated fluxes up to tens of km3 on timescales of 104–105 years [Cox, 2005]. At Taemas, this model would require lateral fluid expulsion from the Warroo-Deakin fault system at depth beneath the Murrumbidgee Group prior to upward migration across the base of the limestone sequence.
5.8. Origins of δ13C Variations in Veins
 The C isotope compositions within most calcite veins are within the range exhibited by the unaltered limestone host rocks (−1‰ < δ13C < 3‰), and are interpreted to be essentially rock buffered. Accordingly, measured δ13C variations within the range −1‰ to 3‰ in individual veins could potentially reflect differences in C isotope compositions of reactive wall rocks along episodically changing flow paths, as was discussed for some within-vein variation in δ18O. This process would be particularly important for low ΣC fluids which would react rapidly with wall rocks along fluid pathways.
 Veins from two sites exhibit 13C compositions which lie well below those of unaltered host rocks. These veins have structural geometries and oxygen isotope compositions similar to those of nearby “isotopically normal” veins, and must have formed at the same time as the rest of the vein system. The 13C depletion at these sites could have been influenced either by the input fluid having a 13C-depleted composition, or by isotopic fractionation during deposition of calcite from initially low ΣC fluids. The second situation occurs only in low fluid:rock ratio regimes and results from preferential 13C partitioning into the calcite, leaving the fluid 13C depleted [Friedman and O'Neil, 1977; Abart and Pozzorini, 2000]. In this case, the precipitating calcite becomes progressively more 13C-depleted with time. For veins C-2 and TS-04-102, the low δ18O compositions are the same as those of more typical nearby veins, and indicate that the veins formed within a part of the hydrothermal system that was dominated by upflow of 18O-depleted fluid. Accordingly, the veins cannot have formed at low fluid:rock ratio, but rather had their δ13C composition governed by the presence of a low δ13C fluid species. Several other veins in the Tate's Straight area have 13C compositions which, although less depleted than C-2 and TS-04-102, are more depleted than unaltered limestones.
 In vein CL-12, the narrow zone of 13C depletion is coupled with an increase in δ18O to a rock-buffered value typical of unaltered limestone (Figure 12d). For 13C depletion due to isotopic fractionation in a low fluid:rock regime, changes in δ18O are unlikely [Abart and Pozzorini, 2000]. Accordingly, the 13C depleted zone is interpreted to reflect a transitory period of influx of low δ13C fluid having a rock-buffered 18O composition.
 A possible source of isotopically light carbon in the limestone sequence is from hydrocarbon-bearing reservoirs. However, isotopic exchange between CO2 and hydrocarbons, such CH4, is extremely slow at the temperatures of vein formation [Ohmoto, 1986]. Incorporation of light C into CO2, followed by production of calcite from HCO3− in solution, requires oxidation of light hydrocarbons via coupled reactions of the form
The most likely places for such fluid oxidation to occur are in the red bed mudstones and sandstones of the Majurgong Formation, immediately overlying the Cavan Bluff Formation, as well as in red bed mudstones in parts of the Sugarloaf Creek Formation underneath the Cavan Bluff Formation. The low 13C veins could therefore have formed from originally hydrocarbon-bearing fluids which subsequently became oxidized by fluid-rock reaction along their flow path. The scarcity, first of reduced mudstone in the red bed sequences and second of 13C-depleted veins, indicates only relatively minor volumes of hydrocarbons were oxidized and involved in calcite precipitation in veins. The isotopically light carbon could have been derived from a hydrocarbon-bearing fluid reservoir in the local limestone sequence. Alternatively, hydrocarbons could have been a component the low δ18O, externally sourced fluid.
5.9. Implications for Growth of Percolation Networks
 The fracture-controlled network of fluid pathways in the Taemas vein field forms a mesh of bedding-discordant faults, bedding-parallel faults, and various types of extension vein arrays. The regular upward increase in vein δ18O over a 20 km2 area indicates that the entire vein system formed a network of fluid pathways having high episodic connectivity with an external fluid reservoir at depth. The lack of rock-buffered veins near the base of the Murrumbidgee Group, especially at the earliest stages of deformation, indicates that vein growth did not commence until infiltration of externally derived fluid occurred. The hydraulic connectivity of the entire vein field to the external reservoir, from the earliest stages of its development, accordingly indicates that fracture growth occurred by an invasion percolation process, rather than by ordinary percolation (Figure 1b). In this model, the major factor repeatedly driving the rock mass to failure is interpreted to have been the migration of successive batches of near-lithostatic pressured fluid hydrofracturing its way upward through the network.
 In general, fluid flow in fracture systems is likely to be discontinuous in circumstances where overpressured fluid reservoirs have their seals episodically, and transiently, breached by fault ruptures. It is suggested that the consequent propagation of cascades of fluid-driven failure events at the leading edge of successive fluid pressure waves is likely to play a key role in generating and repeatedly reactivating fracture-controlled percolation networks. By episodically increasing pore fluid factors and driving failure, migration of fluid pressure waves can transiently decrease crustal strength, drive earthquake aftershocks or swarm activity, and accelerate deformation rates in high pore fluid factor domains above breached reservoirs. Recent seismic observations [Husen and Kissling, 2001; Haney et al., 2005; Kao et al., 2005] indicate that intermittent migration of overpressured fluid pulses, or fluid “burping,” through fracture systems, might be a common mode of fluid redistribution at various crustal levels.
 By integrating structural and microstructural observations with isotopic analysis of veins and wall rocks, it has been possible to explore the development of a long-lived, fracture-controlled fluid flow system which formed during fold growth in a limestone sequence at temperatures of approximately 200°C and depths greater than several kilometers.
 The evolution of fluid pathways was controlled by growth of mesh-like arrays of faults and extension fractures whose development was related to strain accommodation from early during folding until late in the folding history. Transiently supralithostatic fluid pressures were necessary for opening many fractures. Microstructures indicate vein formation typically was associated with repeated episodes of permeability enhancement and destruction. Vein formation therefore occurred in a discontinuous flow regime with episodic hydraulic connectivity between components of the fracture mesh and the fluid reservoir.
 Calcite veins record the evolution of oxygen isotope chemistry in the flow system in space and time within a crustal volume of approximately 40 km3. The Taemas vein field formed from externally derived, low δ18O fluids, having an evolved meteoric or formation water origin. Predominantly upward migration of fluids is indicated by systematic upward change in δ18O of vein calcite from fluid-buffered compositions at the base of the sequence, to rock-buffered compositions at higher structural levels. A sharp isotopic front is present in the vein system in the lower part of the Murrumbidgee Group. Modeling of kinetically controlled, reactive transport suggests that individual fluid pulses likely involved fluid fluxes up to approximately 102 m3 m−2 along fluid pathways. Over the lifetime of the flow system, total fluid flux along pathways could have been higher than 105 m3 m−2.
 Even though flow occurred over a protracted history of fold growth, the isotopic front did not migrate significantly during this period. This is consistent with a discontinuous flow regime in which numerous fluid batches migrated up through the fracture system, each independently reacting with the wall rocks. Discontinuous flow was associated potentially with fault rupture events repeatedly breaching an overpressured fluid reservoir beneath the Murrumbidgee Group. In this scenario, each failure event is interpreted to have initiated upward migration of a fluid pressure wave which drove a cascade of hydraulic extension and shear failure events as it migrated up through the Murrumbidgee Group. Systematic upward changes in vein δ18O in the Taemas vein field indicates that most of the vein system was episodically well connected to the external fluid reservoir, as expected for growth and repeated reactivation of the fracture-controlled flow network by fluid-driven invasion percolation processes. Individual flow episodes were probably terminated by draining the hydraulically accessible regions of the reservoir and by hydrothermal sealing of fluid pathways.
 In this discontinuous flow regime, changes in vein δ18O with time at individual sites in the flow system are interpreted to be related largely to changes in hydraulic connectivity between components of the fracture network, and consequent changes in reactive path length, rates of stable isotope exchange, flow rate, and net fluid flux during successive fluid pulses. Variations in O isotope compositions of reactive limestone wall rocks along episodically changing fluid pathways also may have influenced changes in vein δ18O.
 This research was supported by an ANU Faculty Research Grant. Stable isotope analyses of carbonates were facilitated by H. Scott-Gagan, J. Cowley, and J. Cali at the ANU Research School of Earth Sciences. Oxygen isotope analyses on quartz were conducted by K. Faure at the Institute of Geological and Nuclear Sciences, Lower Hutt, New Zealand. C. Nicholson and R. O'Leary are thanked for assistance in the field. R. Abart, S. Barker, J. Braun, G. Dipple, K. Lassey, S. Micklethwaite, and S. Miller provided valuable discussion. K. and S. Kilpatrick are thanked for very kind hospitality and permission to access the Taemas property. I. Peake, M. Smith, and F. Patmore kindly provided access to other properties in the field area. D. Craw, G. Dipple, and C. Hilgers are thanked for careful reviews.