7/27/2019 Sibs on 1996

1/12

JournalofStruclurol Geology, Vol. 18, No. 8, pp. 1031 to 1042, 1996Copyright Q 1996 Elsevier Science LtdPII: S0191-8141(96)00032-6 Printed in Great Britain. All rights reserved01916141/96%15.00+0.00

Structural permeability of fluid-driven fault-fracture meshesRICHARD H. SIBSON

Department of Geology, University of Otago, P.O. Box 56, Dunedin, New Zealand(Received 27 Jun e 1995; accepted in revisedform 26 March 1996)

Abstract-Fluid redistribution in the crust is influenced by hydraulic gradient, by existing permeability anisotropyarising from bedding and other forms of layering, and by structural permeability developed under the prevailingstress field. Field evidence suggests that mesh structures, comprising faults interlinked with extensional-shear andpurely extensional vein-fractures, form important conduits for large volume flow of hydrothermal andhydrocarbon fluids. Meshes may be self-generated by the infiltration of pressurised fluids into a stressedheterogeneous rock mass with varying material properties, developing best where bulk coaxial strain is symmetricwith existing layering, but they also form under predominantly simple shear. Fluid passage through such structuresgenerates earthquake swarm activity by distributed fault-valve action along suprahydrostatic gradients that mayarise from compaction overpressuring, metamorphic dewatering, magmatic intrusion, and mantle degassing.Within mesh structures, strong directional permeability may develop in the u2 direction parallel to fault-fractureintersections and orthogonal to fault slip vectors. In particular tectonic settings, this promotes strongly focused flowwith high potential for mineralisation. Mesh activation requires the condition Pr _ ~3 to be maintained for thestructures to remain high permeability conduits, requiring fluid overpressuring at other than shallow depths inextensional-transtensional regimes. Favoured localities for mesh development include linkage structures alonglarge-displacement fault zones such as dilational jogs, lateral ramps, and transfer faults. In some circumstances,mesh formation appears to precede the development of major faults. Copyright 0 1996 Elsevier Science Ltd

INTRODUCTIONThe importance of fault-related fracture systems as hoststructures for hydrothermal mineralisation has long beenrecognised (Newhouse 1942, McKinstry 1948). Largevolume fluid flow through an extensive rock volume isneeded to form disseminated hydrothermal deposits ofeconomic significance; for instance, given the lowsolubility of gold transported as complexes (= 10 ppb),a flow of N 1 km3 of aqueous fluid is needed to precipitate10 tonnes of gold in a modest-sized deposit. However,while it is clear that fault-fracture networks form theprincipal avenues for large volume fluid flow in manyareas, it is also notable that the hosting structures aregenerally not large displacement features. Here, weconsider the factors which allow the development ofextensive low-displacement fault-fracture networks andthe flow of large fluid volumes, drawing comparisonswith the distributed faulting and fracturing and accom-panying fluid redistribution believed responsible forearthquake swarms.

of metres (Sibson 1989). Smaller events may, however, goundetected. The predominance of small earthquakes overlarge means that swarms sometimes have rather high b-values in comparison with normal tectonic seismicity.Swarm activity may remain more or less stationary inspace (e.g. Danville, California (Lee et af. 1971)) but mayalso be migratory with the locus of activity changing withtime (e.g. swarms in the Taupo Volcanic Zone, NewZealand (Grindley & Hull 1986)).

Earthquake swarms are most frequently associatedwith areas of recent volcanic or geothermal activitywithin extensional and transtensional tectonic regimes;in the latter case, they tend to be localized withindilational jogs along strike-slip faults (Weaver & Hill1979). Less commonly, swarms have been described fromareas of active compressional tectonics such as the SantaBarbara Channel, California (Shaw & Suppe 1994).Swarm activity has been attributed either to high levelsof material and/or stress heterogeneity within the rockmass (Mogi 1963, Scholz 1968) or, on the basis of tectonicsetting, to the migration of magmatic and/or hydro-thermal fluids (e.g. Sykes 1970, Hill 1977, Gold & Soter1985).

EARTHQUAKE SWARMS AND FLUIDREDISTRIBUTION Hill (3977) mesh modelfor earthquake swarms

Earthquake swarms are a variant of seismic activity To account for the volumetric character of earthquakewhere a large number of small earthquakes occur swarms, the preponderance of small earthquakes overthroughout a substantial volume of rock with the activity large, and their common association with volcanic/waxing and waning through time, often without the geothermal areas, Hill (1977) proposed a mechanicaloccurrence of a distinct principal shock (Sykes 1970). model for swarms involving the migration of fluidsRecorded swarm earthquakes are commonly in the through a honeycomb mesh of interlinked minor shearmicroearthquake range (1~ M < 3) with slip increments fractures (faults) and extension fractures (Fig. 1). Termi-of millimetres to centimetres associated, respectively, nation of individual shears in extension fractures inhibitswith individual rupture dimensions of tens to hundreds the development of large throughgoing faults and

1031

7/27/2019 Sibs on 1996

2/12

1032 R. H. SIBSON

Fig. 1. Hill (1977) mesh model for earthquake swarms comprisinginterlinked shear, extensional, and extensional-shear fractures devel-oped in a triaxial stress field. Diagram represents an extensional normal-fault stress regime when upright, a compressional thrust-fault regime

when viewed sideways, and a strike-slip regime in plan view.

accounts for the dominance of small ruptures. Hillinvoked igneous intrusion as the primary driving forcefor swarm activity, with dykes intruding a set of enechelon extension fractures linked by crack-tip shears.However, noting the occurrence of swarm-like seismicityinduced by artificial fluid injection at Denver andRangeley, Colorado (Healy et al. 1968, Raleigh et al.

1976) he also suggested that earthquake swarm activitymight be triggered by the migration of hydrothermalfluids through similar fault-fracture meshes.The Matsushiro earthquake swarm

The earthquake swarm that occurred from 1965 to1967 near Matsushiro in Central Honshu, Japan (Fig. 2)is an example of particularly intense swarm activityaccompanied by the development of laterally extensivestrike-slip faulting and significant fluid redistribution.Over 700,000 shocks ranging up to M5 were recorded, thewhole being energetically equivalent to one -M6.3earthquake (Hagiwara & Iwata 1968). Most of theactivity occurred within an ellipsoidal volume (dimen-sions - 34 km NE-SW, by - 18 km NW-SE, by - 8 kmdeep) centred around Mt Minakami, a Quaternaryandesitic lava dome. Three peaks in swarm activityoccurred in the periods November-December 1965,March-April 1966 and August-September 1966, the lasttwo being very pronounced (> 500 felt earthquakes perday).Despite strong NE-SW elongation of the swarm alonga well-defined structural grain, surface deformation inthe form of a mesh structure of en echelon tension crackshear zones developed during the third peak in swarmactivity to define a NW-SE trending left-lateral faultzone, some 500 m in width, which could be traced for over5 km (Tsuneishi & Nakamura 1970). The deformationincluded vertical left-lateral shear zones oriented WNW-ESE to E-W and right-lateral shear zones oriented NE-SW to ENE-WSW (Fig. 2). Geodetic measurementsshowed that a total left-slip of - 0.5 m accompanied by-0.3 m of NE-SW extension occurred across the mainfault zone, consistent with focal mechanism analyseswhich yielded double-couple solutions involving eitherleft-lateral strike-slip along steep NW-SE planes, ordextral strike-slip along NE-SW planes (Ichikawa1969). The various observations are consistent with theactivation under E-W compression of a Hill-type strike-

- \\\ \ , 1. 0 km ,

I Aug 1965 Feb 1966II Mar 1966 Jul 1966III Aug 1966 -Dee 1966IV Jan 1967 May 1967V Jun 1967 -0ct 1967

Fig. 2. Time-dependent development of the 1965-67 Matsushiro earthquake swarm and associated surface deformation withcomposite strike-slip focal mechanism (after Hagiwara & Iwata 1968, Ichikawa 1969, Tsuneishi & Nakamura 1970).

7/27/2019 Sibs on 1996

3/12

Structural permeability of fault-fracture meshes 1033slip fault-fracture mesh extending throughout the entirefocal volume.

Swarm development was accompanied by doming ofthe epicentral area to a maximum of over 0.8 m duringthe third peak of seismic activity, after which somesubsidence occurred. Spring flow in the epicentral areaincreased markedly following the two latter peaks inswarm activity and, immediately after the third peak, alarge volume (< 10 m3) of warm Na-Caxhloride brinesaturated with CO2 was discharged from the area ofground fissuring (Morimoto et al. 1967, Tsuneishi &Nakamura 1970). Following the outpouring, seismicitydiminished in what previously had been one of the mostintensely active areas. Subsequent studies have shownabnormally high 3He/4He ratios, possibly derived from ashallow magmatic intrusion, to be present in soil gaseswithin the main shear zone (Wakita et al. 1978).

Explanations put forward to account for the Matsush-iro phenomena include stress-driven prefailure dilatancyand fluid diffusion around the developing fault zone withpostfailure recovery leading to fluid expulsion (Nur 1974,Kisslinger 1975), or an episode of magmatic intrusion atdepth (Stuart & Johnston 1975, Wakita et al. 1978).Nakamura ( 1971 , however, suggested, that the primarycause of the Matsushiro swarm and its accompanyingeffects was the vertical migration of a package ofoverpressured hydrothermal fluid (a water eruption), aview latterly supported by Mogi (1988). Irrespective ofmagmatic involvement, the Matsushiro swarm clearlyrepresents a case of swarm seismicity linked to a majorepisode of hydrothermal fluid migration through alaterally extensive strike-slip fault-fracture mesh. Similarbut less well documented episodes of fluid effusion havebeen noted accompanying swarm activity along the beltof normal faults within the Taupo Volcanic Zone in theNorth Island of New Zealand (Grindley & Hull 1986).

DEVELOPMENT OF FAULT-FRACTUREMESHES

Fluid migration through the crust is influenced by 3-Dinterconnections between existing discontinuities such asbedding, joints, faults, foliation, igneous contacts, etc.Because of crustal heterogeneity with material propertiesvarying widely over a range of scales, particularly in areasof structural overprinting, the local stress state whichaffects the aperture of existing discontinuities may showconsiderable variation from the far-field stress (Fig. 3).As a consequence, the permeability structure is also likelyto be highly heterogeneous. The hypothesis advancedhere is that infiltration of pressurised fluids into stressedheterogeneous crust, perhaps initially along existingdiscontinuities, may induce a range of differently orientedbrittle structures that become progressively interlinkedinto a structural mesh which adds to the permeability ofthe rock mass. Evidence that such mesh structures areprimarily fluid-driven comes from their localisationwithin often uniform rock masses, from hydrothermalalteration and veining indicating their role as fluid

tFig. 3. Schematic illustration of the inception of a fault-fracture meshas a consequence of pressurised fluid infiltration into stressed hetero-geneous crust with inherited structure. Material heterogeneity causeslocal departures from the far-field stress and varying modes of brittle

failure.

conduits, and from the presence of extension veinsindicative of local overpressuring. The resulting bulkstructural permeability within the rock mass can thus beregarded as self-generated by the infiltrating fluids.Str ess contr ols on str uctur alpermeabil it y

In a fluid-saturated rock mass, all normal stresses arereduced by the fluid pressure, Pf , to give effective stresses,0 = (on - Pf ) (Hubbert & Rubey 1959). The effectiveprincipal compressive stresses are then crI= (gl - Pf) >CT~ (02 - Pf) > 03 = (as - P,), and the orientation,with respect to the triaxial stress field, of structuralelements formed by different modes of brittle failure inintact, homogeneous and isotropic rock is reasonablydescribed by classical failure theory involving a compo-site Griffith-Coulomb failure envelope (Fig. 4) (e.g.Brace 1960, Hancock 1985). Different stress-controlledstructural elements affecting rock mass permeability arelisted in Table 1, along with appropriate macroscopicfailure criteria expressed in terms of the fluid pressurecondition for failure. For intact rock, they include: (i)

Fig. 4. Mohr diagram with composite failure envelope for intact rockwith tensile strength, T, illustrating the stress conditions and orienta-tions with respect to the stress field of: (a) extensional failure; (b) hybridextensional-shear failure; and, (c) compressional shear failure, for a

particular rock-type.

7/27/2019 Sibs on 1996

4/12

R. H. SIBSONgrai n-scale tensil e mi crocracks which, as a consequence ofstress heterogeneities arising from grain impingement,form over a range of orientations but statistically show apreferential alignment perpendicular to g3; (ii) macro-scopic extension fractures forming perpendicular to c3 byhydraulic extension fracturing (Secor 1965); (iii) macro-scopi c shear fract ures (faul t s) developing in planescontaining the g2 direction and lying at angles, typically+20 < ~9~< f 30, to o1 (Anderson 1951); (iv) hybridextensional-shear fractures developing in planes whichalso contain the (~2 direction but lie at lower angles(e,, < 0,) to ~1 (Hancock 1985) and, (v) sty l o l i t i c solut ionplanes developing as anticracks (Fletcher & Pollard 198 1)in planes perpendicular to 01 (Fig. 5a). Existing planarcontacts, with or without inherent cohesive strength, mayalso be reactivated.

The type of brittle failure that occurs as a consequenceof pressurised fluid infiltration at a particular point instressed heterogeneous crust therefore depends largely onthe balance between the differential stress and localtensile rock strength, both of which may vary consider-ably. As an individual structure forms, the local stressfield becomes further perturbed through interactionswith existing mesh components (Segall & Pollard 1980)adding to stress heterogeneity within the evolving fault-fracture system. A complex mesh of different brittlestructures may then develop in response to the hetero-geneity of material properties, stress, and fluid pressurewithin the rock-mass, with the orientation of any onetype of mesh component showing considerable variance.

Formation of different structural elements within amesh may affect rock-mass permeability in a variety ofways. For example, while faulting in low-porosity rockwill generally increase fracture permeability as a con-sequence of cataclastic brecciation, faulting in initiallyhigh-porosity sedimentary rock may lead to the forma-tion of reduced permeability deformation bandsthrough porosity collapse and comminution (Aydin &Johnson 1978). Another important point is that much ofthe structural permeability developed in fault-fracturemeshes is likely to be transitory, as microcracks, fracturesand faults may all become infilled with low-permeabilityhydrothermal precipitates, alteration products, and/orclay-rich gouge. Permeability enhancement throughfaulting and fracturing in hydrothermal systems com-petes with permeability reduction as a consequence offluid flow and precipitation (Henley 1984) so thatdeformation within fault-fracture meshes, either con-tinual or intermittent, is probably necessary for them toremain effective as high-permeability structural conduits(Sibson 1994).Facto rs affect in g mesh development

General factors favouring mesh formation includematerial heterogeneity within the rock mass (so that thestress state and/or the tensile strength vary from place toplace) and low effective stress conditions. Layered rocksequences are particularly susceptible to mesh develop-ment, especially in extensional tectonic regimes, because

7/27/2019 Sibs on 1996

5/12

Structural permeability of fault-fracture meshes 1035

conjugatet

extn.fault shear

(a) (b) wFig. 5. (a) Stress-controlled mesh components; (b) segmented fault with linkingcompressional and dilational jogs (after Segall& Pollard 1980); (c) throughgoing fault developed from a Riedel shear mesh (after Tchalenko 1970). Note that the commonintersection of mesh components is the trz direction.

the brittle failure mode is then dictated by the tensilestrength of alternating rock layers (Sibson 1994, Gross1995, Gross & Engelder 1995). In such cases, layerthickness may impose a scale on the mesh structure thatdevelops. Mesh structures incorporating extensional-shear and/or extension fractures require the conditionPf - c3 to be met, at least locally, and are unlikely todevelop where effective stress is high.

In more homogeneous rock-masses, an additionalfactor promoting mesh formation within extensionaland strike-slip regimes is the limited vertical extent ofindividual hydraulic extension fractures. Secor &Pollard (1975) pointed out that while the fluid pressure

at the top of a vertical fluid-filled crack must exceed theleast compressive stress to meet the tensile failurecriterion (Table l), the hydrostatic fluid pressuregradient within the crack will, in general, be less thanthe vertical gradient of g3 in the wallrock. Such crackscan, therefore, only remain open over the limited depthrange where Pf > 03. They suggested an upper limit ofperhaps 100 m or so for individual vertical hydro-fractures in crystalline rock. Outside this restrictedinterval, fluid pressure conditions favour shear failure,with the possibility of an interlinked fault-fracturemesh developing as overpressured fluids migrate alonga hydraulic gradient.

BULK COAXIAL STRAIN(4 ; (4

IIIIIO3 IIIIII

(d)SIMPLE SHEAR (+ AV)

: 03 I (f)

I IFig. 6. Varieties of fault-fracture mesh: Bdk Coaxial Strai n (a) conjugate Coulomb shears; (b) Hill-type mesh in alternatingcompetent and incompetent strata; (c) conjugate tension gash shear zone mesh; Simpl e Shear (+ AV); (d) sheared-out tensiongash shear zone; (e) interlinked shears and dilational jogs; (f) dead-fish shear zone with anastomosing slickenfibre surfacesinterlinked by extension fractures in competent phacoids. Common intersection of mesh components is again the cz direction.

7/27/2019 Sibs on 1996

6/12

1036 R. H. SIBSONVARIETIES OF FAULT-FRACTURE MESH

Characteristic orientations of the basic structuralelements with respect to the local stress field are given inFig. 5(a), along with simple combinations of the differentstructural elements that are likely to develop in individualbrittle to semi-brittle shear zones as a consequence of enechelon segmentation (Fig. 5b) (Segall & Pollard 1980) orthrough the amalgamation of Riedel shear arrays (Fig.5c) (Tchalenko 1970). Different structural elements maycombine in various combinations to build more extensivefault-fracture meshes (Fig. 6). Not all components needbe present in every case.

Fault-fracture meshes accommodating bulk coaxialstrain are most likely to develop when the symmetry ofthe dominant anisotropy in the rock mass (e.g. bedding)and the prevailing stress field coincide. These structuresform a continuum with a varying content of extensionfractures, ranging from arrays of conjugate Coulombshears, through Hill-type fault-fracture meshes, toconjugate tension gash shear zones (Figs. 6aac and7a-d). Hill-type meshes are commonly well developedwhere beds of strong competence contrast are inter-layered, as in the oil-rich Monterey Formation of coastalCalifornia (comprising interbedded dolostones, lime-stones, siliceous shales, and mudstones) which is notablefor the manner in which the mode of brittle failure varieswith lithology (Gross 1995, Gross & Engelder 1995)(Figs. 7a &b).

In situations approximating bulk simple shear with alarge component of rotational strain, the resulting meshstructure is likely to be less regular. Various combina-tions of tension gash shear zones formed by hydraulicfracturing interlinked with discrete shear surfaces maydevelop. Figure 6(e)-(f) are schematic representations ofdifferent styles of shearing in the Chrystalls BeachComplex, interpreted as part of a Mesozoic accretionarymelange by Nelson (1982). In thicker pelitic units withinthe complex, quartz and calcite slickenfibre slidingsurfaces dispersed on scales of centimetres to metresthroughout the complex are interlinked through dila-tional jogs to form a shear zone mesh (Figs. 6e and 7e).Elsewhere, dismembered lenses of comparatively compe-tent sandstone in a highly strained pelitic matrix are cutby arrays of extension veins and bounded by anastomos-ing slickenfibre surfaces (Figs. 6f and 7f).

Orientation of mesh structures is dictated by thevertical stress associated with different tectonic regimes(Anderson 1951). In extensional regimes with (T, = gl,mesh components may include steeply-dipping normalfaults in conjugate sets interlinked with subverticalextension fractures and subhorizontal stylolites, while in

a compressional regime with 0 = CT~, mesh structuremay involve conjugate gently-dipping thrust faultsinterlinked with subhorizontal extension fractures andsubvertical stylolites. In a strike-slip regime with (T, = g2,a mesh may comprise subvertical extension fractures,stylolites and conjugate strike-slip faults. Petit & Mat-tauer (1995) have recently described just such a strike-slipmesh structure, with a mesh dimension of - 10 m perhapsdictated by an inherited semi-brittle fabric, in Jurassiclimestone strata affected by Pyrenean deformation in theLanguedoc region of southern France.

Note that because of crustal heterogeneity, thetendency for individual shears to terminate laterally inen echelon tension crack arrays as well as at their tips(McGrath & Davison 1995) and for complex interac-tions to develop between mesh components, the 3-Dmesh structure will generally be a great deal morecomplicated than the 2-D plane strain diagrams inFigs. 5 and 6 suggest. Additional complexities may arisefrom the formation of orthorhombic fault systems toaccommodate bulk strains in three dimensions (Reches1983). Moreover, with increasing displacement somemesh structures may evolve into throughgoing planarfaults that are favourably oriented in the prevailing stressfield and no longer require the same degree of over-pressuring for reactivation. Mesh development in regionsof high fluid pressure may thus be a precursor to thedevelopment of major new fault structures.g2 directional permeability in fault-fracture meshes

Microscopic and macroscopic extension fracturingenhances permeability in the al/a2 plane, the effectbecoming more pronounced as the fluid pressureapproaches cr3, and the effective least compressive stress,CTJ= ((rs - Pr) tends to zero. Unhealed fault surfacesforming in low-porosity rock, through their intrinsicroughness, also provide permeable channelways. Thecommon intersection of faults and extension fracturesthus enhances rock-mass permeability by adding atubular component in the o2 direction (see Figs. 5 and6) perpendicular to fault slip vectors (cf. Snow 1969).Directional permeability in the a2 direction is also likelyto develop when a conjugate fault mesh forms in rockwith high initial porosity and permeability; reducedpermeability along the faults through comminution andporosity collapse (Aydin & Johnson 1978) then boundstubes of high permeability in the intervening rock-mass(Fig. 6a). Stylolitic solution seams with their residue ofimpermeable clay minerals also represent tabular zonesof reduced porosity (Groshong 1988, Carrio-Schaff-hauser et al. 1990) that restrict flow perpendicular to the

Fig. 7. Photographs of mesh structures: (a, b) varieties of Hill-type extensional fault-fracture meshes (conjugate normal faults. subverticalextension fractures and extensional shears), with infillings of calcite and bitumen from the Miocene Monterey Formation of interbedded dolostones,limestones, siliceous shales, and mudstones, Arroyo Burro Beach, Santa Barbara; (c) conjugate quartz tension-gash shear zone mesh with spacedsolution cleavage (dark striping) developed in a bed of Old Red Sandstone, Marloes Sands, SW Wales; (d) compressional fault-fracture mesh withconjugate thrust faults and associated hydraulic extension fractures infilled with quartz-prehnite and laumontiteecalcite assemblages in steeply-dipping interbedded greywacke sandstones and argillites, east abutment, Benmore Dam, Waitaki Valley, New Zealand; (e, f) quartz and calciteslickenfibre sliding surfaces interlinked with extension fractures and dilational jogs in sheared sandstone-shale melange, Chrystalls Beach

accretionary complex. SE Otago coast, New Zealand.

7/27/2019 Sibs on 1996

7/12

Structural permeability of fault-fracture meshes

Fig. 7. (a)-(c).

7/27/2019 Sibs on 1996

8/12

R. H. SIBSON

Fg. 7. (d)-(f)

1038

7/27/2019 Sibs on 1996

9/12

Structural permeability of fault-fracture meshes 1039c~/rrs plane in which they form (Fletcher & Pollard 198 1).Overall, the combined effect of all these structuralelements is to impart directional permeability along theaverage g2 direction within the rock mass, though theextent of the effect will depend on the 3-D interconnect-edness of the different structural elements.

FLUID PRESSURE CONTROLS ON MESHACTIVATION

The level of fluid pressure at depth, z, in the crust can beexpressed in terms of the pore-fluid factor, 1, = (Pr/a,) =(Pr/pgz), where p is the average rock density and g is thegravitational acceleration. In crust with fluid-saturatedpore and/or fracture space interconnected to the Earthssurface, hydrostatic conditions prevail with 2. = pr/p z0.4. Actively deforming portions of the crust are,however, commonly overpressured at depth with 1, + 1(- lithostatic fluid pressure conditions) in some circum-stances (Fyfe et al. 1978). Overpressuring may arise fromburial compaction in thick sedimentary sequences,metamorphic dewatering in the middle to deep crust,thermal overpressuring from magmatic intrusion, andmantle degassing. All may contribute to suprahydrostaticvertical fluid pressure gradients, but strong lateralhydraulic gradients may also develop, particularly adja-cent to magmatic intrusions. What, then, are the fluidpressure conditions for mesh activation at differentdepths in different tectonic regimes?

To function as high permeability conduits promotingflow of large fluid volumes, mesh structures mustincorporate a significant proportion of gaping extensionfractures and extensional shears. Formation of suchstructures in relatively competent material requiresPf > o3 and (cri - CT~) =c4T locally (Table l), but thestress field throughout the mesh will be very inhomo-geneous. However, on a gross scale, a Hill-type mesh canbe regarded kinematically as equivalent to a singleextension fracture; on these grounds, the condition,Pf E c3 (where (TVs the average least stress at a particulardepth) is adopted here as an approximate criterion formesh activation and large volume fluid flow. Becausehydraulic extension fracturing is integral to the forma-tion of higher permeability meshes and T < 10 MPa formost competent rocks (Etheridge 1983), the upper boundto differential stress during mesh propagation is taken as40 MPa.

Adapting the procedure developed by Secor (1965) thefluid pressure conditions for mesh activation at differentdepths in the crust are expressed on a plot of the pore-fluid factor, A,, versus depth, z, for different stress regimesassuming a maximum differential stress of 40 MPa (Fig.8). Note that for lesser values of rock tensile strength andpermissible differential stress, the 1, values needed forhydrofracturing at a particular depth will be higher.From the 2,/z plot it is apparent that whereas meshactivation in a compressional thrust-fault regime (a, =rr3) requires lithostatic fluid pressures (2, - 1) at alldepths, activation of fault-fracture meshes with large-

0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.6 0.9 1.0, , , ,, , , , , , ,c

STRIKE-SLIP

5 -

Fig. 8. Fluid pressure conditions for activation of fault-fracture meshesin compressional thrust-fault, extensional normal-fault, and strike-slipregimes expressed on a plot of pore-fluid factor, i,,, versus depth for thelimiting condition (01 - 03) < 4T = 40 MPa.

volume fluid flow can occur under hydrostatic or lowerfluid pressure levels (& 50.4) at shallow depths in anextensional normal-fault regime (a, = ai). In strike-slipregimes (fr, = ~r2), he fluid pressure condition for meshactivation can lie anywhere between the extensionalsituation when aV = a2 - a1 in a transtensional setting,and the compressional situation when aV = a2 - a3under transpression. The case illustrated is for theintermediate situation where aV = a2 = 4 ( a1 ax). Atgreater depths in extensional-transtensional settings,fluid overpressures (0.4 < 2, c 1 O) are needed to activatemesh structures as high permeability conduits.

The extent and rate of mesh development andpropagation will be strongly affected by fluid supply, themaintenance of an appropriate hydraulic gradient, andthe permeability characteristics of the rock mass. It seemsprobable that active fault-fracture meshes can beregarded as buoyant, overpressured fluid domains ofthe kind envisaged by Gold & Soter (1985). They pointedout that many of the time-dependent and migratorycharacteristics of earthquake sequences are explicable interms of transitory linkages (burping) within a stack ofoverpressured fluid domains; this concept seems likely tobe especially applicable to swarm activity.

TECTONIC ENVIRONMENTS FOR MESHFORMATION

The predominant association of earthquake swarmswith areas of active extensional and transtensionaltectonics suggests that fault-fracture meshes are widelydeveloped in such settings where they likely play animportant role in the formation of hydrothermal mineraldeposits. Established hydrothermal plumes in such areasare close to hydrostatically pressured, at least in the near-surface (Henley 1984), and may either be localised by

7/27/2019 Sibs on 1996

10/12

1040 R. H. SIBSONexisting fault-fracture mesh conduits or have contributedto their generation. Mesh activation in extensionalregimes is strongly favoured at shallow depths (generally< 1 km) in the vicinity of the hydrostatic boiling horizon(Fig. 8), and must inevitably contribute to fluid focusingand the generation of epithermal mineralisation.

In this regard, it is interesting to note the intenseshallow (z < 5 km) microseismicity from distributednormal faulting, somewhat akin to swarm seismicity,that extends around areas of steam extraction and highflow rate in the Geysers geothermal field of northernCalifornia (Oppenheimer 1986). Towards the base of thefield there is evidence for a transition from a hydrostaticto a strongly overpressured hydrothermal regime(Fournier 1991). It is tempting to believe that themicroseismic activity arises from fluid migration througha fault-fracture mesh developed across this transitionregion, representing a form of distributed fault-valveaction along a suprahydrostatic fluid pressure gradient.Detailed seismicity studies within the San Andreasfault system have revealed other zones of clusteredmicroearthquake activity that, akin to swarms, maydefine zones of overpressured fluid migration (Hill et al.1990). Along creeping segments of the fault system, thefull depth of the seismogenic zone down to lo-15 km isintensely microseismically active. This globally anom-alous slip behaviour has previously been attributed(Irwin & Barnes 1975) to the trapping of overpressuredfluids within the strike-slip faults beneath a cappingserpentinite seal associated with the Coast Range Thrust.The mechanical inference developed here is that thecoupling of microearthquake clusters to aseismic slip insuch regions likely involves simple shear across a tabularfault-fracture mesh, developed in response to a near-lithostatic fluid pressure profile, that extends from thenear-surface to the base of the seismogenic zone.

Elsewhere, along some of the locked portions of theSan Andreas system, the base of the seismogenic zone isdefined by a band of microearthquake activity with avertical extent of a few kilometres. The interpretationhere is that the structural level at the base of theseismogenic zone, defining the onset of greenschistfacies metamorphic conditions, is associated withupwards migration of fluids, accompanied by meshdevelopment, across the transition from a near-lithostaticto a sub-lithostatic fluid pressure environment. Interest-ingly, this is the structural level in the crust wheremesothermal gold-quartz lodes (occupying fault-fracture meshes t2 km in vertical extent) most com-monly develop by fault-valve action across steep hydrau-lic gradients (Sibson et al. 1988). While the extreme fault-valve action that gives rise to such major vein systems isgenerally restricted to steep reverse faults that areseverely misoriented for reactivation, less intense episodicfluid migration at this crustal level seems likely to be afeature of most transcrustal fault zones (Sibson 1994).

Swarm activity is comparatively rare in compressionalsettings but is known from the fold-thrust belt in theSanta Barbara Channel which, at least locally, is knownto be overpressured to near-lithostatic values (Yeats

1983). A recent swarm in 1984 occurred at depths of 7-15 km directly beneath one of the major oil-fields (Shaw& Suppe 1994), apparently along a steep reverse fault; themost favourable structural setting for extreme valvingaction (Sibson 1990).

Mesh structures as fluid conduits in different tectonicsettingsThe direction of fluid flow through a rock mass is

governed by the maximum hydraulic gradient (notnecessarily vertical), existing permeability anisotropy(e.g. bedding, foliation), and superimposed structuralpermeability. Particular combinations of these para-meters may lead to the localised high-volume flowsneeded to develop hydrothermal mineral deposits.Favoured localities for mesh development include short-lived irregularities along large-displacement fault zonessuch as dilational jogs, lateral ramps, transfer faults, andother forms of link structure. In the case of strike-slipfaulting, obvious examples of enhanced permeability infault-fracture meshes parallel to vertical CJ~ nclude theMatsushiro swarm, and the localisation of both magma-tism and associated upwelling hydrothermal plumes inmajor dilational jogs along the San Andreas fault in theSalton Sea area of California (Weaver & Hill 1979,Sibson 1987). Different combinations of hydraulicgradient, intrinsic permeability anisotropy and structuralpermeability may, however, lead to more complex 3-Dflow paths as in the situations outlined in Fig. 9.

Figure 9(a) illustrates a situation where, despite avertical gradient in hydraulic potential, the presence of agently raking dilational jog on a normal fault localisesand deflects flow laterally along the jog structure. Suchfeatures might account for the localisation of hot springactivity at particular sites along normal faults, as occurswithin the Taupo Volcanic Zone in New Zealand. In Fig.9(b), development of a Hill (1977) fault-fracture mesh ina sequence of gently dipping interbedded competent andincompetent strata promotes flow up-dip (cf. Fig. 7a).Dilational jogs associated with reverse slip on bedding-parallel faults during progressive fold tightening andamplification (Fig. SC) induce along-strike migration intohinge-line culminations, as demonstrated for the WattleGully Au-quartz vein system in the Lachlan fold belt ofVictoria, Australia (Cox et al. 1991); this mechanismmay also assist the syntectonic migration of hydrocar-bons into antiformal reservoirs. The effect of a stronglateral hydraulic gradient adjacent to a high-level pluton,coupled with the high (T? permeability of an extensionalfault-fracture mesh, leads to strongly focused lateral flow(Fig. 9d). This situation resembles the structures hostingtin mineralisation around the cupolas of the Cornubianbatholith in SW England (Moore 1975), and may alsohave relevance to the localisation of some geothermalfields in the Taupo Volcanic Zone near the intersectionsof the active belt of normal faults with major ringstructures defining buried calderas (Cole 1990).

These examples, all involving predominantly dip-slipfaulting, illustrate how migration paths for escaping

7/27/2019 Sibs on 1996

11/12

Structural permeability of fault-fracture meshes 1041

.^_.~.^^........_. . . . . . . . ..^_..._. . . . ..^^.......^. . . . . ..^...^......^^...^...._..^..l...._...^....

. . . . . ..__......._^......__..__....

. . ..~..^_.......... .._...___._.

.^^I.__...:.:_:-*. . . . . . .

Fig. 9. Mesh structures in different tectonic settings illustrating the effects of e2 permeability enhancement: (a) gently rakingdilational jog on normal fault; (b) Hill-type fault-fracture mesh in gently dipping interbedded competent and incompetentstrata; (c) dilational jogs and saddle reefs associated with bedding plane reverse-slip, promoting along axis flow into fold

culminations; (d) extensional normal-fault mesh abutting plutonic contact with intense lateral hydraulic gradient.

overpressured fluids may be deflected from the vertical bydirectional stress-controlled permeability. This couldhelp to account for the rapid lateral, as well as vertical,migration of swarm activity that is sometimes observed(Grindley & Hull 1986).

DISCUSSION

This paper has addressed two principal themes. Thefirst is the self-generation of structural permeability indifferent stress regimes by migrating fluids and itsimportance in promoting the large-volume fluid flowneeded to form hydrothermal ore deposits. Stress-controlled structural permeability may also play animportant role in hydrocarbon migration (Du Rouchet1981). The second theme concerns the extent to whichmicroearthquake swarms or clusters may be used todefine active fault-fracture meshes in different tectonicsettings. At other than very shallow depths in exten-sional-transtensional settings, such swarm activity prob-ably represents migration of overpressured fluids alongsuprahydrostatic fluid pressure gradients (cf. Gold &Soter 1985). This is of particular interest given recentrecognition that intermittent breaching of overpressuredfluid compartments in sedimentary basins leads toepisodes of fluid migration (Hunt 1990, Powley 1990)and the expectation of similar behaviour in and aroundcrustal fault zones (Sibson 1981, 1990, Byerlee 1993).Persistent swarm activity at shallow depths (< 4 kmdeep, say) is an attractive target for scientific drilling andother forms of geophysical investigation to test thehypothesis that swarms are associated with the migrationof overpressured fluids. Penetration of swarms by drillingwould allow direct measurement of fluid pressure levels,while sustained measurements might allow correlation of

seismic activity with local changes in fluid pressure. Meshpropagation could be expected to induce localized APffluctuations of the order of the rocks tensile strength(l-10 MPa). Correlations between swarm activity, sur-face deformation, hydrological fluctuations and ground-water chemistry, as at Matsushiro, have the potential toyield information on mesh dimensions and rates of fluidflow.

Important conclusions of this paper are that thedistribution of fluid overpressures in the crust exerts astrong influence on the style of faulting and patterns ofmineralisation, that varying seismic style in differenttectonic settings may be a key to understanding differenttypes of mineralising environment, and that structuralpermeability analysis has the potential to become animportant part of both mineral and hydrocarbonexploration.AcknoM,ledgements-Research leading to this paper forms part of astudy funded from the Public Good Science Fund by Grant UO-0312from the New Zealand Foundation for Research, Science, andTechnology. Many of the ideas developed here arose from conversa-tions over the years with Neville Price and Dave Hill. I thank John Scott,Joe Cartwright and Mike Blanpied for helpful criticism.

REFERENCESAnderson, E. M. 195 1, The Dynami cs of Fault ing and Dyk e Formation

wi th Appl icat ion t o Bri tain (2nd edn). Oliver and Boyd, Edinburgh.Aydin, A. & Johnson, A. M. 1978. Development of faults as zones ofdeformation bands and as slip surfaces in sandstones. Pure & Appl.

Geophys. 116,932-942.Brace, W. F. 1960. An extension of the Griffith theory of fracture torocks. J. geophys. Res. 65,3477-3480.

Byerlee, J. D. 1993. Model for episodic flow of high-pressure water infault zones before earthquakes. Geology 21, 303-306.

Carrio-Schaffhauser, E., Raynaud, S., Latiere, H. J. & Mazerolle, F.1990. Propagation and localisation of stylolites in limestones. In:

7/27/2019 Sibs on 1996

12/12

R. H. SIBSONMoore, J. M. 1975. A mechanical interpretation of the vein and dyke

systems of the SW England oretield. Miner. Deposita 10, 374-388.Morimoto, R., Nakamura, K., Tsuneishi, Y., Ossaka, J. & Tsunoda. N.

1967. Landslides in the epicentral area of the Matsushiro earthquakeswarm-their relation to the earthquake fault. Bull. Eurthq. Res. Inst.Tokyo Univ. 45,241-263.

Nakamura, K. 1971. A hypothesis on the mechanism of the Matsushiro

Deformation Mechanisms, Rheology and Tectonics (edited by Knipe,R. J. & Rutter E. H.). Spec. Publsgeol. Sot. Lond. 54, 193-199.

Cole, J. W. 1990. Structural control and origin of volcanism in theTaupo Volcanic Zone, New Zealand. Bull. ~olcanol. 52,445-459.

Cox. S. F.. Wall. V. J.. Etheridae. M. A. & Potter. T. F. 1991.Deformational and metamorphyc processes in the formation ofmesothermal vein-hosted gold deposits+xamples from the LachlanFold Belt in central Victoria, Australia. Ore Geol. Rev. 6, 391-423.

Du Rouchet, J. 1981. Stress fields, a key to oil migration. Bull. Am. Ass.Petrol. Geol. 65, 74-85.Etheridge, M. A. 1983. Differential stress magnitudes during regional

deformation and metamorphism: upper bound imposed by tensilefracturing. Geology 11, 2311234. _ _

Fletcher. R. C. & Pollard. D. D. 1981. Anticrack model for oressureIsolution surfaces. Geology 9,419-424.Fournier, R. 0. 1991. The transition from hydrostatic to greater than

hydrostatic fluid pressures in presently active hydrothermal systemsin crystalline rock. Geo&s. Res. Left. 18,955-958.

Fyfe, W. S., Price, N. J. & Thompson, A. B. 1978. Fluids in the Earth.sCrust. Elsevier, Amsterdam.

Gold, T. & Soter, S. 1985. Fluid ascent through the solid lithosphereand its relation to earthquakes. Pure & Appl. Geophys. 122,492-530.

Grindley, G. W. & Hull, A: G. 1986. Historical Taupe-earthquakes andearth deformation. Rev. Sot. N. 2. Bull. 24. 173-185.

Groshong, R. H. 1988. Low temperature deformation mechanisms andtheir interpretation. Bull. geol. Sot. Am. 100, 1329-1360.Gross, M. R. 1995. Fracture partitioning: failure mode as a function of

lithology in the Monterey Formation of coastal California. Bull. geol.Sot. Am. 107,779-792.

Gross, M. R. & Engelder, T. 1995. Strain accommodated by brittlefailure in adjacent units of the Monterey Formation, U.S.A.: scaleeffects and evidence for uniform displacement boundary conditions.J. Struct. Geol17, 1303-1318.

Hagiwara, T. & Iwata, T. 1968. Summary of the seismographicobservation of Matsushiro swarm earthquakes. Bull. Eurthq. Res.Inst. Tokyo Univ. 46,485-5 15.

Hancock, P. L. 1985. Brittle microtectonics: principles and practice. J.Struct. Geol. 7,437457.

Healy, J. H., Rubey. W. W., Griggs, D. T. & Raleigh, C. B. 1968. TheDenver earthquakes. Science 161, 1301-1310.

Henley, R. W. 1984. The geothermal framework of epithermal deposits.In: Geology and Geochemistry of Epthermal Systems (edited byBerger, B. R. & Bethke, P. M.). Sot. Econ. Geol. Rev. Econ. Geol. 2.I-24.

Hill, D. P. 1977. A model for earthquake swarms. J. geophys. Res. 82,347-352.

Hill, D. P., Eaton, J. P. &Jones, L. 1990. Seismicity, 1980-86. In: TheSan Andreas Fault System (edited bv Wallace. R. E.). U.S. Geol. Surv.Prof. Paper 1515, 1 i4-151: .

Hubbert, M. K. & Rubey, W. W. 1959. Role of fluid pressure inmechanics of overthrust faulting. BuU. geol. Sot. Am. 70, 115-166.

Hunt, J. M. 1990. Generation and migration of petroleum fromabnormally pressured fluid compartments. Bull. Am. Ass. Petrol.Geol. 74, 1-12.

Ichikawa. M. 1969. Matsushiro earthquake swarm. Geophys. Mag. 34,307-331.

Irwin, W. P. & Barnes, I. 1975. Effects of geologic structure and fluidson the seismic behaviour of the San Andreas fault system in centraland northern California. Geology 3, 7 13-7 16.

Kisslinger, C. 1975. Processes during the Matsushiro earthquake swarmas revealed by levelling, gravity, and spring-flow observations.Geology 3, 57-62.

Lee, W. H. K., Eaton, M. S. & Brabb, E. E. 1971. The earthquakesequence near Danville, California, 1970. Bull. seism. Sot. Am. 61,1771-1794.

McGrath, A. G. & Davison, I. 1995. Damage zone geometry aroundfault tips. J. Struct. Geol. 17, 101 l-1024.

McKinstry, H. E. 1948. Mining Geology. Prentice-Hall, EnglewoodCliffs, New Jersey.

Mogi, K. 1963. Some discussions on aftershocks, foreshocks, andearthquake swarms-the fracture of a semi-infinite body caused byan inner stress origin and its relation to the earthquake phenomenon,3. Bull. Earthy. Res. Inst. Tokyo Univ. 41, 615-658.

Mogi, K. 1988. The mechanism of the Matsushiro earthquake swarm incentral Japan and its relation to the 1964 Niigata earthquake.Tectonophysics 159, 109-l 19.

earthquake swarm. Kagaku-Asahi 10, 127-133.Nelson. K. D. 1982. A surgestion for the oriein of mesoscooic fabric inaccretionary melange, based on features ibserved in the Chrystalls

Beach Complex. South Island, New Zealand. Bull. geol. Sot. Am. 93.6255634.

Newhouse, W. H. 1942. Ore Deposit.r as Related to Structural Features.Princeton University Press, Princeton, New Jersey.

Nur, A. 1974. Matsushiro, Japan, earthquake swarm: confirmation ofthe dilatancy-fluid diffusion model. Geology 2, 2177221.

Oppenheimer, D. H. 1986. Extensional tectonics at The Geysersgeothermal area, California. J. geophys. Res. 91, 11.463-l 1,476.

Petit, J.-P. & Mattauer, M. 1995. Paleostress superimposition deducedfrom mesoscale structures in limestone: the Matelles exposure,Languedoc, France. J. Struct. Geol. 17.245-256.

Powley, D. E. 1990. Pressures and hydrogeology in petroleum basins,Earth Sci. Rev. 29, 2155226.

Raleigh, C. B., Healy, J. H. & Bredehoeft, J. D. 1976. An experiment inearthquake control at Rangeley. Colorado. Science 191, 1230-1236.Reches, Z. 1983. Faulting of rocks in three-dimensional strain fields II.Theoretical analysis. Tectonophysics 95, 133-l 56.

Scholz, C. H. 1968. The frequency-magnitude relation of microfractur-ing in rock and its relation to earthquakes. Bull. seism. Sot. Am. 58,399415.

Secor, D. T. 1965. Role of fluid pressure in jointing. Am. J. Sci. 263,633-646.

Secor, D. T. & Pollard, D. D. 1975. On the stability of open hydraulicfractures in the Earths crust. Geophys. Res. Lett. 2, 510-513.

Segall, P. & Pollard. D. D. 1980. Mechanics of discontinuous faults. J.geophys. Res. 85,43374350.

Shaw, J. H. & Suppe, J. 1994. Active faulting and growth folding in theeastern Santa Barbara Channel, California. Bull. geol. Sot. Am. 106,607-626.

Sibson, R. H. 1981. Fluid flow accompanying faulting: field evidenceand models. In: Earthquake Prediction.. an International Review(edited by Simpson, D. W. & Williams, P. G.). Am. Geophy.s. Vnion,Maurice Eli?ng Series 4, 593-603.

Sibson, R. H. 1987. Earthquake rupturing as a hydrothermal mmer-alising agent. Geology 15, 701-704.

Sibson, R. H. 1989. Earthquake faulting as a structural process. J.Struct. Geol. 11. l-14.

Sibson, R. H. 1990. Conditions for fault-valve behaviour. In: Deforma-tion Mechanisms, Rheology and Tectonics (edited by Knipe. R. J. &Rutter, E. H.). Spec. Publ. geol. Sot. Lond. 54, 15-28.

Sibson, R. H. 1994. Crustal stress, faulting. and fluid flow. In:Geofluids: Origin, Migration and Evolution of Fluirls in Sedimentor,Bastn.s (edited y Parnell, J.). Spec. Publs geol. Sot. Lond. 78, 69-84.

Sibson. R. H., Robert, F. & Paulsen. K. H. 1988. High-angle reversefaults, fluid-pressure cycling, and mesothermal gold-quartz deposits.Geology 16, 551-555.

Snow, D. T. 1969. Anisotropic permeability of fractured media. WaterResourc. Res. 5, 1273-l 289.

Stuart, W. D. & Johnston, M. J. S. 1975. Intrusive origin of theMatsushiro earthquake swarm. Geology 3,63-67.

Sykes, L. R. 1970. Earthquake swarms and sea-floor spreading. J.geophys. Res. 75, 6598-6611.

Tchalenko, J. S. 1970. Similarities between shear zones of differentmagnitudes. Bull. geol. Sot. Am. 81, 1625-1640.

Tsuneishi. Y. & Nakamura. K. 1970. Faultine. associated with theMatsushiro swarm earthquakes. Bull. Earthy. Res. Itwt. Tokyo Univ.48, 29-5 1.

Wakita, H., Fujii, N., Matsuo, S., Notsu, K., Nagao, K. & Takaoka, N.1978. Helium Spots: caused by a diapiric magma from the uppermantle. Science 200, 43&432.

Weaver, C. S. & Hill, D. P. 1979. Earthquake swarms and local crustalspreading along major strike-slip faults in California. Pure & Appl.Geophys. 117, 51-64.

Yeats, R. S. 1983. Large-scale Quaternary detachments in VenturaBasin, southern California. J. grophys. Res. 88, 569-583.