Earthquake rupturing as a mineralizing agent

in hydrothermal systems
Richard H. Sibson
Department of Geological Sciences, University of California, Santa Barbara, California 93106

ABSTRACT
Much fault-hosted epithermal mineralization is localized in dila- a —
tional jogs between en echelon fault segments, as fissure veins or
as hydrothermally cemented, high-dilation wall-rock breccias. Jog
widths may range from millimetres to kilometres; vein textures record
histories of incremental development. Perturbation or arrest of earth-
quake ruptures at dilational jogs has been observed and is believed to 5 cm
involve extensional fracturing at the rupture tip, locally reducing fluid
pressure and inducing suctions opposing rapid slip transfer across the
jog. This forced Assuring leads to brecciation by hydraulic implosion
and to a concentrated fluid influx, allowing delayed slip transfer ac-
companied by aftershock activity. Within the southern San Andreas
fault system, major dilational jogs extend throughout the seismogenic
regime and form loci for magmatic-hydrothermal systems; they act as
vertical pipelike conduits for enhanced fluid flow. Rupture termina-
500 m .
tion at these structures has sometimes been followed by hydrothermal
eruptions, suggesting that high-level boiling events are triggered by
the arrest mechanism. It thus seems probable that episodic mineral
deposition in the top 1-2 km of such jogs is induced by the dynamic
effects of rupturing on the flanking strike-slip faults.

INTRODUCTION
Brittle fault-fracture systems and related zones of brecciation have , 2 0 0 m,
long been recognized as channels for hydrothermal fluid flow and as sites
for epithermal mineral deposition. Whereas in some cases the fracture
systems have been regarded as passive fluid conduits, textural evidence
indicates that the passage of mineralizing fluids is often episodic and pene-
contemporaneous with increments of fault slip (Hulin, 1929; Sibson,
1981). Several observations prompt consideration of a causal link between
earthquake rupture processes and the development of fault-hosted mineral-
ization. (1) Intermittent earthquake rupturing over a broad range of scales
? ? —
is widespread in regions where epithermal mineralization seems likely to
,500 m,
be developing at present. (2) From fossil epithermal deposits, it is apparent
that such mineralization is often concentrated at specific structural sites
where local dilation has occurred within fault systems. (3) Vein textures in
such deposits often record a history of incremental precipitation. (4) Mod- e —
ern seismological studies demonstrate that propagating earthquake rup-
tures are often arrested at similar structural sites.

STRUCTURAL CONTROL OF EPITHERMAL

MINERALIZATION Figure 1. Examples of dilational fault jogs (redrawn for dextral shear
Epithermal mineralization is commonly localized in fault-fracture sense where necessary), a: Small-scale jog in sandstone infilled with
systems developed in the brittle carapace of regions undergoing magmatic locally fibrous quartz (from photograph by M. P. Coward), b: Segment
of surface rupture trace of 1968 Dasht-e-Bayaz earthquake, Iran (after
and related geothermal activity. In recent years it has become increasingly
Tchalenko and Ambraseys, 1970). c: Ore shoots in Camp Bird vein
apparent that much of the mineral deposition in such areas has occurred system, Colorado (after Spurr, 1925). d: Fault and vein system asso-
during episodes of boiling in upwelling hydrothermal plumes (Henley, ciated with Chuquicamata copper deposit, Chile (after Perry, 1952).
1985). Dilational fault jogs, localized zones of extension arising from fault e: General mesh model consisting of linked shear and extensional
curvature or en echelon segmentation, form particularly favorable deposi- fractures (after Sibson, 1985).
tional sites (Fig. 1). They are best developed in normal and strike-slip fault
systems, and the mineralized jogs form ore shoots perpendicular to the tation (Spurr, 1925; Hulin, 1929). Similar breccias have been described
fault slip vector (McKinstry, 1948). Typically, the deposits are character- from the vicinity of recently active faults in present-day geothermal sys-
ized by crustiform fissure veins whose textures record histories of incre- tems (Grindley and Browne, 1976). A notable feature (well exemplified in
mental extensional opening, or by high-dilation, hydrothermally cemented Fig. la) is that many of the driving faults remain barren or nearly barren of
wall-rock breccias showing multiple episodes of brecciation and recemen- mineralization compared with the jog regions themselves.

GEOLOGY, v. 15, p. 701-704, August 1987 701

Downloaded from https://pubs.geoscienceworld.org/gsa/geology/article-pdf/15/8/701/3509933/i0091-7613-15-8-701.pdf

by Petrobras user
 The Martha Hill lode system at Waihi in the Coromandel Peninsula, Prerupture: uniform fluid pressures
New Zealand (Fig. 2), provides a particularly good "fossil" example of a
mineralized dilational jog comparable in scale to the structures effecting
rupture arrest along active strike-slip faults (cf. Sibson, 1985). This epi-
V
thermal quartz-lode deposit of late Miocene age, hosted in hydrothermally
altered andesites, was mined to a depth of —600 m from 1883 to 1952 and
produced ~ 11001 of Au-Ag bullion. The lode system is flanked by strike-
slip faults that are parallel to regional north-south lineaments discernible
on Landsat imagery (Skinner, 1986), and it comprises a honeycomb mesh Postruplure: fluid influx into jogs
of steeply dipping fissure veins and normal faults striking predominantly
northeast-southwest, along with some subsidiary dextral strike-slip shears
(Morgan, 1924; Brathwaite et al., 1986).
Pertirbation
Wellman (1954) inferred structural control of the veins within a
dextral strike-slip fault system, and it is apparent that the structure can be
interpreted as the exhumed upper levels of a dilational fault jog ~ 1 km
wide. Total extension across the lode system, corresponding to the strike Figure 3. Schematic representation of dilational jogs (idealized as sin-
slip transferred, approaches 100 m. Vein textures record a complex history gle linking extension fractures) in dextral strike-slip fault system; pat-
of incremental extension with repeated brecciation of wall rock and preex- terns of fluid inflow (arrows) induced by rupture perturbation and arrest.
Rupture propagation from left to right.
isting veinfill. Fluid-inclusion studies suggest that much of the mineral
deposition within the jog occurred at depths of 400-1000 m and at precip-
itation temperatures of 220-270 °C during intermittent boiling episodes
(Brathwaite et al., 1986). As with smaller dilational jogs, the driving
strike-slip faults remain almost barren of mineralization. genie zone to depths of about 10 km, and form major obstacles to rupture
propagation (Sibson, 1985). Rupture arrest at such sites is often followed
EARTHQUAKE RUPTURES AND DILATIONAL by a concentration of aftershock activity within the jog and by time-
FAULT JOGS dependent slip transfer to the en echelon fault segment. An explanation for
In active strike-slip systems, larger dilational jogs (> 1 km wide) this behavior is based on the premise that slip transfer across a dilational
appear, from microearthquake locations, to extend throughout the seismo- fault jog necessarily involves the opening of a linking extensional fracture
mesh, as suggested by the fossil jog structures (Fig. 1). Although extension
fracturing is favored by deviatoric stress states within dilational fault jogs
(Segall and Pollard, 1980), there are problems in opening a large exten-
sional fracture network in fluid-saturated crust in periods comparable to
earthquake slip durations, typically a few seconds at most.
The passage of a moderate or larger earthquake rupture is likely to be
only slightly iir.peded by a small dilational jog (width < < 1 km; Fig. 3).
However, at depths greater than a few hundred metres, fluid-pressure
differentials developed between the jog interior and its surroundings by
rapid slip transfer may cause local brecciation by hydraulic implosion of
wall rock into the cavity space, and subsequent fluid inflow (Sibson,
1986). At a laiger dilational jog (width >1 km), the suction induced by
incipient opening of the linking fracture mesh may be large enough for the
jog to act as a major energy sink and halt rupture propagation. Fluid-
pressure differe ntials created by the forced opening of the extensional
fractures can again cause wall-rock brecciation by hydraulic implosion.
Following rupi:ure arrest, the lowered fluid pressure within the jog is
restored by infil tration of fluid from the surrounding rocks. Residual strain
energy at the nipture tip may then be bled off by the slow opening of the
extensional fracture system, allowing delayed slip transfer across the jog
accompanied by aftershocks.
Concentrated aftershock activity within dilational jogs is in accord
with the hypothesis of Nur and Booker (1972) who argued that after-
shocks cluster in regions of reduced mean stress and fluid pressure, the
decay of activity being attributed to fluid inflow and restoration of fluid-
pressure levels. The duration of aftershock activity is highly variable. Eas-
ily felt aftershocks (M >3) may continue for days to weeks following
moderate events (e.g., M ~6), but for large events (M >7), the activity
persists for periods of months to years. Evidence from postseismic effusions
suggests that in some instances as much as 107 m 3 of fluid may be
redistributed in the vicinity of main-shock ruptures during an aftershock
Figure 2. Interpretative map and cross sections of Martha lode system,
sequence, major aftershocks causing second-order perturbations in the
Waihi, New Zealand (after Morgan, 1924, and Brathwaite et al., 1986). inflow (Sibson, 1981).

702 GEOLOGY, August 1987

Downloaded from https://pubs.geoscienceworld.org/gsa/geology/article-pdf/15/8/701/3509933/i0091-7613-15-8-701.pdf

by Petrobras user
 SOUTHERN SAN ANDREAS FAULT SYSTEM minor surface breakage extending northwest of the epicenter to Mesquite
Present seismicity patterns and ground deformation indicate that the Lake is attributed to a major aftershock. In contrast, the 1979 (M s 6.9)
southern San Andreas fault is linked to the spreading centers in the Gulf of Imperial Valley earthquake ruptured predominantly northwestward from
California by a cascade of major dilational fault jogs and dextral strike-slip the vicinity of Mexicali, breaking surface just north of the international
faults (Fuis and Kohler, 1984; Fig. 4). Quaternary mafic-silicic magma- border, to stop at the south end of the Brawley jog (Archuleta, 1984). Up
tism and associated high-temperature (jT>300 °C) geothermal fields, some to 1.6 m of coseismic strike slip was inferred at depth. There are indica-
containing highly metalliferous saline brines, are localized within the jogs tions, however, that the rupture may also have extended subsurface some
(Elders et al., 1984). The jog areas maintain a much higher level of distance southeast into the Cerro Prieto jog (Silver and Masuda, 1985).
background microearthquake activity than the strike-slip fault segments, Eight months later, the 1980 (M s 6.4) Victoria earthquake on the Cerro
including fairly frequent earthquake swarms (Hill, 1977; Johnson and Prieto fault ruptured unilaterally northwestward without breaking surface,
Hutton, 1982). to terminate along the southern bound of the Cerro Prieto jog, adjacent to
Recent large ruptures on the Imperial and Cerro Prieto faults also an active geothermal field. Thus, the arrest of strike-slip ruptures at major
display some intriguing structural relations with the jogs (Fig. 4). The 1940 dilational jogs in this region is well established. Ensuing aftershock activity
(M s 7.1) El Centro earthquake ruptured predominantly southeastward tends to be concentrated in the jog areas (Johnson and Hutton, 1982;
from its epicenter; surface displacements were locally as great as 3 - 4 m Wong and Frez, 1984).
and terminated at the northern end of the dilational jog containing the Notable perturbations of the geothermal field at Cerro Prieto have
Cerro Prieto geothermal field (Trifunac and Brune, 1970). Comparatively accompanied these and other historic earthquakes in the area. The geo-
thermal system occurs adjacent to the late Quaternary rhyodacites of
Cerro Prieto volcano and is manifested at the surface by an area of hot-
spring activity (Laguna volcano), apparently derived from a plume of hot
water ascending obliquely from the northeast (Lippman et al., 1984).
Several extensional normal faults, striking north-south to northeast-
southwest and downthrown to the east, have been delineated within the
geothermal field. Elders et al. (1984) inferred the existence of a young
funnel-shaped gabbroic intrusion, some 4 km across, extending beneath
the Cerro Prieto jog at depths greater than 5 - 6 km. From geochemical
measurements, Valette-Silver et al. (1985) inferred a temperature increase
in excess of 10 °C for fluids within the field between the 1979 and 1980
earthquakes, and minor surface effusions of hot water were noted in the
vicinity of the geothermal plant after the 1980 rupture (Suarez et al.,
1982).
The record of prior historical events in the area is fragmentary and
incomplete, but major earthquake sequences in the Imperial Valley-Colo-
rado delta area took place in 1934, 1927, 1915, 1906, and 1852 (Beal,
1915; Anonymous, 1916; Neumann, 1936; Bureau of Reclamation, 1976).
Mud and steam geysers are reported to have broken out in Laguna volcano
immediately following individual large shocks in all but the 1906 series.
From their epicentral locations or intensity distributions, a reasonable case
can be made that these major shocks in the 1934, 1927, and 1915 se-
quences were analogous to the 1980 Victoria earthquake and occurred on
the Cerro Prieto fault. Hydrothermal eruptions after the 1915 and 1852
shocks were particularly violent; columns of mud and hot water reached
estimated heights of 200-300 m. In 1852, geysering continued for several
months after the main earthquake, reintensifying after major aftershocks
(Bureau of Reclamation, 1976).
Historical strike-slip rupturing on the Cerro Prieto fault thus appears
to have repeatedly triggered major boiling episodes within the geothermal
field. This activity is consistent with the mechanism advanced above for
rupture arrest at dilational fault jogs, boiling being initiated by the abrupt
reduction in fluid pressure accompanying forced extensional fracturing at
the propagating rupture tip.

DISCUSSION
The structure and vein textures of the Martha Hill lode system and
similar fault-hosted epithermal deposits are consistent with the mechanism
Figure 4. Dilational fault jogs in southern San Andreas fault system (in inferred for rupture arrest or perturbation at dilational fault jogs, and with
part, after Fuis and Kohler, 1984). Stipple indicates areas of high mi- the historical record of hydrothermal eruptions triggered by strike-slip
croearthquake activity (SSJ = Salton Sea jog; BJ = Brawley jog; CPJ = rupturing along the Cerro Prieto fault. Episodes of boiling and rapid
Cerro Prieto jog). Circles = epicenters of major recent strike-slip rup-
mineral precipitation within dilational jogs are therefore envisaged as fol-
tures; arrows indicate directions of rupture propagation; asterisks =
Quaternary volcanic centers; triangles = geothermal fields (larger trian- lowing major ruptures on the flanking faults, deposition continuing
gles are T >300 °C fields). through the aftershock period of wall-rock readjustment and fluid inflow.

702 GEOLOGY, August 1987

Downloaded from https://pubs.geoscienceworld.org/gsa/geology/article-pdf/15/8/701/3509933/i0091-7613-15-8-701.pdf

by Petrobras user
 Individual aftershocks m a y cause secondary perturbations in deposition. Morgan, P.G., 1924, The geology and mines of the Waihi District, Hauraki Gold-
S o m e mineral deposition m a y also a c c o m p a n y the steady ascent of a field: New Zealand Geological Survey Bulletin 26, 218 p.
Neumann, F., .936, United States earthquakes—1934: Washington, D.C., U.S.
converting hydrothermal plume, but the a b r u p t change in P-T conditions
Department of Commerce Coast and Geodetic Survey, Serial no. 593, 101 p.
engendered b y rupture-induced extensional fracturing must strongly favor Nur, A., and Booker, J., 1972, Aftershocks caused by pore fluid flow?: Science,
concentrated local precipitation. v. 175, p.!>85-887.
T o w h a t depths m a y boiling extend? F o r p u r e w a t e r under hydro- Perry, V.D., 1952, Geology of the Chuquicamata orebody: Mining Engineering,
static fluid pressure, the m a x i m u m boiling depth is a b o u t 2.2 k m , but this v. 4, p. 1166-1168.
Segall, P., and Pollard, D.D., 1980, Mechanics of discontinuous faults: Journal of
is affected b y the presence of dissolved salts or gases such as C O 2 (Henley,
Geophysical Research, v. 85, p. 4337-4350.
1985). Given that dilational jogs in strike-slip systems m a y extend to Sibson, R.H., 1981, Fluid flow accompanying faulting: Field evidence and models,
depths of - 1 0 k m a n d that a b r u p t reductions in fluid pressure are ex- in Simpson, D.W., and Richards, P.G., eds., Earthquake prediction: An inter-
pected t o a c c o m p a n y r u p t u r e arrest at all levels within t h e jogs, boiling national review: American Geophysical Union, Maurice Ewing Series 4,
p. 593-60; I.
episodes m a y begin at substantially greater depths than in the hydrostatic
1985, Stopping of earthquake ruptures at dilational fault jogs: Nature, v. 316,
case. H o w e v e r , sustained boiling a n d mineral deposition still seem likely to p. 248-251..
b e restricted to t h e u p p e r m o s t 1 - 2 k m . T h e m a g n i t u d e s of the suctions 1986, Brecciation processes in fault zones: Pure and Applied Geophysics,
opposing slip transfer will b e s o m e w h a t reduced w h e r e boiling occurs. v. 124, p. 159-175.
F r o m the exploration viewpoint, recognition of the critical structural Silver, P., and Masuda, T., 1985, A source extent analysis of the Imperial Valley
earthquake of October 15,1979, and the Victoria earthquake of June 9,1980:
relations is h i n d e r e d b y t h e b a r r e n n a t u r e of t h e driving faults in compari- Journal of Geophysical Research, v. 90, p. 7639-7651.
son with the mineralized jogs, w h i c h has o f t e n led to t h e erroneous as- Skinner, D.N.B., 1986, Neogene volcanism of the Hauraki volcanic region: Royal
s u m p t i o n that related structures are of different ages. Proper understanding Society of Sew Zealand Bulletin 23, p. 21-47.
of these deposits clearly requires an appreciation of the d y n a m i c s of earth- Spurr, J.E., 1925, The Camp Bird compound veindike: Economic Geology, v. 20,
p. 115-15;:.
q u a k e rupturing; t h e fault-fracture systems c a n n o t simply be regarded as
Suarez, F., Sieh, K.E., and Elders, W.E., 1982, A review of geological effects and
passive fluid conduits. damage distribution of the June 9, 1980 Mexicali earthquake: Earthquake
Engineering Research Institute Newsletter, v. 16, p. 99-105.
REFERENCES C I T E D Tchalenko, J.S. and Ambraseys, N.N., 1970, Structural analysis of the Dasht-e-
Anonymous, 1916, The earthquake at Volcano Lake, Mexico, November 20,1915: Bayaz (Iran) earthquake fractures: Geological Society of America Bulletin,
Seismological Society of America Bulletin, v. 6, p. 181-185. v. 81, p. 41-60.
Archuleta, R.J., 1984, A faulting model for the 1979 Imperial Valley earthquake: Trifunac, M.D., and Brune, J.N., 1970, Complexity of energy release during the
Journal of Geophysical Research, v. 89, p. 4559-4585. Imperial Valley, California, earthquake of 1940: Seismological Society of
Beal, C.H., 1915, The earthquake in the Imperial Valley, California, June 22,1915: America Bulletin, v. 60, p. 137-160.
Seismological Society of America Bulletin, v. 5, p. 130-149. Valette-Silver, N.J., Silver, P.G., and Ricardo, M.M., 1985, The effect of a tectonic
Brathwaite, R.L., McKay, D.F., and Henderson, S., 1986, The Martha Hill gold- disturbance on the Cerro Prieto geothermal field (B.C. Mexico): Geothermal
silver deposit, Waihi, in Brathwaite, R.L., Browne, P.R.L., and Roberts, P.J., Resource Council Transactions, v. 9, p. 515-520.
eds., Proceedings of Symposium 5: Volcanism, hydrothermal systems, and Wellman, H . W , 1954, Stress pattern controlling lode formation and faulting at
related mineralisation: University of Auckland, International Volcanological Waihi Mir e, and notes on the stress pattern in the north-western part of the
Congress, p. 19-23. North Island of New Zealand: New Zealand Journal of Science and Technol-
Bureau of Reclamation, 1976, Record of earthquakes in the Yuma area, Yuma ogy, v. 36E>, p. 201-206.
County, Arizona; Imperial County, California: U.S. Department of the Wong, V., and Frez, J., 1984, Aftershock locations and fault mechanisms: Earth-
Interior, Bureau of Reclamation, Special Report, 73 p. quake Engineering Research Institute Newsletter, v. 16, p. 76-79.
Elders, W.A., Bird, D.K., Williams, A.E., and Schiffman, P., 1984, Hydrothermal
flow regime and magmatic heat source of the Cerro Prieto geothermal system, ACKNOWLEDGMENTS
Baja California, Mexico: Geothermics, v. 13, p. 22-47. Supported by National Science Foundation Grant EAR 86-17445.1 thank Pat
Fuis, G.S., and Kohler, W.M., 1984, Crustal structure and tectonics of the Imperial Browne, Steve Richard, and Frank Spera for comments; Alfred Truesdell and Paul
Valley region, California, in Rigsby, C.A., ed., The Imperial Basin—Tectonics, Bodin for information on the Cerro Prieto geothermal field; Bob Brathwaite of the
sedimentation and thermal aspects: Los Angeles, Pacific Section, Society of New Zealand Geological Survey for alerting me to the significance of the Martha
Economic Paleontologists and Mineralogist, p. 1-13. lode system; and Don McKay of the Waihi Gold Company for granting access to
Grindley, G.W., and Browne, P.R.L., 1976, Structural and hydrological factors the Martha Min ;.
controlling the permeabilities of some hot-water geothermal fields, in Proceed-
ings, United Nations Symposium on Development and Use of Geothermal
Resources, 2nd, San Francisco: Washington, D.C., U.S. Government Printing Manuscript received February 9, 1987
Office, p. 377-386. Revised manuscript received April 20,1987
Henley, R.W., 1985, The geothermal framework of epithermal deposits, in Berger, Manuscript accepted May 12, 1987
B.R., and Bethke, P.M., eds., Geology and geochemistry of epithermal systems:
Society of Economic Geologists, Reviews in Economic Geology, v. 2, p. 1-44.
Hill, D.P., 1977, A model for earthquake swarms: Journal of Geophysical Re-
search, v. 82, p. 1347-1352.
Reviewer's comment
Hulin, C.D., 1929, Structural control of ore deposition: Economic Geology, v. 24,
p. 15-49.
Johnson, C.E., and Hutton, L.K., 1982, Aftershocks and pre-earthquake seismicity: Intriguing ideis, linking earthquakes, ore deposits, rock mechanics, a n d
U.S. Geological Survey Professional Paper 1254, p. 59-76. fluid f l o w — c o n s e q u e n t l y of w i d e interest. Ideas are speculative, b u t will
Lippman, M.J., Goldsten, N.E., Halfman, S.E., and Witherspoon, P.A., 1984, Ex- lead to f u r t h e r w o r k , b y others, including isotopie studies, seismicity, a n d
ploration and development of the Cerro Prieto geothermal field: Journal of
structural geology.
Petroleum Technology, v. 36, p. 1579-1591.
McKinstry, H.E., 1948, Mining geology: Englewood Cliffs, New Jersey, Prentice- Amos Nur
Hall, Inc., 677 p.

704 Printed in U.S.A. GEOLOGY August 1987

Downloaded from https://pubs.geoscienceworld.org/gsa/geology/article-pdf/15/8/701/3509933/i0091-7613-15-8-701.pdf

by Petrobras user