JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 88, NO.

B2, PAGES 1153-1172, FEBRUARY 10, 1983

Mechanics of Fold-and-Thrust Belts and Accretionary Wedges

DAN DAVIS

Department of Earth and Planetary Sciences, MassachusettsInstitute of Technology

Cambridge, Massachusetts 02139

JOHN SUPPE AND F. A. DAHLEN

Department of Geological and Geophysical Sciences, Guyot Hall, Princeton University

Princeton, New Jersey 08544

The overall mechanicsof fold-and-thrust belts and accretionary wedges along compressiveplate
boundaries is considered to be analogousto that of a wedge of soil or snow in front of a moving
bulldozer. The material within the wedgedeformsuntil a critical taper is attained, after which it slides
stably, continuing to grow at constant taper as additional material is encountered at the toe. The
critical taper is the shapefor which the wedgeis on the verge of failure under horizontal compression
everywhere, including the basal decollement. A wedge of less than critical taper will not slide when
pushed but will deform internally, steepeningits surface slope until the critical taper is attained.
Common silicate sedimentsand rocks in the upper 10-15 km of the crust have pressure-dependent
brittle compressive strengthswhich can be approximately represented by the empirical Coulomb
failure criterion, modified to account for the weakening effects of pore fluid pressure. A simple
analytical theory that predicts the critical tapers of subaerial and submarine Coulomb wedges is
developedand tested quantitatively in three ways: First, laboratory model experimentswith dry sand
match the theory. Second, the known surface slope, basal dip, and pore fluid pressuresin the active
fold-and-thrust belt of western Taiwan are used to determine the effective coefficient of internal
friction within the wedge,/x = 1.03, consistentwith Byerlee's empirical law of slidingfriction,/at, =
0.85, on the base. This excessof internal strengthover basalfriction suggeststhat althoughthe Taiwan
wedgeis highly deformedby imbricate thrusting,it is not so pervasivelyfractured that frictional sliding
is always possibleon surfacesof optimumorientation. Instead, the overall internal strengthapparently
is controlled by frictional slidingalong suboptimallyoriented planesand by the need to fracture some
parts of the observed geometrically complex structure for continued deformation. Third, using the
above valuesof/at, and/x, we predictHubbert-Rubeyfluid pressureratios X = Xt,for a numberof other
active subaerial and submarine accretionary wedges based on their observed tapers, finding values
everywhere in excess of hydrostatic. These predicted overpressuresare reasonable in light of
petroleum drilling experience in general and agree with nearby fragmentary well data in specific
wedges where they are available. The pressure-dependentCoulomb wedge theory developed here is
expected to break down if the decollementexhibits pressure-independentplastic behavior becauseof
either temperatureor rock type. The effectsof this breakdownare observedin the abrupt decreasein
taper where wedge thicknessesexceed about 15 km, which is the predicteddepth of the brittle-plastic
transitionin quartz-rich rocks for typical geothermalgradients.We concludethat fold-and-thrustbelts
and accretionarywedgeshave the mechanicsof bulldozerwedgesin compressionand that normal
laboratory fracture and frictional strengthsare appropriateto mountain-buildingprocessesin the upper
crust, above the brittle-plastic transition.

INTRODUCTION present a simple mechanicalmodel of fold-and-thrustbelts

The zones of folding and thrusting that lie along the
margins of many mountain belts constitute one of the most
widely recognizedand best understooddeformationalfea-
tures of the earth [McClay and Price, 1981]. Previousmodels
and theories of the mechanics of fold-and-thrust belts have
been basedlargely on observationsof belts that are no longer
tectonically active, simply because there has been more
geologicalexploration of these inactive areas (for example,
Figures la and lb). Additional insight may be gained by
examining the relatively few active fold-and-thrust belts,
suchas the Himalayan foothills and western Taiwan (Figure
l c). Furthermore, a great deal of geophysicalinformation is
now available for the oceanic analogue of fold-and-thrust
belts, namely, the accretionary wedgesin front of island arcs
(for example, Beck and Lehner [1974], Seely et al. [1974],
Hamilton [1979], and Nasu et al. [1979]). In this paper we
Copyright1983by the AmericanGeophysicalUnion.
Paper number 2B 1741.
0148-0227/83/002B- 1741$05.00
1153
and accretionarywedgesbasedon active examples,particu-
larly western Taiwan.
Although considerablenatural variation exists among the
many fold-and-thrust belts and accretionary wedges of the
earth, severalcommonpropertiesoccur in crosssectionthat
might form the basis of a successfultheory of their mechan-
ics. Chappie [1978] emphasized that fold-and-thrust belts
and accretionary wedges exhibit: (1) a basal surface of
detachment or decollement, below which there is little
deformation, dipping toward the interior of the mountain
belt; (2) large horizontal compressionin the material above
the basal decollement; and (3) a characteristic wedge shape
of the deformedmaterial, taperingtoward the marginof the
mountainbelt (see Figure 1). The overall mechanicsof these
wedgesalongcompressiveplate boundariesis consideredby
Chappie [1978], Davis and Suppe [1980], Stockmal and
Chappie [1981], and othersto be analogousto the mechanics
of wedges that form in front of moving bulldozers or snow
plows. Such bulldozer and snowplowwedgeshave the same
three properties, but on a smaller scale. The soil, snow, or
rock deformsuntil the wedge attains a steady state or critical
1154 DAVIS ET AL.' MECHANICSOF FOLD-AND-THRUSTBELTS

A
w

1o km Canadian Rockies

B
w

10 km J Southern Appalachians

c
E V

x x

10 krn
I !

Western Taiwan

Fig. 1. Crosssections
of several
forelandfold-and-thrust
belts:(a) CanadianRockies[afterBallyet al., 1966],(b)
southern
Appalachians[afterRoedereta!., 1978],and(c) western
Taiwan[afterSuppe,1980a].

taperandthenslidesstably,continuing
to growself-similarly A sequenceof photographsof one experimentalrun is
as additional material is accreted at the toe. A critically shownin Figure 3. As deformationprogresses,the wedge
taperedwedgethat is not accretingfresh materialis the thickens and the locus of active deformation moves away
thinnestbody that can be thrustover its basaldecollement from the rigid buttressuntil the taperreachesits steadystate
withoutany internaldeformation; it is thuson the vergeof critical value. The wedgethen slidesstably. If sandcontin-
shearfailure everywhere.In contrast,a critically tapered ues to be accreted at the toe, the wedge continuesto grow,
wedgethat is accretingfresh materialdeformsinternally maintainingits critical taper. Quantitativeaspectsof these
while slidingin order to accommodatethe influx and to experimentsare discussedlater in this paper.
maintain its critical taper. We choose to model fold-and-thrust belts and accretionary
This wedgebehaviormay be illustratedon the laboratory wedgesas critically tapered deformingwedges of rock
scaleusinga simplemechanicalmodel[Davis, 1978].The analogous to the wedgeof homogeneous sandin the experi-
model consists of sand contained in a bottomless box with ment. Such a model is intended, of course, only to represent
transparent sidewallsthat sitsupona sheetof Mylar, which the gross macroscopicmechanicsand not the complex
in turn lies on a flat rigid base(Figure 2). In most sandbox internal structural details or deformational histories of spe-
deformationalmodels,horizontalcompressionis inducedby cificwedges.In the caseof an accretionarywedgein front of
pushingone wall of the model. Here, the sameeffect is an islandarc, the bulldozerrepresentsthe lithosphericplate
achievedby pullingthe Mylar sheetupon which the sand beneathwhich oceaniccrust is being subducted,and in the
rests;thispullsthe sandagainstthe backwall of the model, caseof a fold-and-thrustbelt, it representsthe inner thicker
which servesas a rigid buttress.The apparatusmimicsthe partof the mountainrange.We shouldalsonotethat a metal
processof plate subductionand allowslarge deformation bulldozeris madeof muchstrongermaterialthan the materi-
with mechanicalsimplicity.Inhomogeneities are minimized al it pushes;this is not true in the earth,in whichrigid, less
by packingthe sandevenly,and sidewall frictionis effec- easilydeformedsegmentsof the lithosphereare generally
tively reducedby coatingwith graphite. strongerlargely becausethey are thicker.
 DAVIS ET AL.: MECHANICSOF FOLD-AND-THRUST
BELTS 1155

Take-up Spool

Fig.2. Schematic
diagram
of laboratory
sandbox
model
ofCoulomb
wedge.
Underlying
Mylarsheet
ispulled
out
beneath buttress at left.

PROPERTIES OF THE MODEL

Here, as well as elsewherein this paper,we have usedan
A fundamentalpremiseof the presentmodelis that rock asterisk
to denote
theeffective
stress
trn*- trn- pf.
deformation
in the upperlithosphere
is governedby pres- Porefluidpressurespf playa majorrolein themechanics
sure-dependent,time-independentCoulomb behavior, i.e., of thrustfaulting[HubbertandRubey,19591, anda proper
by brittle fracture [Paterson,1978] or frictional sliding accounting of their effect in both subaerial and submarine
[Byedee,1978].Macroscopically, the empiricallaw describ- wedges requires some care. It is common in the subaerial
ing boththesemodesof brittlebehavioris the same,only case to assume that the water table is located at the rock
differingin the numericalvalueof the cohesivestrengthSo surface(Figure4a). The porewaterpressureis in that case
andthe coefficientof frictionta= tan 4,whichexpressesthe typically equal to the hydrostaticpressurefor shallow
pressure dependence. The general Coulomb criterion for depths. However, below some point whose location is
shear traction r at failure is of the form stratigraphically controlled, it often rises well above the
hydrostatic
pressure,
typicallyattaininga constant
gradient
somewherein the range betweenthe hydrostaticand the
wheretr is the normaltractionandpf is the porefluid lithostatic
gradient[Fertl,1976].Thisgradient
is commonly
pressure. describedin termsof Hubbertand Rubey'sdimensionless
CohesionSois relativelyunimportantin the mechanicsof parameter
X, theratioof porefluidpressure
pf to thevertical
fold-and-thrust
beltsandaccretionary wedgesthat are com- normal traction exerted by the lithostaticoverburden.
posedpredominantlyof silicatesediments.Typical labora- In manyregionsthe water table may be locatedbelowthe
tory measurementsof Sofrom fractureexperimentson small uppersurfaceof the rock(Figure4b). Fluid pressures
are in
samplesof shalesand sandstones rangefrom 5 to 20 MPa thatcasereduced,andtherockwill berelativelystronger. A
[Hoshino et al., 1972], which is several times less than the third circumstance is that which arisesin dealingwith
cohesionof granitesand other strongerrocks.In addition, accretionary wedges: the rock surface in that case is sub-
these measurementsprobably overestimatethe cohesive mergedbelow sea level (Figure4c). Rocksjust below the
strengths
of in siturockmasses becausetheyarecommonly seafloorwill generally be quite weak, since the effective
done on relatively pristine samples and because rock lithostaticpressuretherealwaysvanishes.To dealwith this
strengthdecreaseswith larger samplescontainingmore submarine caseit is necessary
to introducea slightgeneral-
flaws. At depthson the order of a few kilometersthe effectof izationof HubbertandRubey's subaerialX, andthat is done
this relativelysmallcohesionis negligiblein comparison below; see (6).
withthepressure-dependent termin thefailureequation(see Thelowerlimitof applicability
of thepresentbrittlewedge
Figure 6a, below, for an illustration).The main effectsof theory is the middle or lower crust where pressureand
cohesionon wedgegeometrywill be observednearthe toe of temperaturebecome sufficientlyhigh that common rocks
the wedge,wherecohesioncanadd significantlyto the total beginto displaytemperature-dependent plasticbehavior,
strengthand produce a critical taper narrower than the deforming by motionof dislocations and otherpressure-
correspondingcohesionlesstaper. Farther from the toe independent mechanisms. The expectedeffectsof this brit-
wherethewedgeis thicker,thepressure-dependent termwill tle-plastictransitionon the geometryof largewedgesare
dominate, approach discussed
andthecriticaltaperwill asymptotically furtherat the endof this paper.
the cohesionlessvalue. For simplicitywe shallignorethe
effectof cohesionon toe shapeentirelyin this paperand THEORETICAL ESTIMATION OF CRITICAL TAPER
employ instead of (1) a noncohesivefailure law of the form Considera homogeneous
wedgeof deformablenoncohe-
siveCoulombmaterialbeingcompressed by a pushfromits
(2) thickend, slidingalonga rigidbase(Figure5). Let a be the
1156 DAVIS ET AL..' MECHANICS OF FOLD-AND-THRUST BELTS

Fig. 3. Photographicsideview of stagesin deformationof sandduringan experimentalrun. Initially undeformed

sandmassis increasingly
compressed anddeformedby thrustinguntilthecriticaltaperisattained.Blacksandlayersare
passive markers.

local angleof topographicrelief and/3 be the local dip angle /3)wherepwis the densityof water. For greatergeneralitywe
of the rigid base, as shown. A system of local Cartesian shallconsiderexplicitlythe caseof a submarinewedgein the
coordinatesx, z will be employed, having x parallel to the analysisthat follows, with the knowledgethat the corre-
base in the segmentof the wedge between x and x + dx and spondingresult for the subaerialcase may be recoveredat
with z increasingupward. The local thicknessof the wedge, any point in the argumentby settingpwequal to zero (or
measuredalong the z axis, will be denoted by H. In the case morepreciselyto the densityof air, whichis negligible).The
of a submarine wedge the local thickness of the water third force, which acts on both subaerial and submarine
overburden, measuredalong the direction of gravity, will be wedges,is the frictionalresistanceto slidingalongthe basal
denoted by D. decollement.In termsof the basalsheartractionrbthisforce
The critical taper of such a horizontally compresslye is -r dx, the minus signarisingbecausethe force acts to
wedge is governed essentially by the balance of forces in the resistsliding.Finally, let or,be the normaltractionacting
x direction. Let us enumerate the forces acting on the acrossany face perpendicularto the x axis, with compres-
segment of unit length along strike (i.e., in and out of the sionbeingreckonedpositive.The fourthforceactingon the
page) lying between x and x + dx. There is, first, a segmentbetweenx andx + dx is thenthe resultantcompres-
gravitationalbody force whosex componentis -pgH dx sin sive pushof these normal tractionsactingon the two side
/3 where p, assumedconstant, is the density of the rocks in walls. In contrastto the first three forces,this resultantis the
the wedgeand g is the accelerationof gravity. Second,in the +x direction,for two reasons:not onlyis thex facelargerin
case of a submarine wedge there is another force arising areathanthat at x + dx, but becausethe pushis comingfrom
directly from gravity, namely, the pressure of the water the thick end, cr(x,z) exceedscr(x + dx, z). Balancing
overburden, which has an x component -pwgD dx sin (a + forces requires in the limit dx - 0 that
 DAVIS ET AL.' MECHANICS OF FOLD-AND-THRUST BELTS 1157

ro$$uro
sea level
watertableat rocksurface
rock surface

bottom

::::::::::::::::::::::::::

::::::::::::::::::::::::::::
'..:.:::.:.:.:.:.:.:.:::.:.:..
::::::::::::::::::::::::::::::
.

=================================

...........
:..:.:.:::::::.

A B C
Fig. 4. Relationshipbetweenfluid pressurePt and verticalnormaltractionor.in three cases:(a) water table at
surface;(b) water table below surface' (c) rock submerged.In all three casesthe Coulomb shearstrengthis proportional
to the difference.* = tr: - Pt, shownby shading.

pgH sin/3 + pwgDsin (a + /3) + rb This assumption also depends for its validity on a small-
angle approximation. By defining a dimensionlessquantity
+ o'xdz= 0 (3)
dx
X= (6)
The thin-skinned nature of actual wedges allows us, with Crz- pwgD
very little error, to make use of a small-angleapproximation
we may write the effectivenormaltractiono'z*= rrz- PTin
in which sin a a and sin/3 /3. With this substitution,(3) the form
reduces to the simpler form
Crz*= (1 - X)pg(H- z) (7)

pgH+ pwgD(a
+ ) + rb+xx o'xdz
=0 (4) Equation (6) provides the appropriategeneralizationto the
case of a submarinewedge of the pore fluid pressureratio X
The vertical normal traction, o-z,at any point in the wedge
will be assumedto be solely that due to the lithostatic and = pf/o'zof HubbertandRubey.In essence
the hydrostatic
pressure at the seafloor, pwgD, has been ad9pted as a
hydrostatic overburden, i.e.,
reference level; inspection of Figures 4a and 4c makes it
rrz = pwgD + pg(H- z) (5) clear that this is a logical procedure.

sea level

Fig. 5. Schematicdiagramof a wedge of material subjectto horizontalcompressionand on the verge of Coulomb

failure throughout.The force balanceon an arbitrary columnof width dx is shownand the terminologyusedin deriving
the equations of critical taper is indicated.
1158 DAVIS ET AL.' MECHANICSOF FOLD-AND-THRUST
BELTS

0 C *

Fig. 6. Mohrdiagram illustrating

thestateof stress
(a) at somepointwithinthewedge and(b)at thebaseof the
wedge. Thequantities 4and4,aretheangles ofinternalandbasal friction,
andand%,aretheangles between o-and
thex axiswithinthewedge andat thebaseof thewedge. Thebasalsheartraction r/,isgivenbytheintersection
of the
frictional
failurelaw Irl = ,rr,* withtheMohrstress circlecorresponding
to thebasaldepthH.

The traction'0resistingfrictionalslidingon the basewill the fact that the basal decollementwill usually be a zone of
be written as weakness,either becauseof a lower intrinsic strengthor
because of elevated fluid pressures.For a wedge with
'0 = /XbO'z
*= ,U,b(1- Ko)pgH (8)
uniforminternalproperties/x and X, we must necessarily
where/x0= tan (b0is thebasalcoefficient
offrictionand)tois have(1 - .b)ld,
b (1 - X)/xfor the baseof the wedgeto be a
the generalizedHubbert-Rubeyratio (6) on the base. In throughgoingdecollement.
introducing
thebasalvaluesgoand)towe allowexplicitlyfor To determinethe remainingunknown quantity tr; in the
 DAVIS ET AL.' MECHANICS OF FOLD-AND-THRUST BELTS 1159

force-balance equation (4), we must consider the state of Even in a wedge with uniform properties the angle will
stress within the wedge in more detail. A Mohr-circle vary appreciablywith depth, particularly if ht, h and tab=
representationof the stressat an arbitrary point is shown in
Figure 6a. The maximum and minimum effective compres- For convenience, after removing the factor of (1 - h) from
sive stresseshave been denoted by 0-1' and 0-3*, respective- under the integral sign in (16), let us define a dimensionless
ly. The local angle between the axis of maximum compres- quantity
sive stressand the x axis will be denoted by (see Figure 5).
Since a critically tapered wedge is assumed to be on the
verge of shearfailure throughout, there will be at every point
two planes oriented at angles -+(rr/4 - 6/2) with respect to
K=2H-
If;csc<bsec
dz
2z)- 1 (17)
When (8) and (16) are substituted into (4), we find the
the 0- axis on which the failure criterionIt] = ta0-.* is
theoretical critical taper a + / of a noncohesiveCoulomb
satisfied[Jaeger and Cook, 1969]. By inspection of triangle
wedge to be
BCD in Figure 6a we see that

(o-x*-- O-z*)= (0-1'-- 0'3*)COS

2 (9) (1 - hb)tab+ (1 -
a + /3 = (18)
(1 - Pw/P)+ (1 - X)K
whereas from triangle OAC we find
It is noteworthy that this equation contains no explicit
(0'1' - 0'3*)= -}(o'x*+ O'z*)sinb (10)
dependenceon x so that a wedge with uniform properties
Combiningthese expressions,we find the stressdifference and a planar base should have a constant surface slope.
(0'x- 0'0 = (o'x*- O'z*)to be Note, in addition, that in the limit hb --> h --> 1, (18) reduces
to a --> 0, which is the expected result for a wedge composed
O'z* (1 - h)pg(H- z) of material having negligible strength. For future reference,
(rrx- rrz)=
csc & sec 2- 1 csc & sec 2- 1 we remark that (18) defines a linear relationship between a
and/ of the form
(11)
a+R/3= F (19)
This equation together with the correspondingresult for the where
shear stress
(1 - h)K
R = (20)
(1 - h)pg(H - z) tan 2 (1 - p.,/p) + (1 - h)K
rxz= (0'x- o'z)tan2=
csc (b sec 2q- 1
(1 - hb)tab
(12) F = (21)
(1 - Pw/P)+ (1 - h)K
completely define the state of plane stresswithin a noncohe- To determine the quantity K exactly, we must know the
sive critical Coulomb wedge in terms of a single parameter, angle (z) at every depth in the wedge 0 -< z -< H and this
the stressorientation angle . cannot be found entirely analytically. In a forthcoming
Upon combining (5) and (11) we are now able to write the paper, we will show how (z) can be found numerically by a
integral appearing in the force balance equation (4) in the very straightforwardand rapid procedure, in both the caseof
form
a noncohesivewedgeas well as one with a finite cohesionSo.
In the present paper we shall employ instead a simple

o
O'xdz
=pwgDH
+pgH
2 analytical approximation to K, which we have tested thor-
oughly against our exact noncohesive numerical results and
found to be extremely accurate. To develop this approxima-

+2pg
f; (1-X)(H-z)
dz(13) csc <bsec 2- 1
tion, we shall, for the moment, assume that ht, = h; the
wedge existence condition is in that case simply tat,-< ta and
Invokingthe small-angleapproximationonce again,we set the critical taper equation (18) reduces to
dH/dx = -(a + 13) (14) (1 - h)tat,+ (1 - pw/p)13
a +/ = (22)
dD/dx = a (15) (1 - Pw/P)+ (1 - h)K

so that We note first that at the top surface of the wedge, if there is
the slightest cohesion, 0-1 must be parallel to the local
topography so that t = (H) = a + /. In the small-angle
dx 0-x
dz= pwgHa
- p,,gD(a
+ ]3)- pgH(a
+ approximationa +/ << 1 the integrandin (17) near the top
of the wedge can thus be replaced by
- 2pg(, + /3)
I0
u (- x)
csc <bsec 26- 1
dz (16)

csc&sec2t-
1
1
sin
1 - sin&
(23)

In writing (16) we have assumedthat all of X, b, and O are

independent of x, but they may all in general still depend The angle t,= (0) near the base of the wedge can be
upon depth z. For simplicity, however, we shall henceforth found from the specificationthat frictional sliding governed
considerboth X and bto be strictly constant, regardingthem by a coefficient of friction tat, is occurring there; a Mohr-
as the averaged or effective values for the wedge as a whole. circle representationof the basalstate in the caseht, = h and
1160 DAVIS ET AL' MECHANICS OF FOLD-AND-THRUST BELTS

functionof/ for fixed/0. An internally strongerwedgecan

be narrower and still undergo stable slidingwithout deform-
;'J= 7.75 ing. To obtainan approximationvalid in the caseh0 :/:h, it is
necessaryto replace &0in (28) by arctan [&0(1- h0)/(1 - h)].
In all of the specificcomparisonswith laboratory and field
7.0
observations considered below we have in fact assumed that
h0 = h, so that the simpler forms (22) and (28) are pertinent.

O.85 Before considering any applications of the critical taper

formulae (18) and (28), it is of interest to point out another

0.7' more fundamental form of the force balance

wedge, obtained by inserting (16) into (4). This leads to a
in a critical

- 0.55 formula for the basal shear traction r0, namely,

ro = (p- pw)gHa+ (1 - h)KpgH(a + 13) (29)

This equation shows that the frictional sliding resistanceon

0 I I I i i
0 0.4 08 12 the base is counteracted by two terms, both of which may be
basal friction/ub thought of as driving forces responsible for the tectonic
deformation and overthrusting. The first term is due to
Fig. 7. Sensitivity of coefficientK to basaland apparent internal
coefficientsof friction/xt, and/x. If/xt, is smallin comparisonwith/x, gravity acting on the slopingtop surfaceof the wedge and is
K 2 sin b/(1- sin &), but as/xt,-->/x, K decreasessignificantlyand the same as that which causes a glacier to slide in the
in fact OK/Otb -- -. direction of its surface slope regardlessof the attitude of its
bed. This term has been discussedin the context of gravity
/0 < / is shownin Figure 6b. By applyingthe law of sinesto slidingor spreadingmodelsof overthrusttectonicsby Elliott
the two triangles OAC and OBC it may be shown that [1976].The effectivedensity p - pwappearsnaturallyin (30)
in the case of a submarine wedge, as expected on intuitive

orr
4 &o 1 (sin
&o)
- arccos (24) grounds.The secondterm dependson the taper a +/3 and is
2 2 ksin
& a consequence of the horizontally compressive push as-
sumed to be acting from the rear of the wedge. For typical
and,after'
some
algebra, silicate sedimentary accretionary wedges with geologically
1 sin2&o+ cos&o(sin
2 &- sin2 &o)1/2 reasonablevalues of/x, /20 and X = X0, this secondterm is
larger than the first by a factor of 4 to 5, which implies that
csc & sec 20 - 1 cos2 &o- cos&0(sin
2 d>- sin2 &0)1/2
basal shear tractions in compressive, critically tapered
(25) wedgesare larger by the samefactor than thoseestimatedby
Elliott [ 1976].
In the limiting case &o--> & At this point in our discussion it is also appropriate to
indicate the relationship between our analysis and the well-
1 known Hubbert and Rubey [1959] analysis of the maximum
-->tan2 &o (26) possible lengths of thrust sheets. To illustrate the effect of
csc & sec 2qto- 1
fluid pressure on thrust faulting, they computed the maxi-
If on the other hand the basal layer is very weak, i.e., &o << mum length of a rectangular body (a + /3 = 0), of a given
&, then (z) is small everywhere and

1 sin 4>
= (27) 6
csc 4 sec 20 - 1 1 - sin &
Theoretical Prediction
To obtain an approximate expressionfor K valid in either
limit, we approximate the integrandin the interval 0 -< z -< H
by a linear function passingthrough the two end points (23) 4
and (25); this leads to

sin& sin2 qbo

+ cosqbo(sin
2 qb- sin2 )b)1/2
K= +
1 - sin 4> cos2 &o- cos&o(sin
2 4>- sin2 &o)1/2 2
Experimental

Determinations

(28)

This, together with (22), can be used to find the theoretical

critical taper a +/3 of an actively deformingCoulomb wedge 0o I I I I I I I
0 o 2 4 6
with prescribedphysical properties/, /0, and X0 = X. The
Dtp of Base
nature of the variation of K with / and /0 is illustrated in
Figure 7. An increase in basal friction increases the critical Fig. 8. Mean surface slope measured by linear regressionoE
photographsversus dip of rigid base in sandboxexperiments. Dots
taper becauseof the factor/0 in the numerator of (22), but representthe averageof eight experimental runs at
an increase in the effective internal friction leads to a
= 2, fourteen at /3 = 3, and nine at /3 -- 6. Line is theoretical
decrease in the critical taper, since K is an increasing predictiona = 5.9 - 0.66/3.
 DAVIS ET AL.' MECHANICS OF FOLD-AND-THRUST BELTS 1161

Ir;ornote Sh;rna

ooo

ooo

5000

--------- 6000 TR
N
/
RYUK
Y.._-,ooo

oo

S OUT H L.-..

/
Batan I
CHINA

SEA
Babu yah I

, 1Oo
o

Fig. 9. Tectonic and bathymetricsettingof Taiwan with velocity trianglefor arc-continentcollisionin Taiwan
assumingplate motions of Seno [1977].

thickness, that can be pushed along a basal decollement
withoutundergoinginternalCoulombfailure. Their analysis
is appropriate to the immediate vicinity of the toe of a
deformingwedge, where flat-lyingsedimentsare being ac-
creted. A thrust sheetlongerthan the Hubbert-Rubeymaxi-
mum deforms by thrusting and thickening toward its rear
until its geometryis that of a criticallytaperedwedge.The
once perplexing mechanicalproblems posed by the exis-
tence of large-scale thrust sheets and fold-and-thrust belts
are, in a sense,an artifact of imposinga fixed rectangular
geometry;this point hasbeenemphasizedby Chappie[ 1978]
also. Whereas Hubbert and Rubey calculatedthe maximum
lengthoverthrustattainablewithout a taper, we calculatethe
critical taper requiredfor the emplacementof a thrust sheet
of any length.A morequantitativecomparisonof our results
and theirs, includingthe role of cohesion,will be published
elsewhere.
EXPERIMENTAL VERIFICATION WITH LABORATORY
SANDBOX MODEL
In the sandboxexperiment describedabove (Figures 2 and
3) the sand was dry (k = 0); therefore the equation for the
critical taper (22) reduces to
a + /3 = (30)
I+K
Loosely packed dry sand is a Coulomb material exhibiting
negligiblecohesionand an angleof internalfriction near & =
30, correspondingto/z = 0.58 [Lambe and Whitman, 1979].
We measuredthe coefficient of friction of sand on Mylar to
be /Zb = 0.30 by experimentally observing the inclination
required to initiate gravitational sliding of a confinedcylin-
der of sandresting on a Mylar surface. From (28) we find the
value of K correspondingto/z = 0.58 and/zo = 0.30 to be K
1162 DAVIS ET AL.' MECHANICSOF FOLD-AND-THRUSTBELTS

o
0
o

O- x
 1163

DAvis
ET
AL.:
MECttAIqICS
OF
FOLD.AIqD-TttguST
BELTS
oncovered
byArray
Reg

sea

4
3M-4
2M-3
. IM- 2
O<M_I

10 Feb.--- ;o %. %,..
31Jury
1973 . ...
. * arthquakeS
near
cross
Bkm : showing
hypocenters
ofmc_re
_: onod
indication
- of ac irehem
nfold.
western
and-thrust
belt,
Tawan._
etal.,
Sune,
19791.
Seismicity
1980b;
Wu
atthe
toe
tsa=v
11. Toe }^ Taiwan
taIt,,o
ofical
taper.
B.g_
'_
;.,
secu?,
foothillsof ,,ut52
;', _. ,ff heweug
;,
-o
at
ornear
tt.... Kong.
Asthe
plate
se-
that
the
mtenu--'
t the
future
between
Luzon
in and
turn
onI-long
the
thick
sedimentary
, and
shelf,
the
.92.
The
relation
between
aand 13
predicted
by
the
theory
in impinges
quences
of
the
Chinese
continental
rise,
slope
boundary

accretionary
height
to wedge
expands
become
the continuously
Central in
Mountains
of width
Taiwanand
rising
isthen a= 5.9
- 0.663 nearly
4km
rises above
above
sea sea
level.
level,
it
is
noOnce
the
accrefionary
longer
able
to
grow wedge
unobstruc
This
prediction
mined
is
result
confirmed
that
(Figure
$)
by the
experimentally
deter-
really
because
ofthe
high
(32)kin/re.Y-
ILl,
19761.erosion
rate,
Infact,
the which
ralay
top.oR reaches
of
the
?oav In5
stateto
6
mountain
which
a=(5.70
+--
0'2)
--(0.66
+--
0.14)13 belt
incentral
Taiwan
hasachlevea
aerosion
-"[Suppe
using
Expertme_nts
graphite
...... H11v
and
sandpaper
on
dlarger
smaller
the
tapers tectonic
base
[Davis,
an,_ :.,accord
compression
Taiwan
in
withthe
is
balanced
this by
region
is
an
active '19811.
foreland
fold.
and-thrus
intothetoeas
oroduced
suu"'*o,(
which
isatso belt
with
present-daY
deformation
evidenced
by
seismicextending
activity
(see 11).obser-
FigureThese
978;Goldburg,
theory.
:,o,.,
- '28)contains
length vations
scale; show
that
therefore
criticalthe
taper.
anY_-esisinherently
The interior
of
existence
of
a the
mountain
large
amount belt
of is at
structura
in'and
fluid
pressure
information
collected
duringthe
course
of
Neither
(18)nor . r-^,d,,mb
weu
thetheoryfornoncohesive
"'" theory
scale
independent'
describingtheGiventhis,
behavior
of the
our success
ofthe
controlled
laboratory petroleum
model exploration
active and
deformation
and prodUCtion
steady
statein addition
topography to
make the
west-
andthelow
discussed cohesive
above,
we strengths
are ofsedimentary
encouraged
to rocks,
consider as ern
TaiWan
applications
oftheaparticularly
critical
taper
of appropriate
place
asubaerialto
wedge. test
our
theory
tomore complicated
geological
cases. the

ApPLIcATION
OFTHE
THEORYTOTAIWANTopographic profiles
of Surface
centralSlope
Taiwanshowthat
Taiwan
Tectonic
is
an
Setting
actively
deforming
atmountain
the belt
western mountains
for
situated
edgeon
ofthe have
smallvery
nearly
a
constant
topographic
features
such
as surface
ridgesslope,
and except
valleys
on
a
thecontinental
....
PhillPPme
shelf
ofChina
te(Figure
.Seapla
9).Thesian
plate
.A___ is scale
of
subducting
,uoManila
TrenCh
.- ,ea dateaong
5km
western
.... dopeorless.
Best

ofTaiwanfit
linear
regression
2.5
and
between a
vary
e plate
= o.Oo
+_.
3.40
0.3
o. tothe
(Figure
This
is
,.ailnm with
themean
s.
lope_ eoe} largearea,
and
western
Taiwan
beneath
me foothills
r,,,,
boundary
between
in
the
segment
of
northern
Luzon
and
.th10).12)
Taiwan
(Fgure __,.,d,,
constant
reinarea--
suggesting
that
the
surta
. has_.
region
entire
and
homogeneous material
. ^,, as
decollemem
wellas
In
contrast,
north the
alongsubruction
the is
gyukyUof
opposite
trench
and
topolarity
the both
to
the
south
along
theproperties,
aclosefluid
pressure,
approach
tothe
critical
taper.
Mindanao
The trench
[SenO,
1977;
convergence Tsai
between
the et
al.,
Asian1977;
and Wu,
the 19791.
philippine
Sea
plates
near
cm/yr
inaTaiwan
is
estimated
by
Seno
119771
to
be
about
northwest-southeast
direction. 7 Taiwan
Decollement
the Theis
Dip
regional
of dipof
course the
not
asbasaldecollemen
easily
.'Imeasured
wellas in
thewestern
region
determined
inthe
It
isnoteworthy
that
theplate
continent-ocean
boundary
in boundary
the cuts
across
Taiwan-Luzn
region;
thesurface
slope,
butithas
been
by fairYrenecu,,-
seismic _,:^,,
,,offling,
r deep
.
LuzonArc
whereasand
theManila
Trench
stable are
oriented
continental
margin
of north-south,
China
frontal
southwest
of
part
cross
of
andthe
drilling,
the
sections
wedge
construction
[Suppe, of
1980a,
1981'
Suppe
andgeologic
retrodeforma
Namson,
b,ofvell-constrain
dip
in
Taiwan
is
suits
inoriented
an northeast-southwest'
oblique
collision This
between
the geometry
Luzon
Arc re-
and
the1979;
Namson,
Figure
10 19821.
extendsThe
from area
the
deformationfront
located
near
Chinese
begancontinental
about margin
4m.Y.
agoin(,see
Figure
9).
The
collision
northernmost
Taiwan
[Chi
et
al.,
19811,
is
occurring
nowjust
south
of
Taiwan,
and
will
occur
1164 DAVIS ET AL.' MECHANICS OF FOLD-AND-THRUST BELTS

o ...

2.9

3.3

3km i 20km I

Fig. 12. Topographicprofilesof westernslopeof centralTaiwan in the regionof steadystate topography(see

Suppe[ 1981]for locations).In all casesthe topography
canbe well fit by a constant-slope
regression.
The meanslopeof
the set of profiles is a = 2.9 _+0.3.

Tehchanshan Meiln 0- Chuhuangkeng

oo
calculated from
o
2 km sonic
log
%

3 km 1 km
measurements direct
measurements
o oo

,% 3 km

_ soniclog

e
4 krn- 2 kin-

4 krT1-

o
o
o
'

0 1 2o 0 10 2O 0 10 2O

kPa/m fluid pressure/depth kPa/m kPa/m

Fig. 13. Observedfluidpressure/depth as a functionof depthin threeareasof westernTaiwan[Suppeand Wittke,

1977;Suppeet al., 1981].The Tiehchanshan andMeilin areasare at the edgeof the overthrustbelt andhaveundergone
little erosion.They display a permeablehydrostaticzone overlyingthe less permeableoverpressuredrocks. The
Chuhuangkeng area is deeperin the fold-and-thrustbelt, it has undergonemore erosion,and overpressured fluid
pressuregradientsreachessentiallyto the surface.It is consideredtypicalof the mountainous
areasof westernTaiwan.
The overpressured valuescorrespondto X = 0.7 in thesethree areasand elsewherein westernTaiwan.
 DAVIS ET AL.: MECHANICS OF FOLD-AND-THRUST BELTS 1165

6 o 6 o

basal
dip
fl=5 basal
dip
fl=6' ' Pb
'=085 6 bas
a/dipS=7
o' ' Pb=085
'

4
I OS

115

i i i i i
0 o
O4 06 08 10 04 04 06 08 10

fluidpressureratio fluidpressureratio fluidpressure

ratto

Fig. 14. Surfaceslopea versusfluidpressure

ratioh for threebasaldipsfi = 6- 1,withByerlce'slaw/at,= 0.85
assumedon the basal decollement. Shaded box shows measured a and h = ht, in western Taiwan. The best fitting
effective coefficient of internal friction is/x = 1.03.

Chiai about30 km into the wedge;the measureddip is/3 = 6 wedge.Neither of theseis knownfrom directmeasurements
at the stratigraphiclevel of the basaldecollementnear the in Taiwan. Laboratorystudieshave shownslidingfriction to
baseof the Neogenecontinentalmarginsection.At leastin be remarkably uniform for a wide variety of rock types.
northern Taiwan it is known that this singlelevel of decolle- Accordingto Byerlee [1978], frictional sliding is well de-
ment existsunder the entire width of Taiwan [Suppe, 1980a]. scribedby the relationIrl -- 0.85O'n*
for effectivenormal
We must extrapolateour 6 measurementin the toe region tractionsin the range 5-200 MPa, which encompassesmost
underthe entirewesternslopeof the mountainsto obtainan rocksabove the bdhttle-plastic
transition,especiallyin over-
estimateof the overall taper of the wedge.We believethat pressuredenvironments.In view of this laboratoryevi-
we know the decollementdip within -1 everywhereand dence,we shalladoptByerlee'sempirical'law'/z0 = 0.85 on
within -+1/4 many places near the toe. the basal decollement where it is logical to assume that
essentiallypure frictionalslidingis occurring.We will then,
Fluid Pressures
employthe critical Coulombwedgetheory, equations(22)
Fluid pressuresare well known within the westernfoot- and (28), to infer the effectivecoefficientof friction within
hillsandcoastalplainof Taiwanasa resultof formationtests the wedge, /z. This procedureyields our preferredmodel,
and sonic log measurementsduring petroleumexploration but sincetheremay be objectionsto the useof Byerlee'slaw
[Suppeand Wittke, 1977;Suppeet al., 1981].Typical fluid on the base, we have in addition computed the complete
pressureprofilesat the westernedgeof the fold-and-thrust rangeof values/z0,/z consistentwith the measuredTaiwan
belt where there has been little erosion show an upper field parametersa = 2.9, fi = 6, and h = Xt, 0.7.
permeablezone of hydrostaticfluid pressuregradients(h
0.4) overlyinglesspermeablesedimentsexhibitingoverpres- Fitting the Theory to the Taiwan Data
sured fluid pressuregradients(h 0.7), as seen in the The best-knownparameterof the westernTaiwan wedge
Tiehchanshanand Meilin wells in Figure 13. This overpres- is the surfaceslope a = 2.9. In Figure 14 we show the
suredgradientis very constantthroughoutwesternTaiwan. theoreticalsurfaceslopesof criticalCoulombwedgeshaving
Very soon after the onset of deformation,the permeable decollementdips/3 = 6 -+ 1, plottedversusfluid pressure
Pliocene-Pleistocenesediments are eroded off the growing ratio X for various values of/z, assumingByerlee's law on
structures and overpressuredfluid pressure gradients are the base.The best fitting value of the effectivecoefficientof
exposedessentiallyat the surfaceof the wedge, as for internal friction is /z = 1.03, about 20% greater than the
examplein the Chuhuangkeng field in Figure 13, whichhas assumedlaboratoryfrictional value on the base. Given the
Miocene rocks at the surface. Furthermore, major decolle- uncertainties in a,/3, andX, andassuming that Byerlee'slaw
ment surfacesthat are intersectedby deep drilling in the is valid on the base,the uncertaintyin/z is about 10%, sothe
foothill zone show the same overpressuredfluid pressure inferred differencebetween /z and /z0 is real. An effective
gradientas the overlyingrocks [Suppeet al., 1981].On the internal friction identical to the basal friction /z0 = 0.85 is
basis of all information we have available, we adopt a definitelyprecluded,since it would give rise to a surface
constantfluidpressureratio = 0 = 0.675 - 0.05 for both slopea = 5.9%assuming/3= 6andX = h0 = 0.675(Figure
the wedge and decollement.It shouldbe noted, however, 10). The implicationof this differencebetween/z and/z0 is
that direct fluid pressuremeasurementsare essentiallyre- that, on the average, slightly larger shear tractions are
stdhctedto the western third of the wedge. requiredto produceinternaldeformationwithin the wedge
than to permit frictional slidingalongits base, which is a
Basal Friction
throughgoing decollement.We interpretthisto meanthatthe
We have outlinedthe data on surfaceslopea, decollement wedgematedhal is not so extensivelyfracturedthat preexist-
dip /3, and fluid pressureratio = 0 in the preceding ingslipplanesof all possibleorientations areavailablewithin
sections. The remaining parameters that enter into the it. Much of the internal deformationmust be taken up by slip
critical taper of the mountainwedge are the coefficientof on suboptimallyorientedsurfaces;it is alsolikely that the
friction on the basal decollement/z0and the corresponding commonlyobservedcomplex geologicalstructureswithin
effectivecoefficientof internal friction/z - tan 4within the the wedge may require fracturingfor the deformationto
1166 DAVIS ET AL.' MECHANICS OF FOLD-AND-THRUST
10 20 30
2.0
1.6
1.2 Byerlee's law
Hb= O.$5
.,
.,
0.8
.,
0.4
= b =0.2
= Ub=0.1
0 0.4
Fig. 15. Values of/zb and/z consistentwith the geometryof the westernTaiwan fold-and-thrustbelt, having a =
2.9and/3 = 6.For the measuredfluidpressureratio h 0.7, valuesbetween/z=/zb = 0.2 and/z =/z = 1.75cansatis-
fy the data. With Byedee's law/z = 0.85 on the base, the effective coefficientof internal friction is /z = 1.03. No
solutionshaving/z </z arc possibleas long as
proceed.The ubiquitouspresencethroughoutwesternTai-
wan of fault-bendfoldsproducedby the steppingup of faults
from the basal decollement[Suppe, 1980b;$uppe and Nam-
son, 1979; Namson, 1982] is consistentwith this interpreta-
tion.
In conclusion, then, we find that the brittle rock strengths
commonly measuredin laboratory experimentsare consis-
tent with the observed regional topographic slope of the
Taiwan fold-and-thrust belt, given the observed fluid pres-
sures and decollementdip. No extraordinary rock proper-
ties, suchas extremely elevatedfluid pressuresor extremely
weak basal layers, need to be called on to explain the
processof mountainbuildingin Taiwan or, by implication,
elsewhere.
To assessthe sensitivity of this conclusionto our use of
Byerlee's law on the base, we show in Figure 15 a plot of
versusIx0consistentwith a = 2.9and 3 = 6for three fluid
pressureratios)t = 0.4, 0.7, and 0.85. Values as low as V =
tXb= 0.2 or as high as Ix = tXb= 1.75 can satisfythe Taiwan
data. In general, for any basal friction permitted by uncer-
tainties in the laboratory results, say 0.6 -< V0 -< 0.9, the
correspondingvalue of the effectiveinternalfriction V must
be about 20% larger. If Ix0 = 0.85, the sheartraction on the
basal decollementat a depth H = 10 km is '0 = 65 MPa,
assuming
p = 2.4g/cm3. Thiscanbe reducedto aslowas
= 15MPa if Ix0= 0.2, which cannotbe ruled out by our data.
It should also be remarked that the inferred
effective internal friction over basal friction dependsrather
critically on the assumptionthat )t = )t0. In fact, a solution
having /x = /x0 = 0.85 with )t0 only 10% greater than )t =
0.675 is also consistent with the observed wedge geometry
BELTS
arctan
ldb
40 50 55 60
/J = /Jb=1. 60
=Pb =1.5
55
=0.4
50
40
30
= Pb=0.55
20
10
0.8 1.2 1.6 2.0
basalfrictionPb
= 2.9 and /3 = 6. The essentialrequirement is that the
interior of the wedge be slightly strongerthan the base, for
whatever reason. Since there is no evidence in any of the
well data for even marginally greater overpressureson or
below the decollement, we are led to conclude that/x >/x0.
APPLICATIONS TO OTHER AREAS
In other areas of active plate compressionwe have fewer
constraintson decollementdip and fluid pressuregradient
than in Taiwan. However, some information on decollement
dip is now available through seismicreflectionprofiling or
refraction studies in a number of active submarine accretion-
ary wedges (for example, Beck and Lehner [1974], Seely et
al. [1974], and others). Therefore by assuming the same
values of rock strengthand friction as we found in Taiwan,
namely, ix0 = 0.85 and tx = 1.03, we may compute a
predicted fluid pressure coefficient )t = )to given the ob-
served taper of the wedge (Figure 16 and Table 1). By
assumingthat )t = )toinsteadof allowingfor the possibilityin
some wedges that )t < )to, we may upwardly bias our
predictedvalue of the fluid pressureratio within the wedge
and downwardly bias our prediction on the basal decolle-
ment; usually only the former, if either, is constrained by
drilling. Note in addition that our indirect fluid pressure
determinationis least susceptibleto errors in a or/3 for large
excess of valuesof)t = )to;this is apparentin Figure 17, which showsa
versus/3 for various fluid pressureratios.
The well-known [Beck and Lehner, 1974; Hamilton, 1979]
seismicprofile acrossthe Java Trench southof Bali predicts
)t = 0.7. A profile of the Makran wedgein the Gulf of Oman
 DAVIS ET AL.: MECHANICS OF FOLD-AND-THRUST BELTS 1167

Himalayas o 40km
I I i I

sea level

Japan Trench sealevel Java Trench sea level

Sunda Trench sealevel

Peru Trench sealevel M akra n sealevel

k=0.98

Aleutians sealevel

Barbados sealevel Oregon sealevel

'....................
...............
k=092 k=0.97
Fig. 16. Crosssections
of theHimalayan fold-and-thrustbeltandvariousactivesubmarine accretionary
wedges.
Heavydashed linesarebestfittinglinearprofiles
usedto inferthefluidpressure
ratiosX = ;to,whicharealsoshown.

[White and Ross, 1979]just offshorefrom mud volcanoes
and near a well showingextreme overpressures(J. Harms,
private communication,1982)predictsX = 0.98. The Inter-
nationalProgramof Ocean Drilling (IPOD) transectof the
JapanTrench [Nasu et al., 1979]predictsh = 0.5. Gravity
modelingof the Sunda Arc [Kieckheferet al., 1981] and
seismicreflectionmodelingof the Peru Trench [Keller et al.,
1979] yield tapers corresponding,respectively,to X = 0.7
and X = 0.6. The eastern Aleutians section of yon Huene et
al. [1979]in the Gulf of Alaska predictsX = 0.88; a similar
maximumfluid pressureratio of X = 0.87 was reportedby
Hottman et al. [1979] for nearby wells in shallow water to
the north. The Oregon section of $navely et al. [1980]
predictsX = 0.9, in goodagreementwith fragmentarywell
data on the shelf indicatingh = 0.85 [Moore and yon Heune,
1980]. Using values of decollementdip from $eeber et al.
[1981]and topographicslopesfrom Ohta and Akiba [1973],
we calculate X = 0.76 for the Himalaya. The Middle America
Trench off Guatemala and southern Mexico has a taper
[$eely et al., 1974;Moore et al., 1979]that predictsh = 0.7,
which is consistentwith the strongindirect evidenceof high
overpressures observedqualitativelyat severalsiteswithin
the wedgeduringdrillingoperationsconductedon Deep Sea
Drilling Project (DSDP) leg 84 [Aubouinet al., 1982].The
wedgeof the LesserAntillesextendingeastfrom Barbados
has an overall taper [Westbrook,1975]correspondingto a
predictedfluid pressureratio h = 0.92. Measuredratiosh
reach 0.8 in deeper wells in Barbados,in adequateagree-
ment with this prediction. Near the toe the taper narrows
perceptibly[Moore et al., 1982],indicatingan even greater
fluidpressureratio, namely,h > 0.97. This was confirmed
by direct measurements indicatingh 1 at the bottomof
DSDPhole542(leg78A),whichislocated
only1kminward
from the deformation front. This is the only near-toe direct
fluid pressuremeasurementavailablein any accretionary
wedge.
In summary,the predictedfluidpressurecoefficientsX for
variousaccretionarywedgesof knowntaper are all in excess
of hydrostatic(X > 0.4), which is reasonablein light of
pressuresnormally encounteredin deep drilling [Fertl,
1168 DAVIS ET AL..' MECHANICS OF FOLD-AND-THRUST BELTS

TABLE 1. Geometries of Active Fold-and-Thrust Belts and Accretionary Wedges and the CorrespondingInferred Fluid Pressure
Ratios X = Xt,, Compared With Available Direct Fluid PressureMeasurements--Qualitative Bounds on the Inferred Values of X = Xt,
Due to Uncertainties in a and/3 Can Be Gaged From Figure 17
Predicted
a, deg /3, deg a + /3 Source X = Xt, Observed X X Source
Subaerial
Taiwan 2.9 -+ 0.3 6.0 -+ 0.5 This paper, Suppe ... 0.675 -+ 0.05 extensive well measurements;
[1980b, 1981] Suppe and Wittke [1977],
unpublished data
Himalaya 4.0 -+ 0.5 3.0 -+ 1.0 Ohta and Akiba 0.76 ......
[1973], Seeher et
al. [1981]
Submarine
Japan 4.5 -+ 0.5 5.4 -+ 0.5 Nasu et al. [1979] 0.5 ......
Peru 3.8 -+ 0.3 5.9 -+ 1.0 Keller et al. [1979] 0.6 ......
Java 3.1 -+ 0.3 6.6 -+ 1.0 Hamilton [1979] 0.7 ......
Sunda 4.0 -+ 1.0 5.0 -+ 1.0 Kieckhefer et al. 0.7 ......
[1981]
Guatemala 5.7 -+ 0.7 2.5 -+ 1.0 Seeley et al. [1974] 0.7 high indirect evidence;DSDP Leg
84 drilling; Aubouin et al.
[1982]
Aleutians 3.0 -+ 0.5 4.5 -+ 1.0 von Huene et al. 0.88 -0.87 mud weight in a well in pond-
[1979] ed sediments; Hottman et
al. [ 1979]
Oregon 2.1 -+ 0.5 6.0 -+ 1.0 $navely et al. [1980] 0.90 0.85 + 0.03 mud weight in a well on the
shelf; Moore and von
Huene [ 1980]
Barbados 1.0 -+ 0.5 8.0 -+ 0.8 Westbrook[1975] 0.92 0.8 -+ 0.05 mud weight in a well on Bar-
Overall taper bados; Moore and von
Huene [ 1980]
Near toe 0.7 -+ 0.2 4.5 + 0.5 Moore et al. [1982] 0.97 -1 inadvertent packer experi-
ment in hole 542; Moore et
al. [ 1982]
Makran 1.6 -+ 0.3 2.0 -+ 1.0 White and Ross 0.98 '-1 mud weight in a well offshore
[1979] Pakistan; Harms (private
communication, 1982)

1976]. In addition, the values found in eastern Aleutians,
Oregon, Guatemala, Makran, and Barbados are in good
agreementwith fragmentary nearby well data. Furthermore,
no tapers are observed that lie outside the range predicted by
the theory (Figure 17). These observationsfurther support
our Coulomb wedge analysis as a valid first-order theory.
EFFECTS OF ISOSTASY AND EROSION ON TAPER
Isostatic adjustmentin responseto the overburdenthat is
added by growth of the wedge can result in some degree of
reinitiation of deformation within the wedge. The greater
stackingof thrust sheetstoward the back of the wedge will

10 10 o

8 SUBAERIAL
WEDGES SUBMARINE
WEDGES

6 2 O
. Sunda
L
4o

. I-9....
o
1 Peru
AeutanI

I
.. ,
aran

I
I.X
x'
Oregon
'

ados to

0o 0 i
0o 2o 4o 6o 8o 10 12 0o 2o 4o 6o 8o 10 12

decollement
dipt decollement
dipt
Fig. 17. Theoreticallinear relationshipsa + R/3 = F for variousfluid pressureratios X = Xt,,assuming/at,= 0.85
and/a = 1.03. Boxesindicateobservedgeometriesof active wedges,usedto infer the fluid pressureratios within them.
Heavy outlinesindicatethosewedgesfor which somedirect fluid pressureinformationis available.A rock densityp =
2.4 g/cm3wasusedin the submarine
case;othervalueswouldyieldvery similarresultsasthe sensitivityto p is slight.
 DAVIS ET AL.; MECHANICS OF FOLD-AND-THRUST BELTS 1169

supercritical
8o
t
6
ii!::,I
fxed
wedgeshape

...................................
...............................
/-'":':::[[ :::.. fz
2

0 o -
0 2 4 6 8 10 '2o 0o 2o ao 6o 8o 0o 2o 4o
basaldip basal dip #
A
Fig. 18. Diagram illustratingthe effectson wedgegeometryof (a) surficialerosionor isostaticadjustmentand (b) a
changein basal friction, either in X, or /x,. Both erosion and isostatic adjustmentdrive a critical wedge into the
subcriticalregime, encouragingrenewed deformation.An increasein basalfriction which doesnot violate the condition
(1 - h,)tx,-< (1 - h)tx also encouragesdeformation, to producea steepercritically tapered wedge. A decreasein basal
friction, on the other hand, gives rise to a supercriticalwedge and possiblyto subductionunaccompaniedby accretion
or wedgedeformation. The sensitivityto changesin basalfriction is quite pronounced.The hypotheticalcase depicted
here representsonly a ---2% changein X,of a submarinewedge having X = Xb = 0.8 initially.

tend to result in a greater degree of downward isostatic
adjustmentthere than in the front. This results in an increase
i/3in the dip of the basal decollement. Because total wedge
taper is conservedin the isostaticprocess,i.e., ia + i/3= 0,
the topographic slope a is correspondinglyreduced.
The critical taper equation (19) showsthat the criticality of
a wedge dependson how its taper is partitioned between
topographic slope a and basement dip /3. Because the
coetficient R given by (20) is always less than 1 for reason-
able densitiesp and pw,each degreeof a in the wedgetaper is
less effective in attaining critical taper than is a degreeof/3.
Therefore the exchange of a for /3 during isostatic adjust-
ment causes the wedge to become subcritical and encour-
ages renewed deformation within the wedge until a new
critical taper is attained (Figure 18a).
Erosion also acts to induce deformation within the wedge
by reducingthe topographic slope while leaving the decolie-
merit dip unaltered, thus producing a subcriticaltaper (Fig-
ure 18a). In an active fold-and-thrust belt, the processesof
erosion and internal deformation shouldbe constantly occur-
ring at rates sutficientto counteract each other, as they are in
western Taiwan. Sudden climatic increases in erosive flux or
tectonic decreasesin compressiveflux will causedisequilib-
rium wedge shapes[Suppe, 1981].
Erosion at the upper wedge surface is not likely to be
important in a submergedaccretionary wedge. Accretion at
the front of the wedge should in that case be the major
source of continued wedge deformation. The addition of
extra mass at the front of an already critical wedge results in
a taper-preservingseriesof deformations,which enlargethe
wedge by propagating toward the back. Sedimentation on
top of the wedge may make it supercritical. For these
reasons, as well as those enumerated below, the kinematics
of subaerial and submarine wedges should be substantially
different.
EFFECTS OF FLUCTUATIONS IN BASAL FRICTION
In the precedingdiscussionwe have ignoredpossibletime
variations in the physical properties of either the wedge
and X) or the decollement(/zband Xb).Time variation in basal
shear traction is most likely in long-lived, slowly accreting
submarine wedges where the rate of subduction of the
underlying oceanic plate is fast. It is less likely in steady
state subaerial wedges like Taiwan where, because of the
thick stratigraphicsection and the high erosive fluxes, most
of the accreted material remains in the wedge for at most a
few million years prior to erosion.
Consider a submarine wedge initially at critical taper. If
this wedge experiencesa drop in basal friction, perhaps by
encounteringa stratigraphicchangein decollementmaterial,
the critical taper will decrease and the previously critical
wedge now will be at supercritical taper (Figure 18b). In
contrast,if the wedgeexperiencesa modestincreasein basal
friction, the critical taper will increaseand the now subcriti-
cal wedge will deform until the new critical taper is attained
(Figure 18b). If the new basal material has properties such
that (1 - Xo)/zo> (1 - X)/z, it cannot form a decollementand
a new decollement must form higher within the existing
wedge. Under these conditions the wedge undergoestecton-
ic erosion along its base.
We might expect, from the above considerations,to find
some submarine wedges that are at supercritical taper,
recordinga past period of higher basal friction. If a wedge is
currently supercritical, the fluid pressure ratio X = X we
infer from its taper will be an underestimate. Oceanic
sedimentscan be subducted beneath a supercritical wedge
without offscrapingand accretion. We might also expect,
from the above considerations, to find some submarine
wedges that are undergoing subduction-erosionor loss of
material from their base because of a recent increase in basal
friction beyond (1 - h/,)/x/, - (1 - h)/x. Nonaccreting and
negatively accreting wedges are being increasingly recog-
nized [for example, von Huene et al., 1980; Aubouin et al.,
1982].
REALM OF THE THEORY
A further method of testing the validity of the Coulomb
wedge theory is to examine natural casesin which the theory
would be expected to break down in some expected way.
Two situations will be considered: (1) the basal decollement
1170 DAVIS ET AL' MECHANICS
Strength MPa
500 1OOO
E
12
16
2O
Brittle Deformation
Fig. 19. Rock strength, as measuredby the stressdifference r,
- r3,versusdepthin the earth,assuming frictionalbehaviorIrl =
0.85r,*for r,* < 200MPa andIrl = 50 MPa + 0.6r,*for r,* > 200
MPa [Byedee, 1978] for the near-surface brittle deformation and
quartzplasticflowgovernedby = 5 x 10-6 (r, - r3) exp(-0.19
MJ mol-'/RT), where is in s- and r - r is in MPa [Braceand internally. If the basal friction increases, so does the taper,
Kohlstedt, 1980] at depth. The plastic strengthhas been calculated
for a range of geothermal gradients and geologically reasonable
strain rates. The theoretical strength is greatest at the point of
brittle-plastic transition, where the brittle and plastic strengthsare
equal.
exhibits pressure-independentplastic rather than Coulomb
behavior becauseof rock type, and (2) the thicknessof the
wedge exceeds the depth to the brittle-plastic transition so
that the base no longer displaysCoulomb friction becauseof
high temperature. In either case we may deducethe expect-
ed qualitative behavior by rewriting the critical taper equa-
tion (18) directly in terms of the basal shear traction (8), i.e.,
a + /5 =
rb/pgH + (1 - p.dp)15
(1 - p.Jp) + (1 - R)K
If b is limited by a yield stress that does not depend on
(33)
pressure, the ratio o/pgH and thus the surface slope a
should decrease toward the back of the wedge where the
thicknessH is greater. If the yield stressis low so that 0 <<
pgH, the coefficient K in the denominator of (33) can be
adequatelyapproximatedby K 2 sin b/(1 - sin b).
OF FOLD-AND-THRUST BELTS
The principal situation in which a relatively thin, low-
temperaturewedge might be expected to exhibit basal plas-
ticity is a decollementcomposedof evaporites, as observed
in some ancient fold-and-thrust belts. A well-known active
fold-and-thrust belt with a low-strength salt decollement
occurs in the Zagros Mountains [Stocklin, 1968]; here even
away from the toe we observe an abnormally small surface
slope and taper relative to the typical Coulomb wedge
prediction, as we expect. Similarly, the Salt Range of
Pakistan has an abnormally low surface slope.
As an accretionary wedge or fold-and-thrust belt grows,
its basal decollement may become sufficiently deep that the
assumptionof brittle behavior is no longer valid because
thermally activiated deformation processes prevail. The
depth of the brittle-plastic transition for quartz [Brace and
Kohlstedt, 1980] and feldspar (W. F. Brace, private commu-
nication, 1982) is expected to occur at a depth of roughly 12
to 16 km in regions with moderate geothermal gradients
(Figure 19). Below such depths, slidingresistancealong the
basal decollementdrops rapidly and is independentof pres-
sure. The Taiwan wedge is too small to reach the depth of
brittle-plastic transition, so Coulomb wedge theory is appli-
cable throughout. However, the Coulomb theory is not
expected to be valid toward the interior of very wide
mountain ranges and accretionary wedges. They frequently
show an abrupt drop in surface slope toward their interior,
which is interpreted as the surfaceexpressionof the brittle-
plastic transition along the base (for example, see Figure 20).
The same phenomenon provides a natural explanation for
the break in topographic slope in the Higher Himalaya near
the edge of the Tibetan Plateau and for the edge of the
Altiplano of the Andes.
CONCLUSIONS
The idea that fold-and-thrust belts and accretionary
wedgeshave the mechanicsof bulldozer wedgesat failure in
horizontal compressionhas been verified quantitatively in
this paper. An approximatetheory has been derived for the
minimum or critical taper required for a Coulomb wedge to
slide stably along its basal decollementrather than deform
whereas an increase in the internal strength of the wedge
decreasesthe critical taper. The observed tapers of active
accretionary wedges are consistentwith low apparent cohe-
sive strengthsand with coefficientsof basal sliding friction
PERU-CHILE MARGIN
iI 14s
lat
en 6
u) Cohesionless
o _ ' Coulomb
Theory
'ifeeee
C
o
.,_,
0
Brittle-Plastic Transition
'''b''';'''d
2 0
length in km
0
.......
80 100
Fig. 20. Surface slope of Peruvian continental margin along
latitude 14S as a function of distance from trench. The decrease in
surfaceslopebeyond55 km is interpretedto reflecta reducedbasal
resistance,where the base of the wedge extends below the brittle-
plastic transition.
 DAVIS ET AL..' MECHANICS
#'b = 0.85 and apparent internal friction /x = 1.03. The
roughly20% differencebetween/Xband/x requiredto fit the
observedtaper of the fold-and-thrustbelt in western Taiwan
is interpretedin terms of the two aspectsof deformation
within the wedge. First, the preexistingfracturesare suffi-
cient only to allow frictional slidingon a limited number of
generally suboptimally oriented surfaces, and second, it is
frequentlynecessaryto fracturelockedgeologicalstructures
for continued deformation.
Two important differences exist between subaerial and
submarinewedges.
1. The horizontal gradient in buoyancy caused by the
slopingrock-waterinterfacegives rise to a term involving Byerlee, J., Friction of rocks, Pure Appl. Geophys., 116, 615-626,
1978.
Pw/Pin the criticaltaperequation(22). This termdisappears
Chappie, W. M., Mechanics of thin-skinned fold-and-thrust belts,
discontinuouslyas the wedge emergesabove sea level and Geol. Soc. Am. Bull., 89, 1189-1198, 1978.
shouldproduceabouta 1dropin surfaceslope;sucha drop, Chi, W. R., J. Namson, and J. Suppe, Stratigraphicrecord of plate
although somewhat larger, is observed in eastern Taiwan
(see Figure 10).
2. Erosionrequirescontinueddeformationthroughouta
subaerialwedgeto maintainthe critical surfaceslopeover
the length of the wedge. In contrast, submarinewedges
deform mostly in the toe where new material is encountered
and they may be supercriticalaway from the toe becauseof
sedimentation and fluctuations in basal friction. Observed
instances of subduction without accretion or subduction-
erosionin some submarinewedgesmay also be effects of
fluctuation in basal friction.
The criticaltaper of fold-and-thrustbeltsandaccretionary
wedges is sensitive to the pore fluid pressure ratio h,
particularlyso for valuesof h near 1. The Coulombwedge
theory is tested by computing h for a number of active
accretionarywedgesand fold-and-thrust belts basedon their
observedtapers.Fluid overpressures
(h > 0.4) are predicted
in every case, and the computedvalues of h are in good
agreementwith fragmentarynearby well data where they
exist.
The Coulombtheory is expectedto break down where the
wedgethickensto the depthof the brittle-plastictransition,
below which the basal resistanceto slidingshoulddrop
abruptlywith increasingtemperature.Both the Andean and
Himalayan wedgesdisplay this predicted drop in taper
where they reach a thicknessof about 15 km, which agrees
well with the depthof the brittle-plastictransitionin quartz
for typical geothermalgradients.
We concludethat fold-and-thrustbelts and accretionary
1981.
wedgeshave the overall mechanicsof a bulldozerwedgeat
Lambe, T. W., and R. V. Whitman, Soil Mechanics, SI version, p.
failurein horizontalcompression andthat normallaboratory 69, John Wiley, New York, 1979.
fractureandfrictionalstrengthsare appropriateto mountain- Li, Y. H., Denudation of Taiwan island since the Pliocene epoch,
building processesin the upper crust, above the brittle- Geology, 4, 105-107, 1976.
plastic transition.
Acknowledgments. We are pleased to acknowledge Florian
Lehner and Leonie Davis for their help in early stagesof this work in
1977 and 1978. Brian Evans, Jason Morgan, Sean Solomon, Tom
Crough, David Elliott, and an anonymous reviewer offered many
useful suggestions for which we are grateful. Dave Smith measured
the surface slopesoffshoreof Peru (Figure 20). J. Suppewould like
to thank the Chinese Petroleum Corporation for accessto data and
hospitality in Taiwan and the John Simon GuggenheimMemorial
Foundationfor graciousand unencumberedsupportin 1978-1979.
This work was supported by the Division of Earth Sciences,
National Science Foundation, grants EAR80-18350 and EAR81-
21197 at Princeton and grant EAR78-12936 at MIT.
This paper is dedicatedto W. M. Chappie, who died in February
1981. Our work was stimulatedby a lecture he gave at Princetonin
1977.
OF FOLD-AND-THRUST BELTS 1171
REFERENCES
Aubouin, J., R. von Huene, M. Baltuck, R. Arnott, J. Bourgois, M.
Filewicz, K. Kvenvolden, B. Leinert, J. McDonald, K. McDou-
gall, Y. Ogawa, E. Taylor, and B. Winsborough, Leg 84 of the
Deep Sea Drilling Project, subductionwithout accretion: Middle
America Trench off Guatemala, Nature, 297, 458-460, 1982.
Bally, A. W., P. L. Gordy, and G. A. Stewart, Structure, seismic
data and orogenic evolution of southern Canadian Rocky Moun-
tains, Bull. Can. Pet. Geol., 14, 337-381, 1966.
Beck, R. H., and P. Lehner, Oceans, new frontier in exploration,
Am. Assoc. Pet. Geol. Bull., 58, 376-395, 1974.
Brace, W. F., and D. L. Kohlstedt, Limits on lithospheric stress
imposedby laboratory experiments, J. Geophys.Res., 85, 6248-
6252, 1980.
interactionsin the Coastal Range of eastern Taiwan, Geol. Soc.
China Mere., 4, 491-530, 1981.
Davis, D. M., The mechanics of thrust faults: A sand box model,
B.S. thesis, 53 pp., Princeton Univ., Princeton, N.J., 1978.
Davis, D. M., and J. Suppe, Critical taper in mechanicsof fold-and-
thrust belts, Geol. Soc. Am. Abstr. Programs, 12,410, 1980.
Elliott, D., The motion of thrust sheets,J. Geophys.Res., 81,949-
963, 1976.
Goldburg,B. L., Formation of critical taper wedgesby compression
in a sand box model, B.S. thesis, 70 pp., Princeton Univ.,
Princeton, N.J., 1982.
Hamilton, W., Tectonicsof the Indonesianregion, U.S. Geol. Surv.
Prof. Pap., 1078, 345 pp., 1979.
Hoshino, K., H. Koide, K. Inami, S. Iwamura, and S. Mitsui,
Mechanical properties of Japanese Tertiary sedimentary rocks
under high confiningpressures,Rep. 244, 200 pp., Geol. Surv. of
Japan, Kawasaki, 1972.
Hottman, C. E., J. H. Smith, and W. R. Purcell, Relationshipamong
earth stresses,pore pressure,and drilling problemsoffshoreGulf
of Alaska, J. Pet. Technol., 31, 1477-1484, 1979.
Hubbert, M. K., and W. W. Rubey, Role of fluid pressure in
mechanics of overthrust faulting, I, Mechanics of fluid-filled
solids and its application to overthrust faulting, Geol. Soc. Am.
Bull., 70, 115-166, 1959.
Jaeger, J. C., and N. G. W. Cook, Fundamentals of Rock Mechan-
ics, pp. 87-91, Methuen, London, 1969.
Keller, B., B. T. R. Lewis, C. Meeder, C. Helsley, and R. P. Meyer,
Explosion seismologystudies of active and passive continental
margins, in Geological and GeophysicalInvestigations of Conti-
nental Margins, Mem. 29, edited by J. S. Watkins, L. Montadert,
and P. Dickerson, pp. 443-451, American Association of Petro-
leum Geologists,Tulsa, Okla., 1979.
Kieckhefer, R. M., G. F. Moore, F. J. Emmel, and W. Sugiarta,
Crustal structure of the Sunda forearc region west of central
Sumatra from gravity data, J. Geophys. Res., 86, 7003-7012,
McClay, K. R., and N.J. Price (Eds.), Thrust and Nappe Tectonics,
Spec. Publ. 9, 539 pp., Geological Society of London, Oxford,
1981.
Moore, J. C., and R. von Huene, Abnormal pore pressureand hole
instabilityin forearc regions:A preliminaryreport, report, 29 pp.,
Ocean Margin Drill. Proj., Menlo Park, Calif., 1980.
Moore, J. C., J. S. Watkins, T. H. Shipley, S. B. Bachman, F. W.
Beghtel, A. Butt, B. M. Didyk, J. K. Leggett, N. Lundberg, K. J.
McMillen, N. Niitsuma, L. E. Shephard, J.-F. Stephan, and H.
Stradner, Progressiveaccretion in the Middle America Trench,
southern Mexico, Nature, 281,638-642, 1979.
Moore, J. C., et al., Offscrapingand underthrustingof sedimentat
the deformationfront of the BarbadosRidge: Deep Sea Drilling
Project leg 78A, Geol. Soc. Am. Bull., 93, 1065-1077, 1982.
Namson, J. S., Studiesof the structure,stratigraphicrecord of plate
interactionand role of pore-fluidpressurein the active fold and
thrust belt of Taiwan and a study of manganesedepositsfrom
1172 DAVIS ET AL.: MECHANICS OF FOLD-AND-THRUST BELTS
northern California, Ph.D. thesis, 302 pp., Princeton Univ.,
Princeton, N.J., 1982.
Nasu, N., Y. Tomoda, K. Kobayashi, H. Kagami, S. Uyeda, S.
Nagumo, I. Kushiro, M. Ozima, K. Nakazawa, Y. Takayanagi,
H. Okada, S. Murauchi, Y. Ishiwada, and Y. Ishii, Multi-channel
seismic reflection data across the Japan Trench, IPOD-Japan
Basic Data Ser., No. 3, 22 pp., Ocean Res. Inst., Univ. of Tokyo,
Tokyo, 1979.
Ohta, Y., and C. Akiba (Eds.), Geology of the Nepal Himalayas,
286 pp., Himalayan Committeeof Hokkaido University, Saporro,
Japan, 1973.
Paterson,M. S., ExperimentalRock Deformation: TheBrittle Field,
pp. 16-50, Springer-Verlag, New York, 1978.
Roeder, D., O. E. Gilbert, Jr., and W. D. Witherspoon, Evolution
and macroscopic structure of Valley and Ridge thrust belt,
Tennesseeand Virginia, Stud. in Geol. 2, 25 pp., Dep. of Geol.
Sci., Univ. of Tenn., Chattanooga, 1978.
Seeber, L., J. G. Armbruster, and R. C. Quittmeyer, Seismicityand
continental subduction in the Himalayan arc, in Zagros-Hindu
Kush-Himalaya Geodynamic Evolution, GeodynamicsSer., vol.
3, edited by H. K. Gupta, and F. M. Delany, pp. 215-242, AGU,
Washington, D.C., 1981.
Seely, D. R., P. R. Vail, and G. G. Walton, Trench slopemodel, in
The Geology of Continental Margins, edited by C. A. Burk, and
C. L. Drake, pp. 249-260, Springer-Verlag, New York, 1974.
Seno,., The instantaneous
rotationvectorof the PhilippineSea
plate relative to the Eurasian plate, Tectonophysics,42,209-226,
1977.
Snavely, P. D., Jr., H. C. Wagner, and D. L. Lander, Geologic
crosssectionof the central Oregoncontinentalmargin, Geol. Soc.
Am. Map Chart Ser., MC-28J, 1980.
Stanley, R. S., L. B. Hill, H. C. Chang, and H. N. Hu, A transect
throughthe metamorphiccore of the Central Mountains, southern
Taiwan, Geol. Soc. China Mem., 4, 443-473, 1981.
Stocklin, J., Structural history and tectonicsof Iran: A review, Am.
Assoc. Pet. Geol. Bull., 52, 1229-1258, 1968.
Stockmal, G. S., and W. M. Chappie, Modelling accretionarywedge
deformation using a rigid-perfectly plastic rheology (abstract),
Eos Trans. AGU, 62,397-398, 1981.
Suppe, J., A retrodeformable cross section of northern Taiwan,
Geol. Soc. China Proc., 23, 46-55, 1980a.
Suppe,J., Imbricatedstructureof westernfoothillsbelt, south-
central Taiwan, Pet. Geol. Taiwan, 17, 1-16, 1980b.
Suppe,J., Mechanicsof mountainbuildingand metamorphism
in
Taiwan, Geol. Soc. China Mem., 4, 67-89, 1981.
Suppe,J., andJ. Namson,Fault-bendoriginof frontalfoldsof the
western Taiwan fold-and-thrust belt, Pet. Geol. Taiwan, 16, 1-18,
1979.
Suppe,J., and J. H. Wittke, Abnormalpore-fluidpressuresin
relationto stratigraphyand structurein the activefold-and-thrust
belt of northwesternTaiwan, Pet. Geol. Taiwan, 14, 11-24, 1977.
Suppe,J., J. Namson,andA.M. Pytte,Roleof fluidpressurein an
active overthrust belt: Taiwan, Geol. Soc. Am. Abstr. Programs,
13,562, 1981.
Tsai, Y. B., T. L. Teng, J. M. Chiu, and H. L. Liu, Tectonic
implicationsof the seismicityof the Taiwanregion,Geol. $oc.
China Mem., 2, 13-41, 1977.
von Huene, R., G. W. Moore, and J. C. Moore, Crosssectionof
Alaska Peninsula-Kodiak Island-Aleutian Trench, Geol. Soc.
Am. Map Chart Ser., MC-28A, 1979.
von Huene, R., M. Langseth,N. Nasu, and H. Okada, Summary,
JapanTrenchTransect,Initial Rep.DeepSeaDrill. Proj., Legs56
and 57, 473-488, 1980.
Westbrook,G. K., The structureof the crustand uppermantlein
the regionof Barbadosand the LesserAntilles,Geophys.J. R.
Astron. Soc., 48, 201-242, 1975.
White, R. S., and D. A. Ross, Tectonicsof the westernGulf of
Oman, J. Geophys.Res., 84, 3479-3489, 1979.
Wu, F. T., Recenttectonicsof Taiwan, in Geodyndamics of the
WesternPacific,Adv. in Earth and Planet. $ci. 6, editedby S.
Uyeda,R. W. Murphy,and K. Kobayashi,pp. 265-299,Center
for Academic PublicationsJapan, Tokyo, 1979.
Wu, F. T., Y. H. Yeh, and Y. B. Tsai, Seismicityin the Tsengwen
reservoirarea, Taiwan, Bull. Seismol.Soc. Am., 69, 1783-1796,
1979.
(Received January 25, 1982;
revised August 24, 1982;
accepted November 5, 1982.)