seismic interpretation

1Shaw, Connors, and Suppe

2

StructuralInterpretation

Methods

Part 1: Structural Interpretation Methods Seismic Interpretation of Contractional Fault-Related Folds

3

Basic concepts

Folds are bends or flexures of layered rock that form in response to motion along faults,diapirism, compaction, and regional subsidence or uplift. Folds are expressed in seismicreflection profiles as one or more regions of dipping reflections (dip domains) that corre-spond to inclined stratigraphic contacts.

Folds come in three basic types:

Folds are composed of one or more dip domains, and may have angular or curved foldshapes:

Dip domains are separated by axial surfaces; imaginary planes which, when viewed in two dimensions,form axial traces. Anticlinal axial surfaces occupy concave-downward fold hinges; synclinal axial surfacesoccupy concave-upward fold hinges.

Axial surfaces often occur in pairs that bound fold limbs, which are also called kink bands:

1A-1: Defining folds

monoclines anticlines

anticlines

synclines

synclines

fold limbs crest

single hinge

single hinge

curved hinge

curved hingemultiple hinges

curved hinge

single hinge curved hinge

angular hinge multiple angular hinges

multiple angular hinges

anticlinal axial surfaces

axial trace

synclinal axial surfaces

pairedaxial surfaces

two sets of pairedaxial surfaces

kink bandkink bands

multiple hinges

Shaw, Connors, and Suppe Part 1: Structural Interpretation Methods

4

Monocline,San JoaquinValley,California,U.S.A.

Syncline,Santa BarbaraChannel,California,U.S.A.

Single Hinge Anticline, Niger Delta, Nigeria

Multiple Hinge Anticline, Permian Basin, Texas, U.S.A.

Folds in seismic sections


5

Folds and bedding thicknessFolds are classified based on whether or not the thickness of stratigraphic layers changes in dip domains or across axialsurfaces.

Parallel folds preserve layer thickness, and are common in strata that deformed predominantly by flexural slip (see insetat right). Axial surfaces bisect inter-limb angles in parallel folds.

Parallel fold model Parallel fold, synclinal axial surface

Layer thickness is conserved: Bed thickness T1 equals bed thickness T2.Bisecting axial surfaces: Interlimb angle γ1 equals interlimb angle γ2.

Non-parallel fold model

Various types of folds exhibit non-parallel behavior, where the thick-ness of stratigraphic layers changes gradually in dip domains orabruptly across axial surfaces. These thickness changes may be causedby various deformation mechanisms, including ductile flow withinincompetent beds. Alternatively, thickness changes may be deposi-tional in origin. Axial surfaces do not bisect interlimb angles in non-par-allel folds. Rather, axial surface orientations are governed by the mag-nitude of the change in bed thickness.

Non-Parallel fold, anticlinal axial surface

Parallel folds commonly form by a deformationmechanism called flexural slip, where folding isaccommodated by motions on minor faults thatoccur along some mechanical layering — usuallybedding. Flexural-slip surfaces, which can beobserved in core or outcrop, may vary in spacingfrom a few millimeters to several tens of meters inspacing.

The amount of off-set on flexural-slipfaults increases asthe fold tightens(note slip increasefrom models 1 to2), and is a func-tion of the spacingof slip surfaces.

Slip changesinstantaneouslyacross axial sur-faces in angularfolds (models 1, 2);whereas, slipincreases alongbedding surfacesthrough the hingein curved-hingefolds (model 3).

slip surfaces


6

Balanced model

Overlapping reflections occur in synclines (1) on this stacked section;similar patterns persist in under-migrated sections. The steep limb is notimaged and diffractions are present (2).Synthetic seismic

Stackedsection

Migratedsection

Proper migration removes overlapping reflections and collapses diffractions, but the steep limb remains un-imaged.

Locating axial surfaces in seismic sectionsMigration moves dipping reflections upward and laterally to properly image the fold geometry,but reflections on non-migrated or under-migrated sections do not accurately represent foldshape. However, axial surfaces can be inferred on these sections by mapping the truncations ofhorizontal reflections.

Shortcomings in seismic images of foldsFolds can be distorted or only partially imaged in seismic sections. Two common shortcomings are:

Overlapping reflections in non-migrated or under-migrated sections; and

poor imaging of steeply dipping fold limbs.Model Stacked section (synthetic)

Migrated section (synthetic) Stacked section (synthetic)

Step 1: Pinpoint truncations of horizontalreflections as they enter the poorly imagedzone. Note that diffractions, dipping towardthe fold, may emanate from these trunca-tions.

Step 2: Fit an axial surface that best matchesthe aligned truncations. Note that the inter-preted axial surface matches closely withthe axial surface defined in the migrated sec-tion (left).

reflectiontruncations

diffractions

axial surface


7

Interpreting folds in poorly imaged zonesPoorly imaged zones on folds are commonly caused by, and interpreted as, faults orsteep limbs. Both solutions are often permissible and should be evaluated. Here, wedescribe a method of interpreting parallel folds in poorly imaged zones.

A: Is the poorly imaged zone a fault or steep fold limb?

C: Interpretation using the parallel fold method

B: Method for interpreting parallel folds in poorly imaged zones

Step 1: Pinpoint truncationsof reflections as they enterthe poorly imaged zone.

Step 2: Fit parallel axial sur-faces that best match thealigned truncations. Measurethe average dip outside of thefold limb and measure γ1.

Step 3: Define the dip of bedsin the kink band by making γ2

equal to γ1.

D: Confirmation of fold geometry with dipmeter log and 3-D seismic image

In this example, 3-Dseismic data and a dip-meter log confirm thepresence of steeply dip-ping beds in the poorlyimaged zone. The pri-mary test of the foldinterpretation, however,is whether or not thehorizons correlateproperly across thepoorly imaged zone. Ifthey do, a parallel foldinterpretation is per-missible. If they do not,a non-parallel fold orfault likely occupies thepoorly imaged zone.

fault

fold

2-D, post-stack time migration displayed in depth

Data courtesy of Texaco, Inc.


3-D, post-stack time migration displayed in depth

water-bottommultiples

reflectiontruncations


8

Faults are identified in seismic reflection profiles through:fault cutoffs — terminations of reflections or abrupt changes of reflectionattributes (e.g., amplitude, polarity) at fault surfaces;

terminations of fold limbs or kink bands; and

direct fault-plane reflections, produced by changes in velocity and density acrossor within fault zones.

Cutoffs and fault plane reflections (criteria 1 and 3) directly constrain fault positions. Thrustfaults and their cutoffs, however, are generally difficult to image and identify, and thus the recog-nition of kink-band terminations (criterion 2) is a vital component of interpreting these faults. Inthis section, we describe how these criteria can be used together to identify and interpret thrustand reverse faults in seismic sections.

Fault cutoffs and kink-band terminationsbalanced model

Incipient fault with markers alongfault surface.

in outcrop

Fault cutoffs in outcrop, Mississip-pian Joana limestone, Nevada, U.S.A.

in synthetic seismicSeismic forward model showing faultcutoffs (1) and downward terminatingkink-bands (2).

Recognizing and interpreting faults in seismic sectionfault cutoffsAbrupt terminations (cutoffs)and duplications of prominentreflections constrain the posi-tion of a gently dipping thrustfault. (2-D seismic data, Permianbasin, Texas, U.S.A.)

Downward terminating kink band (2) definesthe position of a gently dipping thrust. (3-Dseismic data, Permian basin, Texas, U.S.A.).

Downward terminating kink band (2) andfault-plane reflection (3) define the positionof a thrust fault that shallows to an upperdetachment. (3-D seismic data/Niger Delta).

Fault with offset markers and cutoffs.Note that hanging wall kink bands termi-nate downward into the fault surface.

1A-2: Recognizing thrust and reverse faults

kink-band terminationsThrust faults and bed-parallel detachments can be identified by the abrupt, downward termina-tions of kink bands. Terminations are generally marked by regions of dipping reflections abovehorizontal or more gently dipping reflections, and may contain fault cutoffs. Dipping reflectionsin kink bands represent strata folded in the hanging wall of a thrust/reverse fault or detachment;whereas, horizontal or more gently dipping reflections represent footwall strata below the faultor detachment. Thus faults and/or detachments should be interpreted at the transition betweenthese two dip domains.


Data courtesy of Texaco, Inc. Data courtesy of Mabone, Ltd.

fault-planereflection

inferreddetachment

inferred fault

inferred faultposition cutoff

cutoff

hanging wallcutoffs

footwall cutoffs

fault


9

Interpreting thrust ramps on seismic sectionsSeismic Example: Peruvian AndesCombinations of the three fault recognition

criteria are employed to interpret thrustfaults on the seismic section presentedhere.

This section images structures that involvetwo large thrust faults, which can be inter-preted using the fault recognition criteria.The top panel is an uninterpreted sectionacross a fold and thrust belt in the Andeanfoothills, Ucayali basin, Peru. Faults in thelower section are interpreted using:Cutoffs (1), kink-band terminations (2), andfault-plane reflections (3). Note how aseries of cutoffs and kink-band termina-tions can corroborate, and be used toextrapolate beyond, the fault-plane reflec-tions. (2-D seismic data, reprinted fromShaw et al., 1999, and published courtesyof Perupetro).

dipping overhorizontal reflections

dipping overhorizontal reflections

VE of 1:1interpreted faults


10

Recognizing detachmentsDetachments are faults that run along bedding or other stratigraphic horizons, and thusgenerally are horizontal or dip at low angles. In fold and thrust belts, detachments arecommonly referred to as decollements. Detachments are generally not imaged directlyon seismic sections, but rather are interpreted at the base and/or top of thrust ramps.Basal detachments can be located in seismic sections by defining the downward termi-nations of kink bands, as described on the preceding pages.

These two seismic sections have prominent detachments. In thesection at right, the detachment is located at the base of a single-thrust thrust ramp. The fold in the hanging wall of the thrust isproduced by slip across the fault bend that is formed at the con-nection of the thrust ramp and detachment. This class of fault-bend fold is described in section 1B-1. In the section below, aregional detachment forms the base of several thrust ramps.

Seismic Example: Sichuan basin, China

Seismic Example: Nankai Trough, Japan

ramp

ramp

detachment

detachment


11

Growth or syntectonic strata are stratigraphic intervals that were deposited during defor-mation. The ages of growth strata therefore define the timing of deformations. In contrac-tional fault-related folds, growth strata typically thin across fold limbs toward structuralhighs. The geometries of growth structures are controlled primarily by the folding mecha-nism and the relative rates of sedimentation and uplift. Thus, growth fold patterns imagedin seismic data are often considered diagnostic of folding mechanism and sediment-to-uplift ratio. In this section, we describe common patterns of growth strata in fault-relatedfolds that are imaged in seismic reflection data.

Growth strata in seismic section:Sedimentation exceeds uplift

In cases where the sedimentation rate exceeds the uplift rate, growth strata are typically char-acterized as sequences, bounded by two or more seismic reflections, that thin toward thestructural high. Growth strata are generally folded in one or more limbs of the structure. In thisseismic section, growth strata thin onto the fold crest, with the lowermost growth units exhibit-ing the greatest thickness changes. (2-D seismic data, reprinted from Shaw et al., 1997).

Growth strata in seismic section:Uplift exceeds sedimentation

In cases where the uplift rate exceeds the sedimentation rate, growth strata typicallythin toward, and onlap, the structural high. Growth strata are generally not presentabove the fold crest, but are folded in one or more limbs of the structure. In this seis-mic section, growth strata onlap the backlimb and forelimb of a fault-related fold. Thegrowth strata are overlain by post-tectonic strata, which are described later in this sec-tion. (This structure is interpreted more completely in sections 1B-1 and 1B-4).

1A-3: Recognizing growth strata

growth strata

pre-growth strata

pre-growth strata

onlappinggrowth strata



12

Growth strata as records of fold kinematicsTwo folding mechanisms — kink-band migration and limb rotation — are commonly ascribed to contractional fault-related folds. These folding mechanisms typically yield distinctive patterns ofdeformed growth strata above fold limbs. Thus, seismic images of growth folds can be used to identify the folding mechanisms, which in turn can dictate the kinematic theory (e.g., fault-bend fold-ing or detachment folding) that is most appropriate to guide the structural interpretation of the seismic data.

In fault-related folds that develop purely by kink-band migration, fold limbs widen through timewhile maintaining a fixed dip (Suppe et al., 1992), as illustrated in the sequential model involv-ing pre-growth strata only (above left). Material is incorporated into the fold limb by passingthrough an active axial surface, which at depth is generally pinned to a bend or tip of a fault(Suppe, 1983; Suppe and Medwedeff, 1990). The fold limb in growth strata is bounded by theactive axial surface and the growth axial surface, an inactive axial surface that defines the locusof particles originally deposited along the active axial surface. In these sequential models, thesynclinal axial surface is active, and the anticlinal axial surface is inactive.

In the case where sedimentation rate exceeds uplift rate (above center), strata are foldedthrough the synclinal axis and incorporated into the widening fold limb. The dip of foldedgrowth strata is equal to dip of the fold limb in pre-growth strata. The width of the dip panelfor each growth horizon corresponds to the amount of fold growth that occurred subsequentto the deposition of that marker. As a result, younger horizons have narrower fold limbs thando older horizons, forming narrowing upward fold limbs or kink bands in growth strata (growthtriangles). In the case where uplift rate exceeds sedimentation rate (above right), each incre-ment of folding produces a discrete fold scarp located where the active axial surface projectsto the Earth’s surface (Shaw et al., 2004). Subsequent deposits onlap the fold scarp, producingstratigraphic pinchouts above the fold limb. Fold scarps and stratigraphic pinch-outs are dis-placed laterally and folded as they are incorporated into widening limbs.

Contractional fault-related folding theories that exclusively invoke kink-band migration includefault-bend folding (Suppe, 1983), constant-thickness and fixed axis fault-propagation folding(Suppe and Medwedeff, 1983), and basement-involved (triple junction) folding (Narr and Suppe,1994).

Folding by kink-band migrationpre-growth strata only sedimentation > uplift sedimentation < uplift

Folding by progressive limb rotation

In fault-related folds that develop purely by limb rotation with fixed hinges (i.e., inactive axialsurfaces), the dip of the fold limb increases with each increment of folding as illustrated in thesequential model involving pre-growth strata only (left). In the case where sedimentation rateexceeds uplift rate (center), strata are progressively rotated with each increment of folding.Thus, older growth horizons dip more steeply than do younger horizons, yielding a pro-nounced fanning of limb dips in growth strata. Fold limb width, however, remains constant. Inthe case where uplift rate exceeds sedimentation rate, growth strata also exhibit a fanning oflimb dips. However, growth strata typically onlap the fold limb.

Contractional fault-related folding theories that exclusively invoke limb rotation include cer-tain classes of detachment folds (Dahlstrom, 1990; Hardy and Poblet, 1994).

pre-growth strata only sedimentation > uplift sedimentation < uplift

inactive axial surfaceinactive axial surface

active axial surface

growth axial surface

fold scarps

growth triangle

growth axial surface

pre-growth strataonlaps

growthstrata

growthstrata

fanning of limb dips

pre-growth strata onlaps


13

Growth structures in seismic dataGrowth structures imaged in seismic sections commonly exhibit patterns that are similar tothe kink-band migration or limb-rotation models that were described on the previous page. Inother cases, folds may develop by a combination of kink-band migration and limb rotation,resulting in hybrid patterns of growth structure. This section presents three seismic profilesas examples of kink-band migration, limb rotation, and hybrid growth structures.

(top) The seismic section above shows a narrowing upward fold limb, or growth triangle,where bed dips within the fold limb generally do not shallow upward, consistent with foldingby kink-band migration. Dipmeter data in the wells corroborates the reflector dips. (upperright) In this section, a fanning and upward shallowing of limb dips within growth strata areconsistent with folding by progressive limb rotation. The core of the anticline is filled withsalt, which presumably thickened during deformation, leading to progressive rotation of theoverlying fold limbs. (lower right) The growth structure in this section contains both a growthtriangle and a fanning of limb dips, suggesting folding by a combination of kink-band migra-tion and limb rotation mechanisms. Kinematic theories that employ hybrid folding mecha-nisms include shear fault-bend folds (Suppe et al., 2004; see section 1A-4), certain classes ofdetachment folds (Dahlstrom, 1990; Hardy and Poblet, 1994; see section 1B-3), and trishearfault-propagation folds (Erslev, 1991; Hardy and Ford, 1997; Allmendinger, 1998; see section1B-2).

Folding by progressive limb rotationSeismic Example: offshore Angola

Folding by both kink-band migration and limb rotationSeismic Example: San Joaquin basin, California, U.S.A.

Folding by kink-band migrationSeismic Example: Santa Barbara Channel, California, U.S.A.

growth triangle

growth triangle


salt mound


detachment

growthstrata

kink-band migration model

growthstrata

growthstrata

limb rotation model

hybrid model


14

The top model shows a post-tectonic drape sequence above a rigid basementhigh. The drape sequence thins toward the crest of the structure, withyounger strata having less relief than older units. The lower model showsgrowth strata above a fold developed by progressive limb rotation. The twostratigraphic patterns are similar, and in some cases difficult to distinguish.Incorrect interpretations of drape and growth sequences can lead to flawedestimates of structural timing and kinematics. Thus, care should be taken intrying to distinguish between drape and growth sequences.

One common difference between drape and growth sequences is the orien-tation of axial surfaces. Axial surfaces in drape sequences often are verticalor dip away from the structural crest, reflecting a state of tension and due, insome cases, to compaction (Laubach et al., 2000). In contrast, axial surfacesin contractional folds generally dip toward the structural crest, reflecting astate of compression. Thus, careful interpretation of axial surfaces, alongwith consideration of regional tectonic history, can, in some cases, help todistinguish between drape and growth sequences.

Distinguishing drape from growth strataSedimentary drape sequences are stratigraphic intervals that were deposited above a structure after deformation ceased, yet they are warped due to primary sedimentary dip and/or compaction.Drape sequences exhibit a wide range of patterns depending on the sedimentary environment and facies. In some cases, drape sequences have patterns that are similar to those of growth stratadeformed by limb rotation. In this section, we illustrate the potential similarity of drape and growth patterns, and show an example of a drape sequence in a seismic section.

Kinematic modelsDrape sequence

Growth fold

Drape foldingSeismic Example: offshore California Borderlands, U.S.A.

This seismic section images a siliciclastic drape sequence that onlapsand overlies a ridge of metamorphic basement rocks.

drape

basement

drape

axial surfacesdips away from crest

axial surfacesdips toward crest

growthstrata


15

Synclinal fault-bend foldsSynclinal fault-bend folds form at concave-upward fault bends. Synclinal axial surfaces arepinned to the fault bend and are generally active; whereas anticlinal axial surfaces are inactiveand move with the hanging wall block. Figures below show a kinematic model, a field example,and a seismic example of synclinal fault-bend folds.

1B-1: Fault-bend foldsBasic conceptFault-bend folds form as hanging wall-rocks move over bends in an underlying fault. This sec-tion describes the geometry and kinematics of fault-bend folding after Suppe (1983) and intro-duces basic techniques for interpreting these structures in seismic data.

To describe the basic concept of fault-bend folding, we will consider the hypothetical case of afault in cross section with one bend joining upper and lower segments. Rigid-block translationof the hanging wall parallel to the upper fault segment produces a void between the faultblocks; whereas, translation of the hanging wall parallel to the lower fault segment produces an“overlap.” Both of these cases are unreasonable.

Rigid-BlockTranslation

In contrast, folding of the hanging wall block through the development of a kink band accom-modates fault slip without generating an unreasonable overlap or void. This fault-bend folding(Suppe, 1983) is localized along an active axial surface, which is fixed with respect to the faultbend. After strata are folded at the active axial surface, they are translated above the upperfault segment. The inactive axial surface marks the locus of particles that were located alongthe active axial surface at the initiation of fault slip. The inactive axial surface moves away fromthe active axial surface with progressive fault slip, and thus the width of the intervening kinkband is proportional to the amount of fault slip.

Fault-BendFolding

Kinematic Model

Field Example

Seismic Example: Argentina

axial surface

axial surface

fault

fault


16

Anticlinal fault-bend foldsAnticlinal fault-bend folds form at concave-downward fault bends. Anticlinal axial surfacesare pinned to the fault bend and are generally active; whereas, synclinal axial surfaces areinactive and move with the hanging wall block. Figures below show a kinematic model, afield example, and seismic examples of anticlinal fault-bend folds.

Kinematic Model

Field Example

Seismic Example: Niger Delta

Seismic Example: Permian basin, U.S.A.

axial surface

fault

axial surface

fault

axial surface

fault


17

Quantitative fault-bend foldingBased on assumptions of conservation of bed length and thick-ness during folding, the shape of a fault-bend fold is related tothe shape of the fault by:

Where θ is the hanging wall cutoff angle before the fault bend; φis the change in fault dip; β is the hanging wall cutoff after thefault bend, and; γ is one half of the interlimb angle, such that theaxial surfaces bisect the interlimb angles and bed thicknessesare preserved. If two of these values are known, the remainingtwo values can be determined.

The fault-bend fold relations are displayed in the graph below.The left side of the graph describes anticlinal fault-bend folds,where the fold is concave toward the fault; the right side of thegraph describes synclinal fault-bend folds, where the fold is con-vex toward the fault.

When interpreting seismic sections, typically the interlimb angle(γ) can be observed (see section 1A-2) and one of the hangingwall cutoff angles (θ or β) can be specified. Using the graph, thechange in fault dip (φ) and the remaining cutoff angle can bedetermined.

For anticlinal fault-bend folds there are two fold solutions foreach θ and φ value; first mode solutions produce open folds thathave been shown to effectively describe many observed foldgeometries; whereas, second mode solutions are geometricallyvalid but have not been shown to effectively describe naturalfold shapes.

Anticlinal fault-bend folds Synclinal fault-bend folds


18

Fault slip and fault-bend foldsThe magnitude of fault slip can change across fault bends, as slipis consumed or produced by fault-bend folding. In cases wherethe initial cutoff angle is not equal to zero (θ � 0), anticlinal fault-bend folds consume fault slip and synclinal fault-bend folds pro-duce fault slip. The change in fault slip is described by theparameter R, which is the ratio of slip magnitude beyond (S1)and before (S0) the fault bend.

In cases where the initial cut-off angle (θ) equals zero, then Requals one (R=1). When the initial cut-off angle (θ) does not equalzero, R can be determined if any two of the four geometric param-eters (θ, φ, β, γ) are specified using fault-bend fold theory (Suppe,1983). The graph below plots R as a function of initial cut-off angle(θ), interlimb angle (γ), and change in fault dip (φ), and is of thesame format used to describe fault-bend fold geometry.

R varies greatly as a function of the tightness of the fold, whichis reflected in part by the interlimb angle (γ). Tight (perhaps with

steep limbs) anticlinal folds generally consume larger amountsof slip (hence they have lower R values) than do gentle anticlinalfolds. Tight synclinal folds generally produce larger amounts ofslip (hence they have higher R values) than do gentle anticlinalfolds.

In both synclinal and anticlinal fault-bend folds with a single faultbend, the width of the fold limb measured along the fault equalsthe slip (S1).

Anticlinal fault-bend folds Synclinal fault-bend folds


19

Seismic interpretation of a synclinal fault-bend foldThis section describes the interpretation of a synclinal fault-bend fold imaged in seismic reflec-tion data. The lower portion of the fault and the syncline are well imaged, and fault-bend fold-ing theory is used to predict the orientation of the upper portion of the fault.

In Figure 1, fault-plane reflections define the position of a thrust ramp locatedbeneath a syncline. Based on the imaged fold shape and fault ramp, the initial cut-off angle (θ) and interlimb angle (γ) can be measured as:

θ = 15°; γ = 82°

Using the synclinal fault bend fold graph (Figure 2), γ and θ are used to determinethe change in fault dip (φ) and the hanging wall cutoff after the fault bend (β):

φ = 15°; β = 14°

φ and β values are used to model the structure in Figure 3. Note that the predictedupper fault segment agrees closely with the fault position as constrained by reflec-tion terminations and potential fault-plane reflections.

Synclinal fault-bend fold, Argentina

2. Synclinal fault-bend fold graph

3. Prediction1. Observations/Initial Interpretation


20

Seismic interpretation of an anticlinal fault-bend foldThis section describes the interpretation of an anticlinal fault-bend fold imaged in seismicreflection data.

In Figure 1, fault-plane reflections and reflection truncations define the position of a thrustramp located beneath an anticline. Based on the imaged fold shape and fault ramp, the initialcut-off angle (θ) and interlimb angle (γ) can be defined as:

θ = 24°; γ = 80°

Using the anticlinal fault bend fold graph (Figure 2), γ and θ are used to determine the changein fault dip (φ) and the hanging wall cutoff after the fault bend (β):

φ = 16°; β = 28

φ and β values are used to model the structure in Figure 3. Note that the predicted upper faultsegment agrees closely with the fault position as constrained by reflection terminations andthe downward termination of the forelimb.

In this example, slip below the fault bend (S0) is also interpreted based on offset reflections.Based on the slip ratio R predicted for this fault-bend fold (obtained using the graph present-ed in the previous section), the slip above the fault bend (S1) is calculated as follows:

R = (S0/S1) = 0.87; given S0 = 1.7 km, then S1 = 1.5 km

2. Anticlinal fault-bend fold graph

Anticlinal fault-bend fold, Niger Delta1. Observations / Initial Interpretation 3. Prediction


21

Composite fault-bend folds: Ramp anticlinesThe most common representation of a fault-bend fold involves deformation above a thrust ramp connecting upper and lower detachments, often referred to as a ramp anticline. In fact, this structureconsists of two fault-bend folds — one related to each fault bend — and thus is part of a class of “composite” fault-bend folds. This section describes the kinematic evolution of a simple ramp anti-cline after Suppe (1983), the geometry of which is governed by the quantitative fault-bend folding theories described in the preceding pages.

Kinematic development of a composite fault-bend fold Seismic Example: Pitas Point, Santa Barbara Channel, California, U.S.A.

0: An incipient thrust fault and axial sur-faces in undeformed strata.

1: Fault slip causes folding of the hangingwall block along active axial surfaces Aand B that are pinned to the two faultbends. Inactive axial surfaces A� and B�form at fault bends and are passivelytranslated away from active axial surfacesby slip. Kink-band width A-A� or B-B� mea-sured along bedding equals slip on theunderlying fault segment. The differencein kink-band width between back andfront limbs reflects slip consumed in fold-ing.

2: Progressive fault slip widens both kinkbands. Models 1 and 2 are in the crestaluplift stage because the fold crest elevateswith increasing fault slip.

3: When the axial surface A� reaches theupper fault bend, material from the backlimb is refolded onto the crest and thefront limb kink-band B-B� is translatedalong the upper detachment. In model 3,A and A�� are active axial surfaces; B andB� are inactive axial surfaces. Model 3 is inthe crestal broadening stage because thefold crest widens without producing addi-tional structural relief with increasingfault slip. In the crestal broadening stage,slip exceeds the width of the fold limbs,and is equal to the distance between axialsurfaces A-A� measured along the fault.

Uninterpreted section

Interpreted section


22

“Multi-bend” fault-bend foldsIn addition to simple ramp anticlines, composite structures include multi-bend fault bend folds (Medwedeff and Suppe, 1997), which contain two or more bends of similar concavity or convexity.Initially, slip across each bend produces a distinct kink band; however, with progressive fault slip, kink bands merge and interact. These interactions can be highly complex, spawning many new axialsurfaces and dip domains. Thus, multi-bend fault-bend folds are generally characterized by the presence of multiple dip domains in backlimbs and forelimbs. Figures below show kinematic modelsof multibend fault-bend folds and a seismic example.

Kinematic development of multi-bend fault-bend folds Convex upward (anticlinal) bends Concave upward (synclinal) bends

0: Incipient fault with two convex upward bends.1: Slip yields two kink bands associated with thetwo fault bends. 2: Kink bands widen with pro-gressive slip. 3: Portions of kink bands are refold-ed, yielding a steeply dipping fold panel.

0: Incipient fault with two concave upward bends.1: Slip yields two kink bands associated with thetwo fault bends. 2: Kink bands widen with pro-gressive slip. 3: A portion of the lower kink band isrefolded as it moves onto the upper fault ramp.

Seismic Example: Niger DeltaMulti-bend fault

Interpreted section

refolded refolded

bends

fault

axial surface

axial surface

fault


23

Modeling curved fold hingesFolds generally exhibit some curvature in their hinges. Most fault-related fold analysis techniques approximate these curved hinge zones as perfectly angular folds or as multi-bend foldscomposed of two or more planar hinge segments (Medwedeff and Suppe, 1997). In many cases, these approximations adequately describe large folds, with small zones of hinge curvatureseparating long, planar fold limbs of the scale typically imaged in seismic data. Moreover, these approximations are useful because they allow for rigorous area and line length balancing.In some cases, however, it may be necessary to more accurately describe curved hinge zones. Here we introduce a curved-hinge fault-bend fold model after Suppe et al. (1997), which obeysfault-bend folding relations but imparts fault curvature on the fold shape using the concept of entry and exit axial surfaces. Other techniques of modeling curved fold hinges (e.g., trishearfolding — Erslev, 1991) are described in later sections.

Sequential models of a syncli-nal fault-bend fold with anangular hinge.

Sequential models of a multi-bend synclinal fault-bend foldwith two fault ramp segments.

Synclinal fault-bend folds Angular Hinge Multibend Hinge Curved Hinge

Sequential models of a curved hinge syn-clinal fault-bend fold. 0: Two incipient ac-tive axial surfaces bound the zone of cur-vature on the fault. 1: Slip causes foldingof the hanging wall rocks. Folding beginsas rocks pass through the entry activeaxial surface (A), and ceases as rockspass through the exit active axial surface(B). 2: Progressive slip widens the kinkband, as inactive axial surfaces (A� andB�) are passively translated up the faultramp.

Seismic Example: Sichuan basin, China

Uninterpreted section

Interpreted section

incipient axial surface incipient axial surface

active axial surface

inactive axial surfaceactive axial surfaces

inactive axial surfaces

incipient entry axial surfaceincipient exit axial surface

entry active axial surfaceexit active axial surface


24

Seismic Example: Santa Barbara Channel, California, U.S.A.Growth fault-bend folds — high sedimentation ratesFault-bend folds develop by kink-band migration, where fold limbs maintain a constant dip but gen-erally widen as fault slip increases. When sedimentation rate exceeds uplift rate, folds that devel-op by kink-band migration have syntectonic (growth) strata that form narrowing upward dipdomains, or growth triangles, above fold limbs (see section 1A-3). Below, we use kinematic modelsto describe how these growth structures develop in a composite fault-bend fold, and show exam-ples of growth structures in seismic sections.

Fault-bend fold with growth strata

Sequential model of a growth fault-bend fold (Suppe et al., 1992; Shaw et al., 1996) with sedimenta-tion rate > uplift rate. Model 1 consists of a composite fault-bend fold developed above a rampbetween detachments. The fold is in the crestal uplift stage of growth (Shaw et al., 1994b), as faultslip is less than ramp width. In Model 2, additional slip widens the kink bands, which narrowupward in the growth section (see section 1A-4). In Model 3, fault slip is greater than ramp width.Thus, strata are refolded from the back limb (A-A��) onto the crest of the structure, which widenswith fault slip (crestal broadening stage, Shaw et al., 1994b). Growth strata are also folded abovethe crest, as they pass through active axial surface A��. Forelimb axial surfaces (B-B�) are releasedfrom the fault bend and passively translated above the upper detachment, and thus do not deformyoung growth strata.

Seismic Example: Los Angeles basin, California, U.S.A.

Active synclinal axial surface — backlimb FBF

Active anticlinal axial surface — forelimb FBF

crestal uplift stage


crestal broadening stage

growth

pre-growthforelimb


25

Growth Fault-Bend Folds — low sedimentation ratesIn cases where sedimentation rate is less than or equal to the uplift rate, fault-bend folds develop patterns in growth strata that are distinct from growth triangles (see section 1A-3). In limbs withactive synclinal axial surfaces, growth strata are folded concordantly with the underlying kink band; whereas, in limbs with inactive synclinal axial surfaces growth strata simply onlap kink bands.Below we describe how these growth patterns are expressed in a composite fault-bend fold after Medwedeff (1989) and Suppe et al. (1992).

Fault-bend fold with growth strata

Sequential model of a growth fault-bend fold (Medwedeff, 1989; Suppe et al., 1992) with a sedi-mentation rate equal to the uplift rate. Model 1 consists of a composite fault-bend fold developedabove a ramp between detachments. Growth strata in the backlimb are folded concordantly withthe underlying kink band. In contrast, undeformed growth strata onlap the forelimb. In Model 2,additional slip widens kink bands and the growth pattern is maintained. In Model 3, fault slip isgreater than ramp width. Thus, strata are refolded from the back limb (A-A��) onto the crest ofthe structure, which widens with fault slip. Growth strata are also re-folded above the crest, asthey pass through active axial surface A��. Formerly inclined growth strata from the backlimbbecome horizontal. Coeval deposition above the fold crest forms a time trangressive angularunconformity. In Model 3, the sedimentation rate is held constant and equal to the uplift rate ofparticles within the back limb.

Seismic Example: San Joaquin basin, California, U.S.A.Composite Fault-Bend Fold with Growth Strata

Seismic reflection profile across the Western San Joaquin basin (Lost Hills anticline) showingcontrasting patterns of growth strata between backlimb (west) and forelimb (east) that areconsistent with fault-bend folding where sedimentation rate is less than or equal to uplift rate(see model 2, left). The fanning of limb dips above the front limb may be due to sedimentarydrape and compaction, or may reflect a component of limb rotation in fold growth (see section1A-3).




growth

pre-growth

backlimbforelimb

foldedgrowth strata


time transgressiveangular unconformity


foldedgrowth strata


26

1B-2: Fault-propagation foldsBasic conceptFault-propagation folds form at the tips of faults and consume slip. These folds are gener-ally asymmetric, with forelimbs that are much steeper and narrower than their correspond-ing backlimbs. Several modes of folding at fault tips have been described to explain thesestructures, including: constant thickness and fixed axis fault propagation folding (Suppe andMedwedeff, 1990); trishear folding (Erslev, 1991; Hardy and Ford, 1997; Allmendinger, 1998);and basement-involved (triple junction) folding (Narr and Suppe, 1994). In this section, wedescribe these kinematic theories, emphasizing their common characteristics, and introducebasic techniques for interpreting fault-propagation folds in seismic data.

Schematic fault-propagation fold modelTo describe the basic concept of fault-propagation folding, we will consider thehypothetical case of a fault ramp in crosssection that propagates upward from adetachment (note that fault-propagationfolds may originate from faults with orwithout detachments). As the fault ramppropagates upward in sequential models0 to 2, an asymmetric fold develops in thehanging wall with vergence in the trans-port direction. The fold consumes slip onthe ramp, with slip being greatest at theramp base and zero at the fault tip. As slipincreases, the fault tip advances and thefold grows larger while maintaining thesame basic geometry.

Common characteristicsAlthough fault-propagation folds exhibit awide range of geometries, several charac-teristics are common to most structures,including:

1) folds are asymmetric, with forelimbs thatare generally much steeper and more nar-row than their corresponding backlimbs;

2) synclines are pinned to the fault tips;

3) folds tighten with depth; and

4) slip on the fault decreases upward, ter-minating within the fold.

ExamplesFault-propagation folds are common inoutcrop and at scales typically imagedby seismic reflection data. This fieldexample (right) has several characteris-tics of fault-propagation folds, includingasymmetry, the presence of a narrow,steeply dipping forelimb, and the down-ward increasing tightness of the fold.

The seismic example is a fault-propaga-tion fold at the southern margin of theTanan Uplift in the southern Tarimbasin. In this example, a thrust rampdelineated by fault-plane reflections ter-minates upward into the forelimb of anasymmetric fault-propagation fold.

Field Example

Professor Bill Brown highlighting a fault-propagation fold in Cambrian Fort Sillslimestone, Arbuckle Mountains, OK, U.S.A. (S.C. Hook)

Seismic Example: Tarim basin, China

fault

fault

purple arrows denote slip on thebase and top of the green unit

fault tip


27

These graphs show therelationships betweenfault shape (θ2) and foldshape (γ and γ*) for con-stant thickness fault-prop-agations folds. The spe-cial case of ramping froma detachment is shown asthe lines θ2 = φ. Theserelations will be used tointerpret a fault-propaga-tion fold imaged in a seis-mic profile later in thissection.

Constant thickness fault-propagation folds Suppe and Medwedeff (1990) present a general relationship between fold shape and faultshape for parallel (constant thickness) fault propagation folds assuming angular foldhinges and conservation of bed length. This section describes the kinematic develop-ment of a constant-thickness fault-propagation fold, and the quantitative relations thatcan be used to model or interpret these structures.

Constant thickness fault-propagationfolds develop as a fault propagatesupward from a bend. An active, syncli-nal axial surface is pinned to the faulttip. As strata pass through this axialsurface, they are folded into the fore-limb. Depending on the fault geome-try, strata may also pass through theanticlinal axial surface into the fore-limb, or from the forelimb onto thefold crest. The backlimb developsmuch like a fault-bend fold, althoughthe limb width is typically greater thanfault slip.

Fault-propagation folds have severalgeometric relations that are useful inconstructing models and interpretingstructures, including:

1) The distance between the faultbend and the point where the anticli-nal axial surface meets the fault equalsthe fault dip-slip at the bend.

2) The bifurcation point of the anticli-nal axial surface occurs along thesame bedding horizon as the fault tip.

Kinematic Model

FPF terminologyThe following terms are used in the derivationand graphs that describe fault-propagation folds.

θ1 = hanging wall cut-off (lower fault segment)θ2 = footwall cut-off (upper fault segment)φ = change in fault dipγ = forelimb syncline interlimb angleγ* = anticlinal interlimb angle�b = backlimb dip�f = forelimb dip


28

Fixed-axis fault-propagation foldsSuppe and Medwedeff (1990) present a second, general relationship between foldshape and fault shape called fixed-axis fault-propagation fold theory. This theory issimilar to the constant thickness theory, except that it allows for bed thinning orthickening in the forelimb (see also Jamison, 1987). These thickness changes areinduced because the forelimb anticlinal axial surface is fixed, meaning that materialdoes not pass through it. The style and magnitude of bed thickness changes are dic-tated by the initial fault shape and cut-off angles. This section describes the kine-matic development of a fixed-axis fault-propagation fold, and the quantitative rela-tions that can be used to model and interpret these structures.

These sequential mod-els (0–2) illustrate thatfixed-axis fault propaga-tion folds develop in asimilar manner to con-stant-thickness fault-propagation folds.However, the anticlinalaxial surfaces are fixed(inactive), causing fore-limb thickening or thin-ning. Folds with low cut-off angles generallyexhibit forelimb thicken-ing, whereas, folds withhigh cutoff angles gener-ally exhibit forelimbthinning.

Kinematic Models

FPF terminology Fixed-axis theory redefines the axial angles (γvalues) associated with a fault-propagationfold. The remaining parameters (θ, φ, δb, and δf)are the same as in constant thickness fault-propagation folds.

γe = forelimb syncline exterior axial angleγi = forelimb syncline interior axial angleγe*= anticlinal exterior axial angleγi* = anticlinal exterior axial angle

These graphs show the relationships between fault shape (θ2) and fold shape (γe, γe*, γi, and γi*) forfixed-axis fault-propagations folds. The special case of ramping from a detachment is shown on the twographs at left as the lines θ2 = φ. Note that separate graphs must be used to define the interior (γi, andγi*) and exterior (γe and γe*) axial angles.

forelimb thickens forelimb thins

fixed axial surfacefixed axial surface

with forelimb thickeningwith forelimb thinning


29

Seismic interpretation using fault-propagation fold theoryThis section presents an interpretation of a structure imaged in seismic reflection data as a fault-propagation fold as described by Suppe and Medwedeff (1990). The seismic profile shows a highlyasymmetric fold, with a poorly imaged forelimb, which are characteristics of many seismic images of fault-propagation folds.

The seismic section shown below is interpreted in five steps on this and the following page. To help distinguish between the two alternative theories, the graph below (from Suppe and Medwedeff,1990) shows the relationship of forelimb to backlimb dips for both constant thickness and fixed axis fault propagation folds. Pairs of limb dips that plot along the “Fixed-Axis Theory” curve indicatethat the structure may be interpreted using this theory. Limb dips that plot along, or to the left of, the φ = θ2 curve may be interpreted using constant-thickness theory. The two theories are coinci-dent along the portion of the “Fixed-Axis Theory” curve that lies on, or to the left of, the φ = θ2 curve.

Limb dips estimated from seismic profile

Step 1: Limb dips are estimated in the seismic profile by interpre-tation of the reflector dips on the backlimb, and by correlation ofhorizons 1 and 2 across the poorly imaged forelimb.

Limb dips in fault-propagation folds

Step 2: Based on the forelimb (δf = 58°) and backlimb (δb = 11°) dips estimated on the seis-mic profile, the fold is inconsistent with fixed-axis theory. However, the structure may beinterpreted as a constant thickness fault-propagation fold with a change in fault dip (φ) of 7°and an initial cutoff angle (θ1) of 42°. On the following page, these values are used to predictthe fold shape (γ and γ*) and cutoff (θ2) angles, and to generate an interpretation of thestructure.


30

Step 5: The interpretation is completed by extending the fault down from its tip at an angle of49° (based on θ2) to the point where it intersects the backlimb synclinal axial surface. At thispoint, the fault shallows by 7° (based on φ) to a dip of 42°. The interior anticlinal axial surfacebisects the interlimb angle between the forelimb and backlimb, and extends down to the fault.The distance between the point where this axial surface intersects the fault and the fault bendequals the fault slip at the bend.

In summary, this model-based interpretation provides an internally consistent, area balanceddescription of the structure that honors the seismic data. In general, constant-thickness andfixed-axis fault-propagation fold theories are most applicable to structures with pairs of dis-crete, parallel axial surfaces bounding fold limbs with roughly constant bed dips. Bed thicknesschanges in the forelimb, relative to other parts of the structure, are best explained with fixed-axis theory. Comparisons of the forelimb and backlimb dips can also be used to distinguishbetween these two alternative theories. On the following pages, we describe other modes offolding that may better describe structures with broadly curved fold hinges, variable forelimbdips, non-parallel axial surfaces, and/or substantial footwall deformation.

Step 3: To interpret the structure using con-stant-thickness fault propagation fold theory,the upper portion of the fold is interpretedusing the kink method, where axial surfacesbisect the interlimb angles (see section 1A-1).This interpretation yields a forelimb interlimbangle (γ) of 61°.

The tip of the fault is located by projecting theaxial surfaces that bound the fold crest totheir point of intersection. From this intersec-tion point, follow bedding along the forelimb(as defined by δf) until it intersects the fore-limb synclinal axial surface. This intersectiondefines the tip of the fault.

Initial Interpretation

Step 4: The remaining fault-propa-gation fold parameters (θ2 and γ*)are then determined from one ofthe two constant thickness fault-propagation fold graphs. Given a γvalue of 61° and a change in faultdip (φ) of 7° (from precedingpage), the theory predicts aninterlimb angle (γ*) of 55.5° and acut-off angle (θ2) of 49°. These val-ues are used to complete theinterpretation.

Complete Interpretation


31

Trishear fault-propagation foldsErslev (1991) proposed an another mode of fault-propagation folding, known as trishear. Trishear folds formby distributed shear within a triangular (trishear) zone that expands outward from a fault tip. Folds developin the trishear zone and cross sectional area, but not bed thickness or length, and are preserved throughdeformation. The displacement field, and thus fold shape, is straightforward to calculate. However, it must bedone iteratively. Hence, the method cannot be applied graphically or analytically (Allmendinger, 1998). Here,we describe some of the basic characteristics of trishear folds, and use the theory as implemented by Hardyand Ford (1997) and Allmendinger (1998) to model and interpret these structures.

Seismic section

Trishear interpretation

The fault-propagation fold in this seismic section has a broadening upward zone offolding and a fanning of forelimb dips (1). These patterns are forward modeledusing trishear, based on parameters derived through an inversion method(Allmendinger, 1998). The best fitting model is displayed on the seismic section inthe lower panel.

In summary, trishear folds are easily distinguished from constant-thickness andfixed-axis fault-propagation folds, in that they display an upward-widening, curvedfold limb ahead of the fault tip, which leads to an upward decrease in limb dip.

This sequential model (0 - 2) shows the development of atrishear fault-propagation fold at the tip of a thrust rampthat steps upward from a detachment. The backlimb of thestructure is a simple fault-bend fold. The geometry of theforelimb is a function of the apical angle, the fault dip, andthe P/S ratio. Small apical angles generally yield tight, high-ly strained forelimbs, whereas large apical angles generallyyield broad, gently strained forelimbs. At a given apicalangle, the steepness of the forelimb increases with pro-gressive slip. The steepness of the forelimb also increasesdownward. This pattern is characteristic of trishear folds,and contrasts with the constant forelimb dips exhibited byconstant-thickness and fixed-axis fault-propagation folds.

Theory Kinematic model

The trishear zone (a-b-c) is bound by two sur-faces that define an intervening apical angle.The surfaces may or may not be symmetricwith respect to the fault (Zehnder andAllmendinger, 2000). To preserve cross sec-tional area (a-b-c = a-a�-b-c) during deforma-tion, there must be a component of displace-ment toward the footwall, as reflected by thevelocity vectors. To model a trishear fold, theapical angle, the fault dip, and the propagationto slip ratio (P/S) of the fault are specified.(after Erslev, 1991; and Allmendinger, 1998).

Propagation to slip ratio

Fault propagation to slip ratio (P/S) has an important influence on fold shape. Low P/S ratios generallyyield steep, tightly folded forelimbs with pronounced bed thickening. High P/S ratios generally yield shal-low, gently folded forelimbs with less bed thickening (from Allmendinger, 1998).


32

Basement-involved (drape) folding with migrating triple junctionsFault-propagation folds that involve basement (crystalline) rock are commonplace, and tend to have shapes that differ from those described by constant-thickness and fixed axis fault-propagationfold models (Suppe and Medwedeff, 1990). Several geometric and kinematic theories have been developed to explain these structures, including models with forelimb shear distributed in triangularzones, such as the trishear model described in the preceding section (Erslev, 1991; Mitra and Mount, 1998). This section describes another kinematic theory proposed by Narr and Suppe (1994), inwhich fold growth is governed by the migration of a fault-fault-fold triple junction. The theory is then applied to interpret a fault-propagation fold in seismic data.

Kinematic model

Triple junction fold terminologyFive parameters describe basement-involved triple junc-tion folds, three of which must be specified to derive theremaining two values:

θ1 = hanging wall cutoff of lower fault segmentε = dip of upper fault segment (generally = 180°- δf)β = dip of footwall monoclineφ = dip of footwall shear orientationψ = footwall angular shear

Fold and fault shape Seismic Example: Orito Field, Putamayo basin, Colombia

Seismic profile of a basement-involved fault propagation. Thefootwall monocline and steep (poorly imaged) forelimb arecharacteristic of triple junction fault-propagation fold models.In the interpretation, the shear orientation (φ) and angle (ψ)were estimated from the graph at left using: 1) the maximumforelimb dip value (δf), estimated from oriented well core andsurface dips, to define ε (120°); 2) the reflection truncations toestimate the fault dip (θ = 60°), and; 3) the dip of the footwallmonocline (β = 9°). The interpreted section involves additionaldeformation induced by a breakthrough of the main fault, a pro-cess which is described later in this section, but neverthelessthe structure maintains the basic geometry described by themigrating triple junction theory of Narr and Suppe (1994).

These graphs describerelations among thefive parameters thatdescribe triple-junc-tion folds. Each graphis for a specific ε value.When modeling struc-tures imaged in seis-mic sections, ε is gen-erally selected byinterpreting the fore-limb dip value (δf). Thedip of the footwallmonocline is also com-monly resolved onseismic sections, leav-ing one additional pa-rameter to be deter-mined (φ or θ) before aunique solution can beobtained. From Narrand Suppe (1994).

In the Narr and Suppe (1994) basement-involved model, folding is driven by themigration of a fault-fault-fold (axial sur-face) triple junction. The triple junctionmoves upward with progressive fault slip,causing shear of the footwall that forms amonocline. Uplift of the hanging wall alsoinduces folding of the sedimentary cover,producing a forelimb with bed dips thatare parallel to the dip of the upper faultsegment. Stages 0–2 show progress devel-opment of a migrating triple junction foldmodel.


33

Growth fault-propagation folding Syntectonic (growth) strata are folded in distinctive patterns above fault-propagation folds. Forelimb growth struc-tures, in particular, vary among the different fault-propagation fold models and thus can be diagnostic of the foldingmechanism. In this section, we contrast growth patterns developed above fault propagation folds as described bySuppe and Medwedeff (1990) and trishear folds (Erslev, 1993), using kinematic models and examples imaged in seismicsections.

Kinematic models

Fault-propagation folds of Suppe and Medwedeff, (1990)grow by kink-band migration, with two active axial sur-faces bounding the backlimb, and one or two active axialsurfaces bounding the forelimb. Syntectonic strata abovethe fold limbs form growth triangles. When sedimenta-tion rate exceeds uplift rate, as in this model, two growthtriangles develop on the backlimb. Fixed-axis fault-prop-agation folds have a single forelimb growth triangle,whereas, constant thickness fault-propagation folds mayhave one or two forelimb growth triangles depending onthe fault geometry. This sequential model (0–2), with a29° fault ramp and a decollement, is a case where bothconstant-thickness and fixed axis theory converge toyield the same geometry.

Sedimentation rate relative to uplift rate can have a pronounced impact on resultant growth geometries. These threeexamples (a-c) show the effects of local non-deposition and erosion on growth structures in fault-propagation folds(after Suppe et al., 1992).

Growth trishear foldGrowth fault-propagation fold

Trishear folds (Erslev, 1993) generally develop bya combination of kink-band migration and limbrotation mechanisms, and these fold kinematicsare reflected in growth strata. Progressive fore-limb rotation during the formation of trishearfolds generally yields an upward shallowing ofbed dips in growth strata. This fanning of limbdips in trishear growth folds contrasts markedlywith the growth triangles predicted by the con-stant-thickness and fixed axis theories. Thissequential model (0–2) (after Hardy and Ford,1997) has a sedimentation rate that slightlyexceeds the uplift rate. The backlimb of thismodel forms by fault-bend folding, yielding a sin-gle backlimb growth triangle.

Effects of low sedimentation rates

Seismic Example: Bermejo anticline, Argentina

Seismic example ofa forelimb growthtriangle in a fault-propagation foldfrom the Bermejoforeland basin, cen-tral Argentina fromZapata and Allmen-dinger (1996).Reproduced cour-tesy of the Ameri-can GeophysicalUnion. Seismic Example: Tarim basin, China

Seismic example of fanning forelimb dips in growth strata from theTanan Uplift, Tarim basin, China. Section is overlain by a modeled tris-hear fold, described in the trishear folding section, that includes modeledgrowth horizons (yellow).

truncations

Growth axial surface Time transgressive unconformity

Active axial surface


34

Breakthrough fault-propagation folds At any stage of fold growth, faults may cut through fault-propagation folds, altering the geometries of these structures. The shapes of these “breakthrough” structures are influenced by the path ofthe fault, which often breaks through the forelimb or shallows to an upper detachment, as well as the folding mechanism. In cases where the slip on the breakthrough fault is substantial and/orstructures are deeply eroded, only remnants of the original fault-propagation fold geometries may remain. In this section, we use several kinematic models to describe styles of breakthrough faultpropagation folding, and show an example of this type of structure in a seismic section.

This seismic section illustrates a common forelimb breakthrough pattern. Although the forelimbis poorly imaged, reflection truncations and the hanging wall and footwall positions of the corre-lated horizon suggest that the fault extends through the structure. Nevertheless, the basic geom-etry of the fold is consistent with a fault-propagation folding mechanism, implying that this is abreakthrough structure.

This sequential model (1–2) shows a constant-thickness fault propagation fold (1) wherethe fault breaks through the middle of the forelimb (2). The fault modifies the original foldgeometry by offsetting the hanging wall portion of the forelimb, and producing an addi-tional kink band within the backlimb that develops by fault-bend folding.

Forelimb breakthrough

Breakthrough styles

Kinematic models

Models showing possible types of breakthrough structures after Suppe and Medwedeff(1990). a and b) decollement breakthroughs; c) synclinal breakthrough; d) anticlinal break-through; e) high-angle (forelimb) breakthrough; and f) low-angle breakthrough.

Trishear fold breakthrough Triple junction fold breakthrough

Faults in trishear and triple-junction fault-propagation folds may also breakthrough at any stageof fold growth. These models are examples of synclinal fault breakthroughs in: a) trishear foldafter Allmendinger (1998); and b) a triple junction model after Narr and Suppe (1994). The geome-tries of breakthrough structures in all classes of fault-propagation folds vary substantially basedon the fault path and, if the fault is non-planar, on folding kinematics after breakthrough.

Seismic Example: Argentina


35

1B-3: Detachment foldsBasic ConceptDetachment folds form as displacement along a bedding-parallel fault is transferred into foldingof the hanging wall layers. Although detachment folds may share some geometric similarities withfault-bend and fault-propagation folds, they differ from these structures because they are notdirectly related to thrust ramps but rather to distributed deformation above detachments. In thissection, we describe basic aspects of the geometry and kinematics of detachment folds. Theseinsights are used to guide the interpretation detachment folds in seismic images.

Styles of Detachment FoldsDetachment folds form at a variety of scales, as isolated structures or in long fold trains, andmany names are used to describe them. The term detachment fold is commonly applied to sym-metric or asymmetric folds that develop above a relatively thick ductile unit and basal detach-ment. If folds are symmetric, have steep limbs, and develop above a relatively thin ductile unit,they are often called pop-up or lift-off folds (Mitra and Namson, 1989; Mount, 1990). Lift-off foldsdevelop by isoclinal folding of the detachment in the core of the anticline, and when they haveflat crests they are referred to as box folds.

Kinematic models of detachment folds

Common characteristicsDetachment folds generally share the followingcharacteristics:1) An incompetent, ductile basal unit thickened in

core of fold, with no visible thrust ramp.

2) A detachment that defines the downward termi-nation of the fold.

3) Competent pregrowth units that, if present, gen-erally maintain layer thickness.

4) Growth units, if present, that thin onto the foldcrest and exhibit a fanning of limb dips.

ExamplesDetachment folds are common in outcrop and at scales typically imaged by seismic reflectiondata. They have been documented in the foreland of fold and thrust belts such as the Jura,Appalachian Plateau (Wiltschko and Chapple, 1977), and Tian Shan (Ferrari et al., section 2-14,this volume). Detachment folds are also common in passive margin fold belts, such as theMississippi Fan (Rowan, 1997) and Perdido Fold Belts (Carmilo et al., section 2-24, this volume)in the Gulf of Mexico, and in the Campos Basin, Brazil, (Demercian et al., 1993), and the NigerDelta (Bilotti et al., section 2-12, this volume).

The field and seismic examples shownhere have many of the common charac-teristics of detachment folds describedat lower left.

Field Example: Canadian Rockies

Seismic Example: Gulf of Mexico


36

Geometry and kinematics of detachment foldsThere is no unique, quantitative relationship between fold shape and underlying fault shape fordetachment folds, due in part to the ductile thicknening occurring in the fold core that gener-ally does not preserve bed length or thickness. Thus, it is often difficult to uniquely constrainthe geometry of these structures unless they are completely imaged. Nevertheless, several geo-metric and kinematic models have been developed (Dahlstrom, 1990; Ephard and Groshong,1995; Homza and Wallace, 1995; Poblet and McClay, 1996) that can serve as guides for inter-preting detachment folds in seismic images.

In this section, we present a geometric and kinematic model of detachment folding developedby Poblet and McClay (1996) that is particularly useful when analyzing growth strata associat-ed with detachment folding that involves a competent unit. These authors propose three dis-tinct mechanisms by which a fold can develop above a propagating detachment. In each of themodels, it is the geometry and kinematics of folding in the competent layer (in particular, limblengths and limb dips) that controls the folding. The incompetent, or ductile layer, is able toflow into, or out of, the fold core as deformation progresses. Layer thickness, line length, andarea are conserved in the competent layers. If the detachment level is allowed to change, or ifdifferential shortening occurs in the incompetent unit, then area is conserved in the ductilelayer as well.

Poblet and McClay (1996) present three modes of detachment fold growth that are illustratedin the figure to the upper right (models 1–3), and differentiated based on their folding mecha-nisms as follows:

1) Primarily Limb Rotation. In this model, the limb lengths remain constant but the limbsrotate to accommodate shortening. A small amount of material must move through the axialsurfaces, inducing a minor component of kink-band migration, as folding progresses. Theincompetent unit is area balanced only if the detachment level varies or differential shorteningoccurs in the incompetent unit.

2) Kink-band Migration. In this model, limb dips remain constant, but their lengths increase toaccommodate shortening. Material moves through the synclinal axial surfaces as folding pro-gresses. The incompetent unit is area balanced only if the detachment level varies or differen-tial shortening occurs in the incompetent unit.

3) Limb Rotation and Kink-band Migration. In this model, limb dips vary, as do limb lengths,but the ratio of the limb lengths remains the same. Strata moves through axial surfaces (pri-marily the synclinal surfaces), and rotate to accommodate shortening. The incompetent unitarea is balanced.

Two fundamental equations relate the shortening and uplift to the limb lengths and limb dipsof these detachment folds: (equations)

S = Lb (1 - cos ϑb) + Lf (1 - cos ϑf + Lt sin ϑf)u = Lb (sin ϑb) = Lf (sin ϑf)

based on the detachment fold terminology defined in the figure to the lower right.

Kinematic models of detachment folds

Detachment fold terminology

Lf = Front limb length

Lb = Back Limb length

S = Slip

ϑf = Front limb dip

ϑb = Back limb dip

u = Uplift


37

Growth strata associated with detachment foldsIt is usually not possible to determine the folding mechanism of a detachment anticline from thegeometry of pregrowth strata alone. For example, the three models on the previous page haveidentical final geometries, but the paths they took to get there (i.e., the fold kinematics), and thefolding mechanisms, were quite different. Growth strata are, however, typically diagnostic of fold-ing mechanism because they record the kinematic history of fold growth (see section 1A-3). Thus,growth strata can be used to distinguish between the modes of detachment folding described byPoblet and McClay (1996).

As illustrated in section 1A-3, kink-band migration causes growth strata to form narrowing-upward kink bands, or growth triangles, with bed dips that are parallel to those of the underlyingpregrowth strata. Growth triangles are bounded by at least one active axial surface. In contrast,limb rotation causes progressive changes in limb dips that result in a fanning of limb dips ingrowth strata. In limb rotation structures, a minor amount of material may still move throughaxial surfaces that are continuously changing orientation, resulting in a minor amount of kink-band migration. Poblet and McClay (1996) refer to these as “limited-activity axial surfaces.”

These models define the activity of axial surfaces that are involved in the three types of detach-ment folds defined by Poblet and McClay (1996):

Axial Surface Activity

Based on these fold kinematics, growth strata have distinctive patterns in each type of detach-ment folds that are shown in the models (1–3) at upper right, which are described as follows:

1) Primarily Limb Rotation. In this model, growth strata predominantly display fanning of dips,recording the progressive rotation of the fold limbs. Small growth triangles form that definegrowth strata which migrated through the limited-activity axial surfaces.

2) Kink-band Migration. In this model, growth strata form growth triangles because strata havemigrated through the active synclinal axial surfaces.

3) Limb Rotation and Kink-band Migration. In this model, growth strata display some fanning ofdip due to rotation of the fold limbs as well as growth triangles that record the migration of stra-ta through the active synclinal axial surfaces.

Kinematic models of growth detachment folds

This seismic line images a detachment anticline with patterns of growth strata that reflect fold-ing by both limb rotation and kink-band migration, suggesting that the structure is compatiblewith model 3 shown above.

Seismic Example: Gulf of Mexico


38

In this section, we describe the interpretation of a detachment fold imaged in a seismic reflection profile based on the foldmodels presented on the preceding pages.

Initial Observations. The fold from offshore west Africa shown at right is symmetric, with units that conserve layer thick-ness (3) and other units that clearly do not (1, 4). There is no obvious thrust ramp present, although reflectors underlyingthe fold are essentially flat suggesting the presence of a detachment (2).

Structural Interpretation. Based on the initial observations, this structure is interpreted as a detachment fold in the sec-tion at lower right. The detachment is interpreted to separate folded layers above from undeformed strata below. Abovethe detachment, a poorly imaged stratigraphic interval is thickened in the core of the fold (1). This incompetent unit rep-resents an Aptian salt bed. The units directly above the salt broadly conserve layer thickness (3), indicating these stratahave acted competently during deformation, probably deforming by flexural slip (see section 1A-2). The constant thicknessof the units also indicates that they were deposited prior to folding. Above these units, layers that thin onto the crest ofthe fold (4) are growth strata. The growth strata generally fan above the fold limbs, with only small panels in the limbs hav-ing the same stratigraphic thickness that they do in the synclines. Thus, the fold grew mostly by limb rotation with onlyminor kink-band migration, similar to the model 1 detachment fold of Poblet and McClay (1996).

Seismic interpretation of a detachment fold: Angola continental slope

Several techniques (e.g., Chamberlin, 1910; Epard andGroshong, 1993; Homza and Wallace, 1995) have beendeveloped to determine the depth-to-detachment beneathanticlines based upon balancing the area uplifted in thefold with the displaced area as shown below in model A. Incases where the detachment depth is know independent-ly, several authors have pointed out that the predictedand observed detachment depths do not always match(Wiltschko and Chapple, 1977; Jones, 1987; Dahlstrom,1990; Groshong and Epard, 1994; Homza and Wallace,1995; Poblet and Hardy, 1995). (In the case of the Angolandetachment fold interpreted in this section, the predicteddepth-to-detachment is greater than 15 km!). These dis-crepancies arise because balancing the uplifted area withdisplaced area has two implicit assumptions, namely that:1) The thickness of the ductile unit outside of the fold ismaintained, and; 2) All of the material in the thickened

zone comes from within the plane of the section. One orboth of these assumptions may be invalid for detachmentanticlines as well as other types of fault-related folds, asshown below in model B. In particular, detachment foldswith highly ductile cores involving salt or over-pressuredmuds often show withdrawal of material in the synclines(and away from fold), causing local thinning of the ductileinterval and subsidence of overlying strata. Withdrawnmaterial is presumably moved into the core of the fold.Alternatively, or in addition, material in the thickened coreof the fold may be derived from out of the plane of section.Both processes invalidate the assumptions of classicdepth-to-detachment calculations, leading to predicteddetachment depths that are generally far too deep. Thus,care should be taken to avoid applying these methods ofcalculating depth-to-detachment in detachment folds withductile cores.

Calculating detachment depth: Why doesn’t it always work?

Structural interpretation

Initial observations

Depth-to-detachment calculations

Model A shows the classic method of calculating the depth to detachment, based on the assumption that the uplift areais equal to the displaced area. The shortening, which is typically determined by unfolding a layer while conserving linelength, and the uplift area are used to calculate the detachment depth by:

depth-to-detachment = displaced area / shorteningModel B shows a typical detachment fold where the uplift area greatly exceeds the displaced area. In these cases, stan-dard depth-to-detachment calculations inaccurately predict detachment depths.


39

Basic conceptShear fault-bend folding produces ramp anticlines with very distinctive shapes thatreflect a significant non-flexural-slip component to the deformation. The structural styletypically shows long back-limbs that dip less than the fault ramp, in contrast with clas-sical fault-bend folding. This section describes the geometry and kinematics of shearfault-bend folding after Suppe, Connors, and Zhang (2004) and introduces basic tech-niques for recognizing and interpreting these structures in seismic images.

Recognizing the structural styleThe typical structural style for ramp anticlines produced by shear fault-bend folding hasback limbs that dip less — in many cases very much less — than the fault-ramp (1). If asignificant stratigraphic section is deposited over the back limb during fold growth ittypically shows evidence of limb rotation (2). These ramp anticlines also commonlyshow front limbs (3) that are quite narrow relative to their long back limbs.

1B-4: Shear fault-bend folds

Seismic Example: Cascadia Canada

Seismic Example: Niger Delta


40

Shear is the essenceClassical fault-bend folds (section 1B-1) deform by flexural slip of the beds as they slide overfault bends (A), conserving layer thickness. In contrast, shear fault-bend folds undergo addi-tional distortion of the hanging wall or footwall, that is they undergo additional shear. Thisadditional shear usually is concentrated in a weak detachment interval such as shale or evap-orite that deforms by bedding-parallel simple shear — like the geometric model below (B).Alternatively, shear may be more distributed as in the analog model from David Elliott (1976)based on sheets of paper (C) or it may involve a bedding-parallel shortening and thickening,which is called pure shear. Shear fault-bend folds can also form by some combination of pureand simple shear or by more heterogeneous deformation as shown below in the distinct-ele-ment mechanical simulation by Luther Strayer (D).

Models

Shear in a seismic example: Cascadia CanadaFlexural-slip unfolding of a shear fault-bend fold yields a hanging wall shape that doesn’t match thefootwall because there has been deformation in addition to flexural slip. In this example from theCascadia accretionary wedge, offshore western Canada, the hanging-wall fault shape is deter-mined by unfolding the layers while conserving line length. The difference between the unfoldedhanging-wall fault shape and the actual fault shape yields the shear profile, showing that there hasbeen layer-parallel simple shear. The shear is concentrated in the yellow and red basal layers.

A: Classic fault-bend fold B: Shear fault-bend fold

C: Analog model of shear fault-bend fold

D: Mechanical model of shear fault-bend fold

Interpreted section

Flexural-slip unfolding gives the shear profile


41

Fold typesEnd-member shear fault-bend foldingEnd-member shear fault-bend folding. We can understand the fundamentals of shear fault-bendfolding and quantitatively check our seismic interpretations by using two simple end-member theo-ries, both involving a weak basal decollement layer of thickness h (shown in yellow). In the simple-shear end member, the decollement layer undergoes bedding-parallel simple shear with no actualbasal fault, just a distributed zone of shear. In the pure-shear end member, the decollement layerslides above a basal fault and shortens and thickens in a triangular area above the ramp. Mixturesbetween these end members are possible, as shown at right, but many actual folds are close to theend members. Classical fault-bend folding is also an end member, with a basal layer of zero thickness(h = 0).

The shape of the fold shows us which stratigraphic interval is the decollement layer. The anticlinalaxial surface terminates at the top of the decollement interval at (A). The synclinal axial surface ter-minates at the bottom (B). Also, if there is pure shear, the synclinal axial surface (C) doesn’t bisectwithin the decollement layer because the latter is thickened above the ramp. These properties areuseful in seismic interpretation.

Graphs of end-member theory. These end-member shearfault-bend fold graphs give the balanced relationship betweenramp dip θ, back limb dip δβ, and shear (αe or α) across thebasal layer. The shear is tan d/h, where d is the displacementat the top of the basal layer and h is its thickness. The dip ofthe back syncline in the basal layer (ψ) is useful in the pure-shear and mixed cases.

The inset drawing of the simple-shear graph shows a modelshear fault-bend fold that corresponds to the yellow square (θ= 23°, δb = 6.5°, and αe = 42°). The drawing of the pure-sheargraph corresponds to the angles shown by the red square (θ= 34°, δb = 15.5°, α = 68°, and ψ = 30°).

Curiously, these shear fault-bend fold graphs also encompassclassical no-shear fault-bend folding, which is reached in thelimit of zero thickness h to the basal layer. Shear (d/h)becomes infinite (αe or α = 90°) and the limb dip becomesparallel the fault (θ = δb) (D).

Simple-shear end-member Pure-shear end member


42

Seismic interpretation of a simple shear fault-bend fold: Cascadia, Canada

Initial assessment. The structure imaged in this seismic section fromoffshore western Canada (Hyndman et al., 1994) shows the character-istics of a shear fault-bend fold, especially the steepness of the fault dip(35-40°) relative to the back limb dip (5-13°). A front limb much nar-rower than the back (1) is also typical of shear fault-bend folds.

Interpreting the ramp geometry. The fault picks (shown below in red)constrain the fault geometry and rule out strongly listric fault interpre-tations. Also, note that there is a downward dying out of the fault throw(2), with throw going to zero at the base of the ramp (3). This is char-acteristic of shear fault-bend folds, in contrast with classical fault-bendfolds.

Significance of synclinal geometry. The back syncline is planar,bisects the inter-limb angle (4), and terminates at the base of the faultramp (3), indicating a simple-shear rather than a pure-shear fault-bendfold (see models previous page).

Fault picks

Timing of growth. Onlapping shallow reflectors (5) show that 120 m ofgrowth strata have accumulated. Deformation began soon after termi-nation of slip on the shallow hinterland thrust to the east, as defined bya seismic horizon (6) that is folded in the backlimb of the shear fault-bend fold but is undeformed above the thrust tip in the hinterlandstructure. Thickness and dip variations in growth strata record defor-mation by limb rotation and kink-band migration (5), consistent withshear fault-bend folding.


43

Refining the interpretation. This structure is morecomplex than the simple models shown previouslybecause the fault ramp is not straight but composed oftwo segments dipping 35° and 40°. Furthermore thebacklimb has two kink bands ab and bc of different dips(1 and 2).

Testing the interpretation. Let us begin by treatingeach kink band of the backlimb (1 and 2) separately,predicting two shear amounts from the two limb dips.Then we will compare the predicted shear with theshear determined from unfolding the hanging wall tosee if our interpretation is consistent.

Applying the simple-shear graph (shown at far right),we find that a backlimb dip δb of 11-12° within the lowerkink band ab and a lower ramp dip θ of 35° predict anexternal simple shear αe of 31-32° (1'). This agrees withthe shear αe of 31° determined by unfolding the hang-ing wall while conserving bed length as shown below(1"). The backlimb dip δb of 5° within the upper kinkband bc and an upper ramp dip θ of 40° predict anexternal simple shear ae of about 8° (2'), which alsoagrees with shear determined by the unfolding (2").These quantitative tests give us more confidence thatour seismic interpretation of this ramp anticline as ashear fault-bend fold is reasonable.

Refining and testing the seismic interpretation: Cascadia, Canada

Two segments of the back limb

Two intervals of shear

Interpreted depth section


44

Note that in this heterogeneous simple-shear fold that the highest shear intervaldefines the base of the backlimb panel that most closely approaches the ramp dip.

Growth strata. The combination of limb rotation and limb lengthening thatoccurs in shear fault-bend folding is recorded by growth strata, as illustrated inthe sequential kinematic models (A1-A3) shown below. Fanning of dips record-ing limb rotation (1) and growth triangles recording kink-band migration (2)(see section 1A-4). Growth strata in the example from the Niger delta at rightshow evidence of limb rotation.

As mentioned above, the fold geometry in pre-growth strata approaches thegeometry of classical fault-bend folding, with bed dips (3) approaching theramp dip, in the limit of large shear (i.e., displacement). The sequential largeshear model at right (B1–B2), however, demonstrates that the component oflimb rotation is recorded in growth strata (4), and thus can be used to distin-guish large shear fault-bend folds from classical fault-bend folds.

Evolution of shear fault-bend foldsKinematic evolution. Both simple- and pure-shear fault-bend folds develop bycombinations of limb lengthening (kink-band migration) and limb rotation. Thegraphs at right show the relationship between limb dip and shear for both foldtypes. In the limit of large shear (i.e., displacement), the fold geometry in pre-growth strata approaches the geometry of classical fault-bend folding, with aback-limb dip that approaches the ramp dip (θ approaches δb). However, evenin these cases folds will grow with a component of limb rotation, recordingtheir shear fault-bend fold heritage.

Heterogeneous simple shear

Niger delta limb rotation

Large shear (displacement) fault-bend folds

Relationships of backlimb dip (δb) to shear (αe and a)

Given a constant ramp dip, the backlimb dip (δb) steepens as shear (αe and α) increases. Points A1 to A3 correspondto models presented at lower left.

Limb rotation plus kink-band migration


45

Initial assessment. This line shows ramp anticlines developed in overpressuredShikoku basin turbidites above the master detachment (D) of the Nankai troughaccretionary wedge. Note that the degree of shortening in the structures increas-es from south to north. Notice the qualitative characteristics of shear fault-bendfolds, including backlimb dips that are less than ramp dip (A). Nevertheless thesestructures are more complex than the end-member models because of superposedlow-amplitude detachment folding and secondary deformation, seen in both foot-walls and hanging walls (B).

This depth-migrated dip line passes through Ocean Drilling Project holes ODP-808and ODP-1174, which reach to the top of oceanic crust (C) (line NT62-8 Moore etal., 1990, 1991, 2002). The 19-meter-thick master detachment was cored in ODP-808just above transparent pelagic sediments of the Shikoku basin (D).

Seismic interpretation of pure-shear fault-bend folds: Nankai trough, Japan

Nankai trough, Japan


46

Strategy. We can test our qualitative interpretation by comparing the seismic geom-etry with the end-member theories.

Fault and limb geometry. In the seismic section shown at upper right, the fault-ramp is located based on reflector terminations shown as red arrows and by corefrom the ODP-808 hole (1). This gives a remarkably straight ramp, dipping at θ = 35°,which is much greater than the average dip of the irregular backlimb (δb = 11-13°),suggesting that this is a shear fault-bend fold. The back syncline in the strong reflec-tors (2) is displaced substantially to the hinterland of the base of the ramp, whichfavors pure-shear or mixed-shear models that we now test.

Comparing with the end-member theory. Plotting the backlimb dip δb = 13° andramp dip θ = 35° on the pure-shear graph at far right (3) predicts a back synclinal dipψ = 31° in the basal decollement layer, which quantitatively agrees with the seismicimage at right. In theory, the location of the top of the decollement layer (in orange)is at the inflection in the back syncline, which agrees with the location indicatedindependently by the fault cutoff of the back anticline (4) — supporting our pure-shear fault-bend fold interpretation. A complete interpretation is shown on the seis-mic image at lower right (see also Suppe et al., 2004).

Fault slip. The back-dip and ramp angles plotted on the graph (3) also give us theshear α = 69° of the basal layer. From this we can calculate the fault slip d = 390 m,based on a basal layer thickness h of about 230 m (tan α = d/h = 1.7).

Refining the seismic interpretation: Nankai trough, Japan

Pure-shear end member

Depth section Interpreted depth section

Picking the fault

Comparing the seismic with anend-member model


47

Basic conceptImbricate structures form by the stacking of two or more thrust sheets and are common in foldand thrust belts worldwide. Imbricate structures can form by break-forward propagation of thrustsheets, by break-backward thrusting, or with coeval motion on both deep and shallow faults. In thissection, we describe the basic characteristics of imbricate structures, and outline an approach tointerpret these structures in seismic profiles using imbricate fault-bend fold theory (Suppe, 1983;Shaw et al., 1999).

1B-5: Imbricate fault-bend folds

Imbricate structures develop where two or more thrust sheets are stacked vertically. These thrustfaults may or may not involve detachments, but imbricate structures are more common in regionswith detachments. In the sequential break-forward model (0–2) shown above, slip on the deepthrust fault produces a fault-bend fold that refolds the overlying thrust sheet. In the sequentialbreak-backward model (0–2), a pre-existing fault-bend fold is cut by a shallow, younger thrust ramp.

Common characteristicsImbricate fault-bend folds typically contain:

1) Two or more vertically stacked thrust ramps;

2) Bedding dips that change across thrust ramps; and

3) Fold limbs at high structural levels with multiple dip domains, reflecting refolding caused by multiple ramps.(Note: multiple dip domains may also be produced bymulti-bend fault-bend folds, see section 1B-1).

Break-forward imbricate Break-backward imbricate

These seismic sections show the three common characteristics described in the modelat left, including (1) multiple ramps, (2) changes in bedding dip across ramps, and (3)multiple dip domains in fold limbs

Seismic Example: Alberta Foothills, Canada

Seismic Example: Niger Delta, Nigeria

Kinematic Models


48

Interpreting break-forward imbricate structures using fault-bend fold theorySuppe (1983) presents a strategy for interpreting break-forward imbricate structures based on thefact that each lower imbricate increases the dips in the overlying imbricates by fixed or quantumamounts that are predictable using fault-bend fold theory. Here we assume that bed-length andthickness are conserved and that all faults step up from a detachment at the same initial step-upangle (ramp dip). This section describes how to implement this approach to interpret imbricatestructures imaged in seismic sections.

TheoryImbricate fault-bend fold theory describes the increas-es in dip order caused by refolding of shallow thrustsheets by younger and deeper faults. In model 0, with asingle thrust ramp A, the forelimb and backlimb dip val-ues are first order (-I and +I), because each limb wasformed by strata passing over a single fault bend.Incipient thrust B is shown in the footwall of thrust A. Inmodel 1, slip on fault B refolds the shallow thrust sheet,producing second order (-II and +II) dip panels. Thesesecond order panels were folded once by thrust A, andagain by thrust B. The dips of the forelimb and backlimbpanels (-I, +I, -II, and +II) are prescribed by fault-bendfold theory based on the initial cutoff angles (θ).

Forelimb and backlimb dip values are based on the initial cutoff angle (θ) and the number of imbri-cated thrusts. This table shows the prescribed forelimb and backlimb dips for first- through sev-enth-order (I-VII) panels based on 8 to 24° fundamental cutoff angles. The order of the dip panel (I-VII) generally corresponds to the number of imbricated faults.

Dip panels are typically measured on seismic sections, and then compared with rows of prescribedvalues. If a general match between observed and prescribed dip values is obtained, then the struc-ture can be interpreted using this table. If a match is not obtained, it may suggest that the initialcutoff angles of the ramps are not equal, requiring use of values different that those on this table(see Mount et al., 1990). These more complex situations can be interpreted using the folding vec-tor technique presented on the next page.

Dip values measured on seismic profile

Two backlimb dip values are observed in this seismic section near the well. The lesser value (-I = 13°)occurs between faults A and B, and in the hanging wall of fault A to the right of the well. The steepervalue (-II = 25°) is restricted to the hanging wall of fault A. These two backlimb dip values are com-pared with the values shown in the table at lower left, to determine if they are consistent with imbri-cate fault-bend fold theory.

Interpreted section

The two backlimb dip values (-I = 13° and -II = 25°) correspond to a 13°initial cutoff angle based on thetable at left (see row highlighted in yellow). Thus, the geometries of faults A and B can be interpretedas part of a break-forward thrust sequence. The lower fault (B) dips at 13°, corresponding to the pre-scribed initial cutoff angle. It shallows to upper and lower detachments based on simple fault-bendfold theory (see section 1B-1) with θ = φ = 13°. The upper fault (A) dips at the second-order value (-II= 25°) where it lies above the backlimb kink band formed by fault B. Where fault A extends beyond theunderlying backlimb kink band, it dips at -I = 13°, corresponding to the prescribed initial cutoff angle.The geometries prescribed by the table match the reflection patterns closely. Note, however, thatother faults in the section further complicate some aspects of the geometry.


49

Interpreting break-forward imbricate structures using folding vectorsHere we describe a method of interpreting break-forward imbricate structures using folding vectors(Shaw et al., 1999). This method can be applied to a wide range of structures, including imbricatesystems where initial cutoff angles of faults vary, bed thickness changes occur, or faults do not soleto detachments. Folding vectors describe the relative displacement of bedding or other surfaces,such as faults, across a fold limb or kink band. Thus, folding vectors can be used to describe therefolding of overlying thrust sheets due to imbrication. In this section, we describe how to deter-mine folding vectors and use them to interpret a break-forward imbricate structure imaged in aseismic section.

Using folding vectorsTo describe how folding vectors are used to interpret break-for-ward imbricate structures, we will consider the case of a shal-low thrust sheet (above fault A) being refolded by a deeperthrust (B). In model 1, slip on the deep thrust B has produceda backlimb kink band that must refold the overlying thrustsheet (A). Hence, the orientation of fault A, and beds in its hang-ing wall, will change as the thrust sheet passes over the under-lying kink band. In model 2, the deflection of bedding across thedeep kink band is used to determine the folding vector (U).Folding vectors are measured parallel to axial surface orienta-tions. The deflection of thrust A across the deep kink band isdescribed by vector X, which is equal to the folding vector U.This results in shear, and hence line length, being preservedparallel to the axial surface orientation. The orientation of bed-ding that is refolded in the hanging wall of fault A can be deter-mined using fault-bend fold theory (see section 1B-1), or byusing folding vectors as shown in model 3. However, in this(and perhaps many) cases, the axial surface orientationchanges between the footwall and hanging wall of fault Abecause bed dips change. Thus, the new hanging wall axial sur-face orientation must be used to measure a new folding vector(Y), which is equal to the deflection of fault A. This folding vec-tor, in turn, equals the deflection of bedding in the hanging wallof fault A that is described by vector Z.

This method also applies in cases where axial surfaces do notbisect interlimb angles, and thus bed thickness is not pre-served. In all cases, however, proper use of folding vectorsresults in area-balanced interpretations.

Note: This method can also be used to model the folding of angularunconformities, sedimentary growth wedges, and other cases wherebed dips within a kink band are not parallel.

Measuring a folding vector

Interpreting a folded thrust

The folding vector method is used tointerpret this seismic section, in whichfault A is refolded by an underlyingkink band bounded by axial surface S�.Fault A enters the left side of the kinkband at a dip of 22°. The folding vectorU is measured as the deflection of abed* across axial surface S� in the foot-wall of fault A.

*Note that folding vectors must be mea-sured parallel to, but not necessarily along,axial surfaces. In this case, the pairedaxial surface corresponding to S� is locatedoff the right side of the section, so the fold-ing vector is measured at an arbitrarypoint in the direction parallel to axial sur-face S�.

The folding vector U is then used topredict the deflection of the fault (A)across the kink band (U = X). The pre-dicted fault position is consistent withreflection terminations that appear torepresent fault cutoffs. Moreover, thefolded fault dips about 30°, roughly par-allel to beds in the overlying kink band(T-T�).


50

Recognizing break-back thrustingThis section describes fault and fold patterns that are common in break-backward imbricate structures, and shows examples in seismic sections.

Patterns of fault cutting older fold limbsTo describe structural patterns common in break-back imbricate structures, we will consider somesimple patterns for a shallow, break-backward thrust ramp (model A1) and detachment (model A2)cutting across a fold limb (S-S’) related to an older and deeper thrust. The shallow thrust ramp maycut across and offset a part of the fold limb without changing fault orientation (model B1).Alternatively, the shallow thrust could change its orientation across the fold limb, offsetting andrefolding parts of the structure (model C1). In the case of model C1, note that the deep folding vec-tor (U) need not equal the deflection of the break-backward thrust (X), in contrast to the break-for-ward example described on the previous page. In the case of the detachment, the shallow faultcould follow bedding planes across the fold limb (model B2). Based on fault-bend fold theory(Suppe, 1983), slip on this shallow detachment would not modify the fold shape. Alternatively, theshallow detachment could follow bedding across the fold limb but cut up section beyond the fold(model C2). In this case the shallow fault conforms to one axial surface and offsets the other.

Patterns of break-backward thrusting in seismic dataThese seismic sections show patterns that reflect thrusting sequence. In section A,axial surface S terminates upward into a thrust that is overlain by gently dipping stra-ta. This pattern is comparable to that shown in model B1 (at left) and reflects break-back thrusting. In sections B and C, axial surfaces S’ are offset by shallow thrust faults.These patterns are comparable to model C2 (at left) and are consistent with break-backward or coeval, but not break-forward, thrusting.

A: Permian Basin, Texas, U.S.A.

B: Peruvian Andes

C: La Puna, Argentina

Patterns in models B1 and C1 are generally diagnostic of break-backward imbricate thrusting.However, patterns in models B2 and C2 are more ambiguous. A detachment that conforms to bed-ding across a fold, as in model B2, can be either a break-backward fault that followed beddingplanes or a folded detachment. Similarly, the pattern shown in model C2 reflects break-backwardthrusting only if the offset axial surface is considered active (i.e., it is pinned to a bend or tip of theunderlying thrust). In contrast, if the offset axial surface is inactive (S’), then the pattern may reflecteither break-backward thrusting or coeval motions on the deep and shallow faults. Thus, some pat-terns are diagnostic of thrusting sequence while others are not. Care should always be taken ininterpreting thrusting sequence based on fault and fold shapes.


51

Determining thrusting sequence using growth strataGrowth strata can be used to determine the thrusting sequence in cases where two or more growth structures can be related to separate faults. Associating growth structures with specific faults canbe difficult in cases where thrust sheets are everywhere vertically superimposed, but it is straightforward where faults are separated, at least in part, horizontally. This section presents seismic pro-files with examples of break-forward and break-back thrust systems interpreted using growth strata.

These seismic sections both image two faults (X and Y) that are sepa-rated horizontally at shallow levels, but vertically overlap one anotherat depth. In section A, the fold associated with fault Y does not deform,and thus pre-dates, the annotated horizon. The fold related to fault Xclearly deforms, and thus post-dates this horizon, reflecting a break-forward thrusting sequence. In section B, the fold associated with faultX does not deform, and thus pre-dates, the annotated horizon. The foldrelated to fault Y clearly deforms, and thus post-dates, this horizon,reflecting a break-backward thrusting sequence. Both seismic imagesare from the deepwater Niger Delta, Nigeria.

A: Break-forward thrusting

B: Break-back thrusting


52

1B-6: Structural wedgesBasic conceptStructural wedges contain two connected fault segments that bound a triangular, or wedge-shaped fault block. The two fault segments, which typically include two ramps or one ramp andone detachment, merge at the tip of the wedge. Slip on both faults accommodates propagationof the wedge tip and causes folding (Medwedeff, 1989). Wedges occur at a variety of scales. Atlarge scales associated with mountain fronts, wedges are typically referred to as triangle zones(Gordy et al., 1975). In this section, we describe common types of wedges and illustrate howthese structures are interpreted in seismic sections.

Conjugate faulting theory Kinematic Model

(above left) Brittle failure of rocks in compression commonly leads to the development of two conjugatethrust faults that dip in opposite directions (Anderson, 1942). Planes of weakness, such as bedding, canalso lead to the development of detachments. In cross section (above right), two conjugate thrustsbound a wedge-shaped fault block and merge at the wedge tip (model 0). Slip on both bounding faultscausing propagation of the wedge (model 1). In this case, the wedge propagates along a detachment, andcauses folding of the hanging wall block. The lower thrust is commonly referred to as the forethrust orsole thrust, and the upper thrust is called the back thrust or roof thrust (Boyer and Elliot, 1982).

Common characteristicsWedges exhibit a wide range of geometries.However, several characteristics are commonto most wedge structure, including:

1) presence of coeval fore- and back-thrusts;

2) folding localized along an active axial sur-face pinned to the wedge tip; and

3) folds may exist in the footwall of the backthrust that produce structural relief.

Examples

This seismic section images a large structural wedge, or triangle zone, at the eastern front of theCanadian Rocky Mountain fold and thrust belt. The common characteristics of structural wedges,(1–3) as described at left, are present in this structure. Note that a second, smaller back thrust ispresent within the main wedge block.

When the back or roof thrust and its hanging wall are gently tilted or warped, but not deformed tothe extent exhibited within the wedge block, the term passive roof thrust is sometimes used. Passiveroof thrusts are common in triangle zones, as shown in this example.

Seismic Example: Alberta Foothills, Canada

Field Example

Structural wedge in Carboniferous Rundle Formation, Front Ranges of theCanadian Rockies. Note the highly deformed rocks near the wedge tip.Several smaller wedges are contained within the larger wedge structure.(J. H. Shaw and F. Bilotti)


53

Wedge models developed using fault-bend fold theoryStructural wedges exhibit a variety of shapes and styles that reflect initial fault geometries, propagation direction, and folding mechanisms. In this section, we present a series of kinematic models thatdescribe basic types of structural wedges governed by fault-bend fold theory (Suppe, 1983; Medwedeff, 1989; see section 1B-1). Models A through C involve detachments, whereas model D does not.

A (0–2): Simple wedge with a detachment andback thrust. Propagation of the wedge tip formsa kink band above the back thrust that is bound-ed by an active axial surface, which is pinned tothe wedge tip. Strata in the kink band are paral-lel to the back thrust (β = 0) because the faultrises from a detachment (θ = 0).

B (0–2): Wedge with a lower forethrust rampand an upper detachment that acts as the backthrust. With slip, the wedge tip propagatesalong the detachment surface. Strata withinthe wedge are folded in an anticlinal fault-bendfold that deforms the detachment or backthrust. A kink band develops above the backthrust with strata that are parallel to theunderlying fault and fault-bend fold. The syn-clinal axial surface pinned to the wedge tip isactive, as is the anticlinal axial surface withinthe wedge block. The anticlinal axial surfaceabove the back thrust, however, is inactive.

C (0–2): Wedge formed by a dipping forethrustand back thrust. With slip, the wedge tip propa-gates along a detachment surface. Strata withinthe wedge are folded in an anticlinal fault-bendfold that deforms the back thrust. A kink banddevelops above the back thrust with strata thatare parallel to the underlying fault, but that dipmore steeply than the beds within the wedgeblock. Both the synclinal axial surface pinned tothe wedge tip and the anticlinal axial surfacepinned to the fault bends are active. The anti-clinal axial surface in the hanging wall of theback thrust is active (in contrast to model B)because a small amount of strata is folded fromthe crest into limb, thus passing through theaxial surface. These kinematics facilitate theconservation of bed length. Alternatively, asmall amount of shear or bed-parallel extensioncould accommodate fault slip without movingstrata from the fold crest into the limb.

D (0–2): Wedge formed by a dipping forethrustand back thrust. With slip, the wedge tip prop-agates along the trajectory of the forethrust.Strata within the wedge are not folded, as theydo not pass over a fault bend. A kink banddevelops above the back thrust with stratathat dip more steeply than the fault. The geom-etry of the kink band (θ) is governed by fault-bend fold theory (see section 1B-1), with �equal to the acute angle between the backthrust and the propagation direction, and β asthe hanging wall cutoff angle relative to thepropagation direction.

Note that in this wedge the roof thrust locallycuts down the stratigraphic section as itextends upward. This is an unusual relation-ship for thrust faults, but nevertheless mayoccur in non-decollement wedges.

Note: green dashed lines are active axial surfaces,red dashed lines are inactive axial surfaces. Seesection 1B-1 for description.


54

Seismic examples of structural wedgesHere we present examples of structural wedges imaged in seismic reflection data.

C: Dashen structure, Sichuan basin, China

Section A images a simple structural wedge that involves a backthrust extending upward from a forethrust ramp. The wedge tippropagation direction is along the path of the forethrust. Note thatin this case, the back thrust has very little displacement relative tothe forethrust. Section B images a wedge comprised of a gentlydipping back thrust that extends from a forethrust ramp. Thewedge propagation direction is along the path of the forethrust,which corresponds with an angular unconformity. Folding at thewedge tip is consistent with the pattern displayed in model D onthe previous page. Section C images a complex structural wedgecomprised of a back thrust extending upward from a foldeddetachment similar to model A on the previous page. The detach-ment level is constrained by the discordance of strata and thebase of the thrust ramp located east of the wedge tip. The wedgestructure, including the detachment, overlies an anticline that isrelated to a deeper level of faulting.

A: Niger Delta, NigeriaB: Santa Barbara basin, CA, U.S.A.


55

Growth structure in wedgesThis section describes growth structures above wedges that are modeled using fault-bend fold theory, afterMedwedeff (1989). Growth structures can be very helpful in distinguishing structural wedges from othertypes of fault-related folds.

Kinematic models Seismic Example: Sumatra, Indonesia

Interpreted section

In wedges that are governed by fault-bend fold theory (see section 1B-1), folds grow by kink-band migration.Folding generally occurs along an active axial surface that is pinned to the propagating wedge tip. In caseswhere sedimentation rate exceeds uplift rate, syntectonic strata form growth triangles above the wedge tipthat are bounded by a planar synclinal (active) axial surface and a curved anticlinal (inactive) axial surface(model W1). In contrast, simple forelimb fault-bend folds have growth triangles bound by a curved synclinal(inactive) axial surface and a planar anticlinal (active) axial surface (model F1). In cases where uplift rateexceeds sedimentation rate, the contrast between wedges and simple fault-bend folds is even more distinct.In a structural wedge, growth strata are folded about an active synclinal axial surface and are parallel to theunderlying forelimb dip (model W2). In contrast, syntectonic strata are not folded above the forelimb of asimple fault-bend fold (model F2), because they have not passed through an active axial surface. Growth stra-ta, therefore, are horizontal, or maintain a primary sedimentary dip, and onlap the forelimb.

(right) This seismic section images a structure with characteristics of a growth wedge. The structure con-sists of a forelimb developed above a south-dipping forethrust. Growth strata thin onto the crest of the struc-ture, and are folded above the forelimb. The synclinal axial surface is roughly planar and folds the growthstrata. In contrast, the anticlinal axial surface is curved, with an abrupt change in orientation at the contactbetween pre-growth and growth strata. Based on this growth pattern, which is similar to model W1 above,the structure is interpreted as a wedge. (For more details on this interpretation, see Shaw and Brennan, sec-tion 2-23, this volume).

Wedges Forelimb fault-bend folds


56

Shear fault-bend fold wedgesStructural wedges can form with non-flexural-slip components of deforma-tion, resulting in fold geometries that differ from those presented on theprevious pages. Here, we describe a class of these wedges that form byshear fault-bend folding (Suppe et al., 2004; see section 1B-4), and show anexample in a seismic section.

The seismic section shown aboveimages two thrust ramps rising from adetachment. The ramp on the left dipsin the same direction as the majorityof faults in the region, and thus is con-sidered a forethrust. The ramp on theright is a back thrust. Slip on the backthrust produces a hanging wall struc-ture that has the characteristics of ashear fault-bend fold. However, giventhat this is a back-thrust above adetachment, the structure is a shearwedge. Based on the fault cut-off angle(θ) and back-limb dip (δb), the struc-ture is interpreted as a pure-shearwedge in the section shown at right.Based on shear fault-bend fold theory(Suppe et al., 2004, see section 1B-4),the fault cutoff angle and backlimb dipyield a 27° dip of the synclinal axialsurface (ψ) in the basal layer and ashear angle (α) of 67°. Growth strataexhibit a fanning of limb dips that isconsistent with the shear wedge inter-pretation.

Simple-shear wedges (model A) have shear in the footwall of the backthrust. This shear folds, and induces slip, on the fault, producing a forelimbthat is similar to the back-limb fold produced by the forward-thrust, simple-shear fault-bend fold equivalent (see section 1B-4). In this model, growthstrata are eroded above the fold crest. Pure-shear wedges (model B) haveshear in the hanging wall of the back thrust that occurs as the wedge tippropagates. The back thrust is not folded, and slip produces a forelimb thatis similar to the back-limb fold produced by the forward-thrust, pure-shearfault-bend fold equivalent (see section 1B-4). In both shear wedges, the fore-limb beds dip less than the underlying back thrust, and the growth struc-tures record folding by a combination of limb rotation and kink-band migra-tion (see section 1A-5). In contrast, classical fault-bend fold wedges (modelC) generally have hanging wall beds that are parallel to the back thrust, andgrowth structures that record folding dominantly by kink-band migration.

Kinematic models

Seismic example: Niger Delta, Nigeria

Interpreted section


57

1B-7: Interference structuresBasic conceptInterference structures form when two or more monoclinal kink bands intersect, often yielding distinctive pat-terns in cross section with anticlines perched above synclines. Interference structures have been documented inthe field and laboratory (e.g., Dewey, 1965; Paterson and Weiss, 1966; Stewart and Alvarez, 1991), and have beenproposed as the origin of structures imaged in seismic profiles (e.g., Mount, 1989; Novoa et al., 1998; Camerlo etal., section 2-24, this volume). In this section, we describe a simple style of interference structure comprised oftwo kink bands with opposing dips, and present examples of these structures imaged in seismic sections.

Seismic Example: Gulf of Mexico, U.S.A.

This seismic section images an interference structure from the Perdido fold and thrustbelt (after Mount, 1989; Novoa et al., 1998). The structure is comprised of two mono-clinal kink bands that intersect at about 5.2 seconds (TWTT). The interfering kinkbands produce an anticline that is perched above a syncline, similar to the modelsshown at left. The sense of shear in the interference structure appears to be counter-clockwise, similar to model B. This section is displayed in TWTT, with a V.E. of about1:1 for a velocity of 2000 m/s, which is representative of the shallow section.

These models (A and B) illustrate interference structures formed by the intersection of two kink bands (1 and 2) that dipin opposite directions. Model A forms by clockwise shear of the through-going kink band (2), whereas model B forms bycounter-clockwise shear of the through-going kink band (1). In both models the through-going kink band separates the otherkink band into two pieces that are joined along two shear surfaces that are parallel to bedding. As a result, the shear sur-faces connect points where the axial surfaces bifurcate. The axial surfaces in these models bisect the interlimb angles (seesection 1A-1), and thus bed length and thickness are preserved. The most distinctive aspect of these structures is that theyyield anticlines perched above synclines.

Kink-band interference can result from many different structural configurations, involving various types of fault-relatedfolds (Mount, 1989; Medwedeff and Suppe, 1997; Novoa et al., 1998). These three models (C–E) illustrate general struc-tural configurations that can yield kink-band interference. The interfering kink bands are developed: C) above two bendsin the same fault; D) by imbrication of two faults; and E) as forelimbs developed above faults that dip in opposite direc-tions. Note that the shallow fold geometries are identical in each of these models. Thus, the geometries of interferencefolds are not always diagnostic of the underlying fault configurations. The different structural configurations do, howev-er, involve different patterns of active (green) and inactive (red) axial surfaces, which may, in some cases, be distin-guished using growth structures (Novoa et al., 1998; see section 1A-3).

Kinematic Models


58

Complex interference structureInterference structures that are faulted and/or involve more than two kink bands may have very complex geometries. In this section, we describe a complex, faulted interference structure imagedin a seismic section. We use a partial restoration of the structure to document its origins as an interference fold.

C: Geologic section

D: Partially restored section

The seismic profile shown in panel A images acomplex fold from the Sichuan basin, China. Thestructure exhibits the basic pattern of an anticlineperched over a syncline that is characteristic ofinterference structures. The structures differ fromthe simple models shown on the previous page,however, in that the core of the fold is cut by athrust. A narrow monocline appears to be offsetby this fault.

In panel B, the section is interpreted with a simpleinterference fold below the main thrust. Folds inthe hanging wall of the thrust are interpreted tobe displaced elements of the interference foldthat, in part, are refolded by a steepening upwardsplay of the fault. Panel C shows the same inter-pretation of the structure without the seismicimage. Restoration of slip on the main fault andthe associated folding in panel D yields a simpleinterference structure.

This example is intended to illustrate that inter-ference structure may have complex geometries.Nevertheless, these structures can generally beinterpreted using a combination of fault-relatedfolding theories. This interpretation invokes thebasic patterns of interference folding with the kinkmethod (section 1A-1) and fault-bend folding (sec-tion 1B-1) to describe the hanging wall structure.The hanging wall portion of the offset monoclineis refolded using the concept of folding vectorsdescribed in section 1B-5.

Interference structures also generally exhibit verydistinct patterns in map view and three dimen-sions. For a description of these patterns, seeNovoa et al. (1998) and Camerlo et al. (section 2-24, this volume).

A: Uninterpreted section

B: Interpreted section

Part 2: Case Studies Seismic Interpretation of Contractional Fault-Related Folds

59

2CaseStudies

Shaw, Connors, and Suppe Part 2: Case Studies

60

2-1: Pitas Point Anticline, California, U.S.A.John H. Shaw, Department of Earth and Planetary Sciences, Harvard University, Cambridge, Massachusetts, U.S.A.Stephen C. Hook, Texaco Inc., Houston, Texas, U.S.A.John Suppe, Department of Geological and Geophysical Sciences, Princeton University, Princeton, New Jersey, U.S.A.

Location: Eastern Santa Barbara Channel, California, U.S.A.Topics: Fault-bend folding, growth structure, map patternsReserves: Gas in Pliocene clastic reservoirs

Figure 2: Post-stack, time-migrated, and depth converted 3-D seismic reflection profile across the Pitas Point trend, with formation tops and dipmeter from the Texaco 234 #7 well. Downwardterminating kink bands (2) indicate a detachment at about 5 km depth (see section 1A-2, Recognizing thrust and reverse faults). Shallow gas sag is documented by Mastoris (1990).

Figure 1: Map of fold trends in the eastern Santa Barbara Channel, California, showing locationsof the Pitas Point trend and seismic profile shown in Figure 2. RT = Rincon trend; ORT = offshoreOak Ridge trend; ORF = Oak Ridge fault; MCT = Mid-Channel (Blue Bottle) trend.

The Pitas Point anticline is located in an active fold and thrustbelt in the eastern Santa Barbara Channel, California (Figure 1)(Namson and Davis, 1988; Shaw and Suppe, 1994). The fold hasa flat crest separating gently north- and south-dipping limbsthat are bounded by the Rincon and offshore Oak Ridge trends,respectively (Figure 2). Both fold limbs terminate downwardsat about 5 km depth, suggesting the presence of a detachmentin the Miocene Monterey Formation. Upper Pliocene andQuaternary strata thin onto the crest of the anticline, suggest-ing that these are growth or syntectonic units. Moving upwardin section, the crest of the fold narrows and migrates to thenorth. Thus, shallow strata penetrated by the Texaco 234 #7well dip gently to the south, whereas, deeper units are hori-zontal or dip gently to the north. In the following discussion, wepresent an interpretation of this structure as a growth fault-bend fold that is compatible with these basic observations.

Pitas Point anticline


61

Figure 4: Post-stack, time-migrated and depth converted 3-D seismic reflection profile across the Pitas Point trend, with formation tops and dipmeter from the Texaco 234 #7 well. Downward termi-nating kink bands, that are highlighted in Figure 2, indicate a detachment at about 5 km depth (see section 1A-2, Recognizing thrust and reverse faults). Labeled axial surfaces correspond to thosemodeled in Figure 3.

Fault-bend folds in the crestal broadening stage exhibit dipping over horizontal strata in the growth section on the fold crest(Figure 3). This pattern is observed in the seismic image of the Pitas Point anticline and in the dipmeter of the Texaco 234 #7well. In Figure 4, we interpret the anticline as a simple fault-bend fold developed above a north-dipping (13°N) thrust ramp thatconnects detachments in the Miocene Monterey Formation. Based on fault-bend fold theory, where θ = φ = 13°, the forelimbshould dip 14°S (β = 14°) and be slightly narrower than the backlimb (R=.95). These values were used to guide the interpreta-tion, which generally conforms to reflection geometries. The fold is slightly modified by slip on the shallow Montalvo thrust,which is described by Shaw et al. (1996).

2-1: Seismic interpretationWe propose a simple growth fault-bend fold model (Figure 3)and interpretation (Figure 4) to describe the geometry andkinematic evolution of the Pitas Point anticline.

Figure 3: Sequential models of a growth fault-bend fold (Suppe et al., 1992). Model 1 containstwo fold limbs developed above a ramp between decollements. The fold is in the crestal upliftstage of growth (Shaw et al., 1994), as fault slip is less than ramp width. In Model 2, additionalslip widens kink bands, which narrow upward in the growth section (Suppe et al., 1992). In Model3, fault slip is greater than ramp width. Thus, strata are refolded from the back limb (A-A��) ontothe crest of the structure, which now widens with fault slip (crestal broadening stage, Shaw et al.,1994). Growth strata are also folded above the crest, generating a pattern of dipping over horizon-tal beds and offsetting the shallow crest from the deep crest of the fold. These patterns, as well asthe fault cutoffs, are observed in the 3-D seismic data from the center of the Pitas Point anticline(Figure 2). Model 4 includes minor displacement on a shallow detachment producing subtle fold-ing that is similar to patterns observed above the Pitas Point anticline in Figure 4.

Pitas Point anticline

1

2

3

4

growth

pre-growth

dipping overhorizontal strata

shallow detachment

slip on shallow detachment





62

Figure 8: Enlarged (2X) portion of the axial surface map superimposed on a time slice (2.3s TWTT) from the 3-D seismic survey. Note that in the zone of crestal broadeningthe trend of axial surfaces B and B� are parallel to the seismic reflections in the forelimb. The wide fold crest is imaged as a broad negative (white) amplitude surrounding plat-form Habitat.

Conclusions:• The Pitas Point anticline is a south-vergent, fault-bend fold developed above a thrust ramp and detachment

within the Miocene Monterey Formation. Maximum slip on the fault is about 3.5 km. • Upper Pliocene and Quaternary strata are syntectonic units folded by displacement on the thrust.• The fold is in the crestal broadening stage of growth in the center of the trend beneath platform Habitat.

2-1: Map-view analysisTo describe the three-dimensional geometry of the Pitas Point anti-cline, we present a structure contour and axial surface map at thetop of the Pliocene Repetto Formation. The axial surface map is gen-erated using the vertical projection method of Shaw et al. (1994), asdescribed in Figure 5. We interpret that the fold along section X-X�(Figure 4) is in the crestal broadening stage of growth. Folds in thecrestal broadening stage have a distinct axial surface map pattern(Figure 6). This pattern is observed in the axial surface map of thePitas Point trend (Figure 7), and in a time-slice from the 3-D seismicsurvey (Figure 8).

Figure 5: Perspective view of a plunging fault-bend fold. (top): Between sections 3 and 2, fault slip is greater thanthe ramp width and the fold is in the crestal broadening stage. As slip decreases to the right of section 2, the foldenters the crestal uplift stage. Fold plunge is denoted by converging pairs of axial surfaces. (bottom): Axial surfacesare mapped by projecting their intersections with the mapped horizon vertically to a horizontal datum.

Figure 6: The axial surface map pattern of a doubly plunging fault-bend fold is characterized by pairs of axialsurfaces that converge at the fold terminations. The zone where the fold is in the crestal broadening stage isdefined by the deflection of the forelimb kink band (B-B�) away from the backlimb kink band (A-A�), yielding awider fold crest.

Figure 7: Axial surface map at the top of the Pliocene Repetto Formation, superimposed on a structure contour map of the same horizon that was generated independentlyfrom well control. The plunge of the fold is reflected by pairs of axial surfaces (A-A�and B-B�) that converge toward the fold terminations. In the center of the trend, the forelimbaxial surfaces (B-B�) are deflected southward. This pattern is consistent with the crestal uplift stage of growth (Figure 6) in the center of the trend and along section X-X�. For amore detailed discussion of the map pattern, see Shaw et al. (1994).


63

2-2: Toldado Anticline, Upper Magdalena, ColombiaAlexis Rosero, HOCOL S.A., Bogota, ColombiaJuan Carlos Ramon, HOCOL S.A., Bogota, Colombia

Location: Upper Magdalena basin, Tolima, ColombiaTopics: Fault-bend folding, growth strataReserves: Oil in the Albian Caballos sandstones

The Toldado anticline is part of the buried NNE trendingOrtega fold and thrust belt (Figure 1). This belt is locatedbelow and to the west of the Avechucos syncline. TheToldado anticline is a NNE-SSW-trending anticline with awide, low-angle crest separating gently east- and west-dip-ping limbs (Figure 2). The structure is interpreted as a fault-bend fold. Two models are geometrically possible based onthe fold shape (in Cretaceous strata) and partially con-strained fault geometries (Figures 3 and 4). The geometry ofgrowth strata is used to constrain the degree of fold evolu-tion and to distinguish the structural interpretation. Thenear-horizontal growth strata (Paleocene) across the foldcrest indicates that the fold is on the Crestal Uplift Stage (seeShaw et al., section 2-1, this volume). Paleocene growth stra-ta gets thinner along the crest of the fold. This is partly dueto variable uplift over the fold and partly due to erosion onthe Eocene unconformity. These data indicate that sedimen-tation rate during the Paleocene was close to, or slightlyhigher than, the uplifting rate.

Figure 1: Location map of the Toldado anticline, showing the main structural features of thestudy area. Note that the Toldado anticline is located close to the trace of the Avechucos syn-cline. Location of the seismic line on Figure 2 is shown.

Figure 2: Post-stack, time-migrated, 2-D seismic reflection profile across the Toldado anticline. The line is in TWT but is displayed in 1:1 scale using the velocity function of the Toldado-3 well.Toldado-3 well and formation tops are shown. Note thinning of Paleocene growth strata (1) across the fold crest. Minor erosion occurs along the crest of the fold associated with the Eocene uncon-formity. The forelimb downward termination (2) defines an intra-Villeta detachment.

Avechucos Syncline / Toldado oilfield


64

2-2: Toldado AnticlineIn this interpretation (#1) we consider that the upper detachment is parallel to strata in its foot-wall and the top of the ramp is located at the base of axial surface A�. We observe a forelimb cut-off angle (β) of 32°. However, before this value is used to calculate the ramp dip, the forelimb mustbe “unfolded” across the 8° bend in the upper detachment (φ). Using fault-bend fold theory, weobtain a 35° dip of the forelimb before it was folded across the bend in the detachment. This valueserves as the forelimb cut-off angle (β0 = 35°) that, along with the “unfolded” inter limb angle ( γ0

= 74°) is used to calculate the change in fault dip (φ = 23°) an initial cut-off angle of the ramp (θ =26°). This yields a 30° dipping ramp.

The resulting interpretation implies that the fault-bend fold is in the crestal broadening stage. Themodels presented by Shaw et al. (section 2-1, this volume) show that at this stage there shouldbe dipping growing-strata on top of horizontal crestal beds (see Figure 3, insert). The seismicdoes not support this geometry and thus this model is discarded.

Figure 3: Interpreted seismic section assuming a fault-bend fold inthe crestal broadening stage (Interpretation #1).


dipping overhorizontal strata


65

2-2: Toldado AnticlineIn this interpretation (#2) we use the observed forelimb hangingwall cutoff angle, (β = 32°), the dip of the upper detachment (13°SE) and the forelimb interlimb angle (γ = 79.5°) to predict thechange in fault dip (φ), the initial cutoff angle (θ), and the slipratio (R). Using the anticlinal fault bend fold graphs (see section1B-1) we derive a φ value of about 16°. With this angle, we obtainan initial cutoff angle of about 27°. The slip ratio R is calculatedas 0.85. This agrees with the slip ratio (S1/S0) measured on theseismic section.

Note that in this interpretation the shortening and slip is small-er than in interpretation #1, and that the axial surface (A�) is notfixed to the top of the ramp (where the ramp meets the upperdetachment). This implies that this fold is on the crestal upliftstage (Suppe, 1983; Shaw et al., 1999). In this case, the horizon-tal growth strata seen on the seismic above the fold crest agreeswith the crestal uplift model (see insert), making this secondinterpretation more plausible than the previous one.

In conclusion, based on fault and pre-growth fold geometries,two structural models are possible for the Toldado anticline.Growth strata are used to distinguish between these alternativemodels, and support our interpretation of the Toldado anticlineas a fault-bend fold in the crestal uplift stage of growth.

Figure 4: Interpreted seismic section assuming a fault-bend foldin the crestal uplift stage (Interpretation # 2).



66

2-3: Sequatchie Anticline, Tennessee, U.S.A.: A small displacement fault-bend fold

Shankar Mitra, University of Oklahoma,Norman, Oklahoma U.S.A.

Location: Appalachian Plateau, Tennessee, U.S.A.Topics: Fault-bend fold, multiple-bend ramp

Figure 2: Part of time-migrated seismic profile through the Sequatchie anticline, Tennessee.

The Sequatchie anticline (Figures 1 and 2) is the frontal structure of the Southern Appalachian thrust beltin southern Tennessee, Georgia, and Alabama. An interpretation of the northern part of the structure inCumberland and Rhea counties, Tennessee, is presented based on surface data, a seismic profile, and datafrom the ARCO-Ladd #1 well. The structure has a low relief and exposes Mississippian to Pennsylvanianunits on the crest of the structure (Figure 1). Farther south, the relief increases, and Middle Ordovician toDevonian units are exposed at the surface (Hardeman, 1966; Harris and Milici, 1977).

Figure 1: Generalized geological map of the Sequatchie anticline in Cumberland and Rhea Counties,Tennessee (modified from Hardeman, 1966), showing the location of the seismic profile shown in Figures 2and 3. Omu-S = Middle to Upper Ordovician and Silurian. D-M = Devonian to Mississippian. Pg-Pco =Pennsylvanian Gizzard Group and Crab Orchard Mountains Group. Pl = Lower Pennsylvanian units abovethe Crab Orchard Mountains Group.


67

2-3: Sequatchie Anticline

Figure 3: Uninterpreted (a) and interpreted (b) time-migrated sections through the Sequatchie anticline. Active axial surfaces areshown in green, and passive axial surfaces in orange. The seismic section and the interpretation do not correlate 1:1 with the structuralcross section because of crooked line effects.

Seismic interpretation is based on both a time-section (Figure2), and a post-stack depth migration (Figure 3), based on avelocity model constructed from a preliminary depth model.The structure is related to a thrust fault that originates at thebase of the Cambrian Rome Formation, and climbs to the baseof the Pennsylvanian Gizzard Group. The seismic data, and theARCO Ladd #1 well indicate that the fault has a low dip (approx-imately 5°) in the Cambrian Rome and Conasauga Formations,but has a much steeper dip (15°) in the Cambro-OrdovicianKnox Group and the remaining Ordovician to Devonian units.The steeper fault dip in the Knox Formation is probably relatedto the higher competence of this unit. Surface data suggest thatthe front limb of the structure has a very steep dip, ranging from30 to 85°.

The Sequatchie anticline is interpreted to be a low-displace-ment fault-bend fold, related to a multi-bend ramp (Figure 4a).There are four bends in the fault, each of which defines anactive axial plane. Movement of the hanging wall over the faultbends results in the development of a series of passive axial sur-faces, which originate at the active axial surfaces and migrateaway from them. The active and and passive axial surfaces sep-arate panels of relatively constant dip, which can be identifiedfrom surface and seismic data, and from the dipmeter data inthe ARCO-Ladd #1 well.

The seismic data show a low westward dip of the basementbetween shot points 390 and 485, and a steeper eastward dipbetween shot points 390 and 325. This geometry is partly due toa seismic pull-up under high velocity carbonates in the hangingwall of the Sequatchie thrust, which was apparently uncorrect-ed in the depth model used for the post-stack depth migration.However, there appears to be a very low-dipping eastward rampunder the Sequatchie thrust, which drops all units down to theeast. The presence of the ramp is also indicated by dipmeterdata in the footwall. This ramp may have formed along a zone ofweakness in the basement during loading associated with theemplacement of Valley and Ridge thrusts. The formation of theSequatchie thrust fault may have been influenced by the loca-tion of this ramp.

DATUM = 1100 feet

DATUM = 1100 feet


68

The cross section presented below (Figure 4a) is based on theseismic interpretation and was restored using line-length bal-ancing (Figure 4b). The restoration shows that the shorteningfor the base of the Rome Formation is approximately 5100 ft.The fault displacement decreases from 5100 ft at the base of the

Rome Formation to 4500 ft at the top of the Knox Group and2200 ft at the top of the Mississippian units. The forward shearof the loose line in the restored section suggests a small amountof differential penetrative strain at the mesoscopic and microscopic scales within the Silurian to Mississippian units.

This inclined shear profile and proposed penetrative deforma-tion is consistent with the steep front limb of the fold, and thesmall fault displacement in the Mississippian units.

Conclusions• The Sequatchie anticline is a fault-bend fold related to a multi-

bend fault ramp connecting major detachments in theCambrian Rome Formation and the Pennsylvanian GizzardGroup.

• Macroscopic shortening associated with the formation of thestructure is approximately 5100 ft.

• The front limb of the structure is fairly steep, suggesting pene-trative deformation within the Silurian to Mississippian units,which have been transported onto the upper detachment.

2-3: Sequatchie Anticline

Figure 4: Structural cross section through the Sequatchie anticline, Tennessee, based on seismic data (Figure 2), surface data, and data from the ARCO-Ladd #1Jewett Heirs well. Active axial surfaces are shown in green and passive axial surfaces in orange. b. Line-length restoration of the structural cross section in a.

ARCO-LADD #1JEWETT HEIRS


69

2-4: El Furrial Field cross section, Eastern Basin, VenezuelaEnrique Novoa, Departamento de Analisis Exploratorio Integrado, Gerencia de Exploracion y Produccion, PDVSA-INTEVEP, VenezuelaLászló Benkovics, Departamento de Analisis Exploratorio Integrado, Gerencia de Exploracion y Produccion, PDVSA-INTEVEP, VenezuelaClaudia Fintina, Departamento de Analisis Exploratorio Integrado, Gerencia de Exploracion y Produccion, PDVSA-INTEVEP, VenezuelaJavier De Mena, Departamento de Delineacion y Caracterizacion de Yacimientos, Gerencia de Exploracion y Produccion, PDVSA-INTEVEP, Venezuela

Location: Eastern Venezuela Basin, VenezuelaTopics: Fault-bend fold, growth strataReserves: 2.0 billion barrels

El Furrial Trend is located in the deformation front of the Serranía delInterior fold belt to the south of the Pirital Fault, Eastern Venezuela Basin(Figure 1). It is divided into three giant oil fields: El Furrial, Carito, andSanta Barbara fields. The boundaries between these fields are tear faultsand/or lateral ramps. This structural trend contains actual recoverablereserves of about 2.0 billion barrels of medium gravity oil. A balancedcross section through El Furrial field is presented. The structure is asym-metrical, with the backlimb much wider than the forelimb. The backlimbincludes two inclined dip domains while the front limb is composed ofone domain (Figures 2 and 3). We interpret the fold as developing abovea two-bend thrust fault that accommodates about 14 km of shortening(Figure 4). Growth strata suggest that the fold started growing during theearly-to-middle Miocene. We present a kinematic model that shows howthis structure may have developed.

Figure 2: A depth-migrated 3-D seismic reflection profile that images El Furrial structure. Notice that the image deteriorates between X� and X��. The blueticks show the top of Oligocene picks in the wells. Profile provided by PDVSA E&P.

Figure 1: El Furrial trend (B) is located in the Eastern Basin of Venezuela (A).

El Furrial Oil Field


70

2-4: El Furrial Field cross sectionA balanced cross section across El Furrial field is shown in Figure3. The structure is characterized by a very narrow forelimb andwide backlimb. The backlimb is made of two inclined dip domains(A-C and C-C�). The kink band A-C dips 27° NW and C-C� dips 12°NW. The forelimb is composed of a dip domain (B-B�) that dips 20°SE. The kink bands A-C and C-C� are parallel to the fault plane. ElFurrial fold started to grow in the early-to-middle Miocene asshown by a growth triangle (Suppe et al., 1992) in the sequence ofthis age interpreted on top of the kink band B-B�. This observationis supported by geochemical data which show biodegradation ofoil (Talukdar et al., 1987). Presumably, this early oil was accumu-lated when the reservoir was at shallow depth during an earlystage of fold development. Normal faults have been interpreted inboth the hanging wall and footwall. These normal faults may berelated to the development of the foreland basin in front of thefold belt. A kinematic model (Figure 4) shows the evolution of theFurrial Fold through time.

Figure 3: A balanced, retrodeformable cross section (X-X��)across El Furrial Trend that integrates seismic reflection (Figure 2) and well data. El Furrial trend develops above a two-bendthrust fault, which causes a very wide backlimb and a narrow forelimb. The steeper portion of the thrust fault is short and the gentle part is very long. The backlimb is interpreted to becomposed of two inclined dip domains (A-C and C-C�) which are parallel to the El Furrial fault. On the other hand, the forelimb is composed of a single inclined dip domain (B-B�). Theseismic data illuminate the kink bands B-B� and C-C� very well, however the dip panels A-C and C�-A are not well defined by the data. Notice the growth axial surface (G�) on top of thekink band B-B�� which shows a growth triangle in the early-middle Miocene sequence. This structural trend accommodates around 14 km of total slip.

Conclusions:• El Furrial field is located within a very asymmetric fault-bend fold where the backlimb is much wider than the forelimb.• Growth strata and geochemical data suggest that it started to grow during early-middle Miocene times.• The fault plane has two bends and is divided into two sections: a narrow, steep ramp and a long, more gentle ramp. • The fault accommodates about 14 km of shortening.

Figure 4: A balanced, kinematic model of development of the El Furrial trend. a: Incipient fault andactive axial surfaces (A and B) in undeformed strata. b: Slip on the two-bend thrust fault generatesinactive axial surfaces A� and B�that are rigidly translated away from active axial surfaces A and B.Once axial surface A� arrives at the convex bend of the fault, an incipient active surface (C) is generat-ed. Moreover, axial surfaces A�, B, and B� become inactive and will be rigidly translated along theupper portion of the thrust fault. c: Additional slip on the fault causes the development of an inactiveaxial surface C�and the kink band C-C�starts to grow until the present geometry is reached. Similarkinematic models for multibend faults are shown by Medwedeff and Suppe (1997).

El Furrial Oil Field


71

2-5: Rosario Field, Maracaibo Basin, VenezuelaTed Apotria, ExxonMobil Development Company, Houston, Texas, U.S.A.M. Scott Wilkerson, Department of Geology and Geography, DePauw University, Greencastle, Indiana, U.S.A.

Location: Maracaibo Basin, VenezuelaTopics: Fault-related fold termination geometry and kinematicsReserves: ~50 MMB (Molina, 1992)

Overview: Fault-related fold terminations typically form due to loss of dis-placement on the genetically-related thrust fault, an along-strike change infault attitude, or both. Constraints on footwall cutoffs and along-strikedisplacement are needed to determine the termination mechanism, whichcan often be determined from reflection seismic data. Fold geometry froma single profile does not uniquely establish kinematics.

The Rosario structure is a contractional fault-related fold located in thewestern Maracaibo Basin, Venezuela (Figure 1). The plunging southern ter-mination is constrained by industry reflection seismic and well data, and isinterpreted to be due to an along-strike decrease in displacement. The faultgeometry changes from a flat-ramp-flat at the crest of the structure wheredisplacement is greatest, to simply a ramp near the lateral fault tip. Theseobservations suggest a kinematic model in which the structure initiated asa modified fault-propagation fold with an isolated fault ramp within the“stiff” layer. With increased shortening, the fault grew to link with upperand lower detachments in the weaker shale units resulting in a hybridizedfault-bend fold. The geometric elements of a single profile at the crest areconsistent with the Suppe (1983) fault-bend fold model. However, inter-pretation of the structure in 3-D suggest different kinematics.

Note: A full presentation of the seismic data and interpretation is inApotria and Wilkerson (2002). A .mov-format animation of the 3-D struc-tural model of Rosario Field can be downloaded at the AAPG Datashareweb page (http://www.aapg.org/datashare/) and is on the CD-ROM accom-panying this book.

Figure 1: Principal structural features of the western Maracaibo Basin, Venezuela. The Rosariooil field (white box) was discovered in 1954, with production from fractured Cretaceous carbon-ates and Eocene fluvial clastic reservoirs (Molina, 1992). The La Luna Quarry is highlighted inred.

Generalized stratigraphy of the western Maracaibo Basin issummarized in Figure 2. Interpretation of the Rosario structureis constrained by 2-D time-migrated, 1985- and 1990-vintageseismic lines (Figure 3). Interpretations from these lines wereconverted to depth (e.g., Figure 4) using interval velocities cal-culated from seismic well ties from the CR-12 well (Figure 3C).Our discussion will focus on seismic lines CCT-90c-14 and CAT-85-1 (Figure 3A–C), which cross the crest of Rosario, where the

primary geometric elements are best imaged. Two high-impedance and continuous reflections mark the top and bottomof the carbonate section (Figure 3A, B). The first reflectionoccurs between the Colon Shale and the top of the carbonates,and marks the mechanical transition between the “stiff” unitbelow and the “weak” clastic unit above. A second strongimpedance contrast occurs at the base of the carbonate sectionand the top of the underlying Rio Negro clastic section. These

reflections bound a total carbonate section that is about 547 mthick in the CR-12 well (Figure 3C). The Tertiary section consistsof alternating sands, silts, and shales and exhibits parallel fold-ing. The Cretaceous Colon Shale normally has a thickness ofabout 550 m, except when structurally thickened where faultsemerge from the underlying Cretaceous carbonate package(Figure 3).

Figure 2: Generalized stratigraphy of the western Maracaibo Basin. Formation tops indepth (meters below a 33m KB) and interval velocities used for depth conversion post-ed from the CR-12 well. Qualitative mechanical stratigraphy and the location of inferreddetachments is also depicted.


72

2-5: Rosario Field

The Rosario Fault exhibits a flat-ramp-flat geometry at the crest(Figure 3A–C). La Luna and portions of Rio Negro strata are trun-cated and displaced onto a footwall flat near the base of the ColonShale. Near the crest, the forelimb dips are slightly greater thanthose on the backlimb, defining a weak asymmetry toward the fore-land (Figures 3, 4). The exact positions of the La Luna and Rio Negrofootwall cutoffs are not well-imaged and are obscured by velocitypull-up beneath the hanging-wall carbonates (dashed line in foot-wall in Figure 3). However, we estimate that apparent displacementonto the upper flat is a maximum of 2.4 km in the plane of this sec-tion (Figure 3B), or approximately 2.0 km if projected into the trans-port plane (see transport direction in Figure 4E).

Deformation interpreted at Rosario occurred during a middle-Miocene and younger Andean Orogeny (Roure et al., 1997). Basedon present-day topographic relief, the structure remained activeinto the Pleistocene and recent.

Figure 3: 2-D time-migrated seismic lines over the Rosario structure (see Figure 4 for location). Seismic lines are about 1:1 in the vicinity of the Cretaceous section.(A) uninterpreted line CCT-90c-14. Tie-line locations are labeled in gray. (B) interpreted line CCT-90c-14. Formation tops are labeled on the right; the approximatemiddle-Miocene surface is dashed. RF = Rosario Fault.; REF = Rosario East Fault. (C) line CAT-85-1. (continued on the next page).


73

Other regional seismic data in the vicinity ofRosario suggest that shortening of the Terti-ary section is not transferred to the foreland.Instead, we interpret a wedge structure thattransfers displacement back towards the hin-terland (Figure 3A–D), similar to structuresseen in nearby outcrops at La Luna quarry(Figure 5).

South of line CCT-90c-14 (Figure 3A, B), sig-nificant differences in the geometry of theRosario structure exist relative to the crest.Over a distance of 4 km, the fold loses a well-developed backlimb and a hanging-wall rampon footwall flat (compare Figure 3D, E). TheRosario Fault also changes from a flat-ramp-flat geometry to simply a ramp (Figure 4E). Alower flat may accommodate the shorteningobserved near the ramp, but there is nodirect evidence for it based on fold shape.Apparent offset of the La Luna Formationdecreases from about 2 km at the crest(Figure 3A, B), to 1.5 km (Figure 3C), to 1 km(Figure 3D), to about 100 m near the termina-tion (Figure 3E). This loss of displacement isprimarily accommodated by transfer to theRosario East Fault, which gains displacementto the south. This is evident from the top LaLuna structure map as two discrete en eche-lon anticlines separated by the Rosario Fault(two dashed lines in Figure 4A, B). Where theColon Shale dampens the displacementtransfer between the two faults, the relaybetween the two folds becomes less evidentin the shallower Tertiary section, and is onlyreflected as a subtle change in fold axis trendabove the transfer zone (single dashed line inFigure 4C, D). In addition, dip magnitude nearthe crest is less on the Colon Shale reflectorcompared to the top La Luna (Figure 4).

Figure 3 (continued): (D) CAT-85-2. (E) CAT-85-3. (F) CAT-85-4. (G) CAT-85-5. The top of La Luna (blue) is mapped in Figure 4A. The top of Colon Shale (green) is mapped in Figure 4C. SeeFigure 4 for location.

2-5: Rosario Field


74

2-5: Rosario Field

Figure 4: (A) Top La Luna Formation sub-sea depth structure (blue reflector in Figure 3). Solid red lines represent the position of seismic lines published in this section of this volume. Other lines are published in Apotria and Wilkerson (2002). Dashed lines mark fold hinges for Rosario andRosario East. Contour depths range from -4925 m (blue) to -4025 m (red) with a contour interval of 50 m. (B) Top La Luna dip magnitude map with superposed structure contours. Dip is a maximum of 22° (red) and a minimum of 0° (blue). (C) Top Colon Shale sub-sea depth structure map (greenreflector in Figure 3). Dashed line represents a single fold axis for both Rosario and Rosario East. Contour interval is 50 m. (D) Top Colon Shale dip magnitude map with a maximum of 18° (red) and a minimum of 0° (blue) with superposed structure contours. (E) Sub-sea depth structure-contourmap of the Rosario Fault. Contour interval is 100 m. Red arrow indicates the assumed regional transport direction perpendicular to the Perija Mountain Front (Figure 1). Dashed lines represent boundaries between the ramp and the two flats. The flats die out to the south, with only a ramp near thetermination. The Rosario Fault also changes attitude toward the north, defining an oblique ramp. The oblique ramp is associated with fold closure to the north, but does not appear to directly influence the fold termination to the south.

The present-day geometry at the crest of the Rosario structure has the essential characteristicsof a fault-bend fold (Figure 3A–C). However, the lateral variation in fold and fault geometry sug-gests that a flat-ramp-flat is not present near the termination, and may have been absent duringthe structure’s early development. Eisenstadt and DePaor (1987) proposed a 2-D model for faultgrowth in which a fault ramp initially nucleates in the “stiff” layer with associated tip strains

accommodated by folding. As shortening accrues, the ramp grows up and down section, eventu-ally linking with upper and lower stratigraphically-controlled flats. In the kinematic model that fol-lows, we extend Eisenstadt and DePaor’s (1987) 2-D model to 3-D, and assert that spatial variationin geometry is also a proxy for temporal variation.


75

2-5: Rosario Field

Assuming that the spatial variation in the fault-fold geometryalso represents the temporal variation of the fold’s develop-ment, we suggest the following kinematic model for the Rosariostructure. Each interpreted fold growth stage is consistent withobserved geometry from tip to crest.

Stage 1 (pre-middle-Miocene, Figure 6A). Cretaceous andyounger strata are essentially undeformed. Regional uplift anderosion occurred within the western Maracaibo Basin in theEocene, but no local fold developed at Rosario during this time.

Stage 2 (middle-Miocene, Figure 6B). Shortening of the sectioninitiates with small reverse offset near the top of the Cretaceouscarbonates. Folding of the Colon Shale occurs with fault propa-gation through the “stiff” carbonate interval. Sub-seismic scaledeformation of the Tertiary section is manifested by layer-par-allel shortening. No evidence for a backlimb is observed. Thisline resembles Figure 3F and 3G observed near the present-daysouthern termination.

Stage 3 (Figure 6C). The fault ramp links with an upper detach-ment in the Colon Shale with potential structural thickeningnear the upper flat (analogous to Figure 5). Increased displace-

ment places the hanging-wall ramp onto the upper flat, and theforelimb begins to steepen. This is supported by the decrease inforelimb dip toward the present-day termination (Figure 4B, D).

Stage 4 (Figure 6D). As displacement accrues, the Rosario Faultcontinues to propagate downward and eventually connects tothe basal flat within the Rio Negro or La Quinta Formations. Thisproduces a discrete backlimb-lower flat transition that isobserved at the present-day fold crest (e.g., Figure 3A–D). Whenboth upper and lower flats are operative, additional fault dis-placement is accommodated by fault-bend folding.

Figure 5: Outcrop analog from the La Luna quarry (see Figure 1 for location). The lower massive unit is the Maraca Member of theCogollo Group carbonates, which is overlain by the thin-bedded, La Luna Formation. The deformation style that occurs at outcrop scalewhere the fault emerges from the Maraca is similar to that seen at seismic scale where faults emerge from the “stiff” carbonate sectioninto the “weak” Colon Shale (Figure 3). This style of deformation at the tip of an emerging thrust fault could account for some of theapparent thickening in the Colon Shale seen on seismic sections near the ramp upper-flat transition.

Figure 6: Model for the 3-D development of the Rosario structure. Each profile represents stages in both the temporal and spatial evolution of thestructure from (A) earliest/least displacement to (D) latest/most displacement. See explanation in text below.

Kinematic Model of the Rosario Structure


76

2-5: Rosario FieldThe 3-D interpretation of the Rosario structure highlights the importance of distinguishing fault-bend fold geometryfrom fault-bend fold kinematics. At the crest, the Rosario structure exhibits characteristics of a fault-bend fold (e.g.,a lower and upper detachment, intervening ramp, and a hanging-wall ramp on footwall flat geometry). However, thediagnostic geometric elements of a single profile do not uniquely establish the kinematic development. The Suppe(1983) fault-bend fold model assumes flexural slip as the deformation mechanism, and results in passive foldingabove a pre-existing fault. The prescribed kinematic model results in fold geometry that is a function of the under-lying fault geometry, and a forelimb dip that remains constant with slip. Given the assumptions (Suppe, 1983), onecan predict the orientation of one element (e.g., the ramp dip) given two other elements (e.g., forelimb dip and theaxial angles).

We measured the same geometric elements near the crest of the Rosario structure on seismic line CAT-85-1 (Figure7) and compared them to theory (Suppe 1983). The forelimb dip (β) is 22°, consistent with the dip map on the top ofthe La Luna (Figure 4B). The axial angle (γ) is more difficult to determine due to smooth, parallel folding, but our esti-mate is 80°. Using these measurements, the Suppe (1983) model predicts a ramp step-up angle (θ) of 17°, which isour observation on CAT-85-1, if backlimb dip is used as a proxy for the ramp dip. The natural example matches theSuppe (1983) model prediction of the geometry of a single profile. However, our observations of the structure in 3-Dsuggest the kinematics of the Rosario structure are different than in the Suppe (1983) model.

Our observations are consistent with the model in Figure 6, in which the Rosario structure evolved from a “fault-propagation fold” into a “fault-bend fold” (geometric rather than kinematic description). We also note that the fore-limb dip decreases along strike (Figure 4B), further supporting a hybridized model. Although the Rosario crest haspresent-day geometry consistent with a simple fault-bend fold, the kinematics are more complicated than a single 2-D profile would suggest. Preservation of growth strata, poorly defined in this study area, would be of further use toconstrain the kinematic development of the fold.

Conclusions• The southern termination of the Rosario structure likely formed due to an along-strike decrease in displacement.

Key elements of the interpretation include: a) the fault geometry changes from flat-ramp-flat at the crest to a faultramp near the southern tip of the structure, b) forelimb dip decreases toward the southern termination, and c) thebacklimb is indistinct toward the southern termination.

• These observations suggest kinematics in which the structure initially developed as a simple fault ramp in the“stiff” layer (fault-propagation fold stage) and later propagated to connect with upper and lower detachments(fault-bend fold stage). Our model is a 3-D extension of a 2-D model proposed by Eisenstadt and DePaor (1987) inwhich fault ramps nucleate in “stiff” units.

• Rosario provides a natural example of a structure where spatial differences may reflect temporal stages in the evo-lution of a fault-related fold. The model departs from previous models of rigid self-similarity and permits variationsin fold style and deformation mechanisms influenced by mechanical stratigraphy.

AcknowledgmentsPermission to reproduce the seismic data was provided by Elsevier Press and The Journal of Structural Geology. Wealso thank ExxonMobil Exploration Company, ExxonMobil Upstream Research Company, and PDVSA (Venezuela) forpermission to publish.

Figure 7: (A) Seismic line CAT-85-1 (Figure 3C) with the addition of interpreted dip domainboundaries (red). (B) Angular measures of fold geometry where γ = axial angle, β = forelimb dip, θ = ramp dip. The geometric elements of this single profile are consistent with the Suppe (1983)fault-bend fold model. However, based on interpretation of the structure in 3-D, the inferred kine-matic development is different. Instead, the structure develops from a fault-propagation fold (activefold above a buried fault tip) into a fault-bend fold (passive fold above an existing fault) as slipincreases (Figure 6).


77

2-6: Medina Anticline, Eastern Cordillera, ColombiaMark G. Rowan, Rowan Consulting, Inc., Boulder, Colorado, U.S.A.Roberto Linares, Ecopetrol, Instituto Colombiano del Petroleo, Piedecuesta, Santander, Colombia

Location: Llanos foothills, Eastern Cordillera, ColombiaTopics: Fault-bend fold, axial-trace map, fold-evolution matricesReserves: Giant fields along trend (e.g., Cusiana, Cupiagua)

Figure 1: Map showing the location of the Medina anticline along the eastern border ofthe Eastern Cordillera of Colombia, between the Quetame basement massif and the frontalAguaclara fault. Insert shows the location of the larger map in the northwestern corner ofSouth America.

Figure 2: Uninterpreted 2-D time-migrated seismic profile across the Medina Anticline, the adjacent Rio Amarillo Syncline, and the frontal Aguaclara Fault (location shown on Figure 4). The foldgeometry, with symmetrical limbs, a horizontal crestal domain, and sharp hinges separating planar dip domains, suggests a fault-bend fold origin. Seismic data courtesy of Ecopetrol.

The Medina Anticline is located in the Llanos Foothills province along the border of the Eastern Cordillera, Colombia, approxi-mately 100 km southwest of the giant fields of Cusiana and Cupiagua (Figure 1). It is interpreted as a simple fault-bend foldbecause of its symmetrical shape, kink-band geometry, and horizontal crestal domain (Figure 2). Shallow structural levels are wellimaged, but the deep geometry and the detachment level and trajectory of the underlying fault are unknown. In order to addressthese issues, we use a grid of time-migrated, 2-D seismic data to generate an axial-trace map (Shaw et al., 1994) of the anticline.We then generate fold-evolution matrices and models, which illustrate the effects of two independent variables on fold geometry(Rowan and Linares, 2000), to determine the factors controlling the three-dimensional geometry of the fold. This allows us toidentify active and inactive axial planes, construct the three-dimensional fault geometry, and complete the structural interpreta-tion to depth. Axial-surface analysis shows that the three-dimensional geometry of the Medina Anticline is compatible with a fault-bend fold interpretation in which displacement increases to the northeast, the ramp dip decreases to the southwest, and thelength of an intermediate flat increases to the northeast.


78

2-6: Medina AnticlineAxial-trace maps display structure contours andthe plan-view traces of fold hinges (Wilkerson etal., 1991; Shaw et al., 1994; Shaw and Suppe, 1994).Models of fault-bend folds in which displacementincreases along strike produce a pattern of activeand inactive axial traces and a three-dimensionalgeometry in which the crestal domain narrowsand then widens (Figure 3). Axial traces interpret-ed on individual seismic profiles across theMedina Anticline (Figure 4) were projected verti-cally and connected to create the axial-trace mapfor the fold (Figure 5). Although this map is broad-ly similar to the model pattern (Figure 3), thereare also significant complications. Furthermore, itis possible to generate comparable patterns byvarying parameters other than displacement(Rowan and Linares, 2000). Thus, further analysisis needed to identify active and inactive axialtraces, understand the critical variables, and com-plete the interpretation.

Figure 4: Partial interpretation of the line shown in Figure 2. The red dashed line is the top of the upper Eocene Mirador sandstone (the main reservoir in the area),which is constrained by nearby well control in both the hanging wall and footwall (see Figure 5). The steep grey lines are axial traces along the fold hinges separatingplanar dip domains; the offset of axial traces in the Medina Anticline is caused by a minor detachment at the base of the Oligocene to lower Miocene Carbonera shales.

Figure 3: Perspective view and axial-trace map of fault-bend fold inwhich displacement increases along strike (after Shaw et al., 1994). Activeand inactive axial traces are green and red, respectively.

Figure 5: Time-structure map of the top of the Mirador Formation (contour interval is 400 msec; depth is relativeto arbitrary datum near surface). Contours are in black (tick marks point downdip), seismic lines are in grey, faultsare in red, wells are in orange, and erosional truncation is shown by the thick dashed line. The blue lines are theaxial traces at this structural level, and the arrows point in the dip direction. The Medina Anticline is bounded bythe broad Nazareth Syncline to the northwest and the tight Rio Amarillo Syncline to the southeast. The crestaldomain is most narrow at the fold culmination and plunges to the southwest and then south to where it intersectsthe Aguaclara Fault where it curves west. Similarly, the backlimb curves and becomes less steep toward thesouthwest and south.


79

2-6: Medina AnticlineFold-evolution matrices show the profile geometry of folds with two varying parameters and areused to construct model axial-trace maps and corresponding perspective views (Rowan andLinares, 2000). Figures 6, 7, and 8 show the fold-evolution matrix, axial-trace maps, and perspec-tive views, respectively, for a model in which both displacement and ramp dip vary linearly.Changes in ramp dip result in curved structure contours (Figures 7c, d and 8c, d) rather than thestraight contours produced by varying displacement (Figures 7a, b and 8a, b). However, this dif-ference is distinctive only for the linear gradients used; for example, a nonlinear displacement gra-dient would result in curved axial traces. A more reliable criterion for distinguishing betweenchanging displacement and changing ramp dip is limb dip. As the ramp dip decreases, both fore-limb and backlimb dips decrease (Figures 7c, d and 8c, d). Thus, parallel structure contours(Figures 7a, b and 8a, b) indicate only changing displacement, whereas divergent structure con-tours (Figures 7c, d, e, f and 8c, d, e, f) show that the ramp dip is varying along strike.

Figure 6: Fold evolution matrix for linear increase in displacement and linear decrease in ramp dip (28.5, 22.5, 16.5, and 10.5degrees). (a) through (e) indicate six different combinations of four profile geometries used to construct the corresponding axial-trace maps (Figure 7) and perspective views (Figure 8).

Figure 7: Axial-trace maps of the six panel combinations indicated in Figure 6, in which displacement and/or ramp dip vary linearly alongstrike. Black lines are structure contours, and dip symbols show the orientation of dip domains.

Figure 8: Perspective views corresponding to the axial-trace maps of Figure 7, in which displacement and/or ramp dip vary linearly alongstrike.


80

2-6: Medina AnticlineThe presence of the Rio Amarillo Syncline between the Medina Anticline and the immediate hang-ing wall of the Aguaclara Fault suggests that there is a flat between the Medina ramp and theAguaclara ramp. Such a flat could change length along strike, so we illustrate a fold-evolutionmatrix, axial-trace maps, and perspective views (Figures 9, 10, and 11, respectively) in which bothdisplacement and flat length vary linearly. The resulting axial-trace patterns and three-dimen-sional geometries are complicated by the presence and interference of new axial surfaces (C, C�,D, D�) associated with the syncline and frontal ramp. The effects are best seen in the third columnof Figure 9, where displacement increases over a flat of fixed length. B� migrates toward the footof the upper ramp, intersecting with C to form D. When B� reaches the upper ramp, it becomesfixed and C now migrates up the ramp. In the meantime, B is also migrating forward; when it reach-es the foot of the upper ramp, B and B� are eliminated and replaced by a new set of A and A� axialtraces associated with the frontal ramp. Farther up the ramp are C�, D�, and an offset B� (e.g., topof second column).

Figure 9: Fold evolution matrix for linear increase in displacement and linear increase in flat length. (a) through (e) indicate sixdifferent combinations of four profile geometries used to construct the corresponding axial-trace maps (Figure 10) and perspec-tive views (Figure 11). Red numbers (1–6) indicate geometries used to construct the model axial-trace map in Figure 12; 3 isintermediate between the middle two profiles in the top row, and 4 is intermediate between the top two profiles in the second col-umn.

Figure 10: Axial-trace maps of the six panel combinations indicated in Figure 9, in which displacement and/or flat length vary linearly alongstrike. Thin black lines are structure contours, thick black lines are faults, and dip symbols show the orientation of dip domains.

Figure 11: Perspective views corresponding to the axial-trace maps of Figure 10, in which displacement and/or flat length vary linearlyalong strike.


81

To model the Medina Anticline, selected profiles from Figure 9 are joinedin map view to create the synthetic axial-trace map in Figure 12a. Thismodel accurately depicts most of the features of the Medina Anticline,so that axial traces can now be identified (Figure 12b). However, themodel has a narrow backlimb and horizontal crest to the southwest(Figure 12a), whereas the observed geometry shows a widening back-limb and dipping, curving crestal domain (Figure 12b). We infer that thisis caused by a southwestern decrease in ramp dip, as modeled in Figures7c and 8c.

The model-constrained map interpretation is then used to complete theinterpretation. Where the crestal domain is narrowest (location 1, Figure12b), axial traces A� and B should intersect at the top of the lower ramp(Shaw et al., 1994). The fault geometry on each profile is then deter-mined as shown and explained in Figure 13.

Figure 12: (a) Synthetic axial-plane map constructed using geometries 1–6 inFigure 9; and (b) corresponding interpretation of the Medina Anticline. In (a),the six profile geometries from Figure 9 were spaced equally and rotated tomatch the orientation and approximate the scale of the Medina Anticline. Thus,from southwest to northeast: flat length first increases as displacement is heldconstant (1–3), flat length then decreases as displacement increases (3, 4), flatlength again increases as displacement is held constant (4, 5), and then both flatlength and displacement increase (5, 6). The dashed line in (a) is the approxi-mate location of the Aguaclara Fault, and the number 1 in (b) indicates the nar-rowest point of the crestal domain where axial traces A� and B switch betweenactive and inactive.

2-6: Medina Anticline

Figure 13: Finished interpretation of the line shown in Figure 2. The axial-trace analysis and comparison of the observed map to the model map (Figure 12) allows the axial tracesto be identified. The level of the flat is determined by the intersection of A� and B where the crestal domain is narrowest (location 1 on Figure 12b) and then correlated along strike.The length of the flat is determined by the intersection of active axial traces B and C at the top of the lower ramp and base of the upper ramp, respectively. Note that the length of thehanging-wall flat (B�-C�) approximately balances that of the footwall flat (B-C).


82

2-6: Medina AnticlineThe fault geometry constructed on each profile is shown in map view inFigure 14. Also shown are deeper normal faults visible in the southwest;their continuation to the northeast is not imaged but is likely, and we spec-ulate that the Medina fault-bend fold formed where the Aguaclara Faultramped up over an underlying basement normal fault (Figure 15). Both thethick-skinned uplift of the Quetame basement massif and the thin-skinneddevelopment of the Medina Anticline are interpreted to have formed duringTertiary inversion of a Jurassic rift basin (Figure 15) (Rowan and Linares,2000; see also Cooper et al., 1995).

Figure 14: Map of the fault geometry underlying the Medina Anticline as constructed using the axial-surface analy-sis. Most of the fault consists of a lower ramp, an intermediate flat that widens to the northeast, and an upper ramp.To the southwest, the fault curves westward, forming an oblique ramp and thus a lower ramp angle. This is apparentlyin response to curving traces of deeper, rift-related normal faults that offset prerift basement (blue). Contour interval is400 msec; depth is relative to an arbitrary datum near surface.

Conclusions:• The Medina Anticline is a fault-bend fold, probably formed as the Aguaclara Fault ramped up over a

basement normal fault during Tertiary inversion of a Jurassic rift basin.• The three-dimensional geometry is controlled by: (1) An increase in displacement to the northeast;

(2) An increase in flat length to the northeast; and (3) A decrease in ramp dip to the southwest.• Axial-surface analysis is a useful tool for constraining subsurface geometry where axial traces are easily

defined, but must be used in conjunction with other data/techniques to avoid model-driven interpreta-tions.

Figure 15: (a) Regional 1:1 cross section through the culmination of the Median Anticline showing its relationship to the inverted Quetame basement massif; and (b)Schematic reconstruction (not to scale) showing the infilled rift geometry at the end of the Cretaceous. Aguaclara fault is shown as an out-of-the-syncline thrust, but itcould also be rooted in basement. Tan = prerift basement; blue = Jurassic synrift; green = Lower Cretaceous; orange = Upper Cretaceous; yellow = Tertiary/Quaternary.


83

2-7: Three-dimensional interpretation of the El Furrial Trend,Eastern Venezuela Basin, Venezuela

Miguel Morales, PDVSA, EPM, Venezuela.Enrique Hung, PDVSA, EPM, Venezuela.Richard Bischke, Subsurface Consultants & Associates, LLC.,

Houston, Texas, U.S.A.

Location: Eastern Venezuela Basin, VenezuelaTopic: Fault-bend fold foldingReserves: 11,100 MMBO (in place)

The super giant El Furrial field is a fault-related anticline located in theEastern Venezuela Basin. The field was discovered in 1986, based ongood quality 2-D seismic data. The El Furrial trend continues to the westas a series of fault-related structures that make up the Serrania delInterior southeast-verging fold and thrust belt (Aymard et al., 1990). TheEl Furrial trend defines the frontal edge of this fold and thrust belt. Thestratigraphy of the fold belt consists of a 3- to 5-km thick Cretaceous toPaleogene passive margin section and a 0- to 8-km thick sequence ofNeogene to recent syntectonic and post compressional foredeep filldeposits. The large Pirital fault overthrusts the westernmost structures,repeating 5000 m (16,000 ft) of Cretaceous section. Fault surface mappingand structural interpretations indicate that the major faults and theirassociated anticlines form a linked en echelon system related to dextraltranspression south of the El Pilar right lateral strike-slip fault system.

Figure 1: Simplified regional map showing the main structural elements of the Serrania del Interior fold andthrust belt. See Figure 2 for a depth-corrected profile.

Figure 2: Depth-corrected profile.

El Furrial is the easternmost of three major structures (El Furrial, Carito, and Tejero, from east to west) that form theNorth Monagas fields in the Serrania del Interior fold and thrust belt (Figure 1). The trend has 11 billion bbl of oil inplace and presently produces 400,000 bbl/day. The structures trend northeast-southwest across northeasternVenezuela, and are offset in a dextral en echelon relationship to each other (Figure 3). These offsets are caused bynorthwest trending lateral ramps in the underlying major thrust faults (Bischke et al., 1997). These structures are theresult of mid to late Miocene dextral transpressional displacements south of the El Pilar strike-slip fault (Figure 1).

In the northern part of the South American Plate, the transpressional displacements produced a series of north-northwest-trending dextral tear faults and lateral ramps that turn to the east-northeast to become ramp and flatthrust faults (Figure 2). Maps constructed of the fault surfaces indicate that many of the faults interconnect to forma linked fault system (Boyer and Elliott, 1982). Figures 4 and 5 describe these general relationships. In Figure 4, theTejero ramp branches off the Urica Fault trend, and the offset Carito ramp creates another lateral ramp that trendssubparallel to the Urica Trend. In turn, the Carito ramp is offset from the Furrial ramp by another lateral ramp.


84

The three ramp systems link to each other and to the Urica Fault trend along a lower fault flat (Figure 4).Displacement of the hanging wall over the three ramps creates the three offset folds, which form the structuraltraps for the fields. Based on interpretation of high quality 3-D seismic data (samples of which are shown on thefollowing pages), we interpret that these structures developed mainly as fault-bend folds (Suppe, 1983, 1985).

The El Furrial trend is overlain by the Pirital thrust, one of the largest faults in the regional system. The Piritalthrust dips to the north over a horizontal distance of 20 km (Figure 6, sp 100 to 600 between 3 to 10 s). In the west,the Pirital branches off the northwest-southeast trending Urica lateral ramp system, forming the western flank ofthe Serrania del Interior fold and thrust belt (Figure 1) (Bischke et al., 1997). The Pirital fault overthrusts theOligocene Naricual reservoir unit, repeating about 500 m (16,000 ft) of the Cretaceous San Juan Formation (Figure6). The main reservoir unit in the area is the Oligocene Naricual Formation, which contains fluvial deltaic to shal-low marine sands (Prieto et al., 1990). The Naricual sands are approximately 500 m (1700 ft) thick (Figure 6), andcan contain 250 m (800 ft) of net pay. This northeastward prograding sequence of sands is contemporaneous withthe trailing shelf margin of the South American Plate. Later overthrusting loaded and down warped the plate form-ing a foredeep basin and most likely an outer rise, similar to the outer rise and gravity high observed seaward ofoceanic trenches (Watts and Talwani, 1974). Seaward of the trenches normal faults tend to occur on theupwarped highs, which extend due to flexure. The Naricual Formation contains many normal faults that may haveoriginated in a similar fashion when the overthrust sheets of the Serrania del Interior advanced toward the south,loading and flexing the South American Plate.

2-7: 3-D El Furrial Trend

Figure 3: Simplified depth map of the El Furrial trend showing offset fields.

Figure 4: Block diagram illustrating a linked ramp-flat and lateral ramp system.

Figure 5: Block diagram showing the hanging wall above the El Furrial, Carito, and Tejero ramps forming threeoffset fault-bend folds.

Figure 6: Regional cross section and stratigraphic column showing the main tectonostratigraphic elements from the Caribbean plate to the Orinoco tar belt (modified fromPDVSA Report).


85

Seismic OverviewHere we present seismic profiles from a 3-D survey thatdefine the geometry of the El Furrial trend. The profiles aremigrated and displayed in time (Figures 7, 8) and depth(Figures 9, 10). Figure 7 is an uninterpreted dip section (A)that trends northwest-southeast across the structure. On dipline A, near the trace of strike section B, there is a panel ofreflections that dips to the southeast (between 3 and 4.3 s)that defines the forelimb of the El Furrial structure. Thissouth dipping panel overlies a prominent horizontal reflec-tor at about 4.3 s. We interpret that this horizontal reflectoris a fault-plane reflection originating from the upper flat(detachment) of a ramp-flat system. The fault is located atthe downward termination or discontinuities in the dip panel(Dahlstrom, 1969; Tearpock and Bischke, 2002; see section1A-2, this volume). The horizontal reflector can be followedto the north where it joins a group of north-dipping reflec-tions. We interpret these north-dipping reflections to repre-sent the backlimb of a fold, which is thrust to the southeastabove the frontal ramp (Suppe, 1983, 1985).

Figure 8 is a time section along the strike of the El Furrialstructure. Note that at about 4.3 s a near-horizontal reflectorextends across the strike profile. This feature correspondswith the fault-plane reflection described in dip section A.Above the fault, a panel of reflections that dips to the eastrepresents the folded hanging wall of the El Furrial structure.The nearly horizontal reflections below the thrust corre-spond to the relatively undeformed footwall. Figure 8 imagesa strike or lateral ramp (Tearpock and Bischke, 2002), ormore precisely the inverted portion of the lateral ramp thatis thrust up a frontal ramp and onto an upper flat.

2-7: 3-D El Furrial Trend

Figure 7: Section A — Seismic time profile images dip panels forming a south-verging anticlinal fault-bend fold. Intersection with section B is shown in black line.Arrows highlight fault position.

Figure 8: Section B — Strike seismic profile in time showing dip panels formed above the main detachment surface.


86

2-7: 3-D El Furrial TrendStructural InterpretationIn this section, we present interpreted depth profiles(Figures 9 and 10). The high impedance package above 5 kmrepresents the Oligocene Naricual sands (orange horizon),and the horizontal reflection just above 6 km is the upperflat of the thrust fault. In section A (Figure 10), the fold con-tains a flat crest separating a narrow, southeast-dippingforelimb and a wide, northwest-dipping back limb. Basedon the fold and fault geometry, the structure appears to bea fault-bend fold (Suppe, 1983, 1985; Novoa et al., section 2-4, this volume).

The El Furrial thrust fault repeats a minimum of 2.0 km ofsection and suggests at least 50% shortening (≈ 4 km)(Figure 10). We can define only a minimum estimate of slipon the El Furrial fault because the backlimb of the struc-ture extends beyond the three-dimensional data set.

Naricual equivalent sands produce to the south of thetrend. Our interpretation suggests that the Naricual sandsproject beneath the El Furrial structure at about the 7.0- to8.0-km level.

In summary, the El Furrial anticline is a well-defined exam-ple of fault-bend fold, similar to structures initiallydescribed in the Appalachian Mountains, U.S.A. (Rich,1934) and the Canadian Rockies (Bally et al., 1966). Thisfold style is common in other parts of South America (e.g.Dengo and Covey, 1993) and across Venezuela. Thus, fault-related folding techniques serve as powerful tools fordescribing many of the hydrocarbon-producing structuresin these regions.

Conclusions• The super giant El Furrial trend is formed by three off-

set fault-bend folds.• The folds are related to a linked dextral en echelon

ramp-flat and lateral ramp system.• Shortening is estimated at 50%.

Figure 9: Interpreted strike line B.

Figure 10: Area-balanced interpreted of dip line A.


87

2-8: Shear fault-bend fold, Deep-Water Niger DeltaFreddy Corredor, Department of Earth and Planetary Sciences, Harvard University, Cambridge, Massachusetts, U.S.A.John H. Shaw, Department of Earth and Planetary Sciences, Harvard University, Cambridge, Massachusetts, U.S.A.John Suppe, Department of Geosciences, Princeton University, Princeton, New Jersey, U.S.A.

Location: Niger Delta, West Africa, NigeriaTopics: Shear fault-bend folding, growth sedimentation

Figure 1: Uninterpreted, migrated, and depth-converted 2-D seismic profile through a fault-related fold in the deep-water Niger Delta. We observed three basic structural patterns that are consistent with pure shear fault-bend folding kinematics: First, a long planar backlimb that dips less than the fault rampwith increasing shallower dips of growth strata, second, a short forelimb compared to the backlimb, and third, a synclinal axial surface that does not bysect the syncline. Seismic data courtesy of MABON LTD.

Using fold shapes, fault plane reflections, and patterns of growthsedimentation, we model a fault-related fold in the deep-waterNiger Delta using shear fault-related folding theory. The NigerDelta offers a unique opportunity to study fault-related folds, asthe structures are well imaged at deep levels in seismic reflectionprofiles and because they preserve growth strata that record foldkinematics. Individual fault-related folds are characterized by long

planar backlimbs with increasingly shallower dips to growth stra-ta, suggesting a component of progressive limb rotation.Forelimbs are short compared to backlimbs, but growth stratashow more consistent dips that suggests a component of foldingby kink-band migration. Combined mechanisms of kink-bandmigration and limb rotation are thus invoked to model the kine-matics of this fault-realted fold.


88

2-8: Shear fault-bend fold, Niger DeltaThe Niger Delta is situated in the Gulf of Guinea (Figure 2) on the mar-gin of West Africa. Sourced by the Niger River, it is one of the largestregressive deltas in the world with an area of roughly 300,000 km2, asediment volume of 500,000 km3, and a sediment thickness of morethan 10 km in the basin depocenter. The northern delta boundary is theBenin flank, an east-trending hinge line south of the West Africa base-ment massif. Cretaceous outcrops on the Abakaliki Fold Belt define thenortheastern delta boundary. The offshore boundary of the delta isdefined by the Cameroon volcanic line to the east, the border of theDahomey basin to the west, and the 4000 m bathymetric contour. Fromthe Eocene to the present, the delta has prograded southwestward intothe Gulf of Guinea. The Niger Delta basin consists of Cretaceousthrough recent marine clastic strata that overlie oceanic and fragmentsof continental crust. The compressional fault-related fold structures inthe deep-water Niger Delta are the product of contraction due to grav-ity-driven extension on the shelf.

Figure 2: High-resolution shaded relief and seafloor bathymetry image of the Niger Delta showing the approxi-mate location of the seismic line used in this study (1).

Figure 3: 2-D seismic section through the fault-related fold interpreted in this contribution showing some important characteristics including (1) sea floor reflection, (2)top of oceanic crust reflector, and thrust fault plane seismic reflection indicated by red arrows. Notice how the backlimb dips much less than the fault ramp. See text fordetail of (3) and (4).

The stratigraphic sequences imaged in the seismic profile shown above (Figure 3) correspond to Tertiary deep-marine and deltaic sediments. At the bottom of this sequence, the Akata Formation, which can be observed abovethe Top of oceanic crust reflection (2), is up to 3000 m thick in this portion of the delta, and is composed of thickdeep marine shale sequences (potential source rocks), and may contain some interbedded turbidite sands (poten-tial reservoirs in deep water environments). On seismic sections, the Akata Formation is generally devoid of inter-nal reflections (3), and exhibits low P-wave velocities that produce a pull-down velocity effect in time sections, andmay indicate regional fluid overpressures. This Formation corresponds to the weak decollement layer that under-goes an externally imposed shear deformation in this fault-related fold. We use shear fault-bend fold kinematics (sec-tion 1B-4, this volume) to interpret this structure. Shear fault-bend folds are characterized by long planar backlimbsthat dip less, or much less than the fault ramp, as observed in Figure 3 (4), and shows increasingly shallower dipsto growth strata suggesting a component of folding by limb rotation. A fault plane reflection is clearly observed (redarrows) that constrains the fault geometry and its planar nature. The fault ramp dips at an angle of 26°. The long pla-nar backlimb dips at an angle of 7.5°, which is much less than the dip of the fault. Also notice that, unlike conven-tional fault-bend folds, the length of the backlimb does not represent the amount of slip along the fault, and that isrepresented by the distance between the green dots. This difference between the fault displacement and backlimblength is due to the combined limb rotation and kink-band migration folding mechanisms that occur in shear fault-bend folding kinematics. Two end-member interpretations are possible: Simple shear and pure shear fault-bend fold-ing. We will discuss the main structural and stratigraphic features to distinguish between these two end members.


89

2-8: Shear fault-bend fold, Niger DeltaThis fault-related fold can be modelled using the pure shear or simple shear fault-bend folding kinematics (section 1B-4,this volume), (Figures 4 and 5). In simple shear fault-bend folding a weak decollement layer of finite thickness (The AkataFormation) at the base of fault ramps undergoes an externally imposed bedding-parallel simple shear with no basal fault.In pure shear fault-bend folding the deformation of a weak decollement layer of finite thickness is locally confined to therock volume in the inmediate vicinity of the fault ramp where stresses are high. The basal decollement layer slides abovea basal fault and shortens and thickens above the ramp with no externally applied bed parallel simple shear. The slipalong the basal detachment decreases to zero at the bottom of the fault ramp. The total slip, then, is accomodated by slipalong the fault ramp, and by thickening of the weak decollement layer. In simple shear fault-bend fold kinematics the syn-clinal axial surface at the bottom of the fault ramp bysects the syncline, while in the pure shear fault-bend fold kinemat-ics this synclinal axial surface is not the angle bisector of the syncline.

Figure 4: Two kinematic models of simple and pure shear fault-bend folds constructed using the end member theory graphs of Figure 5. A) Simple kinematic model of a pure shear fault-bend foldshowing downward propagation of shear with the resulting patterns of growth strata, where the slip rates along the fault ramp are equal to the rates of growth sedimentation. The distance between thebottom of the growth axial surface and the synclinal axial surface at the top of the pre-growth sequence is equal to the maximun slip along the basal fault. B) Simple kinematic model of a simple shearfault-bend fold with patterns of growth strata, where the slip rates along the fault ramp are equal to the rates of growth sedimentation. The final geometry of the fault-related fold is the same in bothmodels. Pure shear fault-bend folding kinematics require a shallower detachment level compared to the calculated detachment using simple shear fault-bend folding.

Figure 5: A) Pure shear fault-bend folding end member theory graph (section 1B-4, this volume)showing the relationship between ramp dip, back dip, and dip of the syncline axial surface within theweak decollement layer. The yellow square in the graph corresponds to the fault-related fold inter-preted in this contribution. B) Simple shear fault-bend folding end member theory graph (section1B-4, this volume). The yellow squares in the graphs correspond to the fault-related fold interpretedusing the backlimb and cut-off angles interpreted in this section (2-8).

axial surfaces axial surfaces

Growth axial surface Growth axial surface


90

Figure 6: Simple shear fault-bend fold interpretation of the migrated 2-D seismic profile in the deep-water Niger Delta. A shear profile is included thatshows the deformation of a line originally perpendicular to bedding before deformation. This profile shows how the shear decreases upwards. The shear isconcentrated between the bottom of the fault ramp and the yellow horizon. An overall simple shear (αe) of 40° is interpreted in the lower 1000 m that ter-minates in the top of the kink-band (a-b), which agrees with the value predicted via theory from the back-limb dip (δb) of 7.5° for kink-band (a-b) and afault dip (θ) of 26°. A simple shear (αe) of 15° is interpreted in the next 500 m that terminates at the fault in kink-band b-c, which agrees well with ashear predicted via theory from the back-limb dip (δb) of 6° for kink-band (b-c) and a fault dip (θ) of 26°. Notice that fault slip decreases from a maximumat the top of the ramp to zero at the base of the ramp. Shallow growth strata over the backlimb suggests limb rotation. The synclinal axial surface in thiscase was interpreted at the point of maximum curvature between the synclinal dip domains. It bisects the syncline across the weak decollement layer. Alower detachment is interpreted at 6500 m depth where the synclinal axial surface intercepts the bottom of the fault ramp. Notice how the length of thebacklimb does not reflect the amount of slip along the fault as predicted by conventional fault-bend fold theory. The forelimb is interpreted using multi-bend fault-bend folding theory. The growth strata onlap the forelimb according to the theory when the rate of growth sedimentation is lower than the rateof structural growth. The gentle dips of the growth strata could be the result of differential compaction and drape.

2-8: Shear fault-bend fold, Niger Delta

Simple Shear Fault-bend Fold


91


Pure Shear Fault-bend FoldFigure 7: Pure shear fault-bend fold interpretation of the migrated 2-D seismic profile in the deep-water Niger Delta. A shear profile is included thatshows the deformation of a line originally perpendicular to bedding before deformation. This profile also shows how the shear decreases upwards. Theshear is concentrated between the bottom of the fault ramp and the yellow and green horizons. An overall pure shear (αe) of 60° is observed in the lower700 m that terminates in the top of the kink-band (a-b), which agrees well with the value predicted via theory from the back-limb dip (δb) of 7.5° for kink-band (a-b) and a fault dip (θ) of 26°. An additional pure shear is observed in the next 500 m that terminates at the fault in kink-band (b-c), which producesa back-limb dip (δb) of 6.0° for kink-band (b-c). Notice that fault slip goes to zero at the base of the ramp. A much higher detachment is interpreted in thiscase at 5700 m depth where the synclinal axial surface also intercepts the bottom of the fault ramp. Notice how the length of the backlimb does not reflectthe amount of slip along the fault as predicted by conventional fault-bend fold theory, and requires less slip than the simple shear case. The synclinal axialsurface in this case was interpreted at the location of maximum change in dip domain. It does not bisect the syncline across the weak decollement layer.The synclinal back angle (ψ) is 23.5°, which agrees well with the value predicted via theory for the observed back-limb dip and ramp angles, and the cal-culated shear angle.


92


Figure 8: Close-up view of the forelimb. The forelimb is interpreted using multibend fault-bend folding theory. The growth strata onlap the fore-limb as predicted by the model when slip rates are greater than growth sedimentation rates. The gentle dips are probably produced by differentialcompaction and drape.

Figure 9: Close-up view of the syncline showing the interpreted picks of the synclinal axial surface across different stratigaphic levels (greendots). The synclinal axial surface, in this interpretation, does not bisect the syncline. Thickening along the decollement layer (Akata Formation)can be observed above the fault ramp, on the left flank of the syncline. These two observations suggest a pure shear fault-bend fold.

Conclusions:• A pure shear fault-bend fold is described in the outer fold

belt of the deep-water Niger Delta where the weak decolle-ment layer corresponds to the deep marine Tertiary AkataFormation.

• The pure shear fault-bend fold described in this section (2-8) is characterized by a long planar backlimb with increas-ingly shallower dips to growth strata, a short forelimb com-pared to the backlimb with onlapping growth strata, and asynclinal axial surface that does not bisect the syncline due

to the thickening of the weak decollement layer across theaxial surface.

• The forelimb is interpreted using the multibend fault-bendfolding theory.

• The length of the backlimb does not reflect the amount ofslip along the fault ramp.

• The patterns of growth sedimentation suggest increasinglimb rotation by progressive increase of shear along thebacklimb.

• Rates of syntectonic growth sedimentation are lower than

rates of uplift along the fault ramp during initial stages offold growth producing onlap over the forelimb, and areincreased later such that no bathymetric relief develops.

• The main feature that allows differentiation between singleand pure shear fault-bend folds in seismic sections is thesynclinal axial surface. This axial surface is an angle bisec-tor in simple shear fault-bend folds, but not in pure shearfault-bend folds due to the thickening of the weak decolle-ment layer across the axial surface as illustrated in thestructure interpreted in this section (2-8).

2-9: Basil anticline, Northern Apennines, ItalyFabrizio Storti, Dipartimento di Scienze Geologiche, Università degli Studi “Roma Tre,” Roma, ItalyStefano Tavani, Dipartimento di Scienze Geologiche, Università degli Studi “Roma Tre,” Roma, ItalySaverio Merlini, ENI/Agip Division, San Donato Milanese, Milano, ItalyAlessandro Mosconi, ENI/Agip Division, San Donato Milanese, Milano, ItalyFrancesco Salvini, Dipartimento di Scienze Geologiche, Università degli Studi “Roma Tre,” Roma, Italy

Location: Northern Adriatic Sea, ItalyTopics: Fault-propagation folding, growth structure, foreland flexureReserves: Gas in Pliocene clastic reservoirs


93

The Basil anticline is located at the toe of the Apennines fold and thrust belt, in the northern Adriatic Sea (Figure 1). In the regional seismic line (Figures 2, 3) the pre-, syn-, and postorogenic sedi-mentary architectures are well imaged, as well as two major thrust-related structures and their overlying growth section. An outstanding feature in the preorogenic succession is the transition from amiddle Eocene-Miocene carbonate platform (easternmost sector) to a basinal sequence, through a slope domain. An upper Messinian unconformity marks the onset of foreland flexure and the sedi-mentation of Pliocene synorogenic deposits in the sinking foredeep. A Pleistocene unconformity marks the end of the major contractional event, followed by the progressive filling of the depocenter.

Figure 1: Location of the seismic profile.

Figure 2: Post-stack, time-migrated 3-D seismic reflection profile across the Basil anticline and the Apenninic foredeep. The presenceof gas is indicated by the pull-down effect in Pliocene sediments.

Figure 3: Interpretation of the seismic profile. Basic sedimentary and tectonic features are highlighted. The lateral transitions among middle Eocene-Miocene sediments are well imaged, as well as theoutstanding Pleistocene unconformity and the overlying progradational sedimentary structures.


94

The Basil anticline provides a spectacular example of a thrust-related anti-cline. Almost the entire fold shape and its interactions with surface pro-cesses (syntectonic sedimentation and erosion) are very well recorded.Deformation terminated before the occurrence of any fault breakthroughwithin forelimb and this prevented any distortion induced by further fore-landward translation. Upper Pliocene strata thin onto the crest of the anti-cline, suggesting that they are growth strata (e.g. Suppe et al., 1992).Erosion of part of the crest and the forelimb indicates that the late Plioceneevolution of the anticline progressed in subaerial conditions. The probablepresence of wedge geometries in the Pleistocene sediments may support alate reactivation of the fault-fold pair. We interpret this anticline as agrowth fault-propagation fold (Figure 5). Details of the basic observationsdiscussed above are provided in Figure 6.

Figure 5: Numerically modeled (HCA; Salvini et al., 2001) cartoon showing the reconstructed evolutionary stepsfor this sector of the Apenninic foreland system and the interpretation of the Basil anticline as a growth fault-propa-gation fold.

2-9: Basil anticline

Figure 4: Seismic image of the Basil anticline.

Outward propagation of the sole thrust along the bottom of the synorogenic sediments.Displacement on the upward migrating frontal ramp is accommodated by the develop-ment of the Basil fault-propagation anticline. Folding occurs in a high sedimentation envi-ronment and well developed growth wedges form on both limbs.

Basil anticline

Flexural sinking of the foreland and deposition of synorogenic clastic sediments in a fore-deep environment.

Preorogenic succession


95

2-9: Basil anticlineThe Basil anticline has a gently dipping backlimb and a steep forelimbtruncated by the thrust ramp. The shape of the anticline is well round-ed in the upper part and becomes more angular in the fold core. Thetwo axial surfaces bounding the flay-lying crest can be traced in thisregion of the fold and their downward prosecution indicates that theymerge at a point located near the upper Messinian unconformity. Thissuggests that the basal decollement is located at the bottom of thePliocene sediments. Upper Pliocene growth wedges are well imaged inboth limbs, supporting their rotation during fold evolution (e.g. Hardyand Poblet, 1994). In particular, the large syntectonic fan in the fore-limb indicates that most of its total rotation occurred in the latePliocene. The occurrence of limb rotation in fault-propagation anti-clines is predicted by the trishear kinematis (Erslev, 1991; Hardy andFord, 1997; Allmendinger, 1998). Wedge geometries can be imaged inthe lower Pleistocene sediments overlying the forelimb. They providea reliable evidence of a limited fold activity post-dating the earlyPleistocene erosional event. Two other sedimentary wedges are tenta-tively imaged in younger strata deposited on both limbs.

Conclusions:• The Basil anticline is a notheast-verging fault propagation

fold developed at the tip of a thrust ramp that soles downinto the upper Messinian unconformity.

• Upper Pliocene and, possibly, lower Pleistocene strata are syntec-tonic units folded during fault motion.

Figure 6: Interpreted seismic image of the Basil anticline showing basic features that have been used for the reconstruction of fold kinematics.


96

2-10: Salt weld detached fault-propagation foldsFrank Bilotti, Timothy Brickner, Thomas Elliott, Chip Morgan, Richard Redhead, Yusri,Unocal, Sugar Land, Texas, U.S.A.

Location: Deepwater Espirito Santo Basin, BrazilTopics: Fault-propagation folding, salt welds, detachments

Figure 1: Regional map of the Espirito Santo Basin.

Figure 2: Pre-stack time migrat-ed seismic profile converted todepth. This line images two con-tractional structures that detachfrom the welded autochthonoussalt level.

An early Tertiary, north-south–oriented compressional event inthe Espirito Santo Basin formed a mixture of salt-weld detachedfault-propagation folds and compressed salt walls. The largerstructures preserve a combination of salt-deflation stratigraphicgeometry and contractional fold geometry.

In this example we model one of these asymmetric contractionalstructures as a constant thickness fault-propagation fold. Wefind that two solutions fit the data depending on where onedefines kink bands and measures the limb dips. Of the two solu-tions, a low-angle breakthrough solution that honors the deepfold geometry fits the data better.

Figure 3: Summary of sub-regionalstructural history of the deepwaterEspirito Santo Basin. Horizontal weldsare less extensive in more distal partsof the basin.

autochthonous salt water


97

Figure 6: Axial surface maps at 2 seismic depth slices. The pattern of a broadening flat crest with decreasing slip predicted by fault-prop-agation folding is supported by the data. The deeper slice at 6400 ms shows the intersection of the crestal axial surfaces, which is alsoconsistent with the fault-propagation folding model.

2-10: Salt weld detached fault-propagation folds

Discussion:We employ a fault propagation fold model to interpret this structure based on its first-order struc-tural geometry as an asymmetric fold with a steep or faulted forelimb. The well-defined basaldetachment and the lack of structural relief across the structure indicate that the structure solesto a horizontal detachment.

Figure 5: Model for a simple fault-propagation fold from a flat detachment.

Figure 4: Main kink bands defined by axial surfaces.

forelimb inactiveaxial surface

backlimb activeaxial surface

single crestal axial surface


98

Solution 1: Using shallow fold geometryWe utilize fault-propagation folding theory to provide balancing constraints for the poorly imaged core of thisstructure. Seismic loop ties and bisecting axial surfaces are the basis for the geometric interpretation.Disagreement between the model structure and the seismic image are due to either geologic complexity notbeing accounted for in the model or shortcomings of the seismic image.


This balanced section works well for the shallow geometry that weused to constrain the fault-propagation fold model; however, it doesnot agree well with the bed dips in the deeper part of the structure. Infact, the steeper bed dips at depth suggest that the model fault geom-etry would actually cut down section with respect to the hanging-wallrocks. This leads us to explore another interpretation of the structure.

Figure 7. Interpretation of the structure using fault-propagation fold (FPF) theory.

a. Predicted forelimbdip based on bisect-ing axial surfacesmatches seismic tie,probably a steepfold limb.

b. Seismic tie is extended across structureand bisecting axial surfaces projected down-ward. There is no net structural relief acrossthe structure so we postulate a flat basaldetachment.

c. FPF theory predicts a basal step-up angle φ = 15° for γ = 30°. Since θ2 = φ, backlimb dipsshould be equal to the basal step-up angle. Inthis section the backlimb dips at 17°. Usingthese parameters we predict the location ofthe fault.

Figure 8. Fault-propagation fold solution for the structure using theshallow geometry as the main constraint.


99

Solution 2: Using deep fold geometryWe again utilize fault-propagation folding theory to provide balancing constraints for the poorlyimaged core of this structure. In this case, however, we choose to respect the geometry of thedeeper fold. The deep fold limbs are steeper and have more inflections in dip, yielding a more com-plex solution. Here we invoke a low-angle breakthrough of a FPF.


Figure 10. The final fault geometry resultsfrom the addition of the breakthrough fault aswell as folding of the existing fault in the core ofthe fault-propagation fold. The shallow foldgeometry generally reflects the fold shape, butnot the exact dip angles. The disagreement canbe due to the changing thickness of strata in thesection or mechanical thickening or thinning ofbeds. Because the history of salt deflation caus-es variation in stratigraphic thickness we pro-pose that this fold forms in rocks with pre-exist-ing thickness variations.

Conclusions:• After nearly complete deflation and welding, the autochthonous salt level still provides a

sub-horizontal detachment surface for thin-skinned contractional structures• This structure fits the basic geometry and kinematics described by fault-propagation

folding theory.• Two models were tested; of these, a more complex model utilizing deep geometry as the

primary constraints provides a better fit to the data.

b. Using theFPF foldgeometry wecan predictthe first-order faultgeometry.

c. We postulate a low-angle fault break-through to explain the additional kink band.Using the model developed in Suppe andMedwedeff, 1991, we balance the modelusing an additional kink band whose widthis the same as the amount of the break-through slip.

a. Defining kink bands from the deep fold geom-etry yields 7 dip panels. The overall geometrystill fits the FPF model; however, another detailmust be added to explain the extra kink band. Using δf=60° and θ2=φ, constant thickness FPFtheory predicts backlimb dip δb=34°. This pre-dicted dip matches the dip of the deep reflectiv-ity of the backlimb.

Figure 9. Interpretation of the structure using fault-propagation fold theory andgeometric constraints from the deeper part of the structure.

predicted fault geometry

kink bands low-angle breakthrough

balanced model section


100

2-11: Structural inversion along the Sakala Fault, East Java Sea, Indonesia

Shankar Mitra, University of Oklahoma, Norman, Oklahoma, U.S.A.

Location: East Java Sea, IndonesiaTopics: Inversion, fault-related folding

The Sakala structure (Figure 1) is a fault-related inversion structure in the East Java Sea (Figure 2), located in a back arc settingbehind the Java trench. Along this trench, the Australian plate is subducted under the Eurasian plate, along a north-dipping sub-duction zone (Hamilton, 1979). Inversion structures in this area resulted from north-south extension in the Eocene and Oligocene,followed by compression in the same general direction, in the early Miocene.

Figure 1: Time section through the Sakala structure.

Figure 2: Generalized map of the East Java Sea, showing the location of the Sakala inversion structure, and a seismic reflection profile through the structure.


101

2-11: Sakala inversion structure

Figure 4: a. Interpreted depth profile throughthe Sakala structure. b. Restoration of the inter-preted seismic profile to the pre-compressionalstage, using antithetic inclined shear. Part ofthe compressional fault-propagation fold is notrestored, so that the true restored geometry ofthe top of the Prupuh Formation is given by thesolid gray line. If A1=A2, the structure is areabalanced.

Figure 3: Pre-stack depth-migrated seismic profilethrough the Sakala structure. a. Uninterpreted profile. b.Shaded area represents the interval between the tops ofthe Ngimbang and Prupuh Formations. Note that thethickened section is in the uplifted block, suggestingstructural inversion.

In order to interpret the detailed geometry and evolution of thestructure, a pre-stack depth migrated section (Figure 3) wasused. The structure is interpreted as an inversion structureformed along the south-dipping Sakala fault. The interval

between the Ngimbang and Prupuh Formations shows a signifi-cant increase in thickness from the footwall to the uplifted hang-ing wall across the fault. The thickness also increases graduallyaway from the fault zone. The structural geometry of the units

closely resembles that formed by the compressive reactivationof an extensional fault-propagation fold with fault breakthrough.The fold geometry was used to model the fault geometry, whichis poorly imaged on the seismic sections (Figure 4a).


102

Forward modeling and restoration was used to decipher thedetailed evolution of the structure (Mitra, 1993). An experimen-tal clay model (Figure 5; Mitra, 1993; Mitra and Islam, 1994)demonstrates the development of an extensional fault-propaga-tion (drape) fold, and the drape dip resulting from break-through of the fault. The experiment simulates the deformation

in pre-extensional units, modeled by a clay layer, above a base-ment fault dipping 45° degrees. Extension initially results in theformation of a broad fault-propagation fold (Figure 5a). Thedeformation occurs by the sequential development of a largenumber of small normal faults, which are progressively rotatedto steeper dips with increasing extension. The extension even-

tually results in a major fault breaking through the clay unit at asteeper angle than the basement fault. The fault-propagation ordrape dip is preserved in the hanging wall, and is rotated as itpasses through the synclinal hinge (Figure 5b).

2-11: Sakala inversion structure

In late Eocene and Oligocene time, extension resulted in the development ofa fault-propagation fold above a deep-seated planar fault dipping approxi-mately 40° (Figure 6a and b). The fault propagated at a steeper angle (55°)through the fold in the Ngimbang and older units and subsequently throughthe synextensional Prupuh Formation. The propagation of the fault throughthe extensional fault propagation fold resulted in a basinward drape dip A-B(Figure 6b). This drape panel was rotated to a shallower dip (B-C) as it passedthrough the synclinal hinge. Basinward of C, units show a small dip into thefault. The synextensional growth units deposited in the hanging wall showeda progressive increase in thickness into the basin through the three major dippanels.

Compressive deformation in the early Miocene resulted in folding of units asthey passed through fault bends (Figure 6c). The panel B-C was folded to itsoriginal steeper dip, and horizontal beds basinward of C were folded to thedipping panel C-D. In addition, a tight fold developed within the synexten-sional units at the tip of the fault. The compressive folding in the possiblyunconsolidated synextensional sediments may have resulted in some arealoss, although the structure can be area balanced by assuming some pene-trative deformation.

Figure 4b shows the restoration of the interpreted seismic profile to a post-extensional stage, using the model described above. Variable inclined shearwas used to restore parts of the hanging wall deformed by fault-bend folding.The restoration did not remove the effects of compressive fault-propagationfolding at the leading edge of the structure. The post-extensional top of thePrupuh Formation possibly had the geometry shown by the dark gray line inFigure 4b. The leading edge of the structure was folded to the geometryshown by the dashed line during compressive deformation, with the area A1= A2.

Figure 5 (above): Clay model showing the development of an extensional fault-propagation fold, and the subsequent breakthrough of a major fault. Note the develop-ment of a drape dip panel in the hanging wall.

Figure 6 (right): Evolution of the Sakala structure. a. Pre-extensional geometry. b.Post-extensional geometry. Note the development of a drape dip due to extensionalfault propagation folding in the hanging wall. c. Final structure, resulting from com-pressive reactivation of the Sakala fault.

Conclusions:• The Sakala structure in the East Java Sea is interpreted to be an inversion structure formed by Miocene compressive reactivation of an Eocene-Oligocene extensional structure.• Extension along the Sakala fault resulted in an extensional fault-propagation (drape) fold with subsequent fault breakthrough, resulting in the preservation of a drape dip in the hanging wall.• Compressive reactivation along the fault occurred by fault-bend folding, accompanied by fault-propagation folding at the leading edge of the structure.

Top Ngimhang Formation

Top Prupuh Formation


103

2-12: Detachment fold, Niger DeltaFrank Bilotti, Texaco, Bellaire, Texas, U.S.A.John H. Shaw, Department of Earth and Planetary Sciences, Harvard University, Cambridge, Massachusetts, U.S.A.Ronald M. Cupich, Texaco, Bellaire, Texas, U.S.A.Roisin M. Lakings, Texaco, Bellaire, Texas, U.S.A.

Location: Deepwater Niger DeltaTopics: Detachment folding, growth strata, restoration

Figure 2: Migrated 2-D seismic profile through the detachment fold. We observetwo basic structural patterns in the seismic profile (top) that are consistent withdetachment folds (left): 1) symmetric, dipping fold limbs situated over flat reflectorsin the Akata Formation and basement; and 2) syntectonic growth strata with beddips that shallow upward toward the seafloor. These observations, and the lack ofan obvious thrust ramp beneath the fold, indicate that the structure is a detachmentfold formed primarily by limb rotation. The structure grew during the Pliocene andQuaternary.

Figure 1: Bathymetry of the offshore Niger Delta showing the major structural belts and the loca-tion of the study area. Modified from Connors et al. (1998).

We describe a large detachment fold located between the inner andouter fold and thrust belts of the deepwater Niger Delta (Figure 1).

Seismic lines define the structure as a broad, symmetric anticline involv-ing Miocene and lower Pliocene deltaic strata (Figure 2). The structureis overlain by syntectonic growth strata that show an upward fanning oflimb dips in the upper Pliocene and Pleistocene section, indicating thatthe fold grew by limb rotation. A continuous, relatively flat basementunderlies the fold indicating the presence of a sub-horizontal detach-ment surface in the pro-delta Akata Formation. Detachment foldingrequires ductile thickening of the Akata Formation above the basaldetachment in the core of the fold.


104

2-12: Niger Delta detachment fold

Conclusions:• This structure is a detachment anticline that grew primarily by limb-rotation since the early Pliocene.• The basal detachment is located in the Akata Formation marine shales, which thickened in the core

of the fold during growth.

We present a kinematic model (Figure 3) and interpreted seismic reflectionprofile (Figure 4) across the detachment fold. The kinematics of fold growthare recorded by the geometry of syntectonic strata. In Figure 3, we comparethe growth of a model detachment fold with a restoration of the seismicinterpretation using heterogeneous inclined-shear. The restoration demon-strates that the structure grew primarily by limb-rotation, with a minor com-ponent of limb widening between restoration steps B and C.

The structure is cored by pro-delta marine sediments of the Akata Formation(Figure 4). Detachment folds require that material in their cores deform andthicken to accommodate fold amplification. The Akata Formation, a marineshale and the probable hydrocarbon source rock, exhibits this increasedthickness.

Figure 3: Sequential model (0-3) of a detachment fold (left) with fixed limb widths that grows by limb rotation. Themodel is compared with a balanced restoration of the structure (right) derived using variable inclined-shear (Novoa et al.,1999).

Figure 4: Interpreted seismic section and geologic cross-section through the detachment structure.


105

2-13: Mississippi Fan Fold Belt, Gulf of MexicoMark G. Rowan, Rowan Consulting, Inc., Boulder, Colorado, U.S.A.Frank J. Peel, BHP Billiton Petroleum, Houston, Texas, U.S.A.

Location: Offshore Louisiana, northern Gulf of MexicoTopics: Salt-cored detachment folds, reverse faults, growth strataReserves: Giant fields along trend to west (e.g., Mad Dog)

Figure 1: Map of the northern Gulf of Mexico showing the distribution of allochthonous salt (black), bas-inward- and landward-dipping faults (blue and red), and deepwater folds (green). Modified from Diegel etal. (1995) and reprinted by permission of the AAPG.

The Mississippi Fan fold belt is one of several deepwater contractional provinces that formed in responseto gravitational failure of the northern Gulf of Mexico passive margin (e.g., Diegel et al., 1995; Peel et al.,1995; Rowan et al., 2004). It comprises salt-cored detachment folds and associated reverse faults thatdeveloped principally during the late Miocene (e.g., Weimer and Buffler, 1992; Rowan, 1997). Although allfolds were originally thought to be cored by the autochthonous Louann salt, modern data show that thefrontal folds are detached above an Upper Jurassic to Lower Cretaceous allochthonous nappe (Peel, 2001;Rowan et al., 2001, 2004).

In this section (2-13), we examine the three-dimensional geometry of a composite frontal fold using aseries of 3-D time-migrated seismic profiles and structure maps. The profile geometry varies considerablyalong strike from a relatively simple, symmetric, unfaulted detachment fold (Figures 2, 3) to an asym-metric, faulted fold that is vergent either basinward (Figures 4, 6) or landward (Figure 5). Also, an earlier(Mesozoic) deformation phase complicates the deep geometry. Thus, no simple 2-D or 3-D structuralmodel adequately explains the relationship between the fold and associated faults, and geometric and/orquantitative models are of minimal use in aiding seismic interpretation in this case.

Figure 2: Uninterpreted and interpretedviews of Profile A showing symmetric,rounded, unfaulted detachment fold cored bysalt. Blue — top and base of allochthonoussalt and equivalent weld (indicated by pair ofdots); purple — undated horizon, possiblyUpper Jurassic; green — MCU (mid-Cretaceous unconformity), alternatively iden-tified as top Cretaceous using new well data(T. Dohmen, 2001, personal communica-tion); red — top Oligocene; orange — intra-Miocene; yellow — time-transgressivegrowth unconformity/onlap surface. Horizoncorrelation around the plunge termination ofthe fold shows that strata truncated by theunconformity at (1) are age-equivalent tothose onlapping the surface at (2); deepergrowth strata are thinned and rotated on bothflanks (3). 3-D data courtesy ofWesternGeco; location shown on Figures 7and 10.


106

2-13: Mississippi Fan fold belt

Figure 3: Uninterpreted and interpreted views of Profile B showing a rounded detachment fold with a very slightasymmetry and a minor, high-angle reverse fault on the forelimb. Again, note the differences in growth strata betweenthe two limbs. Horizons as in Figure 2; location shown on Figures 7 and 10. 3-D data courtesy of WesternGeco.

Figure 4: Uninterpreted and interpreted views of Profile C showing an asymmetric detachment fold with a long, planar backlimb and a steep forelimb cut by a basinward-vergent, high-angle reverse fault zone. The early deformation stage is clearly shown by the structural thinning and thickening (4) between the top salt (blue) and the topOligocene (red). Horizons as in Figure 2 (dashed where approximate, dotted where uncertain); location shown on Figures 7 and 10. 3-D data courtesy of WesternGeco.


107

2-13: Mississippi Fan fold belt

Figure 5: Uninterpreted and interpreted views of Profile D showing a broadly symmetric fold with a larger reverse fault on the basin-ward limb but a deep-level crest on the landward side. Again, note the differences in growth strata between the two limbs. Horizons asin Figure 2 (dashed where approximate, dotted where uncertain); location shown on Figures 7 and 10. 3-D data courtesy ofWesternGeco.

Figure 6: Uninterpreted and interpreted views of Profile E showing an asymmetric fold with a long, gentle backlimb and a steeper, faulted fore-limb. Again, note the differences in growth strata between the two limbs and the early deformation visible at depth on the backlimb. Horizons asin Figure 2 (dashed where approximate, dotted where uncertain); location shown on Figures 7 and 10. 3-D data courtesy of WesternGeco. Thewell was dry.


108

2-13: Mississippi Fan fold beltThe three-dimensional geometry of this salt-cored fold is complex, consisting of three en-echelon segments (I, II, andIII) and a smaller segment (IV), with individual culminations separated by saddles (Figure 7). Thus, any given profilethrough the fold is likely to display significantly different geometries (Figures 2-6). Furthermore, the relationshipbetween the fold and its associated faults is highly variable: Segments I and II each have a frontal reverse fault (linkedat prominent cusp), segment IV has a small landward-vergent fault, and segment III is unfaulted (Figures 7 and 8).

Initial growth strata are thinned and rotated on both limbs (3 on Figure 2), consistent with detachment folding withprogressive limb rotation (e.g., Hardy and Poblet, 1994). However, shallow growth geometries (backlimb truncationat 1 and forelimb onlap at 2 in Figure 2) are similar to those modeled for fault-bend folds with synkinematic erosion(Figure 9) (Suppe et al., 1992; Hardy and Poblet, 1995). Although the detachment (base of salt nappe) does indeedramp up, the studied fold is a detachment fold with no higher-level flat or wedge thrust. Thus, a growth pattern ofbacklimb truncation and forelimb onlap does not necessarily define a fault-bend fold. In this case, it was generatedby a salt-cored detachment fold in which the forelimb locked up as the backlimb continued to rotate.

Figure 7: Time-structure contour map of the studied fold (segments I, II, III, and IV) and more landward structures. Yellows and reds are highs, blues and purples are lows; thinblack lines are reverse faults and black blobs are salt diapirs. The three-dimensional geometry is very complex: individual fold segments may have different orientations, plungeangles, fold-fault relationships, and diapiric influence. Grey lines show seismic profiles of Figures 2-6.

Figure 8: Variation of shortening along the strike of the fold, divided into faulting and foldingcomponents. A through E are the five profiles illustrated in Figures 2 through 6, respectively.Modified from an earlier interpretation (Rowan, 1997), with shortening values determined fromline-length restoration of fourteen equally spaced profiles. There is no direct correlation betweenfault and fold geometries because faults are secondary structures that may or may not developand modify preexisting detachment folds.

Figure 9: Modeled fault-bend fold with synkinematic erosion and sedimentation (modifiedfrom Hardy and Poblet, 1995). The red horizon is a time-transgressive growth unconformity(analogous to the yellow horizon in Figures 2-6), with time-equivalent strata truncated on thebacklimb (1) and onlapping the forelimb (2). The resulting growth geometry is very similar tothat observed in salt-cored detachment folds of the Mississippi Fan fold belt, which do not con-tain an upper detachment and a connecting ramp (compare this figure with Figure 2b). Thus,backlimb truncation and forelimb onlap do not uniquely define a fault-bend fold, but simplyshow that the two limbs behaved differently.


109

2-13: Mississippi Fan fold beltAnother complicating factor is an earlier stage of deformation. Thickness variations in the sectionbetween the top salt and the top Oligocene (e.g., 4 in Figure 4) reflect a dominantly Cretaceous tec-tonic event. An isochron map of an interval immediately above salt shows a complex pattern of paleo-highs and paleo-lows (Figure 10). A related feature is that the suprasalt Cretaceous and Paleogene sec-tion is generally thinner than in the area basinward of salt (e.g., between the purple and red horizonsin Figure 2) because of distal salt inflation during the early history of the margin (Hall, 2000).

The variable profile geometry of this fold is thus partly due to the multi-phase deformation historyillustrated in Figure 11. Deformation began almost immediately after salt deposition due to differentialthermal subsidence and the consequent basinward tilt. This resulted in a combination of distal infla-tion, nappe extrusion, and folding beneath a thin overburden. The complex geometry of these struc-tures (Figure 10) is interpreted as an interference pattern during convergent gliding off both theFlorida and Louisiana margins. The early structures then served as buckling instabilities for the later(Neogene) deformation, but only some were reactivated because of the thicker overburden, and thuslonger wavelength, of the detachment folding.

Figure 10: Isochron map of an interval immediately above salt showing the geometry of the early (dominantly Cretaceous) deformation. Thins correspond-ing to paleo-highs are in yellow and red; thicks corresponding to paleo-lows are in blue and purple. The complex pattern influenced the development of thelater (Miocene) structures, shown by the black lines with arrows, resulting in the larger wavelength, variable fold geometries observed today.

Conclusions:• Salt-cored detachment folds in the Mississippi Fan fold belt have complex three-

dimensional geometries with significant variations along strike caused by variablefold-fault relationships and the effects of an earlier deformation phase.

• Patterns of growth strata are ambiguous and cannot always be used to determine thefold style and nature of underlying faults.

• In the case of salt-detached fold belts on passive margins, therefore, applying simplegeometric and quantitative models to shallow horizons in order to constrain thedeeper interpretation is often inappropriate.

Figure 11: Schematic evolution based on quantitative restorations and regional considerations (modified from Rowan et al., 2000):(a) Upper Jurassic salt deposition; (b) gravity gliding caused by Cretaceous thermal subsidence and basinward tilting results in distalinflation, nappe extrusion, and small-wavelength folds; (c) relative quiescence during the Paleogene as thermal subsidence and tiltingwane; (d) gravity spreading of the Neogene progradational margin, resulting in larger-wavelength folding; and (e) cessation of defor-mation as the Pleistocene deepwater Mississippi fan is deposited. Sections are not drawn to scale, and the effects of salt withdrawaland diapirism are not shown.


110

2-14: Yakeng detachment fold, South Tianshan, ChinaAurélia Hubert-Ferrari, Institut de Géologie, Université de Neuchâtel, Neuchâtel, SwitzerlandJohn Suppe, Department of Geosciences, Princeton University, Princeton, New Jersey, U.S.A.Xin Wang, Department of Geosciences, Zhejiang University, Hangzhou, ChinaChengzao Jia, PetroChina, Beijing, China

Location: Kuche, Tarim Basin, Xinjiang, ChinaTopic: Analysis of a detachment fold in the thickness domainReserves: Exploration region

Figure 2a: Uninterpreted seismic (horizontal scale equals vertical scale, topography exaggerated x4).

Figure 2b: Interpreted seismic (horizontal scale equals vertical scale, including topography) showing the 27 horizons used in analysis of Yakeng. Two major detach-ments bound the thickening yellow and orange interval.

Key Point: Yakeng anticline illustrates the importance of working inthe thickness domain when interpreting detachment folds. Measure-ments in the thickness domain show that Yakeng has 1.2 km shorten-ing above a basal two levels of major detachment, basal diapirism,basement folding, and a 2.4-km-thick growth sequence.

Structural Setting: The active Yakeng anticline is topographicallyexpressed by deformation of the alluvium at the front of the southernTianshan thrust belt (Figure 1). Seismic imaging and drilling (Figure 2)show it to be a classic detachment fold lying above a decollement inthe evaporite-rich Tertiary Jidikuh Formation, which roots northward(below horizon 4) into the massive 200-km-long Quilitak anticline(Figure 1). Just to the south of Yakeng anticline is the Yanan anticline,which is a basement-involved inversion structure whose north flankinterferes with the south flank of the Yakeng anticline.

Figure 1: Yakeng fold at the front of the southern Tianshan. (Landsat TM)


111

Geomorphic expression: The morphology of the 50–150-m-hightopographic anticline illuminates the most recent increment of foldgrowth and sedimentation. It is a deformed and incised alluvial sur-face for which prior through-going drainage systems are still visible(Figures 3 and 4), showing that deposition previously exceededuplift, similar to the present situation at Kuche where Yakeng islargely buried (Figure 1). Limb dips (3–4°) in the valley east of theseismic line (Figures 1, 2b) are a significant fraction of the seismicdips (4–6°), indicating the extreme youth of Yakeng anticline.

Drainage and sedimentation: The topographic anticline is a bar-rier to the river networks (Figures 1, 3); only regionally importantrivers can now cross Yakeng. Smaller streams previously crossedYakeng anticline as demonstrated by the numerous well pre-served wind gaps (Figure 3) and by southward merging channelnetworks that are continuous across the wind gaps (Figures 3, 4).This implied reorganization of drainage networks is an effect ofdecreasing stream power caused by decreasing stream gradientsassociated with fold growth. As a result, sediment is preferential-ly trapped north and south of Yakeng (Figures 5, 6), producing atopographic expression that is narrower than the anticline atdepth, especially on the north flank (Figures 2, 3, and 12).

2-14: Surface expression of fold growth and sediment trapping

Figure 3: The active Yakeng anticline forms a 6-km-wide rounded topo-graphic ridge that few rivers can incise, as shown by the many wind gaps(w). Northward tilting of the north flank of Yakeng and alluvial depositionboth decrease stream gradients, which favors the development of chan-nels on the sides of the alluvial fans and along the northern limb ofYakeng (1). Others channels have a converging pattern (2) which increas-es their stream power sufficiently to keep incising Yakeng anticline. Theincrease of meander amplitude and wavelength across Yakeng also reflectthese changes in gradient.

Figure 6: A facies change on the northern limb of Yakeng anticline is visible in the field (top) and in the seis-mic reflection profile (bottom). The northern edge of the anticline is mainly formed by thick dark conglomerate(Xiyu F.) whereas its top is composed mainly of yellow-grey sandstone. Most coarse dark conglomerates beganto be deposited during the glacial period (1.8 Ma to present). They progressively filled the basin betweenQuilitak and Yakeng anticlines.

Figure 5: Lowrounded morphologyof the Yakeng anticline.

Figure 4: Southward-converging drainage networks are interrupted by wind gaps at the crest ofYakeng anticline. Flow is now to the north on the north flank of Yakeng. These southward converg-ing networks formed before Yakeng anticline developed its present topographic expression. Seismicline in black.


112

2-14: Seismic characteristics and folding mechanism

Figure 7: Yakeng anticline dies out downward in height and width, indicating a basal detachment (1, horizon 4), which extends to the north under Quilitak anticline.Yanan anticline is a basement-involved inversion structure that is young, as shown by changes in structural relief on its south flank (2). Yanan interferes with Yakeng anti-cline (3), making analysis of Yakeng more challenging.

Initial assessment: Yakeng anticline dies out downward (Figure7), suggesting it is a classic detachment fold that can be analyzedquantitatively for shortening and timing (Figure 8). HoweverYakeng is too complex because of regional variation in strati-graphic thickness below horizon 15 (Figure 9) and interferencewith Yanan anticline (Figure 7). This forces us to move our analy-sis of Yakeng from the depth domain to the thickness domain(Figures 10–14).

Figure 9: Measurement of area of structural relief (A11) following the model of Figure 8 is ambigu-ous since the undeformed regional gradient (4) is hard to determine because the basement is foldedand thickness varies regionally. Therefore we move our analysis to the thickness domain (Figures10–14).

Figure 8: Classical detachment folds are characterized by a linear upward increase of area of struc-tural relief A = hs within pregrowth strata (Epard and Groshong, 1993). By measuring the area ofstructural relief of many horizons the magnitude — and the timing — of shortening can be deter-mined s = A/h. Shortening can also be determined for each layer from bed-length measurements s = δL = L2 – L2�, but only if bed length is conserved. Yakeng anticline is significantly more complexthan this model.


113

By flattening the structure to appropriate horizons we can view the structure in thethickness domain and more easily determine the regional stratigraphic gradients(Figures 10–12), which are needed to measure areas of structural relief (Figures 10,12). The analysis shows us that interval 4-15 has undergone significant shortening(1200 m) and interval 4-5 has undergone additional diapiric flow (0.8 km2). The over-lying strata (15-27) show modest thinning over Yakeng and nearly constant thick-ness relief, which can be modeled as the beginning of growth. Strata above horizon27 are more strongly thinned, showing a recent acceleration of growth of Yakeng,preceeding its emergence as a topographic feature.

2-14: Analysis of Yakeng in the thickness domain

Figure 12: Yakeng anticline flattened to horizon 4 (h=v). The analysis given below shows that horizons 5–14 have undergone 1200 m of shortening and thickeningabove an evaporitic detachment. There is and additional 0.8 km2 of diapirism in the basal layer (4-5). Horizons 15–27 show a nearly linear upward dearease in shorten-ing. After horizon 27 time shortening and uplift has accelerated leading to topographic emergence.

Figure 13: Area of thickness relief increases linearly from layer 5 to 15 indi-cating a nearly constant shortening of 1200 m (compare Figure 8). The non-zero intercept indicates an additional 0.8 km2 of diapiric flow in the basalevaporitic interval (4-5). The interval of nearly constant relief (15–27) can bemodeled as a growth internal (δS/δH = 0.2, assuming diapirism is after hori-zon 27.)

Figure 14: Shortening is calculated from area of relief minus the diapiric area(see Figures 8 and 13). The nearly linear shortening within the growth interval(15-27) suggests that diapirism is late, leading to the topographic emergence ofYakeng. The larger apparent shortening of layers 5-6 may suggest a small addi-tional diapiric component.

Conclusions: Yakeng anticline the value of analysis in the thickness domain:

• Thickness analysis clearly identifies the growth, pregrowth, and diapiric intervals.• Beds in the pregrowth sequence have shortened by 1200 m.• There a significant diapiric component in the basal evaporitic layer (0.8 km2).• The growth of Yakeng between horizons 15–27 shows a nearly linear rate of shortening, followed by an

acceleration of growth and topographic emergence.• The topography shows folding of previously through-flowing stream valleysThis study was supported by NSF EAR-0073759, NSFC 49832040, TPEDB-PetroChina, and Princeton 3-D Structure Project.

Figure 10: Section produced by flattening on horizon 4. It iseasier in this thickness display than in the depth display (Figure9) to identify the regional stratigraphic gradients (1, also seeFigure 11) for measuring area of structural relief (A11) and unde-formed height (h11). Note that Yanan anticline largely disappearsin this thickness display because it is a flexural-slip fold conserv-ing layer thickness, whereas Yakeng is visible because it hasgrown by layer thickening. Yanan anticline appears only as adepression (2) because of stratigraphic thinning due to growth(see interval 15–22 Figure 11), but at deeper levels (3) the slightdepression is produced by flow of the basal evaporitic layer (4-5).

Figure 11: The keys to interpreting Yakeng anticline arerevealed by its thickness variations. The basal evaporiticinterval (5–4) shows thickness variation largely caused bydiapiric flow (Figure 13). Interval 5-15 shows large regionalnorthward stratigraphic thickening reflecting syndeposition-al flexure of the basement, plus local structural thickeningat Yakeng. The overlying interval 15-27 shows strati-graphic thinning over both Yanan and Yakeng anticlines,indicating growth. The uppermost interval (topo-27) showslarge thinning over Yakeng, indicating accelerated growthincluding diapirism.


114

2-15: Odd geologic structures of southern Oklahoma revisited, Oklahoma, U.S.A.

IntroductionC.W. Tomlinson (“Odd Geologic Structures of SouthernOklahoma,” 1952) observed that late Paleozoic “Structures oftypes somewhat unusual for the Mid-Continent region occur in theArdmore district of Oklahoma.” Despite structural peculiarity,application of fault-related fold theory to a modern 2-D seismicprofile can explain the geometry of Fox-Graham Field (Figure 2),one of Tomlinson’s “odd structures.” This section (2-15) illus-trates 1) application of fault-bend fold theory to produce a modelof fault shape and footwall structure, 2) how concepts of struc-tural wedging and fault propagation folding combine to produce amodel explaining the geometry and kinematics of the “rabbit-ear”fold (Figure 2) and, 3) how these models are synthesized to pro-duce a retrodeformable, kinematically-viable, forward model thatevolves to approximate the present geometry of the structure.

Figure 2: Uninterpreted, depth-converted 2-D seismic profile across Fox-Graham Field and the Harrisburg Trough. Besides the well-imaged fold that dominates the profile, of particularimportance in constraining a fault-related fold interpretation are the geometry of the pre-Pennsylvanian unconformity (1), recognition of fault plane reflections (2) and, recognition of footwallstructure (3). Tomlinson (1952) described the rabbit-ear fold along trend of the one imaged here. No vertical exaggeration.

Figure 1: Location map.

Paul Genovese, Grizzly Energy Resources, LLC., Columbia Falls, Montana, U.S.A.

Location: Ardmore Basin, Oklahoma, U.S.A.Topics: Fault-bend folding, fault-propagation folding, structural wedging, kinematic forward models, growth strata

approximate 2-Dseismic line location


115

2-15: Odd structures

Figure 3: Subdivision of the fold into dip domains (regions of equal dip, as in Suppe, 1983) in preparation for fault-bend fold analysis . Dips and the fault-planesegment are regarded as “hard” constraints for the purposes of interpretation. Question marks denote uncertainty in the downward continuation of the fault planeand fold axial surfaces (dashed black lines).

Figure 4: Constrained fault-bend fold solution fullypredicts fault-plane geometry.

Fault-Bend Fold Analysis: The true utility of fault-bend fold theory (Suppe,1983), is its capability to provide a complete fault/fold solution using limit-ed data. Figures 3 and 4 illustrate data constraints and fault-bend fold solu-tion for the large fold. The solution (Figure 4) is a multiple bend fault-bendfold. By itself, the solution is largely geometric with few explicit kinematicimplications. Incorporating other observations into the interpretationmakes it more robust. Recognizing the unconformity as an originally hori-zontal isochron, for instance, constrains possible kinematic solutions bydefining the timing and location of folding. As an example, the small fore-limb (+I in Figure 3) is not folded sympathetically with the unconformity,and is therefore interpreted as an older structure, as opposed to a limbformed by slip through Bend 3 (Figure 4).


116

2-15: Odd structuresModels of Footwall Deformation: The shape of the unconfor-mity provides critical information about the kinematic evolu-tion of the fold. Using the folded unconformity as a “straingauge” and restoring fault slip, it is demonstrated that thehanging wall fold cannot be the product of slip on a singlefault (Figure 5) but must result in part from folding in its foot-wall. Further, the folded unconformity helps constrain mod-els of footwall folding (Figures 6 and 7).

Figures 6a and 6b are models of footwall imbricatestructures proposed to explain the present geometry ofthe folded unconformity in Figure 5a. In each case,fault Bend 1 (and consequently Bends 2 and 3) aredeactivated by, and passively transported by, the foot-wall imbricate. Folding is generated only at Bend 4, asynclinal bend in the imbricate fault. These models donot explain the shape of the folded unconformity inFigure 5a, and are discarded as possible solutions.

Figure 5b: Model restored by removing fault slip does not restore the unconformity to hori-zontal above the red panel. A footwall fold of some kind (indicated by the question mark)must exist below the red panel to account for the discrepancy. The width of kink band A-A�represents the total slip of the hanging wall before the unconformity was formed. B-B�, whichlocally refolds A-A�, represents slip on an unspecified footwall fault after the time the uncon-formity formed.

Figure 5a: Kinematic model based on the fault-bed fold solution of Figure 4. Gray panelsrepresent rock folded through fault bends after the unconformity was formed. Their width isconsistent with fault slip applied at lower left. The red panel cannot be explained by the sameslip.

Figures 7a through 7e: Sequence of steps in a kinematic forwardmodel that, unlike those shown in Figure 6, reproduces the shape ofthe folded unconformity in Figure 5 and the seismic profile. In thismodel, a structural wedge folds the footwall and refolds the overlyinghanging wall. Slip on the upper fault in Figure 7e completes the defor-mation, giving the unconformity its present shape. Note the “shoul-der” produced, in part, by “rolling” a flat part of the unconformitythrough Bend 3 and tilting it forward onto the crest of the fold.Compare this feature with that shown on the seismic line in Figure 2.Folding generated in the footwall is shaded differently from that gen-erated in the hanging wall for clarity in Figure 7d, 7e (refer to key).While there is no direct evidence for a thrusted footwall wedge on theseismic data, southwest-verging thrust and reverse faults are notuncommon in the Ardmore Basin (e.g. Overbrook Thrust, CaddoFault).

active axial surface (anchored to fault bend/tip,beds fold as they move through)inactive axial surface (anchored to the rock, formersite of active folding)

folding generated by footwall wedge

folding generated by slip on upper fault

KEYunconformity

Bend 1

Bend 2

Bend 3

imbricate

“shoulder”


117

“Rabbit-Ear Fold:” An “ odd” type of structure described byTomlinson (1952) is the rabbit-ears anticlinoria. Figure 8 com-pares Tomlinson’s rabbit-ear fold with one imaged on the seis-mic profile a few miles away. Figure 9 shows a generalizedmodel of how this type of rabbit-ear fold could form as a typeof structural wedge.


Figure 9 shows the kinematic develop-ment of a fold similar to the rabbit-ear foldsin Figure 8, without the complications ofunconformities and pre-existing structure.

Figure 9a shows the incipient thrust andbackthrust (dashed red lines) that definethe wedge tip. Fault-bend fold parameters qand f are used with standard fault-bendfold theory (Suppe, 1983) to calculate thedip of kink band (a) in Figure 9b. This the-oretical treatment is identical to that of a“nondetachment wedge thrust” (Medwedeff,1988).

Figure 9b and 9c show how fault slip onthe backthrust is consumed by the fault-propagation fold. Theoretically, the dip offorelimb (b) is equivalent to that predictedby fixed-axis fault-propagation fold theory(Suppe and Medwedeff, 1990) where q = f= dip of (a).

Figure 10a: Initial conditions for kinematic model. (1) is a fold limbrelated to deep thrust, perhaps the Arbuckle Thrust, inferred by differentauthors to sole between ~ -30,000’ (Crawford et al., 1990) and -60,000'under the Ardmore Basin. (2) is an incipient thrust fault. (3) is a fault bendthat is the locus for development of a fault-propagation fold.

Figure 10b: (1) is a fault-propagation fold, with slight backward shear (2)applied to achieve balance. (3) is an incipient fault that will decapitate the fault-propagation fold consistent with the observation that the internal angle (4) of thefault-propagation fold (g* of Suppe and Medwedeff, 1990) is not found on the present-day hanging wall.

Figure 10c: Slip on fault (1) translates the hanging wall. Pre-Atokanunconformity (2) (erosion exceeds uplift) is folded forward (3) as it passesover fault bend (4). Atokan sediments onlap unconformity near (3) andfault (1) reaches the seafloor at (5) as a contractional growth fault.

Figure 10d: Emplacement of structural wedge (1) produces kink band (2) in the foot-wall and folds the overlying hanging wall. Subsequently, the pre-Pennsylvanian uncon-formity (3) truncates the hanging wall, preserving a remnant (4) of the pre-Atokanunconformity. Deposition of the Pennsylvanian Deese group follows erosion (5).

Figure 8 compares a kinematic model of a rabbit-ear fold superimposed on seismicprofile (Figure 8a, enlarged from Figure 2), with the rabbit-ear modified fromTomlinson (1952) constructed from well control (Figure 8b). Although the interpretedstructures are a few miles apart, they bear some similarity, most notably the foldedpre-Pennsylvanian (1) and pre-Atokan (2) unconformities in the core of the structure.

8a

8b

9a

9c

9b

10a

10d10c

10b

Forward Kinematic Model: Figure 10a through 10i shows stages in a balanced, kinematic forward model. Figure 10j shows thatthe model result is a good fit with the seismic data giving validity to the solution. It must be noted that the solution is notunique, especially for the footwall structure, but is kinematically viable and therefore more robust than a balanced cross sec-tion. Models like this are useful for considering any time-space dependent features in the petroleum system such as fracturedistribution and intensity, migration pathways and traps, and source-rock/reservoir juxtaposition during generation.


118


10e

10h

10f 10f

10i10j

Figure 10e: Initial wedge block fails and second wedge block (1) is emplacedproducing kink band (2) and slight bend in fault (3).

Figure 10f: Renewed slip on deep (Arbuckle?) thrust widens initial kinkband (1) by an amount (2) folding the entire section.

Figure 10g: Renewed slip fault (1) produces hanging wall folding at fault bends(2)–(5). Rabbit-ear fold (6) begins to form above fault tip (7).

Figure 10h: Continued slip on fault (1) amplifies growth of rabbit-ear fold. Figure 10i: Final increment of slip on fault (1) results in present struc-tural geometry.

Figure 10j: Final stage of kinematic forward model from Figure 10i, enlarged and superimposedon seismic profile. Note in particular the good fit between the model and the unconformity (1),“shoulder” (2), rabbit-ear fold (3), and footwall reflections (4).


119

Conclusions:This case study synthesizes basic observations from seismic and well data(Figure 2), fault-bend fold analysis (Figures 3 and 4), kinematic constraintson models of footwall deformation (Figures 5–7), principles of rabbit-earfolding (Figures 8 and 9), and a fully-retrodeformable kinematic model(Figure 10) to reasonably match the shape and explain the origin and tim-ing of structures observed on a seismic profile (Figure 11, Table 1). It isdemonstrated that application of fault-related fold theory can even yieldtractable geometric and kinematic solutions for odd structures, like thosefound in southern Oklahoma.

Acknowledgements:The author thanks Texaco Exploration and Production Inc. (in particularFrank Gaines) for providing the 2-D seismic line used in this study. Much ofthis study was completed as part of the author’s Ph.D. thesis research atPrinceton University, special thanks to advisor John Suppe and colleaguesJohn Shaw, Frank Bilotti, and Chris Connors. This section (2-15) benefittedfrom thorough and thoughtful reviews by Stephen Hook and Peter Brennan.

2-15: Odd structuresTiming ofstructuralevents

Table 1: Timing of structural events interpreted in kinematic model and on 2-D seismic profile. Letters correspondto labels in Figure 11. Orogenies are adapted from Lang (1957) and Tomlinson and McBee (1959), based mainly onthe presence of conglomerates in the stratigraphic section. Events A-G support punctuated orogenesis, whereas Hdemonstrates continuous deformation. The discrepancy in the thickness of Mississippian strata in well 3 vs. well 4indicates that uplift related to slip on fault D probably began prior to deposition of basal Atoka, as opposed to afteras illustrated in the kinematic model. Observation of growth folding (as in Suppe et al., 1992) constrain displace-ments in the kinematic model to show that the “shoulder” and rabbit-ear folds (Figure 11) formed coevally with G.Both of these folds trap and produce significant quantities of oil, demonstrating that migration occurred after thePennsylvanian.

Figure 11: Cross-section interpretation of the seismic profile from Figure 2, incorporating the final stage of the kinematic model from Figure 10i (boundary shown ingray) plus additional well control (2, 3, and 4). Stratigraphy in the kinematic model is modified to better fit well control. Refer to Table 1 for labels A–H.


120

2-16: Fault-bend folds in the southern Caribbean Ranges, San Carlos, Venezuela

P. E. Kraemer, Pecom Energia, S.A., Neuquen, Argentina J. Silvestro, Pecom Energia, S.A., Neuquen, Argentina

Location: Caribbean Ranges, San Carlos,VenezuelaTopics: Fault-bend folds structures, thrusting sequence, restoration

Figure 2: Uninterpreted and interpreted seismic section. Numbers 1, 2, and 3 indicate the suggested sequence of deformation of thrust sheets. Horizontal scale equalsvertical scale. Section trace shown in Figure 1. Time migrated seismic section displayed in depth.

The San Carlos fold belt is located at the southern tip of the alloc-thonous thrust front of the Caribbean Ranges, Venezuela (Figure 1). Theseismic example (Figure 2) shows a buried fold and thrust belt overly-ing a normal faulted authoctonous platform. The main structures arethree folds (α, β, γ) with typical kink geometries (Suppe, 1983) overlainby a Quaternary unconformity (U). The anticlines β and γ are linked toa common decollement folded by the δ anticline and deep kink panelsprobably related to footwall shale flow. The shortening on the mainthrust ramp is transferred to a structural wedge duplex at the front ofthe fold belt (d). The γ anticline is interpreted as a multi-bend-fault-bendfold (Medwedeff and Suppe, 1997) with two foot-wall (FWR1-2) andhanging-wall (HWR1-2) ramps. The β anticline is a folded single rampfault-bend fold. The syncline (α) is interpreted as an early thrust sheet,folded by a late thrust sheet (ε).

Figure 1: Location map a) Geologic map of the Southern Caribbean thrust front, Venezuela. b)Main structural features at the base of Quaternary unconformity.


121

Conclusions:• The seismic example shows typical kink-fold geometry.• Folds are interpreted as single- and multi-bend fault-bend folds.• Based on geometric constraints, the sequence of deformation is interpreted as shallow foreland propagat-

ing thrust sheets (α, β, γ) that are refolded by late deeper hinterland anticlines (ε, δ).

Acknowledgements:The authors would like to thank Pecom Energia S.A. for the authorization to publish this section (2-16).

2-16: Structural restoration

To test the validity of our structural interpretation and document thesequence of thrusting, we present a four-stage area balanced restoration ofsection A-A�. The sequence of deformation is summarized as:

Stage 0: Full restoration.

Stage 1: Footwall shale flow was active in the early stages of emplace-ment of thrust sheet α until the late emplacement of thrustsheet ε.

Stage 2: Emplacement of thrust sheet α.

Stage 3: Emplacement of thrust sheet β, γ folded by a late anticline δ.

Stage 4: Emplacement of thrust sheet ε that folds thrust sheet α.

Total shortening is 3.2 km, distributed as follows: thrust sheet ε = 0.76km, thrust sheets δ, β, γ = 1.7 km, thrust sheet α = 0.8 km.

Figure 3: Sequential restoration showing the proposed sequence of deformation. Active faults are indicatedby solid red lines. Red dashed lines show fault trajectory prior to displacement.


122

2-17: Quirk Creek anticline, Alberta, CanadaSteven Lingrey, ExxonMobil Upstream Research Co., Houston, Texas, U.S.A.

Location: Southern Alberta Foothills, CanadaTopics: Fault-bend foldingReserves: Gas nearby in Lower Carboniferous carbonate reservoirs

Figure 3: Fault-bend fold structural interpretation of detailed seismic image. Thrust fault trajectory is shown by dashed red line and its dotted projection offthe end of the seismic data. Fold axes are shown by dashed green lines; long dashes relate to changes between hanging-wall flats and ramps, short dashesrelate to secondary bends in the fault trajectory. Blue lines with arrowheads mark bedding orientations with cut-offs against the fault, blue lines with dashedends mark bedding roughly parallel to the fault. The isochrons between reflections mark an increase towards the anticlinal crest interpreted to be caused bysubseismic small thrust faults.

Figure 2: Detail view of time-migrated seismic image of hanging-wall ramp (cutoff) showing bedding reflection terminations against fault reflection. LowerPaleozoic terminations are clear between A-A� and A��-A��; Upper Paleozoic terminations between B-B� are off the end of the seismic line, but are visible inadjacent seismic data. The geometry of fold axes can be inferred in a fault-bend fold sense.

Figure 1: Location map for the seismic line showing position in inner Foothillsbetween Moose Mountain culmination and Quirk Creek gas field. Paleozoic out-crop is shaded blue.

The Foothills of the Canadian Rocky Mountains have provided many examplesof fold and thrust fault structural features considered to be characteristic ofthe detached contraction of sedimentary layering (Bally et al., 1966;Dahlstrom, 1970; Price, 1981; Boyer and Elliot, 1982). Surface exposures areaugmented by extensive seismic reflection profiling and by oil and gas drilling.Fault-related fold mechanisms (Suppe, 1983; Jamison, 1987; Suppe andMedwedeff, 1990) offer a means of more systematic analysis and prediction ofsubsurface geometry within the context of seismic and well control. Thesouthern Alberta Foothills are particularly well suited for this type of analysisbecause: 1) the area yields very good land seismic data, 2) well control isabundant (mature gas field drilling province), 3) well-documented and uncom-plicated pre-tectonic stratigraphic geometry (Cordilleran miogeoclinal plat-form strata situated east of the hingeline), and 4) an empirically constrainedsystem of bedding detachment horizons. In seismic data acquired over thenorthern Quirk Creek area (10 km south of Moose Mountain culmination;Figure 1) a clear example of a hanging-wall ramp (cutoff) is expressed by itstime-migrated reflection image (Figures 2, 3). The full uninterpreted seismicline is displayed in Figure 4; its interpretation is displayed in Figure 5.


123

Figure 4: Post-stack time-migrated 2-D seismic reflection profile across the northern Quirk Creek area. Display is scaled to be 1:1 at an average interval velocity of 4000 m/s. Three wells indicate key stratigraphic tops and positions of fault repetitions. Threestratigraphically calibrated zones of distinctive reflections guide interpretation away from the well control (yellow boxes): 1) a very high-amplitude continuous reflection event (doublet) characterizes the Jurassic-Lowermost Cretaceous Fernie-Kootenay zone imme-diately overlying the Lower Carboniferous (Mississippian) Rundle Group, 2) a high-amplitude sporadically continuous reflection commonly occurs just above the top of the Devonian Palliser Formation, and 3) a system of higher amplitude reflections, three to fourcycles long, indicates the Cambrian strata.

2-17: Quirk Creek


124

Figure 5: Post-stack time-migrated 2-D seismic reflection profile across the northern Quirk Creek area showing structural interpretation. Blue line marks top of the Rundle Group; dark blue line marks top of the Palliser Formation; pink line marks middle Devonianmarker near the top of the Cambrian. Lines are dashed where seismic reflection imaging becomes uncertain. Thrust fault trajectories are marked by heavy, dark red lines. Bedding and faults allow subdivision of the Paleozoic into three layers: 1) LowerCarboniferous (shaded blue), 2) Devonian (shaded dark blue, and 3) Cambrian (shaded pink). A semi-continuous set of high-amplitude reflections interpreted to be the base of the Cambrian reflection set below the basal decollement are shaded light orange. Theheavy, dashed orange line is the projected regional position for the base of the Cambrian assuming a flat, gentle (2–3 degrees) surface. The discrepancy between the two orange lines indicates a velocity anomaly that can be correlated to the number (net thickness)of repetitions of Paleozoic carbonates. Beneath the exploration well 2-23-21-6W5, the velocity “pull-up” effect reaches 600 m/s.

2-17: Quirk Creek


125

To a first approximation, the Fernie-Kootenay reflection eventseparates the Mesozoic rocks above, which show interval veloc-ities of 4000 m/s, from the Paleozoic rocks below, which showinterval velocities of 6000 m/s. Imbrication of fast rocks abovePaleozoic detachment horizons deflects (pulls up) the unfaultedsub-basal decollement reflection (lowest parts of the Cambrianreflections). The magnitude of pull-up can be matched to theexcess thickness of Paleozoic imbricates contained in the overallstructure. The profile expression of the anticline conforms tofault-bend fold theory in that its western limb dip is caused byfootwall ramps and its eastern limb dips are caused by hanging-wall ramps. Deeper elements show duplex-style deformation thatslightly modifies the overlying hanging-wall ramp. Conversion ofthe time image to depth allows the degree of conformance toideal fault bend fold theory to be measured (Figure 6). The hang-ing-wall ramp steps up from a lower detachment in the Cambrianto an upper detachment in the Fernie-Kootenay. The upperdetachment begins just past the eastern end of this seismic line;its position in the hanging-wall can be traced on other seismiclines to the east where it is observed to place the Jurassic-Cretaceous Fernie-Kootenay strata atop middle and UpperCretaceous strata. As the hanging-wall ramp crosses thePaleozoic strata, it flattens briefly at an intermediate detachmenthorizon located at the base of the Lower Carboniferous layer(Banff detachment). Displacement is larger than the horizontalextent of the hanging-wall ramp (> 5 km) so that the cutoffs fullyoverlie the upper detachment. The corresponding footwall rampmust exist west of the seismic data. Figure 7 shows a restorationof the depth profile and an idealized geometric model usingSuppe’s (1983) mathematical constraints of fault-bend folding.

2-17: Quirk Creek

Conclusions:• Fault-bend fold theory makes a good match with the Quirk

Creek anticline observed in seismic data.• Westerly dips arise from thrust sheet strata overlying foot-

wall ramps (cutoffs).• Easterly dips arise from rotation of the leading-edge cutoffs

(hanging-wall ramp) onto an upper detachment.• Footwall duplexing complicates, but does not obscure the

fault-bend fold.

Figure 6: Depth-converted profile of time-migrated seismic image. Time image velocity distortions due to lateral increase in fast Paleozoic rocks relative to slow Mesozoic rocksare removed. Basal fault trajectory consistently overlies Fernie-Kootenay strata and therefore is a footwall flat; the step in the central part of the fault is a bend caused by footwalldeformation (northern plunge end of Quirk Creek gas trap). A-A� and A��-A�� mark position of lower hanging-wall ramp and B-B� marks position of upper hanging-wall ramp.

Figure 7: Restoration and geometric modeling of fault-bend fold geometry. Top section (A) is a restoration of the depth profile using a flexural-slip mechanism. Restoration recovers theprimary flat-ramp-flat fault geometry. Lower three sections (B, C, D) show the initial, middle, and final states of a forward geometric model using Suppe’s (1983) fault-bend fold theory.The final deformation state nearly matches depth profile in Figure 6. Second order internal shortening of Devonian and Cambrian layers are ignored by the geometric model.


126

2-18: Imbricate fault-related folding, South Caribbean Basin, Colombia

Freddy Corredor, Department of Earth and Planetary Sciences, Harvard University, Cambridge, Massachusetts, U.S.A.Tomas Villamil1, Exploration Vice-President, Ecopetrol, Bogotá, Colombia

1 Present address: Lukoil Overseas Colombia, Ltd., Bogotá, D.C., ColombiaJohn H. Shaw, Department of Earth and Planetary Sciences, Harvard University, Cambridge, Massachusetts, U.S.A.

Location: South Caribbean Basin, offshore northern ColombiaTopics: Conventional and shear imbricate fault-bend folds

Figure 2: Uninterpreted, migrated, and depth converted 2-D seismic profile across the Fuerte Imbricate System in the South Caribbean Basin, offshore northern Colombia. Twostacked thrust sheets are imaged (see detailed description in Figure 4). We observe two structural patterns that are consistent with break-forward imbricate systems: A) The upperthrust fault appears folded by the underlying thrust sheet, and B) younger growth strata are folded above the frontal thrust sheet.

The South Caribbean basin represents an accretionary prism thatresulted from the transpressional collision between the Caribbean andSouth American plates during the Tertiary. An imbricate thrust systemin the southern portion of the basin is clearly imaged with 2-D seismicreflection data, with which we interpret fold and fault geometries andpatterns of growth sedimentation. We model this imbricate systemusing a combination of conventional and shear imbricate fault-relatedfolding theories (Suppe, 1983; Corredor et al., 2002; Suppe et al., 2003),and trishear kinematics (Erslev, 1991; Allmendinger, 1998). The pat-terns of growth sedimentation that can be observed in this imbricatesystem are used to further constrain the models.

Figure 1:Regional topogra-phy, bathymetry,and tectonic ele-ments of Colombiaand location of theSeismic line (1)used for this study.

Seismic data courtesy of ECOPETROL


127

The Fuerte Imbricate System is located in the Southwest Caribbean Basin, beneath a largedeltaic system. This deltaic system was fed primarily by the paleo-Atrato, Sinú, andMagdalena rivers. The northern boundary of the imbricate system is the east-west trend-ing Canoas fault. The western limit corresponds to the South Caribbean Deformationfront. The deformation in this imbricate system began late during the Miocene(?) andcontinues through the present day, resulting in its pronounced bathymetry expression(Figure 3). The thrust sheets are composed of Miocene marine shales and turbidite sands(Potential Reservoirs). This imbricate system is detached at the bottom of anOligocene(?) fine-grained section, which is a potential source rock. This system pre-serves growth strata that records fault and fold kinematics. These growth sediments aredeposited in piggy-back basins formed over the backlimbs of individual imbricates andas onlapping sequences against the forelimbs (Figures 2 and 3). The piggy-back stratig-raphy consist of distal marine, fine-grained sediments and condense sections. In theuppermost portion of the seismic profile, a spectacular Pleistocene prograding deltaicsequence can be observed.

2-18: Imbricate fault-related folding — Colombia

Figure 3: High resolution sea floor bathymetric image of a region north of the Fuerte Imbricate system interpreted in this contribution,and regional map showing the location of (1) the bathymetric image, and (2) the seismic line across the Fuerte imbricate system. Theridges on the southern portion of the image (3) represent northeast-trending thrust-related folds that are actively deforming the sea floorand controlling the course of meandering turbidity channels (4). The low regions between ridges (5) correspond to the piggy-back basinsformed above the backlimbs of individual fault imbricates. On the upper right corner (6), the southern limit of the Magdalena delta systemis burying these active folds and faults. The limit between these two systems corresponds to the Canoas Fault (7).

The stratigraphic sequences imaged in the seismic profile (Figure 4) correspond to Tertiarymarine and deltaic sediments. At the bottom of the section an Oligocene(?) sequence is inter-preted (3), and is composed of thick deep marine shale sequences (potential source rocks), andmay contain some interbedded turbidite sands (potential reservoirs in deep water environ-ments). On seismic sections, this sequence is generally devoid of internal reflections. This for-mation is interpreted to correspond to a weak decollement layer that undergoes an externallyimposed shear deformation in this imbricate system. Seismic reflections beneath this sequenceare generally continuous laterally, suggesting that the decollement for this system is located atthe bottom of the Oligocene(?) shale. The Oligocene(?) sequence is overlaid by Miocene-Plioceneinterbedded shallow marine shales and sandstones (4) that produce seismic reflections withhigher amplitudes and lower frequencies. In the uppermost portion of the seismic profile, a spec-tacular Pleistocene prograding deltaic sequence can be observed (5). This deltaic sequence isnot folded by the underlying imbricate system constraining the age for the end of deformation inthese particular thrust sheets. Fault plane reflections and cutoffs are clearly observed (redarrows) that constrain the geometry of both thrust faults. The upper thrust fault is folded by thelower fault suggesting a break-forward sequence of imbrication. Break-forward imbricationresults from a new thrust being developed in the footwall of what was previously the activethrust.

Figure 4: 2-D post-stack migrated and depth-converted seismic section through the imbricate system interpreted in this contribution showing someimportant characteristics including: (1) Sea floor reflection, (2) Growth sedimentation, and thrust faults defined by fault-plane reflections and cutoff(red arrows). Notice how the upper thrust fault is folded across the syncline by the lower thrust sheet.


128

2-18: Imbricate fault-related folding — ColombiaFolding vectors describe folding shear strains, which are angular measuresof the change in orientations of beds and faults across a fold limb or kinkband (see section 1B-5). Thus, folding vectors can describe the refolding ofoverlying thrust sheets due to imbrication (Shaw et al., 1999; see section 1B-5). As mentioned earlier, we observe two structural patterns in the seismicimage that are consistent with this being a break-forward imbricate system:A) The upper thrust fault appears folded by the underlying thrust sheet,and B) younger growth strata are folded above the frontal thrust sheet.Using folding vectors, we test the idea that the shallow thrust sheet (redarrows in Figure 5A) is folded by a deeper thrust.

Slip and shear on the deep thrust has produced multiple kink bands thatshould have refolded the overlying thrust sheet if this is a break-forwardsystem. Hence, the orientation of the shallow thrust, and beds in its hang-ing wall, should change as the thrust sheet passes over the underlying kinkbands. The folding vector method is used to predict the amount of deflec-tion of the shallow thrust as it is refolded by the two underlying kink bands,which are bounded by axial surfaces (A-A�; B-B�). If the predicted shape ofthe fault is consistent with the observed fault shape, it will confirm that thisis a break-forward imbricate system.

The deflections of bedding across the deep kink bands are used to deter-mine the folding vectors (U and V). Folding vectors are measured parallel toaxial surface orientations. In a break-forward system, folding vectors U andV should be equal to the deflections of the shallow thrust described by vec-tors X and Y, respectively. (Note: This method is based on conservation ofshear, and hence line length, parallel to the axial surface orientation). Thethrust fault on the right side of the section, before entering the kink band A-A�, has a dip of approximately 10°. The folding vector U is measured as thedeflection of the light blue layer (Figure 5B) across axial surface A in thefootwall of the thrust fault. The folding vector U is then used to predict thedeflection (X) of the shallow thrust fault (U = X). The predicted dip of thefolded fault above kink band A-A� is 19°, consistent with the dip of the faultplane observed in Figure 5A. Moving to the left, the thrust fault next entersthe kink band B-B� at its dip of 19°. The folding vector V is measured as thedeflection of the light blue layer across the axial surface B, and is used topredict the deflection (Y) of the shallow thrust fault across the kink band B-B� (V = Y). The predicted dip (31.5°) of the refolded fault above kink band B-B� is also consistent with the dip of the fault plane seismic reflection. Thisconfirms that these thrust sheets form a break-forward imbricate system.

Figure 5. Uninterpreted (A) and interpreted (B) close-up view of the Fuerte Imbricate System, offshore northern Colombia, toillustrate how folding vectors (see section 1B-5) are used to interpret this break-forward imbricate system.


129


This imbricate system can be modelled using combined conventional and shear imbricate fault-bendfolding theories. Using folding vectors we have shown on the previous page (Figure 5) that this systemhas a break-forward sequence of imbrication. Additionally, we observe three structural patterns thatsuggest shear fault-bend folding (see section 1B-4) is an important mechanism in this imbricate system(Figure 6). The thrust sheets show gentle back-limbs that dip less than the fault ramps, growth sedi-ments show evidences of limb rotation, and a broad anticline in the shallow thrust sheet overlies a syn-clinal bend on the thrust fault. In simple shear fault-bend folding, the weak decollement layer(Oligocene?) at the base of fault ramps undergoes an externally imposed bedding-parallel simple shear.The total slip produced by the shear is accommodated by increasing slip along the fault ramps, and byrotation along the back-limbs. Further frontal imbrication and the transfer of shear produce a decreasein the ramp and bedding angles in younger and shallower thrust faults (Figure 7), occasionally produc-ing folds not directly related to a fault-bend.

Figure 7: Forward model of a break-forward sequence of imbrication by forward distributed transfer of shear showing theresulting patterns of growth sedimentation. Imbricate fault-bend fold theory describes refolding of shallow thrust sheets byyounger and deeper faults. A) an incipient thrust. B) and C) a simple shear fault-bend fold grows by increasing simpleshear across the weak decollement layer. The growth strata show evidences of limb rotation and kink band migration. D-E)a frontal thrust sheet is formed by simple shear fault-bend folding. This additional shear produces forward (counterclock-wise) rotation of the shallower and younger thrust sheet, effectively decreasing the dip values of the fault ramp and foldingbedding. A portion of the flat crest of the fold rotates forward, forming a forelimb with no associated fault-bend comparableto that observed in Figure 6. The growth sediments deposited in the earlier stages are also refolded by the younger thrust.

Figure 6: 2-D depth-converted seismic section through the imbricate system interpreted in this contribution showing the characteristics that suggest this systeminvolves shear imbricate fault-bend folding: (1) Backlimbs dip less than fault ramps, (2) Small forelimbs compared to backlimbs, (3) Growth sedimentation show evi-dences of limb rotation, and (4) Anticline is underlaid by a synclinal bend (5) in the associated thrust fault. Notice also that the upper thrust fault is folded across thesyncline by the lower thrust sheet, as described on the previous page.

Part 2X: Imbricate Systems — offshore Colombia

Incipient thrust Fault


130


Figure 8: Interpreted depth converted seismic pro-file across the Fuerte Imbricate System, combiningthe results from the structural analyses presentedon the preceding pages. The stratigraphic correla-tion is based on regional interpretation of seismicfacies across a large seismic survey. The complexgeometry of the lower thrust ramp is interpreted toresult from an externally imposed simple shear inits footwall (Shear Profile 2) caused by thrustsheets that lie to the northwest of this image. Theshallower thrust sheet shows an anticline in theright portion of the section, which is underlaid by asynclinal fault bend. This relationship is interpretedas the result of forward rotation of the bedding andthe thrust fault due to the simple shear fault-bendfolding in the underlying thrust sheet (Shear Profile1). Shear faultbend folding is consistent withgrowth sediments showing evidence of limb rota-tion, and with back limbs dipping less than thefault ramps. Shearing in the lower thrust sheet hasalso refolded the shallower thrust fault and the bedsin its hanging wall, indicating that this is a break-forward imbricate system.

Conclusions:• The Fuerte structure is a break-forward, shear imbricate fault-bend fold system in the southern Caribbean basin, offshore northern Colombia.• Folding vectors are used to interpret the thrusting sequence.• Several characteristics allow the interpretation of shear imbrication in this system, including: A) beds on the backlimbs dip less than fault ramps, B) growth sediments show evidences of both

limb rotation and kink-band migration, and C) an anticline in the shallower thrust sheet is underlied by a synclinal bend in the associated thrust fault.• The Fuerte structure was active early during the Miocene-Pliocene with thrust faults emerging in the sea floor. Further foreland thrusting has sheared this system and passively transported it for-

ward along an Oligocene(?) basal detachment during the Pliocene to the present time.


131

2-19: Oligocene fold belt, western Gulf of Mexico, U.S.A.Thomas W. Bjerstedt1, ChevronTexaco Exploration, Bellaire, Texas, U.S.A.

1Present address: U.S. Department of the Interior, Minerals Management Service, New Orleans, Louisiana, U.S.A.

Location: Western Gulf of Mexico; Port Isabel and Alaminos Canyon OCS areas

Topics: Detachment anticlines among salt-withdrawl mini-basins, syndepositional tectonism

Reserves: Rank exploration area

Figure 1: (Left) W-E schematic cross-section along Oligocene depositionaldip from Texas outcrop belt to deepwa-ter. Updip fluvio-deltaic and deltaic sys-tems transitioned down dip to slopeand basin floor fans in Port Isabel andAlaminos Canyon OCS areas.Exploration targets are detachment anti-clines cored by upper Oligocene sand-prone fan systems in the Frio. Saltdiapirism was contemporaneous withlate Oligocene deposition and alsoinfluenced facies patterns.

Figure 3: (Left) Map showing the general location of theOligocene fold belt and the Nickoli structure in AlaminosCanyon OCS blocks 51 and 52. Dots show location of1996 exploration wells in Port Isabel OCS area.

Figure 2: (Above) SW-NE cross section through 3-D post-stack seismic volume. The location of the arbitrary line is shown in Figure 5. Shownare over-thickened anticline core (1), backlimb erosional unconformity, and main growth phase of Nickoli structure (2). No vertical exaggeration.

The Oligocene fold belt is a series of NE-SW–trending detachment anticlinesin the western Gulf of Mexico. Detachment folds are down-dip elements of acoupled extensional to compressional transition. Deposition updip of upperOligocene Frio fluvio-deltaic systems on the Texas coast loaded prodeltaand slope environments to the E–SE (Figure 1). Detachment anticlines arecommon at terminations of low-angle thrusts that cut early Oligocene sec-tion. Nickoli is a well-imaged example of such a fold in this structural trend(Figure 2). Two exploration wells were drilled in southern Port Isabel OCSarea in 1996 (Figure 3). One well lacked a reservoir and the other lackedcharge. 3-D basin modeling suggests that the northern part of the fold-belttrend has more favorable timing relationships for hydrocarbon generation,reservoir deposition, and trap and seal formation.

water bottom

erosional truncation


132

2-19: Oligocene fold belt, western Gulf of Mexico, U.S.A.

Figure 4: Annotated seismic showing regional detachment surface and detachment anticline formed by thrust faults, (1) overthickend core of regressive Frio sand-prone deposystems(middle to late Oligocene), and erosional unconformity at the top of the Frio, and (2) onlap of Frio by Anahuac transgressive shale-prone deposystems (latest Oligocene). Early Miocenedeposits drape the fold until completely buried.

A detachment anticline forms Nickoli, a typical Oligocene fold belt structure. A deep seateddetachment surface, probably late Eocene, is mappable on 2-D regional seismic throughout thefold belt. The oldest thrusts cut the early Oligocene section. The history of the Nickoli fold includ-ed low-angle thrusting that produced an overthickened core, rotation of the thrust backlimb anderosion, the main growth phase, onlap by younger deposits, draping, and eventual burial. Thestratigraphic section indicated by (1) in Figure 4 is the overthickened core of the detachment anti-cline. The section indicated by (2) onlaps the eroded surface and indicates that the main growthphase of Nickoli fold began at the horizon interpreted to be near top of the late Oligocene Friointerval. Frio deposystems are expected to be sand-prone, low-stand basin floor fans.Transgressive, shale-prone, high-stand systems overlying the Frio are recognized in the westernGulf of Mexico as the latest Oligocene Anahuac interval. The youngest thrust was detached alongthe top Frio erosional surface, possibly within condensed zone(s) at the sequence boundary.Early Miocene deposystems that drape Nickoli and other fold belt detachment structures areexpected to be shale prone as sand was diverted into lows until folds were buried completely.Figures 5 and 6 show mapped structure and seismic time slice.

Figure 6: Showing about the same area as Figure 5. Seismic time slice near 3,050 m (10,000 ft) showingstructural configuration and position of the youngest thrust fault.

Figure 5: Structure contour map on the top Anahuac (Figure 4 labeled surface). The map contourinterval = 500 ft and the grid lines represent OCS blocks. The buttressing effect of the salt diapir tothe west rotated the Nickoli anticlinal axis to an east-west orientation. Salt withdrawl from diapirsthat surround the Nickoli mini-basin overprinted the basin-center structures with normal faults.

water bottom

Figure 5 mapped surface

regional detachment surface


133

2-20: Edge-Sigsbee Folds, Gulf of Mexico, U.S.A.Robert J. Alexander1, Thomas W. Bjerstedt2, and Sharon L. Moate3, ChevronTexaco, Bellaire, Texas, U.S.A.

1Present address: BHP Billiton, Houston, Texas, U.S.A.; 2Present address: Texaco Minerals Management Service, New Orleans, Louisiana, U.S.A; 3Present address: Consultant, Bellaire, Texas, U.S.A.

Location: Western Gulf of Mexico, Keathley Canyon OCS areaTopics: Detachment anticlines, Sigsbee Escarpment Reserves: Rank exploration area

Figure 1: (Above) 3-D pre-stack depth-migrated seismic line down the axis of KeathleyCanyon and showing a fold and the leading edge of the Sigsbee salt canopy (light blue).

Figure 2: (Right) Bathymetry of the deepwater Gulf of Mexico showing the location ofKeathley Canyon, the Sigsbee Escarpment and the seismic line. Red color is ~600' (183 m) and blue color is ~9000' (2744 m) water depth.

Low-relief folds at the edge of the Sigsbee Escarpment(toe of the bathymetric slope) have prompted questionsabout their origin and seismic imaging. There are twolikely hypotheses for how these structures formed: 1)they are not “real,” and their seismic expression resultsfrom velocity anomalies caused by abrupt changes inthickness of the salt canopy at the edge of the Sigsbeeand rapid water deepening onto the abyssal plain; and 2)they are formed by a combination of lateral compression,detachment folding, and isostatic response to salt sheetand sediment loading, (i.e., foreland bulge analogy). Avelocity anomaly origin (hypothesis #1) is tenuous wherethese folds are continuous inboard and outboard of thesalt edge. Furthermore, these folds do not occur every-where along the salt edge and Sigsbee Escarpment. Anorigin due to gravity-gliding and lateral compression withdown-dip lateral shear on a Paleogene detachment sur-face (hypothesis #2) is more likely. Good quality 3-D pre-stack depth-migrated seismic data in the salt canopy re-entrant of Keathley Canyon show that a low-relief edge-Sigsbee anticlinal structure is cored by a subtle, incipientimbricate duplex which thickens the section, arching theoverlying sediments into a probable detachment anti-cline. There is seismic evidence for coupling of the later-ally prograding salt canopy with subjacent sediments.Compression at the Sigsbee Escarpment (toe of theslope) is caused by salt and sediment movement towardthe abyssal plain, which forms a duplex of thrusts abovean over-pressured basal detachment surface. The overly-ing sediments arch in response to this structural thick-ening as is seen in foreland “triangle zones.” Although wedon’t offer this explanation for all edge-Sigsbee struc-tures, the interpretation should be considered when well-imaged pre-stack depth-migrated seismic data is avail-able to test hypotheses.


134

2-20: Edge-Sigsbee Folds, Gulf of Mexico, U.S.A.

Figure 3: Velocity model along annotated seismic line showing velocity inversion and inferred over-pressured section;A) Normal model, B) High velocities are clipped to better show a subtle velocity inversion in the Tertiary age section. InFigure 3A, red color is ~14700 ft/s (4480 m/s) and blue color is ~4980 ft/s (1518 m/s). Datum is near base Paleogene.

Figure 4: Enlarged inset from Figure 5. Note duplex structure in the sediments below the base of salt (blue). Compressionoccurs where the base of salt is “stepped,” which confirms some degree of coupling with subjacent sedimentary units.

Figure 5: Interpreted and annotated seismic line showing a low-relief, detachment anticline in subsalt sediments at the edge of the Sigsbee Escarpment. Alsoshown are interpreted faults, and a late Paleogene shale unit that is interpreted as a detachment surface at the base of an inferred over-pressured zone. Velocityanalysis modeling for pre-stack depth migration and pore pressure prediction analysis identified the velocity inversion, which probably continues further north-ward under the salt than is shown in Figure 3.

Conclusion:Detachment anticlines with incipient duplex structures can form the cores of low-relief, four-way closures at theedge of the Sigsbee Escarpment, western Gulf of Mexico. Our model invokes a detachment surface in an over-pressured section coupled with south-directed compression due to gravity gliding of the overlying salt+sedimentload. The suprajacent section is arched by the developing duplex structure in the core of the fold (i.e., betweenpurple and orange lines). There is no evidence for compressional deformation south of this seismic line (leftside) indicating this is analogous to a triangle zone in the foreland of fold and thrust belts.

Velocity modeling of 3-D pre-stack depth-migrated seismic line throughKeathley Canyon shows a velocity inversion and inferred over-pressuredzone. A detachment surface occurs at the base of a late Paleogene shale unit.Edge-Sigsbee detachment anticlines are formed by lateral compression of thesedimentary section as the salt canopy expands onto the abyssal plain gen-erating an over-pressured zone of low strength.


135

We present seismic examples that illustrate basin inversion processes inthe Inner California Borderland (ICB), offshore southern California. The ICBdeveloped by Miocene crustal-scale extension dominated by a pair ofregional low-angle detachments (Crouch and Suppe, 1993; Nicholson et al.,1993). During the Pliocene, the onset of the modern transpressional regimegenerated several large contractional fault-related fold trends spatially con-trolled by reactivation of these detachments (Rivero et al., 2000). We use adense grid of industry seismic reflection profiles (Figure 1), and fault-relat-ed fold theories to analyze an anticline structure within these trends. TheSan Mateo anticline developed by the upward propagation of reverse slipduring the inversion of Miocene half-grabens. Based on the analyses ofkink-band panels, and growth and pre-growth sequences, we propose astructural interpretation for this fold consistent with seismic and well data.In addition, the structural interpretation provides insight into the kinemat-ics of the basin inversion processes.

Figure 1: Location ofthe San Mateo trend inthe Inner CaliforniaBorderlands. The studyarea is defined by thegrid of seismic reflec-tion data. Yellow cir-cles are well locations.L.A. = Los AngelesBasin.

2-21: Fault–related folding in reactivated offshore basins, CaliforniaCarlos Rivero and John H. Shaw, Department of Earth and

Planetary Sciences, HarvardUniversity, Cambridge,Massachusetts, U.S.A.

Location: Inner California Borderland, California, U.S.A.Topics: Basin inversion fault reactivation, fault-bend folding, structural

wedges, blind-thrust

Figure 2: Time-migrated seismic reflection profile across the San Mateo Anticline. Note the regional oceanside detachment (1) extending beneath the San MateoAnticline (2). This detachment is not folded by the contractional structures; thus we interpret that the San Mateo Anticline is formed by thrusting ramping up fromthis detachment surface. A preserved extensional rollover is also visible on the left-hand side of the section (3).

Figure 3: Time-migrated seismic reflection profile across the San Mateo Anticline. Stratigraphic tops are correlated from the San Clemente CH–1 well. Well-illumi-nated cutoffs and fault plane reflections (1) constrain the location of a reactivated normal fault and overlying thrust ramps that form the San Mateo structure. Note theMiocene syn-rift section penetrated by the well that expands toward the normal fault in a rollover structure. The rollover structure shows evidences of bivergent tec-tonic inversion or “bipolar extrusion” (Copper and Williams, 1989; Hayward and Graham, 1989), with both fore (1) and backthrust (2) anticlines developed by theinversion. Gently dipping continuous reflections and three-dimensional mapping define the location of the Oceanside Thrust, one of the regional Miocene detach-ments reactivated in the Pliocene (3). Syn-extensional deposits and unconformities define the presence of other normal faults (4) that are not inverted.


136

We interpret the San Mateo Anticline as an imbricated fault-bend fold produced by theupward propagation of contractional slip from an inverted normal fault into multipledetachment levels (Figures 4, 5, and 6). The backlimb geometry of the anticline exhibit-ing multiple dip-domains indicates the presence of a deeper structure. This sub-thruststructure refolds the shallow thrust sheet of the San Mateo Anticline in a way consistentwith a break-forward system (Figure 6). The thrust front terminations of the San Mateothrust and the underlying thrust are defined by two structural wedges that propagateslip back to the hinterland. At this location, the interaction between the synclinal axialsurfaces of the upper detachments produces a complex geometry of the thrust front.

2-21: Fault-related folding in reactivated offshore basins, CaliforniaSeismic interpretation

Figure 4: Balanced structural interpretation of the San Mateo Anticline. The interpretation highlights the relationship of the contractional structure to pre-existing normal faults reactivated during the phase of basin inversion. Kink-band domains in the back-limb ofthe anticline, and direct fault plane reflections constrain the geometry of the San Mateo thrust from shallow to deeper levels, where it is linked to an older normal fault. The seismic image also indicates that the San Mateo ramp is refolded by a younger, deeper fault.We interpret that both thrust faults terminate in structural wedges, as no foreland structures that could account for the transfer of slip exist beyond the San Mateo anticline. Formation tops from the well San Clemente #1. Labeled axial surfaces correspond to thosemodeled in Figure 6.

Figure 5: Restoration of the proposed structural interpretation for the San Mateo anticline to the top of the PlioceneRepetto Formation. The restoration highlights the role of the extensional system controlling the geometry of theMiocene depocenters, and locating the Pliocene compressional structures. Estimated total shortening is 2.5 km.

Stage 1: The San Mateo Thrust forms andslip produces the kink-bands A-A� and B-B�that define the shallow San MateoAnticline. A structural wedge in the thrusttermination of the upper detachmenttransfers slip back to the hinterland.

Stage 2: Development of the sub-thruststructure with minor displacement thatgenerates incipient kink-band C-C�. A lowerstructural wedge is also formed in thethrust front position, analogous to SanMateo thrust.

Stage 3: Final configuration of the imbri-cated system, consistent with a break-for-ward model (Shaw et al., 1999). Displace-ment on the deeper thrust refolds the SanMateo thrust sheet, and forms a structuralwedge (triangle zone) at the western bor-der of the fold belt.

Figure 6: Balanced sequential model of the development of the San Mateo structure.


137

Conclusions:• The San Mateo anticline is an imbricated fault-bend fold origi-

nated by basin inversion processes. The San Mateo thrustreactivated a segment of a northeast-dipping Miocene normalfault.

• The phase of basin inversion also reactivated a Miocene low-angle detachment as the oceanside thrust. The oceansidethrust transferred contractional slip to associated syntheticand antithetic normal structures, inverting a major graben-boundary fault, and generating a regional structural wedgedefined by the oceanside thrust and a backthrust zone. Thisstructural wedge controls the location of a prominent mono-cline with bathymetric expression.

• The structural style varies considerably across these invertedbasins. In some areas, the pre-inversion geometry of theMiocene basins has not been modified, as it is expressed inwell-developed rollovers preserved in the hanging wall of low-angle and high-angle normal faults. In contrast, uplifting andfolding of the sedimentary fill, and reactivation of half-gra-bens, document the later phase of basin inversion. Footwalland hanging wall short-cuts associated with reverse andthrust faults are also documented by the seismic data.

2-21: Fault-related folding in reactivated offshore basins, California3-D modeling

Figure 7: (a) Oblique view of a three-dimensional model incorporating a representation of the San Mateo thrust, and the structural wedge defined by the Oceansidethrust and the back thrust fault. Seismic image corresponds with profile Y-Y� shown in Figure 3. The blue surface is the top of the syn-rift sequence (MonterreyFormation). (b) Seismic dataset used in the definition of the surfaces shown in the 3-D model. Contours represent bathymetry of the seafloor. (c) Same view as in (a)with the seismic image removed to highlight the lateral continuity of the structural wedge, as well as the contractional folding of the Monterrey Formation.

Integration of seismic data, fold-related folding theories, and3-D visualization techniques are used to illustrate the com-plex of basin reactivation along the San Mateo trend.Modeling of the shallow and deep structural styles, repre-sented by the San Mateo thrust, the backthrust, and the basalOceanside thrust highlights the role of the regional structuralwedge in generating the contractional foreland and hinter-land-directed structures.

San Mateo Anticline

San Mateo Thrust

Back-Thrust

Oceanside Thrust

Anticline trends associated with the Inversion


138

2-22: Coalinga anticline, San Joaquin basin, California, U.S.A.Chris A. Guzofski, Department of Earth and Planetary Sciences, Harvard University, Cambridge, Massachusetts, U.S.A.John H. Shaw, Department of Earth and Planetary Sciences, Harvard University, Cambridge, Massachusetts, U.S.A.

Location: Western San Joaquin basin, California, U.S.A.Topics: Structural wedges, backthrusting, imbricationReserves: 906 Mbbl of hydrocarbons from lower to middle

Miocene Temblor Formation

Figure 2: Migrated and depth-converted seismic profile with several wells showing formation tops across the Coalinga anticline. The structural relief between the synclines (1and 2) bounding the anticline (3) provides evidence that one or more southwest-dipping thrust ramps underlie the structure. The asymmetry of the central anticline (3) demon-strates that an additional northeast-dipping thrust ramp underlies the structural crest. The absence of Tertiary deformation east of the Coalinga anticline (beyond this section)provides evidence that the southwest-dipping fault ramp or detachment does not extend basinward of the Coalinga anticline. This argues for the presence of a structural wedge,where slip is sent back to the hinterland on the inferred backthrust beneath the Coalinga structure.

Figure 1: Landsat TM image of the Coalinga anticline showing the locations of theseismic lines used in this study. Locations of wells 1. Pleasant-Valley #1; 2. Leavitt-Hintze #1; and 3. PVF-11X are shown from Meltzer (1989) and Bartow (1990).

The Coalinga anticline is located in an active fold and thrust belt inthe western San Joaquin basin, California (Figure 1) (Namson andDavis, 1988; Stein and Ekström, 1992). At the surface the Coalingastructure is expressed as a southwest-vergent anticline that is definedby a narrow forelimb with a broad backlimb (Figure 2). We interpretthat this structure developed above a northeast-dipping thrust ramp.At depth, the anticline is northeast-vergent and structural relief acrossthe anticline provides evidence that a deeper, southwest-dippingramp has uplifted the anticline. Herein, we use these observations tointerpret this structure as a structural wedge, having grown throughmultiple stages of fault-bend folding.


139

Figure 3: Balanced structural interpretation of the Coalinga anticline, in which several imbricated faults generate the main fold. Slip on the Coalinga and San Joaquin ramps generates twoanticlinal fault-bend folds, where slip is sent back to the hinterland on folded backthrusts. The width of the forelimb of the Coalinga structural wedge is constrained by a pair of axial surfaces(B0 and B1), where the wedge tip is pinned by the active synclinal axial surface (B0). Imbrication of two older and shallower thrusts by the Coalinga wedge is demonstrated by the “capture” offold limbs associated with these older faults by the Coalinga structural wedge. The forelimb of one of these older structures is constrained by axial surfaces C0 and C1. Growth strata within thiskink band indicate that slip on its causative fault occurred at some point between the deposition of the Moreno shale (~ 65 Ma), and the deposition of the Kreyenhagen shale (~ 37 Ma), clearlybefore the development of the broad limb (B0-B1) that refolds it. A shallow thrust that branches off the main detachment generates the prominent anticlinal fault-bend fold defined by kinkbands A0 and A1. The dip of this thrust ramp was determined based on the forelimb dip using fault-bend folding theory. However, the observation that the backlimb dips less than the faultramp suggests that the backlimb is deforming by shear fault-bend folding mechanisms (see section 1B-4). Deformation of a shear band pinned by axial surface A2 (shaded yellow), leads to aminor rotation of the backlimb. The location of a regional angular unconformity is shown by yellow arrows. The axis of this unconformity (i.e. where rocks change from horizontal to dipping)is shown by a dashed yellow line. Formation depths are from Bartow (1990).

2-22: Coalinga anticline

We interpret the Coalinga anticline as being comprised pri-marily of a stack of imbricated structural wedges (Figure 3).Two structural wedges, with separate dipping forethrustramps and a common upper detachment surface forming abackthrust, generate the gross morphology of the Coalingaanticline (Figures 3 and 4). The structural relief across theCoalinga anticline (between Pleasant Valley and the SanJoaquin basin; Figure 3) is due to the accumulation of slipand uplift on the Coalinga thrust ramp, where the tip of thewedge is pinned by an active synclinal axial surface (B0).Similarly, structural relief across fold A4 to A5 in the SanJoaquin basin is due to slip on the San Joaquin thrust ramp(Figure 4). The prominent forelimb of the Coalinga anticline(defined by A0 to A1 in Figure 3) records slip on a fault thathas branched off of the upper detachment surface.

Figure 4: Migrated seismic line showing the location of the active synclinal axialsurface (A4) in the San Joaquin basin, which is used to constrain the tip of thestructural wedge associated with the San Joaquin thrust ramp (inset).


140

Here we present a balanced sequential model of the development of the Coalinga structure (Figure 5).

2-22: Coalinga anticline

Conclusions:• The Coalinga structure is underlain by two independent southwest-dipping thrust ramps that generate two structural wedges that sole

into a common backthrust/roof thrust.• The Coalinga structural wedge refolds at least two older structures, a Tertiary structural wedge with well-defined growth strata and a

younger southwest-vergent anticline that has accumulated slip from the San Joaquin thrust ramp. While this southwest-vergent anticlineis the most prominent surface feature of the Coalinga structure, slip on the two underlying thrust ramps produce the deeper fold archi-tecture.

Figure 5A: Initial geometry of theCretaceus (and older?) sedimentarysequence beneath the Coalingastructure. The development of a dip-ping panel beneath the shallowangular unconformity (shown byyellow arrows) is possibly related toslip on an unimaged fault (shownby queried red dashed line) thatsteps up to a local detachment.

Figure 5D: Slip on the Coalingathrust ramp generates a structuralwedge (whose forelimb is definedby axial surfaces B0- B1) that cap-tures and refolds the kink bandsassociated with the deformationmodeled in panels B and C. Thebackthrusts of the Coalinga wedgeand the San Joaquin wedge mergeat the regional detachment, as thesummed slip is sent back to thehinterland.

Figure 5B: Initiation of slip(between 65 and 37 Ma) on theCoalinga ramp makes a structuralwedge involving a detachment and abackthrust. Slip generates an anti-cline above the backthrust, withgrowth triangles associated withboth the forelimb of the wedge(defined by axial surfaces C0-C1)and the forelimb above the upperdetachment/backthrust.

Figure 5C: Initiation of the SanJoaquin thrust ramp and develop-ment of the southwest-vergent anti-cline (defined by axial surfaces A0-A1) by slip on the backthrust associ-ated with the nascent San Joaquinstructural wedge. The low angle ofthe backlimb of the southwest-verg-ing anticline, relative to the underly-ing ramp, is due to simple shearfault-bend folding where �e = 61°(see section 1B-4).


141

2-23: Wedge structure, Nias Basin, Sumatra, IndonesiaPeter A. Brennan, Tellumetrics LLC, Sugar Land, Texas, U.S.A.John H. Shaw, Department of Earth and Planetary Sciences, Harvard University, Cambridge, Massachusetts, U.S.A.

Location: Nias basin, Sumatra, IndonesiaTopics: Structural wedge, growth structure, inversion

We describe a complex structure located in the Nias basin, which lies between NiasIsland and the southwest coast of Sumatra (Figure 1). Nias Island lies along the present-day plate margin between the Indonesian and Indian Ocean plates. The basin containsMiocene and younger sedimentary rocks deposited over a basement composed of anearlier Tertiary subduction complex. The basin underwent a period of extension duringthe early and middle Miocene, and a subsequent phase of contraction during the lateMiocene, Pliocene, and Pleistocene. The structure we describe has been structurallyinverted, such that it reflects both extensional and contractional components.

We interpret the structural geometry and kinematics of this anticline using patterns ofsyntectonic growth strata, structural relief, and fault cutoffs (Figure 2).

Figure 1: Map of Sumatra showing the location of the Nias basin.

Figure 2: Post-stack, time-migrated and depth converted seismic reflection profile of the Nias basin that images a contractional fault-related fold. The structureis composed of a monoclinal fold limb that is underlain by a fault, which appears to offset basement and uplift the southern portion of the fold. Two distinctstratigraphic sections (1 and 2) thicken to the north across the fold limb, suggesting that they are syntectonic (growth) strata.


142

Figure 4: Seismic reflection profile across the Nias basin structure with: A) basic interpretations of the faultposition, axial surface traces, and growth sections; and B) structural interpretation. The geometries of the syn-clinal and anticlinal axial surface resemble those in Figure 3B, indicating that the structure is a decollementwedge. In this interpretation, the synclinal axial surface is pinned to the wedge tip, and the backthrust generallyconforms to the bed dips in its hanging wall. This interpretations is consistent with the pattern of growth section1, but does not yet explain the origins of growth section 2.

We propose a structural wedge model (Figure 3) to explain the geometry of the structureand the pattern of growth section 1 (Figure 4).

Based on interpreted fault cut-offs and structural relief, the main thrust ramp beneath theNias structure dips to the south (Figure 4A) indicating that the monoclinal fold limb is aforelimb. The forelimb is bound by a roughly linear synclinal axial surface that extendsupward through growth section 1, and by a curved anticlinal axial surface that has dif-ferent orientations in growth and pre-growth sections (Figure 4A). Based on this axial sur-face pattern, we interpret the forelimb as a growth structure developed by kink-bandmigration, with an active synclinal axial surface and an inactive anticlinal axial surface(see section 1A-3). Given the fault dip direction, this growth patterns is inconsisitent witha simple forelimb fault-bend fold model (Figure 3A), but consistent with a decollementwedge model (Figure 3B). Thus, we interpret the structure as a decollement wedge(Figure 4B).

2-23: Structural interpretation and growth section 1

Kinematic models

A: anticlinal fault-bend fold B: decollement wedge

Figure 3: Balanced kinematic models of an anticlinal fault-bend fold (A) and decollement wedge (B). In model A, the fault-bend fold isdeveloped above a ramp that flattens to an upper decollement. The anticlinal axial surface is active, and thus linearly extends through pre-growth and growth sections. The synclinal axial surface is inactive, and thus changes orientation at the boundary between growth and pre-growth section (see section 1B-1). In model B, slip on the upper detachment is transferred to a backthrust forming a structural wedge(Medwedeff, 1989) (see section 1B-6). The synclinal axial surface is active and the anticlinal axial surface is inactive, in contrast to model A.Thus, simple anticlinal fault-bend folds and structural wedges can be readily distinguished based on patterns of growth strata.

A: Seismic Example: Sumatra, Indonesia

B: Interpreted section


143

2-23: Structural interpretation and growth section 2

The structural wedge interpretation presented on the previous page (Figure 4B)explained the pattern of growth section 1, but did not address the origin and pattern ofgrowth section 2. Two scenarios may explain this older growth structure. First, the oldergrowth structure may reflect an early phase of contractional folding above the fore-thrust, developing as a structural wedge or perhaps a fault-propagation fold.Alternatively, the early growth structure could represent synrift fill of an extensional halfgraben developed by normal motion on the fault. Growth section 2 is middle Miocene inage, corresponding to a period of regional extension. Thus, we prefer the second sce-nario to explain the origin of the older growth structure. This implies that the structureis inverted, with a middle Miocene phase of extension followed by an upper Miocenephase of contraction. This structural inversion in modeled in Figure 5 and interpreted onthe seismic section in Figure 6.

Conclusions:• Nias anticline formed by inversion of a Miocene normal fault and associated half

graben.• Thrust motion on the inverted normal fault is transferred to a backthrust at the

base of the post-rift sequence, forming a structural wedge.• Patterns of folded syntectonic growth strata were used to decipher the inver-

sion history, and to support our kinematic interpretation of this structuralwedge.

Kinematic model

Figure 5: Sequential kinematic model (stages 0 through 5) of the development of the Nias anticline. Model 0 shows an incipient normalfault and active axial surface. Slip on the normal fault (models 1–2) generates a roll-over panel and half graben, which is filled with synriftstrata equivalent to growth section 2. In model 3, strata are deposited above the half graben after rifting has ceased. In models 4 and 5, thelower segment of the normal fault is reactivated as a thrust, which propagates up dip and shallows to a detachment at the base of the post-rift sequence. Slip is transferred to a backthrust that is parallel to the overlying strata, forming a structural wedge.

Interpreted section with inverted normal fault

Figure 6: Interpreted seismic profile, showing an inverted half graben in the core of the Nias anticline. Growth section 2 isinterpreted as synrift strata, similar to the model shown in Figure 5.


144

2-24: Interference Structure, Gulf of Mexico, U.S.A.Rion H. Camerlo, ChevronTexaco, New Orleans, Louisiana, U.S.A.Thomas W. Bjerstedt, Minerals Management Service, New Orleans, Louisiana, U.S.A.Edward F. Benson, NuTec Energy Services, Stafford, Texas, U.S.A.

Location: Perdido fold belt, western Gulf of Mexico, U.S.A.Topics: Kink-band interference and linkage, mechanical

stratigraphy, kink-fold class detachment foldsReserves: Rank exploration area

The Perdido fold belt (Figure 1) is a deep-water fold belt created byupdip extension and sediment loading on the Texas Gulf Coast. LineA-A� (Figure 2) shows two large anticlines within the Perdido fold beltthat are bounded by tabular bands of angular folded units (kinkbands) along both limbs (Figures 3 and 4). The inner kink bands inter-sect, producing an interference structure (Figure 5). This structure isstrikingly similar to the model geometry developed by Medwedeff andSuppe (1997) of a counter-clockwise interference structure, and thatmodeled by Mount (1989) and Novoa et al., (1998). The natural struc-ture deviates from this ideal parallel geometry due to the mechanicalstratigraphy of the deforming units. Shortening across the kink bands(Figure 6), in general, shows two shortening maxima, one to the north-east on the western kink band and one to the southwest on the east-ern kink band. The maxima are likely initiation points of the bandsand demonstrate that kink band shortening is acting in relay, where-in the western kink band loses displacement to the south and theeastern kink band gains displacement to the south. The interferencestructure formed in the overlapping zone of the relay. In detail, theinterference structure is a zone of diminished shortening that may bethe result of inhibition of kink band growth in the zone of interference.The steep shortening gradient northeast of line B-B� may result fromthe difficulty in forming an interference structure above the weakunit. A second shortening minima is seen in the western kink band atline D-D� as well as a distinct swing in orientation. These are the resultof linkage of two synthetic (with respect to dip) kink bands duringkink band growth (Line C-C�). This is similar to the interaction offaults, fractures, and folds in plan view. Both kink bands in the struc-ture are more narrow above the interpreted green horizon in Figure 7than below it, and consequently accommodate less shortening. Theunit between the yellow and green horizon also has appreciable defor-mation induced thickness changes. This unit is interpreted as a sec-ondary detachment horizon in the lower Eocene (a known detach-ment level across south Texas).

Figure 1: Perdido fold belt and area of interest for kink-band interference structure, western Gulf of Mexico.Uninterpreted profile plane vertical seismic section A-A� stretched to depth, located in Figure 2. 3-D post-stacktime migrated seismic courtesy of WesternGeco.


145

2-24: Interference structure, Gulf of Mexico, U.S.A.

Figure 2: Map of green horizon on seismic sections. Dashed blue lines are axial surfaces (kink plane intersections with the mapped horizon), the two kink bands discussed in this section (2-24) are in bold. Contour interval is 200 ft.

Figure 3: Simplified fence diagram of lines A-A�through D-D� (Figures 4 and 7). The axial surfaces of the three kink bands are shown in different colors for clarification.

Figure 4: Profile-plane vertical seismic sections B-B� through D-D�, uninterpreted and interpreted. Colored dashed linesare axial surfaces. All normal faults are omitted for simplification. Line C-C� shows the linkage of kink band “X” and kinkband “Z”. Line B-B’ shows distortion of the kink bands above the lower Eocene detachment at their intersection point priorto crossing in line A-A� (Figure 7). No vertical exaggeration.


146

Vertical seismic profile shows interpreted structure resulting fromthe interference of two kink bands and it is similar to structuresobserved in nature. Displacement gradients along the kink bands inmap view are modified by the interaction of the interfering kinkbands and influenced by the mechanical stratigraphy.

Figure 6: Color-filled contours of amount of shortening across kink bands, in ft. Totaldisplacement across both kink bands is shown in red text. Blue arrows indicate the loca-tion of measurement locations in addition to those of the cross sections.

Conclusions:• The reflection seismic data illuminates megascopic-scale kink bands.• The geometric fold model of kink-band interference agrees well with the observed structure and is a very useful tool in

interpretation.• Appreciable deviations from the geometric model result from effects of the mechanical stratigraphy. Weak layers influ-

ence, and may control, the location of kink-band intersections.• Kink bands intersect and interact in relay similar to published examples of faults, folds, and fractures.

Figure 7: Line A-A� interpreted. Dashed blue lines are axial surfaces (kink planes). Kink band annotation follows the naming convention of Medwedeff and Suppe (1997). The two tan col-ored horizons were added to show details of the kink bands’ intersection. An additional kink band (labeled 33X11T), and requisite branch points P5 and P6, are deviations from the modelgeometry. No vertical exaggeration.

2-24: Interference structure, Gulf of Mexico, U.S.A.

Figure 5: Line drawing of a natural kink-band interference structure inblack Carboniferous slates in the Beara Peninsula, Cork, Ireland, afterDewey, 1965. Dewey reported disharmonic folding at interference points.


147

2-25: End members of gravitational fold and thrust belts (GFTBs) in the deep waters of Brazil

Pedro Victor Zalán, PETROBRAS/E&P/E&P-CORP/TSP,Rio de Janeiro, Brazil

Location: Deep waters of BrazilTopics: Gravitational fault-related folding, growth strata

Figure 2: Depth-migrated seismic section from the Pará-Maranhão Basin illustrating a complete linked extensional-compressional system. See interpreted section in Figure 3.

Figure 1: Location map of the gravitational fold and thrust belts(GFTBs) studied in this work. AM = Amazon Mouth Basin, PM = Pará-Maranhão Basin, BA = Barreirinhas Basin, PE = Pelotas Basin.

Gravitational fold and thrust belts (GFTBs) associated to linked exten-sional-compressional systems occur in the deep waters offshoreBrazil, and show two end members regarding structural and synde-positional styles. One end member, related to longer-lived linkedextensional-compressional systems, is dominated by low rates of sed-imentation. Thus, deformation rates are also low, giving rise to foldbelts with growth folds topped by a series of younging- and stepping-upward time-transgressive unconformities separating stronglydeformed (below) and non-deformed (above) strata of the same age(growth strata). This fold belt type is well illustrated by a seismic sec-tion from the Pará-Maranhão Basin, as well as by another seismic linefrom the Barreirinhas Basin (Figure 1). The other end member, relat-

ed to short-lived, linked extensional-compressional systems, is dominated by high rates of sedimentation. Thus, deformationrates are also higher, giving rise to very thick, harmonically folded and thrusted sedimentary strata, displaying simpler syn-growth relationships. In this case, thick syn-tectonic packages are deposited in the synclines and thinner (or absent) correl-ative packages on the anticlines. Time-transgressive unconformities are markedly absent. This type is illustrated by seismicsections from two major Miocene-Recent progradational sedimentary cones: the Amazon Mouth and the Rio Grande (Figure1). The four cases presented in this section (2-25) are shale-detached/shale-cored fold belts.

Development of Passive Margins and GFTBs Continental margins build outward into deep and ultra-deep waters via denudation of the adjoining shields and deposition ofthe resulting debris, forming the continental shelves and slopes. The rifted/thinned edge of the continental plates cool expo-nentially as they move away from the heat source (mid-oceanic ridges) that initiated break-up of the continental plate. Thesecontinuous events create a very unstable situation since large volumes of sediments pile up at the margin of the continentalshelves, in the upper slope, at the same time the whole area is gradually tilting oceanward due to thermal flexural bending.Large deltaic deposits of major rivers may create similar unstable conditions. Gravity failure occurs and allochtonous mass-es of sediments slide down the slope, over a ductile lithology that detaches the traveling rocks above from the autochtonousrocks below. When the frontal parts of the allochton diminish their velocity due either to a decrease in the gradient of such


148

detachment zone or to a physical barrier (commonly a volcanicedifice) the incoming allochtons collide and strong contrac-tion/compression occurs.

The sliding of huge masses of recently deposited, slightlyindurated sedimentary rocks takes place along well defined,seismically evident, closely spaced detachment zones, nucle-ated in ductile beds with regional distribution; salt or thicklaminated shales. The detachment zones provide the linkagebetween the extensional and compressional provinces. Whenin significant volumes, the ductile beds may be involved in thefolding, giving rise to huge diapir-nucleated folds. The natureof the detachment zone is the main factor determining thestructural style of the associated GFTB. They may be of twostrikingly different types: (a) salt-detached/salt-cored foldbelts, e.g. the Perdido and Mississippi Fan fold belts (GOM)(Trudgill et al., 1999), and (b) shale-detached/shale-cored fold-belts, e.g. the Mexican Ridges (GOM) (Trudgill et al., 1999),Amazon Cone (Silva and Maciel, 1998) and Niger Delta(Hermann, 1998) fold belts. The failure occurs when verticalstresses due to overburden are weakened in relation to sub-horizontal stresses due to several possibilities, including over-pressure in shales (due to petroleum generation or any otherclassical overpressure mechanism) or ductile flow in salt.Shear stresses develop parallel to the slightly dipping beddingand overcome the vertical stresses.

These deformed masses of allochtonous rocks are referred toas linked extensional-compressional systems and havebeen found in the deep/ultra-deep water regions of most con-tinental margins around the world. It is easily understood, andvery well displayed in modern seismic sections, that these sys-tems are composed of three major tectonic domains, each onepresenting different and peculiar deformation (Figure 3).

The extensional domain comprises highly strained subsidedterrains at the upper continental slope, dominated by arcuatedlistric normal faults that sole out at the detachment level. Majorlistric faults present significant associated rollover anticlinesand growth depositional wedges that thicken from the crest ofthe anticline towards the listric fault. Subsidiary listric normalfaults, antithetic to the major downdip listric faults, are alsoabundant, as well as crestal fracturing/faulting in the rolloveranticlines.

The translational domain is a predominantly non-deformed region that passively traveled over the detachment zone. Weakarching may affect the rocks present in this area. Usually, increasing amounts of detachment folding occur oceanwards/bas-inwards, marking the passage of the translational domain into the compressional realm.

The compressional domain may present spectacular deformation, with all kinds of reverse and thrust faults and fault-relat-ed folding (detachment, fault-propagation, and fault-bend folding). When detached on shales, the structural styles, the struc-tural relief, and the overall dimensions may resemble those found in truly orogenic belts (Zalán, 1998). When salt is the lubri-cant, or is otherwise involved, deformation is more complex and salt tongues and canopies (Rowan et al., 2001) or nappes(Hudec et al., 2001) develop. The specific name gravitational fold and thrust belts (GFTBs) has been applied to such enti-ties. Zalán (1999) studied some Brazilian GFTBs in detail and devised a tripartite structural model that predicts an orderlysuccession, from the internides to the foreland, of detachment folding, followed by closely spaced high-angle reverse faultsand associated tight fault-propagation folds (also referred to as toe thrusts), ending in more widely spaced, low-angle ramp-flat thrusts with associated more open fault-bend folding. Important oil discoveries have been achieved in these compres-sional provinces in deep waters off GOM, Nigeria, Angola, and Brazil.

The dimensions of these three domains may vary greatly. Usually the extensional and compressional domains are the widestbut it is very difficult to exactly balance the amount of extension updip with the amount of contraction downdip, because ofthe details of the severe deformation that is usually non-resolvable by seismic data. Since they cover huge areas, on the orderof several thousand square kilometers, it is difficult to have them all covered by 3-D seismic, and it is not unusual that exten-sion and compression are divided into two or three belts of deformation.

Figure 3: Depth-migrated seismic section from the Pará-Maranhão GFTB. The major components of a linked extensional-compressional system are clearly visible: The extensional, trans-lational, and compressional (GFTB) domains and the linking detachment zone.

2-25: GFTBs in the deep waters of Brazil


149


Petroleum Potential of the GFTBsThe potential of these compressional structures for petroleum explo-ration seems to be very high. Structural closures are usually four-wayand on the order of tens of square kilometers, and vertical reliefs are onthe order of hundreds of meters. Common reservoir targets include ver-tically stacked, laterally confined, porous turbidite sandstones, deposit-ed in meandering channels, that are encased within shales in the heartsof the anticlines. Detachment seems to preferentially take place inweakened, highly pressurized organic-rich shales located in oil genera-tion windows. Reverse/thrust faults that splay upwards from thedetachment zones into the folds serve as migrating routes for theascending, released hydrocarbons. When salt is the detachment media,salt windows are required to allow migration from sub-salt sourcerocks. Traps are the fault-related folds and petroleum fields in deepwater GOM, Nigeria, Angola and Brazil have been found in all threemajor types of contractional fault-related folds (detachment, fault-prop-agation and fault-bend anticlines).

GFTBs with Growth FoldingWhen the process of gravity sliding/contraction is long-lasting (20–50m.y.) and takes place in areas with low rates of sedimentation, growthfolding occurs. Since GFTBs develop in exclusively submarine environ-ments, they are never subaerially exposed, sedimentation takes placeconcomitantly with the compressional deformation leading to the depo-sition of syntectonic growth strata that thin up onto the upper parts ofthe fold belt, in the same way the growth wedges develop in the down-thrown sides of the normal faults in the extensional domains.

Medwedeff (1989) unraveled complex growth stratigraphic relation-ships between coeval sediments deposited in the forelimb and back-limb of a fault-bend fold in California. Numerous wells and seismic dataallowed the author to deduce that syntectonic sedimentary strata onlapa time-transgressive unconformity on the forelimb but are folded belowthe unconformity on the backlimb.

The same mechanism seems to be applicable to GFTBs in the BrazilianEquatorial Atlantic margin, such as the Pará-Maranhão and BarreirinhasGFTBs. Figure 3 shows a depth-migrated seismic section from the Pará-Maranhão Basin, where a complete, fully developed linked extensional-compressional system can be seen.

Figure 4: Detailed view of theGFTB shown in Figure 3, display-ing the internal architecture andstructural styles of the compres-sional domain. Notice predomi-nance of fault-bend folding in themore external part of the fold belt(yellow arrows), fault-propagationfolding in the middle part (pinkarrows), and detachment folding inthe more internal zone of the foldbelt (green arrow). Also noticeonlapping pattern and thinningupward of sub-horizontal sedimen-tary packages deposited upon thefrontal (right) part of the GFTB, andthe thinning upward and deformed(folded) nature of the depth-equiv-alent packages on the back (left)portion of the GFTB.

Figure 5: Geologicalinterpretation of sectiondisplayed in Figure 4.Detachment zone andreverse/thrust faults areshown. Two reflectors (a-orange and b-light green)were tracked within theinterpreted pre-tectonicsection, and three reflectors(c-blue, d-purple, and e-dark green) were trackedwithin the interpretedgrowth section, which isalso highlighted by a graytransparent mask.Stepping- and younging-upward unconformities (U)are displayed in yellow.


150


Within the GFTB fault-bend folds are concentrated in the frontalpart while fault-propagation folds occur in the middle part (Figure4). Detachment folds occur in the innermost part of the fold belt.Three reflectors can be tentatively correlated across four time-trans-gressive unconformities (dark yellow) (Figure 5). These reflectors(dark green, purple, and blue) are interpreted to encompass thegrowth strata associated to this GFTB. They thin onto the highesttopography of the fold belt and thicken away into the lower sur-rounding areas. In contrast, lower beds involved in the compres-sional deformation (light green and orange reflectors) show con-stant thickness throughout the fold belt and are interpreted as pre-growth strata.

The geometry of the deformation in this GFTB suggests that thefolding and uplift of the thrust strata was a long-lived process. Sincethere is a major time-transgressive unconformity (as well as threeminor ones) that separates the non-deformed strata above from thedeformed strata below (Figure 5) it is plausible that the rates of sed-imentation were low. We estimate that the time span involved in thesedimentation of the growth strata is around 45 m.y. Deformationstarted slightly below reflector c (Figure 5) (roughly Late Maastrich-tian) and ended slightly above reflector e (Figure 5) (roughly LateOligocene) (sedimentation rate about 35–40 m/m.y). Thus, the de-formed strata were left exposed at the sea bottom several times andsubmarine erosion (currents) could take place.

In a less spectacular manner, the Barreirinhas GFTB display fault-bend folds covered by a major unconformity (there is also a minorassociated stepping-upward unconformity), upon which sub-hori-zontal beds onlap and thin upwards (Figures 6 and 7). The geome-try suggests that the same growth fault-bend folding mechanismdescribed in the Pará-Maranhão GFTB may work here.

The same pattern of a series of younging- and stepping-upward time-transgressive unconformities separating non-deformed onlappingstrata above from thrusted and folded strata below can be seen inseveral other GFTBs in Brazil (e.g. Touros, in the Potiguar Basin;Zalán, 2001) and elsewhere in the world (for instance, in theKrishna-Godavari Basin, in India, Stuart and Hickman, 2001). We sug-gest that these patterns are diagnostic of gravity sliding/contractionaccompanied by growth folding in areas dominated by low rates ofdeformation and sedimentation.

Figure 6: Time-migrated seismic sectionfrom the Barreirinhas GFTB. Two majoropen folds associated to thrust faults (redarrows) can be clearly seen and are cov-ered by sub-horizontal onlapping sedi-mentary packages (yellow arrows).

Figure 7: Geological interpretationof section displayed in Figure 6.Detachment zone and two majorthrust faults are shown in the centerand right portions (external portionof foldbelt) of the section. The twomajor anticlines are interpreted asfault-bend folds. Active (blue) andinactive (pink) axial surfaces areshown in each fold. Towards themore internal parts of the fold belt,deformation is more complex, con-sisting of tight higher-angle reversefaults and associated folding (sug-gestive of fault-propagation folds).Deformed strata are topped by amajor and a secondary upwardclimbing unconformity (yellow).Interpreted growth section is high-lighted by a gray transparent mask.Depth scale is valid for the centralportion of the seismic line.


151

GFTBs Without Growth Folding Some GFTBs do not display the complicated pattern of time-transgressiveunconformities as described above. They involve thick packages of sedimentsthat are folded and thrusted harmonically. Syntectonic sedimentation seemsto follow a simpler pattern of being confined to intervening synclines betweenanticlines. In this case, the syncline packages typically thicken downwardtowards the depocenter and thin upward towards the anticlines. Such is thecase in the GFTBs related to the Amazon Mouth (Figures 8, 9, and 10), to theRio Grande Cone (Figures 11, 12, and 13) and to the Niger delta, where allallochtonous sediments are harmonically folded and thrusted up out in thesea bottom. They are situated in front of young (Miocene) and huge deltaswhere enormous piles of sediments accumulated very quickly while gravitysliding was taking place during the same short time.

The pattern of harmonically folded and thrusted sediments, with thick syn-tec-tonic packages in the synclines (Figures 10 and 13) and thinner correlativepackages on the anticlines, and more importantly, the absence of time-trans-gressive unconformities, are here interpreted as being diagnostic of gravitysliding/contraction in areas dominated by high rates of deformation and sedi-mentation.


Figure 10: Geological interpretation of sectiondisplayed in Figure 9. Detachment zone, reversefaults, and shale diapirs are shown. Growthstrata (highlighted by a gray transparent mask)are ponded in synclines, thinning upwardtowards the flanks of anticlines. Notice theremarkable absence of extensive time-transgres-sive unconformities throughout the whole sedi-mentary section, in clear contrast with the sec-tions illustrated in Figures 5 and 7. Only a veryminor unconformity (yellow) can be seen in theeasternmost flank of the fold belt.

Figure 9: Detailed view of the GFTB shown inFigure 8, displaying the internal architecture andstructural styles of the compressional domain.Tight fault-propagation folds associated withreverse faults dominate the external (right) partof the fold belt, while detachment folds nucleat-ed by shale diapirs constitute the dominant stylein the internal (left) part of the fold belt.

Figure 8: Time-migrated seismic section from the Amazon Mouth GFTB. Extensional domain is only partly shown.Translational and compressional domains are fully displayed.


152


Conclusions:• Gravitational fold and thrust belts associated to extensional-compressional

systems linked via detachment zones nucleated in shales, in the Brazilian deepand ultra-deep waters, show two end members as related to structural stylesand syntectonic sedimentation.

• GFTBs developed in continental margins dominated by low rates of sedimen-tation/deformation display a zonation of fault-bend folding in the more exter-nal parts passing through fault-propagation folding and to detachment foldingas one moves backwards into the internal zones. Stepping-upward time-trans-gressive unconformities cover the folded/thrusted assemblages and areonlapped in the frontal parts of the folds by sub-horizontal growth strata,whose time-equivalent packages are involved in the compressional deforma-tion below the unconformities in the back limbs of the innermost folds.

Figure 13: Geological interpretationof section displayed in Figure 12.Detachment zone and reverse faultsare shown. Growth strata (highlight-ed by a gray transparent mask) areponded in synclines, thinningupward towards the flanks of anti-clines. Notice the absence of exten-sive time-transgressive unconformi-ties throughout the whole sedimenta-ry section, in clear contrast with thesections illustrated in Figures 5 and7. Only a very minor unconformity(yellow) can be seen in the eastern-most flank of the fold belt, similarlyto the Amazon Mouth example(Figure 10).

Figure 12: Detailed view of theGFTB shown in Figure 11, displayingthe internal architecture and struc-tural styles of the compressionaldomain. A train of harmonically fold-ed anticlines dominate the fold belt.Fault-propagation folds associated toreverse faults are preponderant in theexternal (right) part of the fold belt,while a detachment anticline, partlyruptured by reverse faults, can beseen in the internal (center) part ofthe fold belt.

Figure 11: Time-migrated seismic section from the Rio Grande Cone. Extensional domain is only partly shown.Translational domain is practically non-existent. Compressional domain is fully illustrated.


153


Conclusions: (cont.)

• GFTBs developed in areas dominated by high rates of sed-imentation/deformation, usually associated to majordeltas (as is the case in the examples shown in theAmazon Mouth and Rio Grande Cone), display intensefolding, dominated by detachment and fault-propagationfolding. These structures typically have very high struc-tural relief, and seafloor expression. The intense, rapid,and continuous process of sedimentary loading/sliding/contraction does not allow the development of unconfor-mities chiseling the higher parts of the foldbelt. Conse-quently, there are practically no sub-horizontal strata cov-ering the deformed rocks. Syntectonic sediments are con-centrated in the synclines, situated between the interven-ing trains of anticlines.

The major implications for such differences in the deposi-tional/structural styles of the growth strata are in the loca-tion of the turbidite beds and the related hydrocarbontraps. In the first case, syntectonic turbidites should be pre-sent as onlapping strata (stratigraphic traps) above uncon-formities in the frontal parts of the GFTB and in the folds(structural traps) in the internal parts of the GFTB, belowunconformities. In the second case, all syntectonic turbiditedeposits will tend to be channelized bodies that are thickerin the synclines and on the flanks of the anticlines (mixedstratigraphic/structural traps), and thinner or absent up onthe crests. Eventually, inverted depocenters due to shiftingof deformation locus, from the outer to the inner parts, mayuplift such turbidite channels into the core of younger anti-clines. Pretectonic turbidites may be present in the core offolds anywhere in the GFTB.

AcknowledgmentsI would like to thank Petrobras for the permission to pub-

lish this work and my colleagues Haroldo M. Ramos, SergioRogerio P. da Silva, Alvaro Henrique A. de Castro, DesiderioP. Silveira, Sergio de O. Guimarães, and Marcia de B.Pimentel for their help in the processing, interpretation, anddrawing of the seismic sections.


154

ReferencesAllmendinger, R. W., 1998, Inverse and forward numerical modeling of

trishear fault-propagation folds: Tectonics, v. 17, no. 4, p. 640–656.Apotria, T. G., and M. S. Wilkerson, 2002, Seismic expression and kine-

matics of a fault-related fold termination: Rosario structure, Mara-caibo Basin, Venezuela, in M. S. Wilkerson, M. P. Fischer, and T. G.Apotria, eds., Fault Related Folding: The Transition from 2-D to 3-D:Journal of Structural Geology Special Issue, v. 24, no. 4, p. 671–687.

Aymard, R., L. Pimentel, P. Eitz, P. Lopéz, A. Chaouch, J. Navarro, J.Mijares, and J. G. Pereira, 1990, Geological Integration and Evalua-tion of Northern Monagas, Eastern Venezuelan Basin: in J. Brooks,ed., Classic Petroleum Provinces: Geological Society of London,Blackwell, Sp., Pub, no. 50, p. 37–53.

Bischke, R. E., M. Morales, B. A. Ascanio, D. C. Metzner, and P. Verrall,1997, The recognition of a duplex with imbricate thrusting in theFurrial-Tejero Trend, North Monagas area, Venezuela: AAPG Abst.p. A15.

Bally, A. W., P. L. Gordy, and G. A. Stewart, 1966, Structure, seismicdata, and orogenic evolution of southern Canadian Rocky Moun-tains: Canadian Society of Petroleum Geologists Bulletin, v. 14, p.337–381.

Bartow, J. A., 1990, A summary of the Cenozoic stratigraphy and geo-logic history of the Coalinga region, Central California, in TheCoalinga, California, earthquake of May 2, 1983, M. J. Rymer, andW. L. Ellsworth, eds., U. S. Geological Survey Professional Paper,Report: P 1487, p. 3–12.

Boyer, S. E., and D. Elliot, 1982, Thrust systems: AAPG Bulletin, v. 66,p. 1196–1230.

Carnevali, J., 1988, Geology of new giant oil fields in mountain front ofnortheastern Venezuela: Proceedings, 28th International GeologicalCongress, Washington, 9–19 July, v. 1, p. 1240–1241.

Connors, C. D., D. B. Denson, G. Kristiansen, and D. M. Angstadt, 1998,Compressive Anticlines of the Mid-outer Slope, Central Niger Delta(abs.): AAPG International Conf., AAPG Bulletin, v. 82, no. 10, p.1883–1984

Cooper, M. A., F. T. Addison, R. Alvarez, M. Coral, R. H. Graham, A. B.Hayward, S. Howe, J. Martinez, J. Naar, R. Peñas, A. J. Pulham, andA. Taborda, 1995, Basin development and tectonic history of theLlanos basin, Eastern Cordiller, and Middle Magdalena valley,Colombia: AAPG Bulletin, v. 79, p. 1421–1443.

Cooper, M. A., and G. D. Williams, Inversion Tectonics: Geological Soci-ety Special Publication 44, p. 375.

Corredor, F., J. H. Shaw, and F. Bilotti, 2002, Imbricate Shear Fault-bendFolding in the Deep-water Niger Delta, [abs}: GSA Annual Meeting,Denver, Colorado, U.S.A.

Corredor, F., J. H. Shaw, and F. Bilotti, 2001, Modeling of Complex

Imbricate Thrust Systems Using Growth Strata in the Deep-WaterNiger Delta, [abs]: AAPG Annual Convention Program, Denver, Col-orado, U.S.A.

Crawford, M. F., W. G. Brown, N. Donovan, and K. Morgan, 1990,Remote sensing techniques applied to structural geology and oilexploration in south-central Oklahoma: Geological Society of Amer-ica Field Trip Guidebook, 116 p.

Crouch, J. K., and J. Suppe, 1993, Late Cenozoic tectonic evolution ofthe Los Angeles basin and inner California borderland: A model forcore complex-like crustal extension: Geological Society of AmericaBulletin, v. 105, p. 1415–1434.

Dahlstrom, C. D. A., 1969, Balanced cross sections: Canadian Journal ofEarth Sciences, v. 6, p. 743–757.

Dahlstrom, C. D. A., 1970, Structural geology in the eastern margin ofthe Canadian Rocky Mountains: Bulletin of Canadian PetroleumGeology, v. 17, p. 326–346.

Dahlstrom, C. D. A.,1990, Geometric constraints derived from the lawof conservation volume and applied to evolutionary models fordetachment folding: AAPG Bulletin, v. 74, p. 336–344.

Dengo, G., 1985, Mid-America: Tectonic setting for the Pacific marginfrom southern Mexico to northwestern Colombia, in A. E. M. Nairn,and F. G. Stehli, eds., The ocean basins and margins, v. 7: New York,Plenum Press, p. 123–180.

Dengo, C. A., and M. L. Covey, 1993, Structure of the Eastern Cordilleraof Columbia: AAPG Bulletin, v. 77, p. 1315–1337.

Dewey, J. F., 1965, Nature and origin of kink bands: Tectonophysics,1/6, p. 459–494.

Diegel, F. A., J. F. Karlo, D. C. Schuster, R. C. Shoup, and P. R. Tauvers,1995, Cenozoic structural evolution and tectono-stratigraphic frame-work of the northern Gulf Coast continental margin, in M. P. A. Jack-son, D. G. Roberts, and S. Snelson, eds., Salt Tectonics: a Global Per-spective: AAPG Memoir 65, p. 109–151.

Eisenstadt, G., and D. G. DePaor, 1987, Alternative model of thrust faultpropagation, Geology, v. 15, p. 630–633.

Elliott, D., 1976, The motion of thrust sheets: Journal of GeophysicalResearch, v. 81, p. 949–963.

Epard, J.-L., and R. H. Groshong, 1993, Excess area and depth todetachment: AAPG Bulletin, v. 77, p. 1291–1302.

Epard, J.-L., and R. H. Groshong, 1995, Kinematic model of detachmentfolding including limb rotation, fixed hinges, and layer-parallelstrain: Tectonophysics, v. 247 (1-4), p. 85–103.

Erslev, E. A., 1991, Trishear fault-propagation folding: Geology, v. 19,no. 6, p. 617–620.

Gutierrez, R., 1986, Nuevas espectativas en la subcuenca de Maturín,Cuenca Oriental de Venezuela. ARPEL, LX, RANE, tomo II Lima Perú.

Hall, S. H., 2000, The development of large-scale sub-salt structures in

the deep-water western Gulf of Mexico [abs.]: AAPG Annual Con-vention Official Program, p. A62–A63.

Hamilton, W., 1979, Tectonics of the Indonesian region: United StatesGeological Survey Professional Paper 1078, 345 p.

Hardeman, W. D., 1966, Geologic Map of Tennessee, Scale 1:250,000:State of Tennessee, Department of Conservation, Division of Geology.

Hardy, S., and M. Ford, 1997, Numerical modeling of trishear fault-propagation folding and associated growth strata: Tectonics, v. 16,no. 5, p. 841–854.

Hardy, S., and J. Poblet, 1994, Geometric and numerical model of pro-gressive limb rotation in detachment folds: Geology, v. 22, p. 371–374.

Hardy, S., and J. Poblet, 1995, The velocity description of deformation,Paper 2: Sediment geometries associated with fault-bend and fault-propagation folds: Marine and Petroleum Geology, v. 12, p. 165–176.

Hardy, S., J. Poblet, K. McClay, and D. Waltham, 1996, Mathematicalmodeling of growth strata associated with fault-related fold struc-tures, in Modern developments in structural interpretation, Valida-tion and modeling, Buchanan, P. G., and D. A. Nieuwland, eds.: Geo-logical Society of London Special Publication no. 99, p. 265–282.

Harris, L. D., and R. C. Milici, 1977, Characteristics of thin-skinned styleof deformation in the southern Appalachians, and potential hy-drocarbon traps: United States Geological Survey ProfessionalPaper 1018, 40 p.

Hayward, A. B. and R. H. Graham, 1989, Some geometrical characteris-tics of inversion, in M. A. Cooper, and G. D. Williams, eds., InversionTectonics: Geological Society Special Publication 44, p. 17–40.

Hermann, L., 1998, Deep offshore West Niger Delta slope, Nigeria:Structural styles and hydrocarbon reservoirs: AAPG InternationalConference and Exhibition Extended Abstracts Volume, Rio de Ja-neiro, Brazil, p. 18–19.

Hudec, M. R., M. P. A. Jackson, L. F. Binga, J. C. da Silva, R. Fraenkl, andW. Sikkema, 2001, Regional restoration in the offshore KwanzaBasin, Angola: Linked zones of extension, translation, and contrac-tion: AAPG Annual Convention Official Program Book and CD-ROM,Denver, Colorado, U.S.A., June, p. A95.

Hyndman, R. D., G. D. Spence, T. Yuan, and E. E. Davis, 1994, Regionalgeophysics and structural framework of the Vancouver Island mar-gin accretionary prism: Proceeds of the Ocean Drilling Program,Part A: Initial Reports, v. 146, p. 399–419.

Jamison, W. R., 1987, Geometric analysis of fold development in over-thrust terranes: Journal of Structural Geology, v. 9, p. 207–219.

Lang, R. C., 1957, The Criner Hills: a key to the geologic history ofsouthern Oklahoma: Ardmore Geological Society Guidebook—Criner Hills Field Conference, p. 18–25.

Mastoris, S., 1989, 3-D seismic interpretation techniques define shal-low gas sand reservoirs: Oil and Gas Journal, v. 87, p. 69–73.


155

References (Con’t.)Medwedeff, D. A., 1985, Lateral terminations of fault-related folds: Ge-

ological Society of America Abstracts with Programs 17, p. 661.Medwedeff, D. A., 1988, Structural analysis and tectonic significance of

Late-Tertiary and Quaternary compressive growth folding, SanJoaquin Valley, California: Ph.D. Thesis, Princeton University,Princeton, New Jersey, U.S.A. 184 p.

Medwedeff, D. A., 1989, Growth fault-bend folding at southeast LostHills, San Joaquin Valley, California: AAPG Bulletin, v. 73, p. 54–67.

Medwedeff, D. A., 1992, Geometry and kinematics of an active, lateral-ly propagating wedge thrust, Wheeler Ridge, California, in S. Mitraand G. Fisher, eds., Structural Geology of Fold and Thrust Belts,John Hopkins University Press. p. 3–28.

Medwedeff, D. A. and J. Suppe, 1997, Multibend fault-bend folding:Journal of Structural Geology, 19, p. 279–292.

Meltzer, A., 1989, Crustal structure and tectonic evolution: Central Cal-ifornia, Ph.D. thesis, Rice University, Houston, Texas, U.S.A.

Mitra, S., 1993, Geometry and kinematic evolution of inversion struc-tures: AAPG Bulletin, v. 77, p. 1159–1191.

Mitra, S., and Q. T. Islam, 1994, Experimental (clay) models of inver-sion structures: Tectonophysics, v. 230, p. 211–222.

Molina, A., 1992, Rosario Field, Venezuela, Maracaibo Basin, ZuliaState, in N. H. Foster and E. A. Beaumont, eds.: AAPG Treatise ofPetroleum Geology, Atlas of Oil and Gas Fields, Structural Traps VI,p. 293–304.

Moore, G. F., D. E. Karig, T. H. Shipley, A. Taira, P. L. Stoffa, and W. T.Wood, 1991, Structural framework of the ODP Leg 131 area, NankaiTrough: Proceedings of the Ocean Drilling Program, Part A: InitialReports, v. 131, p. 15–20.

Moore, G. F., A. Taira, S. Kuramoto, T. H. Shipley, and N. L. Bangs, 1999,Strucutural setting of the 1999 U.S.-Japan Nankai Trough 3-D seismicreflection survey: EOS, Transactions of the American GeophysicalUnion v. 80, p. F569.

Moore, G. F., et al., 2002, Data report: Structural setting of the Leg 190Muroto transect, in G. F. Moore, A. Taira, and A. Klaus, et al.: Pro-ceedings of the Ocean Drilling Program, Initial Reports, v. 190,http://www-odp.tamu.edu/publications/190_IR/chap_02/chap_02.htm (accessed July 19, 2004).

Morgan, J. K., and D. E. Karig, 1995, Kinematics and a balanced andrestored cross-section across the toe of the eastern Nankai accre-tionary prism: Journal of Structural Geology, 17, p. 31–45.

Mount, V., 1989, State of stress in California and a seismic structuralanalysis of the Perdido fold belt, Northwest Gulf of Mexico: Ph.D.dissertation, Princeton University, Princeton, New Jersey, U.S.A.

Mount, V. S., J. Suppe, and S. C. Hook, 1990, A forward modeling strat-egy for balanced cross sections: AAPG Bulletin, v. 74, no. 5, p.521–531.

Namson, J., and T. L. Davis, 1988, A structural transect of the westernTransverse Ranges, California: Implications for lithospheric kine-matics and seismic risk evaluation: Geology, v. 16, p. 675–679.

Narr, W., and J. Suppe, 1994, Kinematics of basement-involved com-pressive structures: American Journal of Science, 294, v. 802–860.

Nicholson, C., C. Sorlien, and M. Legg, 1993, Crustal imaging andextreme Miocene extension of the Inner California Continental Bor-derland: Geological Society of America Abstracts with Programs, v.25, p. A–418.

Novoa, E., J. Suppe, and J. H. Shaw, 2000, Inclined-shear restoration ofgrowth folds: AAPG Bulletin, v. 84, no. 6, p. 787–804.

Novoa, E., V. Mount, and J. Suppe, 1998, Map-view interference of mon-oclinal folds: Journal of Structural Geology, v. 20, no. 4, p. 339–353.

Parnaud, F., Y. Gou, J. C. Pascual, I. Truskowski, O. Gallango, and H.Passalacqua, 1995, Petroleum geology of the central part of theeastern Venezuela basin, in A. J. Tankard, R. Suarez S., and H. J.Welsink, eds., Petroleum basins of South America: AAPG Memoir 62,p. 741–756.

Paterson, M. S., and L. E. Weiss, 1966, Experimental deformation andfolding in phyllite: Geological Society of America Bulletin, 77, p.343–374.

Peel, F. J., 2001, Emplacement, inflation, and folding of an extensiveallochthonous salt sheet in the late Mesozoic (ultra-deepwater Gulfof Mexico) [abs.]: AAPG Annual Convention.

Peel, F. J., C. J. Travis, and J. R. Hossack, 1995, Genetic structuralprovinces and salt tectonics of the Cenozoic offshore U.S. Gulf ofMexico: A preliminary analysis, in M. P. A. Jackson, D. G. Roberts,and S. Snelson, eds., Salt Tectonics: a Global Perspective: AAPGMemoir 65, p. 153–175.

Poblet, J., K. McClay, F. Storti, and J. A. Muñoz, 1997, Geometries ofsyntectonic sediments associated with single layer detachmentfolds: Journal of Structural Geology, v. 19, p. 369–381.

Price, R. A., 1981, The Cordilleran foreland thrust and fold belt in thesouthern Canadian Rocky Mountains, in K. R. McClay, and N. J.Price, eds., Thrust and Nappe Tectonics: Geological Society of Lon-don, Special Publication no. 9, p. 427–448.

Prieto, R., and G. Valdes, 1990, El Furrial Oil Field a New Giant an OldBasin: Abstract.

Rich, J. L., 1934, Mechanics of low-angle overthrust faulting as illus-trated by Cumberland Thrust Block, Virginia, Kentucky and Ten-nessee: AAPG Bulletin, v. 18, p. 1684–1696.

Rivero, C., J. H. Shaw, and K. Mueller, 2000, Oceanside and ThirtymileBank blind thrusts: Implications for earthquake hazards in coastalsouthern California: Geology, v. 28, p. 891–894.

Roure, F., J. Carnevali, Y. Gou, and T. Subieta, 1994, Geometry and kine-matics of the North Monagas thrust belt (Venezuela): Marine and Pe-troleum Geology 1994, v. 11, no. 3, p. 347–362.

Roure, F., B. Colletta, B. DeToni, D. Loureiro, H. Passalacqua, and Y. Gou,1997, Within-plate deformations in the Maracaibo and East ZuliaBasins, Western Venezuela: Marine and Petroleum Geology, v. 14, p.139–163.

Rowan, M. G., 1997, Three-dimensional geometry and evolution of a seg-mented detachment fold, Mississippi Fan foldbelt, Gulf of Mexico:Journal of Structural Geology, v. 19, p. 463–480.

Rowan, M. G., A. Leibold, A. Maccagni, and G. Valenti, 2001,Allochthonous salt, contractional folds, and turtle structures of theLouisiana deepwater province [abs.]: AAPG Annual Convention Offi-cial Program, v. 10, p. A173.

Rowan, M. G., and R. Linares, 2000, Fold evolution matrices and axial sur-face analysis of fault-bend folds: Application to the Medina anticline,eastern Cordillera, Colombia: AAPG Bulletin, v. 84, no. 6, p. 741–764.

Rowan, M. G., F. J. Peel, and B. C. Vendeville, 2001, Gravity-driven foldbelts on passive margins: AAPG Annual Convention Official ProgramBook and CD-ROM, Denver, Colorado, U.S.A., June, p. A174.

Rowan, M. G., F. J. Peel, and B. C. Vendeville, 2004, Gravity-driven foldbelts on passive margins, in K. R. McClay, ed., Thrust tectonics andhydrocarbon systems: AAPG Memoir 82, p. 157–182.

Rowan, M. G., B. D. Trudgill, and J. C. Fiduk, 2000, Deep-water, salt-coredfold belts: Lessons from the Mississippi Fan and Perdido foldbelts,northern Gulf of Mexico, in W. Mohriak and M. Talwani, eds., AtlanticRifts and Continental Margins: American Geophysical Union Geo-physical Monograph 115, p. 173–191.

Salvini, F., F. Storti, and K. McClay, 2001, Self-determining numericalmodeling of compressional fault-bend folding: Geology, v. 29, p.839–842.

Shaw, J. H., F. Bilotti, and P. Brennan, 1999, Patterns of imbricate thrust-ing: Geological Society of America Bulletin, v. 111, no. 7, p. 1140–1154.

Shaw, J. H., S. C. Hook, and J. Suppe, 1994, Structural trend analysis byaxial surface mapping: AAPG Bulletin, v. 78, p. 700–721.

Shaw, J. H., S. C. Hook, and J. Suppe, 1996, Structural trend analysis byaxial surface mapping: Reply, v. 80/5, p. 780–787.

Shaw, J. H., S. C. Hook, and B. Satrio, 1997, Complex structural reacti-vation defined by growth strata: AAPG Annual Meeting Abstracts, v.6, p. 106.


156

References (Con’t.)Shaw, J. H., E. Novoa, and C. Connors, 2004, Structural controls on

growth stratigraphy in contractional fault-related folds, in K. R.McClay, ed., Thrust tectonics and hydrocarbon systems: AAPGMemoir 82, p. 400–412.

Shaw J. H., and J. Suppe, 1996, Earthquake hazards of active blind-thrust faults under the central Los Angeles basin, California: Journalof Geophysical Research, v. 101, p. 8623–8642.

Shaw, J. H., and J. Suppe, 1994, Active faulting and growth folding inthe eastern Santa Barbara Channel, California: Bulletin of GeologicalSociety of America, v. 106, p. 607–626.

Silva, S. R., and R. R. Maciel, 1998, Foz do Amazonas Basin hydrocar-bon system: AAPG International Conference and Exhibition Extend-ed Abstracts Volume, Rio de Janeiro, Brazil, p. 480–481.

Stein, R. S., and G. Ekström, 1992, Sesmicity and Geometry of a 110-km-long blind thrust fault, 2: Synthesis of the 1982–1985 Californiaearthquake sequence: Journal of Geophysical Research, 97, p.4865–4883.

Stewart, K. G., and W. Alvarez, 1991, Mobile hinge kinking in layeredrock sand models: Journal of Structural Geology, 13, p. 243–259.

Storti, F., and J. Poblet, 1997, Growth stratal architectures associatedwith decollement folds and fault-propagation folds: Inferences onfold kinematics: Tectonophysics, 282, p. 353–373.

Stuart, C. J., and R. G. Hickman, 2001, Tertiary linked extensional andcompressional faults and sedimentation in the Krishna-GodavariBasin, India: AAPG Annual Convention Official Program Book andCD-ROM, Denver, Colorado, U.S.A., June, p. A193.

Suppe, J., 1983, Geometry and kinematics of fault-bend folding: Ameri-can Journal of Science, v. 283, p. 684–721.

Suppe, J., 1985, Principles of Structural Geology: Englewood Cliffs, NewJersey, U.S.A., Prentice-Hall, 537 p.

Suppe, J., and D. A. Medwedeff, 1990, Geometry and kinematics of fault-propagation folding, Ecolg. Geol. Helv., v. 83, p. 409–454.

Suppe, J., G. T. Chou, and S. C. Hook, 1992, Rates of folding and fault-ing determined from growth strata, in K. R. McClay, ed., Thrust tec-tonics: London, Chapman, and Hall, p. 105–121.

Suppe, J., F. Sàbat, J. A. Muñoz, J. Poblet, E. Roca, and J. Vergés, 1997,Bed-by-bed growth by kink band migration: Sant Llorenç de Morun-ys, Eastern Pyrenees: Journal of Structural Geology, v. 19, p.443–461.

Suppe, J., C. Connors, and Y. Zhang, 2004, Shear fault-bend folding, inK. R. McClay, ed., Thrust tectonics and hydrocarbon systems: AAPGMemoir 82, p. 303–323.

Talukdar S., O. Gallango, and A. Ruggiero, 1988, Generation and migra-tion of oil in the Maturin Subbasin Eastern Venezuela Basin. Organ-ic Geochemistry, 13, p. 537–547.

Tearpock, D. J., and R. E. Bischke, 2002, Applied Subsurface Geologicalmapping and Structural Techniques, 2nd ed.: Upper Saddle River,New Jersey, U.S.A., Prentice-Hall, 850 p.

Tomlinson, C. W., 1952, Odd geologic structures of southern Okla-homa: AAPG Bulletin, v. 36, no. 9, p. 1820–1840.

Tomlinson, C. W., and W. McBee Jr., 1959, Pennsylvanian sedimentsand orogenies of Ardmore district, Oklahoma: Ardmore GeologicalSociety, Petroleum geology of southern Oklahoma, v. 2, p. 3–52.

Trudgill, B. D., et al., 1999, The Perdido Fold Belt, northwestern deepGulf of Mexico, Part 1: Structural geometry, evolution and regionalimplications: AAPG Bulletin., v. 83, p. 88–112.

Weimer, P., and R. Buffler, 1992, Structural geology and evolution ofthe Mississippi Fan Fold Belt, deep Gulf of Mexico: AAPG Bulletin, v.76, p. 225–251.

Woodward, N. B., S. E. Boyer, and J. Suppe, 1985, An outline of bal-anced cross-sections: University of Tennessee, Department of Geo-logical Sciences, Studies in Geology, v. 11, 2nd ed., 170 p.

Zalán, P. V., 1998, Gravity-driven compressional structural closures inBrazilian deep-waters: A new frontier play: AAPG Annual MeetingExtended Abstracts v. 2, Salt Lake City, Utah, U.S.A., p. A723.

Zalán, P. V., 1999, Seismic expression and internal order of gravitationalfold and thrust belts in Brazilian deep waters (expanded abstract):Sixth International Congress of the Brazilian Geophysical Society, CD-ROM with Extended abstracts, Rio de Janeiro, Brazil, August, 4 p.

Zalán, P. V., 2001, Growth folding in gravitational fold and thrust beltsin the deep waters of the Equatorial Atlantic, Northeastern Brazil:AAPG Annual Convention Official Program Book and CD-ROM, Den-ver, Colorado, U.S.A., June, p. A223.

Zapata, T. R., and R. W. Allmendinger, 1996, Growth stratal records ofinstantaeous and progressive limb rotation in the Precordillera thrustbelt and Bermejo Basin, Argentina: Tectonics, v. 15, no, 5, p. 1065–1083.

seismic interpretation

Documents