14 Erslev 1997 Journal of Structural Geology

8/12/2019 14 Erslev 1997 Journal of Structural Geology

1/15

Pergamon Journal of Structural Geology,Vol. 19, Nos 3-4, pp. 321 to 335, 19970 1997 Elsevier Science LtdPII : S0191-8141 96)00107-1 All rights reserved. Printed in Great Britain0191~8141/97 $17.00+0.00

Multiple geometries and modes of fault propagation folding in the Canadian thrust beltERIC A. ERSLEV and KYLE R. MAYBORN*

Department of Earth Resources, Colorado State University, Fort Collins, CO 80523, U.S.A.Receiv ed 27 Februa ry 1996; accepted i n evi sedform 11 November 1996)

Abstract-A multitude of fold models have been proposed to explain the variety of fold geometries which developin front of thrust faults. Detailed field, fabric, and photogrammetric studies of 4 fault-cored asymmetrical folds inthe thin-skinned Canadian thrust belt were used to test models of fault-propagation folding. Fold geometriesinclude combinations of angular and rounded fold surfaces, highly contorted anticlinal hinge areas, and minimalpenetrative deformation or changes in bedding thickness. Interlimb angles generally decrease with increasingshortening, indicating progressive fold tightening about fixed anticlinal hinges. Extensive flexural slip thrustingtoward the anticlinal axes of angular folds suggests that kink folding in thin-skinned thrust belts is aided by materialtransfer from both fold limbs into hinge areas. Fold geometries change dramatically along the strike of individualstructures, demonstrating the non-uniqueness of fault-propagation fold geometries.

No single mode of fault-propagation folding can explain the diverse fold geometries seen in the Canadian thrustbelt. This geometric variability can be ascribed to the complex interplay of multiple modes of folding. In strata nearthe causal thrusts, oblique shear and flexural slip in triangular shear zones distribute thrust displacements into bothrounded and angular folds. Simultaneous angular folding of overlying strata commonly occurs by progressive kinkfolding where folds tighten by flexural slip on all fold limbs until the thrust breaks through the fold. Regional andlocal differences in the amount of pervasive, top-to-the-craton shear needed for progressive kink folding may bepartially responsible for the variability of fault-propagation fold geometries in thin- and thick-skinned orogensworld-wide. 0 1997 Elsevier Science Ltd. All rights reserved.

INTRODUCTION

Both thin- and thick-skinned thrust belts show theintimate relationship between folding and faultingwhere the steep forelimbs of asymmetrical folds are cutby thrust faults. Many of these structures have beencalled fault-propagation folds, which were defined bySuppe (1985) as folds that represent deformation thattakes place just in front of the propagating fault surface.A multiplicity of conceptual and quantitative models forfault-propagation folding have been proposed to explainthe corresponding diversity of structural geometries.These models have important implications to our knowl-edge of the geometry and internal structure of petroleumreservoirs as well as the seismic hazard of areas (e.g.Northridge, California) underlain by active fault-relatedfold structures.

We recognize that the term fault-propagation foldingwas initially coined to describe structures in thin-skinnedthrust belts and that some authors further restrict theterm to a geometric, kinematic model that explicitlydescribes the evolution of a tip-line fold using equationsrelating fold geometry to fault shape and displacement(Fischer et al. 1992). But we choose to view the term as akinematic concept describing the wave of deformationpreceding a propagating fault (J. Suppe, personalcommunication, 1995), not a rigidly-defined kinematicframework. The similarity of the fault-propagation folddefinition from thin-skinned orogens and the thrust-fold definition (Stone, 1984) from thick-skinned orogens

*Present address: Geology Department, University of California atDavis, Davis, CA 95616, U.S.A.

suggests that the kinematic concept of fault-propagationfolding is equally applicable to basement-involved, thick-skinned thrust structures.

This paper will test models of fault-propagationfolding by defining and examining structural geometriesfrom the Canadian thrust belt in Alberta. Our detailedfield observations, rigorous photogrammetric profiles,and fabric analyses of four optimally-exposed folds fromthe Canadian thrust belt show both the complexity anddiversity of fault-propagation folds. Comparisons of thestructural geometries with the predictions of existingkinematic models provide tests of these models and helpoutline the multiple, competing processes which deter-mine the geometries and kinematics of fault-propagationfolds.

PREVIOUS STRUCTUR L MODELSStructural explanations of asymmetric folds with

variably truncated forelimbs can be roughly groupedinto either sequential or simultaneous models. Willis andWillis (1934) suggested that many structures in thin-skinned thrust belts form sequentially as break-thrustfolds, with flexural slip tightening regularly spaced,asymmetrical buckle folds until they lock, causing thruststo break through their over-steepened forelimbs. Recentfield (Fischer et al. 1992) and experimental studies(Dixon and Liu, 1992) support this sequence of deforma-tion in areas where early buckle folds were truncated bysubsequent thrusting. The analogous fold-thrust modelfor the basement-involved structures of the Laramideorogen (Berg, 1962; Spang et al. 1985; Brown, 1988)

321


2/15

322 E. A. ERSLEV and K. R. MAYBORNinvokes an initial stage of basement and cover foldingprior to the breakthrough of a thrust or reverse fault.

In both settings, however, irregular spacings of manyasymmetrical folds with faulted forelimbs suggest that aninitial stage of buckle folding may not have been anessential element of the process. In addition, manyasymmetrical folds in thin-skinned thrust belts are onlypartially truncated by thrusts, with thrust slip trans-formed to fold tightening upward into the section. Thisindicates a period of synchronous, not sequential,faulting and folding. Similarly, in basement-involvedorogens, the presence of weakly folded basement blocksunder strongly folded cover strata indicates that base-ment faulting is often synchronous with cover folding(Erslev and Rogers, 1993).

Elliott (1976) and Williams and Chapman (1983)proposed a conceptual model for simultaneous faultingand folding where a tip-line fold develops in a ductilebead of material at the fault tip. Subsequent kinematicmodels have diverged in two directions according to theirinterpretation of the deformation style in the ductilebead: those invoking flexural slip during kink-bandmigration (Suppe and Medwedeff, 1984, 1990) andthose invoking shear oblique to the beds (Dietrich andCasey, 1989; Erslev, 1991).

Suppe and Medwedeff (1984) proposed the firstkinematic model for fault-propagation folding, invokingan ingenious array of kink-bands which maintainedconstant bed thickness during thrust propagation up aramp (Fig. la). This self-similar (non-tightening) geome-try predicts relatively simple relationships between theangular elements of individual structures, making iteasily amenable to computer modeling. Subsequentvariations of this model (Suppe and Medwedeff, 1990;Mosar and Suppe, 1992: Fig. 2a) relaxed some of theangular requirements of the original model, allowing it tobe more consistent with numerous conflicting observa-tions, including fixed anticlinal fold hinges (Fisher andAnastasio, 1994). Limited progressive tightening of folds

(0)

can be accomplished by either changing bed thickness inthe forelimb (Jamison, 1987, fig. 1b; Suppe and Medwe-deff, 1990) or varying the thickness of the hanging wall(Mitra, 1990; Fig. lc). Storti and Salvini (1996)attempted to explain the common occurrence of recum-bent fault-propagation anticlines by adding progressiverollover kink folding to the above models.Angular kink-band models have also been developedfor basement-involved structures of the Rocky Moun-tains in the U.S.A. Chester and Chester (1990) modifiedthe Suppe and Medwedeff (1984) fault-propagation foldmodel for faults with constant dip and compared theirgeometries to basement-cored structures. Narr andSuppe (1993) (Fig. Id) approximated rounded foldsurfaces using multiple, interfering kink geometries.McConnell (1994) used angular approximations todevelop a model for progressive rotation in an upward-flaring kink-band emanating from a fault zone at depth.

An alternative kinematic framework is provided bymodels dominated by shear oblique to bedding. Thesequential stretch-thrust model of Heim (19 19) for alpinenappes can be envisioned as a ductile shear zone whichstarts wide, folding the strata, and ends narrow, faultingthe strata. Ramsay and Huber (1987) showed thatheterogeneous simple shear can accomplish simultaneousfolding in areas of low shear strain and faulting in areasof high shear strain. But one basic problem is that thedistributed shear which causes folding often occurs infront of the propagating fault zone. Dietrich and Casey(1989) recognized that the highly deformed zones in theHelvetic nappes converge toward nappe root zones,forming wedge-shaped shear zones. They superposedpure shear flattening on a tabular simple shear zone tomodel these deformation zones (Fig. 2a). The resultingdeformation is consistent with Helvetic nappe geometriesand strain, although there is no evidence of a temporalseparation of the simple shear and pure shear compo-nents of deformation. In their study of two fault-propagation folds from the Iberian peninsula, Alonso(b)

I_\\L (d)

Fig. 1. Kink-band-dominated fault-propagation fold models with (a) self-similar geometries (Suppe and Medwedeff, 1984),(b) variable thickness forelimbs (Jamison, 1987), (c) layer-parallel shortening and heterogeneous shear in the hanging wall

(Mitra, 1990), and (d) multiple kink-bands to approximate basement-involved structures (Narr and Suppe, 1993).


3/15

Geometries and modes of fault-propagation folding 323(a) (b)

m.~~:b.%, +::;


4/15

324 E. A. ERSLEV and K. R. MAYBORNoccurred during Jurassic through Oligocene time andprogressed from west to east (Shaw, 1963; Dahlstrom,1970). In the carbonate-dominated Front Rangesdetachment folds to the north near Jasper, Alberta,change southward to fault-propagation folds and thento fault-bend folds as the proportion of shaly stratadecreases (McMechan and Thompson, 1989).The Abraham Lake study area in the west half of theWhiterabbit Creek quadrangle (Mountjoy et al. 1974)contains 3 well-exposed folds in the Balinhard andSulphur Mountain thrust sheets (Fig. 3). No obvioussecondary structures overprint the original fold geome-tries with the exception of rotation by underlying thrustimbrication. The Cline fold (Fig. 4a) is interpreted as anincipient fault-propagation fold related to a unexposedfault splay off the Sulphur Mountain Thrust. The ElliotPeak fold (Fig. 4b) shows partial truncation of strata by athrust fault within the Sulphur Mountain thrust sheet.The Balinhard fold (Fig. 4c) is a hanging-wall anticlinetruncated by the Balinhard thrust, which has a slip ofapproximately 1 km.

The Barrier Mountain study area in the east half of theBarrier Mountain Quadrangle (Price and Mountjoy,1971) (Fig. 3 insert) exposes the truncated hanging-wallanticline above the Clearwater thrust, a splay off theMcConnell thrust with 15 km of slip. Excellent exposuresof the anticlinal hinge area are provided by the valleywalls of the Panther River and 3 glacial cirques.

All four folds are defined by Devonian and Mississip-pian carbonate rocks. The massive Palliser Formation, agrey, mottled, dolomite-rich, micritic limestone withlocal chert layers, is typically cut by the fault rampresponsible for the overlying fold. The lower part of thePalliser Formation consists of massive, poorly-beddedmicrite and is overlain by more argillaceous and dolomi-tic strata in gradational contact with the Banff Forma-tion. The Banff Formation contains finely interlayeredmicrite, argillaceous dolomite, and crinoidal calcarenite.Micrite layers commonly show spaced, bed-perpendicu-lar stylolitic cleavage that probably formed during earlystages of layer-parallel shortening. Axial planar, anasto-mosing cleavage cross-cuts the earlier stylolitic cleavagein argillaceous layers within hinge areas.The distinctive Pekisko Formation, a light grey,crinoidal calcarenite, conformably overlies the BanffFormation. The overlying Shunda and Turner Valleyformations consist of thinly- to thickly-bedded, light todark grey, fossiliferous limestone and calcarenitic lime-stone. The youngest Mississippian Formation in thestudy areas is the Mount Head Formation which consistsof dark grey limestone and argillaceous dolomite.Methods

Detailed field measurements of over 2400 bedding,cleavage and vein attitudes were combined with photo-grammetric determinations of fold shape to fully char-acterize the folds. Field and photogrammetric

measurements of bedding orientations were used todetermine fold axis orientations for plunge projections.Photogrammetric determinations of bedding orienta-tions used three x, y, z coordinates on large beddingplanes and were important in exposures where steepslopes and loose rocks made access unsafe or infeasible.Several traverses along individual beds were used tocharacterize fold hinge geometries. Samples were col-lected from all lithologies to check for strain markers,with particular emphasis on crinoidal grainstones. Fabricorientations for the Abraham Lake area are more fullydiscussed and tabulated in Mayborn (1993).

Special care was taken to accurately determine foldgeometries in three dimensions using field photographyand photogrammetry using aerial photographs. A spe-cially calibrated Kern PG-2 plotter at the U.S.G.S.Photogrammetric Plotter Lab in Denver allowed thedetermination of x, y, z coordinates (+ 1 m) of keymarker beds from stereo pairs (Pillmore, 1989). The x,y, z coordinates were projected parallel to fold axes inorder to generate plunge projections showing the truegeometry of the folded exposures.Cline fold

The Cline fold is the smallest of the three folds in theAbraham Lake area. It is exposed in the 620 m valley wallof the Cline River (Fig. 3). A photograph (Fig. 4a) andphotolucida drawing (Fig, 5) of the Cline fold indicaterounded anticlinal and synclinal hinges. Eigenvectoranalysis of 315 bedding measurements gives a fold axistrend and plunge of 320-19 (Fig. 5).

Two larger faults are visible from a distance in theCline fold. The upper fault parallels bedding in thebacklimb of the fold and cuts obliquely through thePekisko Formation in the forelimb. The extra material atthis ramp may be a small duplex structure. The arcuate,bed-parallel geometry of this fault suggests that it mayhave been present before folding. The lower fault, in theBanff Formation, cuts across bedding in the core of thefold and is bedding parallel at the synclinal hinge. Thisfault may be either the tip of the main fault or a minorsplay. Minor faults are commonly parallel to bedding.Poles to bedding (Fig. 5) are bimodal, suggesting aplanar backlimb and forelimb. Two bedding traversesrevealed variable fold geometries (Fig. 6). The top of thePalliser Formation has a very angular hinge, with theapparent dips changing from 45SW to 50NE in lessthan 10 m. A distinctive layer in the middle of the BanffFormation showed a similar change in dip over a distanceof 130 m. This increased hinge zone width confirms theupward rounding geometry suggested by field photo-graphs (Fig. 4a).

The only axial planar cleavage in the Cline structure isnear the fault near the synclinal hinge. This cleavage ispreferentially developed in the silty dolomitic rocks of theBanff Formation and may be due to increased deforma-tion near the fault. Vein orientations appear fairly


5/15

Geometries and modes of fault-propagation folding 325

Fig. 4. Photographs of the (a) Cline, (b) Elliot Peak (Elliot 3 exposure), and (c) Balinhard folds.


6/15

3 6 E. A. ERSLEV and K. R. MAYBORNBedding N Cleavage N

+++W+

cl

++++++N= 4

Forelimb

Fig. 5. Photolucida sketch of the Cline fold showing bedding traverselocations and stereonets of bedding and cleavage orientations. Scale baris approximate due to perspective point problems. Formation key forcamera lucidas, photogrammetric profiles and cross-sections: SurveyPeak Fm. (COsp), Flume and Maligne Fms (Dfl), Southest Fm. (Dpx),Mt Hawk Fm. (Dmh), Alexo Fm. (Dax), PaIliser Fm. (Dpa), Banff Fm.(Mbf), Pekisko Fm. (Mpk), Shunda Fm. (Msh), Turner Valley Fm.(Mtv), Mt Head Fm. (Mmh), Rocky Mountain Group (PPrm), SulphurMt Fm. (Trsm), Whitehorse Fm. (Trwh), Jurassic-Cretaceous Fms(JrK).

random and are restricted to the hinge and forelimb. Thissuggests a lack of bed-parallel shear in the backlimbbecause bed-parallel veins are commonly associated withflexural slip faults elsewhere in the study areas.

The down-plunge projection of the Cline fold (Fig. 7a)confirms the upward rounding fold geometry suggestedby the dip traverses. In the plane perpendicular to the foldaxis, the average apparent dip of the backlimb (53SW)appears steeper than the dip in the photograph (Fig. 4a)due to projection corrections for the inclined fold axis.

IT+ 4 Palliser Fm. Middle Banif Fm.

Forelimb

Fig. 6. Apparent dip averages (solid lines) and ranges (shaded areas) ofbedding orientations collected on bed-parallel traverses in the Banff andPalliser formations (see Fig. 5 for locations). Apparent dips are

calculated for the plane perpendicular to the fold axis.

Elliot Peak foldThe Elliot Peak fold (Figs 4b, 7b & c & 8) is a

spectacular fault-propagation fold where the PalliserFormation is cut by a thrust which dies in the foldedBanff Formation. The Elliot 1 exposure has roughly300 m of relief and is the least dramatic of the 3 Elliotexposures because much of the forelimb has been erodedaway. A bed-parallel thrust fault in the backlimbpartially truncates the crest of the angular anticline. Thebimodal bedding orientations of the Shunda Formationgive a fold axis orientation of 332-12 (Fig. S), with aplanar forelimb and a more variable backlimb whose dipsrange from 85 to 45. Axial planar cleavage and veins areconcentrated in the hinge. The near-vertical cleavageorientations do not change significantly from the hinge tothe forelimb. Thin sections show cleavage zones truncat-ing bioclasts, indicating pressure solution.

The plunge projection of the Elliot 1 exposure does notsupply much information because of the limited exposureof the forelimb. The hinge is angular and the forelimb isplanar. The backlimb is slightly undulatory, confirmingthe bedding data in Fig. 8, and has an apparent dip of59SW. Thickness changes in the forelimb cannot bedetermined because the layers traced in the backlimb arenot the same layers traced in the forelimb.

The Elliot 2 exposure is on the 600 m high, northwestface of Elliot Peak (Fig. 3). Oblique and vertical aerialphotographs provided the only data for this part of thefold (Fig. 7b). The hinge is angular and is rounded only atthe lowest exposures. In the core of the fold, a backthrustin the forelimb turns slightly as it enters the core and thendies out. In addition, a small thrust in the backlimb cutsacross bedding. The extra strata brought into the foldhinge by these faults contribute to the angularity of thehinge by filling the space between the angular limbs.

The fold axis (317-6) at the Elliot 2 exposure wascalculated with photogrammetric determinations of bed-ding attitudes. Beds in the core of the anticline curvethrough the hinge and are truncated by the backthrust.The upper beds form an angular hinge. The backlimbdips roughly 7OSW and the forelimb dips 55NE. Theforelimb does not appear to be thickened or thinnedrelative to the backlimb.

The Elliot 3 exposure, with 1100 m of relief over a mapdistance of 1850 m, gives the most complete profilethrough a fault-propagation fold in the area (Figs 4b &7~). A photolucida (Figs 8) of a photograph taken froman overlook next to the exposed thrust in the PalliserFormation shows the complex geometry and deforma-tion in the fold. The lowest beds in the anticline formangular hinges, with more rounded hinges at intermedi-ate levels. At higher levels, a backthrust on the forelimbcuts the angular hinge region. Several other backthrustsin the forelimb are mostly parallel to bedding exceptwhere they form small ramp anticlines. The upperbackthrust appears to truncate the fold created by thelower backthrust. These faults appear to originate near


7/15

Geometries and modes of fault-propagation folding 327(4 Cline Fold

(4 Elliot 3 ExposureNE

(b) Elliot 2 Exposure8

1OOm(4 Balinhard FoldSW

cNE

Fig. 7. Plunge projections of photogrammetric data defining prominent bedding horizons in the (a) Cline, (b, c) Elliot Peak,and (d) Balinhard folds. Formation key in Fig. 5.

the rounded, synclinal hinge which is complicated bymany smaller-scale faults and folds.The plunge projection (Fig. 7c) clearly shows that theanticlinal hinge geometry changes from angular torounded and then back to angular with height in thefold. The angularity of the upper hinge is caused by theupper backthrust which duplicates strata in the hingeregion. The upper portion of the forelimb appears to beslightly thickened (3%) relative to the more planarbacklimb. The lower portion of the forelimb is thinnedby approximately 20%.In summary, the Elliot Peak fold is a tight fold with asingle axial plane. Both angular and rounded anticlinalhinges occur, with angular hinges commonly associatedwith thrusts and backthrusts which transport extra stratainto the hinge region. The synclinal hinge in the Elliot 3exposure is rounded and complicated by intermediatescale folding and faulting.

Balinhardfold

The Balinhard fold (Figs 4c, 7d & 9) is exposed in a600 m high mountainside. It is probably the most evolvedof the three folds in the Abraham Lake area becauseBalinhard thrust splays cut the forelimb of the fold. Thetwo smaller faults (Fig. 9) are mostly bedding-parallelexcept where they cut bedding at the anticlinal hinge.Bimodal bedding attitudes indicate an angular anticli-nal hinge oriented 146-12 (Fig. 9). Backlimb beddingdips change from 3 1SW near the hinge to 46SW fartheraway from the hinge. Forelimb bedding dips uniformly75NE except in the lowermost, backthrust PekiskoFormation where beds range from 80NE to 8OSW(overturned). A bedding traverse within the BanffFormation (Fig. 10) shows a narrow hinge zone lessthan 40 m wide. The backlimb dip shallows immediatelyadjacent to the hinge, possibly due to a hidden backthrust


8/15

328 E. A. ERSLEV and K. R. MAYBORNBedding N

Fig. 8. Photolucida sketch of the Elliot 3 exposure and stereonets ofbedding and cleavage orientations from the Elliot 1 exposure. Scale baris approximate due to perspective point problems. Formation key inFig. 5.

Bedding Cleavage VeinsN N N

.,.. .,,, Mh\Forelimb NE

Fig. 9. Photolucida sketch of the Balinhard fold showing the beddingtraverse location and stereonets of bedding, cleavage and veinorientations. Formation key in Fig. 5.

adding material to the hinge area similar to the Elliot 2and Elliot 3 exposures.

The Balinhard fold contained the most abundantcleavage of any of the folds studied, with both earlylayer-perpendicular and later axial planar cleavage (Fig.9). The backlimb and hinge areas are dominated byvertical cleavage, with greatest abundance of cleavage atthe hinge where it is axial planar to the fold. Forelimbcleavage is more parallel to bedding and is preferentiallydeveloped in the silty dolomite layers of the BanffFormation. Veins in the Balinhard structure are concen-trated in the forelimb (Fig. 9) where they are commonlybedding parallel and associated with slip surfaces.

The plunge projection for the Balinhard fold (Fig. 7d)reveals uniformly angular fold geometries. There is somerounding in the anticlinal hinges below the faults, but itbecomes angular again above the faults. The interlimbangle for the faulted strata is 71 and there is noappreciable thickness difference between the forelimband the backlimb.Fabric analysis of the Abraham Lake structures

One surprise from the photogrammetric analyses isthat forelimb strata show minimal evidence of penetra-tive deformation. The only exposure in the AbrahamLake area which showed variations in thickness was theElliot 3 exposure, where up to 20% thinning was seen onthe lower forelimb relative to the backlimb. This valueshould be treated with suspicion, however, because thefold axis for the plunge projection was based on relativelyfew photogrammetric measurements of bedding. Adifferent fold axis for this part of the fold or a conicalgeometry could easily explain this apparent thicknessvariation.

Field observations of what appeared to be undeformedfossils on the forelimbs of the folds further suggest lowpenetrative strain. To test this possibility, crinoidalgrainstones were collected for strain analysis. Twentythin sections were cut in the plane of the cross-sections

90iP Middle Banff Fm.c 60

Fig. 10. Apparent dip averages (solid lines) and ranges (shaded areas)of bedding orientations collected on bed-parallel traverses in the BanffFormation (see Fig. 9 for locations). Apparent dips are calculated for

the plane perpendicular to the fold axis.


9/15

Geometries and modes of fault-propagation folding 329and examined for signs of deformation. Poor packing inmany samples made center-to-center strain analysisimpractical, so object ellipticities of crinoidal bioclastswere determined with the mean object ellipse method(Erslev and Ge, 1990). Initial ellipticities were evaluatedfor three orthogonal planes in one apparently unde-formed backlimb sample. The low fabric ellipticities ofthis sample (X/r: 1.14, 1.13, 1.14) suggests low regionalfabric anisotropies probably due to both depositionaland deformation processes.The fabric data in the plane of the cross-sections for 3Abraham Lake localities are plotted in Fig. 11. Details ofthe treatment are given in Mayborn (1993) and are notreproduced here because the results showed that thefabric anisotropies ranged from 1.03 to 1.28, very close tothe regional fabric values. In addition, there was littleconsistency or reproducibility relative to fold position,showing that penetrative strains, as measured by crinoidbioclasts, were probably very low in these rocks.

Cline Strain825

Elliot 1 Strain

Balinhard Strain

Fig. 11. Fabric ellipses derived from mean object ellipse calculations ofcrinoidal grainstones plotted on profiles of the Abraham Lake

structures. Formation key in Fig. 5.

Barrier M ountain fold

Due to their varying degree of development, the foldgeometries from the Abraham Lake area (Fig. 12) can beviewed either as the progressive evolution of a singlestructure or as normal variations in geometry due to localfactors. The trade of space for time suggested by theevolutionary hypothesis is a traditional approach instructural geology and requires that local conditions inthe structures remained the same, with the only differencebeing the amount of slip represented in the individualstructures. The cross-sections in Fig. 12 show anticlinalinterlimb angles tightening from the Cline (74 to 109) tothe Elliot Peak (45 to 59) folds, consistent with non-self-similar fold models. But the fact that the Balinhard foldhas a larger interlimb angle (72 to 84) than the ElliotPeak fold suggests at least a component of localvariability because it is unlikely that the interlimb angleswould get larger with increased displacement.The diversity of fold angularity, from the rounded foldsurfaces in the Cline fold to the wholly angular geome-tries at the Balinhard fold, is difficult to understand. Onepossibility is that rounded folds (e.g. Cline fold) becomemore angular with time, perhaps to accommodate moreshortening in the structure. This would require thestraightening of the forelimb during folding, for whichwe see no evidence. Alternatively, fold curvature could bedue to local effects, such as the position of majorbackthrusts and the local dominance of different modesof folding. To test these possibilities, multiple exposuresof a truncated fault-propagation fold at Barrier Moun-tain, Alberta (Fig. 13) were studied to see whethervariations in fold geometry, like the occurrence ofrounded and angular hinges, are due to local effects orprogressive stages of deformation.The Barrier Mountain fold north of the Panther Riverrepresents an even further evolved structure, with Priceand Mountjoy (1971) interpreting approximately 15 kmof slip on the underlying Clearwater thrust. This studyexamined exposures of the folds hinge area and forelimbnorth of the Panther River in the northeastern side ofBarrier Mountain. Bedding orientations from the hingeand forelimb of the fold (Fig. 13) are scattered in a greatcircle about an average fold axis oriented 322-04. Foldaxes for individual exposures give similar values, indicat-ing a common fold axis for all exposures. The forelimb ofthe fold varies from highly overturned (commonlyaround 45SW) to near vertical (Figs 14 & 15). Areas ofhigh-angle dips in the forelimb show relatively littlebedding plane distortion but the hinge areas are intenselycontorted. Axial planar cleavage is largely limited to theargillaceous Banff Formation in the anticlinal hinge.Veins are most common in the anticlinal hinge andforelimb where they parallel backthrust orientations orformed complex arrays of gash fractures.Fold form changes dramatically along strike. Thesouthern-most profile, exposed in the Panther Rivercanyon (Figs 14a & 15~) shows a highly angular fold,


10/15

33

04

24

2rW

zI

16

E. A. ERSLEV and K. R. MAYBORN

Cline Fold

B Elliot 3 Exposure BN

Balinhard Fold CNE

Fig. 12. Cross-sections through the (a) Cline, (b) Elliot Peak, and (c) Balinhard fault-propagation foldFig. 5. S

Formation key in


11/15

Geometries and modes of fault-propagation folding 331

Fig. 13. Location map for the Barrier Mountain study area with an insetstereonet of bedding orientations measured along the northeast side ofBarrier Mountain.

with hinge distortion by major bedding-parallel fore-thrusts and backthrusts. Within the fold hinge where thestrata rolls over, many bedding parallel surfaces on bothlimbs abruptly ramp upward through bedding, revealingsubstantial layer-parallel slip directed toward the anticli-nal hinge. Further north, however, the fold is morerecumbent and much more rounded (Figs 14b & 15b).Extreme contortion in the Banff Formation within thefold hinge area appears to be related to rotated back-thrusts in the forelimb. These folds are cut by planarfaults emanating from bedding planes in the backlimb.Even further north, the backlimb of the thrust sheetforms a box-like geometry (Fig. 15a), possibly due tofault-bend folding after fault-propagation folding. Theanticlinal fold suddenly becomes angular and the fore-limb is again nearly planar and close to vertical.Photogrammetry was limited to the southern area dueto the availability and angles of aerial photographs (Fig.16). The Panther River exposure shows a rounded andrecumbent fold geometry in the upper Palliser. Above theBanff Formation, the Pekisko Formation is highlyangular, with extra material transported on forethrusts

Fig. 14. Photographs of the (a) Panther River and (b) Ice Lake exposures of the Barrier Mountain fold.


12/15

332 E. A. ERSLEV and K. R. MAYBORN

6000

1 KilometerI IFig. 15. Northeast-trending cross-sections through the Barrier Mountain fold showing the variations in fold geometry along

strike. Formation key and lines of section in Fig. 13.

and backthrusts filling in the hinge area. Just to the north,however, the Pekisko Formation forms a rounded,recumbent fold consistent with field observations at IceLake cirque.

Thin sections from the Barrier Mountain fold confirmthe general lack of penetrative deformation in limestonesfrom the forelimb and hinge. Of 27 thin sections, only 4show measurable deformation: three micritic samples ofPalliser Formation and one crinoidal grainstone from theTurner Valley Formation. The Palliser samples werecollected in the arcuate hinge and forelimb of the PantherRiver and all showed short pressure shadows on pyritecubes indicating minor extension parallel to cleavage.These cleavage-parallel elongations in the more tightlyfolded Palliser Formation do suggest penetrative defor-mation but the irregular distribution of pyrite made asystematic study impractical. Of the crinoidal samplessectioned, only one shows clear penetrative strain, withelongation in the dip-direction of the Clearwater thrust.All other samples, including one taken 5 m from theabove mentioned sample, show no clear evidence ofmeasurable strain anisotropy. This indicates minimalpenetrative strain in the forelimb, with localized zones ofhigher strain whose scale would necessitate much moredetailed sampling for an adequate definition of the strain.

Our unsuccessful attempts to measure penetrative strainsin the planar forelimb and angular hinge areas of thesefault-propagation folds confirms the dominance offlexural slip deformation observed in the field anddeduced from the lack of significant changes in unitthicknesses.

9cmBarrier Mountain Fold, Alberta

8ooO- Biz6 Cirque-z$

7ooo- Panther River

0 2ooo 4000Horizontal Distance ft)

Fig. 16. Photogrammetric profiles through the Panther River exposureand folded Pekisko Formation near Ice Lake.


13/15

Geometries and modes of fault-propagation folding 333DISCUSSION

Fault-propagation folds from the Canadian thrust beltshow a variety of geometries ranging from recumbent,curved folds to more open, highly-angular folds. In all thecases studied here, anticlinal hinges define single planeswhich diverge upward from zones of faulting and tightfolding in the Palliser Formation. Anticlinal hinge areascommonly show highly contorted beds due to foldtightening of the hinge and transport of extra strata intothe hinge by bedding-parallel thrusts and backthrusts.Comparison of the intense deformation in anticlinalhinges with the minimal deformation in adjacent fore-limbs indicates that anticlinal hinges were fixed duringfolding because it is difficult to imagine a process whichcould unfold the contorted hinge areas into planarforelimbs.Cross-sections from the Abraham Lake area (Fig. 12)generally show a progressive tightening of the unfaultedstrata, from the rounded Cline fold, with interlimbanticlinal angles of 74 to 109, to the Elliott Peak foldwith interlimb angles of 45 to 59. Interlimb angles forfault-truncated strata also vary, from 72 to 84 in theBalinhard fold to 27 to 70 in the Barrier Mountain fold.The variability of fold geometries along strike in theBarrier Mountain fold shows that these geometries arenot simply a function of fault angle, fault slip and/orlithologic controls as postulated by some models of fault-propagation folding.Internal deformation in the folds is dominated byflexural slip on bedding planes, with thrust motionsdirected toward anticlinal hinges. Photogrammetricmeasurements show minimal consistent differencesbetween forelimb and backlimb thicknesses. Bed attenua-tion is relatively rare, with local thickening common inanticlinal hinges due to thrust duplication of strata in thehinge areas. Penetrative strain is minimal and mostlylimited to cleavage-related flattening in fold hinges andlocalized simple shear zones related to thrust propagationand flexural shear on fold limbs. The prevalence ofbedding-parallel slip indicates that overall distortionsmay be quite large, particularly in the forelimb and hingeareas, but the scale of strain heterogeneity would havenecessitated much more detailed sampling in order toquantify these distortions.Comparisons of the fold geometries with previouskink-band and oblique shear models (Figs 1 & 2) showthat no individual model can adequately explain thekinematic development of these folds. Some arcuate foldsurfaces can be explained by an initial stage of bucklefolding, but the curved, recumbent surfaces of the BarrierMountain fold indicates that curved fold surfaces are alsogenerated during fault-propagation folding. Obliqueshear models (Fig. 2) such as trishear can explain thearcuate anticlinal (e.g. Panther River exposure of theBarrier Mountain fold) and synclinal (Cline and Elliot 3exposures) fold surfaces in the massive Upper PalliserFormation. But they are contradicted at higher structural

levels by the highly angular fold surfaces and minimalstrains recorded in the forelimbs of these structures. Aninherent problem with applying the trishear model tostructures involving well-bedded sedimentary strata is theinability of the trishear algorithm to model the beddinganisotropy responsible for flexural slip folding.In any case, the narrow zones of faulting at depth needto be linked with the angular folding in the upper levels ofthe structures. We suggest that distributed shear in atriangular shear zone can bridge this gap, with roundedfold surfaces occurring where massive units deform byoblique shear, and with angular fold surfaces occurringwhere better-layered units deform by flexural slip.The dominance of angular fold surfaces at higher levelsin the folds suggests the application of kink-band fault-propagation fold models (Fig. 1). But these modelspredict box-like anticlines with multiple, migratinghinges in unfaulted strata. These geometries are incon-

sistent with the single, fixed anticlinal fold hinges in theunfaulted strata of the Cline and Elliot Peak folds as wellas other examples from Williams and Chapman (1983)and Fisher and Anastasio (1994). Self-similar folddevelopment, where fold shapes and angles remain thesame as kink-band migration creates a bigger structure, iscontradicted by the variability of forelimb orientationsand evidence for fold tightening with increasing struc-tural displacement. In addition, the curvature of manyfold surfaces is not predicted by kink-band models.One alternative to explain the angular folds is to modelthem with progressive kink folding by flexural slip (Fig.17). In this model, analogous to the fixed-hinge chevronfolding of Stewart and Alvarez (1991), fixed fold hingescause progressive limb rotation during folding until thelimbs lock, requiring thrusting either through the fore-limb or along a fold hinge to accommodate additionaldeformation. This model requires flexural slip through-out the section, with synthetic thrusting on backlimbbedding planes and antithetic backthrusting on forelimbbedding planes. In addition, synthetic thrusting forwardof the synclinal axis is generated by the progressive(clockwise in Fig. 17) rotation of the synclinal hinge lines.As a result, progressive kink folding provides a mechan-ism for transmitting regional layer-parallel shear throughfolds in the direction of general tectonic transport.The combination of triangular shear zone folding nearpropagating thrusts with progressive kink folding athigher levels is illustrated in Fig. 18. Figure 18(b) wasgenerated by balancing a generalized version of the Elliot3 cross-section (Fig. 12b) and rotating it so that footwallbedding is horizontal. The angular anticlinal foldgeometry in the upper strata combined with our inter-pretation of fixed anticlinal hinges requires flexural slipand/or flexural shear to transport material into theanticlinal hinge area, consistent with our field observa-tions. A delicate balance between the competing foldmechanisms appears to govern the location of thetransition from triangular shear zone folding to progres-sive kink folding. Variability in the stratigraphic level of


14/15

E. A. ERSLEV and K. R. MAYBORN

Fig. 17. Schematic diagram showing progressive kink-band folding andthe necessity for flexural slip on all fold limbs.

(a>

I after Elliot Peak Fold)

Hinge-aiding

Fig. 18. Balanced, multi-mode fault-propagation fold model for thin-skinned thrust belts based on the Elliot 3 exposure (b), retrodeformed toan earlier configuration (a), and truncated by a subsequent break-thrust

(c).

this transition within individual structures can explainwhy the Pekisko Formation in the Barrier Mountain foldforms both angular and rounded fold surfaces.

An increment of retrodeformation applied to the ElliotPeak section (Fig. 18b) gives a possible earlier stage ofdeformation (Fig. 18a), which is similar to manydetachment fold geometries (Homza and Wallace,1997). Further progressive kink folding of the ElliotPeak structure, as modeled in Fig. 18(b), would probablytighten the folded layers uniformly until they locked andbroke, resulting in break-thrust folding (Fig. 18~). Thus,different modes of fault-propagation folding shouldresult in different fault-propagation rates. In triangularshear zone portions of folds, the master thrust maypropagate proportional to fault slip. In the areas ofprogressive kink folding, where folds tighten uniformlyand lock, the master thrust may propagate very rapidly asa break thrust.

Progressive kink folding requires synthetic flexural slipin the backlimb and thus the presence or absence ofregional flexural slip may control its development. In thecase of the curved fold surfaces in the Cline fold, if foldingoccurred with relatively little backlimb flexural slip (asindicated by field observations), then progressive kinkfolding may have been inhibited, resulting in a morearcuate fold geometry. Similarly, the general lack ofregional bed-parallel slip in the basement-involved foldsof the Laramide Orogen (U.S.A.) may have suppressedprogressive kink folding in the many examples of base-ment-cored folds with arcuate fold hinges. Thus, thepresence (in many thin-skinned structures) or absence (inmany thick-skinned structures) of regional bed-parallelslip may be responsible for some of the differences infault-propagation fold geometries from these orogens.

These two modes of fault-propagation folding, trian-gular shear zone folding and progressive kink folding, arecompetitive processes, with their development dependenton the degree of bedding anisotropy and presence ofregional bedding-parallel slip. In addition, buckling bothbefore and during fault-propagation folding undoubt-edly has important contributions to fold geometries.Additional fold mechanisms generating the self-similarkink-band geometries of Suppe and Medwedeff (1990)may also occur, perhaps at shallower levels in the crust,but no evidence for collaborative geometries was seen inthis study. For the areas studied here, self-similar kink-band models are useful as reproducible approximationsthat allow regional extrapolations. However, detailedanalyses of reservoir and fracture geometries in asym-metric thrust-cored folds will need to consider thecombination of triangular shear zone folding andprogressive kink folding which occur in fault-propaga-tion folds of the Canadian thrust belt.

Acknowledgements-Reviews comments and suggestions by DavidAnastasio, Peter Hennings, Christopher Hedlund, and G. J. Rait weregreatly appreciated. Steve Schultz assisted with field work andpetrographic analyses. We are indebted to Henry Charlesworth and


15/15

Geometries and modes of fault-propagation folding 335Philip Simony for suggesting the field areas in the Canadian Rockies.Charles Pillmore and Jim Messerich at the U.S.G.S. PhotogrammetricPlotter Lab in Denver provided access to and help with photogram-metric equipment. This research was supported by the donors of thePetroleum Research Fund, administered by the Petroleum ResearchFund, and N.S.F. grant EAR-9304547. Additional funding wasprovided by grants from ARCO, Chevron and the Geological Societyof America.

REFERENCESAlonso, J. L. and Teixell, A. (1992) Forelimb deformation in somenatural examples of fault-propagation folds. In Thrust Tectonics ed.K. R. McClay, Chapman and Hall, London, 17>180.Berg, R. R. (1962) Mountain flank thrusting in Rocky Mountainforeland, Wyoming and Colorado. Bulf etin of the American Associa-

tion of Petroleum Geologists 46 2019-2032.Berg, R. R. (1976) Deformation of Mesozoic shales at Hamilton Dome,Bighorn basin, Wyoming. Bull etin of the American Association ofPetroleum Geologists 60 1425-1433.Brown, W. G. (1988) Deformation style of Laramide uplifts in theWyoming foreland. In In teraction of the Rocky Mountain For elandand the Cordill eran Thr ust Belt eds C. J. Schmidt and W. J. Perry, Jr,Geological Society of America Memoir 171, 53-64.Chase, R. B., Genovese, P. W. and Schmidt, C. J. (1993) The influenceof Precambrian rock compositions and fabrics on the development ofRocky Mountain foreland folds. In Basement-Cover Ki nematics ofLaramide Foreland Uplifts eds C. J. Schmidt, R. B. Chase and E. A.Erslev, Geological Society of America Special Paper 280,45-72.Chester, J. S. and Chester, F. M. (1990) Fault-propagation folds abovethrusts with constant dip. Journ al of Structural Geology 12 903-910.Dahlstrom, C. D. A. (1970) Structural geology in the eastern margin ofthe Canadian Rocky Mountains. Bull etin of the Canadian PetroleumGeologists 46 332-406.Dietrich, D. and Casey, M. (1989) A new model for the Helveticnappes. In Alpine Tectonics eds M. P. Coward, D. Dietrich and R.G. Park, Geological Society Special Publication 45,47-63.Dixon, J. M. and Liu, S. (1992) Centrifuge modelling of the propaga-tion of thrust faults. In Thrust Tectonics ed. K. R. McClav, DD 53-70. Chapman and Hall, London.

-.__Elliott, D. W. (1976) The energy balance and deformation mechanismsof thrust sheets. Phil osophical Tr ansactions of the Royal Society of

London A283 289-3 12.

Erslev, E. A. and Ge, H. (1990) Least-squares center-to-center andmean object ellipse fabric analysis. Journ al ofStructural Geology 121047-1059.

Erslev, E. A. (1991) Trishear fault-propagation folding. Geology 19617-620.

Erslev, E. A. and Rogers, J. L. (1993) Basement-cover geometry ofLaramide fault-propagation folds. In Laramide Basement Deforma-tion in he Rocky Mountain For eland of the Western Uni ted States edsC. J. Schmidt. R. B. Chase and E. A. Erslev. Geoloaical Societv ofAmerica Special Paper 280,125-146. Evans, J. P. (1993) Deformation mechanisms and kinematics of acrystalline-cored thrust sheet: the Washakie thrust system, Wyoming.In Lar amide Basement Deformation in the Rocky Moun tain F orelandof the Western Un itedStates eds C. J. Schmidt, R. B. Chase and E. A.Erslev, Geological Society of America Special Paper 280, 147-162.Fischer, M. P., Woodward, N. B. and Mitchell, M. M. (1992) Thekinematics of break-thrust folds. Journ al of Structural Geology 16451460.Fisher, D. M. and Anastasio, D. J. (1994) Kinematic analysis of a large-scale leading edge fold, Lost River Range, Idaho. Journal of Struc-tural Geology 16 337-354.Heim, A. (1919) Geologie der Schweiz. Tauchnitz, Leipzig.Hennier, J. and Spang, J. H. (1983) Mechanisms for deformation ofsedimentary strata at Sheep Mountain anticline, Bighorn Basin,Wyoming. Wyoming Geological Association Guidebook 34 961ll.Homza, T. X. and Wallace, W. K. (1997) Detachment folds with fixedhinges and variable detachment depth, northeastern Brooks Range,Alaska. Journ al of Structur al Geology 19 3377354.

Jamison, W. R. (1987) Geometric analysis of fold development inoverthrust terranes. Journ al of Structural Geology 9 207-220.Mayborn, K. R. (1993) Geometry and deformation of three fault-propagation folds in the Front range of the Canadian thrust belt.Unpublished M.Sc. thesis, Colorado State University, Fort Collins,Colorado.McConnell, D. A. (1994) Fixed-hinge, basement-involved fault-propa-gation folds, Wyoming. Bull etin of the Geological Society of America106 1583-1593.McConnell, D. A. and Wilson, T. G. (1993) Linkage between deforma-tion of basement rocks and sedimentary rocks in basement-involvedfolds. In Basement-Cover Kinematics of Lar amide Foreland Uplif ts.eds C. J. Schmidt, R. Chase and E. A. Erslev, Geological Society ofAmerica Suecial Paoer 280.319-335.McMechan, M. E. and Thompson, R. I. (1989) Structural style andhistory of the Rocky Mountain fold and thrust belt. In WesternCanada Sedimentar y Basin ed. B. D. Ricketts. Canadian Society ofPetroleum Geologists, pp. 47-72.Mitra, S. (1990) Fault-propagation folds: Geometry, kinematic evolu-tion, and hydrocarbon traps. Bull etin of the American Association ofPetroleum Geologists 74 921-945.Mosar, J. and Suppe, J. (1992) Role of shear in fault-propagationfolding. In Thrust Tectonics. ed. K. R. McClav. DD 123-132. Chao-man and Hall, London. - . .

Mountjoy, E. W., Price, R. A., Aitken, J. D. and Cook, D. G. (1974)Geologic map of Whiterabbit Creek (west half), Alberta. GeologicalSurvey of Canada map 1389A.Narr, W. and Suppe, J. (1993) Kinematics of basement-involvedcompressive structures. American Journ al of Science 294 302-360.Pillmore, C. L. (1989) Geological photogrammetry in the U.S. Geolo-gical Survey. Photogrammetri c Engi neering and Remote Sensing 551185-1189.Price, R. A. and Mountjoy, E. W. (1971) Geology Barrier Mountain(east half). Geological Survey of Canada map 1273A.Ramsay, J. G. and Huber, M. I. (1987) The Techniques of ModernStructur al Geology. Vol. 2. Folds and Fau lts. Academic Press,London.Schmidt, C. J., Genovese, P. W. and Chase, R. B. (1993) Role ofbasement fabric and cover-rock lithology on the geometry andkinematics of twelve folds in the Rocky Mountain foreland. InLar amide Basement Deformation in the Rocky M ountain For eland ofthe Western Un ited States eds C. J. Schmidt, R. B. Chase and E. A.Erslev, Geological Society of America Special Paper 280, 144.Shaw, E. W. (1963) Canadian Rockies-orientation in time and space.In Backbone of the Ameri cas eds 0. E. Childs and B. W. Beebe, pp.23 l-242.Spang, J. H., Evans, J. P. and Berg, R. R. (1985) Balanced cross-sections of small fold-thrust structures. The Mountain Geologist 224146.Stewart, K. G. and Alvarez, W. (1991) Mobile-hinge kinking in layeredrocks and models. Journ al of Structural Geology 13 243-259.Stone, D. S. (1984) The Rattlesnake Mountain, Wyoming, debate: areview and critique of models. The Mountain Geologist 21 3746.Stone, D. S. (1986) Seismic profile: South Elk Basin. In SeismicExpression Structural Styles eds A. W. Bally, Vol. 3, pp. 2.2G2.24.American Association of Petroleum Geologists. Studies in Geology15.Storti, F. and Salvini, F. (1996) Progressive rollover fault-propagationfolding: A possible kinematic mechanism to generate regional scalerecumbent folds in shallow foreland belts. Bull etin of the AmericanAssociati on of Petroleum Geologists 80 174-l 93.Suppe, J. (1985) Principl es of Structur al Geology. Prentice Hall, Engle-wood Cliffs, New Jersey.Suppe, J. and Medwedeff, D. A. (1984) Fault-propagation folding.Geological Society of Ameri ca Abstracts with Progr ams 16 670.Suppe, J. and Medwedeff, D. A. (1990) Geometry and kinematics offault-propagation folding. Ecl ogae Geologicae Helvetiae 83 409-454.Williams, G. and Chapman, T. (1983) Strains developed in the hangingwalls of thrusts due to their slip/propagation rate; a dislocationmodel. Journ al of Structur al Geology 5 563-571.Willis, B. and Willis, R. (1934) Geological Structures. McGraw-Hill,New York.

Documents

14 Erslev 1997 Journal of Structural Geology