Extensional Fault-Bend Folding and Synrift Deposition: An Example from the Central Sumatra Basin, Indonesia*

John H. Shaw¹, Stephen C. Hook¹, and Edward P. Sitohang²

Search and Discovery Article #40004 (1999)

¹Texaco Exploration and Production Technology Dept., 3901 Briarpark, Houston, Texas 77042.
²PT CALTEX Pacific Indonesia, Rumbai, Pekanbaru 28271, Indonesia.

* Published in AAPG Bulletin, V. 81, No. 3 (March 1997), P. 367-379; Figures 2,4, and 10 revised for online presentation.

Abstract

We describe an analytical method for interpreting the geometry and structural history of asymmetric half grabens in rift basins with extensional fault-bend fold theory. Using seismic reflection profiles from the Central Sumatra basin and balanced forward models, we show how local variations in tectonic subsidence relative to deposition rates yield distinctive patterns of folded synrift strata and unconformities that record basin history. If the deposition rate exceeds the local subsidence rate, folded growth strata form upwardly narrowing kink bands that have been described previously as growth triangles. In contrast, if the deposition rate is less than or equals the local subsidence rate, growth strata are folded and truncated at the surface on half-graben margins. Subsequent increases in deposition rate relative to subsidence rate form angular unconformities near the half-graben margins. These unconformities develop without the necessity of erosion and are folded by continuing fault slip. Strata above and below the unconformities are concordant in the deeper parts of the half grabens. Thus, angular unconformities on half-graben margins are helpful for defining sequence boundaries that may reflect changes in deposition and tectonic subsidence rates. In addition, fault-bend fold interpretations yield fault geometry and measures of horizontal extension, both of which control three-dimensional half-graben geometry and accommodation space. We show how along-strike variations in fault geometry produce intrabasinal structures that may form prospective fairways or local depocenters.

Introduction

Half grabens form during crustal extension that is accommodated by normal faults, which commonly flatten with depth, causing collapse of the hanging wall and formation of inclined rollover panels (Hamblin, 1965). Many workers have presented geometric and physical models of hanging-wall collapse along vertical or steeply dipping shear surfaces (e.g., Gibbs, 1983; Jackson and Galloway, 1984; White et al., 1986; Rowan and Kligfield, 1989; Groshong, 1990; Nunns, 1991; White and Yielding, 1991; Withjack et al., 1995), including Coulomb shear along active fold hinges (Xiao and Suppe, 1992). The theory of Xiao and Suppe (1992) described how these active fold hinges, called active axial surfaces, are pinned at depth to fault bends and extend upward through prerift and synrift sections. As strata pass through these active axial surfaces due to fault slip, they are deformed into kink bands or inclined rollover panels. In areas of continuously curved or listric normal faults, where fault geometry can be strongly affected by sedimentary compaction, hanging-wall shear is generally distributed throughout the hanging-wall block (Figure 1A). Typically, syntectonic hanging-wall strata thicken gradually and fan toward the fault. In contrast, slip along normal faults composed of two or more planar segments produces hanging-wall shear along discrete axial surfaces related to fault bends (Figure 1B). Above faults composed of planar segments, rollovers are composed of planar segments, and growth strata thicken abruptly toward the fault above rollover panels.

Figure 1 (A) Broadly curved rollover panel above a listric normal fault (from Xiao and Suppe, 1992). (B) Planar rollover panels above a normal fault that is composed of planar segments. Rollover panels are bounded on the left by active axial surfaces, which are pinned to fault bends.

We present examples from Indonesia of extensional fault-bend folds above faults composed of roughly planar segments in which discrete axial surfaces are consistent with rollover formation by Coulomb collapse (Xiao and Suppe, 1992). We use Xiao and Suppe's (1992) theory to interpret the geometry and kinematic history of these structures, thus contributing a well-documented example of an extensional fault-bend fold in a basement-involved rift system with lacustrine strata. This structural and depositional environment contrasts with the gravity-driven normal faults and dominantly marine sediments of the Gulf Coast margin, thus establishing a broader applicability for extensional fault-bend fold theory (Xiao and Suppe, 1992). In addition, we explore the effects of extensional fault-bend folding on patterns of synrift strata where deposits both overfill and underfill the developing accommodation space.

Patterns of folded synrift strata and unconformities record ratios of deposition relative to tectonic subsidence rates and are used to model basin development. Finally, we show how measures of horizontal extension and fault geometry derived in the fault-bend folding analysis can be used to infer three-dimensional basin architecture.

Extensional Fault-Bend Folding

Purely rigid-block translation of the hanging wall over a normal fault that flattens with depth produces a large void between fault blocks that cannot be supported at depth. Collapse of hanging walls into these voids forms inclined fold limbs or "rollovers" above nonplanar faults (Hamblin, 1965); these rollovers have been observed in rift basins worldwide (e.g., Bally, 1983; James, 1984; Nunns, 1991). Xiao and Suppe (1992) modeled this hanging-wall collapse by Coulomb shear along inclined axial surfaces (Figure 2). During progressive fault slip, the hanging wall is sheared through active axial surfaces that are pinned to bends in the fault.

Figure 2 Extensional fault-bend fold models developed above a normal fault that flattens with depth (after Xiao and Suppe, 1992). (A) Incipient fault with a concave-upward bend; (B) antithetic hanging-wall rollover panel developed by shearing along the active axial surface in response to fault slip; (C) additional fault slip widens the rollover panel, which narrows upward into syntectonic (growth) strata forming a growth triangle. Growth strata deposited in the hanging-wall block are thickest above the more inclined fault segment.

This hanging-wall shear along active axial surfaces is often accommodated in rocks and analog models by secondary faults that form above bends in the master normal fault and are translated away from these bends by slip on the underlying detachment (Dula, 1991; McClay and Scott, 1991; Xiao and Suppe, 1992; Withjack et al., 1995). Above concave-upward fault bends where the dip of the fault lessens with depth, rollover panels and active axial surfaces are generally oriented antithetic to the master fault (Figure 2) (Xiao and Suppe, 1992). Rollover panels are bounded by active axial surfaces and by parallel inactive axial surfaces, which mark the rocks that were initially along the active axial surface prior to fault slip.

During progressive fault slip, inactive axial surfaces are translated away from active axial surfaces and, thus, intervening rollover panels widen as fault slip increases (Figure 2).

The geometries of normal faults and associated rollover panels control the size and shape of accommodation spaces in half grabens where sediments can be deposited. Hanging-wall subsidence induced by fault slip produces an accommodation space in the half graben, which is defined by the maximum structural relief of the top pregrowth horizon between hanging-wall and footwall blocks (Figure 3A). In situations where sediments evenly fill or overfill this accommodation space above a single normal fault that flattens with depth, synrift strata are thickest above the most inclined fault segments and thin in the direction that the fault shallows (Figure 2). If normal faults are composed of two or more planar segments, separate compartments develop above each fault segment in the half grabens (Figure 3B). Compartments are separated from each other by active axial surfaces that are pinned to fault bends. Each half-graben compartment has a distinct subsidence rate that is controlled by the dip of the underlying fault segment.The rollover accommodation space is defined by the structural relief across a rollover panel, which represents the maximum structural relief between adjacent half-graben compartments (Figure 3B).

Figure 3 Extensional fault-bend fold models showing basin compartments and accommodation spaces developed above normal faults that flatten with depth. (A) Half-graben accommodation space is the area defined by the maximum structural relief of the top pregrowth horizon between hanging-wall and footwall blocks and the half-graben shape. (B) Hanging-wall compartments 1 and 2 correspond to the two segments of the underlying fault and are separated by the active axial surface, which is pinned to the fault bend. Each basin compartment has its own subsidence rate induced by fault slip that is controlled by the dip of the underlying fault segment. The rollover accommodation space is the area defined by the maximum structural relief between adjacent compartments and the half-graben shape.

If sediments are coevally deposited in adjacent compartments (i.e., sediments are deposited on both sides of an active axial surface), sediments overfill the rollover accommodation space (see Figure 4C, compartments 1 and 2). Due to differences in subsidence rate and rollover accommodation space, deposition rates and strata thicknesses typically change between adjacent half-graben compartments. During progressive fault slip, however, strata are translated between compartments as they migrate through active axial surfaces. In addition to this translation, strata are folded around active axial surfaces and incorporated into kink bands or rollover panels that widen with progressive fault slip. Rollover widths of growth strata reflect the amount of fault slip that has occurred since their deposition. Sediments deposited early in the rift history, therefore, record wider rollover widths than do sediments deposited later. As a result, these syntectonic strata form upwardly narrowing rollover panels called growth triangles (Figure 4) (Xiao and Suppe, 1992).Growth triangles are bounded by active axial surfaces and inactive axial surfaces in growth strata, which are called growth axial surfaces. Growth axial surfaces record the positions of sediments initially deposited along active axial surfaces and, therefore, record paleoboundaries between adjacent half-graben compartments. As a result of different subsidence and deposition rates between compartments, strata abruptly change thickness across growth axial surfaces (Figure 4).

Figure 4 Sequential models (A-D) of half-graben development above a normal fault that flattens to horizontal through two bends. In (B) and (C), growth strata slightly underfill the half-graben accommodation space and are folded by active axial surfaces (green dashed lines). Growth sediments are deposited in compartments 1 and 2, and form a distinct growth triangle above rollover panel 1. However, in (B) and (C), sediments are not deposited in compartment 3; growth strata in compartment 3 have been folded and translated to their present positions and crop out at the surface. In (D), growth strata overfill the half-graben accommodation space, forming an angular unconformity above rollover panel 2 in compartment 3; however, strata above and below the unconformity become concordant in the deeper parts of the basin.

Note that the growth axial surface above rollover panel 1 dips more steeply in strata that overfilled the half-graben accommodation space and dips more gently in strata that slightly underfilled the half-graben accommodation space.

In contrast, where sediments underfill or exactly fill the rollover accommodation space, other fold geometries result (see Figure 4C, compartments 2 and 3). Under these conditions, deposition is confined to the more rapidly subsiding compartment 2, which is separated from the adjacent compartment 3 by an active axial surface (Figure 4). Growth strata deposited in the more rapidly subsiding compartment 2, however, are translated into the adjacent compartment 3 due to horizontal motion of the hanging wall; this motion is induced by fault slip. As these growth strata are sheared through the active axial surface, they are folded into the rollover panel and crop out in angular fashion at the surface. Although subsequent erosion may further alter the geometry of growth strata at the surface, the angular exposure is initially developed by folding and translation of strata into areas of nondeposition. Subsequent deposition of either postrift or synrift sediments above the truncated growth strata generates an angular unconformity. Typically, angular unconformities are interpreted to reflect distinct periods of deformation, erosion, and then deposition; however, the growth fault-bend fold models in Figure 4 demonstrate that angular unconformities can develop in half grabens without erosion or a hiatus in deformation due to increases in deposition rate relative to subsidence rate, where half-graben compartments change from sediment-underfilled to overfilled conditions. In Figure 4D, strata both above and below the angular unconformity are syntectonic and become concordant in the deeper parts of the half graben.

Examples from the Central Sumatra Basin

Figure 5 Map showing the location of the Central Sumatra basin on the Island of Sumatra, Indonesia.

In the Central Sumatra basin (Figure 5), growth triangles and unconformities, similar to those generated in our fault-bend fold models, are observed in seismic images of Tertiary lacustrine, fluvial, and marine strata (Figure 6). Using a trough in Central Sumatra as an example for our model, we interpret the structural geometry and history of half grabens as extensional fault-bend folds. Distinct axial surfaces separating inclined from near-horizontal strata in this basin (Figure 6) suggest that the underlying normal faults are composed of planar segments. Furthermore, the migrated seismic reflection profiles in Figure 6 image strata above and below the angular unconformities that become concordant toward the center of the troughs. Based on extensional fault-bend fold models (Figure 4), these lateral changes from discordant to concordant strata suggest significant increases in deposition rates relative to subsidence rates through time. Collectively, these patterns of folded strata enable us to decipher the structural and depositional history of these half grabens using extensional fault-bend fold theory (Xiao and Suppe, 1992).

Figure 6 Examples of growth triangles and angular unconformities in half grabens that are imaged in migrated seismic reflection profiles from the Central Sumatra basin. Similar growth triangles and unconformities are modeled in Figure 4 and are used to decipher the underlying fault geometry and structural history of the basin. Note how strata above the angular unconformities in the east become concordant to the west in the deeper parts of the half grabens. Datum (0 km) is sea level.

The migrated seismic reflection profile in Figure 7 , which is displayed in depth, images a half graben in the Central Sumatra basin where Oligocene strata thicken westward above an east-dipping normal fault that is locally defined by a prominent fault-plane reflection. In the uppermost part of the synrift section, at least three axial surfaces separate horizontal strata on the left (west) from inclined strata in rollover panels on the right (east) (Figure 7B). In the extensional fault-bend fold models (Figure 4), the steeply dipping axial surfaces that deform the synrift section are pinned at depth to bends in the basin-forming normal fault. Therefore, we interpret these fold hinges in Figure 7 as active axial surfaces that are each pinned at depth to a discrete bend in the underlying normal fault. Active axial surfaces are best located by identifying changes in the dip of reflections in the uppermost growth sequences (Figures 6, 7). These dip changes should be consistent with the sense of simple shear induced by the fault bend. For the concave-upward fault shapes described here, the bed dip should be antithetic to the fault dip and should steepen in the direction that the fault deepens. Alternatively, convex-upward fault bends may yield panels that are synthetic to the fault dip (Xiao and Suppe, 1992). In Figure 7, we extended the westernmost active axial surface downward through fold hinges and used this orientation, which likely reflects the Coulomb shear angle (Xiao and Suppe, 1992), to help define the other, more poorly imaged axial surfaces. In other cases, fold hinges may be more curved and less discrete if the fault bends also are curved. Thus, a range of axial surface dips (inclined shear orientations) should be tested (e.g., White et al., 1986; Groshong, 1990). Moreover, subtle dip changes in rollover panels are generally produced from subtle changes in fault dip or other processes (e.g., differential compaction), and the interpreter must decide upon the appropriate resolution of structural dip changes.

Only one segment of the fault is defined by a fault-plane reflection on the seismic profile in Figure 7; the adjacent fault segments are not imaged. The dips of the folded strata, the imaged fault segment, and the axial surfaces, however, can be used to predict the complete fault shape (Groshong, 1990; Dula, 1991; Xiao and Suppe, 1992). In extensional fault-bend folds, the magnitude of deflection of strata in a rollover panel is equal to the magnitude of fault deflection over the same width measured along the hanging-wall shear (axial surface) orientation (Figure 8). In Figure 8, reflections define an axial surface dip of 66°W and a bed dip of approximately 15°W in the westernmost rollover panel. Basing our prediction on the direction and magnitude of deflection of strata in the kink band, we believe that the fault steepens to a dip of about 39°E in the region west of the fault-plane reflection. Similar analyses for the remaining fault segments yield the entire fault trace on the seismic profile (Figure 7B). Along YY' in Figure 7, the fault consists of several segments that generally flatten with depth to a near-horizontal detachment.

Figure 8 Fault geometry derived from rollover shape. Enlarged portion of the seismic line in Figure 7 annotated with a folded horizon, an active axial surface, and the lower fault segment (solid red line) based on fault-plane reflections (see Figure 7A). The dip of the upper fault segment (dashed red line) is derived from the rollover geometry. The deflection of the folded horizon (1), measured at a distance (L) along a line parallel to the hanging-wall shear (axial surface) orientation, equals the deflection of the fault (2) along the same line. Depth is below sea level; horizontal scale equals vertical scale.

By identifying active axial surfaces and determining fault geometry, we have defined two basic geometric elements of extensional fault-bend folds. Also significant, however, are the positions of inactive axial surfaces, which define the widths of rollover panels. In Figure 7, inactive axial surfaces in synrift section (growth axial surfaces) are readily observed in the uppermost synrift section because they bound dip domains of growth triangles. In contrast, the positions of inactive axial surface in pregrowth or basement sections are not as apparent. A fundamental relation between kink-band width and fault slip, however, enables us to define the positions of inactive axial surfaces. In extensional fault-bend folds above a single normal fault that flattens with depth, the true widths of all antithetic kink bands or rollover panels are the same (Figure 9). This kink-band width is a measure of horizontal extension and records the offset of any pregrowth horizon measured between lines parallel to the hanging-wall shear (axial surface) orientation that are pinned to correlative hanging-wall and footwall cutoffs (Figure 9). In the example from the Central Sumatra basin (Figure 7), we define the fault offset of the top pregrowth (basement) horizon based on the fault shape and reflections tied from well control. Based on the fault-bend fold models, this fault offset measured between lines parallel to the hanging-wall shear orientation equals the width of all antithetic kink bands developed in the half graben. Therefore, the offset of the top basement horizon across the normal fault in Figure 7 can be used to define the positions of inactive axial surfaces in the rest of the half graben.

Figure 9 An extensional fault-bend fold model with two rollover panels developed above bends in a normal fault that flattens to a horizontal detachment. The widths of both rollover panels are the same and are equal to the horizontal extension on the detachment, although slip on each fault segment varies slightly based on fault dip. Rollover widths are also equal to the horizontal offset of any pregrowth horizon (e.g., bed X) across the fault measured between hanging-wall and footwall cutoffs projected along the hanging-wall shear (axial surface) orientation.

The recognition of axial surface shapes and positions in growth and pregrowth sections, along with the determination of fault geometry, describes the trough imaged in Figure 7 as a half graben developed by extensional fault-bend folding. Proper application of fault-bend folding theories yields area-balanced and retrodeformable interpretations (Suppe, 1983; Xiao and Suppe, 1992). Retrodeformable sections can be kinematically restored to a reasonable, predeformation state without changes in rock area. To demonstrate the internal consistency of the interpretation in Figure 7B, we generate a balanced-forward model of the trough in Figure 10 using the fault geometry, compacted stratigraphic thicknesses, and shear (axial surface) orientation observed in the seismic profile. The retrodeformable model conserves rock area, avoids gaps between fault surfaces by shear along active axial surfaces, and forms rollover panels that have widths related to fault slip. The final stage of the model in Figure 10 depicts all the major structural elements of the trough imaged in Figure 7, including the shape of the graben, the three growth triangles, and the angular unconformity between Pematang and Sihapas strata. The consistency between the geometries of the reflections and interpretation in Figure 7B and the final model in Figure 10 indicates that our interpretation of the trough as an extensional fault-bend fold is internally consistent and viable.

The first stage of the sequential forward model in Figure 10 depicts the incipient normal fault and active axial surfaces prior to fault slip. Each stage includes deposition of a major stratigraphic unit with fault slip recorded by the width of the folded synrift strata. In the second through fourth stages, Pematang synrift strata are generally confined to the trough and alternatively fill and slightly underfill the half-graben accommodation space. In the final stage of Figure 10, lowermost Sihapas synrift strata are deposited everywhere and overfill the half-graben accommodation space. This change from underfilled to overfilled conditions implies an increase in deposition rate relative to subsidence rate between Pematang and Sihapas sections.

Figure 10 Sequential, kinematic models of the development of the trough imaged in Figure 7. Fault geometry, shear (axial surface) orientation, and compacted stratigraphic thicknesses were taken from Figure 7B and used to construct these forward models. The initial stage represents basement geometry prior to rifting, and subsequent model stages represent the structural and stratigraphic geometries of the trough at the end of deposition of a formation. The final model (stage 5) depicts folded syntectonic sediment patterns (growth triangles and an unconformity) similar to those observed and interpreted on the seismic profile in Figure 7.

This increase generates an angular unconformity on the basin margin even though strata above and below the unconformity become concordant in the deeper part of the trough. In addition, the dips of the growth axial surfaces reflect this ratio of deposition rate relative to subsidence rate. In the Pematang section, which had a relatively low deposition rate relative to subsidence rate, the growth axial surface has a shallow dip (Figure 7). In contrast, the growth axial surface dips more steeply in the lowermost Sihapas section, which had a higher rate of deposition relative to subsidence rate.

The trough imaged in Figure 7 is one of several Tertiary half grabens in the Central Sumatra basin that share a similar structural history (Eubank and Makki, 1981; Heidrick and Aulia, 1993). Our interpretation and modeling of the trough imaged in Figure 7 as an extensional fault-bend fold has important implications for the structural and depositional histories of the basin. We conclude that the master normal fault in the trough flattens with depth and soles to near-horizontal detachment. The growth triangles imaged on the seismic profiles are consistent with deposition of the lacustrine and fluvial Pematang Group during formation of the rift. Based on compacted thicknesses and fault slip recorded in growth triangles, Pematang strata filled or slightly underfilled the half-graben accommodation space. Thus, the Pematang deposition rate was generally equal to or slightly less than the subsidence rate, producing shallowly dipping growth axial surfaces (Figure 7). Most significantly, the Brown Shale member of the Pematang formation, which has sourced the more than 7 billion barrels of oil recovered from the basin (Oil & Gas Journal, 1993), corresponds to a very shallowly dipping segment of the growth axial surfaces (Figure 7B). This shallowly dipping growth axial surface records a low rate of deposition relative to subsidence rate that may reflect sediment-starved conditions in a relatively deep lake, which is an environment suitable for the deposition and preservation of organic materials. The upward extension of these growth triangles in the lowermost marine Sihapas Group also suggests that these sediments were locally deposited during the latest stages of rifting. We demonstrate that the angular unconformity on the eastern side of the basin between Sihapas strata and the dipping Pematang section (Figures 5, 6) could have been generated by a dramatic increase in deposition rate relative to subsidence rate without significant uplift and erosion between deposition of fluvial-lacustrine and marine units. This change in deposition rate relative to subsidence rate may represent either a decrease in slip rate during the waning stages of the rift or an increase in the deposition rate of marine vs. older lacustrine and fluvial sediments.

Controls on Three-Dimensional Basin Geometry

Extensional fault-bend fold models demonstrate that the sizes and shapes of the accommodation spaces in half grabens are controlled by normal fault geometries, slip, and axial surface orientations in the hanging-wall block (Xiao and Suppe, 1992). These controls also affect three-dimensional basin geometry; therefore, we apply the analytical techniques used to interpret and model in two-dimensions to explore and map the three-dimensional geometry of the half graben.

The hanging-wall shear orientation can be defined by observing the dip of axial surfaces in seismic profiles. In many cases, this dip corresponds to the Coulomb shear angle of the rocks in extension and is roughly constant in basins composed of the same rock types (Xiao and Suppe, 1992). Given a roughly constant hanging-wall shear orientation, first-order highs and lows within the basin are controlled by normal fault geometry and slip. Thus, where fault geometry is constant along strike, intrabasinal highs and lows are controlled by fault slip. Regions of greater fault slip will have greater subsidence than areas of less slip. Alternatively, where horizontal extension is constant, lateral changes in fault inclination also form intrabasinal structures. Folds above shallowly dipping fault segments will remain high relative to folds along strike that overlie steeper fault segments. We explore these effects of fault slip and geometry in the Central Sumatra trough half graben by using fault-related fold theory to recognize and map horizontal extension and to define fault geometry.

In extensional fault-bend folds that form by simple shear, the offset of the hanging-wall and footwall cutoffs of any pregrowth horizon is a measure of horizontal extension above a normal fault that flattens with depth. This measure is independent of fault dip magnitude (Figure 9). In Figure 11, we map this horizontal extension across the trough imaged in Figure 7 using the hanging-wall and footwall cutoffs on the top of basement. In general, this horizontal extension is roughly constant at about 3.3 km over the mapped area. Thus, we speculate that variations in horizontal extension along strike do not significantly affect the lateral geometry of the trough and its accommodation space. Given this roughly constant horizontal extension, we expect to see a direct correlation between fault geometry and the shape of the half-graben accommodation space.

Figure 11 Map of the basin-forming normal fault imaged on seismic profile YY' (Figure 7) and fault offset, measured between projected hanging-wall and footwall cutoffs of the top pregrowth horizon (inset). The width of the fault offset varies only slightly along its mapped extent and is approximately equal to the magnitude of horizontal extension (see Figure 9). The normal fault shallows from a maximum dip in the west to a near-horizontal detachment below -15,000 ft (-4.6 km) in the east. The dip of the intermediate fault segment below 12,000 ft is steepest just south of line YY' (denoted by closely spaced contours) and decreases in both directions along strike (denoted by more widely spaced contours).

The fault map in Figure 11 was derived from the fault-plane reflections and rollover geometries in the trough imaged in Figure 7 and six other east-west-trending seismic reflection profiles. A simplified version of this fault geometry with three fault segments is shown in the three-dimensional model of Figure 12. The fault plane dips most steeply at shallow depths and flattens through two major (>10°) bends to a near-horizontal detachment. In the center of the trough along line YY', the intermediate fault panel is steep and the normal fault flattens to a horizontal detachment at about -5.3 km (-17,500 ft). Along strike on the northern and southern edges of the trough, the dips of the intermediate fault panel are less and the horizontal detachment lies at only about -4.6 km (-15,000 ft) (Figures 11, 12). As a result, the normal fault has a cuspate or bowl-like shape, with the steepest part of the intermediate fault segment lying just south of line YY'. A fault-bend fold model based on this fault shape with laterally constant horizontal extension (Figure 12) demonstrates that subsidence and accommodation space are greatest along the center of the trend.

Figure 12 Perspective views of a three-dimensional fault-bend fold model of the trough imaged in Figure 7. (A) Cuspate normal fault that flattens with depth to a near-horizontal detachment, which is simplified from the fault map in Figure 11. (B) Cutaway view of hanging-wall rollover panels formed by plane strain with all transport vectors in plane XY. Horizontal extension, which equals slip on the deep near-horizontal fault segment, is constant along strike. The central fault low corresponds to a central basin fold low along section ZZ' (inset).

Therefore, we suggest that fault geometry, and not variable displacement, is responsible for defining the structurally lowest depocenter in the trough, which is imaged on strike lines through the basin (Figure 13). In addition to defining low spots, these measures of fault slip and geometry can define structural high points in the troughs that may serve to focus hydrocarbon accumulations. By identifying highs and lows along strike on normal faults, interpreters can quickly recognize depocenters and regional highs (Tearpock and Bischke, 1991), which can be tested by more detailed reflection contouring. Extensional fault-bend fold interpretations provide a method of defining fault planes using fold shape in cases without continuous fault-plane reflections or fault cuts in wells.

Figure 13 Migrated seismic reflection profile ZZ' along the strike of the trough shown in Figure 11 that images a central low area bounded to the north and south by structural highs. Basin highs and lows are caused by lateral changes in fault geometry, as shown in the model of Figure 12. The omitted portion of the profile includes an area of younger folding associated with faults other than the normal fault mapped in Figure 11. Trace of seismic line ZZ' is shown in Figure 11. Horizontal scale equals vertical scale; datum (0 km) is sea level.

Summary and Conclusions

We demonstrated through our interpretation of seismic profiles and forward models that extensional fault-bend folding has controlled the structural evolution of a half graben in the Central Sumatra basin. In addition, we recognized that syntectonic deposits of fluvial-lacustrine and marine strata form growth triangles and unconformities, which were produced by variations between the rates of local deposition and subsidence. We present a general method for using these patterns of folded growth strata and fault-bend fold theory (Xiao and Suppe, 1992) to define basin structure, including fault geometry. Fault and fold maps based on our interpretations also demonstrated the effects of variations in fault geometry on half-graben subsidence and accommodation space. Fault highs and lows along strike in the trough are overlain by fold highs and lows that form ridges and depocenters. Collectively, fold and fault shapes provided the basis for forward kinematic models of half-graben evolution that were used to test the viability of our geologic interpretation. General consistencies between fold and fault shapes generated in forward models and imaged on seismic reflection data demonstrate that extensional fault-bend folding (Xiao and Suppe, 1992) is a viable theory for the origin of asymmetric half grabens in basement-involved rift systems. Analytical techniques based on this theory can be applied with limited seismic reflection data to generate geometrically and kinematically reasonable interpretations that define intrabasinal structures prior to contour mapping of seismic reflections. In addition, fault-bend fold models provide new interpretations of unconformities and folded patterns of syntectonic section that help to decipher the structural and depositional histories of rift basins.

References Cited

Bally, A. W., ed., 1983, Seismic expression of structural styles: AAPG Studies in Geology 15, v. 2, variously paginated.
Dula, W. F., 1991, Geometric models of listric normal faults and rollover folds: AAPG Bulletin, v. 75, p. 1609-1625.
Eubank, R. T., and A. C. Makki, 1981, Structural geology of the Central Sumatra back-arc basin: Proceedings of the Indonesia Petroleum Association 10th Annual Convention, p. 153-194.
Gibbs, A. D., 1983, Balanced cross-section construction from seismic sections in areas of extensional tectonics: Journal of Structural Geology, v. 5, p. 153-160.
Groshong, R., 1990, Unique determination of normal fault shape from hanging-wall bed geometry in detached half grabens: Eclogae Geologicae Helvetiae, v. 83, p. 455-471.
Hamblin, W. K., 1965, Origin of "reverse drag" on the downthrown side of normal faults: Geological Society of America Bulletin, v. 76, p. 1145-1164.
Heidrick, T. L., and K. Aulia, 1993, A structural and tectonic model of the Coastal Plains Block, Central Sumatra basin, Indonesia: Proceedings of the Indonesia Petroleum Association, 22 Annual Convention, p. 285-317.
Jackson, M. P. A., and W. E. Galloway, 1984, Structural and depositional styles of Gulf Coast Tertiary continental margin: application to hydrocarbon exploration: AAPG Continuing Education Course Notes Series 25, 226 p.
James, D. M. D., ed., 1984, The geology and hydrocarbon resources of Negara Brunei Darussalam: Muzium Brunei, 164 p.
McClay, K. R., and A. D. Scott, 1991, Experimental models of hanging wall deformation in ramp-flat listric extensional fault systems: Tectonophysics, v. 188, p. 85-96.
Nunns, A. G., 1991, Structural restoration of seismic and geologic sections in extensional regimes: AAPG Bulletin, v. 75,
p. 278-297.
Oil & Gas Journal, 1993, Worldwide production report: Oil & Gas Journal, v. 91, p. 63-65.
Rowan, M. G., and R. Kligfield, 1989, Cross section restoration and balancing as an aid to seismic interpretation in extensional terranes: AAPG Bulletin, v. 73, p. 955-966.
Suppe, J., 1983, Geometry and kinematics of fault-bend folding: American Journal of Science, v. 283, p. 684-721.
Tearpock, D., and R. E. Bischke, 1991, Applied subsurface geological mapping: Englewood Cliffs, New Jersey, Prentice-Hall, 648 p.
White, N. J., and G. Yielding, 1991, Calculating normal fault geometries at depth: theory and examples, in A. M. Roberts, G. Yielding, and B. Freeman, eds., The geometry of normal faults: Geological Society Special Publication 56, p. 251-260.
White, N. J., J. A. Jackson, and D. P. McKenzie, 1986, The relationship between the geometry of normal faults and that of the sedimentary layers in their hanging walls: Journal of Structural Geology, v. 8, p. 897-909.
Withjack, M. O., Q. T. Islam, and P. R. La Pointe, 1995, Normal faults and their hanging-wall deformation: an experimental study: AAPG Bulletin, v. 79, p. 1-18.
Xiao, H., and J. Suppe, 1992, Origin of rollover: AAPG Bulletin, v. 76, p. 509-525.

The authors thank Hongbin Xiao and John Suppe for helpful insights into their extensional fault-bend fold theory, which provided the foundation for this work. Exceptional reviews by M. Scott Wilkerson and Walter F. Dula, Jr., improved the manuscript. In addition, discussions with Karsani Aulia, Richard E. Bischke, Peter A. Brennan, Chris D. Connors, Paul W. Genovese, Tom L. Heidrick, and Elizabeth A. Lorenzetti provided insights into our structural interpretations and presentation of the theory. Seismic reflection data were provided by PT CALTEX Pacific Indonesia.