Passive margins underlain by a salt detachment are typically interpreted as kinematically linked zones of updip extension and downdip contraction separated by a zone of translation above a smoothly dipping base of salt. However, salt flow is affected by the base-of-salt geometry across which it flows, and early-stage gravity gliding induced by basin tilt may be complicated by the presence of salt-thickness changes caused by the pre-existing base-salt relief. We investigate these effects using physical models. Dip-parallel steps generate strike-slip fault zones separating domains of differential downslope translation and structural styles, provided the overburden is thin enough. If the overburden is thicker, it resists breakup, but a change in the structural trend occurs across the step. Steps with mild obliquity to the dip direction produce transtensional and transpressional faults in the cover separating structural domains. Deformation complexity in the overburden increases where base-salt steps strike at a high angle to salt flow, and it is especially dependent on the ratio between the thick (T) and thin (t) salt across the step at the base of salt. Where the salt-thickness ratio (T/t) is high, basal drag generates major flux mismatches, resulting in a contractional thickening of the salt and associated overburden shortening in thin salt above a base-salt high block. Shortening is transient and superseded by extension as the salt thickening allows the flow velocity to increase. When transitioning off a base-salt high block into a low block, the greater flux within the thick salt results in a monocline with extensional and contractional hinges. Structures are further deformed as they translate through these hinge zones. Our physical models demonstrate that extensional diapirs and compressional fold belts can be initiated anywhere on a slope as the salt accelerates and decelerates across base-salt relief. A fold belt from the Campos Basin, offshore Brazil, is used to illustrate these processes.

Passive margins over a salt detachment are typically large (100–500 km wide) gravity-driven systems comprising kinematically linked domains of proximal extension and distal shortening commonly described as connected by a zone of translation (Figure 1a; e.g., Fort et al., 2004; Hudec and Jackson, 2004; Brun and Fort, 2011; Quirk et al., 2012; Peel, 2014). A simple gravity-driven, physical model with a smoothly dipping base of salt clearly illustrates this three-domain structural division (Figure 1b and 1c). Previous physical modeling studies have called into question this simple three-domain structural zonation. For example, Brun and Fort (2011) illustrate the migration and overprinting of extensional and contractional provinces during the evolution of margin-scale physical models. McClay et al. (1998) also document the overprinting of contractional provinces by a seaward-migrating upper-slope and shelf-extensional zone. Dooley et al. (2013) show how coastal uplift in the Gulf of Mexico led to the development of a pillow-fold belt far inboard from the typical toe-thrust province. All these processes may complicate the typical three-domain structural division.

Salt is also sensitive to the geometry of the surface it flows across. Early-stage gravity gliding induced by basin tilt on salt-bearing passive margins may be complicated by base-salt topography due to folding and faulting, or due to erosional rugosity. For example, Hudec et al. (2013) propose that in the Gulf of Mexico, the Jurassic Louann Salt was deposited in depressions above hyperextended continental and transitional crust. After deposition of the salt, the rifting is interpreted to have continued for 7 to 12 million years before initiation of seafloor spreading and basinward flow of salt. Hudec et al. (2013) infer that salt was deposited atop a rugose or stepped topography over which it would later flow, and/or that it would encounter rugose subsalt topography downdip during the early stages of salt flow. In these cases, deformation observed in sediments overlying salt is inferred to be related to salt flow over base-salt relief, rather than being the result of an outside force (e.g., the interior uplift of Dooley et al., 2013) or a gradual migration of structural domains during margin evolution (e.g., Fort et al., 2004). Local perturbations are also likely to be transitory as the salt and its thin overburden move downdip across base-salt rugosity.

In this paper and our companion study (Dooley and Hudec, 2017), we use relatively simple physical models to investigate the effects of base-salt relief on salt flow and suprasalt deformation patterns. Examples of previous modeling studies incorporating base-salt relief include the convergent and divergent radial-gliding models of Cobbold and Szatmari (1991), the study of Gaullier et al. (1993) on the effects of residual topography below a salt décollement on fault orientation in the overburden, and the basin-scale base-salt relief models of Adam and Krézsek (2012).

Our models are, for the most part, very simple and primarily driven by gravity. We begin by investigating base-salt relief that is parallel to, or at low angles to, the mean salt-flow direction. The model complexity is then increased, with the base-salt relief oriented at high angles to the salt-flow direction. Our goal is to address the validity and applicability of the simple three-domain structural zonation for passive margins over salt detachments (Figure 1). Is this structural zonation a result of a smooth base of salt? Can fold belts form in places other than the toe of slope or salt pinchout? Can base-salt relief promote extension where our simple model would predict zones of translation? What happens when the overburden thickens and strengthens as these margins evolve? And finally, can we apply these models to real-world passive margins over salt detachments to help understand their evolution and kinematics?

A series of simple experiments investigated the evolution of surface structures, block velocities, and overburden breakup with differing subsalt topographies (Table 1). Deformation was induced by tilting the model baseboard to initiate gravity gliding. Computer-controlled time-lapse digital cameras captured the obliquely lit surface of the models. This 2D archival imagery was processed using digital image correlation (DIC) software to quantify downdip displacements, strain, and other parameters (for further details of DIC analysis, see Adam et al., 2005).

As is common in physical model studies of salt tectonics, our salt analog was a ductile silicone and its siliciclastic overburden comprised brittle, dry, granular materials. The silicone used in these experiments was a near-Newtonian viscous polydimethylsiloxane. This long-chain polymer has a density of 950 to 980  kgm3 and a dynamic shear viscosity of 2.5×104  Pas at a strain rate of 3×101  s1 (Weijermars, 1986; Weijermars et al., 1993). The thin overburdens were composed of a mixture of silica sand (bulk density of 1700  kgm3; grain size of 300  μm; internal coefficient of friction, μ=0.550.65; see material properties in McClay [1990], Krantz [1991], and Schellart [2000]) and hollow ceramic microspheres (bulk density of 600  kgm3, grain size of 90  μm, and typical μ=0.45; for details, see Rossi and Storti, 2003).

To compare models directly with early-stage rafting processes on salt-bearing passive margins, we minimized the effects of gravity spreading to focus on gliding. To do this, we chose the simplest solution of mixing the silica sand and ceramic beads such that their density was equal to that of the salt. Initial overburden thicknesses in the models varied from 5 to 20 mm.

The nine experiments documented in this study were modeled in deformation rigs having horizontal dimensions of 130×60  cm (Figure 2a) or 130×100  cm (Figures 2b, 2c, and 3). Subsalt topography was built using silica sands to create the basement steps (angle of repose of the granular mixtures), across which the salt thickness changed (Figures 2 and 3; Table 1). Silicone polymer was deposited across basement topography and allowed to settle for several days. Once the silicone had settled, the overburden was emplaced. In all experiments, deformation was initiated by tilting the rig to 3° and removing the confining downdip endwall to create an open-toe system, similar to the models of Gaullier et al. (1993).

Dip-parallel step at base of salt — Effect of overburden strength on deformation patterns, models 1–3

Model overview

The simplest scenario we consider is a dip-parallel subsalt step (Table 1). After 48 h, surface deformations in models 1 and 2 are summarized by overhead views (Figures 4a, 5a, and 6), overhead views after removal of suprasalt sediments (Figures 4b and 5b), and DIC analyses illustrating downdip movements (Figures 4c and 5c). As illustrated in Figure 2a, the experimental setup was identical for both models except for the thickness of the prekinematic overburden. The similarities between the two models are (1) a greater extension above the zone of thick salt, resulting in larger graben and reactive diapirs (Figures 4a, 4b, 5a, and 5b), (2) a major change in structural trend across the subsalt step from curvilinear graben above thick salt to linear oblique fault zones above thin salt (Figures 4 and 5), and (3) greater downdip translation above thick salt due to its greater mobility (Figures 4c and 5c). The primary differences between the two models are (1) the presence of a discrete strike-slip fault separating the domains of thin and thick salt in the thin-overburden model in model 1 (Figure 4a and 4b), whereas in model 2, the basement step is only manifested at the surface as a major change in structural trend (Figure 5), and (2) the greater (almost double) downdip displacements measured above the thick salt in model 1 (Figure 4c), whereas major differences in downdip translations in the thick overburden of model 2 are confined to the open-toe region (Figure 5c).

Effect of overburden thickness on deformation style

Both models illustrate a dramatic change in downdip displacements across the basement step due to the salt-thickness change and the corresponding increased flow across this structure. In the absence of an overburden, there would be a flow boundary above the subsalt step that divides the salt basin into discrete flow domains. When the overburden is thin, and thus weak, salt flow dominates deformation, causing the formation of a dip-parallel strike-slip fault in the overburden that separates the two domains (Figure 4). On removal of the overburden, this strike-slip fault is composed of a discontinuous, dip-parallel salt wall (Figure 4b). In contrast, when the overburden is thick and strong, the effects of the flow boundary are muted, and greater translations above thick salt are accommodated by rotation at the open toe (Figure 5). Removing the overburden reveals the absence of any dip-parallel topography at the top salt (Figure 5b). Model 3 clearly illustrates a change in the surface deformation after the addition of a synkinematic sand layer (Figure 6). After 48 h deformation, the surface-deformation patterns, consisting of a discrete strike-slip fault system in the mid- to lower slope separating the two salt domains (Figure 6a), were very similar to those in model 1. After the addition of the synkinematic layer onto model 3, which strengthened the overburden, the boundary between the two salt domains was no longer marked by a discrete fault zone but by a radical change in fault strike across the basement step (Figure 6b), similar to that seen in model 2 (Figure 5).

Fault-pattern variability

Fault patterns display similarities in models 1–3. Continuous obliquely oriented faults with lateral displacements characterize deformation above the thin-salt regions (Figures 46). These faults are oblique and roughly parallel in the mid to lower slope but become more orthogonal to dip near the breakaway zone (Figures 46). In the thin-salt domain, the sense of relative displacements above the subsalt step and the lateral margin of the basin are the same (Figure 7). Thus, Riedel shear arrays are synthetic on both sides and link up as linear faults across this zone of distributed shear (Figure 7), which is consistent with previous physical models of distributed shear (e.g., Figures 16 and 18 in Dooley and Schreurs, 2012). In contrast, arcuate graben dominate above the thick-salt regions, which is best seen in model 1 (Figure 4a). Above the thick salt, the sense of displacement above the subsalt step and salt margin are opposite, producing Riedel shear arrays with mirrored orientations (Figure 7). These curved geometries are not due to simple “edge effect” artifacts in the models; it should be noted that faults at the updip breakaway zone extend almost from edge to edge of the salt basin with minimal curvature. These geometries, forming because of shear stresses associated with fixed or slower-moving margins, are similar to those documented by Le Calvez and Vendeville (2004).

Enhanced downdip salt flow occurs in the center of the thick salt to form faults at a high angle to the dip direction; these faults link with lower angle segments at the margins of the thick-salt block to form arcuate graben that break the overburden into rafts separated by reactive diapirs (Figure 7). Although the effect of the subsalt step is more muted in model 2 because of the thicker prekinematic overburden, a major change in structural orientation occurs across the step (Figure 5). Riedel shears propagated out from the northern margin of the salt basin across the thin-salt region, with only minor changes in curvature (Figure 5a); they also propagated out from the southern margin, but the more-rapid extension generated above faster-moving, thicker salt (see the DIC analysis in Figure 4c) generated faults at a high angle to the dip direction. These fault systems met with a major angular change across the subsalt step (Figure 5).

Steps with mild obliquity to salt-flow direction — Boundary-fault migration, models 4 and 5

Model overviews

Setup of models 4 and 5 was analogous to models run with dip-parallel steps at the base of salt, except that the steps were mildly oblique, 10°, to the mean salt-flow direction (Figures 2b, 2c, and 8; Table 1). The senses of boundary shear were the same, and the base-salt step separated zones of different flow velocities (Figure 8). The two main differences between models 4 and 5 and the previous dip-parallel step models are as follows:

  1. 1)

    Translation vectors are oblique to the base-salt step trend, and continued translation moves the suprasalt faults off the step.

  2. 2)

    Different flow velocities on either side of the base-salt step introduce a component of transtension or transpression along the boundary fault zones.

These differences are clearly seen in the DIC imagery in Figure 8. Once more, larger downdip translation and associated extension characterizes deformation above thick salt, whereas distributed shear dominates above thin salt.

Transtensional step summary

Surface deformations in model 4 after 48 h are summarized by an overhead view (Figure 9a), an overhead view after removal of suprasalt sediments (Figure 9b), and a DIC analysis illustrating downdip movements (Figure 10). Evolution of the central part of the model is illustrated by a series of overhead views in Figure 11. Overburden deformation was broadly similar to that seen in model 1, with relatively long and continuous oblique faults cutting the thin-salt block in the mid- to lower slope, and arcuate graben that dissected the thick-salt block into a series of raft blocks and reactive diapirs (Figure 9). The pattern of faulting above the thin salt in the mid- to lower slope — again, indicative of distributed shear across this region — consisted of several synthetic (sinistral) shears and high-angle antithetic (dextral) shears in the mid to lower slope. Toward the updip breakaway, the fault pattern was dominated by structures oriented at a high angle to the regional dip (Figure 9). The boundary between the thin- and thick-salt domains was marked by a strike-slip fault system (Figure 9). DIC analysis illustrates a major disparity in downdip displacements between the thin- and thick-salt domains, and it clearly shows that the boundary between these domains now lies in the thick-salt domain rather than above the subsalt step (Figure 10).

Migration of the thin-thick salt boundary off the crest of the basement step and into the thick-salt domain is clearly illustrated in Figure 11. In the early stages of the model runtime, a series of right-stepping Riedel shears characteristic of sinistral strike-slip deformation (relative in this case) formed above the basement step (Figure 11a). A Riedel shear is seen propagating out from this fault zone close to the updip marker grid and into the thick-salt domain. With increasing downdip displacement, a throughgoing fault zone formed and began to pull away from the basement step (Figure 11b). Riedel shears with sinistral displacements were seen propagating out into the thin- and thick-salt domains. In the thick domains, these Riedel shears quickly rotate to a high angle to the dip direction because of deformation being dominantly extensional within the center of the thick-salt domain (Figure 11b). After 40 h of runtime, the strike-slip zone separating the two structural domains had broadened and migrated well into the thick-salt block, creating a “rootless” fault system, i.e., not overlying the base-salt step.

Transpressional step summary

Surface deformation patterns in model 5 after 48 h are summarized by an overhead view (Figure 12a), an overhead view after removal of suprasalt sediments (Figure 12b), and a DIC analysis illustrating downdip movements (Figure 13). Detailed evolutionary views of the central part of the model are shown in Figure 14. As in the transtensional step, overburden deformation — consisting of long, oblique faults cutting the thin-salt domain in the mid to lower slope and arcuate graben systems dissecting the thick-salt domain (Figure 12a and 12b) — is similar to that observed in model 1. Distributed shear — consisting of synthetic sinistral shears with some degree of divergence across them, as well as a few higher angle antithetic (dextral) shears — once again dominated above the thin-salt block. The boundary between the thin- and thick-salt domains consisted of two strike-slip fault zones, one of which is now inactive because it was translated up onto the high block and away from the base-salt step that caused its formation (Figures 12 and 13). DIC analysis again shows a major disparity in downdip displacements between the thin- and thick-salt domains; it also shows that this boundary now lies north of the subsalt step (Figure 13). Note that bulk downdip displacements were smaller in this model than in the transtensional model (compare Figures 10 and 13). Also note that the extension above the thick salt decreased downslope because of constriction by the transpressional step.

Evolutionary views of the central part of model 5 clearly show initial formation and subsequent migration of the strike-slip fault zone up onto the subsalt step (Figure 14). In the early stages of model evolution, the boundary fault zone formed against the base of the subsalt step and consisted of a series of left-stepping wrench folds that were, in some cases, linked by right-stepping sinistral Riedel shears (Figure 14a). With increasing runtime, the strike-slip zone separating the thin- and thick-salt domains gradually moved up onto the crest of the subsalt step, amplifying the left-stepping wrench folds (Figure 14b). Synthetic strike-slip segments propagated from this boundary fault zone into the thin- and thick-salt domains. In the thick-salt domain, these faults rotated and linked with higher angle graben that accommodated the major extensional strains above the thicker and faster moving salt (Figures 12, 13, and 14b). By 40 h of runtime, the initial transpressional fault zone had moved north of the crest of the subsalt step, forming a “rootless” fault system (Figure 14c). A new strike-slip segment formed above the thicker salt, adjacent to the subsalt step, marking the transition between the more rapidly moving thick-salt domain and the slower-moving thin-salt domain (Figures 13 and 14c).

Base-salt relief at a high angle to salt-flow direction — Complex salt-flow dynamics and superposed deformations, models 6–9

In models 4 and 5, the low angle of the subsalt steps to the dominant flow direction induced salt-flow domain boundaries that were relatively simple. This section investigates the greater complexities in models with a subsalt step that is oriented at a high angle (70°) to the mean salt-flow direction, combined with a more-detailed DIC analysis to document the complex strain histories of these models. We first look at the evolution of model 6, which exhibited extreme variations in deformation in time and space, before moving on to the effects of variable overburden thickness (Table 1).

Effects of a base-salt high block oriented at a high angle to mean salt-flow direction

Overhead views and selected DIC strain analyses illustrate the evolution of model 6 (Figures 15 and 16). After 10 h of gravity-driven deformation, the model consisted of five distinct zones from the updip breakaway to the downdip open toe (Figures 15a and 16a): (1) an updip breakaway zone with extensional graben orthogonal to model dip, (2) a zone of shortening and uplift, with associated outer-arc extension on the crest of this uplift that paralleled the trend of the subsalt high block, (3) a distinctive topographic low or monocline above the downdip edge of the subsalt high block, (4) a zone of extension characterized by graben that were parallel to the strike of the subsalt high block, and (5) a zone of open-toe extension with well-developed graben orthogonal to the model dip direction.

During the later stages of the model runtime, a reversal in strain patterns was observed above the updip margin of the subsalt high (zone ii in Figures 15a, 15b, 16a, and 16b). The area of initial uplift and associated shortening strains was subsequently overprinted by extension (Figure 16b). Over time, a series of extensional graben with strikes that paralleled the trend of the subsalt high block formed and were translated downslope (Figure 15b and 15c). The first formed of these graben (G1 in Figure 15b15d) was subsequently squeezed and ejected a salt sheet onto the model surface as it moved off the subsalt high block down into the thicker salt region. At the updip, breakaway major extension resulted in significant separation of rafts. Downdip of the subsalt high block, the early formed graben were translated downslope and separated from the mid-slope extensional zone by a major raft block (Figure 15b15d).

The strain patterns and associated structures highlight intriguing processes at the updip and downdip edges of the subsalt high block (Figure 16). During the early stages of the model run, an anticline developed above the updip edge of the subsalt high block, and it was expressed as a topographic high — with surface strains indicating outer-arc extension — and was flanked by zones of shortening (Figure 16a). With increasing runtime, this uplift began to extend, forming a series of downdip-younging graben that nucleated above the updip edge of the subsalt high block and were gradually translated downslope (Figures 15b15d and 16b). The zone of shortening strains on the downdip side of original uplift was translated downslope ahead of the first graben (G1 in Figure 16b). Thus, on this updip edge of the subsalt high block stood an early stage of uplift, a salt-cored anticline, followed by an extensional collapse of the uplift. Above the downdip edge of the subsalt high block, the strain pattern remained consistent throughout the model runtime (Figure 16). This margin was defined by a topographic low that paralleled the subsalt high block. Linear zones of extension and contraction that flanked this structure defined the updip and downdip hinges, respectively, of this monocline (Figure 16). Graben that formed above the updip edge of the subsalt high block were translated downslope, extended farther across the extensional hinge, and then squeezed and partially inverted as they passed through the contractional hinge of this system (Figure 15d).

The complex overburden strain history associated with this relatively simple model prompts the question: What factors contribute to the structural variation in time and space of overburden translated over a step in base-salt topography oriented at a high angle to the flow direction? The following section attempts to answer this.

Streamlines, basal drag, and salt-flux mismatches

When dealing with a viscous fluid such as salt, it is useful to consider streamlines, flux, and the effects of basal drag, itself dependent on the salt-thickness ratio T/t (Figure 17), on flow velocities within the system. Consider the case in which salt flowing downdip encounters a subsalt high block (Figure 17a and 17b). In general, a decrease in salt thickness means that streamlines converge above the base-salt high block and the salt flow accelerates (Bernoulli’s principle). This is the case in which basal drag is not a factor, i.e., in which the ratio between the normal salt thickness and thickness above a base-salt high block (T/t) is small. Acceleration compensates for the lesser salt thickness above the base-salt high block, and there is no flux mismatch (Figure 17a). However, if the thickness ratio (T/t) is large, then drag is important and there is a resultant flux mismatch (Figure 17b). In this scenario, more salt is being fed onto the base-salt high block than can be accommodated by the flow within and through the thinner salt body, resulting in early-stage salt thickening and contraction above the high block, as we saw in model 6 (Figures 15 and 16).

Now consider the case in which the salt flows off this high block and encounters thicker salt (a base-salt low block; Figure 17c and 17d). Once more, if the salt-thickness ratio (T/t) is low, then the effects of basal drag are minimal. In this scenario, streamlines diverge into the low block and any flux mismatch is accommodated by deceleration and salt thickening (Figure 17c). But if the thickness ratio (T/t; Figure 17d) is high, then basal drag comes into play, resulting in a major flux mismatch between the high and low blocks. The zone of thicker salt downdip of the base-salt high block initially pulls away from the high block under extension. The flux mismatch results in the creation of (1) an extensional monocline with an updip extensional hinge as material is pulled rapidly across the subsalt step and (2) a contractional hinge as material then encounters the thicker and slower moving salt downdip of the subsalt high block (Figure 17d). This result explains the inverted graben and expelled salt sheets seen in model 6 (Figures 15 and 16).

The evolution of overburden deformation in a step-up/step-down system is illustrated in Figure 18a. In the early stages, a thickened body of salt builds up above the updip margin of the high block over time because of the flux mismatch generated by a high salt-thickness ratio and associated basal drag. At the downdip edge of the high block, the thicker salt body pulls away from the step to form an oblique graben system (Figures 15a, 15b, and 18a). Contractional thickening of the salt above the high block allowed its velocity to gradually increase over time as the effects of basal drag became proportionally less important, eventually resulting in extension above this thickened body of salt and the formation of a new breakaway zone and associated graben in the brittle overburden (Figures 15b15d and 18a). As these graben are translated across and off the subsalt high, they undergo rapid extension while passing through the extensional hinge and inversion while passing through the contractional hinge (Figures 15c, 15d, and 18a). This temporal reversal in strain patterns above a subsalt high block is also confirmed by finite-element modeling (FEM; Figure 19). The model reveals an early stage of updip shortening against the subsalt step and downdip extension as the thick salt pulls away from the subsalt high block (Figure 19b), similar to the patterns revealed in model 6. With time, this strain field is reversed and an extensional and contractional hinge system forms at the downdip margin of the subsalt high (Figure 19c).

Effects of variable-overburden thickness above a base-salt high block oriented at a high angle to mean salt-flow direction

Model 7 was run with base-salt relief identical to that in model 6, but it incorporated a laterally variable prekinematic overburden thickness. Overburden thickness increased from ca. 1 mm in the south to 15 mm in the north (Table 1; Figure 20a). Overhead views and 2D DIC analyses document the early- and late-stage evolution and strain patterns of this model (Figures 20 and 21). After 5 h of gravity-driven deformation, the model consisted of five distinct zones from the updip breakaway to the downdip open toe, identical to those observed in model 6 (Figure 20a and 20b). The effect of a thinner, and weaker, overburden in the southern half of the model is immediately obvious, with more focused shortening and the formation of a distinct fold-thrust belt above the updip edge of the subsalt high block (Figure 20a). This fold belt transitions along strike into a simpler anticlinal uplift and weaker zone of shortening as the overburden thickens and strengthens. The opposite is true in extensional zones, where extensional strains become more diffuse as the overburden thins (Figure 20). With increased model runtime, the updip breakaway zone broadened, the fold-thrust belt above the subsalt high amplified and was translated downdip, and a broad extensional zone characterized the downdip domain (Figure 20c and 20d). Note that a zone of mild contraction was still present above the updip margin of the subsalt high (Figure 20d). As seen in model 6, the strain pattern above the downdip margin of the subsalt high defined an extensional and contractional hinge marking the boundary between the thin and thick salt (Figure 20d).

During the later stages of model 7, the same reversal in strain patterns seen in model 6 was observed above the updip margin of the base-salt high block (Figure 21). Extensional strains begin to appear as narrow and diffuse graben in the southern part of the model where the overburden was thin, narrowing along strike to the north as the overburden thickened (Figure 20a and 20b). The updip domain was characterized by major extensional rafting, whereas downdip, the translated fold belt began to amplify as it passed off the high block and encountered the contractional hinge (Figure 21). This fold belt was separated from the downdip extensional domain by a low-strain zone (Figure 21d). In the later stages, major extension dominated deformation above the subsalt high block (Figure 21c and 21d). Closely spaced faults characterize the diffuse extension where the overburden was thin, whereas focused extension formed discrete graben where the overburden was thicker (Figure 21c and 21d). Shortening strains increased within the translated fold belt as it left the downdip edge of this high. Weak zones in the overburden such as graben G1, G2, and G3 were clearly inverted as they passed through the downdip contractional hinge (Figures 21c, 21d, and 22).

Effects of a base-salt low block oriented at a high angle to mean salt-flow direction

The opposite case, a base-salt low block, is considered in Figure 18b and by model 8 (Table 1; Figure 23), in which flowing salt encounters a base-salt low block (step down) and then a downdip high block (step up). In this case, the deformation patterns are reversed from those observed in models with a base-salt high block. At the updip edge of the low block, a monocline begins to develop as the greater flux from the thick-salt region pulls away from the thin-salt region (Figures 18b, 23a, and 23b). As the salt steps up further downdip, the greater flux from the thick-salt region results in contractionally thickened salt because of the flux mismatch (Figure 18) and an associated fold belt (Figure 23a and 23b). Over time, graben that formed updip of the low block were translated through a contractional hinge and inverted within the low block (Figures 18b, 23c, and 23d). As the salt velocity gradually increased above the downdip step because of thickening, the plateau of thickened salt collapsed under extension (Figures 18b, 23c, and 23d).

The suite of models presented in this part of our study is just a small subset of the kinds and shapes of base-salt relief that may be present on a passive margin during early-stage tilting and gravity-induced deformation. Here, we attempt to answer the questions posed in the “Introduction” section.

What are the effects of dip-parallel and mildly oblique base-salt relief on deformation in the overburden?

Dip-parallel subsalt steps were the simplest case studied using our physical models. In thin-overburden models, a discrete strike-slip fault formed in the overburden that separated the faster translating thick-salt domain from the slower moving thin-salt domain. The structures formed above each of these domains are distinct, with deformation above thin salt consisting of fault patterns consistent with distributed shear in the mid to lower slope region (see Figure 6) due to synthetic sense of shear above the subsalt step and the lateral margin of the salt basin. Minor separation of overburden blocks occurs across the oblique structures in this domain. Above the thick salt, the structures are arcuate, curving from oblique-slip/strike-slip faults with opposite senses of slip from the margins of the thick-salt domain to extensional structures near-orthogonal to the dip direction in the center of the domain (Figure 6). Major separation of the prekinematic overburden into raft blocks occurs in the faster moving thick-salt domain (e.g., Figures 4 and 6).

Base-salt steps with mild obliquity to the mean salt-flow direction produce salt-detached transtensional or transpressional strike-slip faults, where the average displacement vector diverges or converges across the relief. These fault zones form the boundary between thin and thick salt (e.g., Figure 8). Divergent or convergent salt flow serves to move the boundary faults away from the subsalt structure that caused their formation, making them rootless (Figures 11 and 14). Arcuate graben with significant separation again characterizes deformation above the thick-salt domain, whereas deformation is characterized by distributed shear above the thinner salt between the base-salt step and the margin of the salt basin. Suprasalt transtensional features in the Campos Basin are documented by Fetter (2009) across the northeast–southwest basement lineaments. Fetter (2009) suggests basement-involved deformation to explain these features, but they may reflect a process similar to those seen in our models, in which the mean salt flow is oblique to base-salt relief separating zones of different salt thicknesses and thus mobility.

What is the impact of base-salt relief at a high angle to the salt-flow direction?

Base-salt relief at a high angle to the salt-flow direction has a major impact on the salt flow and overburden structure (especially at the early stages of margin tilt when the overburden is thin), in which there is a significant flux mismatch (e.g., Figures 15, 16, 20, and 21). Salt flowing up onto a high block is impacted by basal drag, resulting in compression due to the mismatch between the flux that can be accommodated through the thinner salt and that is being fed onto the base-salt high block from the updip region of thicker salt. This process initially results in a zone of contractional uplift of thickened salt. Eventually, the gradual thickening of salt above the high block allows the salt velocity to increase, and the effects of basal drag become proportionally less, resulting in extensional collapse and the formation of graben that migrate across the base-salt high block. At the downdip edge of a base-salt high block, the salt flux flowing off a high block is less than that pulling away from the downdip thicker salt body, resulting in the development of a topographic monocline with an updip extensional hinge and downdip contractional hinge within the thicker salt. Extensional graben translated across and off the base-salt high block are initially rapidly extended through the extensional hinge and then squeezed and inverted as they pass through the contractional hinge.

The effects of a base-salt low block are equally as dramatic, but the positions of the thickened salt body and topographic monocline are reversed with respect to one another in the dip direction (Figures 18 and 23). It is likely that downdip translation across multiple base-salt irregularities means that overburden structures and associated diapirs could undergo multiple stages of extension and compression and may act as potential nucleation sites for further diapiric activity after translation out into deep water, far from the basement structure where they initially formed.

An example of a fold-thrust belt from the Campos Basin is shown in Figure 24. This fold belt, involving originally thin Albian strata, is located on the midslope, some 40–50 km from the toe-of-salt shortening zone. Structural analysis reveals that this fold belt restores back to the base-salt high block seen on the northwest side of the seismic section. The elevation change across this high block is significant, some 1.4 km, and we believe that the fold belt was generated by the presence of this relief, similar to that seen in our physical models, when the suprasalt overburden was thin (Figure 24). This seismic example compares very favorably to our model 9 (Figure 25), which shows a buried fold belt on the seaward side of the model, far downdip from the base-salt relief that caused its formation. In model 9, this fold belt initiated on the updip side of the base-salt high block but was heavily amplified and broadened as it moved off this high block and through the contractional hinge zone of the downdip monocline (Figure 25a and 25b). Strengthening of the overburden in the model and in the seismic example dampened the effect of this contractional hinge such that later deformation was dominantly extensional in the form of a ramp basin, although the squeezing of salt diapirs could absorb some cryptic shortening.

How does overburden thickness influence suprasalt deformation patterns?

The mechanical behavior of salt and sediments is profoundly different. Salt tends to form discrete flow cells, whereas even less compacted to moderately compacted sediments deform as rigid blocks. Our simplest models with dip-parallel subsalt steps (models 1–3) illustrate that where the overburden is thick and strong, no strike-slip fault forms in the overburden; instead, the base-salt relief is marked by a sharp change in fault orientation (Figure 5). Where the overburden is thin and weak, the structures above thin and thick salt are distinct and separated by a discrete strike-slip fault (Figure 4). Synkinematic sedimentation may alter this relationship with the thin-style structure superseded by a thick-style structure because of sediment loading (Figure 6). Suprasalt-deformation style is strongly controlled by the overburden thickness in our models, with base-salt steps at high angles to the transport direction (e.g., Figures 20 and 21). The transient zones of contraction that formed as salt flowed up onto a high block illustrate well-defined fold-thrust belts where the overburden is thin, grading laterally to a zone of diffuse shortening with minor arching where the overburden is thick and strong (model 7; Figure 20). Updip extensional breakaways are dominated by discrete, spaced graben where the overburden is initially thick, gradually changing to a zone of multiple graben and more-diffuse extension where the sedimentary overburden is thinnest (Figure 20). Gradual strengthening of suprasalt overburdens by synkinematic strata further dampens the contractional effects of base-salt relief, with cryptic shortening absorbed by minor squeezing of preexisting diapirs (Figure 25).

Our physical models demonstrate that base-salt relief can have a significant localized impact on salt flow and suprasalt deformation patterns. Dip-parallel base-salt relief with low obliquity to the mean salt-flow direction separates zones of differential downslope displacement along strike- or oblique-slip fault zones where the overburden is thin, and the zones are marked by distinct changes in structural grain where the overburden is thicker and stronger. Models run with base-salt relief at a high angle to mean that salt-flow direction have the greatest impact on deformation patterns. Unlike our simple extension-translation-contraction model in Figure 1, these models show that extensional diapirs and shortening-related fold-thrust belts can form anywhere on a slope as the salt accelerates and decelerates when flowing across the base-salt relief. They also demonstrate temporal changes in the local stress field that can cause numerous episodes of extension or contraction. These processes may have implications for the initiation of diapirs in the deepwater portions of salt basins. Continued seaward flow transports the diapirs and associated suprasalt deformation farther downslope, further complicating the system and making it difficult to identify the basement structure that caused their formation.

T. P. Dooley thanks J. Donnelly, N. Ivicic, J. Lambert, and B. Williamson for logistical support in the modeling laboratories. We are extremely grateful to V. Mount, C. Schneider, K. McClay, T. Hearon, and O. Ferrer for their constructive reviews of an earlier version of this manuscript. A special thank you goes to L. Moscardelli for reviewing the initial submission and to S. Jones for editing the manuscript. The project was funded by the Applied Geodynamics Laboratory (AGL) Industrial Associates program, comprising the following companies: Anadarko, Apache, Aramco Services, BHP Billiton, BP, CGG, Chevron, Cobalt, Condor, ConocoPhillips, EcoPetrol, ENI, ExxonMobil, Freeport-McMoRan, Fugro, Hess, Ion-GXT, LUKOIL Overseas Services, Maersk, Marathon, Murphy, Nexen USA, Noble, Pemex, Petrobras, PGS, Repsol, Rockfield, Samson, Shell, Spectrum, Statoil, Stone Energy, TGS, Total, Venari Resources, and Woodside. The authors received additional support from the Jackson School of Geosciences and the University of Texas at Austin. Publication was authorized by the director, Bureau of Economic Geology, the University of Texas at Austin.

Biographies and photographs of the authors are not available.

Freely available online through the SEG open-access option.