Abstract

The Barreirinhas Basin is an ideal location to study shale-dominated gravity-driven thrusting systems because of the limited areal extent of the deformed areas compared to other areas in the world. Regional seismic reflection profiles across the Barreirinhas Basin on the Brazilian Equatorial margin show two major deepwater fold and thrust belts linked landward to extensional fault systems. Thrust faults are interpreted to be products of shortening caused by gravity-driven extension on the continental margin that involve rocks of both the shelf and the slope. Results show two main deformation events during the Cretaceous (between ca. 89 and 84 Ma) and several episodes during the Later Cenozoic (ca. 55–0 Ma). All events were characterized by displacement along a detachment fault linking a landward system of normal faults to a basinward system of folds and thrust faults. The Cretaceous deformation involved a <1.5-km-thick sequence deformed in a 30-km-wide set of listric normal faults (an extensional domain) on the outer continental shelf and upper slope. Those normal listric faults merge seaward into a bed-parallel detachment surface forming a 30-km-wide translational domain and a 30-km-wide zone of imbricate thrust faults (a compressional domain) at the toe of the slope. The Cenozoic structural system involves a thick (over 4 km) sedimentary sequence of Turonian to Miocene age, which cross-cuts the preexisting Cretaceous deformed sequence. The Cretaceous and Cenozoic deformational events formed two discrete bowl-shaped fault systems that are linked at different stratigraphic levels. Plots of displacement versus time show normal- and thrust-fault movements during the same time intervals, confirming the link between extension on the continental shelf and compression on the slope. Deformation increased dramatically during the past 10 m.y., with movement on all earlier and some newly formed faults. The increased deformation coincided with tectonic paleogeographic and topographic changes in northern South America during the Late Miocene that led to an increase in sediment supply to the Barreirinhas Basin.

INTRODUCTION

Gravity gliding covers a wide range of temporal and spatial scales, from small (up to several square kilometers in area) surficial geologically instantaneous slumps or slides, to long-lived giant submarine fold-thrust belts that can cover thousands of square kilometers (Morley et al., 2011). Because deformation rates can control the generation of mass flows (i.e., turbidites) and displace large masses of water, understanding the mechanisms and kinematics of deformation in these systems can be critical to several fields, notably reservoir presence for hydrocarbon exploration and local sequence stratigraphy (Gilbert et al., 2011).

Small-scale underwater slides are commonly triggered by catastrophic events such as storms, earthquakes, and high rainfall. They contrast with areally extensive gravity gliding in deepwater fold-thrust belts on passive margins that result from long-term sediment loading (Morley et al., 2011). The latter give rise to large-scale compressional provinces or deepwater fold-thrust belts at the toe of the continental slope, linked to shelf extensional provinces by a broad, relatively undeformed zone of lateral transport.

Large-scale gravity gliding systems developed in sediment piles, such as the Niger Delta, can be described in terms of the critical taper angle thrust model (Davis et al., 1983; Dahlen, 1984), where the combination of a thick sediment pile, steep basinward surface slope, and a gentle landward basal slope commonly generates shear stress to promote basinward sliding up the basal slope. Loss of shear stress at the toe of the slope generates compression and results in a deepwater fold-thrust belt at the toe of the slope. Such detachments usually occur in evaporites or thick shales (Rowan et al., 2004), and shale detachments are normally associated with documented fluid overpressure that also reduces the frictional resistance at the base of the sediment pile (Mourgues et al., 2009). Examples include the Amazon fan (Cobbold et al., 2004), Niger delta (Corredor et al., 2005; Billotti and Shaw, 2005; Cobbold et al., 2009), and the Mexican ridges (Weimer and Buffler, 1992). In order to evaluate the mechanisms that drive large-scale gravity glide systems, it is critical to know the geometry and sequence of faulting and their relationship to sedimentation.

In this paper, we present seismic interpretation across a poorly documented region of the Brazilian equatorial margin in order to: (1) assess the geometry and timing of faulting; (2) document the structural relationship between thrust faults at the toe of the slope and normal faults at the shelf edge; and (3) evaluate their relationship to the sedimentation history.

REGIONAL TECTONICS AND BASIN STRATIGRAPHY

The Barreirinhas Basin is one of a set of basins on the Equatorial Brazilian margin, including the Amazon Cone (the deepwater part of the delta), and lies west of the terminus of the Romanche fracture zone (Fig. 1). The Barreirinhas Basin is separated on its southeast margin from the Ceará Basin by the Tutóia High, but the basin’s boundary to the north with the Pará-Maranhão Basin is poorly defined. Here we treat the Barreirinhas and Pará-Maranhão basins as a single basin.

The Brazilian Equatorial margin transitions from a transform continental margin in the Piauí-Ceará basins to an oblique-rifted margin in the Pará-Maranhão basins (Fig. 2). There is a zone of thinned continental transitional crust as narrow as 12 km underlying the continental slope, and a steep seafloor slope where sediment prograde abruptly into deepwater. This forms an instable margin ideal for generating downslope mass transport by both sedimentary and structural mechanisms.

The initial Aptian rifting (ca. 125 Ma) phase along the Equatorial margin had a strong dextral shear component that led to the creation of small pull-apart basins filled with thick, continental sedimentary sequences (Trosdtorf et al., 2007) (Fig. 3). By Late-Albian time (102 Ma), Brazil had broken free from West Africa, ending dextral shear of continental crust along the margin (Antobreh et al., 2009). Oceanic waters invaded the basin from north to south during the Late Aptian (ca. 112 Ma); a lagoonal anoxic sequence, the Codó Formation (Fig. 3), overlain by the Albian (112–100 Ma) marine Canárias and Cajú Groups (Fig. 3) (Trosdtorf et al., 2007).

An oceanic connection between the waters of the Central Atlantic and South Atlantic was established during Cenomanian/Turonian time (ca. 100–90 Ma) (Antobreh et al., 2009). Following establishment of a passive margin, local sedimentation and gravity tectonism was strongly but indirectly influenced by Andean tectonism. Overlying the Cajú Group is the Turonian (ca. 90 Ma) through Oligocene (ca. 22 Ma) Humberto de Campos Group, with deposition of the time equivalent Areinhas Formation on the continent, the Ilha de Santana Formation on the shelf, and the Travossas Formation in deepwater (Trosdtorf et al., 2007) (Fig. 3). Gravity tectonics have deformed the Travossas Formation and overlying units.

Sedimentation increases during the Miocene have been recognized north of the study area. Figueiredo et al. (2009) compared biostratigraphic data with isotopic data to establish provenances and times of erosion and redeposition of sediment on the Amazon Fan and constructed paleogeographic maps for the Miocene (Fig. 4). Figueiredo et al.’s (2009) paleogeographic maps show a change of drainage direction linking the Western Amazonia wetlands to the Amazon Fan at ca. 6.8 Ma (Fig. 4).

DATA AND METHODOLOGY

Ten regional stratigraphic horizons, roughly corresponding to the main sequences described by Trosdtorf et al. (2007) (Fig. 3), were identified in wells on the shelf and one well in deepwater, and tied to distinctive seismic reflections (Fig. 2). The data set consisted of a 4×8 km spaced grid of two-dimensional (2D) seismic lines covering the whole deformational system. Interpreted horizons are (1) top of the sedimentary basement, (2) Albian/ Cenomanian (102–98 Ma), (3) Turonian (93.5–89.3 Ma), (4) Santonian (85.8–83.5 Ma), (5) Middle Campanian unconformity (82–78 Ma), (6) Maastrichtian unconformity (70.6–65.5 Ma), (7) Eocene (Lutetian/Bartonian) unconformity (42–37.2 Ma), (8) Oligocene (Chattian) unconformity (23.03–28.4 Ma), (9) Miocene (Tortonian/Serravalain) unconformity (13.65–7.246 Ma), and (10) water-bottom.

One regional cross section 130 km long was depth converted using constant average interval velocities for each of the map horizons. This cross section was restored using Lithotect (Geo-Logic Systems, Halliburton), assuming plane-strain deformation.

SEISMIC INTERPRETATION OF THE BARREIRINHAS BASIN

Top of Basement

The interpreted top basement surface is a combination of the top of the prerift mega sequence and crystalline basement on continental crust described by Trosdtorf et al. (2007), and the top of the oceanic crust as defined seismically (Figs. 5–8). The basement map (Fig. 9) shows that the transition from continental crust to oceanic crust corresponds to the change from basement depths of <3000 ms on the shelf to depths of >6000 ms on the toe of the slope across a transitional zone 10–20 km wide (Fig. 5). The interpretation of the nature of the crust is based on changes in the free-air gravity anomalies (Fig. 1) and basement depths (Fig. 9). Rift-related basement faults formed horsts and grabens on what is now the continental shelf (Fig. 5). Rift faults cut through the top of the basement surface (red surface) and mostly terminate upward below the early drift sequence (yellow). The rift sequence can be approximated as the interval between the basement and the top of the Albian and is relatively thin in the study area, generally <1000 ms (Fig. 5). As previously noted, the abrupt transition zone between oceanic and continental crust is appropriate to a transform to oblique margin (Fig. 1).

Albian (ca. 112–100 Ma)

The mapped Albian top (ca. 100 Ma; deeper yellow on the seismic lines in Figures 5–8) corresponds approximately to the top of the early drift sequence and is the first truly time-correlative sequence in both the shelf and deepwater.

Cenomanian/Turonian (ca. 100–89 Ma)

The top of Turonian (ca. 89 Ma; purple on the seismic lines in Figures 5–8) is a continuous and high-amplitude surface that has a characteristic seismic character that can be easily correlated throughout the Equatorial margin. It corresponds to a maximum flooding surface at the top of the Cenomanian/Turonian condensed sequence, composed mainly of mudstones and marls (Trosdtorf et al., 2007) (Fig. 3).

The detachment surface beneath the Cretaceous deformed rocks lies within the Turonian. Because of the steep shelf margin, there is a large difference in total depth between the continental shelf and the toe of the slope within <20 km (Fig. 5).

Base of Coniacian to Top Santonian (ca. 89–83 Ma)

Two seismic units are included within this age span, the top of the Cretaceous deformed sequence horizon (cyan on the seismic sections) and the top of the Santonian sequence (ca. 84–83 Ma; pink on the seismic sections, Figs. 5–8). The age of the deformed Cretaceous section (cyan) is uncertain, but it lies below the top of Santonian and above the base of the Cenomanian/Turonian (Figs. 5–8) implying an age younger than the basal Turonian (ca. 89 Ma) and older than the basal Campanian (ca. 83 Ma).

Base Campanian to Top Maastrichtian (83.5–65.5 Ma)

The Campanian sequence onlaps the top of the deformed Santonian sequence (Fig. 10). At the end of the Campanian (ca. 83 to ca. 75 Ma), sea level dropped causing a change from mainly transgressive to mainly regressive sequences (Trosdtorf et al., 2007, after Haq et al., 1987). The base of the regressive sequence is the top Campanian horizon (ca. 83 Ma; yellow on the seismic sections, Figs. 5–8) and the top of the first regressive package is the top of Maastrichtian (ca. 65 Ma; green on the seismic sections, Figs. 5–8). The overall regression is characterized by progradation of the shelf break from its position at the top of the Campanian to its more seaward position at the top of Maastrichtian (Figs. 5–7).

The Maastrichtian was a time of tectonic quiescence in the basin and the Maastrichtian sequence buries the Campanian ponded mini-basins. The top of the Maastrichtian is the youngest map horizon deposited prior to the onset of Cenozoic folding and faulting, and is the oldest sequence without growth structures (Fig. 5).

Paleocene/Eocene (ca. 66–37 Ma)

No Paleocene horizon was mapped in this study because of limitations of the seismic resolution, therefore it is unclear if deformation actually commenced during the Paleocene. The top Eocene horizon (ca. 42 Ma; light green horizon on the seismic sections, Figs. 5–8) corresponds to the base of the first sequence that can be identified as a structural growth sequence for Cenozoic fault movement.

Oligocene (ca. 34–23 Ma)

During the Oligocene sea-level rise resulted in a change from regressive in the Rupelian (ca. 34–28 Ma) to transgressive in the Chattian (ca. 28–23 Ma) forming the Upper Oligocene unconformity (Fig. 3) (Trosdtorf et al., 2007, after Haq et al., 1987) (orange horizon; Figs. 5–8).

Miocene (ca. 23–11 Ma)

The “top Miocene” horizon (ca. 11 Ma; cyan horizon on the seismic sections, Figs. 5–8) corresponds to a rapid transgressive event on the whole Brazilian Equatorial margin that starved the region of clastic input and gave rise to a large carbonate ramp (Fig. 3) (Trosdtorf et al., 2007).

Upper-Miocene to Present (ca. 11–0 Ma)

In the latter Miocene the region experienced the onset of another major clastic influx. The sequence above the late Miocene horizon (ca. 11 Ma) is a prograding sequence of deepwater unconsolidated ponded basin mud and clay sequence that fill accommodation spaces created by the folded structures (Figs. 5–7). The present-day seafloor is a folded, faulted, and eroded surface cut by numerous Pleistocene seafloor canyons (Figs. 5–7).

SEDIMENTATION RATES

Sedimentation rates are a critical factor in gravity-driven fold-thrust belts because high sedimentation rates lead to steepening of the continental slope, and downslope movement serves to reduce the oversteepened surface gradient. Sedimentation rates were calculated for the single deepwater well location, highlighted in red (zoom inset, Fig. 2), using unit thicknesses of faunally defined age horizons that correspond to the reflectors in Figure 5. Depth-converted seismic stratigraphic horizons were used to estimate sedimentation rates shallower and deeper than the faunally defined horizons; decompaction was not taken into account. Sedimentation rates were also calculated for a point on the shelf using seismic horizons only. The results for both locations are probably perturbed by slumping and erosion, but are the only data available. The calculated sedimentation rates are summarized in Table 1 and plotted on Figure 11.

Sedimentation rates as high as 200 m/m.y. in deepwater and 90 m/m.y. on the shelf prevailed during the Coniacian (ca. 89–83 Ma; Table 1 and Fig. 11) interval. The Coniacian-Santonian is a mud-dominated interval with a sedimentation rate similar to that of the modern Amazon Fan (Nancy Engelhardt-Moore, 2009, personal commun.). Fossil recovery from well cuttings was poor, consistent with a rapid deposition mass wasting environment. These high rates of sedimentation coincided with the Cretaceous rapid deformation of the Travossas Formation in deepwater (ca. 89–83 Ma).

Maastrichtian (ca. 70–67 Ma) sedimentation rates were also high, 86 m/m.y. in deepwater, and 66.7 m/m.y. on the shelf (Table 1 and Fig. 11), likely associated with an episode of tectonic uplift that affected the whole Brazilian Equatorial margin, as described by Zalan (2004) extending as far south as the Camamu Basin (Cobbold et al., 2010). During Late Cretaceous to Paleocene time two erosional events are observed in the deepwater but not on the shelf, and are represented on Figure 11 as extremely low sedimentation rates. We estimate that the events occurred between 83.5 and 70 Ma (7.5 m/m.y.) and 67–55 Ma (15 m/m.y.), and interpret them as local erosion on the top of large fold structures (Fig. 7).

From the end of the Paleocene to Mid Miocene time (55–10 Ma), rates of sedimentation were very low on the shelf (4–7.1 m/m.y.), but higher in deepwater (22.2–40 m/m.y.) areas as sediment on the shelf was eroded or bypassed, and deposited in the deep basin.

From Late Miocene to the Present (10–0 Ma; Table 1), sedimentation rates increased both on the shelf (40 m/m.y.) and in deepwater (37.5 m/m.y.). A Late Miocene, post–10 Ma, high sedimentation pulse coincides with rearrangement of the drainage system east of the Andes, and the birth of the modern Amazon drainage (Figueiredo et al., 2009), indicating that Late Miocene drainage rearrangement affected areas farther south than previously recognized.

KINEMATIC ANALYSIS OF THE BARREIRINHAS BASIN

The postrift structural evolution of this part of the Barreirinhas Basin is dominated by a series of collapse systems involving detachment surfaces within shale units. Other fold structures comparable to those analyzed have been mapped in the Barreirinhas and Pará-Maranhão basins (Zalan, 2011) and in the Foz do Amazonas (Cobbold et al., 2004; Araujo et al., 2009; Perovano et al., 2009), and each is a highly complex three-dimensional system. Variations in the exact timing of fault movements within the basin seem likely. Our work focuses on one representative set of structures, and does not necessarily depict timing or structural details in other parts of the basin.

Structural Palinspastic Restorations

To better understand deformation rates on the faults, fault propagation, and linkage between faults, a present-day deformed section was restored to four earlier configurations (Fig. 12). The restorations in each case assume a continuous and planar sea-bottom slope from undeformed shelf sediment to undeformed basin-floor sediment, following a seismic time horizon correlated to biostratigraphic data and the seismic-stratigraphic model. The method does not account for possible minor variations in sea-bottom topography, but provides an adequate basis for structural analysis.

Restorations were constructed preserving bed-lengths and assuming flexural slip/flow kinematics. The sections are subperpendicular to the trends of folds, thrusts, and normal faults. Note that there are cross-structures that represent shortening along the b-kinematic axis of thrust sheets, but along-strike strain is generally <1%, and therefore negligible for our purposes.

Present (0 Ma)

At present most of the deformation in the study area is distributed among four normal faults and two large thrust faults. The normal faults (FN-1, FN-2, FN-3, and FN-4) and thrust faults (FR-1 and FR-2) are labeled in Figure 12. There is also a smaller back-thrust fault associated with a fold collapse feature but that is a minor structure (Fig. 12). The present-day shortening measured on our representative cross section is ∼2200 m and the present extension is ∼1500 m (Fig. 12).

Miocene (10 Ma), Oligocene (27 Ma), and Eocene (42 Ma)

Displacement on the four normal faults and the two thrust faults was restored to the paleogeometries of 10, 27, and 42 m.y. ago. Restorations of the normal faults resulted in extensions of 420 m (10 Ma) and 180 m (27 and 42 Ma). Restoration of the thrust faults resulted in shortening of 530 m (10 Ma) and 430 m (27 and 42 Ma). Most of the deformation (75% of the shortening and 72% of the extension) could be restored at 10 Ma (Fig. 12). After the second restoration time-step (27 Ma) 80% of the shortening and 90% of the extension is resolved, no change resulted from 27 to 42 million years (Fig. 12).

Santonian (83.5 Ma)

During the Coniacian-Santonian (Fig. 12) a prograding shelf and a high sedimentation rate in deepwater (226 m/m.y.) caused slope instability and triggered the formation of a set of normal listric faults on the shelf and thrust faults at the toe of the slope. This wide linked extensional-compressional system developed very rapidly, all the deformation and the infilling of the deformed seafloor took place during the Santonian (ca. 86–84 Ma) within the sequences represented in blue and pink in Figure12. Because of limited stratigraphic resolution within this interval, and poor seismic imaging due to subsequent deformation by a younger fault system (Fig. 13), detailed interpretation of individual faults is not possible. Depiction is schematic for this event (Fig. 12).

Cenozoic Fault Analysis

Integration of detailed structural analysis with a well-defined stratigraphic age model provides the opportunity to determine finite (i.e., long-interval) fault motion rates over the Cenozoic history of the basin (Fig. 13). For each fault the fault-parallel displacement was measured for a representative pregrowth section, i.e., the top of Cretaceous (Fig. 14).

The identification of the timing of fault motion also allows us to infer linkages of various faults in the systems; shelfal normal faults moving simultaneously with basinal reverse faults can reasonably be assumed to link across the intervening translational zone.

The displacement versus time plot (Fig. 15) demonstrates a variable deformation rate through time. Deformation that began in the Eocene (ca. 42 Ma) continued during the Oligocene with motion on both normal and thrust faults, but deformation rates were always slow (Fig.15). Deformation rates increased significantly in the Miocene as indicated by the expanded Miocene section on the downthrown side of shelf-margin normal faults (Fig. 3). In post-Miocene time, deformation rates continued to increase (Fig.15), with major normal growth faulting forming synclines on the shelf margin, and fold crests rising toward the sea surface at the toe of the slope (Fig.13). Additional normal faults (FN-3 and FN-4) developed in the footwall of preexisting normal faults (FN-1 and FN-2) (Fig. 13).

DISCUSSION

Coniacian/Santonian Bed-Parallel Gravity Gliding

The very brief time interval of ca. 89–84 Ma (duration ca. 5 Ma) corresponds with a period of eustatic sea-level fall (Trosdtorf et al., 2007, after Haq et al., 1987; Fig. 3) and increased tectonism in the Andes (Zalan, 1998), which resulted in a very high sediment influx of 226 m/m.y. to the basin. The combination of a steep basement slope (10°–15°) above the narrow continental to oceanic crustal transition zone with this period of high sedimentation rate and consequent oversteepening of the surface slope led to instability of the slope. This Santonian slope instability generated a set of linked listric normal faults on the shelf and thrust faults at the toe of the slope. On the shelf a thin deformed sequence is characterized by a 30-km-wide zone of listric normal faults detached within the underlying sequence of Cenomanian to Turonian marls and shales (Fig. 10). The extensional domain is linked by a 30-km-wide translational domain without visible internal deformation to a compressional domain. On the toe of the slope the sequence is deformed by a set of landward dipping thrust faults forming a belt of imbricate thrust sheets with 30 km in the dip direction (Fig. 5) and 30 km in the strike direction (Fig. 8).

Zalan (2011) concluded that Cretaceous gliding and thrusting marked the onset of deformation that was semicontinuous throughout the Cenozoic. However, consistent age of sediment infill of local basins created by normal faults on the shelf and by thrusts in the basin (Fig. 16), and cross-cutting relationships in which Cenozoic faults sole at a deeper stratigraphic level and rotate the Cretaceous thrusts (Fig. 13), indicate two discrete tectonic events separated in time.

Syndeformational Santonian age sediment was deposited on the top of both the rotated blocks of the shelf and the folds at the toe of the slope. This onlapping sequence is highlighted in pink in Figure 10. The isochron map for the interval between the top of the Turonian and the top of the Santonian (Fig. 16) includes the thickness of the Cretaceous syndeformational sequence and part of the postdeformational sequence; most of the sediment, up to 1000 ms, accumulated on the slope in pocket mini-basins similar to those described by Hooper et al. (2002) and Corredor et al. (2005) in the Niger Delta.

The Cretaceous allochthon is ∼1 km thick (Fig. 10) and ∼70 km in downdip extent (Fig. 14). Gliding was facilitated by a combination of a low basal slope and high pore pressures in the Turonian shales, but even with an extremely efficient detachment the area-to-thickness ratio of the allochthon is unusual. Stratigraphic resolution is limited within this thin interval, and seismic imaging is poor due to subsequent deformation by the Cenozoic fault system (Fig. 13). Detailed interpretation of individual faults is not possible, and depiction is schematic for this event (Fig. 12).

Maastrichtian through mid-Eocene Structural Quiescence

During this time interval, the shelf margin continued to prograde into deepwater, but was structurally stable with the exception of reactivation of one basement-involved fault in the narrow zone of extended continental crust (Fig. 6), an observation also made by Zalan (2011).

The isochron map between the Maastrichtian and the Eocene horizons (Fig. 13) shows thick and thin axes of deposition of the Paleocene/Eocene sediment (ca. 65–37 Ma). On the shelf break, sediment eroded from the footwall and redeposited on the hanging wall of the normal faults, forming elongated depocenters parallel to the shelf break (Fig. 17). On the continental slope, sediment was deposited while thrust faults moved, resulting in the establishment of a growth sequence. Sediment deposited on the hanging wall of the thrust faults and formed ponded mini-basins. Thrust fault movement formed folds on their hanging walls that created anticlines on the Paleocene/Eocene seafloor. These anticlines were soon partly eroded and sediment redeposited on the thrust fault footwalls (Fig. 17).

Mid-Eocene Onset of Nonbed Parallel Gravity Sliding

During mid-Eocene the shelf margin collapsed, developing two normal faults (FN-1 and FN-2) and a thrust fault (FR-2 in Fig. 13). The two normal faults linked to thrust faults FR-1 and FR-2, forming concave detachment faults (Fig. 13).

Deformation in the system persisted for at least another 40 m.y., until the Present. The Cenozoic deformed area is less extensive (∼30 km) both downdip and along strike (Fig. 14) than the Cretaceous deformed area (Fig. 18).

The Eocene fault systems cross-cut bedding and form two bowl-shaped fault systems at two different depths (Fig. 13). The bowl geometry is different from the classic shale detached deepwater fold and transform belt observed in Nigeria (Corredor et al., 2005; Cobbold et al., 2009) and described here for the Cretaceous section (Fig. 18).

The isochron map (Fig. 17) of the thin mid-Eocene to Oligocene interval suggests that during the Oligocene transgression, most of the sediment was trapped on the continental shelf and slope. On the continental slope, sediment was trapped in ponded mini-basins created by normal and thrust faults. The largest fold was partly breached by erosion, allowing sediment to bypass to the abyssal plain (Fig. 17).

The Oligocene to mid-Miocene unconformity isochron map (Fig. 17) indicates a time of erosion on the shelf and deposition in deepwater. The isochron map shows erosional channeling of the continental shelf and slope, and breaching of the major anticline continued to allow bypassing of sediment onto the abyssal plain.

Mid-Miocene Acceleration of Deformation

Deformation rates increased dramatically in mid-Miocene time, as indicated by the expanded pre–10 Ma Miocene section on the downthrown side of the shelf-margin normal faults (Fig. 14). This accelerated fault motion corresponds in time to increased sedimentation rate (Fig. 11), and a transition from a carbonate ramp to prograding siliciclastics along the entire Equatorial margin.

The Late-Miocene to Present rapid deposition, highlighted in blue in Figure 13, generated a thick, progradational sequence depicted in the Miocene to Present isochron map (Fig. 17). The accumulation is the thickest sequence deposited in the ponded mini-basins, and caused mini-basins to coalesce. Localized depocenters are thickest near the Miocene shelf break, above the hanging walls of the normal faults, and in basins that formed behind the major anticline on the thrust faults at the toe of the slope (Fig. 17). The fold was locally breached, allowing sediment to be deposited onto the abyssal plain.

It is likely that the high sedimentation and deformation rates observed in the Barreirinhas and other basins during the Late Miocene are an indirect consequence of the major drainage reorganization in northern South America that began ca. 11 Ma (Altamira-Areyan, 2009) and diverted the drainage west of the Purus arch from the Caribbean and into the present-day Lower Amazon Basin (Fig. 4) (Figueiredo et al., 2009). That may also be the case for other large deepwater fold and thrust structures developed on the Brazilian Equatorial margin from the Amazon Cone (Araujo et al., 2009; Perovano et al., 2009) to the Barreirinhas Basin (Zalan, 1998, 2004, 2005, 2011; Gilbert, 2006, 2011; Krueger and Gilbert, 2009; Krueger et al., 2011).

Deformation rates continued to increase, with major normal growth faulting on the shelf margin, and uplift of folds at the toe of the slope (Fig. 13). Approximately 80% of the net strain in the area took place within the last 10 m.y. (Fig. 12).

Two new normal faults (FN-3 and FN-4) developed in the footwall of preexisting normal fault FN-2, and linked to the existing FN-2/FR-1 system (Fig. 13). Fault motion rates increased significantly (Fig. 15) as indicated by the expanded Miocene section on the downthrown side of the shelf-margin normal faults (Fig. 13).

Major Pleistocene to Holocene canyon systems were subsequently incised into the shelf margin, cutting both normal faults and growth folds (Fig. 17). The three-dimensional effects of the deformation can be seen in the Miocene (10 Ma) to Present isochron map (Fig. 17) that shows development of a thick ponded mini-basin bounded by normal faults landward and thrust faults basinward.

Driving Mechanisms

Shale-detached deepwater fold systems, similar to those of the Barreirinhas Basin, have been described at several continental margins, particularly the Niger Delta (Damuth, 1994; Hooper et al., 2002; Rowan et al., 2004; Krueger and Gilbert, 2006, 2009; Sultan et al., 2007), the Pará-Maranhão and Barreirinhas basins (Zalan, 2005, 2011; Gilbert, 2006), and offshore Namibia (Butler and Paton, 2010). Two factors must be considered in the kinematics of these thrust belts, the generally accepted mechanism that allows thrust motion, and the “triggering event” that initiates deformation.

The critical taper wedge model of Davis et al. (1983) and Dahlen (1984) is often cited as the causative mechanism for thrusting, but it must be borne in mind that the ultimate cause of deformation is the condition or conditions that created the key tapered-wedge requirements: a basal detachment that slopes toward the hinterland, and a surface slope toward the foreland. In deepwater fold belts, it is generally accepted that these conditions are largely driven by rapid sediment progradation linked to a sea-level drop or tectonic events. In addition, deepwater fold and thrust belts are invariably linked to gravitational collapse and updip extension on the shelf, and therefore are not strict analogs to the foreland fold and thrust belts and subduction zones from which the model was developed.

In areas of very thick sediment accumulation such as the Amazon Cone (Araujo et al., 2009; Perovano et al., 2009) and Niger Delta (Corredor et al., 2005; Billotti and Shaw, 2005; Cobbold et al., 2009), the sediment weight depress the lithosphere (Morley et al., 2011) causing the slope of the basal contact of the sedimentary pile to dip landward, which is also observed to a lesser extent in the Barreirinhas Basin (Zalan, 2011).

In the Barreirinhas Basin, a steep continental slope combined with rapid sediment progradation generated an unstable surface slope. The evolution of the sedimentary wedge top through time was measured on the seismic lines and on the structural restorations: ∼4° during the Santonian deformation, 3.3° in mid-Miocene time, 3.6° in the mid-Oligocene, 3.8° in mid-Miocene, and steepening abruptly to 5° in the mid-Miocene to Present, the time of major deformation (Fig. 12).

The compression on the toe of the slope is caused by friction at the detachment level and cohesion of the sliding rocks. Zalan (1998) associated compression at the toe of the slope with a slowdown in the gravity-driven movement of the sediment due to either (1) a change in the gradient of the detachment layer or (2) the buttressing effect of a more rigid body, such as an igneous intrusion, an ancient volcano, or a protruding rift-phase domino-type fault block. Where observed, buttressing effects are relatively isolated, and have the localized effect of forming imbricate fans, and locally deflect the trend of the fold belt (Fig. 8). In the absence of any of the other features, we interpret the change in gradient as the causative factor for the compression at the toe of the slope.

At the toe of the slope, compressional deformation by thrusting and folding occurs when the taper conditions are no longer sufficient to support basal shear on the sediment pile (Davis and Kusznir, 2004; Dahlen, 1984). The change in basal gradient typically occurs at the top of the oceanic crust. The loss of water-bottom gradient can be the result of accumulation of sediment at the toe of the slope via thrusting and folding or redeposition of eroded sediment from the shelf (Fig. 6). In the Barreirinhas Basin the basal slope is close to zero, therefore sediment surface is the key element to allow further thrusting.

The ultimate origin of the deformation is more problematic. Zalan (2011) speculated that earthquakes associated with reactivation of nearby oceanic fracture zones could have been a triggering mechanism for deformation. Undoubtedly such events could trigger discrete movement events, but given the long (40 m.y.) duration of the Cenozoic faulting, and the direct temporal correspondence of high deformation rates with high sedimentation rates, sediment loading is favored here as the controlling factor. Distal tectonic events in the Andes may have influenced Equatorial margin gravity deformation indirectly by impacting sediment input.

SUMMARY AND CONCLUSIONS

The structural architecture of the Barreirinhas Basin is dominated by two major deepwater fold and thrust belts linked landward to extensional fault systems, but the Cretaceous and Cenozoic faults and folds have markedly different characteristics. The short-lived Cretaceous system is <1.5 km thick, and involved listric normal faults and small stacked imbricate thrusts linked by a bed-parallel décollement. The Cenozoic fault system cuts through 4 km or more of Cretaceous and Cenozoic sediment, and cross-cut the preexisting Cretaceous deformed sequence. Normal faults connect to the thrust faults at depth, and cut across bedding, forming two bowl-shaped “mega-slumps.”

Integration with an age model developed from sequence stratigraphy and paleontological data allows the determination of finite fault motion rates for the Cenozoic system. Movement on both normal and thrust faults began at the same time, and fault motion rates for both shortening and extension varied simultaneously, suggesting linkage between normal and thrust faults in deformation rate and net strain. This supports the idea that extensional deformation on the shelf is being accommodated by shortening on the toe of the slope. In addition, the results demonstrate a temporal relationship between sedimentation rate and fault motion rates, verifying that gravitational loading by rapid sedimentation is likely the driving mechanism.

Relative fault timing indicates that the extensional province propagated landward, and that thrust faults raised the elevation of the lower slope, lowering the slope gradient. As more faults were introduced into the system the location of the shelf break moved landward, maintaining the slope gradient at or below 5°.

Establishing deformational mechanisms and strain rates helps to fill the gap between our knowledge of small-scale, geologically instantaneous gravity-driven submarine slumps or slides (a sedimentological phenomenon) and the larger-scale, slower gravity-driven thrusts (a structural phenomenon). Closing this gap may aid in understanding the generation of mass-flow deposits at steep margins, with implications for sedimentology, basin analysis, and hydrocarbon exploration.

WesternGeco kindly provided permission to publish the four seismic lines depicted here. Much of the mapping and structural restoration was done through the auspices of Devon Energy, and we especially acknowledge permission to use the interpreted seismic grids that are the basis of the isochron maps. Structural restorations were performed using LithoTect software. Special thanks to Michael Hankins (HRT America, formerly of Devon); to Pedro Zalan, Ivo Trosdtorf, and Jorge Picanzo Figueiredo (Petrobras); and to Dale Bird (Bird Geophysical) for discussions of the local and regional geology. Nancy Engelhardt-Moore of Devon Energy provided key information on the implications of the paleontological data to both age-dating and sedimentation rates. Also special thanks to Laura Unverzagt (University of Houston) for help with ArcGIS, and lastly, the editors and reviewers, especially Peter Cobbold, who provided much-needed suggestions for revisions.