Gravitational failure of passive margins

1

• Gravitational failure of passive margins
• Gravitational failure of passive margins occurs
  by some combination of gravity gliding (the
  translation of a mass down a sloping detachment)
  and gravity spreading (the vertical collapse and
  lateral spreading of a mass under its own weight)
  (Fig. 130). The overall pattern is one of updip
  extension balanced by downdip shortening. The
  deformation may be detach simply on the
  autochthonous salt layer, as in offshore Angola
  (Fig. 131), or it may involve multiple detachment
  layers including the autochthonous salt,
  allochthonous canopies, and shale detachments, as
  in the northwestern GoM (Fig. 132).
• There are exceptions to the general pattern of
  updip extension and downdip contraction. For
  example, since deformation is often driven by
  failure of the continental slope, which progrades
  basinward over time, early contractional
  structures may be overprinted by later
  extensional faulting (Fig. 133). But there are
  also examples where the opposite is true due to
  reactivation during shifts in the drive for
  gravitational failure or renewed basement
  deformation caused by plate reorganization (Fig.
  134).
• Since it is the passive margin that is failing,
  any spatial or temporal variations in margin
  geometry will influence the deformation. For
  example, a margin like that in Brazil has
  segments that are convex towards the ocean and
  those that are concave(Fig. 135). This will
  result in either divergent or convergent
  deformation (Fig. 136). In the case where a
  deltaic lobe builds out, creating a locally
  divergent margin, models show that deformation
  consists, of course, of proximal extension and
  distal contraction, but also distal

• concentric extension as the lobe spreads radially
  (Fig. 137). When the sediment lobe shifts
  laterally due to fluvial avulsion, early
  structures may be reactivated in oblique
  directions or overprinted by structures with
  different trends (Fig. 138).
• Gravitational failure is inherently
  self-limiting. When the gravity potential becomes
  large enough to overcome the resistance of the
  rocks to deform, failure occurs. But the very act
  of failure reduces the gravity potential so that
  deformation will stop unless the gravity
  instability is maintained or increased. In the
  case of gravity gliding, the dominant factor is
  the dip of the basal detachment. Thus, basinward
  translation is accelerated by differential
  thermal subsidence and cratonal tectonic uplift,
  both of which increase the dip of the detachment,
  and is retarded or halted by differential loading
  subsidence, which decreases the basinward tilt
  (Fig. 139). In the case of gravity spreading, the
  dip of the basal detachment is also important,
  but so is the dip of the sea-floor slope. A steep
  slope can cause failure even when the detachment
  dips back towards land. Thus, shelf aggradation
  and upper slope progradation drive further
  shortening because they steepen the slope and
  increase the gravitational instability, whereas
  lower slope and abyssal sedimentation retard or
  stop deformation, both by reducing the slope and
  by thickening and thereby strengthening the
  distal section (Fig. 140).

2

• The tectonic history of different margins can be
  quite different, leading to variable histories of
  gravitational failure. For example, in the Kwanza
  Basin of offshore Angola, gravity-driven
  deformation has been continuous since the Albian,
  shortly after salt deposition (Fig. 141). But the
  rates were not constant there was an initial
  pulse driven by differential thermal subsidence
  that slowed over time (as is typical on passive
  margins), followed by a later increase in
  deformation rates caused by several episodes of
  Tertiary uplift of the craton and basinward
  tilting of the margin (e.g., Duval et al., 1992
  Cramez and Jackson, 2000).
• In contrast, the Mississippi Fan foldbelt of the
  northeastern Gulf of Mexico shows a more complex,
  four-stage history of gravitational failure.
  Interval thickness patterns indicate that there
  were two episodes of shortening, each followed by
  a period with little or no deformation (Fig.
  142). Figure 143 illustrates the proposed
  evolutionary model (1) the usual early gravity
  gliding driven by differential thermal
  subsidence (2) a quiescent stage as thermal
  subsidence waned and landward loading subsidence
  slowed basinward tilt of the margin (3) the main
  stage of shortening driven by gravity spreading
  of the prograding Tertiary clastic margin and
  (4) cessation of deformation due to canyon
  incision and bypass of the Mississippi Fan to the
  abyssal plain.

3
Figure 130. Modes of gravitational failure (a)
gravity gliding, which is the translation of a
rigid body down a sloping detachment (b) gravity
spreading, which is the vertical collapse and
lateral spreading of a body with a surface slope
and (c) a combination of the two processes, which
is most common (Rowan et al., 2003b).
4
Figure 131. Squashed seismic profile and cross
section of the Angolan margin showing relatively
simple updip extension and downdip contraction
detaching on the autochthonous salt layer (Marton
et al., 2000).
5
Figure 132. Regional section through the
northwestern Gulf of Mexico, showing updip
extension and downdip shortening occurring at
multiple levels of both salt and shale (Peel et
al., 1995).
6
Figure 133. Cartoon showing how progradation may
result in early toe-of-slope contraction may be
overprinted by outer shelf extension (courtesy of
B. Vendeville).
7
Figure 134. Seismic profile and interpretation
showing early extension overprinted by later
contraction (Cramez and Jackson, 2000).
8
(b)
Figure 136. Experimental models showing
gravitational failure where there is divergent
gliding (a) or convergent gliding (b) (Cobbold
and Szatmari, 1991).
(a)
9
Figure 137. Deformation associated with the
gravitational failure of a sedimentary lobe
proximal concentric extensional structures,
distal radial extensional structures, and distal
concentric contractional structures (courtesy of
B. Vendeville).
10
Figure 138. Superposed deformation patterns when
sediment lobes shift laterally (courtesy of B.
Vendeville).
11
Figure 139. Gravity gliding is driven by
basinward tilting of the detachment, which in
turn is controlled by various factors (Rowan et
al., 2003b).
12
Figure 140. Gravity spreading is controlled by
the sea-floor slope, so is driven by shelf
aggradation and upper slope progradation but is
halted by canyon incision and bypass
sedimentation on the lower slope and abyssal
plain (Rowan et al., 2003ba).
13
Figure 142. Deepwater fold in the Mississippi Fan
foldbelt showing (1) an early stage of
deformation (2) an interval of
constant-thickness and thus no deformation (3)
the main growth interval deposited during fold
amplification and (4) cessation of deformation
(Rowan et al., 2003b).
14

• Updip extension
• Examples of extensional structures from both the
  western and eastern South Atlantic margins are
  illustrated in Figures 144 and 145. The
  characteristic style is one of basinward-dipping
  listric normal faults that sole into the salt
  layer or equivalent weld, although
  landward-dipping faults are also observed. The
  faults intersect the salt at cusps that become
  triangular salt rollers once the weld forms. Note
  that these are not intrusions or diapirs of salt
  they are simply remnants of the original salt
  layer left in the footwalls of the faults. Early
  faults have a short wavelength and later faults
  have a greater spacing (Fig. 146), a typical
  pattern that is related to the thickness and thus
  strength of the overburden being extended. In
  some places, the trains of normal faults are
  broken up by essentially undeformed blocks called
  rafts (Figs. 147 and 148). Rafts form when the
  hanging wall of a fault separates completely from
  the footwall and moves coherently along the salt
  detachment, with the area between rafts being
  filled by younger growth strata (Fig. 149).
  Although rafts are best known from the South
  Atlantic margins, they also occur in the
  northwestern Gulf of Mexico (Fig. 150).
  Experimental models and observations suggest that
  rafts are favored when the detachment dip is
  relatively low and that asymmetric listric faults
  form when the basinward dip is greater (Fig.
  151).

15
Figure 144. Listric normal faults in the
extensional province of the Angolan margin (Duval
et al., 1992).
16
Figure 145. Extensional structures in the
Espirito Santo basin of Brazil (courtesy of C.
Fiduk and CGG).
17
Figure 146. Regional seismic profile (split into
two parts) across the extensional province of the
Brazilian margin (Mohriak at al., 1995). Note
that older faults have a smaller spacing and
younger faults are more widely spaced.
18
Figure 147. Seismic line and interpretation from
offshore Angola showing an undeformed raft
between two younger depocenters (Duval et al.,
1992).
19
Figure 148. Regional cross section showing a
number of rafts and intervening depocenters
formed during extension (Duval et al., 1992).
20
Figure 149. Model showing the evolution of rafts
from typical listric normal faults as extension
progresses (Duval et al., 1992).
21

• Downdip contraction
• Updip extension is balanced downdip by some
  combination of thrusting, folding, squeezing of
  diapirs, and salt inflation and extrusion (Fig.
  152). The shortening typically takes place in a
  zone just landward of the distal pinchout of the
  salt layer (Fig. 153). But it can also occur at
  the toe of the slope, where contractional
  stresses are high, with the implication that
  areas of contractional folding can shift
  basinward over time as the slope progrades (Figs.
  154 and 155). In addition, contractional
  structures may occur wherever there is a barrier,
  such as a landward-dipping step in the base salt.
• An unusually simple example from the Perdido
  foldbelt in the northern GoM is shown in Figures
  156 and 157, illustrating regular-wavelength,
  highly elongate, symmetrical detachment folds.
  This correlates to a prekinematic section that is
  essentially constant thickness (Fig. 156). In
  contrast, folds in the Mississippi Fan foldbelt
  from the same margin are shorter and have more
  irregular orientations (Fig. 159). The frontal
  fold comprises four segments with separate
  culminations, and serial profiles across this
  composite fold show large along-strike variations
  in fold geometry (Figs. 160 through 162). This is
  probably due to an early stage of deformation
  during the early gravity gliding phase shortly
  after salt deposition (see thickness variations
  in the section immediately above salt in Figs.
  160-162). The early structures would have served
  as nucleation points for the initiation and
  growth of later folds. An important implication
  of this complex three-dimensional fold geometry
  is that one cannot use observed portions of folds
  to predict along-strike fold geometry under
  allochthonous salt.

• An isochron map of the section above the salt in
  the Mississippi Fan foldbelt shows the geometry
  of the early deformation event (Fig. 163).
  Although the Miocene folds have a relatively
  simple geometry (Fig. 159), the older structures
  are highly irregular and might appear to be
  caused by differential loading and salt
  withdrawal rather than by shortening. However,
  the same structures to the southwest (e.g., in
  Walker Ridge) have a more linear NE-SW
  orientation, and they curve around to a NW-SE
  orientation in the eastern part of the foldbelt
  (grey lines in Fig. 164). In other words, they
  are parallel to the Cretaceous carbonate shelf
  margin, which is consistent with their
  development during gravity gliding down the
  paleo-slope. The reason for the complex
  structures in Figure 163 is that this was an area
  of convergent gliding, producing dome-and-basin
  interference structures (see Fig. 136b). Note
  that the absence of early deformation in the
  Perdido foldbelt suggests that early gravity
  gliding was accommodated farther landward.
• The Mississippi Fan foldbelt (as well as analogs
  in other basins such as offshore Angola, the
  Espirito Santo basin of Brazil, and offshore Nova
  Scotia) shows a close relationship between folds
  and diapirs, with all diapirs located along fold
  crests (Fig. 164). A detailed example is shown in
  Figure 165, where the Green Knoll diapir is
  located along the frontal deepwater fold. There
  are two possible explanations for this
  relationship (Fig. 166) (1) the diapirs could
  postdate the folds, forming by breaching of the
  fold crest, in which case the overburden should
  show some record of fold collapse as the salt is
  withdrawn

22

• or (2) the diapirs could predate the folding,
  thereby controlling the location of the folds
  because of the weakness of salt, in which case
  there should be a long-lived growth history
  adjacent to the diapirs. Data show that strata
  that are constant thickness over folds (i.e.,
  prekinematic) often have thickness changes
  adjacent to the diapirs (Fig. 167) or variable
  thicknesses on either sides of these diapirs,
  showing that salt features predated the folding.
  An example of a squeezed diapir from Angola Block
  33 is shown in Figure 168, where the teepee
  structure around the steep weld passes along
  strike to a more typical thrusted structure. A
  different example shows a steep reverse fault
  beneath and adjacent to a shallow salt body, with
  a highly rotated and uplifted hanging wall (Fig.
  169). Note the thinning below the green horizon
  (which marks the onset of shortening), showing
  that there was a long-lived salt structure at
  this location. The geometry is interpreted as a
  weld that was reactivated as a reverse fault
  during diapir squeezing and rejuvenation (Fig.
  169 Green Knoll is interpreted similarly, as
  shown in Fig. 165).
• A key implication of the proposed explanation for
  the observed fold-diapir relationship is that
  shortening can be one of the major causes of
  lateral salt extrusion. Squeezing of a diapir
  greatly increases salt flow rates, which, as we
  shall see later, favors lateral extrusion of
  allochthonous salt. If enough salt is extruded
  and amalgamates to form a regional canopy,
  gravitational failure will then take place above
  this detachment, slowing or even stopping
  shortening above the deeper level (Fig. 168).

• Another, less obvious, manner in which salt can
  reduce the gravity potential is by proximal
  subsidence into a thick salt layer and distal
  inflation of the salt layer (second panel in Fig.
  172 and Fig. 173). Typically, the inflated salt
  breaks through a thin overburden and advances
  basinward as an allochthonous nappe (Fig. 172),
  rafting along the overburden and thrusting it
  over its sub-nappe equivalents (Fig. 173). Distal
  inflation and nappe extrusion was first observed
  in the Angolan deepwater, where the toe of the
  system is marked by massive salt with a thin
  condensed section representing the entire
  post-salt interval, and the last 20-30 km of this
  salt is a nappe that ramps up basinward (Figs.
  174 and 175). More recently, both inflation and
  distal nappes were recognized in the northern GoM
  deepwater. Figure 176 is a restoration of a cross
  section through part of the Mississippi Fan
  foldbelt that shows that the section immediately
  above the salt is thinner than out in front of
  the salt due to early inflation. Moreover, the
  salt underlying a frontal fold in the same
  foldbelt is not actually cored by salt at the
  autochthonous (late Middle Jurassic) level, but
  by salt that is ramping up to a level within the
  Lower Cretaceous (Fig. 177). The extent of this
  nappe as mapped by F. Peel is shown in Figure
  178, who interprets it along the entire foldbelt
  and into Keathley Canyon, with a width ranging up
  to about 50 km. Alternatively, Rowan (2001d)
  showed it more discontinuous and with a maximum
  width of about 20 km. An important implication of
  this geometry is that much of the source rocks
  may lie below the nappe, making both maturity and
  charge higher-risk.

23

• In offshore Angola, inflation and nappe formation
  have been more or less continuous since the time
  of salt deposition, just as obvious shortening
  has been. In contrast, nappe formation in the
  northern GoM ended during the Cretaceous and
  inflation ended variously between the Cretaceous
  and the Miocene. Again, so did the early
  shortening event. Thus, the gravitational
  instability along these margins was accommodated
  near the distal salt pinchout by a combination of
  salt inflation, salt extrusion, rafting of the
  overburden, and contractional folding.
• We have seen that deepwater, salt-cored foldbelts
  are typically characterized by a combination of
  symmetrical detachment folds, sometimes cut by
  relatively minor reverse faults (Fig. 179a). The
  folds are usually polyharmonic because of the
  increasing overburden thickness with time.
  Shortening is also accommodated by the squeezing
  of diapirs (Fig. 179a, b) and the squeezing and
  inflation of distal salt massifs and the
  extrusion of salt from the distal pinchout (Fig.
  179b). What is normally not found are structural
  styles where faulting dominates over folding
  (Fig. 179c). The reason is that the distal salt
  inflation usually provides sufficient salt to
  continue filling fold cores even when salt
  started out thin, so that faulting remains only a
  minor component of the deformation.

• To wrap up this section on distal contraction, we
  examine a case study and introduce a new wrinkle.
  The GoM example shows a complex pattern of
  extensional, contractional, and strike-slip
  structures controlled by a polygonal network of
  remnant Louann salt and scattered diapirs. Figure
  180
• shows the distribution of shallow salt bodies in
  southeastern Mississippi Canyon that include
  isolated diapirs to the east and a canopy to the
  west. The early
• history of the area is shown in two maps. During
  the Late Jurassic through Oligocene, evacuation
  of the Louann salt formed a series of early
  minibasins (Fig. 181). One of these (in the NW)
  was an expulsion rollover structure, but the
  others had more three-dimensional salt movement,
  with salt moving into and inflating a number of
  salt plateaus and ridges (Fig. 181). These
  plateaus subsequently collapsed during the early
  to mid Miocene, possibly triggered by some
  basinward translation of the overburden, forming
  new minibasins with turtle structures as salt
  moved into diapirs (Fig. 182). By the end of the
  middle Miocene, there was an established
  architecture of welded minibasins of different
  ages separated by a polygonal pattern of remnant
  Louann salt ridges, with diapirs located above
  the ridge intersections.

24

• B. Vendeville has modeled the effects of
  shortening on this type of architecture. Figure
  184 shows the initial configuration of
  minibasins, a polygonal network of salt ridges,
  and diapirs (top) and the geometry after
  shortening (bottom). The minibasins, as the
  strongest elements, moved rigidly while the
  diapirs and salt ridges localized the
  contractional strain (Figs. 184 and 185). Diapirs
  were squeezed, forming vertical welds, and
  contractional structures formed over the ridges
  (Fig. 185). What is only subtly apparent in the
  figures is that some of the structures
  accommodated strike-slip or transpressional
  movement. Moreover, some of the minibasins
  rotated about vertical axes, so that deformation
  on one flank could range from contractional to
  strike-slip to extensional.
• Going back to our GoM example, we can see that
  the Miocene deformation comprises a mix of
  contractional, strike-slip, and extensional
  faults and folds located over the salt ridges,
  just as in the analog models (Fig. 186).
  Moreover, during the main upper Miocene
  shortening event, the existing diapirs extruded
  salt tongues laterally, even though surrounding
  minibasins were welded. This could only have
  happened by squeezing of the diapirs again, as
  in the models (Fig. 185).

25
Thrusting
Folding
Squeezing of diapirs
Nappe extrusion
Figure 152. Ways in which distal shortening can
be accommodated (Rowan et al., 2003b).
26
Figure 153. Location of the Perdido foldbelt in
the northwestern GoM just landward of the
basinward pinchout of the Louann salt.
27
Figure 155. Plan-view images of the same model
showing the progression of folding basinward as
the slope progrades (Rowan et al., 2003b model
by B. Vendeville). The red line is the shelf
margin.
28
Base Louann salt
Figure 156. Cascading detachment folds in the
Perdido foldbelt, northwestern GoM (Hall et al.,
2001).
29
Figure 157. Map of the Perdido foldbelt showing a
regular spacing of elongate anticlines (Rowan et
al., 2000c).
30
Figure 158. Line from the Espirito Santo basin in
Brazil showing small-wavelength early folds
overprinted by larger-wavelength later folds
(courtesy of C. Fiduk and CGG).
31
Figure 159. Map of part of the Mississippi Fan
foldbelt showing a more variable geometry with
shorter fold lengths, oblique fold trends, and
diapirs along some of the folds (Rowan and Peel,
2003 map by F. Peel).
32
Figure 160. Section A (see Fig. 159) with
symmetric detachment fold early growth geometry
(3) is symmetrical, whereas late fold geometry is
asymmetric, with strata truncated on the backlimb
(1) onlapping the same unconformity on the
forelimb (2) (Rowan and Peel, 2003).
33
Figure 161. Section C (see Fig. 159) with
asymmetric fold geometry superposed on older,
small-wavelength deformation (4) (Rowan and Peel,
2003).
34
Figure 162. Section E (see Fig. 159) with complex
fold/fault geometry (Rowan and Peel, 2003).
35
Figure 164. Map of the Mississippi Fan foldbelt
showing older, Cretaceous structures (grey lines)
with a different trend in the eastern part of the
area (Rowan et al., 2000c).
36
Figure 165. Series of seismic profiles across the
fold structure that includes the Green Knoll
diapir (Rowan et al., 2001a). Note that Green
Knoll is interpreted as a squeezed diapir with a
weld that has been reactivated as a reverse
fault.
37
Figure 170. Model for the development of the
high-angle reverse fault by squeezing of a
preexisting diapir and reactivation of the weld
during shortening (modified from Rowan et al.,
2000a) a basement step can help but is not
necessary.
38
Figure 171. Simplified model that shows
shortening being accommodated by a combination of
folding and the squeezing of preexisting diapirs
salt extruded from the squeezed stocks
amalgamates to form an allochthonous canopy which
can then serve as a shallower detachment for
gravitational failure and the development of
linked systems of extension and contraction
(Rowan et al., 2001a).
39
Figure 174. Squashed seismic profile and cross
section on the Angolan slope showing inflated
salt massif and allochthonous nappe at toe, both
a measure of shortening in response to updip
extension (Marton et al., 2000).
40
Figure 176. Depth section and restoration of a
profile from the Mississippi Fan foldbelt showing
early inflation of the Louann salt in Restoration
1 (Rowan et al., 2003b interpretation and
restoration by F. Peel).
41
Figure 178. Interpreted extent of the
autochthonous Louann salt and the Mesozoic
allochthonous nappe (courtesy of BHP Billiton and
F. Peel).
42
Figure 179. Summary cartoon of the typical
structural styles in salt-detached foldbelts
note that the thin-salt example is rare because
distal salt is usually inflated (Rowan et al.,
2001a).
43
Figure 184. Experimental models in which (top)
subsidence of elliptical minibasins created a
polygonal pattern of remnant salt ridges with
diapirs (black) at the intersections and
(bottom) the diapirs and salt ridges localized
later contractional strain (courtesy of B.
Vendeville).
model by B. Vendeville
44
Figure 185. Pre-shortening cross section (top)
showing minibasins and diapirs post-shortening
cross section and model (middle and bottom)
showing squeezed diapirs, vertical welds, thrust
faults, and folds (courtesy of B. Vendeville).
model by B. Vendeville
45

• Shale-detached foldbelts
• Before moving on, it is worth making a brief
  comparison between passive margin foldbelts
  detached on salt and those detached on shale. The
  first point is one we have already seen, namely
  that viscous detachments (salt or highly
  overpressured shale) lead to symmetrical
  detachment folds, whereas brittle detachments
  (normal shale) lead to asymmetric,
  thrust-dominated structures (Fig. 188). Two
  end-member cases are illustrated in Figure 189,
  where the symmetrical style in the Mexican Ridges
  foldbelt suggests relatively weak shale and the
  asymmetric, imbricate stack in the Para-Maranhao
  basin suggests relatively strong shale. A
  combined scenario is evident in the Niger Delta,
  where frontal structures are dominated by
  basinward-vergent thrusts but more landward
  structures are more symmetrical (Fig. 190). That
  this is a function of the amount of overpressure
  (and thus strength) can be seen by the velocity
  push-down beneath the symmetrical structures.

• Another difference between salt- and
  shale-detached foldbelts concerns the onset of
  deformation. Translation above salt typically
  begins immediately after salt deposition because
  of the weakness of salt, even at the surface.
  However, shale-detached foldbelts typically have
  a relatively thick prekinematic section (Figs.
  189 and 190) because deformation begins only when
  the shale is buried enough to develop enough
  overpressure. Finally, whereas salt-detached
  foldbelts are located at the toe of the salt and
  the toe of slope, shale-detached foldbelts
  gradually advance basinward as the slope
  progrades and the degree of overpressure
  correspondingly increases (Fig. 191). However,
  out-of-sequence thrusting and folding is common
  as shales in the fold cores periodically build up
  pressure and de-water, so that strength varies
  over time (Fig. 192).

46
Frictional (strong) detachment
Viscous (weak) detachment
Models by B. Vendeville
Figure 188. Experimental models (by B.
Vendeville) of shortening showing asymmetric,
fault-dominated style when detachment is
frictional (shale) and symmetric, fold-dominated
style when detachment is viscous (salt or highly
overpressured shale) (Rowan et al., 2003b).
47
Figure 189. Seismic profiles of shale-detached
foldbelts (top) Mexican Ridges foldbelt with
symmetrical folds and (bottom) Para-Maranhao
basin with asymmetric thrusts (Rowan et al.,
2003b).
48
Figure 190. Seismic profile and cross section
(interpretation by F. Peel) in the Niger Delta
showing asymmetric frontal structures and
symmetric landward structures velocity pushdown
beneath the symmetric folds indicates a high
degree of overpressure (Rowan et al., 2003b).
49

• ALLOCHTHONOUS SALT
• Allochthonous salt is defined as sheetlike salt
  bodies emplaced at stratigraphic levels above the
  autochthonous (evaporative) salt layer (Jackson
  and Talbot, 1991). It is best developed and
  understood in the northern Gulf of Mexico (Fig.
  193), but has also been reported from other salt
  basins such as the Great Kavir of Iran, the
  Yemeni Red Sea, the Precaspian Basin of
  Kazhakstan, offshore west Africa, offshore
  Brazil, and offshore Nova Scotia.
•  
• Allochthonous salt emplacement
• Allochthonous salt sheets were originally thought
  to be emplaced as intrusions at the level of
  neutral buoyancy. But the profile geometry of a
  number of young salt sheets shows that, although
  overlying strata are generally parallel to the
  top salt, underlying strata are truncated by the
  base salt (Fig. 194). Restoration of an intrusion
  model for the Mica salt body in northeastern
  Mississippi Canyon results in unrealistic
  geometries (Fig. 195) in which a diapir in the
  core of an eroded fold extruded salt along, and
  only as far as, the unconformity, while pushing
  underlying beds back down. The preferred model is
  one in which a salt body at the sea floor forms a
  bathymetric high that gradually expands radially
  (Fig. 196) in other words, this is just a
  passive diapir that is growing more laterally
  than vertically. The angle of the salt-sediment
  interface is primarily a function of the ratio of
  salt flow rates to sedimentation rates (Fig.
  197) when the two are balanced,

• the diapir grows vertically when sedimentation
  rate is relatively fast or salt flow rate is
  relatively slow, the diapir narrows upward and
  when the ratio between salt flow and
  sedimentation rates is high, the salt cannot just
  stick up into the water column, so it spreads
  laterally. Thus, a better restoration of the Mica
  salt body shows the progressive spreading of the
  salt at the surface (Fig. 198).
• The base of allochthonous salt often has a
  ramp-flat geometry (e.g., top of Fig. 195). The
  ramps form concentric rings (Fig. 199),
  demonstrating the radial spreading of the salt
  (note also the radial faults). As expected, ramps
  reflect periods of higher sedimentation rates
  (lowstands) and flats reflect slower
  sedimentation (Yeilding et al., 1995). Because
  flats form during highstands, the tip of a
  subhorizontal salt sheet commonly occurs at the
  level of a condensed section/sequence boundary in
  the adjacent minibasin. A condensed section will
  often have high amplitudes, which can mislead
  people into interpreting it as a weld extending
  away from the salt body.

50

• The model in Figure 196 and the restorations in
  Figure 198 suggest that salt is growing right at
  the seafloor as a bathymetric high, but is this
  really the case? Clearly, actively growing
  allochthonous salt creates significant
  bathymetric relief (Fig. 200), but only rarely is
  the salt actually exposed by far the majority
  of allochthonous salt bodies have a thin,
  condensed overburden. In fact, the Mica body
  currently has a scarp, suggesting that it is
  still growing beneath a thin overburden, rather
  than inactive as suggested in Figure 198. An
  examination of the various allochthonous bodies
  in Figure 201 shows that some of the salt bodies
  are currently active, as indicated by a
  bathymetric scarp others that are buried often
  show sediment thickness patterns representing a
  paleo-scarp. Using these data, overburden
  thicknesses above active allochthonous salt range
  up to about 300 meters.
• The overburden typically consists of condensed
  muds and marls deposited on top of the
  bathymetric high as long as the diapir was at the
  surface. But as the salt spreads, the surface
  area of the top salt increases, so that older
  condensed section gets progressively thinned or
  extended (Fig. 202). In addition, the overburden
  rafts along so that it gets emplaced over
  equivalent subsalt strata that are thicker and
  more sand prone. Similarly, suprasalt deformation
  after the salt is buried may further disrupt the
  condensed section, for example by extensional
  faulting (Fig. 203). The net result is
  that any given well can

• encounter different relationships between the
  ages of supra- and subsalt strata, from a normal
  section with an interval of salt, to missing
  section, to repeat section (Fig. 203).
• Some, but by no means all, wells have encountered
  an anomalous section beneath allochthonous salt.
  This so-called gumbo zone is characterized by
  older fauna, overturned strata, and local repeats
  of section. It has been explained as a basal
  shear zone formed by structural disruption as the
  salt is emplaced (Fig. 203). However, even when
  this zone is highly pressured, it is still
  stronger than the salt, and the bulk, if not all,
  of the relative shear takes place within the
  salt. An better interpretation is that inflation
  of a salt body leads to failure of the overburden
  on the bathymetric scarp, with the slumped
  material subsequently being overridden by the
  advancing salt (Fig. 204). This is directly
  analogous to the general model for the growth of
  passive diapirs presented earlier, in which the
  salt body grows by cyclical periods of burial,
  inflation, failure, and salt break-through. An
  example of slumps at the margin of a salt-related
  bathymetric high is seen at the Mica salt body
  (Fig. 205). Some salt bodies have so-called
  Christmas-tree stuctures, which are wedges of
  opaque reflectivity at the edges of diapirs that
  correlate with condensed sections in the adjacent
  basins (Fig. 206). These wedges may be local
  extrusions of salt at times of slow sedimentation
  (salt wings) or they may be slumps of the
  overburden and even salt occurring at the same
  time because bathymetric relief builds up during
  highstands.

51

• Although small-scale ramps and flats are usually
  explained by variations in sedimentation rate,
  there are more diverse explanations for the
  larger-scale extrusion of allochthonous salt.
  First, a significant decrease in sedimentation
  rates, for example when there is a major avulsion
  of the sediment source, should result in lateral
  flow if the diapirs are still being fed from a
  source layer being loaded by existing minibasins.
  Second, a significant increase in sedimentation
  rates and the associated load will drive more
  rapid salt flow and thus lateral extrusion. Just
  such a regional correlation between major updip
  depocenters and the timing of canopy formation
  was proposed for the northern Gulf of Mexico
  (Fig. 207). But it is not just vertical loading
  that increases when there are major depocenters.
  These depositional events also cause gravity
  spreading, which in turn will squeeze existing
  diapirs and drive salt extrusion. The link
  between shortening and salt extrusion is very
  clear in Figure 208 the salt was extruded during
  the growth interval on the contractional fold
  and, more importantly, extrusion stopped at
  exactly the time that shortening stopped. Thus,
  the regional link between large sediment loads
  and canopy formation (Fig. 207) may actually
  reflect increased gravity spreading, i.e., the
  increased lateral load on diapirs may be more
  important than the increased vertical load on the
  adjacent minibasins.

52
Figure 193. Early regional cross section across
the Gulf Basin showing the large extent of
allochthonous salt (Worrall and Snelson, 1989).
53
Figure 194. Near-surface allochthonous salt
bodies drawn at the same scale (Schultz-Ela and
Jackson, 1996). Suprasalt strata are conformable,
whereas subsalt strata are truncated. Active
bodies have bathymetric scarps over one or both
margins and buried bodies have thickness patterns
representing paleo-scarps.
54
Figure 195. Depth section of the Mica salt body
and a restoration assuming that salt was intruded
into the stratigraphy (Fletcher et al., 1995).
The restoration would require that salt was
preferentially intruded along a major angular
unconformity while pushing older strata down, a
highly unlikely scenario.
55
Figure 196. Model for the extrusion of a salt
glacier spreading from a bathymetric high formed
over the feeder diapir (Fletcher et al., 1995).
This is simply a passive diapir that is growing
more laterally than vertically.
56
Figure 198. Better restoration of the Mica salt
body, suggesting that salt was gradually extruded
as a salt glacier at the sea floor (Fletcher et
al., 1995).
57
Figure 200. Bathymetric image of a portion of the
deepwater Gulf of Mexico showing the significant
bathymetric relief over allochthonous salt
(Diegel et al., 1995).
58
Figure 201. Near-surface allochthonous salt
bodies. Note that some bodies have bathymetric
scarps and some have paleo-scarps that are now
buried, showing that the maximum overburden
thickness while the salt was active was about
300m (Schultz-Ela and Jackson, 1996).
59
Figure 202. Salt glacier model, this time with
overburden that gets stretched/faulted as the
upper surface area of the salt body increases
(modified from Fletcher et al., 1995).
60
Figure 203. Composite, model cross section of an
allochthonous salt body showing a suprasalt
condensed section, a proposed subsalt shear zone,
and wells that penetrated different stratigraphy
depending on the relationships between suprasalt
and subsalt deformation (Harrison and Patton,
1995).
61
Figure 205. Image of the seabed just above the
Mica salt body showing slumps along the
bathymetric slump (courtesy of BP and C.
Yeilding).
courtesy of C. Yielding and BP
62
Figure 206. Seismic profile with large vertical
exaggeration showing Christmas-tree structure at
the edge of a diapir representing either slumps
or lateral salt extrusion during times of slow
sedimentation (courtesy of BP and C. Yeilding).
63

• Allochthonous canopies
• There are different types of allochthonous salt
  systems made up of three basic components,
  although it should be remembered that these are
  end-member models (Fig. 209). An asymmetric salt
  tongue has a basinward-leaning feeder, usually
  welded, and a subhorizontal tongue that is
  extruded basinward. A more symmetrical
  bulb-shaped salt stock spreads radially from a
  vertical feeder. A salt nappe extrudes basinward
  from the distal depositional edge of the salt.
  Individual tongues and stocks may amalgamate to
  form salt-tongue and salt-stock canopies,
  respectively (Fig. 210). The surface along which
  two individual tongues or stocks merge is called
  a salt suture. Sutures may be vertical and
  located at a stratigraphic high in the base salt
  where salt extruded from two different directions
  met (Fig. 211). Alternatively, sutures may be
  dipping when the two salt bodies had similar or
  overlapping extrusion directions (Fig. 212). In
  the latter case, the suture is often marked by a
  dipping zone of high amplitudes within the merged
  salt body that represents the condensed section
  originally deposited on top of the lower salt and
  now incorporated into the canopy. Although these
  zones are discrete in a young canopy, salt flow
  during subsequent withdrawal and diapirism or
  gravitational failure of the overburden will
  disrupt the condensed section and disperse exotic
  blocks of old strata into the salt. Sutures are
  often difficult to see directly another
  indication is local sharp lows or highs just
  above salt (dimples and pimples Figs. 211
  and 212). Ultimately, the best indicator of
  sutures is clearly-imaged subsalt truncations
  that show the emplacement history of the
  allochthonous salt.

64
Figure 209. Basic components of allochthonous
salt (a) salt tongue, with salt extruded
basinward from a leaning feeder (b) bulb-shaped
salt stock, with salt spreading radially from a
vertical feeder and (c) salt nappe, with salt
gradually ramping up and basinward with no local
feeders.
65
Figure 210. Individual salt tongues or stocks can
amalgamate to form canopies with multiple subsalt
feeders (Jackson and Talbot, 1991).
66
Figure 211. Pre-stack depth-migrated seismic line
showing two sutures in a salt canopy (data
courtesy of Veritas and L. Liro and M. Kadri).
Note the suprasalt contractional structure over
the left, vertical suture.
67
Figure 212. Pre-stack depth-migrated seismic line
showing a dipping suture marked by high-amplitude
reflectors representing the condensed section of
the lower salt body (data courtesy of Veritas and
L. Liro and M. Kadri). Note the dimple above
the top of the suture.
68

• Salt-tongue systems
• Salt tongues form when a basinward-leaning diapir
  extrudes salt laterally. The extrusion is
  asymmetric, so that the feeder is located at the
  landward edge of the tongue (Fig. 213). A nice
  example from the Precaspian basin is shown in
  Figure 214. Work from the northern GoM suggests
  that tongues can evolve in two end-member styles
  stepped counterregional, in which a subhorizontal
  weld connect two counterregional
  (landward-dipping) features or roho, in which
  basinward-dipping listric growth faults sole into
  the subhorizontal weld (Fig. 213).
•  
• Stepped counterregional
• The schematic evolution of a stepped
  counterregional system is illustrated in Figure
  215. A diapir grows up and basinward, possibly
  because it is forming on the slope and thus
  preferentially spreads slightly downdip. At some
  point, it extrudes laterally to form the tongue,
  with the eventual collapse of the feeder
  resulting in a basinward-dipping and thickening
  growth wedge in the hanging wall of the
  counterregional weld. Evacuation of the tongue
  occurs in two stages (1) initial vertical
  subsidence, usually over the landward portion of
  the tongue, with basinward thinning and onlap
  onto the salt and (2) progressive basinward
  rollover as the salt moves up into a
  basinward-leaning diapir, creating a
  basinward-dipping and thickening monoclinal
  growth wedge. In other words, this is identical
  to an expulsion rollover but takes place above
  allochthonous salt instead of the primary salt.

• An example of a mature stepped counterregional
  system is shown in two profiles, one through a
  shallow dome (Fig. 216) and one just several
  miles away where there is a counterregional fault
  instead of the secondary dome (Fig. 217). The
  subhorizontal weld represents the evacuated salt
  tongue-canopy. In map view, these systems
  comprise one or more elliptical salt domes with
  counterregional faults connecting the domes and
  extending laterally away from them (Fig. 218).
  Modern data show that the counterregional faults
  are arcuate, curving into being tangential with
  the landward edges of the domes. The orange/brown
  area on the map denotes the areal extent of the
  subhorizontal weld, i.e., the original
  salt-tongue canopy. The feeders to this
  two-tongue canopy were located where the weld
  dips off to the NW and NE.
• Restoration of a profile across two stepped
  counterregional systems (the basinward one is
  that in Figs. 216-218 and the landward one is the
  Terrebonne Trough) shows how steep diapirs first
  form salt tongues (Fig. 219). Evacuation of each
  tongue creates a growth wedge in the hanging wall
  of what looks like a landward-dipping normal
  fault. However, these are not true faults over
  most of their length, but instead are salt welds
  that accommodated subsidence into the tongues
  (again, just as in expulsion rollover
  structures). Thus, counterregional fault is a
  misleading term because there is little, if any,
  real extension these are evacuation structures.
  The salt evacuated from the plane of the cross
  section moves laterally into the isolated salt
  domes that occur along the counterregional fault
  trend (Fig. 218).

69

• Roho
• The schematic evolution of the roho style of salt
  evacuation is illustrated in Figure 220. Here,
  two salt tongues are shown to amalgamate to form
  a subhorizontal salt-tongue canopy. Evacuation
  proceeds by the development of basinward-dipping
  listric faults that sole into the salt sheet,
  with landward-dipping and thickening growth
  wedges in their hanging walls. Displaced salt
  moves laterally and basinward where it is
  extruded to form diapirs or shallower tongues.
  The deformation is driven by gravity gliding or
  spreading above the allochthonous detachment.
• Dip and strike profiles through a roho system
  from northern Mississippi Canyon are illustrated
  in Figures 221 and 222, respectively. The dip
  profile shows several deep, landward-dipping
  welds representing evacuated feeder diapirs from
  the Louann level. The sheet is subhorizontal over
  much of its length before it ramps up basinward
  the strike line shows that it also ramps up
  slightly at both edges. A family of listric
  growth faults sole in the remnant salt or weld.
  These faults and the associated basins generally
  young in a basinward direction (Fig. 223).

• The map view shows that the roho system is
  elongate in a downdip direction (Fig. 224). It
  has extensional growth faults above the landward
  portion, local contractional toe structures along
  the basinward edge, and strike-slip structures
  along the margins, all indicating basinward
  translation of the overburden. Roho systems are
  effectively large submarine slumps detached on
  salt-tongue canopies. A small, recent equivalent
  is shown in Figure 225, where the overburden
  above a shallow salt tongue has slid basinward.
  The eastern margin of another is shown in a time
  slice and serial profiles (Fig. 226) the
  overall fault trend is parallel to regional dip
  and the structure in places is that of a flower
  structure.
• Restoration of the profile in Figure 221 shows
  how several salt tongues amalgamated to form the
  canopy and how evacuation of the canopy caused
  further basinward extrusion (Fig. 227). Basinward
  translation of the overburden totaled some 25 km
  on a sheet of 50 km length. This translation was
  balanced only partly by contraction at the toe
  (Fig. 224) most of the translation was taken up
  by lateral movement into the salt, extrusion of
  that salt, and contractional uplift and erosion
  of the overburden (Fig. 227).

70

• Comparison
• The evolution of some salt tongues into stepped
  counterregional systems and others into roho
  systems may be explained in part by the loading
  history above the allochthonous salt. But a more
  important factor is the downdip length of the
  salt sheet (F. Diegel, as quoted in Schuster,
  1995). A relatively long sheet will have enough
  gravity head to cause gravity spreading,
  basinward translation of the overburden, and
  roho-style evacuation, whereas the overburden
  above a short sheet will not move laterally and
  thus forms a counterregional system. This is
  supported by a regional map (Fig. 228), where
  isolated tongues evolved as stepped
  counterregional systems and larger sheets (long
  in the dip direction) developed roho systems.
  Published explanations in which stepped
  counterregional systems develop on the shelf and
  roho systems develop on the slope are
  demonstrably wrong.

• This explanation for the distinction between
  stepped counterregional and roho systems can help
  interpretations in areas where the data are
  equivocal. For example, there are many cases in
  the GoM deepwater where two deep welds are
  separated by a shallow dome, and it is unclear
  whether they are connected or not (Fig. 229). If
  shortening is observed over the basinward tongue,
  the two are probably connected because a longer
  detachment is needed to allow basinward
  translation (roho system). In contrast, if
  shortening is confined to the landward tongue or
  the basinward tongue has counterregional-style
  evacuation, they are likely to be separate. In
  this case, the shortening on the landward tongue
  marks the toe of one canopy and the basinward
  tongue is an isolated stepped counterregional
  system. One reason why this is important is that
  the latter scenario provides a direct migration
  pathway from the deep source rocks, whereas the
  former scenario requires that hydrocarbons move
  through a weld in order to charge shallower
  traps.

71
Figure 213. Evolution of an allochthonous salt
tongue into either a roho system or a stepped
counterregional system (Schuster, 1995).
72
Figure 214. Time-migrated seismic line from the
Precaspian Basin showing a subhorizontal salt
tongue fed from a leaning weld. The yellow
horizon marks the transition from strata that
thin onto the salt/weld to strata that thicken
into the salt/weld.
73
Figure 215. Model evolution of a stepped
counterregional system (modified from Schuster,
1995).
74
Figure 216. Time-migrated seismic profile through
a stepped counterregional system on the Louisiana
shelf (Schuster, 1995).
75
Figure 217. Time-migrated seismic profile located
just a few kilometers from that in Figure 216,
showing that the secondary stock has been
replaced by a counterregional fault (Schuster,
1995).
76
Figure 219. Restoration of a cross section with
two stepped counterregional systems (Schuster,
1995). Note that apparent extensional offset of
horizons on either side of the counterregional
faults is actually caused by vertical
subsidence into salt with no lateral translation.
77
Figure 220. Model evolution of a roho system
(modified from Schuster, 1995).
78
Figure 224. Map view of the roho system of
Figures 221 and 222 (Schuster, 1995).
Allochthonous salt is in green and the
allochthonous weld is in orange/brown.
79
Figure 225. Image and interpretation of a
near-seabed horizon over a recently developed
roho system, showing a linked system of
extension, contraction, and strike-slip
deformation as the overburden above a salt tongue
slid basinward (Peel et al., 1995).
80
Figure 227. Restoration of the profile from
Figure 221, showing the emplacement and
amalgamation of salt tongues and the subsequent
evolution of the roho system (Schuster, 1995).
81
??
This is what you see How do you interpret?
possible charge
If you see shortening here, the tongues are more
likely to be connected (part of roho system) with
continuous salt/weld
If you see shortening here or counter- regional
style withdrawal here, the tongues are more
likely to be separated (toe of roho and basinward
stepped counterregional style)
Figure 229. Using observed geometries to help
distinguish between a connected canopy or
isolated canopies.
82

• Salt-stock systems
• Bulb-shaped salt stocks may coalesce to form a
  salt-stock canopy. The growth and subsequent
  collapse of such a canopy is illustrated in
  Figure 230. Diapirs grow vertically while
  expanding radially. If there is enough salt and
  the diapirs are not too far apart, the stocks
  will amalgamate, creating a canopy with
  significant structural relief of the base salt.
  In three dimensions, the base salt looks like an
  irregular egg carton, so that any 2-D profile can
  look quite different from nearby profiles. Note
  that subhorizontal stratal cutoffs at the base
  salt show that the relief is not secondary, i.e.,
  due to deeper deformation. At some point, the
  bathymetrically high diapirs collapse, thereby
  initiating minibasin formation. This collapse may
  be caused by local gravity spreading after the
  source layer is depleted, shutting off the salt
  fountain (as shown in Fig. 230). Alternatively,
  it may be caused by deep-seated extensional
  widening of the feeder (diapir fall) even before
  the source layer is depleted. In any case,
  evacuation of the bulb-shaped stocks is typically
  a two-stage process (1) initial vertical
  subsidence until local touch-down and weld
  formation and (2) rotational collapse, either
  landward in the hanging wall of a ramp fault or
  basinward in a counterregional style. Salt
  displaced from the stocks forms diapirs and
  tongues over the highs in the base salt or at the
  margins of the canopy.

• Four real examples showing different stages of
  salt-stock evacuation are shown in Figure 231,
  and the bathymetric relief over a young canopy
  and isolated stocks is illustrated in Figure 232.
  A completely evacuated stock has dipping
  suprasalt strata separated by a dipping weld from
  horizontal subsalt strata (Fig. 233) in three
  dimensions, this weld is bowl shaped. Restoration
  of this stock shows vertical and then rotational
  subsidence of a minibasin while the evacuated
  salt was extruded into a flanking salt tongue
  with diapir (Fig. 234). A map from the outer
  Louisiana shelf shows three evacuated stocks with
  tabular salt in between where they amalgamated
  (Fig. 235). The map also shows the salt-stock
  canopy bordered to the west by a roho system. The
  roho overburden moved basinward, but the
  salt-stock canopy overburden did not because of
  the structural relief on the base salt (the
  ultimate detachment), so that the connecting,
  NW-dipping weld accommodated sinistral
  strike-slip movement.
• The basinward margin of the southernmost
  evacuated stock in Figure 235 is bounded by a
  landward-dipping fault due to the counterregional
  style of evacuation. Another common style
  commonly seen is where the minibasin rotates
  landward in the hanging wall of a
  basinward-dipping fault (Fig. 236, left). These
  so-called ramp faults are highly arcuate, curving
  around the updip margins of the evacuated stock.

83

• The geometry and evolution presented so far is
  for a true salt-stock canopy, where the canopy
  forms while salt is still moving up from the deep
  source layer (Fig. 230). An alternative that
  produces a similar final geometry is the apparent
  salt-stock canopy (Fig. 237). In this case, the
  source layer is depleted or the diapirs fall
  before amalgamation and canopy formation it is
  salt evacuated from the collapsing stocks that
  eventually merges to form the highs in the base
  salt or equivalent weld. There are two key
  differences between the two models. First, there
  is never as much salt in the system in apparent
  canopies. Second, the age relationships between
  supra- and subsalt strata are very different. In
  true canopies (Fig. 230), the suprasalt strata
  are all younger than the adjacent subsalt strata
  (ignoring a thin condensed section that was
  originally above the diapir but is now at the
  base of the minibasin) in apparent canopies
  (Fig. 237), there is an overlap in age between
  the oldest suprasalt strata and the youngest
  subsalt strata. Both styles exist, and there is
  no way to tell the difference geometrically age
  control is required.

• There is some uncertainty as to what happens to
  the roots of vertical stocks after the salt is
  evacuated into shallower levels and the source
  layer is depleted. Seismic profiles often show
  keels in the base salt (Fig. 238), but is there
  actually a remaining root connecting the canopy
  to the autochthonous layer? Inclined diapirs can
  become completely welded out simply by vertical
  subsidence (e.g., Fig. 220), but this cannot
  happen in vertical diapirs. We have already seen
  that a vertical stock can be squeezed during
  contraction, thereby forming a vertical weld
  (Fig. 239 see also the example of Fig. 168). But
  is it possible to form a vertical weld without
  contraction, i.e., by collapse of surrounding
  strata into the evacuating feeder? Just such an
  origin has been proposed for Upheaval Dome in the
  Paradox Basin of Utah (Fig. 240). But this
  requires concentric normal faults and radial
  reverse faults (Fig. 241), a pattern that has
  not, to my knowledge, been observed on seismic
  data. Moreover, such an evacuation of a root
  relies purely on a density drive for the salt
  flow, because there is no differential pressure
  on the salt once the source layer is welded. It
  is more likely that, in the absence of
  contractional squeezing, vertical roots remain in
  place below evacuating stocks (see Figs. 230 and
  237). Some of the best modern 3-D, pre-stack
  depth-migrated data are now showing just this,
  for example in the case of the Thunder Horse
  diapir (Pfau et al., 2002).

84
Figure 230. Model evolution of a true salt-stock
canopy, in which salt spreads radially from
vertical feeders and amalgamates (modified from
Rowan, 1995). Subsequent collapse of the stocks
leads to replacement of the salt by elliptical
minibasins. Note that deep evacuation structures,
such as turtles, are not shown.
85
Figure 231. Seismic examples showing different
stages in the evolution of salt stocks, from
unevacuated (upper left) to mostly evacuated
(lower right) (Hodgkins and OBrien, 1994).
86
Figure 233. Time-migrated seismic profile showing
an evacuated salt stock, bounded by dipping
welds, now replaced by a young minibasin (Rowan,
1995).
87
Figure 234. Depth section and restoration of the
same feature, showing how the stock was gradually
replaced by the minibasin as salt was extruded
into the adjacent salt tongue and secondary
diapir (Rowan, 1995).
88
Figure 240. Proposed pinch-off of a vertical
diapir at Upheaval Dome in the Paradox Basin
(Jackson et al., 1998).