A new type of gas chimney exhibiting an unconventional linear planform is found. These chimneys are termed “Linear Chimneys”, which have been observed in 3-D seismic data offshore of Angola. Linear Chimneys occur parallel to adjacent faults, often within preferentially oriented tier-bound fault networks of diagenetic origin (also known as anisotropic polygonal faults, PFs), in salt-deformational domains. These anisotropic PFs are parallel to salt-tectonic-related structures, indicating their submission to horizontal stress perturbations generated by the latter. Only in areas with these anisotropic PF arrangements do chimneys and their associated gas-related structures, such as methane-derived authigenic carbonates and pockmarks, have linear planforms. In areas with the classic “isotropic” polygonal fault arrangements, the stress state is isotropic, and gas expulsion structures of the same range of sizes exhibit circular geometry. These events indicate that chimney's linear planform is heavily influenced by stress anisotropy around faults. The initiation of polygonal faulting occurred 40 to 80 m below the present day seafloor and predates Linear Chimney formation. The majority of Linear Chimneys nucleated in the lower part of the PF tier below the impermeable portion of fault planes and a regional impermeable barrier within the PF tier. The existence of polygonal fault-bound traps in the lower part of the PF tier is evidenced by PF cells filled with gas. These PF gas traps restricted the leakage points of overpressured gas-charged fluids along the lower portion of PFs, hence controlling the nucleation sites of chimneys. Gas expulsion along the lower portion of PFs preconfigured the spatial organisation of chimneys. Anisotropic stress conditions surrounding tectonic and anisotropic polygonal faults coupled with the impermeability of PFs determined the directions of long-term gas migration and linear geometries of chimneys. Methane-related carbonates that precipitated above Linear Chimneys inherited the same linear planform geometry, and both structures record the timing of gas leakage and palaeo-stress state; thus, they can be used as a tool to reconstruct orientations of stress in sedimentary successions. This study demonstrates that overpressure hydrocarbon migration via hydrofracturing may be energetically more favourable than migration along pre-existing faults.
Hydrocarbon migration is directly impacted by structures such as faults and
salt diapirs
Recent studies from the upper slope of the Lower Congo Basin have revealed
the existence of a new type of chimney
The role of stresses in controlling the orientations of venting structures,
hydraulic fractures, and redirecting fluid flow has been well documented
In this case study, Linear Chimneys are associated with networks of
tier-bound, small, densely spaced normal faults which have a polygonal
organisation in map view. Polygonal networks of discontinuities affecting
discrete intervals of fine-grained sediment have previously been linked to
diagenetic processes by
Generally, polygonal faults are considered non-tectonic fault systems
arising due to compactional dewatering of very fine-grained sediments during
the early stages of burial in passively subsiding sedimentary basins
Dip map of referential horizon 5.3 Ma showing distributions of fluid expulsion structures across the study area. Subtypes of fluid expulsion features are shown diagrammatically (see legend below main figure). Palaeo-stress ellipses show relative directions and magnitudes of the horizontal principal stresses and are constructed from the planform geometry of the polygonal fault networks (see Sect. 3.2.3 for more information). The blue axis and red axis on stress ellipses indicate the palaeo-orientation of the intermediate and minimum stresses, respectively. The location of the seismic survey is indicated by a red star on the insert map.
Based on seismic observations, the objective of this study is to constrain the relative timing of fluid flow and polygonal faulting, thereby offering a fluid migration model for the affected interval. This model will be used as a platform to discuss the interactions between fluid flow, faults, and local stress states. Particularly, the following questions are addressed: (1) why are chimneys linear in planform and not circular or elliptical as observed elsewhere? (2) Why do they occur specifically along certain parts of PF planes?
The seismic data presented in this study extend across the outer shelf and
upper slope of the Angolan continental margin (Lower Congo Basin) (Fig. 1).
Two 3-D seismic surveys acquired in 2006 on behalf of Total S.A. have been used
for principal investigation (Appendix Fig. A1). The larger of the two surveys
covers an area of 1310 km
The Lower Congo Basin formed during the rifting and break-up of western
Gondwana,
followed by the opening of the central South Atlantic
Geological setting.
A large seaward-dipping listric growth fault rooted in the crest of a NW–SE-trending salt wall (dashed pink line on Fig. 1) divides the study area into a
landward footwall domain and a seaward hanging wall domain
Four, late Tertiary depocentres named Syncline-0, -1, -2, and -3 occur along the
strike of the salt wall, situated in the hanging wall of the large listric
growth fault (Figs. 1, 2a). These synclines developed during late-stage
salt-detached extension in which the NW–SE-trending salt wall collapsed,
forming the large listric growth fault which transects the survey. Syncline-0, -1, and -2 are located adjacent to two salt diapirs (D1 and D2; Fig. 1),
which are rooted in the salt wall at depth. Syncline-0 subsided from the
Early Miocene (ca. 20 Ma) to Messinian
The fluid flow structures are located within the Middle Miocene to Quaternary
strata, which are mainly composed of hemipelagites
The deepest Tier-1 ranges from 70–130 m thick and contains the Late Miocene
units, whilst the shallower Tier-2 contains the Pliocene units and has a
maximum thickness of ca. 250 m. Tier-1 has a thicker pinch-out toward
Diapir-1, while Tier-2 shows a thinner pinch-out where polygonal faults become
undetectable below 60 ms TWT (Fig. 3). These PFs often extend into strata
above (e.g. interval A in Fig. 3a). The strata immediately overlying the PF
intervals cover the relief of the horst and graben structures below and show
constant thicknesses (e.g. interval B–C in Fig. 3a). Pockmarks associated
with circular PF hosts can often be observed at the base of PF tiers
Arbitrary seismic line showing the pinch-out of polygonal fault tiers (defined by black dotted lines) against the SE flank of Diapir-1. Line location is shown in Fig. 1. The polygonal faults disappear beyond the black vertical arrows, progressively towards the pinch-out of the tier at the transitional boundary where the wedge thickness starts to be less than 60 ms TWT. Note that PFs are absent toward the pinch-out of Tier-2, but are present at the same location in Tier-1 below where this tier reaches its maximum thickness (Ho et al., 2013). This may provide additional support for the theory of minimum thickness determining PF growth (Carruthers, 2012). This observation can serve as a reference example for PF growth. Image adapted from Ho et al. (2013).
In this study area, PFs are organised into different patterns in map view,
such as the isotropic polygonal fault pattern gradually reorganising to a
system comprised of longer faults in a certain direction (i.e. referred to as
anisotropic PFs), with shorter faults orthogonally intersecting them. The
shorter faults are the same length as the standard polygonal fault
segments, whilst the longer ones are up to 20 times longer
Preferred fault alignments or “anisotropic fault patterns” within polygonal
fault networks have been observed in this study area. Concentric faults
surround pockmarks
The orientations of the PFs around or above the aforementioned tectonic
structures are not unusual as the fault patterns mantle the expected stress
state of the structures
Direct evidence for the time when PF activity ceased.
Throughout this paper we will show that stress conditions and polygonal
faulting in this area has had a profound impact on the subsequent phases of
fluid flow by defining a number of interim traps. Consequently, it is
important to outline the nomenclature used when referring to different scales
of stresses and specific parts of the fault planes in this
study.
“Regional stress” refers to stress states in the subsurface driven by
primary tectonic forces, which include gravity and the lateral extension
and contraction occurring above the regional salt detachment. “Local stress” refers to stress state at the scale and within close
proximity of individual tectonic structures where the regional stress field
may be locally perturbed. “In situ stress” refers to stress conditions in place at the location of
individual polygonal faults; this is particularly relevant when trying to
understand the stress conditions at sites of incipient hydraulic fracture
developments which lead to the formation of chimneys. “Lower footwall”, when not specified, refers to the lower part of tilted
PF blocks immediately adjacent to the fault, which moved upward, or
referencing the lower part of horsts in this study area. “Lower hanging wall”, when not specified, refers to the lower part of PF
grabens.
Evidence for fluid flow around salt structures is provided by the occurrence
of chimneys, pockmarks/depressions, positive high-amplitude anomalies
(PHAAs), which are acoustically hard (increase in acoustic impedance)
and interpreted as methane-derived authigenic carbonates, and negative high-amplitude anomalies (NHAAs), which are acoustically soft (decrease in acoustic
impedance) and interpreted as free gas
Morphology of Linear Chimneys.
Chimneys have been observed worldwide in seismic data Type-1 Linear Chimneys terminate upwards into linear PHAAs within depressions, which are shallow and flat-bottomed with relief in the range 3–5 ms TWT (Fig. 8).
The acoustic columns defining the chimneys are often associated with velocity pull-up effects. Type-2 Linear Chimneys terminate upwards into columns of linear NHAAs (Fig. 8). The chimney body is also characterised by push-down reflection zones. Type-3 Linear Chimneys terminate upwards into linear PHAAs with depressions and downwards into linear NHAA columns (Fig. 8).
Linear Chimneys of this type are usually not represented by any reflection distortion zone.
The NHAA columns in Type-2 and Type-3 are situated in the lower part of the
PF tier and are capped by the Intra-Pliocene regional barrier (see seismic lines in Figs. 9 and 10).
Group of Linear Chimneys with flame-liked PHAAs at upward
terminations.
Dip map showing linear depressions at the present day seafloor.
The topmost termination of a chimney is easily distinguishable when
associated with pockmarks or PHAAs
In the study area, Linear Chimneys mainly occur within the Pliocene PF Tier-2
Classification scheme showing types of terminations of Linear Chimneys. Left column: symbols which represent the three groups of Linear Chimneys with different upward and downward terminations. Middle column: the seismic images of the chimneys represented by the symbols. The apparent bases of chimneys are marked with crosses. Right column: the 3-D interpretation of the major intersecting positions between the polygonal faults and the Linear Chimneys shown in the adjacent seismic images.
Linear Chimneys of Type-2 and PF cells filled by gas in Syncline-3.
Few types of gas-charged fluid migration features are found within anisotropic PF networks. In the interval of PF Tier-2 in map view, a kilometric-scale PF area is filled by negative high-amplitude patches in Syncline-3 (Fig. 9b), where NHAA lumps are observed to mimic the PF pattern. The whole NHAA area is limited laterally by the extensional fault of Syncline-3 and vertically by the Intra-Pliocene horizon, below which Linear Chimneys of Type-2 are observed (Fig. 9a).
The linear planform of the topmost and lowest termination of chimney
Type-3, and the alignment of the Linear Chimneys parallel to polygonal faults
in withdrawal Syncline-2.
Dip map of key horizon 5.3 Ma (base of PF Tier-2) showing the distribution of Linear Chimneys in Syncline-1. Isolated positive high-amplitude anomalies (PHAAs) at the topmost terminations of Linear Chimneys at horizon 2.5 Ma (interpreted as methane-related carbonates) are superimposed on the dip map. Green lines highlight the locations of Linear Chimneys underlying the carbonate. Red amplitudes are flame-like PHAAs. The blue axis and red axis on stress ellipses indicate the palaeo-orientation of the intermediate and minimum stresses, respectively.
Linear Chimneys intersect fault planes in different positions within PF
Tier-2. A catalogue and statistical analysis comprising counts of each intersection position has been made by examining 209 detected
chimneys (Fig. 12; see also Fig. A3; sourced from
Pie charts showing the percentage of chimneys intersecting or emanating from different parts of fault planes or adjacent fault blocks. The position of the chimney–fault intersections are illustrated with cartoons in each pie segment. See key at bottom left for drawing codes. Image modified from Ho (2013).
Populations (1) and (2) represent 73 % of the total number of chimneys (see right column in Fig. 8 for summary). In the case of population (2), the Linear Chimneys may also intersect the lower part of the PFs, but the seismic resolution and distortion prevents an accurate determination of their position. Smaller populations include chimneys whose body intersects the middle portion of the PF footwall and hanging wall (9 %) and chimneys occurring in the middle of PF blocks (7 %). The remaining 10 % of chimneys intersect at other various positions (Figs. 12, A3). Among the 73 % (Fig. 12), 23 % and 8 % of the chimneys terminate downwards into negative bright spots in the PF footwall or hanging wall, respectively; these subpopulations all belong to Type-3 Linear Chimneys. Consequently, one-third of the chimneys are associated with free gas stored in the lower part of PF blocks, while the rest only have apparent roots in the lower part of the PF tier or deeper.
Although most linear venting structures occur in PF Tier-2, some exceptions occur. For example, a radial network of a leakage system at a kilometre-scale was found along syncline-related extensional faults in a deeper Late to Middle Miocene interval devoid of PFs (for details, see Fig. 13). This complex network is composed of interconnected linear depressions associated with PHAAs that overlie a network of big Linear Chimneys (Fig. 13a–b). These Linear Chimneys are characterised by push-downs (Fig. 13c), most of which have horizontal lengths around or in excess of a kilometre with the longest ones occurring along the strike of extensional faults (Fig. 13a).
The geometrical coincidence of Linear Chimneys and PFs implies a relationship between these structures. To decipher the genetic relationships the following aspects need to be discussed: (1) the relative timing of PFs and Linear Chimney formations, (2) the gas-charged fluid migration pathways to the nucleated location of chimneys, (3) the mechanisms of preferential gas accumulation, and (4) factors that control the linear planform of the chimneys.
Analysing the timing of polygonal fault formation is essential for the discussion
of whether pre-existing PFs affected fluid migration pathways, i.e. chimneys.
The relationship between the timing of PFs and Linear Chimney formation can
be constrained by several lines of evidence. Polygonal fault nucleation is
widely considered to occur during the early stages of fine-grained sediment
compaction
Kilometre-scale Linear Chimneys along the tectonic faults of
Syncline-0.
Within the study area, new evidence supports the general consensus that PFs grew in sediments very close to
the palaeo-seafloor.
Seismically recorded “gas chimneys” are commonly considered to be the result
of hydraulic fracturing of an impermeable interval
The timing of chimney formation is suggested to be recorded by their
associated pockmarks/depressions, and methane-related carbonates, which formed
at a chimney's topmost terminations when hydrocarbon-charged fluid reached
the palaeo-seafloor. Chimneys connected to pockmarks have been suggested to
have formed during catastrophic blow-out events on the seafloor
It could be argued that the chimneys emanating from the lower part of a polygonal fault plane formed by overpressured gas expulsion at the upper tip of proto-PFs, which were still in their developmental stage. This assumption is, however, inconsistent with the fact that many chimneys are modern and currently active as indicated by PHAAs and pockmarks at their topmost terminations on the present day seafloor (Fig. 7), while the fault planes have already fully developed since the end of the Pliocene. The nucleation point of the chimneys must therefore correspond to a level from which the fluid could not migrate further along the fault plane, and hence it forced the gas to open a new migration path, i.e. chimney.
As the nucleation site of Linear Chimneys is directly linked to the site of gas accumulations, we first investigate the stratigraphic location of gas accumulations by tracing the gas migration pathway prior to the accumulations. This is done by analysing the chimney's downward terminations. Type-3 chimneys (31 %) initiated within the PF tier as indicated by negative high-amplitude columns at their downward termination (Fig. 10c), which are interpreted as residual gas accumulation. In contrast, the downward terminations of the major population of chimneys (Type-1) cannot be determined with precision because of signal attenuation downward. However, they still appear to root in the lower part of the tier or its base, suggesting that overpressured gas-charged fluids occurred around the lower boundary of the tier, which most probably leaked and emptied the reservoirs, leaving no or only weak seismic signals. Therefore, Type-3 chimneys are interpreted as an earlier stage of Type-1, before their gas was exhausted.
Now we investigate how gas migrated specifically into the lower part of the PF tier or below. Because PFs root at different depth levels and the presence of bright spots occurs at different strata (within or below the lower fault tier) (see the profiles in Fig. A4), it is suggested that gas below the PF tier migrates via the long roots of PFs into different permeable layers within the tier and forms multi-layered reservoirs (Fig. 9a).
We do not rule out the possibility that gas was already present within the
carrier bed before polygonal faulting. However, the seismic data clearly
show that the timing of tectonic faults and PF overlapped (Fig. A4a). In
many cases tectonic faults (propagation) post-date PF initiation or formation, and it has
been demonstrated that tectonic faults are the main fluid migration paths for
fluid into the shallow interval in the study area
As the exact stratigraphic levels of gas sources and migration pathways to the base of chimneys cannot be identified, based on the region in which chimneys are rooted, we propose the following scenarios when gas migrated upwards from deeper sources: (1) gas was trapped in strata along sealed tectonic faults below the PF tier; (2) gas migrated laterally, reached carrier beds immediately below the PF tier, and was intersected by long PFs, then accumulated there (Figs. 14a, A4); or (3) gas migrated along the lower portion of the PFs to reach permeable layers inside the lower tier (Fig. 14b). These three processes either happened solely or in combination with each other as a series of steps.
Conceptual model for free gas migration into the (multiple)
permeable layer(s) in the lower part of PF Tier-2 from an underlying carrier
bed. Bold black lines denote the segments of PFs interpreted as impermeable.
Cartoon not to scale.
In conclusion, the rooting position of the majority of chimneys suggests that, before the chimney nucleated, gas migrated to and accumulated preferentially in the lower part of or at the base of the PF tier.
As supported by the statistical analysis presented herein, over 54 % of
chimneys stem from the region around the lower PF footwall; therefore, we
infer that over 54 % of the time gas accumulated in the footwall at the base
of chimneys. It is also the same for the 19 % of chimneys that stem from the
lower PF grabens (hanging wall). As a result, 73 % of the total time gas
preferentially accumulated in the lower part of PF blocks, so we investigate
the cause of this phenomenon. We suggest that two hypotheses in combination
account for the mechanism of preferential gas accumulation in the lower PF
footwalls of tilted blocks, horsts, and lower hanging walls/grabens: (a) the
presence of an impermeable regional seal and (b) an impermeable portion of fault
plane). Two other hypotheses together determine the preferential gas
migration to the lower PF footwall: (c) the differential strain in fault
blocks and (d) the stratigraphic position of permeable layers in fault
blocks. Finally, one hypothesis is for the graben hanging wall: (e) the increase
in
local permeability.
It can be argued that sediments in the lower PF tier are more permeable and
lead gas to preferentially accumulate in such a place. This possibility is
disregarded because of the similarity between the lithologies in the upper
and lower part of the PF tier as indicated by Total S.A. internal well
reports, although it is noted that the permeability measurement of the host sediments is
unavailable. The combination of the above elements is suggested to induce
the formation of PF fault-bound traps in the lower part of a PF tier.
Causes of gas retention within the lower part of PF blocks.
A conceptual model for the formation of Linear Chimneys is proposed below.
The majority of Linear Chimneys stem along the surface of the lower PF
footwalls at various positions (Fig. 12), suggesting that gas-charged fluids could
not migrate along the upper portion of PFs while impermeable (at the moment
when chimneys formed). The permeability of small faults in fine-grained
marine sediments varies upon changes in stress and the resultant strain
around faults
Left column: conceptual models illustrating alternative gas
migration pathways through the PF tier. Right column: seismic lines showing
the critical observations. Free gas on seismic profiles is expressed by
negative high-amplitude anomaly (NHAA), while methane-related carbonates are
expressed by positive high-amplitude anomaly (PHAA).
We would like to emphasise that apart from overpressured fluid (gas) creating
new fractures, overpressured gas may also pass through, filling pre-existing
sub-vertical cracks/fractures in the hanging wall bottoms along the main
fault surface
For chimneys originating within the lower part of PF grabens, gas might be compartmentalised in the damaged graben by the impermeable portion of the PF, which was likely extended downward beneath the Intra-Pliocene barrier; therefore, gas was not able to flow into the adjacent horsts (Fig. 14bii). Consequently, hydraulic fractures initiated in the graben centre and propagated upward along the central axis (Fig. 16c).
For chimneys that do not intersect with any fault, i.e. occurring in the middle of
PF fault blocks, the illustrated model by
The linear planform of chimneys and their evident spatial relationship to
anisotropic polygonal faults suggest that gas migration and hydraulic
fracture propagation are controlled by the alignments of anisotropic PFs.
Anisotropic PFs follow the orientations of salt tectonic structures,
indicating that the PFs are heavily influenced by the stress states resulting
from salt activities
In Syncline-0 polygonal faults are absent, yet the kilometric-scale Linear Chimneys are still present (Fig. 13a). Here, Linear Chimneys are parallel to
deep-seated tectonic faults resulting from salt movement, and the horizontal
stresses are not equal as the intermediate principal stress exceeds the
minimum one
In the smaller scale of polygonal faulted blocks, Linear Chimneys and
anisotropic PFs are often aligned, such as in Syncline-2 and -3 (Figs. 10a,
9c, 5). However, in a particular location above the ridge of Syncline-2,
Linear Chimneys are aligned with a pseudo-isotropic (less anisotropic) PF
network enclosed in a zone between two (strong) anisotropic PF patterns; one
is parallel to the edge of Syncline-2 and the other has a ladder-like
pattern in the centre of Syncline-2 (Fig. 10a–b). In this specific location,
although the PF pattern is similar to isotropic polygonal faulted areas
the stress magnitude remains greater because of the tectonic extension
Finally, the lateral propagations of kilometric-scale Linear Chimneys
rarely impeded by faults are oriented roughly parallel to them and the
chimneys can reach much greater lengths (Fig. 13). In contrast, chimneys
within polygonally faulted areas are much shorter horizontally (
In conclusion, the examples above demonstrate that (1) when tectonic faults are solely presented (without PFs) the planform and orientation of chimneys are affected only by the stress field of tectonic faults; (2) while in areas where PFs occur, tectonic stress controls the orientation of anisotropic PFs, and the in situ stress of the PFs controls the orientation of Linear Chimneys.
Conceptual model for the time steps of gas migration into the PF
tier and the formation of the Linear Venting System (sensu Ho et al., 2012a).
Conceptual models for two styles of gas migration into intervals with
or without polygonal faulting.
Linear Chimney formation can be summarised in six steps (Fig. 17).
During the Pliocene, anisotropic PFs formed and developed under the
influence of an anisotropic stress field induced by adjacent (salt) tectonic
structures. Gas-charged fluids migrated vertically from deeper intervals along
tectonic faults and laterally into the permeable beds below or at the base
of the PF tier (Fig. 17a). Gas-charged fluids migrated upwards along the root of PFs, then flowed
into the lower part of the tier and filled the highest permeable layers in
the horst or the fractured apex of grabens where the permeability was higher
than in the undamaged sediment (Fig. 17a–b). The pressure of gas-charged
fluid was not strong enough to allow gas to intrude the upper part of the PF
plane (which is to referred as impermeable). Further upward migration of the
gas-charged fluids within strata was prevented by the Intra-Pliocene
impermeable interval. Overpressure of gas-charged fluids attained the threshold value for
hydraulic fracture propagation but was insufficient to reactivate the fault. Hydraulic fractures (i.e. chimneys) propagated upward from the lower part
of the PF footwall or hanging wall (Fig. 17c) throughout the end of the
Pliocene to the Quaternary. These fractures were affected by the stress field
around the closest fault and developed a linear planform parallel to adjacent
faults (along the direction of the intermediate principal stress). The linear outlet of chimneys on the seafloor was eroded by gas venting,
producing a linear depression in which methane-derived authigenic carbonates
precipitated and are expressed by PHAAs in seismic data (Fig. 17d).
This analysis of Linear Chimneys has revealed information about the
palaeo-activities of buried hydrocarbon systems, especially how gas-charged fluid
interacted with pre-existing geological structures while migrating upward to
the subsurface. Based on the analysis of linear venting structures, we
attempt to reconstruct the hydrocarbon leakage regime in this study area.
Linear venting structures and gas concentrations occur predominantly in the
synclines, indicating they are sites of active fluid flow (Figs. 9b, 10b). The
reason why gas preferentially concentrates within synclines in the Pliocene PF
interval in this study area may be because of coarser-grained sediments trapped
in the syncline depocentres during that period. It is also known that
synclinal faults cut down to deep turbidite channel reservoirs in this study
area
In contrast, where anisotropic PFs are absent, no Linear Chimneys occur. Therefore, gas migrations are likely unaffected by the surrounding stress state because the horizontal principal stresses are too weak or too similar, and instead gas may migrate in random directions until it reaches a permeable bed or mechanically weak zone to break through (Fig. 18b).
To summarise, the direction of fluid leakage in areas of anisotropic PFs can
be predicted by analysing fracture and fault directions
The anisotropic stress attributed to perturbations of the regional stress
field by faults and salt diapirism controls the orientation of PFs, which in
turn impacts gas-charged fluid accumulation, migration, leakage pathways, and
ultimately the geometry of gas leakage conduits and associated expulsion
features at the seafloor. The mechanism of Linear Chimney formation is
summarised as follows.
Fluid expulsion features making the upper termination of chimneys at the
palaeo-seafloor (pockmarks/depressions, and seep carbonates) date chimney
formation from the end of the Pliocene to the present. Polygonal faulting initiated
in the shallow depth range from 50 to 100 ms TWT below the seafloor during
the early Pliocene predates Linear Chimneys. PF blocks form fault-bound gas traps in the lower part of PF tiers. The location of these traps determines the site of gas leakage and hence
the nucleation site for vertical chimneys. Linear Chimneys nucleating along the lower part of polygonal fault planes
document gas-charged fluids that did not migrate along the upper portion of
PF planes, which therefore appear to be impermeable. The linear planform of chimneys is mainly determined by the orientation of
the intermediate principal stress around the closest fault. Overpressured
gas-charged fluids break through the host rock by pushing aside the host rock
towards the direction of minimum principal stress; consequently, Linear Chimneys developed aligned and parallel to the intermediate principal stress
and hence tectonic and/or polygonal fault strike. In isotropic stress fields, under the same spectrum of fluid expulsion
dynamics, the morphologies of chimneys and associated fluid expulsion
features at the seafloor (depressions/pockmarks, seep carbonate bodies)
are circular, while they are linear in anisotropic stress fields surrounding
tectonic faults, salt structures, and in anisotropic PF networks. In situ stress fields of isotropic PFs alone are not sufficient to induce
Linear Chimneys, and anisotropic tectonic stress fields must be involved. In areas experiencing a transition of two stress fields, Linear Chimneys
follow the trend of less anisotropic PFs rather than the nearby tectonic
structures. Therefore, the development of Linear Chimneys is interpreted to
have been predominantly affected by the in situ stress field of anisotropic
PFs (which are dominated by the anisotropic tectonic stress). Linear Chimneys can be used as a tool to reconstruct previous stress directions in
the same way as using preferentially orientated PFs.
This study is based on the data of statistical analysis shown in Fig. 12. Hence, the data are accessible to the public. The entire 3-D seismic surveys are not accessible to the public because they contain sensitive information. In addition, they are property of the government of Angola, Total S.A. and their affiliates.
The areas investigated in previous studies (Ho, 2013; Ho et al., 2012, 2013, 2016) are shown on the dip map of horizon 5.3 Ma. Superposition of the high-resolution survey (pink area) and the regional survey (grey area).
Different patterns of anisotropic PF networks in which Linear Chimneys are found.
Examples of different groups of chimneys intersecting and emanating from different parts of polygonal faults (PFs). The linear vents of Type-1, 2, and 3 are labelled with black circles. The percentages correspond to the numbers of chimneys intersecting fault planes at specific positions (see pie chart in Fig. 12). The planform dimensions of the chimneys are shown on maps below each section. The apparent bases of chimneys are marked with crosses. See description in Sect. 4.1.1.
Linear patches of negative high amplitudes interpreted as gas
accumulation at the base of PF Tier-2.
The supplement related to this article is available online at:
The authors declare that they have no conflict of interest.
We thank Total S.A. for providing data and funding, Total S.A. and its partners for
publication permission, and the Ministry of Science and Technology of Taiwan
for the grant MOST1052914I002069A1. Our work is based on and extended from
chapter 6 of Sutieng Ho's PhD. The scientific work was fully carried out at
Total S.A., France, and completed under their direction. Sutieng Ho thanks Benoit
Paternoster for his supervision in geophysics since 2007. Sutieng Ho thanks
Mads Huuse for reviewing Sutieng Ho's PhD, and Cardiff University for partial PhD funding. We are grateful for
enormous support and valuable advice from Timothy Byrne, David Hutchings, Quentin Vannelle, Ludvig Löwemark, and
Char-Shine Liu. Special thanks go to our English editor,
Sebastian Czarnota Konrad, for his English language guidance and proofreading services in the last few years.
We thank Andreia Plaza-Faverola and two reviewers for their helpful comments. This work had
previously been submitted to