Abstract
Buried circular collapse structures above a tabular evaporitic body are recorded by recently acquired 3D seismic data on the Levant Basin and continental margin, offshore Israel (Eastern Mediterranean). The structures formed during the Pliocene as buried Messinian (late Miocene) evaporites underwent extensive dissolution in a submarine, deep-water setting. Three-dimensional seismic analysis is used to describe the detailed morphology of the structures and the associated overburden, allowing the reconstruction of their origin and development. It is proposed that evaporite dissolution led to the collapse of the weakly lithified overburden, and this deformed with a series of concentric extensional faults. From the structural analysis of the overburden, the estimated maximum duration of the dissolution event is 0.75–1 Ma. The mechanism proposed for the creation of the circular collapse structures is subjacent dissolution of the more soluble evaporites in the Messinian evaporites, as a result of focused vertical fluid flow at the base of the evaporitic series. Rapid release of overpressured fluids, as, for example, during an earthquake, is thought to have initiated the focused fluid flow, which impinged on the evaporitic seal to the point where dissolution occurred, creating the localized circular collapse structures in the overburden.
Evaporite dissolution is considered an important process in many evaporite-bearing basins worldwide (see, e.g. Warren 1999). Examples of evaporite dissolution have been described in the North Sea (Lohmann 1972; Jenyon 1983; Cartwright et al. 2001), Western Canada (Anderson & Knapp 1993), Gulf of Mexico (Rezak et al. 1985; Hossack 1995), US Permian Basin (Anderson & Kirkland 1980) and West Africa (Hudec & Jackson 2002). Evidence of dissolution is usually represented by discordant geometrical relationships of strata, hiatuses or the presence of residue breccias (Warren 1999). In this paper, we provide clear morphological evidence for the occurrence of evaporite dissolution in the Levant Basin of the Eastern Mediterranean. Three-dimensional mapping of the late Miocene–Pliocene stratigraphic interval has revealed the presence of a series of kilometre-scale circular depressions. We interpret these structures as having formed during the Pliocene in response to the dissolution of buried Messinian evaporites within a fully submarine setting. The circular structures developed in a non-halokinetic setting, producing collapse and/or subsidence of the overlying deep-water sediments.
Circular depressions linked to evaporite dissolution have been widely documented in the literature (Sugiura & Kitcho 1981; Davies 1983; Kastens & Spiess 1984; Clark et al. 1999). However, the lack of resolution in subsurface imaging has generally hampered their detailed morphostructural analysis and therefore the precise mechanism by which they form is not clear. The exceptional imaging of the Messinian evaporite sequence and overburden by recently acquired 3D seismic data in the Levant Basin, provides a unique opportunity to investigate the processes that drive both evaporite dissolution and the deformational response to dissolution in this area.
The primary aim of this paper is to describe the morphology of evaporite dissolution structures in the study area and their associated structural setting. The structures are bounded by a series of extensional ring faults that exhibit similarities to salt withdrawal basins (Ge & Jackson 1998; Maione 2001), caldera collapse features (Branney 1995; Cole et al. 2005) and impact craters (Stewart & Allen 2002). The ring faults observed in the study area are extensional and exhibit a domino geometry, dipping towards the centre of the structure. The detailed analysis of the deformation of the overburden is used in this paper as a tool to assess the timing of the dissolution process (i.e. its initiation, rate and termination).
The second aim of this study is to develop a model for the evolution of the localized dissolution structures, explaining how and when they formed, and what controls their distribution in the Levant Basin. Based on the detailed 3D seismic interpretation, we propose a mechanism for dissolution involving focused fluid expulsion from the thick basin fill succession beneath the evaporite sequence. This mechanism of subjacent dissolution may have occurred in other evaporite-bearing continental margins worldwide.
Regional framework
The Levant Basin and its continental margin are situated in the easternmost Mediterranean Sea, at the zone of interaction of the Anatolian, African and Arabian plates (Fig. 1; Vidal et al. 2000). This passive continental margin was formed as a result of rifting of the Tethys Ocean during the Late Triassic to Early Jurassic (Garfunkel & Derin 1984). A carbonate platform setting dominated in the area during most of the Jurassic and Cretaceous (Bein & Gvirtzman 1977; Druckman et al. 1995).
(a) Schematic map of the Eastern Mediterranean, at the zone of interaction among the Anatolian, African and Arabian Plates. The main tectonic lineaments and the location of the Nile delta are indicated. The study area is shown by the dark grey box (3D and 2D seismic data). The contour lines indicate the depth in metres of the Mediterranean sea floor. The position of the present-day shelfbreak is approximately indicated by the 200 m contour line. Modified from Tibor et al. (1992), Robertson (1998) and Vidal et al. (2000). (b) Schematic regional cross-section through the Eastern Mediterranean basin (location shown in (a); modified from Garfunkel 1998). C, Cretaceous; J, Jurassic.
Commencing in the Late Cretaceous, the formation of the Syrian Arc fold belt created a series of NE–SW-oriented compressional structures along the Levant continental margin (Fig. 1; Eyal 1996; Buchbinder & Zilberman 1997; Garfunkel 1998). A system of submarine canyons developed on the slope during the Oligocene (Druckman et al. 1995). The most prominent of these are the Afiq (or Gaza–Beer Sheva), the El Arish and the Ashdod Canyons (Druckman et al. 1995; Buchbinder & Zilberman 1997). During the Miocene the shelf was intermittently emergent, and the submarine canyons were extended to the shelf area through headward erosion (Buchbinder & Zilberman 1997).
At the end of the Miocene, the connection between the Mediterranean Sea and the Atlantic Ocean became restricted, causing the onset of the Messinian Salinity Crisis (Hsü et al. 1978). The resulting sea-level fall led to a major erosional phase on the Mediterranean margins and to the deposition of a thick series of evaporites in the entire basin (Hsü et al. 1978). The Levant Basin records the lateral transition from the erosional unconformity to the evaporitic series (named the Mavqi'im Formation in Israel; e.g. Cohen 1993) which is c. 2 km thick in the deepest part of the basin. During the Pliocene and Pleistocene, the re-establishment of normal marine conditions resulted in the deposition of a thick wedge of siliciclastic sediments in the Levant Basin and continental margin (Yafo Formation).
Dataset and methods
The main dataset used in this study consists of c. 6200 km2 of 3D seismic data acquired in 2000 by BG-Group and its joint venture partners (Fig. 2). The whole coverage is represented by the Med Ashdod, Levant A and Gal C seismic surveys (Fig. 2). The Levant A survey is the focus of this study, whereas the Med Ashdod and Gal C survey are used only for correlation and general analysis of the evaporitic depositional system. The survey was acquired with an in-line trace interval of 6.25 m and a cross-line spacing of 25 m. Final processing yielded a time-migrated 12.5 by 12.5 m grid (i.e. 6400 bin cells km2).
Map showing the details of the seismic surveys used in this study, together with the location of selected exploration wells. The 3D seismic surveys (Levant A, Gal C and Med Ashdod) are outlined by the dashed rectangles. Black dotted lines indicate the 2D seismic lines. The location of the seismic sections presented in Figure 4 is also indicated.
The seismic data are near zero phase with SEG normal polarity (i.e. an increase in impedance is represented by a positive amplitude). The 3D seismic dataset was migrated with a single-pass 3D post-stack time migration. The main focus of this study spans the 2.5–3.5 s two-way travel time (TWT) interval of the seismic data. The dominant frequency content of the data varies with depth, but is c. 50 Hz in the interval of interest. The vertical and lateral resolution for this interval is estimated to be respectively 10 m and 40 m, using an average velocity value of 2000 m s−1, as derived from velocity checkshot data in the Gaza Marine-1 well (GM1 in Fig. 2; Frey-Martinez et al. 2005b).
A set of 2D seismic lines and wells complete the dataset used in this study. Approximately 6000 km of 2D multichannel seismic profiles with a grid spacing of c. 10 km × 10 km (Fig. 2) were interpreted for mapping the regional distribution of the depositional units. Wireline logs and unpublished commercial stratigraphic reports, mainly based on cutting analyses, were available from nearby exploration wells (Fig. 2). The well data were used for stratigraphic and lithological analysis, for correlation of the depositional units and time to depth conversion.
The analysis of evaporite dissolution was undertaken focusing on the timing of deformation of the overburden to the circular structures. A series of parameters were measured to analyse the variation of the vertical relief (ΔZ) and the expansion index (EI) of the structure with depth (Fig. 3a). The EI is a parameter used to visualize the variation in thickness above synsedimentary structures, initially developed for growth faults (Thorsen 1963). The same concept is here applied to circular structures that collapse or subside incrementally, and display synkinematic deposition of the overburden (Fig. 3b and c).
(a) Idealized circular collapse depression showing the parameters measured for quantitative analysis of the dissolution structures studied in this paper (modified from Branney 1995). The EI was defined by Thorsen (1963) as the ratio between the thickness of deposits downthrown (b, directly above the collapse structure) and the thickness upthrown (a, i.e. undisturbed sediments laterally bounding the collapse structure) measured on successive discrete stratigraphic intervals. The vertical relief (ΔZ) is measured on a selected stratigraphic horizon, as the difference in elevation between the centre and the rim of the collapse structure. (b) Schematic representation of the expected geometry resulting from pre-sedimentary growth v. synsedimentary growth of the circular collapse structure. (Note the difference between concentric parallel onlap in the first case, and concentric onlap coupled with divergent strata configuration and thickness variation in the second case.)
Seismic-stratigraphic framework
The stratigraphic focus of this study extends from the Upper Cretaceous to Recent. The seismic-stratigraphic context of this study interval is described with reference to two representative seismic sections (parallel and perpendicular to the progradation direction of the Levant continental margin, Fig. 4) and summarized in a stratigraphic chart (Fig. 5). The 3D and 2D seismic data were calibrated with the well data, allowing the definition of three main units for description of the local seismic-stratigraphic context: Unit 1, Unit 2 and Unit 3 (see Figs 4 and 5).
(a) Composite seismic section across the Levant Basin and continental margin (see Fig. 2 for location). The three seismic-stratigraphic units defined in this study (Units 1–3) are shown, together with the main interpreted horizons. Unit 2, the focus of this study, is represented by a thick wedge of evaporites pinching out towards the Levant continental margin. Marginal extensional faults within Unit 3, detaching at Unit 2, are marked by dashed lines. The crossover point of (b) is indicated at the top of the section. YSM, Yafo Sand Member. B.S., Base Senonian horizon; L.E., Late Eocene horizon; M, Horizon M; N, Horizon N; B.P., Base Pleistocene horizon. (b) Seismic section along the direction of the Levant margin. The presence of the Oligo-Miocene Afiq Canyon, deeply incising within Unit 1, should be noted. In Unit 3, interpretation of slump deposits is after Frey-Martinez et al. (2005a). Localized downwarping of seismic reflections is observed within Unit 3 and highlighted by the black arrows. The crossover point of (a) is indicated at the top of the section. CS-1, circular structure CS-1; YSM, Yafo Sand Member.
Stratigraphic chart showing the main formations observed in the study area (after Garfunkel & Almagor 1987; Druckman et al. 1995), their age and the correlation with the seismic-stratigraphic units described in this paper. The lithological data are derived from unpublished stratigraphic well reports. Fm, Formation; YSM, Yafo Sand Member.
Unit 1
Unit 1 is bounded at the base by the Base Senonian horizon and at its top by Horizon N (Fig. 4a and b). This seismic-stratigraphic unit includes sediments deposited on the Levant continental margin from the Late Cretaceous to the late Miocene. They are composed of limestones in the marginal lower part of the unit (En Zetim, Taqiye and Zora Formations, Fig. 5), and of dominantly marls and shales in its upper part (Bet Guvrin and Ziqim Formations, Fig. 5).
The most remarkable feature of Unit 1 is represented by a system of submarine canyons (Afiq, El Arish and Ashdod canyons) incising the shelf and continental margin in a SE–NW direction (Fig. 4b; Druckman et al. 1995; Buchbinder & Zilberman 1997). The Afiq submarine canyon is the most prominent of these features (Fig. 4b). Its base is defined on seismic section by truncation of the underlying horizons (Fig. 4b). The canyon fill lacks detailed stratigraphic control, as none of the wells penetrate its maximum thickness in the canyon axis. However, some observations can be made based on the stratigraphic relationship with the wells located on the flanks of the canyon, integrated with previous work undertaken on the same canyon in the onshore region (Druckman et al. 1995). These observations suggest that the canyon fill sediments date back to the Oligocene, and unconformably overlie Lower Cretaceous to Upper Eocene deposits (Druckman et al. 1995). The lithology of the infill is represented by hemipelagic marls intercalated with debrites and fine- to medium-grained sandstone layers (Druckman et al. 1995).
Unit 2
This seismic-stratigraphic unit is bounded at the base by Horizon N and at the top by Horizon M. These horizons are regional seismic reflections, which represent the base (Horizon N) and the top (Horizon M) of the Messinian evaporites across the entire Mediterranean Basin (Hsü et al. 1973; Garfunkel & Almagor 1987). Horizon N is represented in the study area by a strong negative seismic reflection (black event on the seismic sections), which exhibits a remarkable continuity (Fig. 4). This horizon is mainly conformable with the underlying reflections, showing only localized truncation in the marginal and basinal part of the study area. Horizon M is a strong positive seismic reflection, highly continuous across the study area (Fig. 4a and b). The internal geometry of Unit 2 is characterized by the alternation of high- and low-amplitude seismic reflections with a transparent seismic facies (Fig. 4a). The thickness of Unit 2 varies from >1700 m towards the Levant Basin (Fig. 4a) to a few metres near the Levant continental margin, where eventually Horizons N and M merge (Fig. 4a and b).
Well calibration of Unit 2
The stratigraphy of Unit 2 is known only from stratigraphic reports that are mainly based on cutting analyses and petrophysical interpretation. However, there are only a limited number of full penetrations of the Messinian evaporites within the offshore Levant, with most of the wells terminating either within the upper part of the unit or above it. The upper part of the Messinian evaporites is better calibrated, as where it has been cored (Fig. 6) is found to consist of a layered evaporitic series that can be correlated to the Mavqiim Formation (Cohen 1993; Druckman et al. 1995; Buchbinder & Zilberman 1997). This part of Unit 2 in wells Gaza Marine-1, Gaza Marine-2, Or-1 and Or South-1 is composed of a thin layer of cryptocrystalline anhydrite passing upward to interbedded anhydrite and claystone, weakly calcareous, with traces of pyrite and chert (unpublished well reports). The lithologies within the near-basal part of the Messinian evaporites are inferred from well log data, and are composed of a thick halite interval, interbedded with a claystone–anhydrite layer that is a few metres in thickness (Fig. 6; unpublished well report).
Correlation scheme of the Messinian evaporites in the Levant A seismic survey. (a) Seismic section near the Levant margin, crossing the wells (location of wells shown in Fig. 2), and showing the seismic-stratigraphic units and the interpretation of the main seismic horizons. (b) Schematic representation of the lithology and stratigraphic relationship of the Messinian evaporites as described in unpublished well reports. The main seismic horizons have been tied where possible to the lithological and stratigraphic units of the Mavqiim Formation.
Unit 3
Unit 3 is bounded at the base by Horizon M and at the top by the present-day sea bed (Fig. 4). The basal part of Unit 3 is expressed as a 50 m thick package of high-frequency, continuous high-amplitude seismic reflections that are restricted to the areas underlain by the Afiq and El-Arish canyons (Frey-Martinez et al. 2005b). According to the well data, this package consists of Lower Pliocene sandstone interbedded with claystone and marls, deposited in a basin floor turbiditic fan (Yafo Sand Member in Figs 4b and 5; Frey-Martinez et al. 2005b). The El Arish and Afiq Canyon have no seismic expression in the interval post-dating the Yafo Sand member, and at present they are buried features (Druckman et al. 1995; Frey-Martinez et al. 2005b).
Overlying the Yafo Sand Member is a c. 1700 m thick interval of continuous, moderate- to high-amplitude seismic reflections defining a package of Plio-Pleistocene progradational–aggradational clinoformal reflections (Fig. 4a), which comprises the remaining part of the Yafo Formation (Fig. 5; Druckman et al. 1995; Buchbinder & Zilberman 1997; Frey-Martinez et al. 2005b). These deposits are composed of hemipelagic and turbiditic claystones, alternating with sandstones, siltstones and marls. Their depositional environment varies from outer neritic to upper or middle bathyal.
Unit 3 is extensively affected by downwarping of reflections and displacement by steeply dipping faults, detached at the level of the Messinian evaporites (Fig. 4a and b). These structures have been previously interpreted as detached normal faults related to gravitational gliding of Unit 3 above the Messinian evaporites and dated as late Pliocene or Pleistocene (Almagor 1984; Garfunkel & Almagor 1987; Tibor & Ben-Avraham 1992).
Description of the circular structures
General features
Initial reconnaissance mapping of the study area using only the widely spaced grid of 2D seismic profiles showed the top of the Messinian evaporites to exhibit a fairly featureless structure, with gentle dips and subtle topographic relief. However, more precise mapping of the top of the Messinian evaporites using the 12.5 m line spacing of the 3D seismic data has led to the recognition of a series of localized depressions exhibiting a well-defined circular planform geometry. These structures (named progressively CS-1 to CS-10) are visible at the top of the Messinian evaporites and are evident on the time–structure map of Horizon M in the Levant A seismic survey (Fig. 7). They are situated a few hundred metres to a few kilometres basinward of the present day pinch-out of the Messinian evaporites (Fig. 7). The map shows the occurrence of the largest of the circular structures (CS-1) and a number of smaller structures (CS-2 to CS-10; Fig. 7). Their diameter varies from a few hundreds of metres (CS-2 to CS-9 and CS-10) to 1–2 km (CS-1 and CS-7). The occurrence of the circular structures is limited to the area outlined by the flanks of the Oligo-Miocene El Arish and Afiq submarine canyons (Fig. 7). One circular structure (CS-8) is located in a more basinward position, at c. 10 km from the eastern margin of the Messinian evaporites (Fig. 7). This structure does not have a well-defined morphology on the time–structure map of Horizon M, but it is clearly circular on time–structure maps of the overlying horizons.
Time–structure map of Horizon M in the Levant A seismic survey. The zoom shows the distribution of the circular structures analysed, named progressively CS-1 to CS-10. The presence of the linear depression and the extensional faults near the pinch-out of the Messinian evaporites should be noted. The main deep structures of the study area are represented by the anticlines (axes highlighted on the map) related to the Syrian Arc foldbelt system (Neev & Ben-Avraham 1977; Tibor & Ben-Avraham 1992). The location of the seismic sections displayed in Figures 8, 9 and 13 is indicated.
In cross-section, the circular structures (Figs 8 and 9) are bowl-shaped depressions with a gently concave-upwards geometry. The stratal configuration of the deformed units is generally parallel stratified. However, some interbedded onlapping reflection-bound units are observed above the depressions (Fig. 8b and c). The depressions are rooted at the upper part of the Messinian evaporites (Figs 8 and 9). The deformed stratigraphic interval lying above the depressions comprises the lower part of the Yafo Formation, from its base to the Upper Pliocene interval (Fig. 8b and c). This stratigraphic interval is defined as the overburden to the circular depressions. The largest of the circular depressions exhibits intense localized faulting of the overburden. Similar faulting may occur above the smaller circular depressions, but if so, is considered to be beneath seismic resolution. The thickness of the Messinian evaporites below the circular structures ranges from 150 m (Fig. 8) to 800 m (Fig. 9).
(a) Seismic section perpendicular to the Levant continental margin, crossing radially the circular structure CS-1 through its centre (location of the section shown in Fig. 6). The section shows the general stratigraphic context of the structure CS-1. The proximity of CS-1 to the extensional faults and the linear depression near the pinch-out of the Messinian evaporites should be noted. The main interpreted seismic horizons are marked with dotted lines. The seismic package from the Base Pleistocene to Horizon Q60 shows consistent thickness variation across the extensional faults, defining the growth phase of these faults. The crossover point of (b) is indicated in the lower part of the section. (b) Seismic section approximately parallel to the Levant continental margin, crossing radially CS-1 through its centre, and (c) interpretation. The seismic horizons interpreted within the overburden of CS-1 are indicated with their progressive number (ME20 within Unit 2, PL20 to PL50, and Base Pleistocene within the Unit 3). Horizon N and the underlying seismic reflections within Unit 1 are downsagged below CS-1. We interpret this as an apparent geometry caused by a seismic ‘push-down’ resulting from the seismic velocity contrast between the Messinian evaporites (Unit 2, average seismic velocity 4 km s) and the marine clastic sediments of Unit 3 (average seismic velocity 2 km s). This seismic section shows the thickening of the stratigraphic package PL20–PL50 and onlap of reflections within the same interval above CS-1. The set of extensional and subvertical faults, steeply dipping toward the centre of CS-1 and deforming its overburden, should be noted. YSM, Yafo Sand Member. (d) Depth cross-section showing vertical to horizontal aspect ratio of CS-1.
(a) Variance time slice (2512 ms) showing the circular appearance of CS-2, CS-9 and CS-10 in plan view. The location of the seismic cross-section of (b) is indicated. (b) Seismic section across CS-2 (see Fig. 7 for location). A minor push-down effect is present at Horizon N beneath CS-2. (c) Variance time slice (2256 ms) showing the circular appearance of CS-8 in plan view. The location of the seismic cross-section of (d) is indicated. (d) Seismic section across CS-8 (see Fig. 7 for location). (e) Seismic section across CS-3, CS-6 and CS-7 (see Fig. 7 for location). (f) Seismic section across CS-4 and CS-5 (see Fig. 7 for location). M, Horizon M; N, Horizon N. (g) Salt outlier located in the northern area of the pinch-out of the Messinian evaporites (see Fig. 7 for location).
Two other types of structures that relate to the deformation of the Messinian evaporites are evident on the map of Horizon M: a linear peripheral depression and marginal extensional faults (Fig. 7). Both the linear depression and the faults are detached at the level of the Messinian evaporites. The linear depression extends almost continuously along the locus of pinch-out of the Messinian evaporites and exhibits a straight to undulating pattern in plan view (Fig. 7). This depression is associated with downwarping of the overlying reflections (Figs 7 and 8). This feature resembles peripheral sinks and rim synclines observed adjacent to salt diapirs; that is, the depression created in the surface as a result of the flowage of the mobile layer into the positive feature (Trusheim 1960). The same linear peripheral depression has been described at the edge of dissolving thick evaporitic beds, as a result of edge-inward meteoric dissolution (Warren 1997). The occurrence of a linear depression similar to that observed in the study area has been recorded in the Eratosthenes Seamount (Major & Ryan 1999) and in the Bannock Structure, Mediterranean Ridge (Von Huene 1997), nearer the study area. These linear depressions are filled with brines that are the result of dissolution of evaporites by fluids travelling up the flank of the structure (Von Huene 1997; Major & Ryan 1999).
The second type of evaporite-related features is represented by a system of extensional faults deforming Unit 3 at the eastern margin of the Messinian evaporites (Figs 7 and 8). They are interpreted as growth faults based on sediment thickness variation from the hanging wall to the footwall of the fault (Fig. 8; Cartwright 1992). These faults have been related to gravitational gliding of the Plio-Pleistocene sediments above the evaporite detachment level (Almagor 1984; Garfunkel & Almagor 1987; Tibor & Ben-Avraham 1992). They are analogous to the peripheral evaporite-related structures described in the North Sea (Jenyon 1986; Coward & Stewart 1995; Huuse 1999). In some parts of the study area, the process of marginal deformation of the Messinian evaporites created salt outliers that are observed landward of the locus of pinch-out of the Messinian evaporites (Fig. 9g).
Circular structure CS-1: detailed 3D seismic interpretation
The most dramatic and best imaged of the circular depressions mapped within the study area is CS-1, as clearly seen on the time–structure map of Horizon M (Fig. 7). Because of the quality of the imaging, this structure is described in detail in the present section. CS-1 is a relatively isolated feature, situated 2 km west of the eastern pinch-out of the Messinian evaporites (Fig. 7).
Morphology
CS-1 exhibits a concave-upward, U-shaped geometry at Horizon M, with a diameter of c. 2 km, as measured on cross-section between its opposite external rims (Fig. 8). The base of this structure is rooted in the upper part of the Messinian evaporites (Fig. 8). The overlying Pliocene reflections display a regular layer-cake stratigraphy of downwarped reflections defining a U-shaped depression (Fig. 8). The morphology of CS-1 is mostly evident on the 3D visualization of the time–structure map of Horizon PL20 (Fig. 10). Half of the circular depression is visible in this display as a sharply defined, nearly half-conical structure (Fig. 10).
Three-dimensional visualization of the time–structure map of Horizon PL20, cutting at the top of a seismic section parallel to the Levant continental margin (see location in Fig. 7), and crossing the circular structure CS-1. Half of CS-1 is visualized on this map, showing the relationship with the underlying stratigraphy. In Unit 2 (Messinian evaporites) the interval between Horizon ME20 and Horizon M appears to thin out toward the flanks of CS-1.
Stratigraphic architecture
Seismic sections crossing CS-1 (Fig. 8) clearly show that this structure is associated with structural deformation of the overburden, which is represented by the horizons overlying CS-1 from Horizon M to PL60 (Fig. 8b and c). The base of the Messinian evaporites (Horizon N) exhibits an apparent sag beneath CS-1 (Figs 8b and 10). This feature is interpreted as a negative seismic velocity anomaly (push-down; see Fig. 4 for explanation). Restored to its original depth, Horizon N appears undeformed beneath CS-1 (Fig. 8d).
Horizon M displays a partly chaotic geometry across CS-1 (Fig. 8a and b). The flanks of the depression are oversteepened, and exhibit a maximum dip of 10–16°. The chaotic and oversteepened character of Horizon M in this area indicates collapse of the upper parts of the Messinian evaporites in the circular area defined by CS-1. This collapse is associated with a vertical negative relief of c. 180 m, defined as the maximum downwards deflection below the regional datum.
Importantly, there is evidence of volume loss within the Messinian evaporites, in the area defined by CS-1, when compared with the original depositional thickness in nearby areas where there are no circular structures or signs of evaporite depletion (Fig. 8a). As a result of the clearly defined conical morphology of the circular depression, it is possible to calculate the discrete loss of volume associated with CS-1 in the upper part of Unit 2 as c. 0.188 km3. There is evidence from the reflection configurations within the Messinian that volume loss was not uniform over the full vertical extent, but was restricted to the upper part of the Messinian evaporites (Figs 8 and 10). As it appears in Figure 8d, the thickness of the seismic package between Horizon N and Me20 is nearly constant whereas the package from Me20 to Horizon M thins considerably across CS-1. This observation argues strongly that this lower interval (Horizon N–Me20) was virtually unaffected by the depletion.
Structural analysis
The overburden to CS-1 displays a sequential deformation pattern that constrains the timing of the development of this structure. The overburden can be divided into three packages based on the geometry of the seismic reflections (Fig. 8b and c). (1) The first package is bounded at the base by Horizon M and at the top by Horizon PL20. This package is composed of parallel-stratified downsagged reflections, with no considerable thickness variation across CS-1 (Fig. 8b and c). (2) The second seismic package is bounded at the base by Horizon PL20 and at the top by Horizon PL50, and it is the key interval for structural analysis of the overburden deformation. This seismic package exhibits a convergent reflection configuration, coupled with localized but significant thickness variation above CS-1 (Fig. 8b and c). Importantly, in the interval PL35–PL45 there is evidence of concentric onlap against the internal rim of the depression (Fig. 8b and c). The geometry of this onlap relationship and the convergent reflection configuration suggest a synsedimentary growth of the structural depression (Cartwright 1992) during deposition of sediments in the PL20–PL50 interval, as opposed to the infill of a pre-existing structure (Fig. 3b and c). (3) The third seismic package is bounded at the base by Horizon PL50 and the top by the Base Pleistocene horizon. This package is composed of parallel reflections, which exhibit no downsag and thickness variation across CS-1 (Fig. 8b and c).
The transition from a parallel configuration in the overburden immediately overlying Horizon M to a convergent reflection configuration above indicates that the discrete and localized volume loss within the evaporites caused the downsag of the overburden (Horizon M–PL20) and led to the subsequent infill of the depression during the progressive development of its structural relief (PL20–PL50 interval).
Faults
The structural deformation of the overburden to CS-1 is associated with a series of high-angle normal faults exhibiting predominantly domino geometry (Fig. 8b and c). The faults dip inwards toward the downsagged part of CS-1, with throws ranging from 10 to 15 m and dips from 60° to nearly subvertical (Fig. 8d). The faulted interval occurs between Horizon M and PL50 (Fig. 8c), with a thickness of c. 500 m (Fig. 8d). The fault system is interpreted to consist of a set of small growth faults, because the faults must have affected the sea bed at the time of their formation, as indicated by the onlap and thickening pattern of the overburden in the PL20–PL50 package (Fig. 8c). The fault pattern in plan view is particularly well seen on the TWT–dip map of horizon PL20 (Fig. 11) where the faults exhibit a striking concentric pattern. Individual faults are c. 500–1000 m long and traverse about 30–160° of arc around the collapse structure. The faults are more numerous within a radius of 1 km from the centre of the structure (Fig. 11a and b). The planform geometry of the faults is comparable with ‘ring faults’ observed above collapse structures (e.g. karst, calderas, withdrawal basins; Branney 1995; Stewart 1999; Maione 2001). The observations made on the set of faults above CS-1 suggest that they are the result of a horizontally radial extensional stress field, induced by the collapse of the circular depression.
(a) TWT–dip attribute map of Horizon PL20, and (b) interpretation, showing the detailed morphology of this surface and the deformation associated with the circular structure CS-1. The pattern of concentric extensional and subvertical faults around CS-1 should be noted. The dotted lines highlight the position of the seismic sections used for the measurements displayed in Figure 12.
Onset and timing
As documented above, the synsedimentary growth of the circular depression CS-1 has been stratigraphically tied to the seismic interval between Horizons PL20 and PL50. The timing of formation of CS-1 has been quantitatively assessed by measuring two key parameters linked to the structural analysis of the overburden: the vertical relief (ΔZ) and expansion index (EI) (Fig. 3a). The values measured have subsequently been interpolated in diagrams where they are plotted against the depth of the point measured, and tied from wells to the chronostratigraphic chart (Fig. 12).
Diagrams showing the vertical relief (ΔZ) and expansion index (EI, after Thorsen 1963) measured and calculated on random seismic sections across CS-1 (see Fig. 10 for location of sections, and Fig. 3a for explanation of the parameters.) The reference horizon for the measurements is taken at the sea bed that represents a horizontal surface above CS-1. ?z (Fig. 11a–c) and EI (Fig. 11d–f) are plotted (x-axis) in the diagrams against the depth of the reference horizon (y-axis, in metres), and against the stratigraphic chart. The grey areas highlight the maximum observed variation of the vertical relief (Fig. 11a–c) and the maximum values of the expansion index (Fig. 11d–f), which are interpreted as related to the time of maximum synsedimentary growth of CS-1.
The ΔZ associated with CS-1 decreases progressively from Horizon M to PL65 (Fig. 12a–c). In the first part of the curve the vertical relief drops moderately, displaying very limited variation from 180 m to 160 m (Horizon M to PL20). Therefore the vertical relief in the Zanclean to the Lower Piacenzian interval is similar to the relief created by the collapse of the top of the Messinian evaporites. The value of the vertical relief drops rapidly from 160 to 50 m in the interval PL20–PL50, and is stratigraphically confined within the Piacenzian. This reduction mainly coincides with the convergent reflection configuration and onlap fill interval and hence the main phase of growth of the structure. The steepness of the curve then increases again, reaching the zero value at Horizon PL65 (directly above the Base Pleistocene).
The EI diagrams (Fig. 12d–f) exhibit a general trend characterized by the main peaks well within the Piacenzian interval (horizons Pl25 to PL50), where it reaches the value 2.1. The EI decreases both up- and down-section from the main peaks. Minor pulses of the EI (e.g. PL40–PL45 in Fig. 12d and e) could be explained by polyphase structuring of CS-1 within this interval.
The interpretation arising from these diagrams constrains the onset of the collapse related to the formation of CS-1 to within the Piacenzian. Assuming an average depositional rate of 20 cm ka−1 (calculated from the wells) and a thickness of 150–200 m for the deformed deposits (highlighted in grey in Fig. 12) the time span of maximum synsedimentary growth of the structure is calculated to have lasted c. 0.75–1 Ma. The accuracy of this estimate could be affected by a number of factors: the chaotic character of Horizon M and the Yafo Sand Member above CS-1 and their consequent contribution to errors in horizon correlation, the nearby presence of growth faults and slumps, differential compaction, and uncertainties in the chronostratigraphic calibration. Of these factors, the most significant is likely to be differential compaction, which has probably occurred above the dissolution structures. The error in thickness measurements arising from this could have led to a slight shift forward in timing of the end of the synsedimentary growth. Incorporating a realistic error owing to differential compaction would place an upper limit of 1 Ma as the maximum duration of the dissolution process.
Horizon PL20 represented the coeval sea bed at the time of the onset of the formation of CS-1, as shown by subsequent onlap of the overlying reflections (Fig. 8b). This indicates that the top of the Messinian evaporites was located at c. 250 m of burial depth (not corrected for differential compaction). Well reports record an upper to middle bathyal depositional environment of the Yafo Formation overlying the Messinian evaporites. Therefore it is concluded that the process responsible for the growth of the CS-1 structure was in a fully submarine environmental context.
Comparison of the timing of the development of CS-1 and other important structures on a more regional scale along the margin shows unequivocally that the collapse of CS-1 predated the evolution of major gravity tectonic structures. Seismic sections perpendicular to the marginal extensional faults (Fig. 8a) show that although downwarping or displacement of the overburden above these structures is associated with thickness variation and onlap patterns similar to that observed in the overburden of CS-1, the timing of the growth interval is much later, within the Pleistocene (Fig. 8a).
Circular structures CS-2 to CS-10: detailed 3D seismic interpretation
On the time–structure map of Horizon M, a series of smaller circular depressions are identified and named CS-2 to CS-10 (Fig. 7). Their diameter ranges between 200 m and 1 km, and they are localized to a small area west of CS-1, close to the northern and southern locus of pinch-out of the Messinian Unit, or in a more basinward position.
The smaller circular structures are rooted within the upper part of the Messinian evaporites, where evidence of volume loss associated with the structures is observed (Fig. 9). However, the reduced size of the depressions precludes any accurate calculation of loss of volume in each case. The circular depressions exhibit the same general geometrical characteristics to those described for CS-1; that is, the concave-upward geometry and the disruption of Horizon M and of the overburden (Fig. 9). The negative structural relief above the structures ranges between 50 and 100 m as measured on Horizon M (Fig. 9a, e and f).
The deformed overburden to the circular structures appears to be more limited stratigraphically than the deformed overburden to CS-1. The vertical extent of the disrupted zone above CS-3 to CS-6 is confined to the Yafo Sand Member (Fig. 9e and f), whereas above CS-2 and CS-8 the disruption reaches the seismic reflections between Horizon PL20 and PL40 (Fig. 9b and d).
The structural deformation of the overburden above the circular depressions shows variable characteristics, which are generally more difficult to analyse than in the case of CS-1. Thickness variation is observed above CS-2 and CS-8 in the PL20–PL40 stratigraphic interval (Fig. 9b and d). In this interval, onlap of seismic reflections is observed over CS-8, above Horizon PL20 (Fig. 9d). The overburden to CS-3, 4, 5 and 7 (Fig. 9e and f) exhibits a chaotic pattern and structural disruption within the Yafo Sand Member, whereas Horizon PL20 appears to be locally upwarped. No faulting is evident above any of these smaller circular structures. It seems likely, however, that owing to their limited size, faults may well be present but they may simply be too small to be imaged individually.
Despite the imaging problems associated with their smaller diameters, the timing of the onset of the smaller circular structures can be assessed by comparison with CS-1 using the same procedure as applied in the case of CS-1. In general, the thickness variation of the overburden above all the circular depressions is confined to the interval between the Yafo Sand Member and Horizon PL40. The onlap observed above CS-8 (Fig. 9d) appears to constrain the time of formation of this circular structure well within the stratigraphic interval PL20–PL40. Therefore the smaller circular structures appear to have formed between the Zanclean and the Piacenzian (Horizon PL40). The growth interval of the smaller circular structures (Zanclean–Piacenzian) appears more extensive than that of CS-1 (Piacenzian). A discrepancy in the style of the structural deformation has also been observed between CS-1 and some of the smaller circular structures (e.g. the vertical sides of CS-3). These differences could be related to variation in degree of lithification of the overburden, causing it to collapse vertically (CS-3) rather than with more gentle downsagging (CS-1). However, it is equally likely that the discrepancies observed in timing of growth and in structural style could be due to imaging problems at the limits of seismic resolution, such that the real extent of deformation of the smaller circular structures is not correctly portrayed in the seismic data. Whatever the explanation for the minor differences, the overall growth period for all the circular structures overlaps to a high degree within the Pliocene, and this suggests a common mechanism for their origin.
Discussion
Observations and identification
A number of contrasting geological processes can produce circular structures, ranging from catastrophic gas expulsion to form pockmark craters to meteorite impacts (Stewart 1999). No general uniformity of process links this diversity of circular geological structures other than that the action of the process is centred at a point (the centre of the circular feature). As discussed by Stewart (1999), the best approach to diagnose the origin of circular structures is to document their size, cross-sectional geometry and distribution with respect to the structural and stratigraphic context. The key characteristics of the circular structures observed in the study area are: (1) the circular geometry in plan view, and negative (concave-upward) relief, with their diameter ranging between 200 m and 2 km; (2) the downwarped or partly chaotic overburden; (3) the depletion of the underlying tabular Messinian evaporites; (4) the concentric fault pattern and thereby dominantly radial extensional stress field centred on the axis of the depression. Given the location of the collapse structures directly above a depleted evaporite unit, the most plausible interpretation for the genesis of these structures is intrinsically linked to the loss of the evaporites from the region directly beneath the collapse structures. It is highly significant that on CS-1, the volume of missing evaporites is equivalent to the volume of the depression as measured at Horizon PL20, implying a direct volumetric coupling between removal of evaporites and downwarping of the overburden.
Two mechanisms linked to deformation of evaporites can potentially generate circular structures: (1) evaporite withdrawal (Jackson & Vendeville 1995); (2) dissolution (Ge & Jackson 1998). Withdrawal and flowage of mobile evaporites (probably halite) involves a constant-volume assumption, and therefore it should be expected that there would be positive salt structures of comparable excess volume above regional to match the depletion below regional (see, e.g. Jackson & Talbot 1986; Davison et al. 2000). The absence of any positive structural elements (such as pillows or small diapirs) close to the depressions makes it difficult to accept salt withdrawal as a viable process for the genesis of the circular structures. Consequently, it is argued here that the only feasible mechanism for the formation of the circular depressions is the collapse of the overburden as a result of the in situ dissolution of the underlying Messinian evaporites.
The amount of fluid necessary to form the largest dissolution structure (CS-1) can be crudely estimated, assuming that dissolution of pure halite (NaCl) took place in fresh-water solution. Using a solubility of NaCl in fresh water of 359 g l−1 (Davies 1989), the volume of water required to dissolve 0.188 km3 of NaCl (i.e. the volume of CS-1) is calculated as 1.15 km3. Using a time-averaged dissolution rate, and based on the timing of 1 Ma for the growth period for CS-1, the minimum rate of evaporite dissolution can be estimated as 188 m3 a−1.
The interpretation of the circular depressions in the study area as dissolution structures is strengthened by the analogy with other described examples of circular collapse structures created by dissolution of evaporites on the sea bed of the Eastern Mediterranean basin (Ross & Uchupi 1973; Kastens & Spiess 1984) and in the North Sea (Lohmann 1972; Jenyon 1983). Their circular pattern has generally been related to their location at the top of underlying salt diapirs (Ross & Uchupi 1973; Kastens & Spiess 1984; Ge & Jackson 1998; Cartwright et al. 2001). Despite the recognition that such structures occur in salt provinces, the mechanism of formation of circular dissolution structures above a tabular evaporitic body is poorly understood, and is therefore further discussed below.
Genesis
The mechanism invoked for the formation of the dissolution structures needs to account for the circular shape of the dissolution structures, their submarine context and the burial depth of the evaporites at the time of their formation. The key to understanding the mechanism lies in the analysis of the geological and hydrological framework of the basin, coupled with the timing of evaporite dissolution.
Mechanism of dissolution
Three main modes of dissolution of buried evaporites have been described in sedimentary basins (Johnson 1997; Warren 1997, 1999; Cartwright et al. 2001), according to the mechanism of migration of undersaturated fluid: (1) lateral dissolution; (2) superjacent dissolution; (3) subjacent dissolution.
Lateral dissolution results from circulation of undersaturated fluids along the edges of the evaporite body. This process could have promoted linear dissolution along the evaporite pinch-out. However, it cannot be invoked for the creation of the dissolution structures described here, because of their localized and circular geometry, and their location at some distance from the edge of the evaporites.
Superjacent dissolution results from circulation of relatively undersaturated fluids above the evaporites, either from an overlying aquifer or from fluids penetrating at the evaporites through structural pathways such as faults (Cartwright et al. 2001). In this area, the unit immediately overlying the evaporites is the Lower Pliocene Yafo Sand Member, which is a sand-dominated body. However, this permeable unit is strictly confined to the thalweg of the Afiq Canyon and sealed laterally, and is thus unlikely to have acted as an extensive aquifer. Moreover, at least one of the dissolution structures is located basinward of the most distal reaches of the Yafo Sand Member (Fig. 7) and could not therefore have been formed by fluids flowing along this potential aquifer. The rest of the Yafo formation overlying the evaporites is dominantly composed of fine-grained siliciclastic sediments. Therefore this low-permeability unit could not have acted as an extensive aquifer and it could not have provided the undersaturated fluids necessary for the dissolution of the Messinian evaporites. It could be argued that the marginal growth faults may have provided the pathways for penetration of undersaturated fluid at the top of the evaporites. However, synsedimentary growth of the faults is recorded only from the Lower–Middle Pleistocene (Fig. 8a), and therefore the timing of activity of the faults clearly post-dates the formation of the circular dissolution structures. Moreover, this mechanism of fault-related downward penetration of undersaturated fluid does not provide a simple explanation for the circular and localized geometry of the dissolution structures.
Subjacent dissolution results from undersaturated fluids migrating along the base of the evaporites. Examples of this type of dissolution have been recorded in the North Sea (Cartwright et al. 2001) and Khorat Plateau in Thailand (Warren 1997; El Tabakh et al. 1998). Evaporite beds are impermeable once buried to depths of few hundred metres, where they act as an aquiclude (Warren 1997). In the study area, the thick evaporitic series could have acted as an efficient seal for upward migration of deep fluids in the clastic deposits of Unit 1, underlying the evaporites. It is therefore acceptable that relatively low-salinity fluids within the pre-evaporitic sediments, migrating upward, could have promoted subjacent dissolution of the more soluble evaporites. Importantly, this mechanism could account for the circular and localized geometry of the dissolution structures, as explained further below.
The Levant Basin experienced considerable subsidence and sediment accumulation during the Pliocene (Tibor et al. 1992). It is probable that rapid burial of the pre-evaporitic units would render them prone to overpressuring, with the evaporitic series defining the top of the zone of overpressure even at shallow depths, because of the high sealing capacity of the evaporites (Warren 1997). In support of this suggestion, evidence of focused vertical fluid flow processes related to overpressured pre-evaporitic sedimentary units such as mud volcanoes, diapirs and sea-bed mounds have been described in the study area (Frey-Martinez et al. 2005b) and in the nearby deep Nile cone area (Loncke et al. 2004). Most significantly, a series of kilometre-scale conical mounds developed within the Yafo Sand Member has been attributed by Frey-Martinez et al. (2005b) to highly focused vertical fluid flow along the axis of the Afiq Canyon, providing evidence of sediment remobilization of the canyon fill and the Yafo Sand Member sediments. The striking similarity in the geometry and timing between the dissolution structures and the conical sea-bed mounds strongly supports the hypothesis of a common mechanism for their origin, driven by vertical and localized expulsion of fluids in the overpressured pre-evaporitic sediments. The location of the dissolution structures (Fig. 7) and of the mounds along the Oligo-Miocene depositional fairway of the Afiq Canyon points to a pivotal role of the El-Arish and Afiq submarine canyons on the pathways and localization of upward escaping pore fluids.
It is probable that where the Messinian evaporites were too thick to be pierced, the focused fluid flow partially dissolved and fractured the base of the evaporites, causing the fluids to penetrate into the upper parts of the evaporites. We thus propose a model for the evolution of the circular dissolution structures in the study area, which is illustrated and explained in detail in Figure 13.
Schematic illustration of the successive phases of formation of the circular dissolution structure CS-1. The dissolution process started in the early Piacenzian (b), when vertical focused fluid flow begun corroding the lower evaporitic unit and dissolving its upper and more soluble part, causing collapse of the overburden and successive onlap of sediments (b and c) at the coeval sea bed. In this interval the formation of the concentric faults directly related to the collapse of the sediments above the depleted evaporitic unit should be noted. The process terminated by the late Piacenzian (d), with the deposition of Horizon PL50. N, Horizon N; M, Horizon M.
Previously it has been argued that the depletion of evaporites to form the depressions above was focused mainly within the upper part of the Messinian evaporites. At first sight, this might be taken to suggest that suprajacent dissolution was a more likely model. However, in our view it is equally plausible that fracture conduits for the upwelling of undersaturated fluids ascending from below (Warren 1999) could easily have preferentially dissolved the most soluble lithologies, located at the top of the layered evaporitic package. The fracture conduits would have been opened by the ascending high-pressure fluids, acting ‘corrosively’ on the lower part of the evaporite succession. This ‘corrosive’ action might then also explain the extraordinary circular planform of the dissolution structures themselves.
Origin of fluid flow
The origin of the fluids responsible for dissolution of the evaporites and creation of the circular structures can conceivably be attributed to three potential sources: (1) gypsum to anhydrite conversion; (2) sediment compaction; (3) seismic pumping related to earthquakes on deep basement faults. The source of fluid supply needs to account for the timing, duration, salinity and, most importantly, the circularity of the dissolution structures.
(1) Burial of gypsum results in its dehydration into anhydrite and this conversion releases fluids that are undersaturated in Na and Cl, and thus capable of dissolving halite (Warren 1991). This conversion could have occurred at the base or within the evaporitic unit in the study area. However, the temperature and pressure conditions required (Warren 1991) were reached only within the Pleistocene, therefore much later than the formation of the circular structures. Gypsum to anhydrite conversion is therefore rejected as a possible mechanism for generation of the dissolving fluids.
(2) The second possible source of relatively undersaturated fluids is the compaction of the thick sedimentary sequences beneath the evaporites. Pore fluids expelled during compaction would most probably be undersaturated in composition (Magara 1978). However, compaction of the pre-evaporitic sediments in the study area represents a gradual and areally widespread form of fluid expulsion, and it is therefore difficult to reconcile this scale of laterally pervasive dewatering with the formation of circular dissolution structures, which by their discrete geometry suggest a more focused supply of fluid.
(3) The third process potentially able to produce the requisite flow of undersaturated fluid sufficient to dissolve the evaporite sequence from below is seismic pumping related to earthquakes on deep basement faults. The tectonic framework of the study area shows clear evidence of the existence of deep-seated structures below the Messinian evaporites (Fig. 14). The principal structures are anticlines and faults linked to the Syrian Arc foldbelt (Fig. 14; Neev & Ben-Avraham 1977; Garfunkel et al. 1979). The area is seismologically active at present and this level of activity could be expected to have occurred throughout much of the Neogene (Garfunkel & Almagor 1985; Garfunkel 1998; Vidal et al. 2000). Major earthquakes are well known to lead to expulsion of basinal fluids (Sibson et al. 1978), therefore the activity of deep-seated structures could have generated focused fluid flow by seismic pumping.
Seismic section crossing the Levant continental margin in a west–east direction, showing the deep structural setting of the study area (location of seismic section shown in Fig. 7). The relative position of the structure CS-3 above the Afiq submarine canyon (dashed line) and the axis of the Syrian Arc anticline, and related fault system (dotted lines), should be noted. B.S., Base Senonian horizon; M, Horizon M; N, Horizon N. The interpretation of the faults post-dating the deposition of the Messinian evaporites, within Unit 3, is indicated with dotted lines.
Although it could be argued that the deep-seated tectonic structures would probably represent a linear fluid flow at depth, the flow could have become focused during its ascent. For example, previous workers have linked the occurrence of mud diapirs, mud volcanoes and slumps to earthquake activity in the southeastern Mediterranean region (Loncke et al. 2004; Frey-Martinez et al. 2005b).
Based on the arguments presented above, the most probable trigger mechanism for the initiation of the dissolution process is considered to be seismic activity. The circular dissolution structures are strikingly similar in shape, timing and location to mounded structures caused by mud diapirism at this level of the Yafo Sand Member described just 50 km away, landward of the pinch-out of the Messinian evaporites (Frey-Martinez et al. 2005b). This similarity argues strongly for a common origin. It therefore seems entirely plausible that similar earthquake-triggered and focused vertical fluid flow could have impinged on the base of the evaporitic seal within the study area to the point where dissolution occurred with the resultant localized circular dissolution structures in the overburden. It is interesting to consider that were it not for the presence of the highly efficient seal of the Messinian evaporites, this form of vertical fluid expulsion might simply have resulted in mud diapirism or extrusion as is the case elsewhere within the region updip of the pinch-out of the Messinian evaporites (Frey-Martinez et al. 2005b).
The dissolution mechanism proposed in this paper may be applicable worldwide in other evaporite-bearing basins where similar conditions for expulsion of focused fluid flow are present (e.g. West Africa and Gulf of Mexico). Ultimately, as demonstrated here, such a mechanism can lead to breaching of basinwide seals by depletion of considerable thicknesses of evaporites, thus creating effective ‘windows’ for vertical hydrocarbon migration. An in-depth understanding of the processes presented here is therefore critical for a better evaluation of the efficiency of evaporitic seals, the presence of stratigraphic traps, and pathways for the vertical migration of hydrocarbons.
Conclusions
This study records the occurrence of buried circular evaporite dissolution structures above a tabular evaporitic body in the Levant Basin, Eastern Mediterranean. Dissolution occurred in buried evaporites and in a deep-water setting during the Pliocene. The estimated maximum time of duration of the event is 0.75–1 Ma.
Evaporite dissolution led to the collapse of a weakly lithified overburden, which deformed with a series of extensional concentric faults. Other less conclusive indications of evaporite dissolution and evaporite-related deformation are observed close to the present margin of the Messinian evaporites; namely, salt outliers, faults detaching above the evaporite unit, and a linear depression along the evaporite pinch-out.
The mechanism of dissolution proposed for the creation of the circular collapse structures is subjacent dissolution of the more soluble evaporites in the Messinian evaporites, as a result of focused vertical fluid flow at the base of the evaporitic series. Instantaneous release of overpressured fluids, as occurs, for example, during an earthquake, could have triggered and focused the fluid flow, which is thought to have impinged on the base of the evaporitic seal to the point where dissolution occurred with the resultant localized circular collapse structures in the overburden.
Acknowledgements
We are grateful to BG-Group for supplying the data and for permission to publish this paper, and in particular to I. Campbell for suggesting important changes to the manuscript. We gratefully acknowledge Editor I. Alsop and I. Davison and an anonymous referee, whose reviews significantly helped improve the manuscript. M. Huuse, R. Davies and J. Frey-Martinez are thanked for discussion and comments.
- © 2005 The Geological Society of London
References
3D seismic analysis of circular evaporite dissolution structures, Eastern Mediterranean
