Folds, salt diapirs and their relationship to faults

Quick refresher

An anticline bulges upwards, with older rocks in its core. A syncline is the opposite. If we do not know the age of the layers, we call the folds an antiform and a synform.
We describe a fold by its limbs (the two sides), its hinge (the line of greatest bending), its axial surface, its interlimb angle and the plunge of its hinge. The crest is the highest point of a layer. In an inclined fold it does not coincide with the hinge.
The axis of a cylindrical fold is the line where its limbs intersect. We find it as the cross product of the normals to the limbs (the lines at right angles to them). The axial plane bisects (cuts in half) the interlimb angle. The crest of a deeper horizon is shifted in the direction in which the axial surface dips.
Folds grow by buckling, when the layers are squeezed along their length, or by bending over basement blocks (blocks of the old rock beneath the sediments), salt and reefs. When a layer is bent, its outer arc is stretched: $e = y/R$ . Slip along layer boundaries reduces the stretching of each layer.
Salt is almost incompressible (it cannot be squeezed into less volume) and it flows. Compacted sediments deeper than about 1–1.5 km are usually denser than salt. Salt flows mainly because the load on it is uneven (differential loading). The salt pierces its roof where stretching or erosion has weakened it, and then it goes on growing while sediment piles up around it.
Traps beside salt are: crests over pillows and domes; layers that end against the flank of the salt; traps beneath overhangs; and traps between domes. Where the salt has been squeezed out completely (a salt weld), “windows” remain where fluid can cross from the subsalt rocks to the suprasalt rocks.

Key relationships

\Theta = 180° - \angle(\mathbf n_1, \mathbf n_2),\qquad \mathbf l = \mathbf n_1\times\mathbf n_2

interlimb angle and direction of the fold axis from the normals to the limbs

\Delta x = \frac{\Delta z}{\mathrm{tan}\,\delta_a}

shift of the crest of a horizon lying $\Delta z$ deeper, in a similar fold whose axial plane dips at angle $\delta_a$

e = \frac{y}{R},\qquad \frac{1}{R} \approx \frac{\dd^2 z}{\dd x^2}

stretching at a distance $y$ from the neutral surface of a layer (the surface that is neither stretched nor squeezed), and the curvature at the crest from depth values

\Delta p = \sigma_v(z_1) - \sigma_v(z_2) - \rho_s\,g\,(z_1 - z_2)

pressure difference in the salt beneath a minibasin (a depression between domes; top of salt at $z_1$ ) and beneath a dome (at $z_2$ )

Common pitfalls

Placing a well for a deep horizon directly under the crest of a shallow one: in an inclined fold the crest shifts with depth.
Mixing up the hinge, which is the line of greatest bending, with the crest, which is the highest point of a layer.
Taking the shape of a fold on a vertical section or a map for its true profile: the true profile lies in the plane perpendicular to the hinge.
Explaining the growth of salt domes only by salt being lighter than the sediments: what decides is uneven loading (differential loading) and the strength of the roof.
Treating salt as a continuous seal everywhere: where it has been squeezed out, windows for crossflow remain.

After this lesson you can

describe a fold by its parts (limbs, hinge, axial surface, crest, interlimb angle and hinge plunge), and classify it by its interlimb angle, the position of its axial surface and the shape of its layers
find the axis and axial plane of a fold from the attitudes of its limbs (the direction and angle of their slope), and estimate how far the crest of a deeper horizon is shifted
explain buckling (squeezing of layers along their length), bending and slip along layers, and estimate how much the rock at the crest is stretched from the curvature of the fold
explain why and how salt flows, how pillows, domes and diapirs grow, and what traps and risks go with them, including in the Pre-Caspian Basin
recognise folds that are linked to faults, and understand what this means for the trap and for wells

Elements of a fold

Folds come in every size, from millimetre-scale folds in core to folds tens of kilometres across. Large folds, kilometres to tens of kilometres long, form traps, but small folds are usually found beside them. A description of a fold is geometric: it says what is bent and how, not why. The causes of bending are covered in the section on how folds grow.

Limbs and hinge zone: The parts of a layer with the least curvature are the limbs, and the place of greatest bending is the hinge zone. The line of greatest bending on the surface of a layer is the hinge line, or simply the hinge.
Axial surface: The surface that passes through the hinges of all the layers of a fold. It roughly cuts the interlimb angle in half. A flat axial surface is called an axial plane.
Fold axis: The direction parallel to the hinge. Move a straight line of this direction along a curved surface: if this can draw the fold, it is a cylindrical fold. Over a stretch a few kilometres long many folds are close to cylindrical, but not at their ends, where the hinge plunges.
Interlimb angle: The angle between the limbs, measured in the profile of the fold. The smaller the angle, the tighter the fold.
Crest, crest line, trough line: The highest point of a bent layer is the crest, and the line joining such points along the fold is the crest line. In a syncline the lowest points are joined by the trough line.
Plunge of the hinge: The angle between the hinge and the horizontal. At the ends of an anticline the hinge plunges (dips down), and the fold closes periclinally (the layers dip away from the centre). In a syncline the fold closes centriclinally (the layers dip towards the centre). The highest point of the hinge is the culmination, and the lowest is the depression.

SchematicThe parts of an asymmetric anticline in profile (schematic): the gentle and steep limbs, the hinges of the layers and the axial surface through them, the crest of the top layer and the interlimb angle. The crest does not lie on the hinge. It is shifted towards the gentle limb. The layers are similar to one another, and the hinge of each next layer is shifted along the axial surface.

How folds are described

Folds by interlimb angle

Fold: Gentle
Interlimb angle: 180–120°

Fold: Open
Interlimb angle: 120–70°

Fold: Close
Interlimb angle: 70–30°

Fold: Tight
Interlimb angle: Less than 30°

Fold: Isoclinal
Interlimb angle: Limbs parallel, about 0°

Fold	Interlimb angle
Gentle	180–120°
Open	120–70°
Close	70–30°
Tight	Less than 30°
Isoclinal	Limbs parallel, about 0°

The class boundaries are only conventions. The value of the angle itself matters more.

We group folds by the position of the axial surface. Upright folds have an almost vertical axial surface. Inclined folds have a tilted one. In overturned folds, one limb has been tipped over and its layers are upside down. Recumbent folds lie on their side, with an almost horizontal axial surface. By the position of the hinge, folds are horizontal, plunging or vertical. By their shape in plan view (seen from above), we tell apart linear folds, many times longer than they are wide; brachyanticlines and brachysynclines, which are short, only a few times longer than wide; and nearly equidimensional domes and basins. In a symmetric fold the limbs have equal length and dip. In an asymmetric fold, one limb is shorter and steeper.

By the shape of the layers, we tell apart two kinds of fold. In parallel folds, the thickness of a layer measured at right angles to the bedding is constant. (Concentric folds, which have a common centre of curvature, are a special case.) In similar folds, the thickness measured along the axial surface is constant. The layer is thicker in the hinge zone and thinner on the limbs, and all the layers repeat the same shape. A more precise description of the shape of a layer uses dip isogons: lines joining points of equal dip on the top and the base of the layer (classes 1A–1C, 2 and 3). Folds that form traps are usually gentle and open, upright or slightly inclined, with layers close to parallel.

The difference matters when we predict what lies at depth. A similar fold repeats its shape at any depth and only shifts along the axial surface. In a parallel fold the curvature changes with depth. In the core of an anticline the layers bend more and more tightly, and somewhere below, the fold must die out, above a ductile layer or a detachment surface (a surface along which the upper rocks slide over the lower ones). So we cannot simply carry the map of a shallow horizon over to a deep one.

The profile shows the true shape of a fold. It is the cross-section made by a plane at right angles to the hinge. On a vertical section that is not at right angles to the hinge, and on a map, the shape is distorted. A horizontal surface cuts a fold whose hinge plunges at $p$ so that the “nose” of the fold on the map is $1/\sin p$ times longer than in the profile. At a plunge of 5° this is 11.5 times. This is why the plunging ends of anticlines look like long, gentle tongues on structure maps.

Worked exampleFold axis and axial plane from dipmeter data

Given (assumed data)

West limb (well 1, dipmeter survey): dip azimuth 285°, dip 18°
East limb (well 2): dip azimuth 95°, dip 52°
Task: find the fold axis, the interlimb angle and the axial plane; find how far, and in which direction, the crest of a horizon lying 900 m deeper is shifted

Solution

Normals to the limbs (east; north; up), as in the lesson on structural analysis: $\mathbf n_1 = (-0.2985;\ 0.0800;\ 0.9511)$ , $\mathbf n_2 = (0.7850;\ -0.0687;\ 0.6157)$ .
The axis is the line where the limbs intersect: $\mathbf l = \mathbf n_1\times\mathbf n_2 = (0.1146;\ 0.9304;\ -0.0423)$ , $|\mathbf l| = 0.9383$ . The trend is $7.0°$ and the plunge is $\arcsin(0.0423/0.9383) = 2.6°$ . So the hinge is almost horizontal and plunges gently to the north.
The angle between the normals: $\cos\gamma = \mathbf n_1\cdot\mathbf n_2 = 0.3457$ , $\gamma = 69.8°$ . The interlimb angle is $\Theta = 180° - 69.8° = 110.2°$ , so the fold is open and markedly asymmetric (limbs of 18° and 52°).
The axial plane bisects the interlimb angle. Its normal points along $\mathbf n_1 - \mathbf n_2 = (-1.0835;\ 0.1487;\ 0.3354)$ . The dip is $\arccos(0.3354/1.1439) = 73.0°$ and the dip azimuth is 277.8°: the axial plane dips steeply to the west, towards the gentle limb.
The crest of the horizon 900 m deeper is shifted in the direction in which the axial plane dips, by $\Delta x = 900/\mathrm{tan}\,\delta_a = 900/3.261 = 276$ m.

Answer: The fold axis plunges at 2.6° towards an azimuth of 7.0°. The interlimb angle is 110° (an open, asymmetric fold). The axial plane has a dip azimuth of 278° and a dip of 73°. The crest of the horizon 900 m deeper lies about 276 m west of the crest of the upper one.

What it means. We run seismic lines for such a fold across the axis, at azimuths of about 97–277°. Only on these lines can we see its true profile. A planned well for the deep horizon is moved to the west, towards the gentle limb. If it were placed under the crest of the upper horizon, it would reach the deep horizon on the steep eastern limb, below its crest. We calculated the shift for a similar fold. In a parallel fold the shape changes with depth, and we map the deep horizon separately from seismic data. The axis, the line where the limbs intersect, is the same as in the problem of the intersection of two planes (stereographic projection). The data in this example are assumed for illustration.

CalculatedEqual-area lower-hemisphere projection for the example, drawn with the formulas of the lesson. It shows the arcs of the fold limbs and their poles (points of the same colour), the fold axis (the point where the arcs meet, trending about 7° and plunging about 3°) and the dashed arc of the axial plane, which passes through the axis.

How folds grow and where the rock fractures

Folds arise in two ways. In buckling, the sequence is squeezed along the layers. A strong layer between weaker ones (a “competent” layer) loses its stability and bends into waves. This is how chains of folds grow in fold belts. The wavelength is proportional to the thickness of the competent layer, and it grows with the viscosity contrast (how much stiffer the strong layer is than the rocks around it; see the advanced section). Buckle folds form chains and are usually elongated in plan view. In many of them the length is about 10 times the width, although the scatter is large. In bending, a load acting across the layers flexes them. This happens over uplifted basement blocks, over salt pillows and domes, and over reefs and sand bodies around which the clays have compacted more. Such folds follow the shape of whatever lies beneath them, and they do not necessarily show compression.

A layered sequence does not bend like one thick beam. It can slip along the layer boundaries (flexural slip), and each layer bends almost independently. A pack of cards behaves in the same way. It is easy to bend while the cards slide over one another, and almost impossible to bend once they are glued together. Slip along layers greatly changes how strain is spread inside a fold.

Extension and compression of a layer in bending

e = \frac{y}{R},\qquad e_{max} = \frac{h}{2R},\qquad \frac{1}{R} \approx \frac{\dd^2 z}{\dd x^2} \approx \frac{z_1 - 2z_0 + z_2}{a^2}

where:

$e$ relative elongation along the layer (strain); positive for extension, negative for shortening, —
$y$ distance from the neutral surface of the layer, which keeps its length: positive towards the convex side (the outer arc), negative towards the centre of curvature, m
$R$ radius of curvature of the neutral surface, m
$h$ thickness of a layer that bends as a whole (no slip within it), m
$z_1,\ z_0,\ z_2$ depths of the surface at three points of a profile perpendicular to the hinge, equally spaced by a distance $a$ ; $z_0$ is at the crest, m

Units: lengths in the same units; $e$ is dimensionless (multiply by 100 to get per cent)

Holds when

The neutral surface lies in the middle of the layer, strains are small and the layer does not slip within itself.
The curvature from the second difference (the formula with $z_1 - 2z_0 + z_2$ ) is valid when the slope at the middle point is small (at the crest $\dd z/\dd x \approx 0$ ). Otherwise use $1/R = z''/(1 + z'^2)^{3/2}$ .
We take the profile at right angles to the hinge. In any other direction the curvature comes out smaller.

The outer arc of a layer is longer than the neutral surface by the fraction $y/R$ , and the inner arc is shorter. The stretching grows with the thickness of the layer and with the curvature. A thin layer in a gentle fold is hardly deformed, while a thick layer in a tight hinge is stretched by several per cent. When layers slip along their boundaries, we put the thickness of an individual layer into the formula, not that of the whole stack.

Worked exampleStretching of the rock at the crest, from the curvature of the fold

Given (assumed data)

Depths of the top of a carbonate layer in three wells along a profile across the hinge: 1846, 1820 (crest) and 1838 m; spacing between wells 400 m
Layer thickness: 40 m
Case with slip along layers: the 40 m layer as a stack of 10 layers of 4 m, slipping along their boundaries
Tensile strength (the stress at which the rock breaks when pulled) and Young's modulus (its stiffness) of the rock: 5 MPa and 25 GPa (order of magnitude for tight limestone)

Solution

Second difference: $z'' = (1846 - 2\cdot 1820 + 1838)/400^2 = 2.75\cdot 10^{-4}$ m⁻¹. The slope at the crest, $(1838 - 1846)/800 = -0.01$ , is small, so $R = 1/z'' = 3636$ m.
The layer bends as a whole: $e_{max} = 20/3636 = 0.55$ %.
The 4 m layers slip along their boundaries: $e_{max} = 2/3636 = 0.055$ %.
The strain at which the rock breaks in tension (the failure strain) is $5/25{,}000 = 0.02$ %. The ratios of the two bending strains to it are 27.5 and 2.75.

Answer: Without slip along layers, the top of the layer at the crest is stretched by 0.55%, which is 27.5 times the failure strain. With slip along the layers every 4 m, it is stretched by 0.055%, which is 2.75 times the failure strain.

What it means. In both cases we expect extension fractures (joints) at the crest, but the picture differs. A thick layer that bends as a whole is heavily fractured at its top and squeezed at its base, where stylolites and small thrusts are possible. A stack of thin slipping layers is fractured less and more evenly, because each fracture stays inside its own layer. So we predict the fracturing of a crest from the curvature together with the thickness of the mechanical layers. The curvature of the top of a horizon, taken from seismic data, is a common attribute (a quantity measured from the data) for predicting fractures (Fractures and naturally fractured reservoirs). The limbs, where the curvature is small, are only slightly stretched. The calculation is simplified: we know the failure strain of the rock only as an order of magnitude, and real layers are not uniform. The data in this example are assumed for illustration.

CalculatedMaximum extension (stretching) at the top of a bent layer against the radius of curvature, from the formula

e = h/(2R)

. One line is for a 40 m layer bending as a whole, and one is for 4 m layers slipping along their boundaries (logarithmic scales). The dashed line shows the order of magnitude of the failure strain of tight limestone in tension. The points are the crest from the example.

Where the rock fractures also depends on how the fold grew. In buckle folds the stretching is concentrated in the outer arcs of the hinge zones: in the upper part of the layers in the hinge of an anticline, and in the lower part in the hinge of a syncline. The limbs are only slightly deformed. In folds linked to faults, fractures also gather at the fault itself and on the steep limb (see the section on folds and faults).

Folds as traps

The dip of the beds closes an anticline on all sides: this is four-way closure (closure and spill point). Across the hinge, the dip of the limbs closes the fold. Along the hinge, the plunge of the hinge towards both ends closes it. The crest of the trap is at the culmination of the hinge. The spill point is usually at the higher of the saddles at the ends of the fold, where the hinge rises again towards the neighbouring fold. So when we assess a trap, not only the limbs matter but also the hinge. When the plunge is gentle, the closure along the hinge may be smaller than the closure across the fold.

Two things discussed above matter for deep horizons. First, the crest of an inclined fold shifts with depth, in the direction in which the axial surface dips. Second, the shape of a parallel fold changes with depth, and below it may die out above a detachment layer (a layer of ductile clay or salt). So we map the structure horizon by horizon.

Sometimes a fold grows while sediment is being deposited. This is a syndepositional fold, also called a growth fold. Less sediment was deposited over its crest than on the limbs and in the neighbouring synclines, so the beds thin towards the crest. This thinning shows when the fold grew, and so whether the trap existed by the time of migration (timing of trap formation). On the shallow-water crest of a growing fold, the best reservoirs could be deposited, such as wave-washed sands and carbonate buildups. But erosion could also take place.

Salt: why it flows and how diapirs grow

Rock salt (halite) is a special rock. Its density is about 2.16 g/cm³ and hardly changes with depth, because it has no pores and it compresses very little. Sediments are the opposite. Near the surface they are loose and light, about 2.0 g/cm³. As they are buried, they compact and become heavier, up to 2.4–2.6 g/cm³ (compaction trend). Also, at the usual temperatures of a basin, salt flows like a very viscous (thick) liquid (brittle and ductile behaviour). Below a certain depth the salt lies under denser rocks. This is an unstable position, like oil under water.

But buoyancy alone is not enough. The roof above the salt is a strong, brittle plate, and for the salt to rise, this plate must be pushed apart or pierced. The main cause of salt flow is differential loading (uneven load): where there is more sediment above the salt, the pressure in the salt is higher, and the salt flows to where the load is smaller. Wedges of sediment at deltas and slopes, folds and faults in the roof, and tectonic compression and stretching create the unevenness.

Pressure difference in the salt beneath a minibasin and beneath a dome

\Delta p = \sigma_v(z_1) - \sigma_v(z_2) - \rho_s\,g\,(z_1 - z_2) = g\int_{z_2}^{z_1}\bigl(\rho_b(z) - \rho_s\bigr)\,\dd z

where:

$\Delta p$ excess of the pressure in the salt beneath the minibasin over the pressure in the salt beneath the dome, at the same depth $z_1$ , Pa
$z_1$ depth of the top of the salt beneath the minibasin (the thickness of sediment above the salt there), m
$z_2$ depth of the top of the salt at the dome, m
$\sigma_v(z)$ overburden pressure of the sediments at depth $z$ (overburden pressure), Pa
$\rho_b(z)$ bulk density of the sediments at depth $z$ , kg/m³
$\rho_s$ density of the salt, about 2160, kg/m³
$g$ acceleration due to gravity, 9.81, m/s²

Units: SI; in metric oilfield units, replace $g$ by 0.00981 (densities in g/cm³, pressure in MPa)

Holds when

Salt behaves as a fluid: the pressure in it is the same in all directions and increases with depth according to its density.
The minibasin and the dome are wide enough for the weight of the column above a point to set the pressure. We ignore the strength of the roof and the friction against the walls, although they slow the flow.
We compare the same depth $z_1$ . Below it both columns consist of the same salt, so the difference does not change.

The difference is the weight of the extra sediment above the minibasin, minus the weight of the salt that it replaces. As long as the sediments between $z_2$ and $z_1$ are on average lighter than salt, the load does not drive the salt towards the dome. Once the sediments have compacted and become heavier, every new metre of sediment in the minibasin increases the difference. The difference vanishes when the salt beneath the minibasin has been squeezed out completely.

Worked exampleWhy salt flows from beneath a minibasin towards a dome

Given (assumed data)

Compaction trend of the sediments (as in the lesson on overburden pressure): $\phi_0 = 0.40$ , $L = 2000$ m, grain and water densities 2.65 and 1.05 g/cm³
Salt density: 2.16 g/cm³
Minibasin: sediments above the salt down to a depth of 3000 m
Dome: top of the salt at a depth of 500 m

Solution

The bulk density of the sediments is $\rho_b = 2.65 - 1.60\cdot 0.40\,\e^{-z/2000} = 2.65 - 0.64\,\e^{-z/2000}$ . It equals the density of the salt when $\e^{-z/2000} = (2.65 - 2.16)/0.64 = 0.7656$ , that is, at a depth of $z = 534$ m. Below this depth the sediments are denser than the salt.
The mean density of the whole column above a point, $\sigma_v/(g z)$ , reaches the density of the salt deeper, at about 1120 m. The light upper hundreds of metres keep pulling the average down for a long time.
Overburden pressure from the trend: $\sigma_v(3000) = 68.23$ MPa, $\sigma_v(500) = 10.22$ MPa. The column of salt from 500 to 3000 m: $0.00981\cdot 2.16\cdot 2500 = 52.97$ MPa.
The difference is $\Delta p = 68.23 - 10.22 - 52.97 = 5.04$ MPa. This is equivalent to a column of salt $5.04/(0.00981\cdot 2.16) = 238$ m high.

Answer: At a depth of 3000 m the pressure in the salt beneath the minibasin is 5.04 MPa higher than beneath the dome, and the salt is squeezed towards the dome.

What it means. A difference of a few megapascals is small compared with the pressures themselves (about 68 MPa). But it acts for millions of years on very viscous salt, and so the domes grow while the minibasins deepen as they receive new sediment. Suppose the minibasin were filled with young, loose sediment less than a kilometre thick. On average it would be lighter than the salt, and loading by itself would not squeeze the salt out. This is why the first pillows and diapirs are more often linked to stretching, compression or a pile of sediment that has already built up. The porosity at the surface in the trend from the lesson on overburden pressure, 0.40, is modest. In clay-rich sediments it is usually higher. At 0.5–0.6 both depths increase, to roughly 1–1.3 km for the density at a given depth and 2–3 km for the average. So the example understates the thickness of sediment beyond which loading begins to squeeze the salt out. The calculation shows the driving force, not the speed: the strength of the roof and the friction against the walls slow the flow. The data in this example are assumed for illustration.

CalculatedDensity of the sediments on the compaction trend of the example, shown at a given depth and as an average over the whole column above it, together with the density of the salt. The points mark the depths below which the sediments are denser than the salt: about 530 m for the density at a given depth, and about 1120 m for the average.

Salt pierces its way upwards in stages. First, stretching breaks and thins the roof with normal faults, and the salt fills the space beneath the graben (a down-dropped block between two faults) that forms. This is reactive diapirism. Once the roof is thin, the salt lifts and pushes it aside by itself. This is active diapirism. When the diapir reaches the surface or the sea floor, it grows passively (downbuilding). Its top stays at the surface while sediment piles up around it, and it “grows” up through an ever thicker pile. The faster the salt rises compared with the accumulation of sediment, the more steeply the beds are turned up at the flank of the diapir. Compression also helps: salt rises into the cores of growing anticlines.

Salt structures and traps beside salt

Salt pillow: A gentle uplift of salt. The roof above it is bent into a fold but not pierced.
Salt wall, salt ridge: An elongated uplift of salt. Above large walls, and along them, normal faults break the roof.
Stock, diapir, overhang: A column of salt that has pierced the roof. Near its top the salt can spread sideways and form an overhang. Where it has reached the surface or the sea floor, it forms salt sheets and canopies (wide, flat sheets of salt).
Minibasin: A depression between domes. The salt has moved out of it, and a thick pile of sediment has built up in it. Along its edges, next to the domes, are rim synclines (troughs around the domes).
Salt weld: A place where the salt has been squeezed out completely and the suprasalt rocks rest on the subsalt ones. For the subsalt tier it is a “window” in the salt seal: fluids pass through it from one tier of the section to the other.
Turtle structure: An anticline in the middle of a former minibasin. The thickest middle part of the minibasin squeezes out the salt beneath it earlier than the rest, and it “sits down” on the subsalt floor. The edges keep sinking, because the salt escapes from under them towards the neighbouring domes. As a result the middle ends up higher than the edges, and the former syncline turns into an anticline.

SchematicSalt structures and traps beside salt (schematic): a pillow with an anticline above it, a diapir with an overhang, a minibasin with a turtle structure, a salt weld (a place where the salt has been squeezed out completely) and a subsalt buildup beneath the salt. Oil is shown in orange.

Traps beside salt

Trap: Crest above a pillow or dome
How it is closed: Anticline above the salt, closed on all sides
Main risk: Faults and fractures above the growing crest

Trap: Against the flank of a dome
How it is closed: Beds are turned up and truncated by the salt, which acts as the seal
Main risk: Complex shape, distorted seismic image

Trap: Beneath an overhang
How it is closed: The bed butts against the overhanging salt
Main risk: Finding the trap beneath the salt and drilling through the salt

Trap: Blocks above a dome
How it is closed: Faults above the crest divide the beds into blocks
Main risk: Sealing by faults (Faults)

Trap: Turtle structure in a minibasin
How it is closed: Anticline formed as the salt moves out from beneath the edges of the minibasin
Main risk: Timing of formation relative to migration

Trap: Subsalt traps
How it is closed: Regional salt above buildups and folds
Main risk: Windows in the salt, overpressure (fluid pressure much higher than normal), great depth

Trap	How it is closed	Main risk
Crest above a pillow or dome	Anticline above the salt, closed on all sides	Faults and fractures above the growing crest
Against the flank of a dome	Beds are turned up and truncated by the salt, which acts as the seal	Complex shape, distorted seismic image
Beneath an overhang	The bed butts against the overhanging salt	Finding the trap beneath the salt and drilling through the salt
Blocks above a dome	Faults above the crest divide the beds into blocks	Sealing by faults (Faults)
Turtle structure in a minibasin	Anticline formed as the salt moves out from beneath the edges of the minibasin	Timing of formation relative to migration
Subsalt traps	Regional salt above buildups and folds	Windows in the salt, overpressure (fluid pressure much higher than normal), great depth

Traps beside salt are often combined: the crest, truncation against the salt flank and faults work together.

Salt also causes problems. While drilling, the salt flows and narrows the hole. So casing across salt (the steel pipe that lines the hole) is designed for a pressure close to the overburden pressure, the pressure of the rock above (Casing: purpose, types, properties and standards). Salt dissolves in fresh drilling fluid, so crews drill it with salt-saturated or oil-based muds (Oil-based muds). Within the salt there are layers and inclusions with abnormally high pressure. Seismic waves usually travel faster in salt than in the surrounding sediments, and the image beneath domes is distorted. This is why the shape of subsalt traps is uncertain. Salt conducts heat well, so beneath domes the rocks are colder, and above them warmer, than in their surroundings (heat flow and salt).

Salt of the Pre-Caspian Basin: the subsalt and suprasalt tiers

The most striking example of salt tectonics in Kazakhstan is the Pre-Caspian Basin (how the basin formed). The salt of the Kungurian Stage, the upper part of the Lower Permian, divides its section into two tiers. The original thickness of the salt is estimated at 4–5 km. The salt began to deform soon after it was deposited, most strongly in the Late Permian and Triassic. But the domes went on growing later. The deformation continues today, and some domes reach the surface. More than a thousand salt domes have been identified in the basin. Near the margins they are arranged in rows parallel to the margins. In the minibasins between the domes, sediments of the Late Permian and Early Triassic built up to thicknesses of several kilometres. Some Jurassic–Neogene intervals locally reach thicknesses of up to 3 km where a dome was growing nearby at the time.

The two petroleum-bearing tiers of the Pre-Caspian Basin

Tier: Subsalt (Upper Devonian – Lower Permian)
Rocks: Carbonate platforms and buildups around the margins; in the centre, deep-water black shales, the main source rocks
Traps and examples: Reefs and buildups: Tengiz, Kashagan, Karachaganak; anticlines of the eastern margin formed by compression from the Urals side in the Permian–Triassic: Zhanazhol and its neighbours

Tier: Suprasalt (Upper Permian – Neogene)
Rocks: Mainly sandstones and shales
Traps and examples: All productive traps are associated with salt: anticlines with a salt core, and beds sealed by faults and by the flanks of domes; the main accumulations are in Jurassic and Cretaceous sandstones

Tier	Rocks	Traps and examples
Subsalt (Upper Devonian – Lower Permian)	Carbonate platforms and buildups around the margins; in the centre, deep-water black shales, the main source rocks	Reefs and buildups: Tengiz, Kashagan, Karachaganak; anticlines of the eastern margin formed by compression from the Urals side in the Permian–Triassic: Zhanazhol and its neighbours
Suprasalt (Upper Permian – Neogene)	Mainly sandstones and shales	All productive traps are associated with salt: anticlines with a salt core, and beds sealed by faults and by the flanks of domes; the main accumulations are in Jurassic and Cretaceous sandstones

A brief, simplified summary. The largest reserves of the basin are in the subsalt tier; in its Russian part the Astrakhan field belongs to it as well.

Salt is the regional seal of the subsalt tier, and the two tiers are separate in terms of fluid flow and pressure. The subsalt beds, except for some of the accumulations along the eastern margin, are overpressured (their fluid pressure is typically 1.8–2.1 times the hydrostatic pressure, the pressure of a column of water) and contain more saline waters. Above the salt the pressure is hydrostatic (Abnormal formation pressures: origin, prediction and recognition). The suprasalt accumulations are concentrated mainly in the south-east of the basin, in the areas north and south of the Emba river. The most productive Jurassic and Cretaceous horizons lie at depths of less than 1000 m. The porosity of their sandstones is 25–35%, and the permeability is usually hundreds of millidarcies (mD, a unit of permeability). Commercial oil was first obtained in the South Emba area in 1911, and by 1970 more than a hundred fields had been discovered in the suprasalt tier. In the 1990s, more detailed seismic surveys made it possible to map structures in the Upper Permian rocks as well, in the minibasins between the domes: crests within the minibasins themselves and “half-crests” on the slopes of domes, formed by uneven outflow of salt.

Geological and geochemical data show that the oil of the suprasalt accumulations came from the same subsalt source rocks. It rose where the salt between the domes has been squeezed out completely or almost completely, that is, through such “windows” in the seal. The literature on the basin calls these areas salt-free troughs. Hydrocarbons could also have escaped from subsalt traps through such windows, which is why some high-relief subsalt buildups are probably water-bearing. Salt also affects temperature. In the east of the basin, beneath a salt dome 3.5 km thick, the top of the subsalt rocks is 18 °C colder than at the same depth in the neighbouring minibasin. Beneath the largest domes the difference can reach 60 °C. This changes both the maturation of source rocks and the prediction of reservoir temperature.

Folds and faults together

Folds and faults often grow together, and the type of fold suggests where to expect faults and fractures.

Fault-bend fold: A thrust (a low-angle fault along which rock is pushed over other rock) runs along a weak bed, then cuts up through the sequence on a ramp, and then runs along a bed again. The thrust sheets ride over the bends in the fault surface and are bent by them, so the shape of the fold follows the shape of the fault surface.
Fault-propagation fold: The fold grows ahead of the tip of a thrust that has not yet broken through the sequence. The forelimb (the front limb) is steep and sometimes overturned, and the fault itself often cuts it. The crest is shifted towards the forelimb.
Detachment fold: The sequence is crumpled above a ductile layer (salt or clay) along which it slides. The fold dies out downwards, towards the detachment layer.
Rollover anticline: The hanging wall (the block above the fault) of a listric normal fault slides down a fault surface that flattens with depth, and it bends towards the fault. The crest is often broken by antithetic faults (faults that dip the opposite way to the main fault) and small grabens (fault types).
Drape fold, forced fold: The layers drape over a basement step above a fault, over a basement high or over a reef. The fold follows the shape of what lies beneath it and dies out upwards.
Inversion anticline: A depression bounded by normal faults is compressed later. The faults are reactivated as reverse faults, and the fill of the former trough is squeezed upwards into an anticline.

SchematicFolds linked to faults (schematics): a fault-bend fold, a fault-propagation fold with a steep forelimb, a rollover anticline in the hanging wall of a listric normal fault, and a drape fold over a basement step.

An example is South Mangyshlak. Before the Jurassic, the northernmost of the Late Permian–Triassic rifts (stretched, sunken zones of the crust) was already compressed and uplifted (inverted). This is the Mangyshlak fold zone. Farther south, on the Zhetybai Step, almost all the reserves of the main productive beds, which are Lower and Middle Jurassic sandstones, are held in elongated anticlines. They are arranged in three lines roughly parallel to the fold zone. The folds range in length from a few kilometres to 45 km (Uzen), and in closure from tens of metres to more than 300 m. The southern limbs are noticeably steeper than the northern ones, and faults often cut them. According to the interpretation of seismic data, the leading edges of thrust sheets in Triassic rocks probably lie beneath the folds, and about 70% of the growth of the structures took place from the Middle Miocene to the present. The seal above Uzen is thinner than 100 m, yet it holds an accumulation more than 300 m high.

Three rules for wells follow from this. First, in asymmetric folds above thrusts, faults often cut the steep limb, and the crest shifts with depth. Second, fractures are concentrated in the hinge zone and on the steep limb, and also near faults. Third, we must not place wells on fault-related folds using the map of a single horizon. We need a three-dimensional structural model that agrees with the seismic data and the wells (Practical structural geology).

Terms and symbols

Anticline and syncline: Folds that bulge upwards (older rocks in the core) and folds that bulge downwards (younger rocks in the core).
Hinge line and hinge zone of a fold: The line of greatest bending of a layer and the area around it.
Axial surface: The surface through the hinges of all the layers of a fold, which cuts the interlimb angle in half.
Interlimb angle $\Theta$: The angle between the limbs in the profile of the fold. Folds are classed by it as gentle, open, close, tight and isoclinal.
Plunge of the hinge: The angle between the hinge and the horizontal. At the ends of an anticline the hinge plunges, closing the fold.
Fold profile: The cross-section made by a plane at right angles to the hinge. It shows the true shape of the fold.
Parallel and similar folds: Parallel folds have a constant layer thickness measured at right angles to the bedding. Similar folds have a constant thickness measured along the axial surface.
Buckling and bending: Buckling is the growth of a fold under compression along the layers. Bending is the growth of a fold under a load across them.
Neutral surface: A surface inside a bent layer that keeps its length. Outside it the layer is stretched, and inside it the layer is squeezed.
Differential loading: A difference in the weight of the rocks above the salt. It makes the salt flow.
Salt pillow, dome, diapir: An uplift of salt beneath a bent roof, a dome of salt, and a column of salt that has pierced the roof.
Minibasin: A depression between domes from which the salt has gone and in which a thick pile of sediment has accumulated.
Salt weld: A place where the salt has been squeezed out completely and the suprasalt rocks rest on the subsalt ones — a “window” in the salt seal.

Common mistakes and misconceptions

Wrong

“The crest of a deep horizon lies directly beneath the crest of a shallow one.”

In fact

Only in an upright, symmetric fold. In an inclined fold the crest shifts with depth, in the direction in which the axial surface dips, by $\Delta z/\mathrm{tan}\,\delta_a$ . In a parallel fold the shape also changes with depth.

Wrong

“An anticline always indicates compression.”

In fact

Folds also grow over salt pillows, over basement highs and reefs, and in the hanging wall of listric normal faults when the crust is stretched. Buckle folds and thrusts show compression.

Wrong

“Salt rises because it is lighter.”

In fact

Buoyancy is only part of the cause. In the upper kilometre the sediments are on average lighter than salt, and the roof has to be pierced. What matters most is differential loading and the weakening of the roof by stretching or erosion.

Wrong

“Any trap beneath regionally continuous salt is securely sealed.”

In fact

Where the salt has been squeezed out into domes, windows remain beneath the minibasins, and hydrocarbons escape upwards through them. This is how the suprasalt traps of the Pre-Caspian Basin were filled, and probably why some of the subsalt buildups are water-bearing.

Wrong

“Fractures are distributed evenly across a fold.”

In fact

Stretching is concentrated in the outer arcs of the hinges, and it grows with curvature and layer thickness. On gentle limbs it is small. Slip along layers shares the strain out among the thin layers.

In depth: advanced detail

Where the formulas for the axis, interlimb angle and axial plane come from

The limbs of a cylindrical fold are planes that contain its axis. So the axis is at right angles to the normals of both limbs, and it points along their cross product $\mathbf l = \mathbf n_1\times\mathbf n_2$ (vector calculations). Take both normals pointing upwards, and suppose neither limb is overturned. Then the angle $\gamma$ between the normals equals the angle between the limbs measured outside the fold, and the interlimb angle itself is $\Theta = 180° - \gamma$ . For a fold with a horizontal hinge, whose limbs dip at $\delta_1$ and $\delta_2$ in opposite directions, this gives $\Theta = 180° - \delta_1 - \delta_2$ . In the example, $180 - 18 - 52 = 110°$ , which agrees with the vector calculation up to the effect of the plunge of the hinge.

Two planes cut the angle between two planes in half. Their normals are $\mathbf n_1 + \mathbf n_2$ and $\mathbf n_1 - \mathbf n_2$ . In a gentle fold the first plane is close to horizontal. The second passes between the limbs through the hinge, and this is the axial plane. Both planes contain the axis. So with a horizontal hinge their strike coincides with the direction of the axis. With a plunging hinge the strike deviates from the axis, and the deviation is larger when the plunge is steeper and the dip of the plane is gentler (for the axial plane of the example, it is less than 1°). In a similar fold, a deeper layer is the same bent layer shifted along the axial plane. So its crest is shifted horizontally by $\Delta z/\mathrm{tan}\,\delta_a$ in the direction in which the plane dips.

Wavelength of buckle folds

Suppose a competent layer of thickness $h$ and viscosity (resistance to flow) $\eta_l$ is compressed along its length in a medium of viscosity $\eta_m$ . Folds of one particular wavelength, the dominant one, grow fastest. For a viscous layer in a viscous medium, the theory of buckling (Biot and Ramberg) gives:

Dominant wavelength of buckling

L_d = 2\pi h\,\sqrt[3]{\frac{\eta_l}{6\,\eta_m}}

where:

$L_d$ dominant wavelength of the folds, m
$h$ thickness of the competent layer, m
$\eta_l,\ \eta_m$ viscosities of the layer and of the surrounding rocks, Pa·s

Units: $L_d$ and $h$ in the same units; only the ratio of the viscosities matters

Holds when

A single viscous layer in a uniform viscous medium; small amplitudes at the start of fold growth.
The viscosity contrast is large (tens or more); with a small contrast the layer thickens instead of bending.
Multilayer sequences, slip along layers and brittle rocks give other wavelengths. The formula shows a tendency, not an exact value.

The wavelength is proportional to the layer thickness, and it depends only weakly on the viscosity contrast (as a cube root). A 40 m layer with a contrast of 1000 gives a wavelength of about 1.4 km. With a contrast of 100 it gives about 0.64 km. A 10 m layer with a contrast of 1000 gives about 0.35 km. So the sizes of the trap-forming folds in a fold belt hint at the thickness of the layer that controlled the folding.

Connections to other topics

Stress, strain and structural analysis — attitudes, vectors and strain, and brittle and ductile behaviour of rocks.
Traps and timing — closure, spill point and the timing of trap formation.
Faults — faults, their displacements and sealing.
Plate tectonics and sedimentary basins — how the Pre-Caspian Basin and other basins of Kazakhstan formed.
Pore pressure and overburden pressure — overburden pressure and the compaction trend of sediments.
Reservoir and seal — salt and shales as seals, and what limits the height of an accumulation.
Fractures and naturally fractured reservoirs — fractures in folds and fractured reservoirs.
Structural styles and tectonic history — combinations of folds and faults in different tectonic settings.
Casing: purpose, types, properties and standards — casing wells through intervals of flowing salt.

Check yourself

1
The limbs of a fold dip at 25° to the west and at 25° to the east, and the hinge is horizontal. What is such a fold called according to its interlimb angle?
1. A.Open
2. B.Isoclinal
3. C.Gentle
4. D.Tight
Show the answer
Correct answer: C. Gentle
The interlimb angle is $\Theta = 180° - 25° - 25° = 130°$ , which is more than 120°, so the fold is gentle. Trap-forming folds are usually like this. The dip angles alone do not tell us whether it is an anticline or a syncline: we must know whether the limbs dip away from the hinge or towards it.
2
The axial plane of an anticline dips west at 65°. The crest of the horizon at a depth of 1800 m is known. Where should we look for the crest of the horizon at a depth of 3000 m, if the fold is similar?
Show the answer
The shift is $\Delta x = 1200/\mathrm{tan}\,65° \approx 1200/2.14 \approx 560$ m to the west, in the direction in which the axial plane dips. If the fold is parallel, its shape changes with depth, and we must map the deep horizon separately.
3
Why does salt gather into domes and diapirs?
1. A.Salt expands when heated and rises
2. B.Only because it is lighter than any sediment
3. C.Because of differential loading: under thick sediments the pressure in the salt is higher, and the salt flows to where the load is smaller, especially where the roof is weakened
4. D.Salt dissolves at depth and is redeposited higher up
Show the answer
Correct answer: C. Because of differential loading: under thick sediments the pressure in the salt is higher, and the salt flows to where the load is smaller, especially where the roof is weakened
Salt flows like a very viscous liquid, and a difference in pressure drives it. That difference comes from the different weight of the rocks above the salt. Buoyancy helps, but in the upper kilometre the sediments are on average lighter than salt. The salt pierces a strong roof where stretching or erosion has weakened it, or where the roof is thin.
4
A layer 20 m thick is bent at the crest of a fold with a radius of curvature of 5 km. How much is the top of the layer stretched without slip along layers, and with slip along the layer boundaries every 2 m?
Show the answer
Without slip, $e = (20/2)/5000 = 0.002$ , that is, 0.2%. With slip, each layer 2 m thick bends on its own: $e = (2/2)/5000 = 0.0002$ , that is, 0.02%. This is 10 times less, so a stack of thin slipping layers fractures less than a thick single layer.
5
What is a salt weld and why does it matter for exploration?
1. A.The part of the top of a dome that has been eroded at the surface
2. B.A fracture in the salt healed by calcite
3. C.A place where the salt has been squeezed out completely and the suprasalt rocks rest on the subsalt ones: a path for crossflow from the subsalt beds into the suprasalt ones
4. D.A sandstone interbed within the salt sequence
Show the answer
Correct answer: C. A place where the salt has been squeezed out completely and the suprasalt rocks rest on the subsalt ones: a path for crossflow from the subsalt beds into the suprasalt ones
Under minibasins the salt can be squeezed out completely. There the seal of the subsalt tier is broken: hydrocarbons can leave the subsalt traps and fill the suprasalt ones. This happened in the Pre-Caspian Basin, where the oil of the suprasalt accumulations came from the subsalt source rocks.
6
Salt with an original thickness of 4 km (assumed) has gathered into domes that occupy 40% of the area, and beneath the minibasins it has been squeezed out completely. What has the average height of the salt in the domes become, and what does it mean if the subsalt floor lies at a depth of about 10 km?
Show the answer
The volume of salt stays the same, so the height is $4/0.4 = 10$ km. Domes of this height reach almost to the surface. This is what happens in basins with thick salt: some domes reach the surface, and the salt near their tops dissolves and spreads. The calculation is rough. It ignores dissolution, erosion and the spreading of salt into sheets.

Updated: 19 October 2026