Perched and tilted contacts, trapped water

Quick refresher

In an accumulation in equilibrium, all the wells see one oil line and one water line, and the free water level (FWL) is the same. If the oil line is common and the water lines differ, each well has its own FWL. For each extra kilopascal of water pressure, the FWL is higher by $1/(G_w - G_o)$ metres.
Perched water is shut in under the accumulation, in a closed volume. Its FWL lies above the general one. Usually it is at the depth of the lowest point through which the block is connected to its neighbour. This is the base of a window in a fault (a part of the fault where the formations of the two blocks face each other). Or it is the base of the formation on a saddle (a low point between two domes, two raised parts of the structure). The pressure of perched water is higher than that of the main aquifer by $(G_w - G_o)\,\Delta z$ .
Trapped water means closed bodies of water inside the accumulation. In them a log sees water among the oil, and their pressure does not fall on the oil line. This is not a contact of the accumulation.
When formation water flows, its datum pressure (the pressure converted to a common reference depth) changes over the area. The contact goes down in the direction of the flow, $\rho_w/(\rho_w - \rho_o)$ times more steeply than the potentiometric surface (the surface formed by the levels to which water would rise in wells). The direction and size of the gradient come from the datum water pressures in at least three wells that do not lie on one line.
We tell the cases apart from a combination of signs. Does the FWL equal the depth of the window or the saddle? What head gradient would a flow need (the head is the height to which water would rise in a well), and does such a gradient exist in the area? Does the contact change smoothly from well to well? Is the water in the samples the same?
The accuracy of comparing different wells is about 15–18 kPa: the absolute accuracy of the gauges and the depth matching. A difference of datum water pressures of less than two such errors proves nothing.

Key relationships

z^*_{\text{B}} - z^*_{\text{A}} = \dfrac{\Phi_{w,\text{A}} - \Phi_{w,\text{B}}}{G_w - G_o}

shift of the FWL between wells with a common oil line

\Phi_w = p_w + G_w\,(z_{\text{ref}} - z)

water pressure reduced to the datum $z_{\text{ref}}$ ; the difference of heads is $\Delta\Phi_w/G_w$

\Delta p_{\text{perch}} = (G_w - G_o)\,(z^*_{\text{main}} - z^*_{\text{perch}})

excess of the pressure of perched water over the main aquifer

\tan\beta = \dfrac{|\nabla\Phi_w|}{G_w - G_o} = \dfrac{\rho_w}{\rho_w - \rho_o}\,|\nabla h|

tilt of the contact in a flow of formation water

Common pitfalls

Explaining different FWLs of wells by separate blocks without checking whether their oil line is common.
Taking a difference of water levels at the level of the error of comparing wells for a tilt of the contact.
Looking for the direction of the tilt from two wells, or from three that lie almost on one straight line.
Reducing the water pressures of different wells to one depth with one density, when the salinity of the water changes over the area.
Drawing the contact of a whole accumulation along a trapped lens of water.

After this lesson you can

explain why, with a common oil line, the water level of each well is set by its water pressure, and convert a difference of datum water pressures into a difference of depth of the free water level
describe how perched and trapped water form, and recognise them from pressure, logs and the structure of the block
find from the datum water pressures in three or more wells the slope of the potentiometric surface and the tilt of the contact, with their errors
tell perched water and a tilted contact from a capillary shift, separate fluid systems and measurement errors
estimate what each picture means for the volume of an accumulation, the placing of wells and edge water support

Common oil, different water

In an accumulation that is not being produced, the fluids have come to equilibrium over geological time. Suppose the formation water is at rest and connected over the whole area. Then all the wells see one oil line and one water line, and their crossing is one free water level for the accumulation (contact from the crossing of lines). This rule has a condition: the water must be one body, and at rest. If it is divided into closed volumes, or if it flows, its pressure at the same depth is different in different wells, and each well has its own FWL.

We compare the pressures of different wells after converting them to a common datum depth $z_{\text{ref}}$ with the gradient of their own fluid. This is the datum pressure, or potential: $\Phi = p + G\,(z_{\text{ref}} - z)$ , where $G = \rho g$ and $z$ is the true vertical depth below sea level (depth reference systems). In a fluid phase that is at rest and connected, the datum pressure is the same at any depth and in any well. A fluid flows from a higher potential to a lower one (Pressure measurements in producing reservoirs: barriers and accuracy). For water, the difference of datum pressures divided by its gradient is the difference of hydraulic heads: the heights to which water would rise in the wells. The heads, plotted over the area, form the potentiometric surface.

Free water level with a common oil line

z^* = z_{\text{ref}} + \frac{\Phi_o - \Phi_w}{G_w - G_o},\qquad z^*_{\text{B}} - z^*_{\text{A}} = \frac{\Phi_{w,\text{A}} - \Phi_{w,\text{B}}}{G_w - G_o}

where:

$z^*,\ z^*_{\text{A}},\ z^*_{\text{B}}$ FWL: the depth where the oil line crosses the water line; in wells A and B, m
$z_{\text{ref}}$ depth of the datum, m
$\Phi_o$ oil pressure reduced to $z_{\text{ref}}$ with the oil gradient; common for connected oil, MPa
$\Phi_w,\ \Phi_{w,\text{A}},\ \Phi_{w,\text{B}}$ water pressure reduced to $z_{\text{ref}}$ with the water gradient; in wells A and B, MPa
$G_o,\ G_w$ pressure gradients of oil and water, $\rho g$ , MPa/m

Units: any consistent system (MPa, MPa/m and m)

Holds when

The oil is at rest and connected between the wells, so its datum pressure is one. At the contact in each well the water is continuous.
The densities of the phases are constant. For gas, the gas gradient is taken instead of $G_o$ .
The depths are vertical, from one level. The pressures are absolute, from calibrated gauges.

Each kilopascal by which the water in well B is “stronger” (at a higher pressure) than in well A raises its FWL by $1/(G_w - G_o)$ metres: for light oil and formation water about 0.3 m, and for heavy oil 0.6 m or more. The formula does not say why the water is stronger: whether it is shut in separately or flowing. Other signs decide that.

Different water levels with a common oil have several causes. We check them in order, starting with the most frequent and the cheapest to check.

Why the FWL differs between wells

Cause: Errors of depths and gauges
Pressure lines: the oil and water lines of a well are shifted equally
FWL: by its own lines it is the same, to the accuracy of the depth matching
Check: calibration, depth matching (Forced-gradient technique and multi-well analysis)

Cause: A separate system of oil
Pressure lines: the oil line is shifted, the water is common
FWL: the block has its own
Check: separate systems

Cause: Perched water
Pressure lines: the oil is common, the water of the block is to the right
FWL: higher than the general one, as a step
Check: structure of the block, head gradient in the area

Cause: Flow of formation water
Pressure lines: the oil is common, the water pressure changes smoothly over the area
FWL: tilted along the flow
Check: water pressures in three or more wells

Cause: Depletion of the common aquifer
Pressure lines: the water pressure is below the initial one near a producing neighbour
FWL: changes with time
Check: dates of measurements, production of neighbours

Cause	Pressure lines	FWL	Check
Errors of depths and gauges	the oil and water lines of a well are shifted equally	by its own lines it is the same, to the accuracy of the depth matching	calibration, depth matching (Forced-gradient technique and multi-well analysis)
A separate system of oil	the oil line is shifted, the water is common	the block has its own	separate systems
Perched water	the oil is common, the water of the block is to the right	higher than the general one, as a step	structure of the block, head gradient in the area
Flow of formation water	the oil is common, the water pressure changes smoothly over the area	tilted along the flow	water pressures in three or more wells
Depletion of the common aquifer	the water pressure is below the initial one near a producing neighbour	changes with time	dates of measurements, production of neighbours

If it is not the FWL that differs but the oil–water contact (OWC) from logs, with a common FWL, the cause is the capillary properties of the rocks (FWL and OWC from logs). If a log sees oil below the FWL, it is a paleo-contact (a trace of an old oil–water boundary).

Perched water: a closed volume of water

A trap fills from the top downwards. The oil comes along the formation and displaces water, and the water must leave downwards along the formation, into the water-bearing part. If the path of the water is blocked, the oil cannot go further. This is how perched water arises. Take two neighbouring blocks separated by a fault. The formation of block B meets the formation of block A across the fault only over part of its height, in the juxtaposition window (the part of the fault where layers of the two blocks face each other). Above and below the window, the formation of one block faces the shales of the other (sealing by faults). On its other sides the water-bearing part of block B is closed too: by its base, by the pinching out of the formation (it gets thinner and ends), or by another fault.

While the FWL of block A is above the base of the window, the water of block B leaves through the lower part of the window, and the level in both blocks is common. When the oil in block A goes below the base of the window, the whole window is filled with oil. Almost no water flows through an oil-bearing formation, because its relative permeability to water there is close to zero (Relative permeability). The water of block B is shut in. Its FWL stops at the depth of the base of the window, and the FWL of block A goes deeper as the trap fills. A saddle between two domes works in the same way, if the formation on it is thin. The FWL of the farther dome stays at the base of the formation on the saddle, which is the deepest point through which water could still pass.

SchematicPerched water behind a window in a fault (schematic, not to scale). The oil of the blocks is connected through the window, and the water of block B is shut in below the base of the window. The FWL of block A, which is open to the regional aquifer, is 32 m deeper.

The oil of the blocks is one system: it is connected through the window, and its pressure is common. But the water of block B is in balance with the oil at its own FWL. So its pressure is higher than the pressure of the main aquifer at the same depth, by the amount that the oil column between the two FWLs creates:

Excess of the pressure of perched water over the main aquifer

\Delta p_{\text{perch}} = (G_w - G_o)\,(z^*_{\text{main}} - z^*_{\text{perch}})

where:

$\Delta p_{\text{perch}}$ by how much the pressure of perched water is above the pressure of the water of the main formation at the same depth, MPa
$z^*_{\text{main}}$ FWL of the block that is open to the main aquifer, m
$z^*_{\text{perch}}$ the perched FWL, m
$G_w,\ G_o$ gradients of water and oil, MPa/m

Units: any consistent system

Holds when

The water of the block is closed on all sides, and since it was separated it has not leaked through seals and faults.
The densities of the phases in the blocks are the same.

Perched water is always “stronger” (at a higher pressure) than the main water: on a pressure plot its line lies to the right. The deeper the FWL of the main block went down after the water became separated, the greater the excess. The shut-in water holds it until it seeps out through the surrounding rocks. So perched water is a sign that the water-bearing part of the block is well isolated.

It is most convenient to look at such data on an excess pressure plot over the common oil line (excess pressure plot). The oil points of both wells lie on the vertical line of zero. The water-bearing points lie on two parallel straight lines with the slope $G_w - G_o$ , and each of them comes to zero at its own FWL.

Pressure lines of wells A and B from pretests

Well, zone: A, oil
Points: 5
Gradient, kPa/m: 7.052
Scatter, kPa: 0.9
Pressure, MPa: 26.6997 at 2530 m

Well, zone: A, water
Points: 4
Gradient, kPa/m: 10.496
Scatter, kPa: 0.9
Pressure, MPa: 26.7000 at 2530 m

Well, zone: B, oil
Points: 4
Gradient, kPa/m: 7.054
Scatter, kPa: 0.9
Pressure, MPa: 26.4919 at 2500 m

Well, zone: B, water
Points: 4
Gradient, kPa/m: 10.504
Scatter, kPa: 0.8
Pressure, MPa: 26.5003 at 2500 m

Well, zone	Points	Gradient, kPa/m	Scatter, kPa	Pressure, MPa
A, oil	5	7.052	0.9	26.6997 at 2530 m
A, water	4	10.496	0.9	26.7000 at 2530 m
B, oil	4	7.054	0.9	26.4919 at 2500 m
B, water	4	10.504	0.8	26.5003 at 2500 m

The lines are fitted by least squares (a line from pretests). The pressure is given at a reference depth. All the pretests are valid, the pressures are absolute, and the depths are vertical, below sea level. The data are assumed for illustration.

Worked examplePerched water in a block behind a fault

Given (assumed data)

Lines of wells A and B: from pretests (the table above)
Fluids: oil 0.72 g/cm³ from a sample, water 1.07 g/cm³ from analyses in both wells: $G_o = 7.063$ , $G_w = 10.497$ kPa/m
Accuracy of comparing wells: in each well a gauge of ±7 kPa and a depth of ±1 m (one standard error)
OWC from the resistivity log: A is 2528.7 m, B is 2496.4 m
Structure: a window in the fault between the blocks, down to a depth of 2498 ± 4 m; 2.4 km between the wells; the head gradient of the regional aquifer from the wells of the area is not more than 0.3 m/km

Solution

The oil lines at a depth of 2490 m: A is 26.4176 and B is 26.4214 MPa, and B is higher by 3.8 kPa. The accuracy of comparing the oil lines of two wells is $\sqrt{2\cdot 7^2 + 2\cdot(7.063\cdot 1)^2} = 14.1$ kPa. The oil in the blocks is the same.
The water lines at a depth of 2520 m: A is 26.5950 and B is 26.7104 MPa, and B is higher by 115.4 kPa. The accuracy is $\sqrt{2\cdot 7^2 + 2\cdot(10.497\cdot 1)^2} = 17.8$ kPa, so the difference is 6.5 times the error. The water in the blocks is different.
The FWL from the lines of each well itself. A: $2530 + (26.6997 - 26.7000)\cdot 1000/3.444 = 2529.9$ m. B: $2500 + (26.4919 - 26.5003)\cdot 1000/3.450 = 2497.6$ m. The FWL of block B is higher by 32.3 m. A gauge error is common to both lines of a well and does not affect its FWL. An error in the depth matching moves the FWL together with the log of the same well, and into the difference of the FWLs of two wells it brings only ±1.4 m.
The distance from the FWL to the OWC from logs: A is 1.2 m, and B is 1.2 m. The capillary picture is the same, so the shift is not capillary.
Could such a shift come from a flow of water? A difference of FWL of 32.3 m corresponds to a difference of datum water pressures $3.434\cdot 32.3 = 110.9$ kPa, which is $110.9/10.497 = 10.56$ m of head. Over 2.4 km this is 4.4 m/km, which is 15 times the regional gradient. Besides, a flow would tilt the contact smoothly, not as a step between blocks.
The FWL of block B (2497.6 m) coincides with the base of the window (2498 ± 4 m) within its error.

Answer: The oil of the blocks is one system, and the water of block B is shut in. Its perched FWL is 2497.6 m, which is 32.3 m higher than in block A (2529.9 m), at the depth of the base of the window in the fault.

What it means. The hydrocarbons in place of block B are calculated with its own contact. In it the interval from 2498 to 2530 m is water-bearing. So the volume of the accumulation of the block, from the map of its top (the top of the formation), is smaller than with a common contact (Gross rock volume from maps). A well placed in block B to a depth of 2510–2525 m, counting on oil, would find water. For development something else matters too: under the oil of block B there is a small closed volume of water. There is no edge water support there. The pressure of the block will be held only by the elastic expansion of the fluids (they expand as the pressure falls) and by the flow of oil through the window (Lumped multi-tank material balance model). Water samples give independent arguments: water that was separated long ago from the regional water may differ in salinity. The data in this example are assumed for illustration.

CalculatedExcess pressure over the oil line of well A for both wells of the example. The oil falls on the vertical line of zero, and the water on two parallel lines that come to zero at their own FWLs. The band is the depth of the base of the window in the fault, with its error.

Trapped water inside an accumulation

Water stays inside an accumulation wherever it had no way out when the oil came. This happens in sand lenses (small lens-shaped bodies of sand) surrounded on all sides by shales, in blocks bounded by sealing faults that the oil did not enter at all, and in closed depressions between domes. This water is called trapped water. It must not be confused with the residual water in the pores. That water is immobile and spread through all the oil-bearing rock (Fluid saturation), while here we mean a whole volume of mobile water.

On a log, trapped water shows as a water-bearing formation among oil-bearing ones: a low resistivity at ordinary porosity, with oil above and below. The danger is to take the top of such a formation for the OWC and to “cut off” the lower part of the accumulation, or to perforate water. The pressure gives the lens away: its pretest does not fall on the oil line. A lens that closed before the oil arrived kept the water pressure of that time, and it usually lies on the regional water line. In the accumulation of the example above, such a lens at 2480 m would show about 26.175 MPa. This is 172 kPa below the oil pressure at the same depth: exactly the excess of the oil pressure over the water pressure 50 m above the FWL of block A.

In calculating hydrocarbons in place, we subtract the trapped bodies from the oil-bearing volume, but we do not draw the contact of the accumulation along them: the oil below a lens goes on to the common FWL. In development, a lens that is penetrated gives a short inflow of water without support, because its closed volume is quickly depleted. The oil around it behaves as usual.

A tilted contact: formation water flows

Water in a sedimentary basin (a large area where thick layers of sediment have piled up) is rarely entirely at rest. Two causes make it flow. The first is compaction (the pressing together of sediments under the weight of the layers above). Sinking sediments squeeze out pore water, and it moves to where the pressure is lower, towards the edges of the basin and up the formations. The second is the relief of the land: a formation that comes to the surface in mountains is filled with water up to a high level and is refilled by rain and snow. The water runs down it to the lowlands, where the formation reaches the surface lower down, or where water can seep upwards through thin seals (Abnormal formation pressures: origin, prediction and recognition). In different periods of the life of a basin, one cause or the other dominates. The speeds of such a flow are centimetres and metres a year. But the water pressure changes over the area, and this is enough to tilt a contact.

The oil in a trap is at rest, and its datum pressure is the same everywhere. But the water pressure beneath it changes from well to well with the head. By the formula for the FWL, the contact goes down in the direction in which the water flows. This gives a tilted contact. The derivation of the formula for the tilt, and an example with a given slope of the potentiometric surface, are in the topic on traps. Here the more important question is how to find this slope from pressure measurements.

Tilt of the contact from the gradient of the datum water pressure

\tan\beta = \frac{|\nabla\Phi_w|}{G_w - G_o} = \frac{\rho_w}{\rho_w - \rho_o}\,|\nabla h|

where:

$\beta$ tilt angle of the contact; in practice $\tan\beta$ is given in metres per kilometre, m/km
$\nabla\Phi_w$ areal gradient of the datum pressure of water: its change per unit horizontal distance, in the direction of the greatest change, kPa/km
$\nabla h$ gradient of the potentiometric surface (of the heads), m/km
$G_w,\ G_o$ pressure gradients of water and oil, kPa/m
$\rho_w,\ \rho_o$ densities of water and oil in the formation, g/cm³

Units: a gradient in kPa/km and pressure gradients in kPa/m give $\tan\beta$ in m/km

Holds when

The oil is at rest, the flow of water is steady, and the densities are constant.
The datum water pressures are compared on one datum and at one water density (otherwise, see the advanced section).
Capillary forces are ignored: this is about the FWL. The OWC from logs is parallel to it, at the height $h_d$ (transition zone).

The contact goes down in the direction of the flow, to where the datum water pressure is lower, and it is $\rho_w/(\rho_w - \rho_o)$ times steeper than the potentiometric surface: for oil usually 3–10 times. A gas–water contact tilts less, with a factor of 1.2–1.3 for gas. A gas–oil contact is horizontal, because both phases are at rest.

The gradient $\nabla\Phi_w$ is a vector: it has a size and a direction. Two wells give only its component along the line that joins them. We need three wells that do not lie on one straight line: a single plane $\Phi_w(x, y) = \Phi_{w1} + b\,x + c\,y$ passes through three points. Wells that penetrated only water are just as useful here as oil wells, because only the datum water pressure matters. The accuracy grows with the distance between the wells and with the area of the triangle. Over a few kilometres the difference of water pressures is tens of kilopascals, and the error of each well is about 12 kPa. If the wells stand almost on one line, the direction of the gradient across that line stays unknown.

Wells of the field in the example

Well: 1
$x$ ; $y$ , km: 0.0; 0.0
Top, m: 1985.0
Penetrated: oil and water
Water pressure, MPa: 21.5710 at 2060 m

Well: 2
$x$ ; $y$ , km: 0.5; 4.5
Top, m: 2086.9
Penetrated: water only
Water pressure, MPa: 21.9810 at 2100 m

Well: 3
$x$ ; $y$ , km: 4.5; −1.0
Top, m: 2040.6
Penetrated: oil and water
Water pressure, MPa: 21.8892 at 2100 m

Well	$x$ ; $y$ , km	Top, m	Penetrated	Water pressure, MPa
1	0.0; 0.0	1985.0	oil and water	21.5710 at 2060 m
2	0.5; 4.5	2086.9	water only	21.9810 at 2100 m
3	4.5; −1.0	2040.6	oil and water	21.8892 at 2100 m

The $x$ axis points east and the $y$ axis north, from the crest. The pressure is on the water line at the depth shown; the lines are from the pretests of each well. The depths are vertical, below sea level, and the pressures are absolute. The measurements were made before production started in the area. The data are assumed for illustration.

Worked exampleTilt of the contact from three wells

Given (assumed data)

Wells: the table above; wells 1 and 3 penetrated oil and water, and well 2, on the northern flank, only water
Fluids: oil 0.82 and water 1.05 g/cm³ from samples: $G_o = 8.04$ , $G_w = 10.30$ kPa/m, $G_w - G_o = 2.26$ kPa/m
FWL from each well's own lines: well 1 is 2045.0 m, well 3 is 2087.0 m; the oil lines of the wells coincide within the accuracy of comparison
Accuracy: gauge ±7 kPa and depth ±1 m in each well
Datum: 2000 m below sea level
Planned wells: well 4 on the eastern flank (6.0; 0.4), top 2075.8 m; well 5 on the western flank (−4.0; 0.4), top 2025.8 m

Solution

Datum pressures of water $\Phi_w = p - 0.01030\,(z - 2000)$ : well 1 gives $21.5710 - 0.01030\cdot 60 = 20.9530$ MPa, well 2 gives $21.9810 - 0.01030\cdot 100 = 20.9510$ MPa, and well 3 gives 20.8592 MPa. Relative to well 1: $\Delta\Phi_2 = -2.0$ and $\Delta\Phi_3 = -93.8$ kPa.
A check: by the formula for the FWL, well 3 must have its FWL deeper than well 1 by $93.8/2.26 = 41.5$ m. By their own lines it is deeper by 42.0 m. Agreement within the accuracy of comparing wells confirms that they have the same oil.
A plane through three wells: $0.5\,b + 4.5\,c = -2.0$ and $4.5\,b - 1.0\,c = -93.8$ . Hence $b = (-2.0 + 4.5\cdot(-93.8))/20.75 = -20.44$ and $c = 4.5\,b - (-93.8) = 1.82$ kPa/km.
The gradient is $|\nabla\Phi_w| = \sqrt{b^2 + c^2} = 20.52$ kPa/km. Water flows to where $\Phi_w$ is lower, along the azimuth 95° (measured from north), that is, to the east. The head gradient is $20.52/10.30 = 1.99$ m/km, and the tilt of the contact is $20.52/2.26 = 9.1$ m/km.
The error. Each $\Phi_w$ is known with an accuracy of $\sqrt{7^2 + 10.30^2} = 12.5$ kPa. Carrying these errors through the formulas of step 3 gives $20.52 \pm 4.1$ kPa/km, a tilt of 9.1 ± 1.8 m/km, and a direction of ± 11°. The tilt differs reliably from zero.
The contact in the planned wells relative to well 1: $\Delta z = -(b\,x + c\,y)/2.26$ . Well 4: $-(-20.44\cdot 6.0 + 1.82\cdot 0.4)/2.26 = 53.9$ m, and the contact is 2098.9 ± 12 m, with the top at 2075.8 m. Well 5: $-(-20.44\cdot(-4.0) + 1.82\cdot 0.4)/2.26 = -36.5$ m, and the contact is 2008.5 ± 7 m, with the top at 2025.8 m.

Answer: The water flows to the east (azimuth about 95°), and the contact is tilted by 9.1 ± 1.8 m/km. In well 4 we expect oil: the contact is at about 2098.9 m, 23 m below the top, although a horizontal contact at 2045.0 m promised water there. In well 5, on the contrary, the contact is 17 m above the top. This is 2.4 times the error of the forecast, so there is almost certainly no oil, although a horizontal contact would have given 19.2 m of oil.

What it means. A tilt of a few metres per kilometre cannot be noticed in one well, but on the flanks of an accumulation it moves the contact by tens of metres. The error of the forecast grows with the distance from the triangle of wells. Both planned wells lie outside it, and the contact of well 4, which is the farthest from well 1, is known only to ±12 m: oil is likely there, but not guaranteed. These ±12 and ±7 m allow only for the errors of the gauges and depths in three wells. An error in the densities (±0.01 g/cm³ for each phase) adds about 6% to the shift of the contact. An error of the depth of the top from the map at the planned point enters the height of the accumulation in full. With it, the margin of well 5 is less than 2.4 times the error. The picture is refined by datum water pressures in any wells of the same aquifer, including those outside the accumulation, and by a map of heads for the area. In well 4 itself, its own pressure lines will give the contact. A plane is a first approximation: over large distances the potentiometric surface bends. The data in this example are assumed for illustration.

How to tell them apart: pressure, logs, water, structure

No single sign settles the matter. The combination of signs does. The table brings together the main signs of the three pictures that are most often confused.

Perched water, water flow and capillary shift

Sign: FWL from pressure
Perched water: different, as a step between blocks
Water flow: changes smoothly over the area
Capillary shift: common

Sign: OWC from logs
Perched water: follows its own FWL
Water flow: tilted together with the FWL
Capillary shift: changes with the quality of the rocks

Sign: Water lines
Perched water: the water of the block is to the right of the main water
Water flow: the water pressure changes over the area
Capillary shift: common

Sign: Head gradient needed
Perched water: much larger than the regional one
Water flow: agrees with the regional one
Capillary shift: not needed

Sign: Structure
Perched water: a window, a saddle, a closed block
Water flow: recharge and discharge areas (where water enters and leaves the formation)
Capillary shift: change of rock, shale content

Sign	Perched water	Water flow	Capillary shift
FWL from pressure	different, as a step between blocks	changes smoothly over the area	common
OWC from logs	follows its own FWL	tilted together with the FWL	changes with the quality of the rocks
Water lines	the water of the block is to the right of the main water	the water pressure changes over the area	common
Head gradient needed	much larger than the regional one	agrees with the regional one	not needed
Structure	a window, a saddle, a closed block	recharge and discharge areas (where water enters and leaves the formation)	change of rock, shale content

Errors of depths and gauges shift both lines of a well equally. A gauge error does not change the FWL by its own lines, and a depth error moves it together with the log of the same well. A separate system of oil gives a shifted oil line (separate systems or an error).

Logs add their own signs of water flow. Suppose the flow began after the trap was filled. On the flank from which the water flows, the contact rose, and below it residual oil was left in the rock. This is a paleo-contact, which a log sees but pressure does not. On the opposite flank the oil moved down, and no traces were left. Water from the recharge area is usually fresher. On this flank its resistivity $R_w$ changes, and the water saturation from logs is calculated with different $R_w$ over the area (Determining formation water resistivity). The oil at the contact on such a flank is often heavier and more viscous: flowing water carries away the light components and brings bacteria that degrade the oil.

A tilt without flow. A contact can be tilted in water at rest too, if equilibrium has not yet been reached. Late tectonic movements (movements of the earth's crust) tilted the trap, and in a tight formation oil and water move so slowly that they have not had time to level the boundary. In good reservoirs this takes a geologically short time. So this explanation is considered for low-permeability formations, and it is checked against the history of the structure.

Order of checking different water levels

1
One reference system
Vertical depths below sea level from the directional survey, and absolute pressures. The measurements were made before production began in the area, or they are brought to one date.
2
Oil
Is the oil line common, within the accuracy of comparing wells? If not, these are separate systems.
3
Significance
We compare the difference of datum water pressures with the accuracy of comparing wells. Less than two errors is not an argument.
4
Capillary effect
The distance from the FWL to the OWC from logs in each well. It is the same for the same rocks.
5
Cause
The structure of the block (windows, saddles, closed volumes) and the head gradient in the area: perched water or flow. Water samples: salinity.
6
Development
The pressure response of the blocks to production, and the water influx, test the explanation that was chosen.

What changes for hydrocarbons in place and development

Volume. Perched water divides an accumulation into parts with their own contacts, and each is calculated from its own “area–depth” curve (Gross rock volume from maps). A tilted contact is not described by one such curve. We build a map of the height of the accumulation, the difference of the depths of the contact and the top, and we sum it over the area. In the dome of the example, a horizontal contact at the level of the FWL of well 1 would give 1.60 billion m³ of gross rock volume (the volume of rock in the accumulation). A tilted contact through the same point gives 29% more, and the centre of the accumulation is shifted by 1.8 km along the flow (advanced section).

CalculatedThe dome of the example on a map. Shown are the structure contours of the top; the outline of the accumulation with a horizontal contact at the level of the FWL of well 1 (dashed), and with a tilted contact (solid line, shaded); the lines of equal head; and the direction of water flow. Filled circles are drilled wells, and empty ones are planned.

Wells. On the flank towards which the water flows, the oil goes deeper than a horizontal contact promises, and on the opposite flank the water stands higher. The perforation intervals, and the distance to the contact on which the coning of water (the drawing up of water towards the well; Water and gas coning) depends, are counted from the local contact, not from one level for the whole accumulation.

Reservoir energy. Perched water is a small closed volume. There is no edge water support in such a block. The pressure in it is held only by elastic expansion (the fluids expand as the pressure falls) and by the flow of oil from the neighbouring block, and often injection is needed. A tilt of the contact, on the contrary, speaks of an extended and permeable aquifer that is connected to the recharge area. These are the conditions for a water drive (the push of water that moves the oil to the wells; water drive).

Gas cap and oil rim. In a flow of water, the gas–oil contact stays horizontal, because the oil and gas are at rest, while the oil–water contact is tilted. So the thickness of an oil rim (the oil layer between the gas cap and the water) changes over the area. On the flank from which the water flows it is thinner and can disappear, and there the gas lies directly on the water.

Contacts change with time. Production from a common aquifer at a neighbouring field lowers the water pressure towards that field. A gradient appears that did not exist before development, and the water levels in wells drilled later diverge. So we compare measurements of different years only after bringing them to one date (Pressure measurements in producing reservoirs: barriers and accuracy), and we first check a difference of water levels in new wells for depletion of the aquifer.

Terms and symbols

Perched contact, perched water: A contact above the general level of the accumulation, and the water under it, shut in a closed volume. The pressure of perched water is higher than that of the main aquifer at the same depth.
Trapped water: Closed bodies of mobile water inside the accumulation: lenses, blocks and depressions that the oil did not enter. It is not the same as the residual water in the pores.
Juxtaposition window: A section of a fault where the formation of one block meets the formation of the other. The blocks are connected through it.
Datum pressure $\Phi$: The pressure converted to the datum with the gradient of its own fluid. In a phase that is at rest and connected it is the same everywhere.
Potentiometric surface, head: The levels to which water would rise in wells that penetrate the formation. Its slope gives the direction and strength of the water flow.
Tilted contact: A contact that goes down in the direction of the flow of formation water, $\rho_w/(\rho_w - \rho_o)$ times more steeply than the potentiometric surface.

Common mistakes and misconceptions

Wrong

“The FWL in two wells differs by 30 m, so these are different accumulations.”

In fact

First compare the oil lines. If they coincide, the oil is one, and it is the water that differs: perched or flowing. A separate accumulation is when the oil line is shifted.

Wrong

“The datum water pressures of two wells differ by 15 kPa, so water flows and the contact is tilted.”

In fact

The accuracy of comparing two wells is about 15–18 kPa. A difference of less than two such errors proves neither a flow nor perched water. We need other wells, a check of depths and the calibration of gauges.

Wrong

“Perched water will support the pressure in the block like edge water.”

In fact

It is a small closed volume with no recharge, and its elastic reserve is small. The pressure of the block is held by the flow of oil from the neighbouring block, or by injection.

Wrong

“Since the water flows, the gas–oil contact is tilted in the same way as the oil–water contact.”

In fact

Oil and gas are at rest, and the boundary between them is horizontal. Only the contacts with flowing water tilt: the oil–water contact $\rho_w/(\rho_w - \rho_o)$ times more steeply than the heads, and the gas–water contact only 1.2–1.3 times.

Wrong

“A water-bearing formation was penetrated under the oil: this is the contact, and there is no oil below.”

In fact

It may be a trapped lens. Its pressure does not fall on the oil line, and the oil below it goes on to the common FWL. We check with pretests below the lens.

In depth: advanced detail

A tilt in a dome: the shift of the accumulation, and the volume

For a dome with a top in the shape of an elliptic paraboloid (a smooth bowl turned upside down, with oval contours), a tilted contact can be allowed for by a formula. Let the top be $z_{\text{top}} = z_{\text{crest}} + a_x x^2 + a_y y^2$ , and the contact a plane $z^* = z_1 + t_x x + t_y y$ that passes through a level $z_1$ above the crest. The height of the accumulation, the difference of the depths of the contact and the top, is again a paraboloid, only shifted:

Height of the accumulation in a dome with a tilted contact

\begin{aligned} h(x, y) &= H_{\text{eff}} - a_x\,(x - x_0)^2 - a_y\,(y - y_0)^2, \\ x_0 &= \frac{t_x}{2a_x},\quad y_0 = \frac{t_y}{2a_y},\quad H_{\text{eff}} = H + \frac{t_x^2}{4a_x} + \frac{t_y^2}{4a_y}, \\ V &= \frac{\pi\,H_{\text{eff}}^2}{2\sqrt{a_x a_y}} \end{aligned}

where:

$h(x, y)$ height of the accumulation from the top to the contact; the accumulation is where $h > 0$ , m
$a_x,\ a_y$ curvature of the top along the axes, m/km²
$t_x,\ t_y$ components of the tilt of the contact, positive if the contact goes down along the axis, m/km
$H$ height of the accumulation at the crest, $z_1 - z_{\text{crest}}$ , m
$x_0,\ y_0$ shift of the centre of the accumulation from the crest, km
$H_{\text{eff}}$ greatest height of the accumulation, at its shifted centre, m
$V$ rock volume of the accumulation, km²·m = million m³, million m³

Units: coordinates in km, depths and heights in m

Holds when

The top is an elliptic paraboloid with axes along $x$ and $y$ . The formation is thicker than the accumulation (there is bottom water), so the rock volume is the volume between the top and the contact.
The contact is a plane. The trap does not spill: the outline of the accumulation does not reach the spill point (the lowest point through which oil can leave).

In such a dome, a tilt does not change the shape of the accumulation. It shifts it along the flow by $t/(2a)$ and makes it taller. The shift grows with the tilt and with how gentle the dome is. In a gentle structure even a moderate flow moves the accumulation to the flank.

Worked exampleShift and volume of the accumulation in the dome of the example

Given (assumed data)

Dome: crest 1985 m, $a_x = 2.5$ and $a_y = 5$ m/km²; the formation is thicker than the accumulation
Contact: through the FWL of well 1 at the crest (2045.0 m), with the tilt from the example of three wells: $b$ and $c$ are the gradient of the datum water pressure

Solution

Components of the tilt: $t_x = -b/2.26 = 9.04$ and $t_y = -c/2.26 = -0.81$ m/km. The contact goes down to the east and very slightly to the south.
Shift of the centre: $x_0 = 9.04/5 = 1.81$ km, $y_0 = -0.81/10 = -0.08$ km.
Height: $H = 2045.0 - 1985 = 60.0$ m, and $H_{\text{eff}} = 60.0 + 9.04^2/10 + (-0.81)^2/20 = 68.2$ m.
Rock volume: with a horizontal contact $\pi\cdot 60.0^2/(2\sqrt{12.5}) = 1599$ million m³, and with a tilted one $\pi\cdot 68.2^2/(2\sqrt{12.5}) = 2066$ million m³, which is 29% more.
The edges of the accumulation along the $x$ axis: with a horizontal contact $\pm\sqrt{60.0/2.5} = \pm 4.9$ km, and with a tilted one $1.81 \pm \sqrt{68.2/2.5}$ , that is, from −3.4 to 7.0 km.

Answer: The accumulation is shifted by 1.81 km to the east, and the rock volume is greater by 29%.

What it means. The position of the accumulation changed more than its volume. The main error of a horizontal contact is not in the volume but in where to look for it. The western edge of the accumulation is in fact 3.4 km from the crest instead of 4.9, and the eastern one is 7.0 km away. If well 1 had stood not at the crest but on a flank, an estimate from its contact would have been wrong in the volume too: a horizontal plane through its level would cut off the accumulation in a different way. The data in this example are assumed for illustration.

Water of different density: which pressures to compare

The formulas of the lesson compare datum pressures at one water density. Suppose the salinity changes over the area, for example the water is fresher towards the recharge area. Then reducing pressures to the datum with a “foreign” density creates a false difference. Horizontal flow is set by the pressure difference at one depth. So the pressure of each well is reduced to the datum with the density of its own water, and only then are they compared. The error is greater the farther the measurement is from the datum: with a density difference of 0.02 g/cm³ and a distance to the datum of 300 m, it is $9.81\cdot 0.02\cdot 300 \approx 59$ kPa. This is more than five metres of head, as much as the whole gradient over three kilometres in the example of the lesson.

This gives two rules. The datum is chosen in the middle of the accumulation, closer to the measurements. Regional data, which are tests of the aquifer in wells outside the accumulation, made at different depths, are used with the water density from the samples of each well. If the water changes noticeably over the area, the gradient is calculated as a difference of pressures at one depth, without a transition to heads.

Connections to other topics

Pressure gradients, fluid contacts and capillary equilibrium — FWL from the crossing of lines, capillary shift of the OWC, separate fluid systems.
Forced-gradient technique and multi-well analysis — accuracy of comparing wells, and the patterns of depth and gauge errors.
Traps and timing — derivation of the formula for the tilt, and hydrodynamic traps.
Faults — windows in faults, and sealing.
Gross rock volume from maps — the volume of an accumulation from maps, with several contacts.
Pressure measurements in producing reservoirs: barriers and accuracy — potential, pressure profiles in producing accumulations, and depletion of an aquifer.
Abnormal formation pressures: origin, prediction and recognition — the potentiometric surface and artesian pressures.
Determining formation water resistivity — resistivity of formation water at different salinity.
Lumped multi-tank material balance model — blocks connected through windows and faults, in a material balance model.

Check yourself

1
The oil lines of two wells coincide, and the water line of well B is 70 kPa to the right of the water line of well A; $G_w - G_o = 3.43$ kPa/m. Where is the FWL of well B relative to well A?
1. A.Higher by about 70 m
2. B.Lower by about 20 m
3. C.At the same depth
4. D.Higher by about 20 m
Show the answer
Correct answer: D. Higher by about 20 m
With a common oil line, the FWL is the crossing with the water line. The water of well B is “stronger”: its line crosses the oil line higher, by $70/3.43 \approx 20$ m. Why the water is stronger, shut in or flowing, is decided by other signs.
2
Water pressures reduced to one datum in two wells 3 km apart differ by 24 kPa; the water is 1.04 and the oil 0.86 g/cm³. What tilt of the contact along this line would that mean, and does the difference prove a flow?
Show the answer
$G_w - G_o = 9.81\cdot(1.04 - 0.86) = 1.77$ kPa/m, and the shift of the FWL is $24/1.77 \approx 13.6$ m. The tilt is about 4.5 m/km along the line, downwards towards the well with the lower water pressure. But the accuracy of comparing two wells is 15–18 kPa. A difference of 24 kPa is less than two errors, and it does not prove a flow. We need a third well, a check of depths and gauges, or regional heads.
3
In the oil-bearing part, 50 m above the FWL, a log shows a water-bearing sandstone 4 m thick between shales. A pretest in it fell on the regional water line, 0.17 MPa below the oil line. What is it?
1. A.The oil–water contact of the accumulation: there is no oil below
2. B.A supercharged pretest
3. C.A closed lens of trapped water; the contact of the accumulation stays at the former FWL
4. D.Perched water that raised the contact of the whole accumulation by 50 m
Show the answer
Correct answer: C. A closed lens of trapped water; the contact of the accumulation stays at the former FWL
The lens became separated before the oil arrived, and it kept the water pressure of that time. Its pretest is on the regional water line, not on the oil line. The oil above and below the lens is one system, and its FWL is the former one. We subtract the lens from the oil-bearing volume and do not perforate it.
4
Why can the direction of the tilt of a contact not be found from two wells, and why is it bad if three wells stand almost on one straight line?
Show the answer
The gradient of water pressure is a vector in a plane, with two components. Two wells give one equation: the component along the line that joins them. Three wells give a plane. But if they are almost on one line, the component across that line is found by dividing by the small area of the triangle. The errors of gauges and depths then blow it up into nonsense, and the direction can “wander” by tens of degrees.
5
Which argument most strongly tells perched water in a block from a tilt of the contact in a flow?
1. A.The water gradient in both wells is 10.5 kPa/m
2. B.The FWLs of the wells differ
3. C.The FWL of the block coincides with the base of the window in the fault, and the head gradient needed for a flow is many times larger than the regional one
4. D.The oil line of the wells is common
Show the answer
Correct answer: C. The FWL of the block coincides with the base of the window in the fault, and the head gradient needed for a flow is many times larger than the regional one
Different FWLs with common oil are given by both cases. They are told apart by structure and scale. A perched FWL stands at the depth of the lowest point of connection between the blocks, and it changes as a step. A flow needs a head gradient in agreement with the regional one, and it tilts the contact smoothly.
6
Above an oil rim there is a gas cap, and formation water flows to the east. What happens to the thickness of the rim, and why does it matter?
Show the answer
The gas–oil contact is horizontal, because oil and gas are at rest, and the oil–water contact goes down to the east. The rim is thinner in the west, from where the water flows, and thicker in the east. In the west it may disappear, and the gas will lie directly on the water. The thickness of the rim decides the perforation intervals, the risk of gas and water breakthrough, and the very possibility of producing oil on the western flank.

Updated: 19 October 2026