Facies, facies successions and sedimentary logs

Quick refresher

A facies describes the rock and the process that deposited it. A single facies does not decide the environment. The combination and order of facies do: the association and the succession.
Walther's law: only deposits of environments that were lateral (side-by-side) neighbours at the time of deposition lie on top of one another without a break. It does not hold at erosion surfaces, unconformities and faults. An abrupt change between facies that are not neighbours points to a hiatus (a gap in deposition), a sudden change of conditions or a fault.
Fining upward (grains getting finer towards the top) is the typical record of a channel fill and of a waning flow. Coarsening upward is the record of a shoreline, bar or delta advancing into the basin, as ever shallower facies move over deeper-water ones.
A graphic log shows depth, lithology, the grain-size profile (the width of the column), sedimentary structures, boundaries, fossils and the interpretation. Core is described at a scale of 1:20–1:50 and compared with the well log at 1:200 after the depths have been matched.
The shape of the GR curve follows the change in shale volume through the section. A bell means shale volume increasing upwards, a funnel means it decreases upwards, a cylinder is a uniform unit with sharp boundaries, and a serrated curve is interbedding. Different environments give the same shape, so we check it against core.
The order of facies sets the distribution of permeability. In a channel sandstone the flow capacity (permeability times thickness) is concentrated at the base, and in a shoreface sandstone at the top. When layers that communicate vertically are waterflooded, this changes the vertical sweep (the share of the thickness that the water flushes), because water is heavier than oil and moves down.

Key relationships

z_c = \dfrac{\sum_i k_i h_i z_i}{\sum_i k_i h_i}

height of the centre of flow capacity of a sand body above its base

r_{ij} = \dfrac{s_j}{N - s_i}

expected probability of a transition from facies $i$ to facies $j$ if the facies succeed one another at random

H \approx (6\ldots 10)\cdot 2.9\,\bar s

river depth from the mean thickness of cross-sets $\bar s$ , to compare with the thickness of a channel cycle

Common pitfalls

Naming a facies by its interpretation (“fluvial facies”) without describing the rock: the interpretation may be revised, while the description stays correct.
Applying Walther's law across an erosion surface or a fault.
Reading a bell on the GR log as a river channel without core: tidal channels, submarine channel fills and transgressive units look the same.
Taking the thickness of a sandstone for the depth of a single channel: thick sandstones are usually made up of several stacked channels.
Comparing a core log with a well log without matching the depths.

After this lesson you can

distinguish a facies as a description of the rock from a facies as an interpretation of the environment, and explain why a single facies tells of a process but not of an environment
state Walther's law, apply it to a section and recognise the boundaries where it does not hold
recognise in core and on the log the fining-upward of channel sandstones, the coarsening-upward of shoreface and delta-front deposits, and turbidite successions
read and draw a graphic log and compare it with the shape of the GR curve — cylinder, bell, funnel or serrated — keeping in mind that such a reading is ambiguous
estimate from a succession the structure of a sand body, the number of stacked channels and the distribution of permeability through the thickness, and explain what this means for waterflooding
test with a transition matrix whether the change of facies is regular

Facies: description and interpretation

When a geologist describes core or an outcrop (rock exposed at the surface), the geologist divides the section into beds and units. They differ from their neighbours in a combination of features: the composition and size of the grains, sorting, sedimentary structures, colour, fossils, traces of life and the character of the boundaries. A unit with a stable combination of features is called a facies, or more precisely a lithofacies. Each facies is a record of one process: scouring of the bed by a strong current, the movement of dunes, settling of mud from still water, or reworking of sand by waves.

The word “facies” is used in two senses, and it is useful not to mix them. In the descriptive sense, a facies is an observation: “very fine-grained sandstone with hummocky cross-stratification”. In the interpretive sense, it is an environment: “beach facies”, “fluvial (river) facies”. In Russian-language literature, the word is more often tied to the conditions of deposition, and “lithotype” and “lithofacies” are used for descriptive units. A description stays correct even if the interpretation is revised later. That is why reports keep them in separate columns.

We also define facies from other data. Seismic facies are defined by the pattern and strength of reflections on seismic sections (Seismic interpretation), and electrofacies by a combination of well-log readings. These too are descriptive units. We establish their link to the lithofacies of core by comparison.

Core lithofacies in the examples of this lesson

Code: F1
Rock and structures: very coarse-grained sandstone with gravel and mudstone pebbles, erosional base
Process: scouring of the bed by a strong current, a coarse lag (leftover gravel) at the base of a channel

Code: F2
Rock and structures: fine- to coarse-grained sandstone with large-scale trough cross-bedding
Process: migration of dunes along the bed under a current

Code: F3
Rock and structures: fine- and very fine-grained sandstone with small-scale ripple cross-lamination
Process: ripples as the current weakened

Code: F4
Rock and structures: siltstone and mudstone with plant roots, carbonaceous interbeds
Process: settling of mud on a waterlogged plain where plants grow

Code: F5
Rock and structures: very fine-grained sandstone with hummocky cross-stratification, mudstone interbeds
Process: storms on the sea floor between fair-weather and storm wave base

Code: F6
Rock and structures: fine- to medium-grained sandstone with low-angle parallel lamination
Process: wave swash on the beach

Code: F7
Rock and structures: bioturbated mudstone with marine fauna
Process: settling of mud below storm wave base

Code	Rock and structures	Process
F1	very coarse-grained sandstone with gravel and mudstone pebbles, erosional base	scouring of the bed by a strong current, a coarse lag (leftover gravel) at the base of a channel
F2	fine- to coarse-grained sandstone with large-scale trough cross-bedding	migration of dunes along the bed under a current
F3	fine- and very fine-grained sandstone with small-scale ripple cross-lamination	ripples as the current weakened
F4	siltstone and mudstone with plant roots, carbonaceous interbeds	settling of mud on a waterlogged plain where plants grow
F5	very fine-grained sandstone with hummocky cross-stratification, mudstone interbeds	storms on the sea floor between fair-weather and storm wave base
F6	fine- to medium-grained sandstone with low-angle parallel lamination	wave swash on the beach
F7	bioturbated mudstone with marine fauna	settling of mud below storm wave base

The scheme is made up for the sections in this lesson. In every project, it is built anew from core. The lesson Primary sedimentary structures describes how structures are recognised, and the lesson Sedimentary rocks and their classification how rocks are named.

Associations, successions and Walther's law

Facies that occur together and belong to one environment form a facies association. For example, cross-bedded sands, rippled sands, floodplain clays with roots and thin crevasse-splay sands (sand spread out where a flood breaks through a river bank) make up the association of a meandering river. The order in which facies follow one another from bottom to top is a facies succession. In older literature it was called a “sequence”. But sequence stratigraphy has taken over that word, and there it means a body of strata between unconformities (sequences). In Russian-language literature, regular successions are also called cyclites.

We find the environment by comparing the association and the succession with facies models. These are generalised descriptions of which facies, and in what order, a river, a beach, a delta or a submarine fan leaves behind. The models are compiled from modern environments and well-studied ancient sections. Walther's law is what makes it possible to compare the vertical section of a well with the lateral (side-to-side) picture of environments.

Walther's law (in Russian-language literature, the Golovkinsky–Walther law): only deposits of environments that were lateral neighbours at the time of deposition can lie on top of one another without a break. The reason is that environments do not stand still. When the shoreline advances into the sea, a point on the sea floor that used to be shelf is covered first by the shoreface, then by the beach, then by the coastal plain. In the section of a well, these deposits lie one on another in the same order in which the belts ran from sea to land.

SchematicWalther's law (schematic): the shoreline advanced into the sea step by step; in the section of the well, the belts that were lateral neighbours lie on top of one another in the same order — from shelf clays at the bottom to coastal-plain deposits at the top. The boundaries of the belts cut across surfaces of equal age.

The law works only for continuous sections with gradual transitions. It does not hold:

at erosion surfaces and unconformities: the deposits that are missing could have belonged to any environments (unconformities)
at faults: rocks that were never neighbours end up next to each other in a well (Faults)
when conditions change abruptly: in a rapid fall of relative sea level (sea level measured against the land), the sands of the upper shoreface and beach lie directly on the shelf clays, with a sharp, often erosional base. In a rapid rise, marine clays cover the coals of the coastal plain. In both cases the intermediate belts are missing
for belts destroyed by the next environment: a channel that cuts through its own floodplain and the sands of the shore destroys their upper parts

Two types of process govern changes of facies. Autocyclic processes are internal to the environment: a river wanders and abandons its channel, a delta switches its distributaries (the branches of a river that flow into the sea), a shoal (a sandbank) migrates. They create successions of local significance, which may not be repeated in the next well. Allocyclic processes are external: fluctuations of relative sea level, tectonic subsidence and uplift, and climate with the supply of sediment that goes with it. In carbonate basins they also include changes in the carbonate factory itself (the organisms that make the carbonate), for example the extinction of reef builders (Carbonate sediments, environments and carbonate porosity). External causes act on the whole basin, so the boundaries they produce, such as flooding surfaces (surfaces where the water suddenly became deeper), can be traced over long distances.

For a reservoir model, we group facies and associations into genetic units (also called architectural elements). These are bodies created by one process, with a recognisable shape: a channel fill, a belt of point bars (sand bars on the inner banks of river bends), a mouth bar (a sand bar at the mouth of a river channel), a submarine fan lobe. We carry their size and shape, not individual beds, from the wells into the space between wells (Reservoir architecture and performance).

Typical vertical successions

Some successions occur so often that we recognise them in core and logs almost at once. Below are their general features. The topics on fluvial and aeolian and on coastal and marine deposits cover the environments themselves in detail.

Typical successions (bottom to top)

Succession: Meandering-river channel
Bottom to top: erosional base, gravel → cross-bedded sand → rippled sand → silt and clay with roots
GR shape: bell

Succession: Shoreface
Bottom to top: shelf clay → interbedding with storm sands → cross-bedded sand → low-angle laminated beach sand → soil or coal
GR shape: funnel

Succession: Delta front
Bottom to top: prodelta clay → silt and sand → mouth-bar sand; a distributary channel is often cut into the top
GR shape: funnel, sometimes a bell on top

Succession: Submarine fan lobe
Bottom to top: thin turbidites → ever thicker and sandier ones
GR shape: funnel or serrated

Succession: Submarine channel fill
Bottom to top: thick sandy turbidites → ever thinner ones
GR shape: bell

Succession: Carbonate shallowing-upward cycle
Bottom to top: marine mud or grainy sand → tidal-flat deposits with fenestrae, sometimes evaporites
GR shape: often uninformative

Succession	Bottom to top	GR shape
Meandering-river channel	erosional base, gravel → cross-bedded sand → rippled sand → silt and clay with roots	bell
Shoreface	shelf clay → interbedding with storm sands → cross-bedded sand → low-angle laminated beach sand → soil or coal	funnel
Delta front	prodelta clay → silt and sand → mouth-bar sand; a distributary channel is often cut into the top	funnel, sometimes a bell on top
Submarine fan lobe	thin turbidites → ever thicker and sandier ones	funnel or serrated
Submarine channel fill	thick sandy turbidites → ever thinner ones	bell
Carbonate shallowing-upward cycle	marine mud or grainy sand → tidal-flat deposits with fenestrae, sometimes evaporites	often uninformative

The thickness of each succession ranges from metres to a few tens of metres, and of carbonate cycles from decimetres to metres. The schemes are generalised: real sections are often incomplete, stacked and eroded at the top.

Fining upward in a channel. The river scours its bed, leaving gravel and clasts (fragments) of eroded clay at the base of the channel. On the inner bank of a bend, a point bar grows. As it grows, each point on it receives an ever weaker current: dunes give way to ripples, ripples to mud, and floodplain clays with roots are laid down on top. The thickness of a complete succession is roughly equal to the depth of the channel. So we compare it with the depth estimated from the thickness of cross-sets (river palaeodepth). Thick sandstones are usually made up of several stacked channels. Each successive channel cuts the top off the previous one, and erosion surfaces with gravel, and sometimes clay interbeds, are left between them. Braided rivers carry little clay, and the stacked sands make a uniform unit.

Coarsening upward at the shore. When the shoreline advances into the sea, the surf zone (where waves break) approaches every point of the sea floor. Clays deposited below storm wave base are overlain by interbedded clays and storm sands with hummocky cross-stratification. Above them come clean cross-bedded sands of the upper shoreface, and then low-angle laminated beach sands (wave bases). If sea level did not change meanwhile, the thickness of the succession is close to the depth at which the shoreface began. In deltas, coarsening upward builds an advancing mouth bar, which a fining-upward distributary channel often cuts off at the top.

Turbidites. Each turbidite bed is a single event with graded bedding (Bouma sequence). Trends show only at the scale of units. When a submarine fan lobe advances, thin distal (far-from-source) turbidites give way to ever thicker and sandier ones. When a channel fills up and is abandoned, the beds become thinner upwards. Units of thin turbidites without a trend are characteristic of the distal parts of a basin.

Shallowing-upward cycles in carbonates. On shallow-water platforms, accumulation easily keeps up with sea level, and the section is built of metre-scale cycles. Subtidal muds (muds that stay under water) or grainy sands are followed upwards by tidal-flat deposits with fenestrae (small gaps left by gas or drying) and desiccation cracks, sometimes with evaporites (salts left by evaporating water). Then the sea floor is flooded again (Carbonate sediments, environments and carbonate porosity). Uniform sandy units with sharp boundaries are produced by the stacked channels of braided rivers, aeolian dunes, submarine channels and carbonate shoals.

CalculatedFour shapes of the GR curve, calculated for assumed successions: bell (channel), funnel (shoreface), cylinder (uniform sandstone), serrated (interbedding). On the left of each panel is a column whose width shows the grain size.

The graphic log

A graphic sedimentary log is a drawing of a section that shows at a glance how the rock, grain size and sedimentary structures change. We draw it from the description of core or an outcrop, and lay it out so that another specialist can reconstruct the main points without the text.

Depth on the left: depth for wells, height above the base for outcrops. For core it is the depth along the drill string, which is later matched to the well log.
Lithology: a fill or symbol for each rock — sandstones, siltstones, clays, carbonates, coals, evaporites.
Grain size: the width of the fill; the right edge of the column is the size profile on the scale “clay – silt – sand from very fine to very coarse – gravel” (size scale). Fining upward looks like a column that narrows towards the top, and coarsening like one that widens.
Sedimentary structures: symbols that look like the structure itself — inclined dashes for cross-bedding, wavy lines for ripples, burrow marks.
Bed boundaries: sharp and flat, erosional (a wavy line, often with pebbles above it), gradational.
Fossils, bioturbation index, palaeocurrents (arrows), sample positions, oil and gas shows; on the right, facies codes and the interpretation.

We describe core at a large scale, 1:20–1:50, so that individual beds fit. To compare it with the well log, we redraw the column at the scale of the log displays, usually 1:200 (log displays). First we shift the core depths to the log depths by matching the gamma activity of core and log (core depth matching). After that we tie the core facies to the shape of the curves and extend them to wells without core.

CalculatedGraphic log of an assumed section and the GR curve calculated from it. The shale volume of the beds is converted to readings, using clean-rock and shale lines of 25 and 125 API, and averaged over 0.5 m. On the right is the interpretation.

Worked exampleWhat the log shows

Given (assumed data)

Section: the log in the figure: 2412–2440 m, 18 beds
Reservoir cut-off in the example: sand or a coarser rock with shale volume below 0.3
Thickness of cross-sets in the channel sandstones: 0.16 m on average, as in the example of the lesson on sedimentary structures

Solution

The interval is 28 m thick; the reservoir is 15.7 m: 6.5 m of shoreface sandstone, 8.6 m of channel sandstones and 0.6 m of crevasse-splay sandstone. The net-to-gross ratio (the share of reservoir rock in the total thickness) is $15.7/28 = 0.56$ .
The lower unit, 2428.0–2437.0 m (9.0 m): above the bioturbated shelf clays, interbedding with hummocky stratification gives way upwards to cross-bedded and low-angle laminated sandstone. The grain size grows from silt to medium sand, and GR falls upwards. This is coarsening upward, the advance of the shoreline.
The upper unit, 2418.6–2427.2 m (8.6 m), begins with an erosion surface with gravel. The grain size falls upwards, and floodplain clays with roots lie on top: a channel succession. At 2422.6 m inside it, a second erosion surface with gravel divides the unit into two cycles — 4.6 and 4.0 m.
River depth from the sets: dune height $h_d \approx 2.9\cdot 0.16 = 0.46$ m, depth $(6\ldots 10)\,h_d$ — from 2.8 to 4.6 m. In the extreme estimates, when the dune height is from 2.2 to 3.6 times the set thickness, the depth is from $6\cdot 2.2\cdot 0.16 = 2.1$ to $10\cdot 3.6\cdot 0.16 = 5.8$ m. Each cycle fits within these limits, but the unit as a whole (8.6 m) does not.

Answer: The section is a regressive succession (the shoreline moves seaward, and shallower facies move over deeper ones): shelf → shoreface and beach → coastal plain, into which two stacked channels are cut. Reservoirs make up 56% of the section and belong to bodies of different origin.

What it means. The bodies have different shapes. The shoreface sandstone is a wide sheet stretching along the ancient shoreline, and we can correlate it confidently between wells. The channel sandstones are narrow ribbons and belts, and a neighbouring well may miss them. The surface between the two channels is a possible weak barrier: clay pebbles or a thin clay interbed may remain on it. The comparison with river depth is rough. The present thickness of a cycle is less than the original, because of compaction and because the next channel eroded its top. Also, individual scours in a channel can be deeper than the mean depth. The data in this example are assumed for illustration.

GR curve shape: bell, funnel, cylinder

In clastic sections (made of rock fragments), shale volume is usually related to grain size: the finer the sediment, the more clay particles it contains. Most of the natural gamma radiation of such rocks comes, as a rule, from clay minerals (what the GR log tells us). So the shape of the GR curve in a sandy unit often repeats the grain-size profile, which is the right edge of the graphic log, only mirrored: the fine material goes to the right, towards high readings. We read the spontaneous potential curve (a log of the natural electrical voltages in the borehole; see Spontaneous potential (SP) log) in the same way. In Russian-language practice, these curves are used to build electrometric facies models.

Cylinder (blocky): low readings through the whole unit and sharp boundaries: uniform sandstone — stacked channels, aeolian dunes, a submarine channel fill, a carbonate shoal.
Bell: a sharp base and readings increasing upwards — shale volume increases, often a fining-upward trend: a channel, a tidal channel, a submarine channel fill, a transgressive unit (deposits laid down as the sea deepens).
Funnel: readings decrease upwards and the top is usually sharp — shale volume decreases, often a coarsening-upward trend: a shoreface, a mouth bar, a crevasse splay, an advancing submarine fan lobe.
Symmetric shape: coarsening upward, then fining: the advance and subsequent retreat of a shoreline or bar.
Serrated curve: frequent oscillations: interbedding of thin sands and clays — a floodplain with crevasse splays, a tidal flat, thin turbidites.

The shape of a curve is a hint, not a diagnosis. Different environments give the same shape, so we check it against core, the structures seen on borehole image logs, fossils and bioturbation. Marine fauna and the burrows of burrowing organisms rule out a river, while roots and coals rule out the open sea. Also, GR measures radioactivity, not grain size. Coarse sands rich in mica or feldspar give high readings, potassium-poor kaolinitic clays give low ones, and in carbonates a rise in the curve is often caused by uranium, not clay (GR pitfalls). Finally, the tool averages about half a metre of section. It smooths out thin beds and the gravel at the base of a channel, and this can be seen in the figures of the lesson.

Succession and reservoir behaviour

The order of facies in a unit also sets the order of permeability. In a channel sandstone, the coarsest and most permeable sands lie at the base. In a shoreface sandstone, they lie at the top. A single number, the height of the centre of flow capacity, conveniently describes where the flow capacity of a unit is concentrated.

Height of the centre of flow capacity of a sand body

z_c = \frac{\sum_i k_i h_i z_i}{\sum_i k_i h_i}

where:

$k_i$ permeability of layer $i$ , mD or D — the same for all layers
$h_i$ thickness of the layer, m
$z_i$ height of the middle of the layer above the base of the body, m
$z_c$ height of the centre of flow capacity above the base, m

Units: any consistent units

Holds when

The layers are horizontal and can be traced over the distance between wells.
Flow is along the layers, and the contribution of a layer is proportional to its flow capacity $k_i h_i$ (averaging over layers).

The ratio $z_c/h$ , where $h = \sum_i h_i$ is the thickness of the body, is less than 0.5 if the flow capacity is concentrated at the bottom and more than 0.5 if it is concentrated at the top. Together with the mean permeability $\sum k_i h_i/h$ it indicates where the injected water will go.

Worked exampleWhere the main flow goes in the sand bodies of the example

Given (assumed data)

Lower channel cycle, 2422.6–2427.2 m (bottom to top): 0.2 m: $M_z = -0.5$ , $\sigma_I = 1.5$ ; 1.8 m: 0.7 and 0.8; 1.4 m: 1.5 and 0.6; 1.2 m: 2.6 and 0.5
Shoreface unit, 2428.0–2437.0 m (bottom to top): 2.5 m: $M_z = 4.3$ , $\sigma_I = 1.1$ ; 2.5 m: 3.3 and 0.7; 2.5 m: 2.6 and 0.45; 1.5 m: 2.0 and 0.35
Permeability of a layer: from the Krumbein–Monk formula $k = 760\,d_g^2 e^{-1.31\sigma_I}$ , $d_g = 2^{-M_z}$ mm; mean size $M_z$ and sorting $\sigma_I$ in φ units

Solution

Permeability of the channel layers from bottom to top: 213.0, 101.0, 43.3 and 10.7 D; of the shoreface unit: 0.5, 3.1, 11.5 and 30.0 D.
Flow capacities of the channel: $42.60 + 181.80 + 60.62 + 12.84 = 297.86$ D·m; of the shoreface unit: $1.25 + 7.75 + 28.75 + 45.00 = 82.75$ D·m.
Channel: $z_c = (42.60\cdot 0.1 + 181.80\cdot 1.1 + 60.62\cdot 2.7 + 12.84\cdot 4.0)/297.86 = 1.41$ m, $z_c/h = 1.41/4.6 = 0.31$ .
Shoreface unit: $z_c = (1.25\cdot 1.25 + 7.75\cdot 3.75 + 28.75\cdot 6.25 + 45.00\cdot 8.25)/82.75 = 7.03$ m, $z_c/h = 7.03/9.0 = 0.78$ .
The lower half of the channel carries 80% of the flow capacity, and the upper half of the shoreface unit carries 91%.

Answer: In the channel cycle, the centre of flow capacity lies at 0.31 of the thickness above the base, and 80% of the flow goes through the lower half. In the shoreface unit, it lies at 0.78 of the thickness, and 91% of the flow goes through the upper half.

What it means. The formula is for loose sand, so it overestimates permeability by orders of magnitude: in the reservoir, the sands are compacted and cemented (initial permeability). For the lowest layer of the shoreface unit, which has clay interbeds, the overestimate is especially large. But the distribution through the thickness is largely preserved. This matters in waterflooding. Water is heavier than oil and sinks. In a channel sandstone, water moves along the permeable base, and gravity holds it there. The water breaks through early to the producing wells, and the top of the reservoir is left poorly swept. In the shoreface unit, water enters the permeable top, while gravity pulls it down into the less permeable layers. The front levels out, and the vertical sweep is better. This is so if the layers communicate vertically and the mobility ratio (how easily the water moves compared with the oil) is favourable. Without crossflow between the layers, their order has no effect on the sweep. The topic on displacement in layered reservoirs treats this quantitatively. The data in this example are assumed for illustration.

Successions also set barriers. Between stacked channels, erosion surfaces with clay pebbles and clay interbeds remain. Between shoreface units, there are the clays of flooding surfaces, which are continuous over a large area. The first kind of barrier is discontinuous, so the flow bypasses it. The second kind divides the reservoir into parts that communicate poorly (Reservoir heterogeneity and permeability averaging). With this in mind, we choose perforation intervals, injection points and the order of developing the layers.

Terms and symbols

Facies, lithofacies: A bed or unit with a stable combination of features — composition, grain size, structures, fossils — that reflects the process of deposition.
Facies association: Facies that occur together and belong to one environment.
Facies succession: A regular change of facies from bottom to top of a section; in Russian-language literature also a cyclite.
Facies model: A generalised description of the facies of an environment and their spatial order, compiled from modern environments and ancient sections.
Walther's law: Only deposits of environments that were lateral neighbours lie on top of one another without a break.
Autocyclic and allocyclic processes: Internal to the environment (wandering of channels, switching of distributaries) and external (sea level, subsidence, climate, sediment supply).
Genetic unit: A body created by one process and having a recognisable shape: a channel fill, a mouth bar, a submarine fan lobe.
Fining and coarsening upward: A decrease and an increase of grain size from bottom to top of a unit.
Graphic sedimentary log: A drawing of a section: depth, lithology, grain-size profile, structures, boundaries, fossils, interpretation.
Bell, funnel, cylinder: Shapes of the GR curve in a sandy unit: shale volume increases upwards, decreases upwards or is constant.
Electrofacies: An interval with a characteristic combination of well-log readings, compared with the core lithofacies.
Centre of flow capacity: The height above the base of a body, weighted by the flow capacities of the layers $k_i h_i$ .

Common mistakes and misconceptions

Wrong

“If there is a sandstone with large-scale cross-bedding in the core, it is a river channel.”

In fact

Cross-bedding tells us about a process (bedforms migrating along the bed), not about an environment. The neighbouring facies and their order show the environment: floodplain clays with roots lie above a channel sandstone, and interbeds with storm sands and marine fauna lie beneath a shoreface sandstone, as in the example of the lesson.

Wrong

“If facies lie on top of one another, their environments were neighbours.”

In fact

Only if there is no break between them. At erosion surfaces, unconformities and faults, facies that were never neighbours end up side by side, so such boundaries must be picked out specially.

Wrong

“A bell on the GR log is a river channel.”

In fact

A bell means only that shale volume increases upwards. Tidal channels, submarine channel fills and transgressive units look the same. We check the environment against core, fossils and structures.

Wrong

“The thickness of a channel sandstone equals the depth of the river.”

In fact

Thick sandstones are usually made up of several stacked channels. The depth of one channel is estimated from the cycle between erosion surfaces and from the thickness of cross-sets.

Wrong

“Only the mean permeability of a unit matters for waterflooding.”

In fact

It also matters where in the unit the flow capacity is concentrated. With the same mean permeability and vertically communicating layers, water in a unit with a permeable base breaks through earlier and sweeps the top less well than in a unit with a permeable top.

In depth: advanced detail

Is the change of facies regular: the transition matrix

The eye readily sees a pattern even where there is none. Counting transitions helps to check that facies really do follow one another in a particular order. We write the section as a chain of facies from bottom to top, and count how many times facies $i$ is followed by facies $j$ . We do not count a transition of a facies into itself, because a boundary within a single facies is arbitrary. So such a chain is called an embedded Markov chain (a chain in which only changes of state are counted).

Observed and expected transition probabilities

p_{ij} = \frac{f_{ij}}{s_i},\qquad r_{ij} = \frac{s_j}{N - s_i},\qquad d_{ij} = p_{ij} - r_{ij}

where:

$f_{ij}$ number of upward transitions from facies $i$ to facies $j$ , —
$s_i$ number of transitions from facies $i$ (the row total), —
$N$ total number of transitions, —
$p_{ij}$ observed transition probability, —
$r_{ij}$ expected probability if the facies follow one another at random but never repeat, —
$d_{ij}$ difference: a positive value means the transition is preferred, —

Units: dimensionless

Holds when

Beds are defined by one rule throughout the section.
The simple estimate of $r_{ij}$ is approximate. Refined methods treat the ban on transitions of a facies into itself more rigorously. We test the significance of deviations with the $\chi^2$ test (the chi-squared test).

A chain of preferred transitions is a facies succession identified from numbers rather than by eye.

Worked exampleFacies transitions in a fluvial core

Given (assumed data)

Section: 40 beds of four facies: F1 — gravel at the base of a channel, F2 — cross-bedded sandstone, F3 — rippled sandstone, F4 — siltstone and clay with roots
Number of upward transitions: F1→F2 — 8, F1→F3 — 1; F2→F3 — 8, F2→F4 — 2; F3→F2 — 1, F3→F4 — 9; F4→F1 — 8, F4→F2 — 1, F4→F3 — 1

Solution

Row totals: 9 transitions from F1, 10 from F2, 10 from F3, 10 from F4; in all $N = 39$ .
Observed probabilities of the main transitions: F1→F2 $8/9 = 0.889$ , F2→F3 $8/10 = 0.800$ , F3→F4 $9/10 = 0.900$ , F4→F1 $8/10 = 0.800$ .
Expected under random succession: $r_{12} = 10/(39 - 9) = 0.333$ , $r_{23} = 10/(39 - 10) = 0.345$ , $r_{34} = 10/(39 - 10) = 0.345$ , $r_{41} = 9/(39 - 10) = 0.310$ .
Differences: 0.556, 0.455, 0.555 and 0.490. All four transitions occur much more often than chance would give. The differences for the other transitions are negative.

Answer: The preferred chain F1 → F2 → F3 → F4 → F1 is a repeating fining-upward cycle with an erosional base, typical of channels.

What it means. The transition F4 → F1 (gravel above floodplain clay) is the erosional base of the next channel, not a gradual change. So we do not apply Walther's law to it, even though the transition is regular. Rare transitions are informative too. F3 → F2 is a renewed strengthening of the current within a channel. F4 → F2 and F4 → F3 are crevasse-splay sands or channels without a gravel base. With three or four dozen transitions, the estimates are rough. We check their reliability with the $\chi^2$ test and by comparison with other wells. The data in this example are assumed for illustration.

Connections to other topics

Primary sedimentary structures — the structures by which facies are defined, and river palaeodepth from cross-sets.
Sediment texture: grain size, sorting and shape — grain size and sorting — the profile of the graphic log.
Stratigraphy and geologic time — unconformities, why lithological boundaries are not of the same age everywhere (diachronous boundaries), and sequence stratigraphy.
Gamma ray log and natural gamma ray spectrometry — the physics of gamma-ray logging and its pitfalls.
Well correlation — how well sections are correlated using facies and marker surfaces.
Carbonate sediments, environments and carbonate porosity — carbonate environments and shallowing-upward cycles.
Displacement in heterogeneous reservoirs — waterflooding of layered reservoirs and the role of the order of the layers.
Reservoir heterogeneity and permeability averaging — measures of heterogeneity and the averaging of layer permeabilities.
Core analysis — core description, scanning and depth matching.
Continental environments: aeolian and fluvial — fluvial and aeolian environments.
Coastal, shallow-marine and deep-marine environments — coastal, deltaic and deep-water environments.
Reservoir architecture and performance — the shape and connectivity of sand bodies.

Check yourself

1
Low-angle laminated beach sandstone lies sharply, with an erosional base, on bioturbated shelf clays; there is no interbedding with storm sands between them. How can this be explained?
1. A.The beach sandstone was deposited before the clays
2. B.It is definitely a fault
3. C.Ordinary advance of the shoreline — Walther's law is satisfied
4. D.The shoreface belts are missing: probably a sharp fall of relative sea level and erosion; it is important to trace such a surface in correlation
Show the answer
Correct answer: D. The shoreface belts are missing: probably a sharp fall of relative sea level and erosion; it is important to trace such a surface in correlation
Shelf clays and beach sands are not neighbours: between them lies the shoreface. Since its deposits are missing and the base is erosional, the change was a jump, usually caused by a rapid fall of relative sea level. A fault is also possible, but that conclusion needs other evidence: a missing or repeated section, or slickensides (polished, grooved fault surfaces).
2
The GR curve in a sandy unit is a bell: a sharp base and readings increasing upwards. Which conclusion is correct?
1. A.Grain size necessarily decreases upwards
2. B.It is a river channel
3. C.Shale volume increases upwards; it may be a channel, a tidal channel, a submarine channel fill or a transgressive unit — the environment is refined from core
4. D.It is a shoreface
Show the answer
Correct answer: C. Shale volume increases upwards; it may be a channel, a tidal channel, a submarine channel fill or a transgressive unit — the environment is refined from core
GR shows a change in radioactivity, usually related to shale volume. Several environments give a bell, and micaceous or feldspathic sands can distort the link with grain size. What decides is core, fossils, bioturbation and structures on image logs.
3
The mean thickness of cross-sets in a channel sandstone is 0.25 m, and a fining-upward unit is 12 m thick. Is it a single channel?
Show the answer
Dune height $h_d \approx 2.9\cdot 0.25 \approx 0.73$ m, river depth $(6\ldots 10)\cdot 0.73 \approx 4.4–7.3$ m, extreme estimates $6\cdot 2.2\cdot 0.25 = 3.3$ and $10\cdot 3.6\cdot 0.25 = 9.0$ m. 12 m is more than even the extreme estimate, so the unit was most probably built by two or more stacked channels. In core, we look for internal erosion surfaces with gravel and jumps in grain size.
4
Two units of the same thickness and mean permeability, with the layers in both communicating vertically: in the first, permeability falls upwards; in the second, it rises upwards. In which of them will water in waterflooding break through sooner along the base and sweep the top worse?
1. A.In the second
2. B.Equally: what decides is the mean permeability
3. C.It depends only on the viscosity of the oil
4. D.In the first
Show the answer
Correct answer: D. In the first
In the first unit the permeable base and the weight of the water act together: the water moves along the bottom and breaks through early. In the second, water enters the permeable top, and gravity pulls it down, levelling the front. The viscosities of oil and water affect both units and can reduce the difference between them, but they do not change which of them breaks through first. If the layers were separated by continuous clays and there were no crossflow, the order of the layers would have no effect on the sweep.
5
A chain of facies has 30 transitions; 6 are from facies A and 9 from facies B. What is the probability of the transition A → B if facies succeed one another at random, and what does it mean if the observed probability is 5/6?
Show the answer
$r_{AB} = s_B/(N - s_A) = 9/(30 - 6) = 0.375$ . The observed $5/6 \approx 0.83$ is greater by 0.46, so the transition A → B is preferred. With only six transitions from A, the conclusion is tentative. We check it with the $\chi^2$ test and on other sections.
6
Two wells are 2 km apart. A thin bed of clay with marine fauna can be traced in both, while a channel sandstone is present in only one. Which boundary is more reliable as a marker in correlation, and why?
Show the answer
The clays with marine fauna were deposited when the whole area was flooded. That is an allocyclic event, and such a surface can be traced far. A channel is the result of the wandering of a river, which is an autocyclic and local process. The channel may be absent from the neighbouring well, and its base cuts different beds. So we take the base or top of the marine clays as the marker, and correlate the channel sandstone as a separate body.

Updated: 15 October 2026