Stratigraphy and geologic time

Quick refresher

In undisturbed bedding, the lower layer is older than the upper one. Everything that cuts across the layers — faults, intrusions (bodies of rock that pushed in as magma), erosion surfaces — is younger than they are.
A section is incomplete: millions of years are missing at unconformities. The ages of the rocks above and below an unconformity show how much time is not recorded.
Radioactive clocks give a numerical age: $t = \ln(1 + D^*/P)/\lambda$ . Volcanic ash beds and intrusions are the most reliable to date. Detrital grains (grains eroded from older rocks) carry the age of the source area, not of the deposit.
Time is divided into eons, eras, periods, epochs and ages. The rocks laid down in those spans are divided into eonothems, erathems, systems, series and stages. The ages of the boundaries are refined from time to time. On the 2026/06 version of the scale, the Phanerozoic began 538.8 million years ago, the Mesozoic 251.902 million years ago and the Cenozoic 66.0 million years ago.
Boundaries between rock types slide through time (they are diachronous: the same boundary is older in one place and younger in another). Boundaries based on fossils (biostratigraphic) and on time (chronostratigraphic) are close to time lines. Correlating “sand with sand” can link bodies of different ages.
The shape of the layers depends on a balance between the space available for sediment (subsidence plus the change in sea level) and the supply of sediment. This balance decides transgressions (the sea advancing over land), regressions (the sea retreating) and sequence boundaries.

Key relationships

P = P_0\,e^{-\lambda t},\qquad t_{1/2} = \frac{\ln 2}{\lambda}

the law of radioactive decay and the half-life

t = \frac{1}{\lambda}\ln\left(1 + \frac{D^*}{P}\right)

age from the accumulated radiogenic daughter isotope

\dot A = \dot S + \dot E

rate at which space for sediment is created: subsidence plus eustatic (worldwide) change of sea level

Common pitfalls

Taking the age of detrital grains — for example zircons from a sandstone — for the age of deposition: it shows only that the rock is not older than its youngest grain.
Correlating beds because the rocks look alike, without checking age: identical sandstones in two wells may be of different ages, and different ones of the same age.
Treating a section as continuous: breaks and erosion hide millions of years, and the average rate of accumulation “across a break” is too low.
Taking the ages of boundaries from old tables: the scale is revised, so a report should state which version it uses.
Confusing “lower — upper” (for rock sequences) with “early — late” (for time): “Lower Jurassic rocks”, but “the Early Jurassic epoch”.

After this lesson you can

apply the principles of stratigraphy and work out from a section the order of deposition, deformation, erosion and faulting
distinguish the types of unconformity, estimate from the ages of the rocks the time missing at an unconformity, and explain why unconformities matter for traps and seals
calculate an age from the law of radioactive decay, check it against two isotope systems and estimate the age of a horizon between dated beds
use the geologic time scale: the ranks of its divisions, the main boundaries and their ages on the current international scale
distinguish lithostratigraphic, biostratigraphic and chronostratigraphic units (units defined by rock type, by fossils, and by time) and explain why the boundaries of beds do not coincide with time lines
relate subsidence, changes in sea level and sediment supply to the geometry of layers — the basis of sequence stratigraphy

The principles of stratigraphy: what came first, what came later

Sediment builds up layer by layer, and this simple fact holds the main rules for reading a section. The Danish naturalist Niels Stensen, known as Nicolas Steno, stated the first three in 1669:

Law of superposition: in undisturbed bedding each layer is younger than the one beneath it and older than the one above it.
Original horizontality: sediment is deposited almost horizontally, so layers that are steeply tilted or crumpled were deformed after they were deposited. The exceptions are cross-lamination (thin layers laid down at an angle inside a bed) and the slopes of deltas and reefs, where the original dip reaches tens of degrees.
Lateral continuity: a layer continues in all directions until it thins to nothing (pinches out), changes into another rock or is cut off. Identical layers on opposite sides of a valley or a fault are parts of one layer.

Several more rules were added later:

Cross-cutting relationships: a fault, fracture, intrusion or erosion surface is younger than all the rocks it cuts.
Inclusions: fragments of rock inside a layer are older than the layer. A granite pebble is older than the conglomerate it lies in.
Faunal succession: fossils follow one another in a section in a fixed order, the same everywhere. So we use them to correlate distant sections (see the section on biostratigraphy).
Uniformitarianism (the present is the key to the past): the erosion, transport and deposition that we observe today also operated in the past, so modern rivers, deltas and reefs help us to understand ancient ones. Rates and scales may have been different, though. Rare events — floods, landslides, eruptions — leave much more evidence in a section than their rarity would suggest.

SchematicA schematic section for working out the order of events. Tilted layers 1–3 and fault F1 are cut off by an erosion surface. Horizontal layers 4–6 lie on it, with a bed of volcanic ash in layer 5. Fault F2 displaces layers 4 and 5 but not layer 6.

How to read the section in the figure

1
The oldest are layers 1, 2 and 3
They were deposited horizontally in this order (the law of superposition) and acquired their tilt later (original horizontality).
2
Fault F1
It displaces layers 1–3 and is cut off by the erosion surface. So it is younger than layer 3 and older than the erosion. The tilting of the layers also happened in this interval. The figure does not show which came first, the tilting or the fault. To find out, you check whether the fault itself is tilted together with the layers.
3
Uplift and erosion
The surface that cuts the tilted layers and the fault is an angular unconformity: after the tilting, the area was uplifted and eroded, and then sank again below the level at which sediment is deposited.
4
Layers 4–6
They lie horizontally on the erosion surface. Sandstone 4 contains pebbles of rocks 1–3, which by the rule of inclusions are older than layer 4. Above it come shale 5 with an ash bed, then sandstone 6.
5
Fault F2
It displaces layers 4 and 5 but not layer 6: it is younger than layer 5 and older than layer 6. The age of the ash, found from isotopes, is the oldest the fault can be: the fault is not older than the ash.

In short: deposition of layers 1, 2 and 3 → tilting and fault F1 → uplift and erosion → deposition of layers 4 and 5 with the ash → fault F2 → deposition of layer 6. The same method, applied to seismic sections and wells, rebuilds the history of traps: when the beds were tilted, when the faults were active and when the seal was deposited (timing of trap formation).

Unconformities and missing time

Sediment does not build up without a break. A surface where younger rocks lie on markedly older ones is called an unconformity. It separates two periods of deposition. Between them, either no sediment collected, or sediment laid down earlier was eroded. The span of time that the rocks of the section do not represent is a hiatus (stratigraphic break).

Types of unconformity

Type: Angular unconformity
What is seen: young layers lie on truncated, tilted or folded layers
What happened: deposition, deformation, uplift and erosion, renewed subsidence and deposition

Type: Disconformity (parallel unconformity with erosion)
What is seen: the layers above and below are parallel, the boundary is uneven, with pebbles and signs of weathering on it
What happened: uplift without tilting, erosion, renewed deposition

Type: Paraconformity (hidden unconformity without signs of erosion)
What is seen: the boundary looks like an ordinary bedding surface
What happened: a long break in deposition; the main clue is the ages of the rocks, because some biostratigraphic zones are missing

Type: Nonconformity (unconformity on basement)
What is seen: sedimentary rocks lie on igneous or metamorphic ones
What happened: the basement was exposed at the surface and eroded, then covered by sediment

Type	What is seen	What happened
Angular unconformity	young layers lie on truncated, tilted or folded layers	deposition, deformation, uplift and erosion, renewed subsidence and deposition
Disconformity (parallel unconformity with erosion)	the layers above and below are parallel, the boundary is uneven, with pebbles and signs of weathering on it	uplift without tilting, erosion, renewed deposition
Paraconformity (hidden unconformity without signs of erosion)	the boundary looks like an ordinary bedding surface	a long break in deposition; the main clue is the ages of the rocks, because some biostratigraphic zones are missing
Nonconformity (unconformity on basement)	sedimentary rocks lie on igneous or metamorphic ones	the basement was exposed at the surface and eroded, then covered by sediment

The type of unconformity can change along a single surface. Where the edge of the basin was uplifted and tilted, it is angular. Towards the centre of the basin the angle between the layers disappears, the break gets shorter, and the surface passes into conformable bedding (layers lying parallel, with no break).

SchematicThe four types of unconformity (schematic): angular, parallel with erosion (disconformity), hidden without signs of erosion (paraconformity) and unconformity on basement (nonconformity). The orange line is the unconformity surface.

In a well, you recognise an unconformity by several signs. The dip of the layers changes abruptly on image logs and dipmeter data (an angular unconformity; how dip is measured). Biostratigraphic zones are missing. At the boundary there are conglomerates, weathering profiles and karst (cavities dissolved in limestone). After a long break without erosion, there is a hardground: a consolidated sea floor with mineral crusts and nodules. The maturity of the organic matter jumps (Source rocks, maturation and migration). Density and wave velocity jump too, if the rocks beneath the unconformity were once buried deeper and compacted more. On seismic sections, truncated reflections can be seen beneath an angular unconformity (Seismic interpretation).

Worked exampleHow much time is missing at an unconformity

Given (assumed data)

Below the erosion surface: sandstones with ammonites of the Bathonian stage of the Middle Jurassic: 168.2–165.3 million years ago
Above: shales with foraminifera of the Albian stage of the Lower Cretaceous: 113.2–100.5 million years ago
Ages of the stage boundaries: from the International Chronostratigraphic Chart, version 2026/06

Solution

The smallest gap: the youngest Bathonian layers are preserved beneath the unconformity, and the oldest Albian layers lie above it: $165.3 - 113.2 = 52.1$ million years.
The largest gap: only the oldest Bathonian layers are preserved below, and the youngest Albian layers lie above: $168.2 - 100.5 = 67.7$ million years.
The section has no Callovian stage, no Upper Jurassic at all (from 161.5 to 143.1 million years ago) and no Lower Cretaceous before the Albian. For comparison, the whole Middle Jurassic lasted $174.7 - 161.5 = 13.2$ million years.

Answer: Between 52.1 and 67.7 million years are missing at the unconformity — several times longer than the whole Middle Jurassic lasted.

What it means. The accuracy is limited by how finely biostratigraphy can date the rocks. Here the unit is a stage lasting 3 to 13 million years; finer zones would narrow the range. Gaps like this are common: at the edges of basins, a section often has less time recorded than missing. There are two practical consequences. Do not calculate an average rate of accumulation from ages that an unconformity separates. And do not trace a bed across an unconformity as if it were continuous, because beneath the unconformity the bed may be cut off. The data in this example are assumed for illustration.

Unconformities in the subsurface of Kazakhstan

In the south of the Pre-Caspian Basin, in the Karaton–Tengiz zone, there are carbonate buildups hundreds of metres high. They are made of rocks from the Upper Devonian to the Bashkirian stage of the Middle Carboniferous. Thin Lower Permian shales lie on top of them with an unconformity. So at the tops of the buildups at least 16 million years are not recorded: from the end of the Bashkirian age (315.2 million years ago) to the beginning of the Permian (298.9 million years ago).
In the south-east of the basin, on the slope of the South Emba Uplift, the pre-Jurassic unconformity cuts off part of the subsalt carbonate platform together with its Kungurian salt seal. For this reason, the resource assessment of the early 2000s rated the prospects for new discoveries there as low.
On Mangyshlak the Jurassic sequence lies with an unconformity on Permian–Triassic rocks that were deformed before the start of the Jurassic. The productive sandstones of Uzen and Zhetybai are in its lower part.

Traps. A reservoir cut off by an unconformity and sealed by shales forms a stratigraphic trap (a trap formed by a change in the rock layers, not by a fold or fault) (traps beneath an unconformity).
Reservoirs. Beneath an erosion surface, water dissolves (leaches) and weathers carbonates and basement rocks. Karst and fractures can make a good reservoir.
Seals. Erosion can destroy the seal, and then the accumulation beneath it will not be preserved.
Migration paths (the routes oil and gas take from the source rock to the trap). Porous rocks at the surface of an unconformity carry oil and gas over long distances.
Burial history. You must add back the eroded section. Otherwise the burial history curve and the maturation of the source rock come out wrong (burial history curves).

Relative and numerical age: radioactive clocks

The principles of stratigraphy give a relative age: what is older and what is younger. Radioactive isotopes give a numerical (or absolute) age: how many years ago. The nuclei of a radioactive isotope, called the parent, decay by themselves into a daughter isotope. In each unit of time, a constant fraction of the remaining nuclei decays. Temperature, pressure and chemical bonding hardly affect this rate. That is what makes decay a clock: radiometric dating.

The law of radioactive decay, and age from the accumulated daughter isotope

P = P_0\,e^{-\lambda t}, \qquad t_{1/2} = \frac{\ln 2}{\lambda}, \qquad t = \frac{1}{\lambda}\ln\left(1 + \frac{D^*}{P}\right)

where:

$P_0,\ P$ number of atoms of the parent isotope in the mineral at the moment it formed and now, atoms (or moles)
$D^*$ number of atoms of the radiogenic daughter isotope accumulated over the time $t$ : $D^* = P_0 - P$ , atoms
$\lambda$ decay constant — the fraction of nuclei that decay per unit time, yr⁻¹
$t_{1/2}$ half-life: the time in which half of the parent nuclei decay, years
$t$ age — the time since the system closed, years

Units: $\lambda$ and $t$ in reciprocal units (yr⁻¹ and years); $D^*/P$ is a ratio of numbers of atoms; ln is the natural logarithm

Holds when

The system is closed: after the mineral formed, neither parent nor daughter isotope entered or left it.
There was no daughter isotope at the start, or its initial amount is known (otherwise the isochron method is used — see the advanced section).
Every decay gives an atom of exactly this daughter isotope; for ⁴⁰K only about 10% of decays turn into ⁴⁰Ar (the rest give ⁴⁰Ca), so for potassium–argon ages $D^*/P$ is divided by this fraction.
The decay constant is known exactly; the age is counted from the moment when the mineral cooled enough for the daughter isotope to stop escaping from it.

The expression for the age follows from $D^* = P_0 - P = P\,(e^{\lambda t} - 1)$ . In every half-life the amount of parent isotope halves: one half-life leaves a half, two leave a quarter, three leave an eighth, and the ratio $D^*/P$ is then 1, 3 and 7. The method works as long as you can measure the ratio accurately. Isotopes with long half-lives date old rocks, and those with short half-lives date young ones.

CalculatedFractions of the parent (

P/P_0

) and accumulated daughter (

D^*/P_0

) isotopes against time expressed in half-lives, according to the decay law. The points for 1, 2 and 3 half-lives are marked.

The radioactive clocks used most often

System: ²³⁸U → ²⁰⁶Pb and ²³⁵U → ²⁰⁷Pb
Half-life: 4.47 and 0.704 billion years
What is dated: zircon from volcanic ash beds and igneous rocks; the most precise dates, to hundredths of a per cent

System: ⁴⁰K → ⁴⁰Ar (and the ⁴⁰Ar/³⁹Ar method)
Half-life: 1.25 billion years
What is dated: volcanic rocks and ash beds (sanidine, biotite); glauconite and illite, with caveats

System: ⁸⁷Rb → ⁸⁷Sr
Half-life: about 50 billion years
What is dated: igneous and metamorphic rocks — by the isochron method

System: ¹⁸⁷Re → ¹⁸⁷Os
Half-life: about 42 billion years
What is dated: black shales rich in organic matter, sulfides; experimental dating of oil

System: ¹⁴C → ¹⁴N
Half-life: 5730 years
What is dated: organic remains younger than about 50 thousand years; almost useless for petroleum geology

System	Half-life	What is dated
²³⁸U → ²⁰⁶Pb and ²³⁵U → ²⁰⁷Pb	4.47 and 0.704 billion years	zircon from volcanic ash beds and igneous rocks; the most precise dates, to hundredths of a per cent
⁴⁰K → ⁴⁰Ar (and the ⁴⁰Ar/³⁹Ar method)	1.25 billion years	volcanic rocks and ash beds (sanidine, biotite); glauconite and illite, with caveats
⁸⁷Rb → ⁸⁷Sr	about 50 billion years	igneous and metamorphic rocks — by the isochron method
¹⁸⁷Re → ¹⁸⁷Os	about 42 billion years	black shales rich in organic matter, sulfides; experimental dating of oil
¹⁴C → ¹⁴N	5730 years	organic remains younger than about 50 thousand years; almost useless for petroleum geology

Half-lives are rounded. Different studies adopt different decay constants for ⁸⁷Rb; ages calculated with the old constant are about 1.6% younger (see the advanced section).

A sedimentary rock is the hardest to date. Its detrital grains (grains eroded from older rocks), for example the zircons in a sandstone, formed long before deposition. Their age only shows that the sandstone is not older than its youngest grain. So we find the age of sedimentary sequences indirectly. We use beds of volcanic ash (tuff, bentonite) that fell at the same time as the sediment. We use intrusions that cut the sequence and are therefore younger than it. We use minerals that grew in the sediment as it was deposited, for example glauconite. And above all we use fossils whose time ranges have been calibrated by such dates (see the section on biostratigraphy).

Worked exampleAge of a volcanic ash bed from zircon

Given (assumed data)

Measured ratio of radiogenic ²⁰⁶Pb to ²³⁸U: 0.04553
Measured ratio of radiogenic ²⁰⁷Pb to ²³⁵U: 0.3266
Decay constants: $\lambda_{238} = 1.55125\cdot 10^{-10}$ yr⁻¹, $\lambda_{235} = 9.8485\cdot 10^{-10}$ yr⁻¹

Solution

Age from uranium-238: $t = \ln(1.04553)/(1.55125\cdot 10^{-10}) = 0.044524/(1.55125\cdot 10^{-10}) = 287.02$ million years.
Age from uranium-235: $t = \ln(1.3266)/(9.8485\cdot 10^{-10}) = 0.282619/(9.8485\cdot 10^{-10}) = 286.97$ million years.
The two ages differ by 0.05 million years, about 0.02%. The date is concordant (the two ages agree), so the zircon has not lost lead.
Sensitivity: an error of 0.1% in the ²⁰⁶Pb/²³⁸U ratio changes the age by 0.28 million years, an error of 1% by 2.8 million years.

Answer: The ash fell about 287.0 million years ago — on the 2026/06 version of the scale this is the Artinskian age of the Early Permian (290.1–283.3 million years ago).

What it means. The two uranium isotopes decay at different rates. So when the two ages agree, you have an internal check: if the zircon lost lead or contains an older core, the ages disagree. Zircon crystallises in the magma shortly before the eruption. Its age therefore matches the age of the ash bed, within a fraction of a million years, and so the age of the sediment around it. Modern instruments measure isotope ratios to hundredths of a per cent. So the ages of many boundaries of the scale are known to within tens of thousands of years. For example, the Permian–Triassic boundary is at 251.902 ± 0.024 million years ago. The data in this example are assumed for illustration.

Dated beds let us estimate the age of the beds between them, and the rate of accumulation, as long as there are no breaks between them.

Worked exampleAge of the top of a bed between two dated ash beds

Given (assumed data)

Ash bed A: depth 1842.0 m, age 287.0 million years (the example above, rounded)
Ash bed B: depth 1960.0 m, age 291.4 million years
Top of the reservoir bed: 1905.0 m, between the ash beds; no signs of a break between them

Solution

Average rate of accumulation in the present-day, compacted thicknesses: $(1960.0 - 1842.0)/(291.4 - 287.0) = 118/4.4 = 26.82$ m per million years.
Age of the top: $287.0 + (1905.0 - 1842.0)/26.82 = 287.0 + 2.3 = 289.3$ million years.

Answer: The top of the bed is about 289.3 million years old; sediment accumulated at a rate of about 27 m per million years, that is, about 27 mm per thousand years.

What it means. The interpolation assumes a constant rate of accumulation. A hidden break between the ash beds would make this wrong. If half of the time was a break, the sediment piled up twice as fast during the rest, and the age of the top could be quite different. Also, the thicknesses are compacted. If the porosity of the sediment was 55% when it was deposited and is 10% now, the layer has shrunk by half, and the original rate of accumulation was about 54 m per million years (decompaction). The data in this example are assumed for illustration.

The geologic time scale

The history of the Earth, about 4.57 billion years, is divided into units of several ranks. Together they form the geologic time scale. The largest units are the eons: the Hadean, the Archaean, the Proterozoic (together called the Precambrian) and the Phanerozoic. “Phanerozoic” means “the time of visible life”: it is the last 538.8 million years, which have left rocks with many remains of organisms. The Phanerozoic is divided into three eras: the Palaeozoic, Mesozoic and Cenozoic. The eras are divided into periods, the periods into epochs, and the epochs into ages. Almost all oil and gas accumulations are in Phanerozoic rocks.

CalculatedThe geologic time scale drawn to scale: at the top, the whole history of the Earth with its eons; below, the Phanerozoic with its eras and periods. The ages of the boundaries are from the International Chronostratigraphic Chart, version 2026/06. Brackets mark intervals important for the subsurface of Kazakhstan.

Periods of the Phanerozoic on the 2026/06 version of the scale

Era: Cenozoic
Period (symbol): Quaternary (Q)
Start, million years ago: 2.58
Duration, million years: 2.58

Era: Cenozoic
Period (symbol): Neogene (N)
Start, million years ago: 23.04
Duration, million years: 20.5

Era: Cenozoic
Period (symbol): Palaeogene (Pg)
Start, million years ago: 66.0
Duration, million years: 43.0

Era: Mesozoic
Period (symbol): Cretaceous (K)
Start, million years ago: 143.1 ± 0.6
Duration, million years: 77.1

Era: Mesozoic
Period (symbol): Jurassic (J)
Start, million years ago: 201.4 ± 0.2
Duration, million years: 58.3

Era: Mesozoic
Period (symbol): Triassic (T)
Start, million years ago: 251.902 ± 0.024
Duration, million years: 50.5

Era: Palaeozoic
Period (symbol): Permian (P)
Start, million years ago: 298.9 ± 0.15
Duration, million years: 47.0

Era: Palaeozoic
Period (symbol): Carboniferous (C)
Start, million years ago: 358.86 ± 0.19
Duration, million years: 60.0

Era: Palaeozoic
Period (symbol): Devonian (D)
Start, million years ago: 419.62 ± 1.36
Duration, million years: 60.8

Era: Palaeozoic
Period (symbol): Silurian (S)
Start, million years ago: 443.1 ± 1.0
Duration, million years: 23.5

Era: Palaeozoic
Period (symbol): Ordovician (O)
Start, million years ago: 486.85 ± 1.5
Duration, million years: 43.8

Era: Palaeozoic
Period (symbol): Cambrian (Ꞓ)
Start, million years ago: 538.8 ± 0.6
Duration, million years: 52.0

Era	Period (symbol)	Start, million years ago	Duration, million years
Cenozoic	Quaternary (Q)	2.58	2.58
Cenozoic	Neogene (N)	23.04	20.5
Cenozoic	Palaeogene (Pg)	66.0	43.0
Mesozoic	Cretaceous (K)	143.1 ± 0.6	77.1
Mesozoic	Jurassic (J)	201.4 ± 0.2	58.3
Mesozoic	Triassic (T)	251.902 ± 0.024	50.5
Palaeozoic	Permian (P)	298.9 ± 0.15	47.0
Palaeozoic	Carboniferous (C)	358.86 ± 0.19	60.0
Palaeozoic	Devonian (D)	419.62 ± 1.36	60.8
Palaeozoic	Silurian (S)	443.1 ± 1.0	23.5
Palaeozoic	Ordovician (O)	486.85 ± 1.5	43.8
Palaeozoic	Cambrian (Ꞓ)	538.8 ± 0.6	52.0

± is the uncertainty in the age of the boundary. In Russian the Carboniferous has two names, kamennougolny (“coal-bearing”) and karbon. Russian-language literature divides it into Lower, Middle and Upper series. The international scale divides it into the Mississippian and the Pennsylvanian. Older literature combined the Palaeogene and Neogene into the Tertiary period. On Russian maps the Palaeogene is marked with the sign ₽.

Each division of time has a matching body of rock that was deposited during that time. So there are two parallel sets of divisions, with different names:

Divisions of time and of rock

Time (geochronologic): eon
Rock (chronostratigraphic): eonothem
Example: Phanerozoic

Time (geochronologic): era
Rock (chronostratigraphic): erathem
Example: Palaeozoic

Time (geochronologic): period
Rock (chronostratigraphic): system
Example: Permian Period — Permian System

Time (geochronologic): epoch
Rock (chronostratigraphic): series
Example: Early Permian epoch — Lower Permian (Cisuralian) series

Time (geochronologic): age
Rock (chronostratigraphic): stage
Example: Kungurian age — Kungurian stage

Time (geochronologic)	Rock (chronostratigraphic)	Example
eon	eonothem	Phanerozoic
era	erathem	Palaeozoic
period	system	Permian Period — Permian System
epoch	series	Early Permian epoch — Lower Permian (Cisuralian) series
age	stage	Kungurian age — Kungurian stage

So we say “rocks of the Permian System” or “Permian rocks”, but “in the Permian Period”; and “Lower Permian rocks”, but “in the Early Permian epoch”.

A division's boundary is fixed not by a number but by a place in a section. The lower boundary of a stage is defined at one reference section in the world, the Global Boundary Stratotype Section and Point (GSSP, the “golden spike”). It is usually placed at the first appearance of a particular fossil species. Such points have already been ratified (officially approved) for most stages of the Phanerozoic. The numerical age of a boundary comes from dates of ash beds near it and from the astronomical cycles recorded in the sediments, so the age gets refined over time. The International Commission on Stratigraphy maintains the scale. For example, scales of the previous decade put the beginning of the Cretaceous Period at about 145 million years ago, while version 2026/06 gives 143.1 ± 0.6 million years ago (the reference section for this boundary has not yet been ratified). In the same version the boundaries of several Permian and Triassic stages have moved by 0.3–0.9 million years. This is why reports state which version of the scale their ages come from.

Lithostratigraphy: formations, members and their boundaries

Lithostratigraphy divides a section by its rocks. The boundaries are drawn where the lithology (rock type) changes: sandstone gives way to shale, shale to limestone. The basic local unit in the practice of the CIS countries is the suite (Russian svita), close to the English formation. It is a body of rock that differs in composition from its neighbours and can be traced on maps and in wells. Suites are combined into groups and divided into subsuites and members. They are named after the place where the unit was first described, and where its reference section, the stratotype, was chosen. Besides the local units there are regional ones, called horizons, which are traced across a whole region. In field practice, the productive beds and horizons of a field are usually numbered or given indices, and the field development plans call them by these.

The main feature of boundaries between rock units is that they are diachronous: they cross lines of equal time. A rock reflects the environment where it was deposited — beach, shelf (the shallow sea floor next to a coast), delta — and environments shift with time. When the sea advances onto the land (a transgression), coastal sands are deposited farther and farther towards the land. Shales from the seaward part of the shelf are laid over them, and open-sea limestones are laid over those. A sand member looks like a single layer in a section, but at the edge of the basin it is younger than in the deep part of the basin. When the sea retreats (a regression), the picture is reversed.

SchematicA transgression (schematic): sands near the shore, and shales and limestones farther from it, are deposited at the same time and shift together towards the land. The boundaries between the rocks cross the time lines

t_1

–

t_4

(dashed); in wells A and B the base of “the same” sandstone is a different age.

How big is the difference? Say the shoreline moves towards the land by 20 km per million years (2 cm a year). Then the base of the sandstone in two wells 30 km apart, along the direction the shoreline moves, differs in age by 1.5 million years. For an estimate of hydrocarbons in place this does not matter, because it is one sandstone and it is connected. But a line drawn along the base of the sandstone is not a time line. A thin shale bed deposited at a single moment crosses such a member and may split it into poorly connected parts. So correlation by lithology is checked against fossils and marker beds (Well correlation). Walther's law links the vertical succession of rocks to the lateral change of environments: only rocks of adjacent environments lie directly on one another without breaks (Facies, facies successions and sedimentary logs).

Biostratigraphy, chronostratigraphy and other ways of dating layers

Biostratigraphy divides a section by the remains of organisms (fossils). Species appear, live for a while and die out, and each interval of a section holds its own set of fossils. The levels where a species first and last appears (first and last appearance datums, FAD and LAD) serve as markers. The intervals between them are biostratigraphic zones (biozones). The time ranges of species have been calibrated by dating ash beds, so a zone also gives an age. The best index fossils evolved quickly, spread widely and did not depend on the environment. In the sea they are planktonic organisms (tiny floating organisms); on land they are the pollen and spores of plants. Species that lived only on one kind of sea floor tell us about the environment of deposition, but they date the rock poorly.

In wells, geologists study microfossils (tiny fossils) that survive in cuttings and core: foraminifera, ostracods, calcareous nannoplankton, radiolarians, conodonts (especially important for Palaeozoic carbonates), spores and pollen. Cuttings have one peculiarity. Fragments from the borehole wall higher up cave in and end up in samples from greater depths. So cuttings reliably give the depth at which a species first appears in the samples as drilling goes from top to bottom, that is, the top of its range. The depth below which the species is no longer found may be pushed downwards by cavings (the fragments that fell from above), so it is checked against core and sidewall core.

Chronostratigraphy picks out bodies of rock that were deposited during a set interval of time: systems, series and stages. By definition its boundaries are isochronous (they follow lines of equal time). They are found from all the available signs of time: fossils, dates, and events that affected a whole basin or the whole planet.

Magnetostratigraphy. The Earth's magnetic field has reversed its polarity hundreds of times, and a rock keeps the magnetisation of the time when it formed. The alternation of normal and reversed magnetisation in a section is compared with a reference polarity scale. The scale is built from dated rocks and from the magnetic stripes of the ocean floor (spreading rate). Polarity reversals happen almost at the same time across the whole planet, so their levels serve as time lines.
Chemostratigraphy. The carbon and strontium isotope composition of seawater changed with time in almost the same way throughout the ocean, and carbonates record it. It is used to correlate distant sections and to refine ages.
Event beds. Beds of volcanic ash that fell within a few days are the best markers. They are isochronous and can be dated from zircon, and on well logs the gamma ray log often shows them.
Cyclostratigraphy. Variations in the Earth's orbit and axis of rotation, with periods of about 20, 40, 100 and 405 thousand years, changed the climate and left rhythms in the sediments. By counting the rhythms, we can measure time to within tens of thousands of years.
Seismic stratigraphy. Reflecting boundaries on seismic sections mostly follow bedding surfaces. So they follow time lines, approximately, and not boundaries between rock types (Seismic interpretation).

Three approaches to dividing a section

Approach: Lithostratigraphy
Defined by: composition and texture of the rocks
Boundaries: diachronous: they slide through time
Units: group, formation (suite), member, bed

Approach: Biostratigraphy
Defined by: assemblages of fossils
Boundaries: close to isochronous if the fossils do not depend on the environment
Units: biostratigraphic zone

Approach: Chronostratigraphy
Defined by: time of deposition
Boundaries: isochronous by definition
Units: system, series, stage

Approach	Defined by	Boundaries	Units
Lithostratigraphy	composition and texture of the rocks	diachronous: they slide through time	group, formation (suite), member, bed
Biostratigraphy	assemblages of fossils	close to isochronous if the fossils do not depend on the environment	biostratigraphic zone
Chronostratigraphy	time of deposition	isochronous by definition	system, series, stage

Sea level, accommodation and the basics of sequence stratigraphy

Three quantities decide where sands and shales are deposited and how the layers lie. The first is tectonic subsidence (or uplift) of the basin floor (why basins sink). The second is eustatic change in the level of the world ocean (eustasy). Glaciers growing and melting, and changes in the volume of the ocean basins, cause it. It is global. Together, these two give the relative change of sea level at a given place, and they create or destroy accommodation (space for sediment). The third quantity is the sediment supply from the land.

Rate at which space for sediment is created

\dot A = \dot S + \dot E

where:

$\dot A$ rate of relative sea-level rise, that is, of the creation of space for sediment; a negative value means a relative fall of sea level, m/Myr
$\dot S$ rate of tectonic subsidence of the floor; uplift is taken with a minus sign, m/Myr
$\dot E$ rate of eustatic change of sea level; a rise is taken with a plus sign, m/Myr

Units: m/Myr (metres per million years, the same as mm per thousand years)

Holds when

Compaction of previously deposited sediment is included in the subsidence.
The bending of the lithosphere under the weight of the sediment and the water is not separated from the tectonic subsidence (isostasy).

If less sediment arrives than the space that is created, the water gets deeper and the shoreline moves back towards the land. This is a transgression, and the layers step back (retrogradation). If supply and space are equal, the layers build upwards (aggradation). If more sediment arrives than there is room for, the shoreline advances into the basin and the layers prograde (progradation, building outwards). This is a regression with a rising or constant sea level. When the relative sea level falls ( $\dot A < 0$ ), the regression is forced: the shelf dries out and erodes, rivers cut into it, and the sediment goes to the foot of the slope.

Worked exampleWill there be a transgression?

Given (assumed data)

Tectonic subsidence of the shelf: 30 m per 1 million years
Sediment supply: enough to build up the shelf by 25 m per 1 million years
Eustatic change of sea level: three cases: a rise of 10, a fall of 5 and a fall of 40 m per 1 million years

Solution

Rise of 10 m/Myr: $\dot A = 30 + 10 = 40$ m/Myr, more than the supply (25). The water depth increases by 15 m per million years, and the shoreline retreats towards the land: a transgression.
Fall of 5 m/Myr: $\dot A = 30 - 5 = 25$ m/Myr, the same as the supply of sediment. The depth does not change, and the layers build upwards.
Fall of 40 m/Myr: $\dot A = 30 - 40 = -10$ m/Myr. The relative level is falling although the floor is sinking: the shelf dries out and erodes, and a sequence boundary forms.

Answer: With the same subsidence and the same supply of sediment, the shelf responds with a transgression, with upward building, or with a forced regression with erosion — depending on the rate of the ocean level change.

What it means. A fall of ocean level of 5 m per million years did not cause a regression, because tectonic subsidence outweighed it. The relative level counts, not the global one. So in basins that sink at different rates, the same changes of ocean level leave different records. Real rates change with time and along the basin (near the shore the subsidence is slower than on the outer shelf), so the geometry of the layers is more complicated. The data in this example are assumed for illustration.

The change from one regime to another is recorded in the geometry of the layers. On seismic sections you can see it in how the reflecting boundaries end:

Onlap: gently dipping layers rest against a steeper, older surface and reach farther and farther up it. This is a sign of growing accommodation, for example in a transgression.
Downlap: tilted layers of a prograding body lie down onto the surface beneath them, thinning out down the slope.
Toplap: the upper ends of the tilted layers meet the surface above them at an angle. At the top, sediment did not accumulate; it passed through on its way elsewhere.
Truncation: layers are cut off by an erosion surface. This is a sign of an unconformity.

Russian translations of these terms differ from one source to another, so people often use the English terms.

SchematicHow layers end, in cross-section (schematic): erosional truncation of tilted layers beneath an unconformity, onlap of horizontal layers onto a slope, and downlap and toplap of prograding clinoforms (inclined layers that build outwards).

Sequence stratigraphy divides a section into sequences: bodies of layers that belong together by origin. Unconformities bound them, and in the deep parts of the basin the conformable surfaces into which the unconformities pass. Each sequence reflects one cycle of relative sea level. It consists of systems tracts, bodies that built up at different stages of the cycle:

Systems tracts (simplified)

Tract: Falling-stage and lowstand (FSST, LST)
When it accumulates: the level falls and then begins to rise slowly
What is deposited: the shelf is eroded, rivers cut valleys; sands go to the foot of the slope — into submarine fans

Tract: Transgressive (TST)
When it accumulates: rapid relative rise: space grows faster than the supply of sediment
What is deposited: retrograding members; at the top, the maximum flooding surface (MFS): thin shales with plankton, often rich in organic matter

Tract: Highstand (HST)
When it accumulates: the rise slows down
What is deposited: prograding deltas and coastal sands

Tract	When it accumulates	What is deposited
Falling-stage and lowstand (FSST, LST)	the level falls and then begins to rise slowly	the shelf is eroded, rivers cut valleys; sands go to the foot of the slope — into submarine fans
Transgressive (TST)	rapid relative rise: space grows faster than the supply of sediment	retrograding members; at the top, the maximum flooding surface (MFS): thin shales with plankton, often rich in organic matter
Highstand (HST)	the rise slows down	prograding deltas and coastal sands

The scheme is idealised. In real basins the tracts are only partly developed, and geologists draw their boundaries from the combined evidence of core, logs and seismic data.

For a petroleum engineer, sequence stratigraphy is a way to predict where reservoirs and seals will be, and how they are connected. The shales of the maximum flooding surface keep their character over large areas. They are regional seals and reliable markers for correlation, and on logs a gamma ray maximum often shows them (Gamma ray log and natural gamma ray spectrometry). Sandstones of incised valleys (valleys cut into the shelf by rivers) and of lowstand submarine fans can form stratigraphic traps. Correlation by sequences links bodies of the same age, whereas correlation of “sand with sand” may link bodies of different ages (Well correlation).

Stratigraphy in the engineer's work

Names in documents. Beds and development units are called by the names of formations, horizons and stages, and by local indices. A composite log of a well shows depths, ages, formations, lithology and logs side by side.
Marker beds. Members of shale, limestone, coal and ash that can be followed over a large area, and have a recognisable log signature, let you correlate wells (Well correlation).
The expected section while drilling. A well plan is based on the stratigraphic column, which shows at what depth to expect salt, limestones and abnormal-pressure zones. Biostratigraphy on cuttings while drilling shows where the bit is relative to the target bed.
Pressure and fluids. Beds separated by laterally continuous shales often have different reservoir pressures and different fluid contacts. Stratigraphic boundaries show where to expect a change in pressure (Abnormal formation pressures: origin, prediction and recognition).
Basin history. The ages of rocks and of breaks are the basis of burial history curves and maturation models (burial history curves, Source rocks, maturation and migration).

Terms and symbols

Stratigraphy: The part of geology that studies the order, age, composition and extent of layers of rock.
Law of superposition: In undisturbed bedding each layer is younger than the one beneath it.
Cross-cutting relationships: Faults, intrusions and erosion surfaces are younger than all the rocks they cut.
Unconformity: A surface that separates two sequences with a break in deposition, or erosion, between them; the types are angular, parallel, hidden and on basement.
Hiatus (stratigraphic break): An interval of time not represented by rocks in the section.
Relative and numerical age: Relative age is the order of events; numerical age is the age in years, found from radioactive decay.
Half-life $t_{1/2}$: The time in which half of the nuclei of a radioactive isotope decay: $t_{1/2} = \ln 2/\lambda$ .
Concordant age: A uranium–lead age that is the same from ²³⁸U and from ²³⁵U: a sign that the mineral has not lost lead.
Geologic time scale: Eons, eras, periods, epochs and ages with the numerical ages of their boundaries; the international scale is revised periodically.
System, series, stage: The bodies of rock deposited during a period, an epoch and an age (chronostratigraphic units).
Global Boundary Stratotype Section and Point (GSSP): The place in a reference section where the lower boundary of a stage is defined.
Suite (formation): The basic local lithostratigraphic unit — a body of rock that differs in composition from its neighbours; suites are combined into groups and divided into subsuites and members.
Diachroneity: The failure of a lithological boundary to coincide with a time line: the same rock has different ages in different places.
Biostratigraphic zone: An interval of a section defined by fossils, for example between the levels of the first and last appearance of a species.
Accommodation (space for sediment): The volume below the level up to which sediment can accumulate; it is created by subsidence of the floor and a rise in sea level.
Transgression and regression: Movement of the shoreline towards the land and towards the basin.
Sequence and systems tract: A body bounded by unconformities and the conformable surfaces corresponding to them, and the part of it that accumulated at a particular stage of the sea-level cycle.
Maximum flooding surface: The top of the transgressive tract: persistent thin shales, a frequent marker and regional seal (MFS).

Common mistakes and misconceptions

Wrong

“A sandstone at the same stratigraphic level in two wells is therefore of the same age.”

In fact

Lithological boundaries are diachronous: in a transgression a coastal sandstone at the edge of the basin is younger than one in its depths. Contemporaneity is checked with fossils, marker beds and correlation by sequences.

Wrong

“The age of zircons from a sandstone is the age of the sandstone.”

In fact

It is the age of the eroded rocks of the source area; the sandstone is not older than its youngest grain. Ash beds, fossils and cross-cutting intrusions give the age of deposition.

Wrong

“Since the ocean level fell, there was a regression on the shelf.”

In fact

What decides is the relative sea level — the sum of the eustatic change and the tectonic subsidence — and the supply of sediment. If subsidence is fast, a fall in ocean level can pass without a regression.

Wrong

“‘Lower Jurassic’ and ‘Early Jurassic’ are the same thing.”

In fact

“Lower” refers to rocks (a series), “early” to time (an epoch): Lower Jurassic rocks accumulated in the Early Jurassic epoch.

Wrong

“A section 2 km thick accumulated over 100 million years, so sediment was deposited at 20 m per million years.”

In fact

That is the average rate, including all the breaks and erosion; while deposition was going on it was much higher. Calculating average rates across unconformities is meaningless.

In depth: advanced detail

The isochron: an age without knowing the initial daughter isotope

Minerals usually contain some daughter isotope from the very beginning. The simple formula treats all the daughter isotope as made by decay, so it overestimates the age. The isochron method gets round this problem. Take several minerals or samples from one intrusion. They formed at the same time from one melt, so they have the same initial strontium isotope ratio ⁸⁷Sr/⁸⁶Sr, but different amounts of rubidium. Divide the numbers of atoms by the number of atoms of the stable, non-radiogenic isotope ⁸⁶Sr:

The isochron of the rubidium–strontium system

\frac{{}^{87}\mathrm{Sr}}{{}^{86}\mathrm{Sr}} = \left(\frac{{}^{87}\mathrm{Sr}}{{}^{86}\mathrm{Sr}}\right)_0 + \left(e^{\lambda t} - 1\right)\frac{{}^{87}\mathrm{Rb}}{{}^{86}\mathrm{Sr}}

where:

$\frac{{}^{87}\mathrm{Sr}}{{}^{86}\mathrm{Sr}}$ measured ratio of strontium isotopes in the sample, —
$\left(\frac{{}^{87}\mathrm{Sr}}{{}^{86}\mathrm{Sr}}\right)_0$ initial ratio, the same for all samples, —
$\frac{{}^{87}\mathrm{Rb}}{{}^{86}\mathrm{Sr}}$ measured ratio of the parent ⁸⁷Rb to ⁸⁶Sr in the sample, —
$\lambda$ decay constant of ⁸⁷Rb, $1.3972\cdot 10^{-11}$ , yr⁻¹
$t$ age, years

Units: ratios of numbers of atoms, dimensionless

Holds when

The samples are contemporaneous and had the same initial strontium isotope ratio.
The system remained closed after it formed.

On a graph of “⁸⁷Rb/⁸⁶Sr — ⁸⁷Sr/⁸⁶Sr” the points lie on a straight line, the isochron (a line of equal age). Its slope equals $e^{\lambda t} - 1$ and gives the age. The intercept on the vertical axis is the initial ratio, which also carries information: what material the melt came from. If the points do not lie on a line, the system has been disturbed.

Worked exampleAge of a granite from an isochron

Given (assumed data)

⁸⁷Rb/⁸⁶Sr in four samples: 0.62; 2.15; 4.80; 8.90
⁸⁷Sr/⁸⁶Sr in the same samples: 0.70823; 0.71670; 0.73136; 0.75405
Decay constant of ⁸⁷Rb: $1.3972\cdot 10^{-11}$ yr⁻¹

Solution

Slope from the extreme points: $(0.75405 - 0.70823)/(8.90 - 0.62) = 0.005534$ ; from all four points by least squares, 0.005534, and the intercept on the vertical axis is 0.70480.
Age: $t = \ln(1 + 0.005534)/(1.3972\cdot 10^{-11}) = 395.0$ million years.
With the decay constant $1.42\cdot 10^{-11}$ yr⁻¹, adopted in studies of past decades, the result would be 388.6 million years.

Answer: The granite formed about 395 million years ago — in the Early Devonian on the 2026/06 version of the scale; the initial ⁸⁷Sr/⁸⁶Sr ratio is about 0.7048.

What it means. The two decay constants give ages that differ by more than six million years. This is why only ages calculated with the same constants are compared. The isochron age refers to the moment when the minerals stopped exchanging strontium. For a granite this is the moment of cooling, not of emplacement (intrusion), and for slowly cooled massifs the difference can be considerable. The data in this example are assumed for illustration.

Connections to other topics

Well correlation — how beds are correlated between wells by lithology, fossils and sequences.
Traps and timing — traps beneath unconformities, burial history curves and the timing of trap formation.
Plate tectonics and sedimentary basins — why basins sink and create space for sediment.
Source rocks, maturation and migration — how the age of the rocks and the burial history determine maturation.
Facies, facies successions and sedimentary logs — facies, Walther's law and vertical successions of rocks.
Primary sedimentary structures — the original structures of sediment, which show the conditions of deposition and which way up a bed is.
Stress, strain and structural analysis — strain, the tilting of layers and how it is measured.
Seismic interpretation — seismic stratigraphy and the tracing of horizons.
Gamma ray log and natural gamma ray spectrometry — marker shales and ash beds on logs.

Check yourself

1
A fault displaces Jurassic layers but does not disturb the Cretaceous layers lying on them. What can be said about its age?
1. A.It is younger than the displaced Jurassic layers and older than the overlying Cretaceous ones
2. B.It is older than the Jurassic layers
3. C.The age of the fault cannot be determined from the section
4. D.It is contemporaneous with the Cretaceous layers
Show the answer
Correct answer: A. It is younger than the displaced Jurassic layers and older than the overlying Cretaceous ones
By the rule of cross-cutting relationships a fault is younger than all the rocks it cuts, while the layers that overlie it were deposited after it. Its age lies between the time of deposition of the youngest displaced layers and that of the oldest overlying ones.
2
What fraction of the uranium-238 and what fraction of the uranium-235 in a zircon has decayed in 287 million years? Why are both systems measured?
Show the answer
What remains is $e^{-\lambda t}$ : for ²³⁸U $e^{-1.55125\cdot 10^{-10}\cdot 2.87\cdot 10^{8}} = e^{-0.0445} \approx 0.956$ , for ²³⁵U $e^{-0.2827} \approx 0.754$ . So about 4.4% of the ²³⁸U and 25% of the ²³⁵U have decayed. The two “parent — daughter” pairs are two independent clocks that run at different speeds. If the ages agree, the zircon has not lost lead. If they disagree, the zircon lost lead or has an older core. Natural uranium contains only about 0.72% ²³⁵U, so young zircons contain very little ²⁰⁷Pb, and the age from ²⁰⁶Pb/²³⁸U is more precise.
3
Beneath an unconformity in a well are limestones of the Visean stage (346.7–330.3 million years ago); above it are shales of the Kungurian stage (283.3–274.4 million years ago). Within what limits is the time missing at the unconformity?
Show the answer
The smallest gap runs from the end of the Visean age to the beginning of the Kungurian: $330.3 - 283.3 = 47.0$ million years. The largest runs from the beginning of the Visean to the end of the Kungurian: $346.7 - 274.4 = 72.3$ million years. The section has no Serpukhovian stage, no Middle or Upper Carboniferous, and none of the Lower Permian below the Kungurian.
4
Sands are deposited near the shore, shales farther out in the sea and limestones farther still. The shoreline advances into the sea. What will be the succession of rocks in a single well, from the bottom up?
1. A.No regular pattern
2. B.Limestone — shale — sandstone
3. C.Shale — sandstone — limestone
4. D.Sandstone — shale — limestone
Show the answer
Correct answer: B. Limestone — shale — sandstone
In a regression, environments closer to the shore move over a point that was open sea. So shale is deposited on the limestone, and sandstone on the shale. Each of these members is diachronous: it is younger in the direction of the basin.
5
A shelf sinks at 50 m per million years, the ocean level falls by 20 m per million years, and the sediment supply is such that it could build the shelf up by 45 m per million years. What happens to the shoreline?
Show the answer
$\dot A = 50 - 20 = 30$ m/Myr, which is less than the supply of 45 m/Myr. There is more sediment than room for it, so the shoreline advances into the basin. This is a regression, even though relative sea level is rising. It is a normal regression, not a forced one: the shelf is not exposed, and the layers prograde while also building upwards.
6
Which of the boundaries on a well log is most likely to be a line of one time over a large area?
1. A.The base of a member of coastal sands
2. B.The top of a reef
3. C.The top of a channel sandstone
4. D.A volcanic ash bed
Show the answer
Correct answer: D. A volcanic ash bed
Ash falls within a few days over a vast area, whatever the environment, and zircon can date it. Channel sands, coastal sands and reefs are bodies tied to their environment of deposition, and their boundaries slide through time.

Updated: 12 October 2026