◀ Geology

Sediment · transport · landscapes

Source to Sink

Follow one grain from the mountain that made it to the seafloor that buries it — and learn the whole surface system on the way.

Everything you weather off a mountain has to go somewhere. This is the machine that moves it: water and wind pick up grains, carry them by size and energy, and drop them in places that leave a signature you can read back. Master this and you've got the core of both Sedimentary Geology and Geomorphology in one system.

The source-to-sink profile, mountain to deep sea. Tap any station to see what happens there — the process, the grain size that settles, the structure it leaves, and the landform it builds. No pressure, just wander downstream.

Whether a grain erodes, travels, or settles comes down to two things: how big it is and how fast the water moves. This is the Hjulström diagram — set the current and watch which zone your grain falls into. Then take the challenges.

30 cm/s a brisk stream
Transport Set a grain and a current to see what happens.

Challenge

Loading…

Solved 0 Streak 0

A deposit is a confession. Grain size records the energy, sorting and roundness record the journey, and structures record the flow. Read the clues and name the environment that made this rock.

The specimen

Read 0Streak 0

“Well sorted” is a useful thing to say and a hopeless thing to compare. Textural analysis turns the same observation into numbers — and once a distribution is numbers, two samples from opposite sides of the world can be put on the same axis.

The phi scale

Grain size spans seven orders of magnitude, so it is worked in logarithms. The phi scale is a negative base-2 log of the diameter in millimetres, which makes each whole phi step a halving of grain size — exactly the Wentworth boundaries — and turns most natural distributions into something close to a normal curve you can do statistics on.

φ = −log2( d in mm )     d = 2−φ
−6φ64 mmcobble
−2φ4 mmpebble
−1φ2 mmgranule
1 mmv. coarse sand
250 µmfine sand
63 µmsilt
4 µmclay
The trap worth knowing about. Phi runs backwards to millimetres — bigger phi means smaller grains. So the percentile labels invert: φ84 is the D16 diameter, and φ16 is D84. Get this the wrong way round and your sorting comes out negative, which is the usual sign it has happened.

Analyse a distribution

Seven sedimentary environments, each a realistic grain-size distribution. Pick one and read what the statistics say about it — then compare the numbers with the environment they came from.

Folk & Ward — the sedimentologist’s set
Mean, Mz
Sorting, σI
Skewness, SkI
Kurtosis, KG

Verbal classification

USCS — the engineer’s set
Gravel  > 4.75 mm
Sand  4.75–0.075 mm
Fines  < 0.075 mm
D10 / D30 / D60
Cu = D60/D10
Cc = D30²/(D10D60)

Grading

How grain size is actually measured

No single technique spans the whole range, so any full distribution is stitched together from two or more — and the joins are where errors live.

MaterialSize rangeMethodWhat it actually measures
UnconsolidatedBoulders – pebblesDirect measurement with calipersThe three axes, one clast at a time
Pebbles – sandSieve stack, shakenThe intermediate axis — whatever squeezes through a square aperture. A blade passes a sieve its length never would
Silt – claySettling: pipette, hydrometer, SedigraphSettling velocity, converted to an equivalent spherical diameter through Stokes’ Law
ConsolidatedConglomerate, brecciaDirect measurement on outcrop or slabAxes of exposed clasts
Sandstone, siltstoneThin section, traverse methodApparent diameters in a 2-D cut — a random slice rarely passes through a grain’s centre, so thin-section sizes are systematically too small and need a correction
Laser diffractionWhole range, small sample, fastDiffraction pattern converted to an equivalent spherical diameter. Fast and reproducible, but it does not measure the same thing a sieve does — results from the two are not interchangeable

Stokes’ Law, and why silt takes all afternoon

Settling methods work because a small sphere falling through a viscous fluid reaches a terminal velocity that depends on the square of its diameter. Halve the grain size and it settles four times slower.

v =  s − ρl) g d2
18 µ
Diameter
Grain density ρs2650
Water temp20 °C

Where it stops working

Stokes’ Law assumes laminar flow around the grain, which holds only while the Reynolds number stays below about 1. For quartz in water that fails at roughly 100 µm — above it, turbulent wake drag takes over and Stokes overestimates the settling velocity badly.

Which is a convenient accident: sieving becomes impractical below about 63 µm, and Stokes becomes valid at about 100 µm. The two methods hand over almost exactly at the sand–silt boundary — the overlap is small, but it exists.

It also assumes spheres. Real platy clays settle more slowly than their mass implies, so a settling method reports them as finer than they are — one more reason sieve, settling and laser results on the same sample do not agree.

Folk & Ward, in full

Graphical measures, computed from percentiles read off the cumulative curve. All values in phi.

StatisticFormulaWhat it tells you
MeanMz = (φ16 + φ50 + φ84) / 3Average size. Uses three percentiles rather than the median alone, so a skewed tail still registers
SortingσI = (φ84−φ16)/4 + (φ95−φ5)/6.6The energy indicator. Spread of the distribution — low means one process worked the sediment hard and repeatedly
SkewnessSkI = (φ1684−2φ50) / 2(φ84−φ16)  +  (φ595−2φ50) / 2(φ95−φ5)Which tail is longer. Positive = a fine tail (fines that settled out); negative = a coarse tail (a lag the current could not lift)
KurtosisKG = (φ95−φ5) / 2.44(φ75−φ25)Peakedness. A sharply peaked curve with wide tails often means sediment sorted elsewhere then dumped somewhere new
σI — sortingSkI — skewnessKG — kurtosis
< 0.35 very well sorted+0.3 to +1.0 very fine skewed< 0.67 very platykurtic
0.35–0.50 well sorted+0.1 to +0.3 fine skewed0.67–0.90 platykurtic
0.50–0.70 moderately well sorted+0.1 to −0.1 symmetrical0.90–1.11 mesokurtic
0.70–1.00 moderately sorted−0.1 to −0.3 coarse skewed1.11–1.50 leptokurtic
1.00–2.00 poorly sorted−0.3 to −1.0 very coarse skewed1.50–3.00 very leptokurtic
2.00–4.00 very poorly sorted> 3.00 extremely leptokurtic
> 4.00 extremely poorly sorted

Two schemes, two purposes

Graphical versus moment methods
Folk & Ward is graphical — it uses six percentiles and ignores everything between them. Moment statistics (Krumbein & Pettijohn) use every class in the distribution, so they are more complete but far more sensitive to the tails. Where a distribution is open-ended — a laser analyser that stops at 0.04 µm, a sieve stack with material in the pan — the moment values shift with where you truncated, while the graphical ones barely move. That is why sedimentology kept the graphical set, and why software such as GRADISTAT reports both and lets you see the disagreement.
Why the engineers use different numbers
USCS asks a different question: not “what environment made this” but “will it hold up a building”. So it splits at sieve sizes that matter mechanically — the No. 4 (4.75 mm) and No. 200 (0.075 mm) — and measures spread with Cu and Cc rather than σI. A well-graded soil packs densely and bears load because small grains fill the gaps between large ones; a uniform one does not. It is the same curve read for a different purpose — and it is the only textural scheme on the ASBOG exam.
Sorting is the environment indicator
Of all these numbers, sorting discriminates depositional environment best, because it records how many times the sediment was picked up and put down. A beach reworks the same grains twice a day for centuries and ends up very well sorted with almost no fines. A glacier carries everything it scrapes up and drops the lot together, so till spans clay to boulders with no sorting at all — and its Cu runs into the tens or hundreds. Between those, a tidal channel alternates energy and traps mud in the slack water, giving a poorly sorted, fines-rich mixture. Run the three through the analyser above and the sorting value alone separates them.

Size tells you about energy. Shape tells you about history — how far a grain travelled, by what, and how many times. And fabric, the way grains sit against one another, records the moment of deposition and everything that happened to the deposit afterwards.

Four things people mean by “shape”

1 · Form
Gross shape, from the ratios between the three mutually perpendicular axes — long (a), intermediate (b) and short (c). The Zingg scheme divides these into spheroid, disc, rod and blade. Form is largely inherited from the parent rock’s fabric — a schist yields blades because mica sheets dictate it — so it says more about source than about transport.
2 · Sphericity
How close the grain is to equidimensional, independent of how sharp its corners are. It governs settling velocity — a sphere settles faster than a plate of the same mass — and how readily a clast rolls in traction transport. The Corey shape factor ψ = c / √(ab) is the usual measure; a sphere gives 1.
3 · Roundness
Sharpness of the corners, judged against the Powers (1953) visual chart: very angular, angular, subangular, subrounded, rounded, well rounded. This is the abrasion record — and it is a function of grain composition and hardness, grain size, transport process, and distance travelled. Independent of sphericity: a grain can be a well-rounded blade.
4 · Surface texture
Features too small to affect gross shape — polished (water-worked), frosted (a matte finish from wind impact or chemical etching), or pitted, scratched and fractured. Frosted sand is a strong hint of an aeolian leg in the journey even when everything else has been reworked since.

The Powers chart, with sphericity

Roundness runs left to right, sphericity bottom to top — the two vary independently, which is the point the grid makes better than any sentence. Click a grain.

Textural maturity

Sorting and roundness improve together as sediment is worked, so they can be combined into a single ladder describing how much reworking a deposit has had. It is the clearest single link between a hand specimen and a depositional environment.

Stage 1

Immature

Poorly sorted, angular grains, and clay-grade matrix still present. Deposited close to source with little reworking — alluvial fans, till, debris flows.

Stage 2

Submature

Matrix largely winnowed out, but sorting is still poor to moderate and grains remain angular to subangular. Some current work has happened.

Stage 3

Mature

No matrix, moderately to well sorted, grains subangular to subrounded. Sustained transport or repeated reworking — most river and shallow-marine sands.

Stage 4

Supermature

Very well sorted, narrow unimodal distribution, grains rounded to well rounded. Beaches and dunes, or sediment recycled through several cycles.

Two cautions. Maturity is not age — a supermature sand can be young if the environment worked it hard, and an immature one can be ancient. And durability confounds it: a quartz-only sand may look supermature simply because everything else dissolved or broke down, not because it travelled further. Compositional and textural maturity usually track each other, but not always.

Fabric — how the grains sit

Orientation, and what it records
In conglomerates, discs and blades stack overlapping like fallen dominoes — imbrication — with their flat faces dipping upstream, which makes them one of the most reliable palaeocurrent indicators available. In sandstones, long axes align parallel to flow under upper-flow-regime conditions. In mudrocks, platy clays settle and then compact parallel to bedding, which is what produces fissility — the tendency to split into sheets that distinguishes a shale from a massive mudstone.
Grain support: who is carrying the load
In a clast-supported deposit the framework grains touch one another and bear the weight; matrix, if any, merely fills the gaps. In a matrix-supported deposit the large clasts float in finer material and never touch — which means the transporting medium had enough strength to hold them apart, so it was a debris flow or ice, not a current. That single observation separates a conglomerate from a diamictite.
Grain contacts record burial
As a sand is buried and compacted, its contacts evolve: point contacts → long (straight) contacts → concavo-convex, where one grain is pressed into another → sutured, an interlocking seam produced by pressure solution. The sequence is a burial-depth gauge, and each step destroys porosity. Distinguish genuine grain-to-grain contacts from cement-to-cement ones, which are chemically precipitated after deposition and mean something quite different.
Porosity is not permeability
Porosity is how much pore space there is; permeability is how well connected it is. Think of a car park: many spaces reached by narrow, disconnected lanes gives high porosity and low permeability — which is exactly a clay, at 50–60% porosity and effectively impermeable. Fewer spaces with wide connecting roads gives moderate porosity and high permeability, which is a clean sand. Fabric, sorting and grain shape all act on the connections rather than on the volume.

Your digital grain-size card and cheat sheet — the Wentworth scale, what sorting and roundness look like, and the structures that give an environment away.

Wentworth grain-size scale

Coarsest at the top. Each step is a doubling (or halving) of diameter — the same φ (phi) steps you'll measure in lab.

Sorting

How uniform the grains are. Well-sorted = a long, selective journey (wind, waves); poorly sorted = dumped fast (glaciers, debris flows).

Roundness

How worn the corners are. Angular = close to source; well-rounded = carried a long way, especially by water.

Chemical weathering & the CIA

Hydrolysis strips cations out of feldspar and leaves clay behind. The Chemical Index of Alteration turns that into a number you can compare between rocks and climates.

Bowen’s series, run backwards — the Goldich sequence

The order minerals crystallise out of a melt is very nearly the order they break down at the surface, reversed. A mineral that formed at 1200 °C and 30 km depth is furthest from equilibrium in cool, wet, oxygenated air — so it goes first. Quartz, last to crystallise and closest to surface conditions, survives to become the beach.

Click least stable first. The series is branched, not a single line — the ferromagnesian chain and the plagioclase chain weather in parallel and only merge at potassium feldspar. Either branch may be advanced at any point.

The A–CN–K diagram

The CIA becomes far more useful plotted than tabulated. Cast a rock as three molar quantities — A = Al2O3, CN = CaO* + Na2O, K = K2O — and its position in the triangle is its CIA: the value equals the point’s height above the base. Fresh igneous rocks cluster along the bottom near 50; weathering drives samples upward toward the clays.

Al2O350
CaO* + Na2O32
K2O18

Molar proportions, any consistent scale — only the ratios matter.

CaO* is the calcium held in silicates only. Carbonate and apatite must be subtracted first, or a limey mudstone will read as unweathered when it is nothing of the kind — the commonest way to get a CIA badly wrong.

Weathering by-products — what goes where

You do not need the balanced equations. You need the four slots: what you started with, what attacked it, what was left behind as a solid, and what was carried away dissolved. Get those and the rest is bookkeeping — because those two outputs become two entirely different rocks.

 
You start with
It reacts with
↓   produces   ↓
Solid by-product
Dissolved by-product
Right 0Asked 0

The distinction that organises all of it:

Congruent dissolution — the mineral goes entirely into solution and leaves no solid residue. Halite, gypsum, calcite.

Incongruent dissolution — part goes into solution and part reprecipitates as a new solid. Every silicate hydrolysis reaction: feldspar leaves clay behind.

A mineral that dissolves congruently contributes nothing to the detrital sediment pile — it leaves as ions and comes back, somewhere else, as limestone or evaporite. That is why karst terrain has so little soil on it.

Where the by-products end up

The point of tracking solid versus dissolved is that they build different rocks in different places. Solid residues travel as particles and stack up as siliciclastics; dissolved loads travel as ions and precipitate as chemical sediments, often an ocean away. After Boggs, Table 1.3.

Weathering processProduct typeExampleUltimate depositional product
PhysicalParticulate residuesResistant silicate grains — quartz, feldspar — and rock fragments of every kindSandstones, conglomerates, mudrocks
Chemical —
hydrolysis
Soluble constituentsSilicic acid; K⁺, Na⁺, Mg²⁺, Ca²⁺Cherts, limestones
Secondary mineralsClay minerals — kaolinite, smectite, illiteMudrocks and shales
Chemical —
simple solution
Soluble constituentsSilicic acid; K⁺, Na⁺, Mg²⁺, Ca²⁺, HCO₃⁻, SO₄²⁻Limestones, evaporites, chert
Chemical —
oxidation
Secondary mineralsFerric oxides and oxyhydroxides; manganese oxidesMinor constituent in siliciclastic rocks — the red in red beds
Soluble constituentsSilicic acid; SO₄²⁻Chert, evaporites
Read the last column twice. Hydrolysis of a feldspar builds two rocks at once — the clay becomes a shale, and the dissolved silica and cations become a chert or a limestone somewhere else entirely. One weathering reaction, two depositional systems, two basins.

The twelve soil orders

USDA Soil Taxonomy. Soil is the weathering interface between rock and sediment — the order tells you the climate and how long it has been forming.

Magnitude & frequency

The organising question of process geomorphology: does the landscape get shaped by the everyday event, or the rare catastrophe?

Glacial & mass-wasting landforms

Domain D territory: what ice and gravity leave behind.

Sedimentary structures

The flow signatures — each one narrows down the energy and the environment.