← Back to the North America 20k map · This simulator uses European taxa, productivity values and a European Holocene history. The physics (fall speed, dispersal, basin size, preservation) is the same in North America; the species and numbers are not.

Palynological inference · forward model

Pollen Rain Simulator

A core sits at one point in a landscape. Every grain in it was produced somewhere else, carried an unknown distance, and survived an unknown length of time. These four models follow pollen from the physics of a single grain to a pollen diagram spanning eleven thousand years.

Scale 1 · micrometres

The grain: how fast it falls

Fall speed is the single grain property that drives every dispersal model further down this page. In still air a small grain falls at the speed where drag balances weight, which is Stokes' law. Size counts twice (weight grows with d³, drag with d), so a maize grain falls about ten times faster than a birch grain and seldom travels far.

One metre of still air, real time. Each column drops grains at its modelled fall speed. The highlighted column is the grain set by the sliders.

Grain

Sacci are hollow, honeycomb-walled bladders: lots of volume for very little mass.
Drops as the grain dries out after release.
1 = smooth sphere. Flattened, lobed or spiny grains have more drag.
–modelled fall speed
–measured value, where known
–Reynolds number (Stokes holds below ~1)
–same grain without sacci

Release

Scale 2 · millimetres

Capture at the ovule: pine and Ginkgo

Gymnosperm ovules catch pollen in a pollination drop, a bead of fluid secreted from the micropyle and later withdrawn. In Pinus the ovule faces down and the drop hangs below it; buoyant saccate grains float up through the drop into the ovule. Ginkgo grains have no sacci and do not float, so capture depends on the drop being drawn back in. Run the season, then change the ovule orientation or the pollen mix to see which combinations work.

Presets

Set-up

Slow withdrawal carries grains in with the fluid. Fast drying leaves grains stuck on the outside.

After capture

Pinus (Pinaceae)Ginkgo biloba
PollenBisaccate: a central body with two air sacs. Very abundant and easy to recognise in sediments.Monosulcate, boat-shaped, no sacci. Hard to separate from cycad and some other monosulcate grains, so fossil grains are usually lumped into broad form-genera such as Cycadopites.
OvulesTwo per ovuliferous scale in an upright seed cone; micropyles point down toward the cone axis.Usually two on a long stalk, exposed on short shoots, with no cone.
CaptureGrains on the drop float up through it into the micropyle (flotation), helped by withdrawal of the drop. Drops are secreted and withdrawn repeatedly over several days.Grains caught in the drop are drawn into the pollen chamber as the drop is resorbed.
Male gametophyteA pollen tube grows through the nucellus and delivers non-motile sperm.A branched, haustorial pollen tube anchors in the nucellus and feeds there for months, then releases two large, many-flagellate swimming sperm.
Pollination to fertilisationAbout 12–14 months: pollinated one spring, fertilised the next summer.About 4–5 months: pollinated in spring, fertilised in late summer.

The animation is a qualitative model: grains move by buoyancy relative to the drop fluid, and a grain left on the surface of a fast-shrinking drop can stick and be stranded. It reproduces the pattern Leslie (2010) found experimentally, where flotation in downward-facing ovules preferentially captures saccate pollen. The drop itself (sugars, proteins, its role in pollen germination) is not modelled.

Scale 3 · kilometres

The landscape: one season of pollen rain

A 10 × 10 km map at 62.5 m resolution. Each plant community releases pollen according to its species' productivity, the wind carries it according to its fall speed, and the basin integrates whatever lands. Paint the vegetation, move the core, and compare the count with the vegetation that produced it.

Click or drag on the map to move the coring site. Switch to Paint to change the vegetation.

Landscape

Sampling site

Air and canopy

0 = wind from all directions equally over the year.
Slows trunk-space wind in forests and filters pollen from afar at sites under trees.
Atmospheric parameters
Sutton's parameters, as used by Prentice and Sugita. Defaults: n = 0.25, Cz = 0.12 (neutral conditions).

Sediment and time

The count

–radius supplying half the land pollen
–radius supplying 80%
–of land pollen from beyond the 10 km map
–of deposited grains destroyed or unidentifiable

Vegetation versus the count

Tracking: all land pollen

Percentages of the total land pollen (trees, shrubs, herbs). Spores, aquatics and unidentifiable grains are given as % of land pollen plus that group, as on a standard diagram. Click a row to map where that taxon's grains came from.

Taxon Cover within 1 km Cover, whole map Pollen deposited Survives Pollen counted REVEALS-style estimate

Pollen source area

Cumulative share of the pollen reaching the core, by distance from the core. The far right step is pollen from beyond the map.

All land pollen

Scale 4 · millennia

The core: eleven thousand years

The same landscape model, run on a Holocene vegetation history: open steppe giving way to birch and pine, then hazel and mixed oak forest, alder on wet ground, beech late, and clearance for farming. Each level is counted under the site, wind and sediment settings above, and every level is as old as its depth. Solid silhouettes are the pollen count; the dashed line is the true land cover on the map.

Hollow outlines are ×10 exaggerations for minor taxa. Things to look for: Tilia forms far more of the forest than its pollen shows; Alnus and Pinus overshoot; cereal pollen hardly registers until fields reach the basin edge; and with the dry phase switched on, robust Lactuceae and fern spores spike with the indeterminable grains.

What the model does

Production

Each taxon releases pollen in proportion to its cover times a pollen productivity estimate (PPE) relative to Poaceae = 1. Tree values follow the European syntheses (Broström et al. 2008; Mazier et al. 2012). Values marked ° are illustrative guesses, mostly for spores, aquatics and Populus.

Dispersal

Sutton's plume as used by Prentice (1985) and Sugita (1993): the share still airborne at distance z is g(z) = exp(−b·zn/2) with b = 4vg / (n·u·√π·Cz). Deposition per m² is −g′(z)/2πz. Heavy grains with a high fall speed vg (Fagus, cereals) drop close to the source.

Wind and canopy

A steady prevailing wind weights upwind cells by 1 + a·cos(θ). Under a canopy, understorey plants release into slow trunk-space air, and a site inside closed forest receives less pollen from beyond ~150 m (Tauber 1965's trunk-space, canopy and rain-out components).

The basin

A lake collects over its whole surface, so no source is counted closer than the basin radius. Small basins are dominated by the vegetation at their edge; large lakes average a wide region (Jacobson & Bradshaw 1981). Cells inside the basin become open water.

Preservation

Each taxon has a corrosion susceptibility based on Havinga's oxidation series. The surviving share is exp(−3s⁶)·exp(−6E·s²·√t), where E is the environment's oxidation rate and t the age in thousands of years. Thin-walled grains (Populus, Ceratophyllum) are lost even in good conditions; robust ones (Lactuceae, fern spores) become over-represented as conditions worsen.

Grain physics

Fall speed v = m·g / (3π·μ·d·χ), with mass from a body of density ρ plus two sacci at 0.1 g/cm³, d the diameter of a sphere of equal volume, μ the viscosity of air and a Schiller–Naumann correction when the Reynolds number nears 1. Densities and shape factors for the presets are set so the model matches measured fall speeds.

Sources

Prentice 1985 (Quat. Res.); Sugita 1993 (Quat. Res.), 2007 (Holocene); Broström et al. 2008 (Veg. Hist. Archaeobot.); Mazier et al. 2012 (Rev. Palaeobot. Palynol.); Tauber 1965; Jacobson & Bradshaw 1981 (Quat. Res.); Havinga 1967, 1984; Doyle 1945; Leslie 2010 (New Phytol.); Friedman 1987 on the Ginkgo pollen tube; Hirase 1896 on Ginkgo swimming sperm.

Inference

The REVEALS-style column divides each count by PPE × the deposition a uniform landscape would deliver to this basin (Sugita 2007). It assumes even vegetation, no directional wind and perfect preservation, so its error shows what the core cannot tell you on its own.

🚧 Under construction