Varves

A varve is a pair of sediment layers deposited in a lake over one year. Counting them gives an age directly, with no reference to radioactive decay.

The Swedish geologist Gerard De Geer named them in 1912 and built the first varve chronology by correlating sequences across Sweden, before any radiometric method existed.

How annual layers form#

In a lake that freezes or experiences strong seasonal contrast, deposition alternates.

The couplet is one year. The distinction is visible under a microscope and measurable by grain size, colour, and chemical composition.

Three conditions are needed for varves to be preserved: strong seasonality, sufficient sediment supply, and an absence of burrowing organisms or currents that would mix the layers. Deep, cold, oxygen-poor bottom water preserves them best, which is why the good sequences come from particular lakes rather than all of them.

The main records#

Lake Suigetsu#

The most important record, because it does something no other archive does: it dates radiocarbon samples against counted years.

The lake has a small catchment, no major inflowing river to disturb the bottom, and anoxic deep water that excludes burrowing organisms. Layers have accumulated undisturbed for tens of thousands of years.

The sediment also contains terrestrial plant fragments — leaves, twigs, seeds — that fell into the lake in a known year. These can be radiocarbon dated, and the result compared against the counted layer position.

Bronk Ramsey and colleagues (2012) published this as a continuous calibration record from 11,200 to 52,800 years. It is now a primary component of the international IntCal20 calibration curve.

The terrestrial material matters. Aquatic plants incorporate dissolved carbonate from old rock and read too old — the reservoir effect described in Radiocarbon Dating. Leaves from land plants take carbon directly from the atmosphere.

Independent verification#

Varve counts are checked against markers that arrive from outside the lake.

Common objections#

"Multiple layers can form in one year"

Correct, and this is the standard objection. A single storm or flood can deposit a distinguishable layer, producing more couplets than years.

The phenomenon is real and is why varve chronologies are not built by counting one core.

Storm layers are distinguishable. They are graded, often coarser throughout, and lack the systematic light-dark couplet structure produced by the seasonal cycle. Micro-analysis of grain size and geochemistry separates them.

Cores are also counted at multiple locations in a lake, and event layers vary between sites while the annual signal does not.

The check that settles it is external. If varve counts systematically overcounted, dated tephra layers would appear at the wrong depth. They do not.

"Mount St. Helens produced thousands of laminae in hours"

Accurate. The 1980 eruption deposited finely laminated sediment rapidly, and Steve Austin has cited this against varve chronologies.

Rapid lamination is not in dispute. Laminae are not varves; a varve is specifically a seasonal couplet, and the distinction is defined by structure rather than by assumption.

The Mount St. Helens laminae are volcanic, graded, and lack the paired light-summer and dark-winter composition. They also contain no seasonal biological signal — no diatom blooms, no annual pollen influx — which the lake records do.

The pollen check is direct: Suigetsu's layers contain the seasonal pollen sequence of the surrounding vegetation, repeating annually.

"Varve counting is calibrated against radiometric dating, so it isn't independent"

The relationship runs the other way. Varves are counted physically and are used to calibrate radiocarbon, not the reverse.

De Geer built the Swedish chronology in 1912, decades before radiocarbon existed. The counting method does not use decay at all.

Where the two are compared, the comparison is a test rather than a calibration in the objectionable sense: the counted position is known first, and the radiocarbon measurement either matches or does not.

"The Green River Formation formed rapidly, not over millions of years"

The Green River varve count implies several million years for the formation, and young-earth authors have argued the laminae are not annual.

The debate here is genuine, and some researchers have argued that not every Green River couplet is annual, particularly in some intervals.

The constraint is what the formation contains. It preserves catfish and other fossils buried in place, in-place burrows at certain horizons, and evaporite minerals requiring the lake to have partly dried. These indicate a lake existing over a long period rather than a single depositional event.

The wider argument on varves does not rest on Green River, since Suigetsu and the Swedish chronologies are directly checkable against tephra and radiocarbon.

What the evidence shows#

Varves are counted, not calculated, and the method predates radiometric dating by several decades.

Lake Suigetsu provides a continuous annually resolved record to about 52,800 years, containing terrestrial plant material that allows radiocarbon to be checked against counted years — which is why it underpins the international calibration curve.

Counts are verified against dated volcanic ash layers, against radiocarbon, and against events shared with ice cores and tree rings. Each of those is an independent opportunity for the counts to be shown wrong.

Related counted records are in Dendrochronology and Ice Cores, and the agreement between methods is examined in Cross-Calibration.