How cold was it, 20,000 years ago?

Cette publication est également disponible en : Français

During the Last Glacial Maximum, about 20,000 years ago, it was cold. But how cold? Estimates of surface air temperature vary between 1.8 and 8°C colder than today. This remains imprecise.  Christoph Schmidt and Georgina King are working on a new SNSF project to develop a method for estimating past temperatures that can be applied globally – at any latitude and any altitude.

A “global” method – from the moon to the foot of Mont Blanc

This method is based on almost ubiquitous material: quartz and feldspar. The idea was first introduced in the 1960s for terrestrial applications, and then in the 1970s for lunar samples from the Apollo 12 mission. It has only recently been revived and developed, in particular by a un groupe de la FGSE : Frédéric Herman and Rabiul Biswas (now a professor in India) put a lot of effort and time into it. “They have developed it to a point where we can now build on it and try to apply it on a larger scale” points out Ch. Schmidt. “There are still a range of open questions and problem we have to face, but we have ideas on how to overcome them.”  This method could be applied in all regions of the world where these minerals are present, and why not, on other planets!

The trace of paleo-temperatures trapped in minerals

How can we reconstruct past temperatures? Quartz and feldspar trap electrons generated by environmental radiation. The team is exploiting the fact that the charge trapped in these minerals – generated by irradiation – depends on the ambient temperature. In the luminescence machine (visible on the video), the sample is exposed to heat or light, which triggers the release of the luminescence signal. A highly sensitive device, called a photomultiplier, is then able to record individual photons released from the mineral – a level of light well beyond the detection limit of the human eye.

We try to measure the relative level of signals in response to two competing processes: radiation and temperature. By doing so, we try to find the thermal history that most likely can explain the signal pattern that we observe experimentally from our sample.

Christoph Schmidt
One of the main challenges is to accurately characterise the behaviour of the quartz and feldspar samples. In order to extrapolate laboratory observations to a larger time scale, it is important to be as accurate as possible: small inaccuracies will seriously affect an extrapolation into the distant past.

We aim to reconstruct absolute temperature in different ranges of time, from 30 to 40,000 years to shorter time scale. But this method will have other possible applications: for example to estimate the temperature of a rock during a volcanic eruption or any type of thermal hazards.

Christoph Schmidt

20,000 years ago, from the equator to the far north

Reconstructing the absolute temperature time series from the Last Glacial Maximum to the present day is one thing. The team also aims to cover as wide a latitudinal gradient as possible. “We start in the North in Norway, it is the northern most piece of bedrock that was not covered by a glacier during the last ice age (as we want to reconstruct air temperature). At the southernmost point, very close to the equator, the Ruwenzori Mountains in Uganda are the only non-volcanic mountain massive in central Africa. So there is not much of a choice! Non-volcanic origin is important, because volcanic samples show very special luminescence properties that we try to avoid.”

Ruwenzori Mountains, Uganda (© Martin Mwaura |
Mont-Blanc, France (© Christoph Schmidt)

Two massifs, one in the tropics and the other in a temperate region will allow estimating how temperatures dropped from 1000 to 4000 m during the Last Glacial Maximum (26,500 to 19,000 years ago). This adiabatic lapse rate is crucial to model the atmospheric climate. The adiabatic lapse rate is the variation of air temperature with altitude, related to atmospheric pressure alone.

Reconstructing past temperatures to better understand the future?

Knowledge of past surface air temperatures as a function of latitude and altitude is important for understanding the Earth’s climate oscillations and atmospheric circulation. In the context of global warming, it is a key element in predicting future scenarios. In particular, these temperature data serve as crucial input parameters for evaluating climate models and determining climate sensitivity. “This information can be fed into climate models that will tell us about our future on this planet, that is what the temperature will be in the next 50 or 100 years.”

There weren’t people with thermometers 20,000 years ago, so we try to extract this information from the rocks, to give that to the climate scientists.

Georgina King


  • Biswas, R.H., Herman, F., King, G.E., Lehmann, B., Singhvi, A.K., 2020. Surface paleothermometry using low-temperature thermoluminescence of feldspar. Climate of the Past 16, 2075-2093.

Leave a Reply

Your email address will not be published. Required fields are marked *